Vous êtes sur la page 1sur 337

Stancetaking in Discourse

Pragmatics & Beyond New Series (P&BNS)
Pragmatics & Beyond New Series is a continuation of Pragmatics & Beyond and
its Companion Series. The New Series offers a selection of high quality work
covering the full richness of Pragmatics as an interdisciplinary field, within
language sciences.

Andreas H. Jucker
University of Zurich, English Department
Plattenstrasse 47, CH-8032 Zurich, Switzerland
e-mail: ahjucker@es.uzh.ch

Associate Editors
Jacob L. Mey Herman Parret Jef Verschueren
University of Southern Belgian National Science Belgian National Science
Denmark Foundation, Universities of Foundation,
Louvain and Antwerp University of Antwerp

Editorial Board
Shoshana Blum-Kulka Susan C. Herring Emanuel A. Schegloff
Hebrew University of Indiana University University of California at Los
Jerusalem Angeles
Masako K. Hiraga
Jean Caron St.Paul’s (Rikkyo) University Deborah Schiffrin
Université de Poitiers Georgetown University
David Holdcroft
Robyn Carston University of Leeds Paul Osamu Takahara
University College London Kobe City University of
Sachiko Ide
Foreign Studies
Bruce Fraser Japan Women’s University
Boston University
Catherine Kerbrat- Sandra A. Thompson
University of California at
Thorstein Fretheim Orecchioni
Santa Barbara
University of Trondheim University of Lyon 2
John C. Heritage Claudia de Lemos Teun A. van Dijk
Pompeu Fabra, Barcelona
University of California at Los University of Campinas, Brazil
Marina Sbisà Richard J. Watts
University of Berne
University of Trieste

Volume 164
Stancetaking in Discourse. Subjectivity, evaluation, interaction
Edited by Robert Englebretson

Stancetaking in Discourse
Subjectivity, evaluation, interaction

Edited by

Robert Englebretson
Rice University

John Benjamins Publishing Company
Amsterdam / Philadelphia

TM The paper used in this publication meets the minimum requirements of

American National Standard for Information Sciences – Permanence of
Paper for Printed Library Materials, ansi z39.48-1984.

Library of Congress Cataloging-in-Publication Data

Stancetaking in discourse : subjectivity, evaluation, interaction / edited by Robert
p. cm. -- (Pragmatics & beyond new series, issn 0922-842X ; v. 164)
Includes bibliographical references and index.
1. Discourse analysis. 2. Conversation analysis. I. Englebretson, Robert.
P302.S65    2007
401'.41--dc22 2007026354
isbn 978 90 272 5408 5 (Hb; alk. paper)

© 2007 – John Benjamins B.V.
No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any
other means, without written permission from the publisher.
John Benjamins Publishing Co. · P.O. Box 36224 · 1020 me Amsterdam · The Netherlands
John Benjamins North America · P.O. Box 27519 · Philadelphia pa 19118-0519 · usa

Table of contents

Acknowledgements vii

Stancetaking in discourse: An introduction 1
Robert Englebretson

Using a corpus to investigate stance quantitatively and qualitatively 27
Susan Hunston

Linking identity and dialect through stancetaking 49
Barbara Johnstone

Grammatical resources for social purposes: Some aspects 
of stancetaking in colloquial Indonesian conversation 69
Robert Englebretson

Subjective and intersubjective uses of generalizations 
in English conversations  111
Joanne Scheibman

The stance triangle 139
John W. Du Bois

The role of I guess in conversational stancetaking 183
Elise Kärkkäinen

Stance markers in spoken Finnish: Minun mielestä 
and minusta in assessments 221
Mirka Rauniomaa

Stancetaking as an interactional activity: Challenging the prior speaker 253
Tiina Keisanen

Positioning and alignment as activities of stancetaking in news interviews 283
Pentti Haddington

Name index 319
Subject index 321


Finland. I would especially like to thank both Mara Henderson and Michelle Morri- son for their painstaking and diligent editorial assistance at various stages of the manuscript. held March 31 through April 3. 2004. Geoffrey Raymond. which have improved the overall quality of the manuscript. Susan Hunston. Elise Kärkkäinen. Pentti Had- dington. I gratefully acknowl- edge the Rice University Office of the President (Presidential Research Award 2006–2007). Robert Englebretson. and especially for the lively and en- gaging roundtable discussion that concluded the conference. and Robin Shoaps. . Joanne Scheibman. at Rice University. I would like to thank each of the presenters and attendees for mak- ing this into a rich and fruitful conference. Presenters (listed in alphabetical order) were John W. For financial support in the completion of this project. Du Bois. I would also like to thank the graduate students and faculty in the Rice Linguistics Department for provid- ing logistical support and a warm welcome (Houston weather not withstanding!) to our guests. We gratefully acknowledge the following organizations for funding the sympo- sium: Rice University Department of Linguistics. Tiina Keisanen. the Stance Research Group at the University of Oulu. Acknowledgements The papers in this volume have their genesis in work presented at the 10th Bien- nial Rice Linguistics Symposium. Mirka Rauniomaa. and the Rice University Humanities Research Center (formerly the Center for the Study of Cultures). We would also like to express our gratitude to the two anonymous reviewers for their input and comments. Barbara Johnstone.


offering new insights into the sociocultural. interactional. scholarship in linguis- tics and related disciplines has witnessed a notable upsurge of interest in stance. discourse-functional linguistics. Shoaps and Kockelman 2002). The present volume joins this burgeoning field. Mushin 2001. based on the contexts within which it occurs. sociocul- tural linguistics. and interactional linguistics). Kärkkäinen 2003. as it is used by actual speakers or writers to act and interact in the real world. Stancetaking in discourse An introduction Robert Englebretson Rice University 1. stance is by no means a monolithic concept. Jaffe 2004. Macken- Horarik and Martin 2003). education. represents an ongoing trend toward under- standing the full social and pragmatic nature of language.g. Definitions and conceptions of stance are as broad and varied as the individual backgrounds and interests of the researchers themselves. and as I will argue throughout this chapter. and numerous journal articles have dealt with this topic either directly or indirectly. Introduction During the initial few years of the twenty-first century. Several published monographs explicitly reference stance in their titles (Gardner 2001. Wu 2004). systemic-func- tional linguistics. . social psychology. As this heterogeneous range of research implies. But what is notewor- thy about the focus on stance from all of these different perspectives. an intersection of subdisciplines within linguistics (among e. panels and symposia have been devoted to various perspectives on stance (Englebretson 2004. is that it marks an orientation toward conceiving of language in terms of the functions for which it is used. This broad array of research represents a convergence. corpus linguistics.. and sociology. cognitive linguistics. however this term is defined. special issues of journals have focused on this topic (Berman 2005. Research on stance. and it highlights a set of overlap- ping interests with closely-allied fields such as anthropology. Hunston and Thompson 2000.

and to begin to develop a model which recognizes the heterogeneous and multifaceted nature of stancetaking. to seek to understand how stance research- ers conceive of stance in their own work. researchers who appear to be working on similar issues may in fact use different terminology to cover what appears to be the same types of phenomena. and to what extent these various approaches may converge with each other overall. burgeoning state of stance research. and this is the approach I have chosen to adopt here. The lack of a consistent definition of stance poses a conundrum for anyone wishing to summarize the state of the field in an introduction such as the present chapter. The overall purpose of this volume is to explore how it is that speakers (and writers) actively engage in taking stances in natural dis- course. Prescrib- ing a narrow definition would needlessly fragment and limit this interdisciplinary field of research at a time when it has only just begun to emerge and take shape. Secondly. more inclusive approach would be to adopt a broader. One approach would be to present a clearly operationalized definition of stance. ethnographi- cally-informed conception of stance. both in this introductory chapter and in the volume as a whole. A reader of this volume who wishes a clear-cut. and would likely end up excluding much of what has been published under the rubric of stance research. Thus two authors may use stance to encompass two seem- ingly disparate types of phenomena. Such an approach would border on academic imperialism. operationalized definition of stance will no doubt find it challenging to follow the disparate threads of ideas as they coalesce throughout the various approaches represented here. definitions and understandings are not necessarily shared as common ground from one scholar to the next. However.1 There are at least two ways in which the use of this term can be problematic. the term stance refers to in the first place. while individual researchers do tend to operationalize stance within their own work. including many of the papers in the present volume. A second. while other language researchers may prefer the term evaluation and may wish to eschew the label stance altogether. exactly. A critical observer of the recent preponderance of work on stance may won- der what. I believe this is the most realistic and fruitful way to approach the current. For ex- ample. This inclu- sive approach has the advantage of promoting dialogue among stance researchers and encouraging interplay among ideas. Robert Englebretson and performative nature of language by addressing stancetaking from a range of perspectives in natural discourse. First. the converse situation is also true. Such a writer is faced with essentially two possibilities. and consequent- ly to exclude from the purview of stance research anything which does not meet this definition. but always intimately bound up with the pragmatic and social aspects . one scholar may use stance to encompass what other scholars refer to as subjectivity. The understanding of stance that emerges is indeed heterogeneous and variegated.

investigate. Section 4 briefly introduces each of the contributions to this volume. This approach takes seriously Hunston’s (this volume) call for both quantitative and qualitative methodologies in the investigation of stance. stancetaking is a gerund. The remainder of this introductory chapter is organized as follows. and write about. set- ting aside for now the case of academic language-related stance research to which we shall return in Section 3.e. which we can observe. In other words. Stancetaking in Discourse. The present section takes up the first proposition just mentioned – the question of the conceptual entity known as stance – by investigating the meaning of the word stance in order to more fully come to terms with stance as an object of research. this section approaches the topic from an ethnographic perspective in everyday language. while highlighting the situated. it claims that stancetaking happens in discourse – in language in its natural habi- tat – and is thus best studied within this context. I survey recent work relevant to stance. Secondly. Section 2 steps back from academically-oriented approaches to stance. and interactional character of stancetaking. entails several proposi- tions. it suggests that stances are something that people actively engage in (i. Stance in discourse The title of the present volume. when speakers and writers use the term stance in their naturally-occurring speaking and writing. and thereby seek to contextualize the present vol- ume within the current intellectual climate of stance-related scholarship. pragmatic. to discuss the semantics and pragmatics of the term stance from a metalinguistic perspective. based on an object-incorporation of stance and the active verb take). Section 3 returns to the question of how stance has been approached within language research. what do they mean? This section seeks an answer to this question by presenting a brief quantitative and qualitative overview of the term stance as observed in two corpora of present-day English. 2. Finally. . Introduction  of human conduct. as observed from how people actually use it in their speaking and writ- ing. and we hope the ideas and claims in the papers will spark further inquiry and new directions in investigating and refining this field. In order to come to an un- derstanding of what stance may be understood to be. I outline trends in the meaning of stance. Based on two corpora of present-day English. We offer the present volume as an exploratory work in stance research. and it also serves to prefigure the kinds of analyses undertaken throughout the rest of this volume.. research. I suggest that the diverse nature of how stance is conceptualized in everyday language use impacts and informs how language researchers have conceptual- ized and appropriated it in academic contexts as well. It presupposes first that there is a conceptual entity known as stance.

2 All SBCSAE examples in this section are excerpted from the corpus directly. Since there is no agreed-upon definition of stance from an academic perspective. all concordance counts and collocations were obtained by using VIEW (Variation in English Words and Phrases). are clearly signified in the quantitative and qualitative analyses of stance in language use. This is firmly in keeping with a usage-based perspective on language (cf. scope of genre. primarily (but not exclusively) on face-to-face conversation.byu. and contains time-aligned audio for each transcript. 2003. which takes as its starting point the idea that language use shapes language form and meaning. the question naturally arises as to why it is that stance researchers have adopted this term to refer to such a broad and variegated set of phenomena. and stance as a word in English more generally. (1993). in what situations. the BNC online search tool provided by Mark Davies at Brigham Young University: http://view.4 the BNC is over 400 times larger. The SBCSAE is transcribed using the intonation unit-based discourse transcription system proposed by Du Bois et al.edu. As I will demonstrate in the present section. For the BNC data presented in this section.3 In addition to representing the English of the United States versus the English of the United Kingdom. and interactional research. The data for this section come from two corpora: the four volumes of the San- ta Barbara Corpus of Spoken American English (SBCSAE) (Du Bois et al. the everyday conceptions of stance serve as a window into how this term has been appropriated by linguistic. Du Bois and Englebretson 2004. these two corpora are vastly different from each other in terms of size. The rather obvious fact that stance researchers are also language users suggests that questions of how often. and transcription detail. The SBCSAE contains approximately 249.5 The SBCSAE focuses entirely on spoken language. the spoken component of the BNC uses an undifferenti- ated orthographic transcription system with minimal attention to interactional or intonational details. Robert Englebretson A metalinguistic analysis of stance in everyday language is crucial in com- ing to terms with the kinds of activities implicated in stancetaking. Barlow and Kemmer 2000. In other words. there is a reflexive relationship between stance as a concept in language research. the heterogeneity of stance as an ob- ject of research. social. and the audio files are neither included nor time-aligned . and in what kinds of interactional and collocational contexts stance is used are all highly relevant in understanding how stance is conceptualized more specifically in re- search contexts as well. inter alia). and the varying uses of this term by scholars. 2000. Bybee 2006. World Edition (BNC) (BNC 2001). Du Bois et al. 2005) and the British Na- tional Corpus. comprising approxi- mately 100 million words. I would suggest that these conceptions of stance are based directly on the wide range of meanings and activi- ties associated with this lexeme in everyday language use. the BNC consists of a written component (90% of the corpus) and a spoken component (10% of the corpus).000 words at present.

it is a fairly infrequent word in both corpora. We observe that tokens of stance are quite rare in both speech and writing. the SBCSAE is especially useful for in- teractional.000 words of talk (and making further quantitative analysis of this lexeme impossible in this corpus). I will begin this discussion in Section 2. It occurs only three times in the SBCSAE.1 Quantitative findings This section addresses the quantitative distribution of stance in both corpora. although more frequent than in the SBCSAE.e. in Section 2.7 The lemma stance occurs a total of 1. the BNC is tagged for parts-of-speech and sentences. Because of its close focus on interactional detail and the availability of the audio files. as well as to our general conception of stance in language research. in order to analyze the semantic prosodies (cf. tags. giving a rate of occurrence of 1 token per every 83.3 discusses the adjectives in the BNC that most frequently collocate with stance. ethnographic. 2. the BNC is an especially useful resource for large-scale quantitative computational and text work (as long as the researcher is not seeking interactional detail. Table 1 shows two important quantitative findings about stance. Hunston this volume) that are most typically associated with this term.2. and in sub-genres within the BNC. while the SBCSAE is untagged. Introduction  with the transcripts. Table 1 below. presents the overall occur- rence of tokens of the lemma stance (i.835 times in the combined written and spoken compo- nents of the BNC. Given the somewhat comple- mentary natures of these corpora.6 Finally. Section 2. each corpus is more suitable for some types of re- search purposes than it is for others. with a rate of occurrence of roughly 1 token per every 53. acoustic-phonetic. the two together provide an excellent vehicle for qualitative and quantitative analyses of stance. if the research does not require data from written English. Obviously. of course. the singular form stance and the plural form stances combined) in both corpora. and can thus be used with parsers and other automated tools. prosodic infor- mation. First. which is the topic of Section 3.4 summarizes the findings of this investigation and outlines what it has to contribute to our un- derstanding of the term stance specifically. Then. Section 2. or audio files in the spoken component). but occur more frequently in writing than in speech. We also find that stance shows skewed distribution across registers of the BNC. I will move to a qualitative investigation of the occurrences of stance in the SBCSAE.198 words of text – also relatively rare. Due to its massive size. and genre-based orientation. and conversation-analytic research (and for traditional corpus work too.1 by presenting a broad quantitative overview of the rate of occurrence of stance in both corpora. and if large-scale frequency counts are not a focus of concern).. .

and is perceivable.070 words in the spoken component.681 words of text. (5) stance is consequential – i. Furthermore. Haviland 1989.000 1 per 83. the two text-types with the highest rate of occurrence of stance are scripted speeches and social science lectures.793 87. The present section provides a qualitative analysis of each of these tokens within their interactional contexts-of-use. taking a stance leads to real consequences for the persons or institutions in- . Du Bois this volume). (2) stance is public. I shall return to the question of what exactly these distributional factors tell us about stance in Section 3.e. but it is necessary to first address this term from a qualitative and interactional perspective. interpretable. and with respect to other stances (Scheibman. at a greater-than five to one ratio. and education.. 2.8 Thus stance occurs over five times more frequently in writing than it does in speaking. this volume. the SBCSAE contains only three to- kens of stance. and stance is social morality. and non-academic writings on law.070 words BNC (Written Component) 1. evoking aspects of the broader sociocultural framework or physical contexts in which it occurs. (4) stance is indexical (cf.835 97.4 below. newspaper writing on science. and available for inspection by others (cf. These analyses bring to light five key conceptual principles of stance as follows: (1) stancetaking occurs on three (often overlapping) levels – stance is physical action.  Tokens of stance in the SBCSAE and BNC Corpus Tokens Total words Rate of occurrence SBCSAE 3 249. The three text types with the highest rate of occurrence of stance in the BNC are university essays.198 words Secondly. the rate of occurrence of stance in the written component is 1 token per every 48. politics.284. this table shows that stance occurs far more frequently in writing than it does in speech. when stance does occur in the BNC.2 Qualitative analyses As shown in Table 1 in the previous section. When the BNC is broken down by written versus spoken texts. Interestingly. there is not a single token of stance in the entire 4-million word BNC sample of conversation.311 1 per 53.681 words BNC (Total) 1.364 1 per 48. Silverstein 1976). stance is personal attitude/be- lief/evaluation.619. (3) stance is interactional in nature – it is collaboratively constructed among partici- pants. as contrasted with roughly 1 token per every 246.000 words BNC (Spoken Component) 42 10.947 1 per 246. makes a similar point about the “relational” nature of stance in discourse).334. Robert Englebretson Table 1. it is broadly correlated with formal registers. In the spoken component of the BNC.

e. I suggest that a qualitative understanding of what stance means in talk-in-interaction serves as a window into the academic concept of stance – both in the papers in this volume. Pressing against you with that left hand (line 8). As we examine stance from a qualitative perspective in everyday talk. In the remainder of this section.. All of these descriptions refer to the opponent’s stance.. (H) While he’s down in a deep jigo. First. In Section 2.. for the relationship between stancetaking and responsibility). the use of stance here is clearly a physical act. The textual and physical contexts of stance in this excerpt illustrate each of the five principles outlined above. 9 . This excerpt begins as Nick is explaining the kinds of situations in which Hane-Makikomi would be useful. 7 .. In this speech event (SBC057. this extreme defen- sive stance refers literally to a physical way in which the opponent is positioning his body. 6 . the reflexive relationship between these uses and the conceptualization of this term in academic stance research becomes apparent. the physical stance of this hypothetical opponent is public. 2 . entitled Throw Me) the sensei. Example (1) comes from a recording of a judo class.. is teaching and demonstrating Hane-Makikomi (springing wraparound throw).. 8 . with his left foot forward.362-966. Pressing against you with that left hand.. As described in Nick’s subsequent explanation. with his left foot forward (line 7). 3 .. Du Bois.. who’s fighting you in that. and in language-related research on stance more generally. I will present quantitative collocational evidence from the BNC which support these qualitative findings. this volume.. extre=me (Hx) defensive stance. 4 → .9 (1) (Throw Me SBC057: 948. I will demonstrate how each of the three SBCSAE excerpts illustrates these five principles.. Keeping you out. Overall. This excerpt presents a hypothetical example – a description of when Hane-Makikomi would be a desirable throw. in this case. But I like the Hane-Makikomi. Nick. The interactants are you (the judo practitioner who will be using Hane-Makikomi) and he (the opponent who is fighting you in that extreme defensive stance). Introduction  volved (cf. extreme defen- sive stance. .3. Secondly. (H) when it’s used on a player (Hx). down in a deep jigo (line 5) (jigo is a judo term referring to ‘self-defen- sive posture’).. We now turn to an analysis of each of the three stance tokens in the SBCSAE. body posture.. i. 5 . The term stance appears at the arrow in line 4 of this example.593) 1 NICK: .... Preferably...

To roughly paraphrase: if your opponent takes a physical stance like this. On one level. part of assessing one’s opponent during a judo match. Similarly. and it is the opponent’s extreme defensive stance that will lead to your subsequent use of Hane-Makikomi. Nick observes and inspects the hypothetical opponent’s stance and describes it in detail. her application interview to attend a private Catholic college in the northeast – a college to which she was not admitted. Gail is describing an event from the past. stance is indexical. Thirdly. then you respond with Hane- Makikomi. Stephanie (Steph) is a senior in high school and is presently applying to colleges.730-303.. You must be there in order for the opponent to be press- ing against you with that left hand (line 8). He then interprets the stance as Keeping you out (line 9). Patty is Stephanie’s mother.e. This stance of a deep jigo has specific mean- ing within the practice of judo. it indexes the specific knowledge systems of judo. The relevant background information for understanding these excerpts is as follows. i. seeing the opponent in this physical stance calls up a host of understandings about the opponents intentions. 2 they asked me some really weird [questions] though. and Maureen (Maur) is Gail’s mother. In this case. 4 I know[2=2]. as well as the sociocultural background in which this martial art is embedded. stance has consequences. In example (2). Robert Englebretson and is indeed being inspected and interpreted.. Here. This kind of overt inspection of stance is. (2) (Hold my Breath SBC035: 285. after all. who is currently a college student. the physical stance points to (indexes) something beyond the textual and physical context: namely. Fourthly. While example (1) illustrates physical stance. These remaining two tokens of stance in the SBCSAE both come from the same speech event (SBC035 entitled Hold my Breath). In this case.10 The conversation is taking place at a family get-together. Finally. Gail is Stephanie’s older cousin.681) 1 GAIL: . 6 GAIL: [3Like3] they -- . the following two examples illustrate personal and moral stance. this excerpt represents the interactionality inherent in taking a stance. 3 MAUR: [I know]. And this relationship between stance and consequence is indeed the upshot of Nick’s instruction in this excerpt. the opponent’s probable next actions. You know what. it is your previous offensive moves that have led to the opponent’s current extreme defensive stance. 5 PATTY: [2They2] [3did3]. the opponent’s stance has the direct and physical consequence of elicit- ing Hane-Makikomi. the physical details of this particular stance require two individuals: you and he (your opponent). and your best move.

-- 13 PATTY: that’s important].. Because of the similarities between this and the next example.. 11 STEPH: (H) Well no. 9 da-da-da-da-da [I mean VOX>. 6 . 11 PATTY: [So what did they ask you]. Unlike in example (1). the stance token occurs in line 12 at the arrow. 10 PATTY: [There’s nothing wrong with that. 17 I was like . 8 → (H) moral stance. 8 you know my li=fe. In this excerpt. and evaluation of a controversial (line 18) and very personal (as cited later in the transcript – not shown here) moral issue. 15 but m. which picks up 25 seconds later in the speech event from the end of example (2). Introduction  7 I expected to go in there= and talk about.. 7 (H) <VOX a ni=ce value= community uh. 14 STEPH: But]. (3) (Hold my Breath SBC035: 328. [2(H)2] 14 PATTY: [2Boy that2] was a dumb question..521) 1 STEPH: (H) Lee applied to Williams. 15 [3that’s one -- 16 GAIL: [3Boom. . the= information they’re sending out is. 12 but what they’re do. but rather the speaker’s beliefs about.on abortion? 13 . But lately.-- 4 . Prompted by Gail’s story of not being admitted to a college after an interview. Stephanie recounts the following experience of her friend Lee who was waitlisted under similar circumstances at a different college... (H) liberal whatever. 18 PATTY: That’s a contro3]versial question. this stance does not concern physical body posture. (TSCH)3]... 10 and what I expected from [the=m and. (H) And Williams used to be real=ly lib. 5 . They asked me what my stance was on abortion. 2 And she got waitlisted. 9 and what I wanted out of a schoo=l. what my stance was abou.955-365. I will discuss them both together after presenting example (3). attitudes toward. 3 .Mom. 12 → GAIL: (H)] They asked me .

They had. 28 . The criticism... 18 STEPH: (H) .. But the most important evidence that Stephanie is criticizing the college’s stance is displayed in the uptake by her mother. and evaluation displayed in this sequence illustrates that stance is indeed observable and avail- able for interpretation.g. and values. In example (2).. 25 (H) they= . this use of stance refers to beliefs. they had <VOX the sit-ins VOX>. Stephanie’s criticism of the college’s stance is suggested by the mocking-sounding prosody in lines 7–9. for purposes of making a decision on her admission to the college. and likewise as a social value – e.. the stance here in example (3) is attrib- uted to a general institution (Williams). evaluated. Taken together. the observable. unlike in the previous example.. Patty herself evaluates the college’s stance... 21 . now they’re going to conservative. 22 (H) real liberal. slowly. Similarly.. 20 . In example (2). First.10 Robert Englebretson 16 what I’m [saying is]. very . both excerpts illustrate the interactional nature of stance. at the same time that her mother is defending it. years ago. examples (2)–(3) illustrate the five principles of stancetaking described earlier. Subsequently. Secondly. and interpreted by the interviewer(s). moral stance. attitudes. 23 . and public nature of stance is highlighted in both cases. took her answers . her stance is presumably being scrutinized. The stance token in this excerpt occurs in line 8 at the arrow. the upshot of Gail’s story is the juxtaposition of her stance on abortion with the (implied) stance of the private Catholic college. stance in these examples is depicted as a personal belief or attitude. 19 . that’s important (line 13). defense. Patty likewise interprets Stephanie’s comments as criticism. However. in example (3). but she didn’t have the political views. As in the previous example.. and in turn seeks to defend the college’s stance. Thirdly. By the very nature of an interview. rather than to a specific person (Gail) as in example (2). moral stance in line 8 of example (3). 26 and they waitlisted her. 24 and when she was interviewed for it. In example . interpretable. There’s nothing wrong with that (line 10). everything you know. and by the overall negative tone of her claim that the college has shifted from liberal to conservative. 27 (H) she had the grades.. Stephanie is observing and criticizing the conservative moral stance of the col- lege. 17 PATTY: [I’m listening].. Gail’s stance was specifically elicited in the con- text of an interview.

Fourthly. Fi- nally. 2. occur one word to the left of) stance. Introduction 11 (3). stance is public and interpretable.. and stance is consequential. a nice value community (line 7) and an unspecified list of other conservative views insinuated by Stephanie’s da-da-da-da-da (line 9). Stephanie’s point is that the institution’s stance was apparently at odds with the stance of her friend Lee. and supplements the quantitative and qualitative findings presented in the previous two sections. In the following sec- tion. stance is interactional. which is never overtly stated here. the college’s “moral stance” likewise indexes a range of other stances. both excerpts demonstrate the indexical nature of stance. Stephanie suggests that the lack of align- ment between Lee’s views with the stance of the college may have led to Lee being waitlisted. In both cases. she had the grades. An analysis of the kinds of adjectives that typically occur with stance leads to a richer understanding of the meaning of this term. but she didn’t have the political views (lines 27–28). it is the dialectic of conflict between the stances of the individual and the stances of the institution that lead to the overall realization of the stance more generally – just as in the case of the physical stance discussed in example (1). where the extreme defensive stance is described in terms of the positioning of two bodies with respect to one another. In example (2). in both excerpts. a so-called “anti-abortion” or “pro-life” stance would stereotypically index Gail as likely being a person who supports more conserva- tive political views and agendas. The qualitative analyses of these three excerpts from the SBCSAE have illus- trated five general principles of stancetaking in discourse. we will return to the BNC and examine how collocational evidence supports these qualitative findings as well. This sec- tion provides a brief overview of the adjectives in the BNC that are immediate collocates with (i. The following table shows the 20 adjectives in the BNC that collocate with stance more than five times in the corpus. Gail’s stance on abortion. . stance is consequential. In example (3). The implications of example (2) are that Gail’s stance on abortion may have led her to not gain admission to this particular school. stance is indexical. is seen as indexing wider social morals and values.3 Collocational evidence The previous section has illustrated five general themes observed in stancetaking in discourse: stance is physical/personal/moral. while in example (3).e. while a so-called “pro-abortion” or “pro-choice” stance would stereotypically index and imply that she holds a wider array of more liberal political views.

in which about 60 per cent of the body weight is on…” (GVF). “You take up a left stance and line yourself up so that your left foot is in front of the opponent…” (A0M).12 Robert Englebretson Table 2. and moral dimensions of stance. strength. and cultured style…” (B2H).  Adjectives in the BNC collocating with stance (frequency > 5) Adjective Number of tokens political 37  aggressive 20  moral 16  upright 15  tough 13  critical 12  neutral 10  positive 10  forward 10  public 10  negative 9  basic 9  particular 8  left 8  ideological 7  conservative 7  anti-abortion 6  previous 6  different 6  right 6 This table shows that there are only 20 adjectives in the BNC that serve as imme- diate collocates of stance with a frequency greater-than five tokens apiece. “For the evolving herbivores the advantage of an upright stance was soon complemented by the ability to rear up…” (C9A). and these categories often overlap. and I will now turn to a summary of these findings with re- spect to the qualitative observations from the previous section. evaluative. First. One token of right is evaluative in nature and reflects moral stance: “…exotic religious sect. and right (with the exception of one token of right meaning ‘morally correct’). collocates of stance reflect the physical. He was both eager to . Geoff ’s powerful running. forward. “The attacker steps forward into the right stance to deliver a right lunge punch. left. A close inspection of the larger co-text shows that there are four adjectives in this list that refer to physical stance only: upright. personal. tall and with a dis- tinctive upright stance.” (GVF). These adjectives are illustrative of how stance is conceptualized in naturally-occurring speech and writing. “From this position the begin- ner steps into the forward stance. For example: “Fair-haired.

“Clearly. For example.” (FAJ). particular.” (CGB). physical stance: “There- fore. personal/moral: “…men are useful but disposable. and this applies to both physical and personal/moral stance. positive. is difficult to assess in terms of the collocational evidence alone. ‘Fuck The Right To Vote’. previous as opposed to current and future. And many American women see Hillary Clinton as much more…” (CEK). Consider for example: “…the group have attracted most attention for their confrontational political stance. in a paper written in 1951. For example: “But this does not mean that it takes the conservative stance of necessarily accepting existing defini- tions of crime. took a different stance from her psychoanalytic contemporaries when she ana- . conservative. “…the anti-abortion lobby had apparently been indicat- ing the ‘weak’ moral stance of protestants on the issue. it does receive sup- port from the collocates basic. and anti-abortion. suggest that stance may often imply a high degree of severity or strength. For example: “…loss of support for the Greens was attributed to their critical stance on unifica- tion…” (HL2). Adjectives that highlight the personal and moral nature of stance include: moral. ideological. the music group’s political stance is apparently suffi- ciently well-known as to attract attention. is overtly supported by the collocational evidence as well – in particular by the adjectives political (the most frequent adjective collocate of stance in the BNC) and the adjective public itself. tough and aggressive. Introduction 13 adopt the right stance and unnerved by the strangeness of it. this kind of legislation reflects a more positive stance to- wards older workers and their role in the labour market. Each of these adjectives implies some sort of comparison with other stances: basic as opposed to more complicated. namely that stance is public. Here. particular as opposed to general.” (B01).” (A07). Consider the use of different stance in the following BNC excerpt: “Geraldine Pederson-Krag. the public nature of the stance is highlighted by the very adjective itself. and negative.” (CRX). by virtue of its very meaning. previous. However. most pertinently summed up on their recent mail order-only single and the LP’s strongest track. The third principle. “…at a time when Bush adopted an anti-abortion stance. Similarly. Two collocates. “I have to report a very negative stance by the British Mountaineering Council to our association…” (ECG). and is summed up (publicly) on a single from a record album. and different. namely the interactional nature of stance. The victims of their aggressive stance are their children. and different. and “it was necessary for the Soviet Union to adopt a more aggressive stance if it were to maintain credibility as a rival to Washington…” (G1R). Other evalu- ative stance collocates include critical. neutral.” (CH1). in “A new Archbishop of Canterbury was selected who took a public stance against clergy in gay and lesbian relationships…” (C9S). The second principle observed in the previous section. which. sets two stances in opposition to one another. Piaroa boys are not constrained into learning the aggressive stance of young warriors to fight men of this world…” (CJ1).

14 Robert Englebretson lysed the system of mass production. we again glimpse the interactional and collaborative nature of stance. this volume.4 Summary The previous three subsections have offered a quantitative and qualitative over- view of how speakers and writers use the term stance. The qualitative analyses of the three tokens in the SBCSAE brings to light five general principles about stancetaking that can be inferred from the use of the term stance in the context of interaction. 2. It is no wonder. that stance has been appropriated for language-related re- search. In this sentence. in terms of the interplay between quantita- tive and qualitative analyses).1 demonstrates that stance is an infre- quent lexeme. as one stance here is being juxtaposed against another – just as in examples (1)–(3) in the previous section. occurs far more often in written discourse than in spoken discourse.” (CBH). These distributional facts. Section 2. which is rec- . to varying degrees. Secondly. the types of language it occurs in. stance is a public act. These both necessitate a broader understanding of the sociocultural and interactional nature of the stances being taken. From this brief extract. personal belief/at- titude/evaluation. and the broader interactional and collocational contexts all play a role in how stance is conceptualized. Hunston. this is in keeping with its formally-skewed distribution and prevalence in specialized genres. suggest that stance is a term that tends to occur in quite specialized genre contexts. First. then. Even the three tokens of stance found in the SBCSAE illustrate the restricted contexts of use: one is a specific reference to posture in a judo class. in the subsequent papers in this volume. stance refers to physical embodied action. and shows skewed distribution across the sub-genres of language recognized in the BNC. Approaching the meaning of stance from a usage-based perspective recognizes that frequency of use. two stances are clearly being pitted against one another: the stance of “Geraldine Pederson-Krag” and the opposing stances of her contemporaries. as in example (2). which would be virtually impossible to achieve by using collocational evi- dence alone (cf. especially the lack of a single token of stance in the 4-million-word conversational component of the BNC. The final two principles. indexicality and consequentiality. and/or the social morality espoused at the institutional level. and the second referring to the institutional views of another school. as in example (3). and the other two are specific to a conversation about academic institutions – one referring to a college admissions interview. would require more than a brief window of textual context. In terms of frequency of use. and these same principles in fact will ap- pear as themes. as in example (1).

In fact. Du Bois this volume. stance is interactional in nature. Section 2. and interested readers should of course consult the cited works for more information and specific details. Scheibman this volume). Keisanen this volume). stance is a relational notion (cf. This section provides a cursory overview to contextualize the papers in the pres- ent volume.e. Lyons. most stance literature makes explicit reference to one or more of these categories.  (1994: 13) . who has long championed the recognition of subjectivity in language. interpretable. Finally. specific stances are indexical. Introduction 15 ognizable. inter- action – these three terms have each been widely associated with stance. defines subjectivity as follows: We can say that locutionary subjectivity is the locutionary agent’s (the speaker’s or writer’s. quite simply. evoking larger aspects of the physical context or the socio- cultural systems in which they are embedded. evaluation. taking a stance has real consequences for the persons or institutions involved. For this reason. self-expression in the use of language. and a fair amount of work exists that deals with stance-related themes but that uses other terminology. although often this focus is only implicit. Du Bois this vol- ume). it is no wonder that the representation of stance in language-related research is also broad and multifaceted. This review is not intended to be exhaustive or comprehensive. Thirdly. as conceptualized in natural discourse. and serves to contextualize the present volume within this broader intellectual climate. Given the diverse nature of stance itself. 3. Approaches to stance from an academic perspective tend to focus to varying degrees on any number of these five principles. stance has been operationalized differently from one researcher to the next. Fourthly. and subject to evaluation by others (cf. stancetaking is consequen- tial (cf. Research on stance: Subjectivity.3 demonstrates that the adjectives that tend to collocate with stance provide further support for these five themes as observed qualitatively in the interactional data. Our analysis of stance in discourse suggests overwhelmingly that it is by no means referring to a monolithic concept. interaction As suggested by the subtitle of the present volume – Subjectivity. The following section summarizes a few relevant ap- proaches to stance from the perspective of language-related research. i. evaluation. collaboratively coming into being among the participants in an exchange and/or by virtue of opposition to other stances.. the utterer’s) expression of himself or herself in the act of utterance: locutionary subjectivity is.

Hunston and Thompson (2000) present work on evaluation from a variety of perspectives (see Thompson and Hunston (2000) for a comprehensive overview). inter alia). For the latter. grammati- cal. Both subjectivity and evaluation are strongly implicated in the following two definitions of stance. In oth- er words. Evaluation has been operationalized by Thompson and Hunston as follows: evaluation is the broad cover term for the expression of the speaker or writer’s attitude or stance towards. Similarly.. Evaluation figures prominently both in corpus linguistics and in systemic-functional linguistics. subjec- tivity remained out of the scope of interest for most linguists. Research in evaluation therefore seeks to explicate the range of lexical. Biber and . while subjectivity refers broadly to “self-expression” (Lyons 1994: 13). or assessments” (1999: 966). value judgments. define stance as: “personal feelings. However. evaluation can be roughly summed up as subjectivity with a focus. in earlier work. 10) explores the complex nature of subjectivity and its central role in language use. and laments that approaches to semantics and pragmatics that fail to fully embrace subjectivity are inherently flawed. evaluation implies self-expression that is focused toward a narrow purview – self- expression about the “entities or propositions” (Thompson and Hunston 2000: 5) present in the very language that the speaker or writer is currently producing. and has been more fully fleshed out over the intervening years (Langacker 1999. until the mid 1980s. attitudes. Bednarek (2006) provides a book-length study of evaluation in newspaper discourse. or feelings about the entities or propo- sitions that he or she is talking about.g. viewpoint on. Other work on evaluation has focused on elucidat- ing its features within specific types of language – e. Langacker (1985) notes the inherent role of subjectivity in construing a scene and profiling various aspects of it. Martin and White (2005) provide a thorough examination of evaluation from an appraisal framework. Athanasiadou et al. As Ben- veniste famously speculates: “language is marked so deeply by the expression of subjectivity that one might ask if it could still function and be called language if it were constructed otherwise” (1971: 225).16 Robert Englebretson Subjectivity. broadly construed. textual. Lyons (1981: Chap. That attitude may relate to certainty or obligation or desirability or any of a number of other sets of values.  (2000: 5) Thus. Biber et al. The subjective nature of point-of-view and scene construal is central to cognitive grammar. and intertextual means by which speakers and writers laminate their language with their attitudes and points-of-view about its content. subjectivity has indeed taken on a more central role in linguistics – especially in cognitive linguistics and other functionalist paradigms. Within the past quarter century after Lyons’ claim. (2006) pro- vides a recent and thorough overview of subjectivity from a variety of linguistic perspectives. is seen as an essential quality of language.

 (1986: 1) Researchers in both corpus linguistics and systemic-functional linguistics have contributed to the identification and description of lexis and grammar that serve as markers of stance. in “the lexical and grammatical expression” (Biber and Finegan 1989: 92).. Scheibman (2002) provides an overview of frequent . research on complement clauses and complement-taking predicates has also been fruitful. Field (1997) outlines the role of factive constructions to index epistemic stance. All sentences encode such a point of view. Both definitions strongly assert the subjective and evaluative na- ture of stance.” “assessments. … and the description of the markers of such points of view and their meanings should therefore be a central topic for linguistics. or commitment concerning the propositional content of a mes- sage” (1989: 92). affect (“personal feelings”) (cf. according to the two definitions cited above. Introduction 17 Finegan define stance as “the lexical and grammatical expression of attitudes. according to this approach. attitudinal. stance is located in form. 1989. Kärkkäinen (2003) similarly approaches epistemic stance by analyzing certain kinds of complement- taking predicates such as I think – which. In particular. i. we observe that stance can be subdivided into evaluation (“value judgments. and Rauniomaa (this volume) for an analysis of two similar epistemic constructions in Finnish. and epistemicity (“commitment”). have grammaticized into epistemic fragments. Conrad and Biber 2000. See also Kärkkäinen (this volume) for an analysis of I guess. following Thompson and Mulac (1991). The grammar of English modals has also been well documented and investigated (cf. judgments. as have evaluative adjectives and nouns (Hunston and Sinclair 2000). minun mielestä (‘in my opinion’) and minusta (‘I think’). It is the goal of the stance researcher. entails a clear form/meaning relationship. Hyland and Tse (2005) provide an overview of what they term “evaluative that constructions” in academic writing. then. Ochs 1989).e. Researching stance. they encode their point of view to- wards it … The expression of such speakers’ attitudes is pervasive in all uses of language. to investigate how lexicon and grammar both encode and reflect the various categories of stance. With regard to epistemic stance. Other recent studies have focused on stance and subjectivity in naturally- occurring conversation. and style stances (Biber and Finegan 1988. feel- ings. Given the categories proposed in these two definitions.” and “attitudes”). Stubbs (1986) has called for a similar research agenda: whenever speakers (or writers) say anything. Downing 2002). adverbials have proven to be a rich source of various types of epistemic. then. or the degree of commitment/certainty of the speaker/writer. demonstrating how writers use these types of complement clauses to display their own stance toward the information they are presenting. review and discussion in Thompson and Hunston 2000: 20–21).

Matoesian’s work is also fairly unique among stance research. and orientation to. These are functional dimensions which apply across texts…”  (Berman et al. Affective). and Generality (of reference and quantification – specific vs. and tenses in American English. Precht (2003) provides a statistical analysis of stance-related lexemes and grammatical con- structions in various genres of British and American English. Berman et al. (2002).) Other recent approaches have begun to address the cross-cultural. cross-lin- guistic. historical. Attitude (Epistemic.18 Robert Englebretson combinations of subjects. papers by Johnstone (this volume) and Englebretson (this volume) similarly deal with how speakers may use stancetaking to index social identities. Sheibman (this volume) addresses the interplay between stance.  (1996: 420) Examples of recent work exploring the construction and realization of “social acts and social identities” (Ochs 1996: 420) through stance include Matoesian. Ochs (1996) writes: linguistic structures that index epistemic and affective stances are the basic lin- guistic resources for constructing/realizing social acts and social identities. general). Shoaps (2004) analyzes Zacapultec ritual wedding councils and explores the various cul- turally-situated practices of stancetaking. This role may account in part for why stance is elaborately encoded in the grammars of many languages. as already discussed. Epis- temic and affective stance has. analyzes subject-verb combinations that serve as epistemic fragments to index subjectivity and stance in American English conversation. Wu (2004) analyzes clause-final particles in Mandarin conversation and describes their use in marking epistemic stance. and developmental nature of stance marking. (See also Bucholtz and Hall (2005) and Benwell and Stokoe (2006) on identity. verb types. the sociocultural dimensions of stancetaking. Recipient). and papers in Berman (2005) investigate what they term “discourse stance”: We consider the notion ‘discourse stance’ as referring to three interrelated di- mensions of text-construction: Orientation (Sender. Another trend in recent research entails a recognition of. an especially privileged role in the constitu- tion of social life. On a more micro-level of identity. Deontic. as it explicitly addresses the role of physical stance and embodied action. then. arguing that grammatical and lexical patterns are shaped by subjectivity and the speak- ers’ needs to personalize their contributions to the discourse. Kärkkäinen (2003). and the wider social discourses and stereotypes that those stances may evoke. who provides an analysis of a focus group meeting of police officers and demonstrates that stance serves to “index broader forms of socio-cultural knowledge embedded in the professional division of labor between academic trainers and police train- ees” (2005: 169). noting cross-cul- tural differences in the expression of stance. Text. 2002: 258) .

espe- cially the contributions by Keisanen and Haddington. the interactional nature of stance is not just evident in conversation. as is the case with many (but not all) of the contributions to the present volume. and so on) is inherently intersubjective – as implied from the interactional in the category label interactional sequence. rather. part of what a stretch of talk is doing or accom- plishing or proposing. includes White. As Edwards (1999) observes. Written texts are broadly understood as interactive and interpersonal. from this perspective. Work on stance. it is a pervasive feature of talk. which takes as a startingpoint the intersubjectivity of lan- guage. intersubjectivity is not so much a special domain of study. Introduction 19 Scholars working in this approach have focused on analyzing and comparing a carefully-collected sample of written language from school children.  (2003: 259) Yet. nor something carried by a special range of devices but. negotiated. assessment sequences. illustrating that many have shifted from a subjective to an intersubjective orientation. for researchers whose work is either based in or informed by conversation analysis (CA). Verhagen (2005) presents one such notable in-depth study of this topic. intersubjective understand- ings are part of talk’s business.  (1999: 131) Thus. who views stance as an inherently dialogic activity. then. and realized in and through interaction. and investigates the textual resources that provide the means for speakers/writers to take a stance towards the various points-of-view or social positionings being referenced by the text and thereby to position themselves with respect to the other social subjects who hold those positions. any interactional sequence (such as adjacency pairs. especially in cognitive linguistics. but is widely recognized in research on stance in writing as well. Heritage and Raymond (2005) provide one recent study along these lines. specifically from an intersubjective approach. also deal with this aspect of stance head-on. Several of the papers in the current volume. of uptakes and recipient design. the theoretical and methodological underpinnings of CA take intersubjectivity as a given: In CA. of the entire machinery of turn-taking for instance. of various ages and in various countries. Furthermore. has begun to take such a focus. intersubjectivity is nothing new. to assess the development of “discourse stance” from a cross-linguistic perspective. Fitzmaurice (2004) presents a historical per- spective on the development of stance markers in English. takes its place among the aspects of language that are jointly constructed. and studies . Stance. However. The final trend that I will address in this section concerns the interactional nature of stancetaking. Recent work. where every word we use is indexical.

the -nya clitic. and to include an appendix of transcription symbols at the end of each paper as relevant. The current and final section briefly introduces each of the contributions to this volume in turn. field linguistics/grammatical description. Conclusion The previous sections of this introduction have contextualized this volume within the broader discourse-based and academic notions of stance. we have chosen not to pursue a uni- fied approach to transcription for the volume as a whole. and discourse linguistics. Hunston demonstrates the essential role of both quan- titative and qualitative analyses. and in preserving the data as originally transcribed. and epistemic-stance moves. In conclusion. as well as how the readers engage with the text. Hyland and Tse 2005. Johnstone explores the discursive construction of local identity. however. (Hyland 2005. White 2003. and the ways in which speakers use epistemic stance to support or undermine each others’ claims about local-sounding speech. This group of papers begins with Hunston’s contribution. the reader will notice a variety of transcription systems and conventions. This research background provides the intellectual context in which the papers in this volume have come into being. local identity. Before getting to the specific papers. 4. In the subse- quent paper. a discussion of stance in large electronic corpora.20 Robert Englebretson of stance in writing focus both on how the author engages the readers. Rather. evaluation. and interaction. In the interests of being true to the authors’ intentions. Each of the first four papers in the volume (after the introduction) addresses stance from within a specific approach to linguistics: corpus linguistics. Throughout this volume. sociolin- guistics. include subjectivity. broadly construed. we have chosen to allow each author to use the transcription system that best suits his or her purposes. exploring the relationship between dialect. and argues that grammatical descriptions of particu- lar languages need to pay attention to how grammar functions in the service of stancetaking. a note on transcription is in order. In the next paper. Johnstone offers a sociolinguistic perspec- tive on stance. this section has provided a brief overview of trends in stance research that. inter alia). and voice . Through an analysis of a sociolinguistic interview with two residents of Pittsburgh. Englebretson presents an overview of three aspects of stancetaking in colloquial Indonesian. and shows how these methods can profitably be used together in the identification and description of stance across and within various text types. Englebretson presents an overview of three aspects of colloquial Indonesian grammar (first-person-singular reference.

This group of four papers concludes with Scheibman’s contribution. In the following paper. are regularly oriented to in naturally-occurring talk. positions the subject. In the first paper in this group. and aligns with other subjects. a study of generalizations in American English conversation. and intersubjective orientations. Kärkkäinen provides a description and account of the epistemic fragment I guess in American English conversation. under the direction of Elise Kärkkäinen. in which the stance-taker evalu- ates an object. Scheibman highlights the interactive and collaborative nature of stancetaking. These three facets of stancetaking are mutually constitutive. known as the stance triangle. She demonstrates the consequential nature of stancetak- . Keisanen demonstrates the collaborative nature of stancetaking. Rauniomaa con- tinues the theme of epistemic fragments in conversation. as she analyzes challenges that are formatted as negative interrogatives. Finland. Keisanen takes up the theme of stance in the sequential organization of talk. and. subjective. Through this analysis of generalizations. demonstrating that generalizations are both subjective and intersubjective within the context of interactive discourse. Each of the remaining four papers in the volume offers a unique perspective on stance in talk-in-interaction – informed by Du Bois’s stance triangle model. providing an analysis of two stance markers in Finnish. These three dimensions relate to the act of taking a stance. and generalizations themselves often index commonly held beliefs and broader cultural values. Scheibman describes the grammatical forms and discursive use of a set of generalizations in her corpus. These four papers are also grouped together because the authors themselves were members of the stance research group at the University of Oulu. and explicates their function in the interactive projection of sequence organization. The contribution by Du Bois presents a unified conceptual framework and model of stancetaking. Introduction 21 alternation) and illustrates that in addition to the traditional functions ascribed to these three grammatical resources. and is intimately bound up with the sequential organization and interper- sonal projections in the conversation. Du Bois shows. in that the act of taking and challenging a stance relies on at least two participants. Du Bois ar- gues that stance is a social action that is characterized in terms of the interrelated dimensions of objective. as well as by methods of conversation analysis. speakers are actively using them to construct their social worlds through stancetaking. Kärkkäinen argues that this expression serves to organize stancetaking among co-participants in talk. Rauniomaa observes the role of these epistemic fragments within as- sessment sequences. and each paper in this group represents significant research based on that project. minun mielestä (‘in my opinion’) and minusta (‘I think’). In the subsequent paper. Participants in conversation use generalizations primarily to index solidarity with one another. and serves to some degree as background for each of the remaining four papers in the volume.

coughing. 4. and a range of data and questions.000 words for the four volumes of the SBCSAE was arrived at after exclud- ing non-word tokens such as speaker labels.934 W-Units according to http://www. stance indicates a mention of this English lexical item. although this number is slightly less when using the VIEW search interface). distributed by Oxford University Computing Services on behalf of the BNC Consor- tium.22 Robert Englebretson ing. by real people. a certain perspective. 5. and it should be understood as a use of this term in the full context of stance research. 2. However. this stance is open for challenge by another participant. no such specific lexical focus is intended. the number of W-Units (tagged words) in this corpus is actually around 97.619. The figure of 100 million words is generally cited as the size of the BNC. and throat-clear- ing. the understanding of stance that emerges out of the papers in this volume is likewise collaboratively constructed. This word-count is thus an accurate reflection of the number of spoken words (including truncated words and so-called filled pauses) in the SBCSAE to date. Notes 1. . just as stancetaking itself is interactive and emergent in dis- course.ac. pauses.natcorp. Throughout this chapter. in their full sociocultural environments. Because concor- dances and queries operate on W-Units. italics serve to differentiate mention versus use of the term stance.619. in that once a speaker has taken a stance. Each of the authors brings to bear a par- ticular approach.xml. 3. All rights in the texts cited are reserved. The volume concludes with Haddington’s contribution on stancetaking in news interviews. To conclude. all tables in this section will use the actual word-count generated by the concordance: 97. in order to focus on it as a specific term and to scrutinize its meaning and use. Haddington shows that stance in news interviews is both collaborative and consequential.uk/corpus/index. The figure of 249. and provide a starting point from which to pursue further exploration and refinement of the burgeoning field of stance research. Haddington observes ways in which interview- ers position interviewees with regard to stance. this figure also excludes non-lexical vocal noises such as laughter. When italicized. and non-vocal noises such as table thumps. Each contribution highlights key facets of language as used in actual discourse. I wish to gratefully acknowledge Mark Davies for creating such a wonderfully accessible and user-friendly search tool for the BNC.6 million (officially 97. and how interviewees in turn dia- logically align or disalign with that positioning.311 words according to the VIEW search tool. Taken together. All data cited from the BNC in this chapter have been extracted from the British National Corpus.ox. When not italicized. and for making it publicly available online for lan- guage research purposes.ID=numbers. the papers in this volume offer a rich overview of this important aspect of sociocultural life and lan- guage use.

2006. Canakis. M. and Finegan. Biber. and Stokoe. E. Berman. R. 10. A reviewer has raised the interesting question as to whether stance is used differently in American versus British English.” “PATTY. 1999. Benwell. Berman.” Written Language and Literacy 5 (2): 255–290.. URL: http://www. VIEW is based on David Lee’s indexing work of the BNC: http://personal. M.natcorp.” References Athanasiadou. and Finegan. and Strömqvist. The transcription conventions used in the SBCSAE are outlined in Du Bois et al. “Identity and interaction: A sociocultural linguistic approach... E. ID=corpus most of the audio recordings are stored in the National Sound Archive at the British Library where they can be accessed for research purposes.ac. FL: University of Miami Press.ac. 2005. M. E. There is some slight variation in the exact word-count of the written and spoken compo- nents. Stanford: CSLI. (eds. this is not a question I can answer here. New York: Mouton de Gruyter. and Kemmer. Barlow. (1993). A. Evaluation in Media Discourse: Analysis of a Newspaper Corpus. uk/ Bucholtz. 1971. and Cornillie. 223–230.A. The actual speaker labels appearing in the pub- lished corpus read “STEPHANIE. Meek (trans). H. Journal of Pragmatics 37 (2). Ragnarsdóttir. D. 2006. 2005. Usage-Based Models of Language. 1999.ZIP which claims to have corrected a num- ber of errors and inconsistencies in the genre classification found in the publication version. 2002.ox. B.” Discourse Studies 7 (4–5): 585–614. 2006. Biber.” and “MAUREEN. 8.hk/~davidlee/ devotedtocorpora/home/BNC_WORLD_INDEX. According to the BNC FAQ at http://www. Bednarek. Leech.xml.cityu. All numbers in this section refer to those calculated by the VIEW search tool. 1988. Johansson. S. Because of the vastly different nature of these two corpora. “Developing discourse stance across adolescence.natcorp. S. 2001. The Longman Grammar of Spoken and Written English. Discourse and Identity. “Adverbial stance types in English.” In Problems in General Linguistics.). version 2 (BNC World).A. R. Edinburgh: Edinburgh University Press. 9. depending on whether one uses the official BNC numbers or the numbers arrived at us- ing VIEW.A. Distributed by Oxford University Computing Services on behalf of the BNC Consortium. London: Continuum International. 1989.. The British National Corpus. E. “Styles of stance in English: Lexical and grammatical marking of evidentiality and affect. Benveniste. two of the speaker labels have been shortened from those presented in the published SBCSAE transcript.E.” “GAIL. S. Coral Gables. London: Longman.edu. B. Biber.ox. G. K. “Discourse stance. Introduction 23 6. Conrad. S.). and Finegan. 7. D. and Hall. “Subjectivity in language. Due to space considerations. (eds.” Special Issue.. M.uk/faq. D. and also appear in this volume as appendices to the papers by Scheibman and Kärkkäinen. C.” Discourse Processes 11 (1): 1–34. . Subjectification: Various Paths to Subjectivity.” Text 9 (1): 93–124. E.

Du Bois. 2004. Du Bois.” Dis- course Studies 7 (2): 173–192.W. C.). Houston.” Functions of Language 12 (1): 39–63.. and Biber. J. D. Part 3. (Panel Organizer). C.A.” Func- tions of Language 8 (2): 251–282.W. VIEW (Variation in English Words and Phrases). S.24 Robert Englebretson Bybee. S. and Tse. “‘Sure. Hunston. S. Meyer. J. “From usage to grammar: The mind’s response to repetition. 2005. Special Session at the Sociolinguistics Symposium 15. “Subjectivity. and Martey. and Englebretson.” Discourse Studies 6 (4): 427–448. sure’: Evidence and affect. J. Du Bois.W. “Adverbial marking of stance in speech and writing..” Social Psychology Quarterly 68 (1): 15–38. Heritage. 2000. Gardner. M. BNC online search tool. N. Downing. Hyland. Newcastle upon Tyne. 2005. 2. 2004. S. J. Philadelphia: Linguistic Data Consortium. “Stance and engagement: A model of interaction in academic discourse.A. TX. W.. When Listeners Talk: Response Tokens and Listener Stance. 1997.. 1989. Cumming. CD. Evaluation in Text: Authorial Stance and The Construction of Discourse. “A local grammar of evaluation. and Thompson. “Outline of discourse transcription.” Text 9 (1): 27–68.” In Pragmatics in 1998: Selected Papers from the 6th International Pragmatics Conference. Hunston and G.” In Talking Data: Transcription and Coding in Discourse Research. M. R.A. 2000.” In Evalua- tion in Text: Authorial Stance and the Construction of Discourse. Chafe. Davies. A. J. “‘Surely you knew!’: Surely as a marker of evidentiality and stance. K. Englebretson. and Paolino.” In Evaluation in Text: Au- thorial Stance and The Construction of Discourse. and Sinclair.L. Schuetze-Coburn. Part 4. Conrad.). 130–141. Stancetaking in Discourse: Subjectivity in In- teraction. intersubjectivity and the historical construction of interloc- utor stance: From stance markers to discourse markers. D. Thompson. DVD. Jaffe. New York: Oxford University Press. DVD. Part 2. April 1–4. Haviland. 1993. “The role of factive predicates in the indexicalization of stance: A discourse perspective. .” Language 82 (4): 711–733. 2004.W. 2005. The 10th Biennial Rice Linguistics Symposium. Ed- wards and M. Stance in Social and Cultural Context. Philadelphia: Linguistic Data Consortium. 45–89. S. K. J.edu. Hillsdale. 2002. “The terms of agreement: Indexing epistemic authority and subordination in talk-in-interaction.. Du Bois. Edwards. Amsterdam: John Benjamins. and Englebretson. Part 1. A. J. Chafe. S. DVD. 2000. “Evaluative that constructions: Signalling stance in research ab- stracts. S. W. New York: Oxford University Press. Hyland. J. Lampert (eds. Fitzmaurice.byu. and Thompson. Meyer. Hunston. R. Santa Barbara Corpus of Spoken American English. vol. Antwerp: International Pragmatics Association. NJ: Lawrence Erlbaum Associates. R. March 31–April 3. and Raymond.. J. 74–101. Field. 2006. 2004. D. J. 56–73. “Shared knowledge as a performative and rhetorical category. Philadelphia: Linguistic Data Consortium. Rice University.). Brigham Young University: http://view. Santa Barbara Corpus of Spoken American English. 2000. G.). R.B. Thompson (eds. Hunston and G.A.. 2003. 1999. S. (Symposium Organizer). 2005. Ver- schueren (ed. Du Bois. S.W. P.” Journal of Pragmatics 27 (6): 799–814. Santa Barbara Cor- pus of Spoken American English. G. New York: Oxford University Press. J. 2001. Thomp- son (eds. S. Santa Barbara Corpus of Spoken American English. Philadelphia: Linguistic Data Consortium.

). Shoaps. Lyons. LA. J. “Struck by speech revisited: Embodied stance in jurisdictional discourse. “Evaluation: An introduction. Precht. University of California. 313–329. 1986. 2003.). 407–437. “A matter of prolonged fieldwork: Notes towards a modal grammar of Eng- lish. 2002. . White. 1996. Langacker. Text 9 (1).). R. Thompson. and White.R.” In Iconicity in Syntax. Lyons. Text 23. E.R. New Orleans.). Mushin.” In Approaches to Grammaticalization.D. Amsterdam: John Benjamins. Point of View and Grammar: Structural Patterns of Subjectivity in American English Conversation. M. 109–150. Wu.” Text 23 (2): 259–284. Panel presented at the 101st annual meeting of the American Anthropological Association. G. Thompson (eds. S. M. New York: Oxford University Press.Syntax.-J.). A. J. Epistemic Stance in English Conversation: A Description of Its Interactional Functions. New York: Oxford University Press. J. Hunston and G. Trau- gott and B. Martin. Heine (eds. G. S. Silverstein. Stubbs. R. Morality in Grammar and Discourse: Stance-taking and the Negotiation of Moral Personhood in Sakapultek (Mayan) Wedding Counsels. 2003. November 21. 2004. E. I. “Beyond modality and hedging: A dialogic view of the language of inter- subjective stance. and Kockelman. “Stance moods in spoken English: Evidentiality and affect in British and Amer- ican conversation. New York: Palgrave Macmillan. New York: Cambridge University Press. “Observations and speculations on subjectivity.R. K. “Subjecthood and subjectivity. Ochs. P. The Language of Evaluation: Appraisal in English. dissertation. Haiman (ed.” In Subjecthood and Subjectivity: Proceedings of the Colloquium ‘The Status of the Subject in Linguistic Theory’. Stance in Talk: A Conversation Analysis of Mandarin Final Particles. 2. “Linguistic resources for socializing humanity. Gumperz and S. vol. verbal categories. “A quantitative perspective on the grammaticization of epis- temic parentheticals in English. Santa Barbara. Levinson (eds. (Panel Organizers). 2005. E. J. J. 2002. 2001.C. Ochs. R.” Journal of Sociolinguistics 9 (2): 167–193. 1–27. (ed.W. Constructions of Intersubjectivity: Discourse. M. Paris: Ophrys. Yaguello (ed. 1989. Amsterdam: John Benjamins. Evidentiality and Epistemological Stance: Narrative Retelling.). 1981. Ph. London: Fontana Paperbacks.” In Meaning in An- thropology. 1999.A. R. 1994. Scheibman. Linguis- tics Department. Amsterdam: John Benjamins. Introduction 25 Kärkkäinen. and Mulac. R. Basso and H. “Shifters. and Martin. 2005. Thompson. Grammar and Conceptualization. Amsterdam: John Benjamins. 1976. 11–55. 2003.” Special Issue. S. E. Langacker. P. 2000.). Amsterdam: John Benjamins. and Hunston. Selby (eds.” Applied Linguistics 7: 1–25. New York: Mouton de Gruyter. Amsterdam: John Benjamins. “The pragmatics of affect.” In Rethinking Linguistic Relativ- ity.R. K. with a Focus on ‘I think’.H. Shoaps. 2003. and Cognition. 1991.). Negotiating heteroglossia: Social perspectives on evaluation. Morality and Epistemology: Stance- taking and the Discursive Constitution of Personhood. Language. 9–17. and cultural description. Special Issue. A. Macken-Horarik. 2004.” In Evaluation in Text: Authorial Stance and the Construction of Discourse. (eds. J.R.” Text 23 (2): 239–257.W. J. P. Matoesian. 1985. M. Albuquerque: University of New Mex- ico Press. Verhagen. 2005. Meaning and Context.R.


both of which demonstrate the benefits of examining not just natu- rally-occurring data. Sinclair 1991). and language acquisi- tion (Biber et al. they allow us to observe multiple uses of a word or phrase in context. Such investigations reveal the “latent patterning” of language (Sinclair and Coulthard 1975: 125). It is true that papers in this collection and elsewhere identify particular forms associated with stance in various languages. or several types of mean- ing. Matthiessen 2006). but large amounts of such data. have been used in increasingly diverse ways to investigate topics such as lexis. Granger 1998. Lindquist and Mair 2004. Such calculations have led to. The phenomenon of stance represents an area of difficulty for corpus linguis- tics. electronic collections of large numbers of texts. Secondly. among other things. grammar. language change.1 There are two complementary aspects of corpus linguistics that are relevant to this paper. Firstly. Introduction Since their development in the 1960s. 1999. an increased concern for the concept of collocation (Sin- clair 1991. rather than a form. In addition. 2004). corpora give us the opportunity to quantify and make it particularly easy to quantify forms. and suggesting that meaning is disseminated across phraseologies rather than belonging to individual words (Sinclair 1991. Standard corpus investigation software allows the researcher to examine the frequency of words. or structures in a given corpus and to calculate comparative frequencies of co-occurrence. Semino and Short 2004. but it is . because stance is a meaning. software allows the researcher to look at each example of a given word or phrase in its immediate context in a collection of texts. stylistics. known as corpora. phrases. 2004) and for the examination of variation between registers (Biber 1988. showing the importance of context and intertextuality to meaning (Teubert 2004). Using a corpus to investigate stance quantitatively and qualitatively Susan Hunston University of Birmingham 1. a type of meaning.

by drawing atten- tion to a number of studies in this area. I focus on the methodological issues involved. the explication of a single instance normally implies that a pattern has been identified. Conrad and Biber examine stance adverbials in four broadly-de- fined registers (see Section 2). The paper then turns to issues of qualitative research and the role of a corpus in that research. in the sense that recurring instances of a phenomenon are noted. In fact. where it is proposed that the combination of such features amounts to a rhetoric of stance. In doing so. and that interpreting the role of stance in discourse entails a deeper understanding of the discourse as a whole than can be obtained from looking at the immediate co-text of an individual lexical item. It is argued that a corpus can help to demonstrate that many items. but simply that statements of the type ‘this is a demonstrably typical occurrence’ are worth making. The issues that such quantitative work raises are discussed in Section 4. numbers themselves are useless without interpretation. The first is that quantitative meth- ods are not irrelevant to discourse studies. and I make two assumptions. of course. In addition. The paper begins with two examples of largely quantitative studies of stance. research in the area of discourse will never be wholly quantitative. though not carry- ing stance themselves. although a comprehen- sive account of stance cannot be wholly quantitative. the availability of quantita- tive data in a corpus can assist the investigation of stance in texts in both qualita- tive and quantitative ways. I shall argue below that corpus methods can make a useful contribution to the investigation of stance. on the other hand. These problems notwithstanding. quantitative does not mean huge. both of which compare the frequency of stance exponents in two or more collec- tions of texts. The second assumption is that. in particular to find an account for the sometimes hazy intuitive sense that an utterance has a stance implication even when there are no obviously evaluative items in it.28 Susan Hunston always acknowledged that identifying stance entails more than simply locating those forms. the numbers themselves are derived from close ex- amination of many or all of the specific instances of the targeted phenomenon. This argument is carried forward in Section 7. This is the case even when the amount of data collected is rela- tively small. Sections 5 and 6 illustrate the use of a corpus to explain subjective responses to individual texts. while Charles compares stance-related patterns in two academic disciplines (see Section 3). and without the grounds for the interpretation being made clear. Section 8 concludes that. often co-occur with stance items. . and the explanation would hold true for other similar instances.

they are also able to examine each group of adverbials in more detail. They divide these adverbials into a number of cross-cutting categories: • Meaning. most surprisingly). forms. which they define as “grammatical devices used to frame a proposition” (Conrad and Biber 2000: 58). and style stance (e... Conrad and Biber proceed to compare three broad registers: conversation. Conrad and Biber distinguish epistemic stance (e. and news reportage with reference to these three variables. prepositional phrases (for a fact). showing which adverbials are most frequent overall. noun phrases (no doubt). They distinguish between single adverbs (evidently). and giving information about the uses of particular meanings. and that style adverbials in initial position comment both . adverb phrases (most surprisingly of all). • single adverbs are more frequent than the other grammatical realizations in all registers. and I bet account for most of the finite-clause adverbials in conversation (Conrad and Biber 2000: 70). pre-verbal. they note that style adverbials occur in news reportage “in articles which review sports or entertainment performances. Using a corpus to investigate stance 29 2. • finite clauses are more frequent in conversation than in the other registers. • prepositional phrases are more frequent in academic prose than in the other registers. • Form. For example. An example: Stance adverbials One of the best-known corpus studies of stance is Conrad and Biber’s (2000) study of stance adverbials.. at- titudinal stance (e. • final position adverbials are frequent in conversation. and in quotations” (Conrad and Biber 2000: 67). • marking of epistemic stance is more frequent overall than marking of attitu- dinal or style stance. and final. They identify four positions for stance adverbials: initial. and non-finite clauses (more simply put).g. • pre-verbal and initial adverbials are frequent in the written registers. I guess. and posi- tions. As well as these general conclusions.g. such as: • stance marking is particularly frequent in conversation. finite clauses (I think). ac- ademic prose. post-verbal. • Position. evidently). that “comment clauses” such as I think. simply put).g. by identifying and quantifying the occurrence of instances of each variable in cor- pora representing those registers. This exercise allows them (Conrad and Biber 2000: 63–72) to draw a number of conclusions.

pattern identifies the frequent be- haviour of a given lexical item. however. . 1996. such as V n ‘into’ -ing (verb followed by noun group and prepositional phrase introduced by into e.. is Charles (2004). It examines only one kind of stance marker and establishes a very limited number of categories of stance meaning. or N ‘at’ n (noun followed by prepositional phrase introduced by at e. Charles’s work draws on the notions of grammar pattern and of meaning group. Those iden- tified by Francis et al. and in terms of the specificity of the corpora concerned. perhaps.. but that moves beyond it methodologically. ‘call him a liar’ or ‘give her a book’). include some traditional complementation patterns such as V ‘that’ (verb complemented by that-clause e. and report verbs. and that compares corpora built according to different criteria. 3. Hunston. their focus on the variability between registers has inspired more detailed corpus studies that look at stance features beyond adverbi- als and break down the broad register categories that they use. Put briefly.g. Their work is a good example of one kind of corpus study.g. Charles restricts her corpora to two discipline areas: political science and materials science. for example. adjectives in patterns be- ginning with anticipatory it. Most seriously. She investigates a number of classes of items that realize stance: stance adverbials. but also some that are less traditional. between different types of academic prose or between different contexts of conversation. Francis et al.. nouns that occur both with appositive that-clauses and as anaphoric nouns (Francis 1994).. ‘believes that’) or V n n (verb complemented by two noun groups e. Bearing these criticisms in mind.. it uses an extremely broad definition of register. ‘…talk him into going…’). or ADJ ‘to’ n (adjective followed by prepositional phrase introduced by to e. Pattern and stance One example of work that is broadly in the tradition of Conrad and Biber. as developed by Francis. and to a particular kind of writing: postgraduate student theses. Hunston and Francis 1999). It is naturally open to criticism in terms of the natural fuzziness of categories of this kind. ‘…his anger at the killings…’). ‘…oblivious to the consequences…’). Conrad and Biber’s work is broad in scope but limited in the amount of detail it incorporates. one that delimits and counts categories. thereby linking two clauses (Conrad and Biber 2000: 71). thereby conflating a large num- ber of different meanings in a single grouping.g.g.g. not distinguishing. and Manning (e.30 Susan Hunston on what is coming up and on what has gone before. expressed as a sequence of elements.g.. 1998.

turn 7. expec- tation. the pattern V ‘on’ n has a number of meaning groups including: 1. Whereas the word possible. hypothesis. etc. depend. enlarge. As an example of such groups. • The evidence group: evidence. embroider. assertion. tell. it’s possible to imagine. lean. impinge. renege 6. observation. sugges- tion.) For each of these patterns. it’s amazing that. chance. prevaricate. grass. weigh 4. (1996. There are a number of advantages to this use of pattern. count. danger.) • it v-link ADJ to-inf (it’s easy to see.) • N that (the assumption that. etc. etc. intrude. deliberate. and these are the ones targeted by Charles as a useful starting point for searching a relatively large corpus. rely. spy. Verbs meaning ‘think’: brood. rat. ponder. rest. etc. the suggestion that. Verbs meaning that someone changes their mind: back-pedal. 1998). nouns with the pattern N ‘that’ are sub-divided into: • The idea group: idea. • The argument group: argument. Although the instances of stance identified are still limited by form – Charles. statement. ride. • The possibility group: possibility. etc. inform. Francis et al. For example. backtrack. infringe. hang. expand 3. hinge. making additions or correc- tions where necessary. Charles uses the same groups where possible. default. conclusion. point. Verbs meaning ‘depend’: bank. etc. com- promise. impression. meaning groups are reported in Frances et al. makes no claim to be exhaustive in her quantification of stance – these instances comprise words in context rather than in isolation. like Conrad and Biber. impact. Verbs meaning ‘interrupt’: encroach. listen in. muse. They include: • it v-link ADJ that (it’s possible that. Verbs meaning ‘spy’ or ‘inform’: eavesdrop. assumption. snoop. Verbs meaning ‘have an effect on’: act. etc. cogitate. rat. trespass 5. believes that. meditate. tell Some patterns have been identified from previous research (e. proof. ruminate. Verbs meaning ‘add details’: elaborate. notion. pivot. press.) • V that (suggests that. Using a corpus to investigate stance 31 One of the key findings in this work is that words sharing a pattern can be grouped according to sense. snitch. belief. for example. speculate 2.g. comment. etc. . re- flect. 1998) as tending to occur predominantly in the context of the expression of stance. indication.

However she also notes more Table 1. The use of meaning groups allows her to identify more specific types of meaning than Conrad and Biber do. Charles links this difference to a contrast in what counts as knowledge in the two disciplines: “knowledge in the natural sciences focuses primarily on universals. Charles’s figures for these four groups are shown in more detail in Table 1.  Meaning group frequency (adapted from Charles 2004: 118) Group Politics corpus Materials corpus (tokens per 10.32 Susan Hunston may have many different functions in a text. the functions of it is possible that or it is possible to are much more limited. Charles acknowledges.2 15. that nouns in the pattern N ‘that’ and in the argument group occur just over 45 times per ten-thousand words in the politics corpus compared with just over 5 times per ten-thousand words in the materials corpus. moving from the quantitative to the qualita- tive. There is space here only to give a few examples of her work. She also identifies a particularly group (particularly and in particular). numbers of words each of which may occur only infrequently. while that in the social sciences is more likely to be concerned with particulars” (Charles 2004: 42). and typi- cally. Furthermore.3 5. in the argument group of nouns occurring in the pattern N ‘that’. and she goes on to make more interesting observations about the differences between the two corpora. yet still to keep the study within manageable bounds. These are about twice as frequent in the materials corpus as in the politics corpus. Charles’s quantification of instances of stance. and 9 in her materials corpus.000 words) Idea 72.0 Possibility 6. which is more frequent in the politics corpus than in the materials corpus. nouns in the evidence group occur with approximately equal frequency.7 Evidence 14. consisting of generally. She notes. however. for example. using this method. Counting these nouns together allows even the very infrequent ones to be included in the quantification.6 13.000 words) (tokens per 10. Cit- ing Becher and Trowler (2001). in general. and so taking account of. In her study of adverbs. using patterns such as N ‘that’ and ‘it’ v-link ADJ ‘that’ allows Charles to incorporate nouns and adjectives into her study. of which 7 occur once or twice. For example. By comparison. that figures such as these give only general information. shows dis- tinctions between the disciplines under investigation.7 Argument 45. there are 36 nouns in Charles’s politics corpus.8 5. one of the meaning groups Charles identifies is the generally group. many occurring only once or twice. usually. as a whole. while at the same time grouping together.7 . as do nouns in the possibility group.

In both cases. such as the use of the individual word generally in the two dis- ciplines. not on research. Examples from her study include (Charles 2004: 120–121): (3) The difficulty with the assumption that ‘middlepowermanship’ was a function of ‘middlepowerness’ was that it was demonstrably untrue (PS) (4) These results are consistent with the idea that the smaller particles not only require less time to oxidise… (MS) Charles notes that the pattern ‘it’ v-link ADJ ‘that’ occurs with similar frequency in both corpora. it is used to comment on actions undertaken by the thesis writer. In addition. In the materials corpus. it is used to comment. But they have been found to be insufficient to explain the decision. using Charles’s methodology. Using a corpus to investigate stance 33 subtle differences. she notes that nouns in the idea group are used in the politics corpus to comment on the work of others. but in the ma- terials corpus to refer to the writer’s own work.2 Although the most frequent adjectives in the . has proved useful in distinguishing between other corpora. however. The main use in materials is to make a general statement within which the current writer locates their work. the other of issues of the Sun and News of the World news- papers (popular British tabloids). Hunston (2004). whereas the main use in politics is to make a general statement against which the current writer locates their work. In the politics corpus. (PS) (6) On removal it was apparent that some of the components of the fibreboard had melted. (MS) This pattern. (PS) As noted above. Examples are (Charles 2004: 95): (5) It quickly became clear. ‘it’ v-link ADJ ‘that’. however. the pattern is used to comment on the object of research. but on the actions of political actors. compares the oc- currence of the same pattern in two corpora: one consisting of issues of the New Scientist magazine. arguably. Charles divides nouns in the pattern N ‘that’ into meaning groups and compares their frequency. Charles (2004: 43) quotes these examples in illustration: (1) It is difficult to calculate this force and it is generally taken into account using a statistical treatment… We make the same assumption for a single particle (MS) (2) Thus far I have concentrated on the factors generally cited as having been responsible for the decision to abandon Britain’s world role. that the Bosnian Serbs were not eager to settle the conflict.

again allows generalisations to be made without saying anything about the specific instances of use. (New Scientist) (8) It is scandalous that the rich can buy the drugs privately. The difference is illustrated by these examples: (7) It was important to establish this because it was possible that strontium and calcium in fossils might have reacted chemically with the rock in which the fossils were buried. sad. . the tabloid newspapers make a judgment as to its desirability. Charles and Hunston have focused on a phraseology that is relatively stable. In selecting a pattern such as ‘it’ v-link ADJ ‘that’ for study. This does not mean that each instance is exactly the same as the others. her concern is the quantitative one of establishing what is usually or typically the case in each discipline. disgusting. rather than of a simple counting of words. and counting the occurrence of each adjective. disgraceful. by looking at each instance of use if necessary. even in quantitative approaches. looking at more instances and a wider range of adjectives divided into meaning groups indicates that in the New Scientist. but tough luck if you are poor. the group including important and vital. but it does suggest that broad generalisations can be made. and scandalous are most significant. and the group including outrageous. that the overwhelming majority of its instances do represent stance. On the other hand.34 Susan Hunston pattern are largely the same in both corpora (9 of the most frequent 12 adjec- tives in the pattern in each corpus are the same). Charles’s findings (outlined above) concerning the difference between her Political Science texts and her Materials Science ones in the typical use of word and phrases such as generally and idea that are the outcome of a close examina- tion of each instance. whereas in the Sun/News of the World. while the New Scientist com- ments soberly on the likelihood of a proposition. in the sense that it can be demonstrated. disappointing. it is worth pausing to consider what ‘quantitative methods’ of the kind outlined above involve. It will be noted that such generalisations are seldom. taken as the end point of the investiga- tion. Distinguishing between adjectives that occur frequently in that pattern in different types of text. In other words. often involving the scrutiny of several paragraphs or pages of text. (Sun) Before continuing to discuss the problems associated with quantitative methods. the group including possible and likely is most significant in occurrence.

as in: . scholarly is indeed used positively in the following examples: (9) …this is a fine. according to Becher’s findings. while poor work is sloppy. by counting particular words are likely to be unsuccessful. One possible conclusion from this research is that it should be possible to measure the quantity of stance marking in a given text or corpus by counting the instances of these adjectives and others that are similar. he finds that they are by no means always used to indicate stance. which occurs 648 times and which is sometimes positive. Using a corpus to investigate stance 35 4. elegant and economical describe good work. who discusses evaluative language in academic disciplines. Groom does find large numbers of the adjectives noted by Becher. on the other hand. but looking more closely. For example. on the other. on the one hand. whereas negative values are associated with thin and trivialising. Av- erage work in history. (10) Siegel has written an informative. The problems of quantification In spite of examples such as the above. where the meaning might be glossed as ‘pertaining to scholars’ rather than ‘demonstrating good scholarship’: (11) This book features the normal scholarly apparatus of a bibliography… (12) There will always be scholarly concern over… (13) …from the viewpoint of provoking scholarly debate… Overall. especially evaluative stance. and stance functions. might be described as sound or well-researched. scholarly. and well-written. original. it will be apparent that quantifying stance is problematic because there is no simple correspondence between individual words. and relatively short book…. He takes as his starting point Becher’s (1989) research establishing groups of adjectives – which according to interviewed re- searchers. scholarly study that is a delight to read… but is unrelated to stance in many more examples. In his corpus of his- tory review articles. Groom interprets only 48 of the 629 instances of scholarly in his corpus as having a positive evaluative meaning. positive values are associated with the concepts of scholarly. This is exemplified by Groom (2004). Groom’s topic is the adjectives used in book reviews in academic journals of history and of English literature. As a result. re- flecting the value-system of the discipline. For example. In physics. attempts to quantify stance. have an evaluative meaning in various disciplines – in each case. and average work is accurate. Becher says that in his- tory. Similarly he looks at the word original.

Channell points out that looking at a large number of examples of this phrase .g. then. if it is a phrase. 5. Groom argues that immediate co-text distinguishes evalu- ative from non-evaluative meanings. Put simply. It militates against the use of the large corpora that are essential for reliable quantitative results.. however. Looking at each instance of a word in a corpus has its own problems. however. and the relations become closer the more specific the form is taken to be (e. counting forms is not the same as counting functions. and that this quantification may lead on to more detailed qualitative work. This is the tradition of examining the typical contexts (co-texts) of individual words and phrases in order to identify their phraseologies and their functions. although it has the potential to lead back into the quantitative. In other words. In some instances. that context is crucial in identifying stance. there are relations between form and function. This tradition is beneficial to the study of stance. Corpora and the qualitative investigation of stance It has been suggested so far in this paper that a corpus may be useful in quanti- fying stance markers. On the other hand.36 Susan Hunston (14) …this volume comprises a detailed work of original research embellished with… but not necessarily so: (15) …his discussion is not based on original research but instead a synthesis of selected secondary sources… (lack of originality is not a problem) (16) Michael Farry’s study of County Sligo may not rival the sweep of Fitzpatrick’s analysis. and it presupposes that the categories of stance are uncontroversially distinguished. Groom suggests that only 258 of the 648 instances actually indicate posi- tive stance. and certainly not the pungency of his writing. or a word in a pattern. but it provides a work- manlike slab of original research on a vital moment in modern Irish history (originality is not enough) Again. The emphasis on context leads to an aspect of corpus work that is more qualitative than quantitative. A good example is Channell’s (2000) study of the phrase par for the course. and that this must therefore be taken into account in quantifying stance. because there is no one-to-one cor- relation between form and function. The point has also been made. and that this must be borne in mind when relatively crude quantitative measures are used. rather than a word). counting forms can be a useful first step in quantifying functions.

3 The notion of typicality (Sinclair 1991) is crucial to this kind of corpus use and has been used to argue for “hidden meanings” in those cases where a phrase is used in an atypical way (Louw 1993). I re in the head. G: Oh shit. She quotes (2000: 48) the follow- ing example. She notes that. which have been obtained by searching for the string wrong end of the stick:4 realises that he got completely the wrong end of the stick. Using a corpus to investigate stance 37 enables the researcher to identify a common context and a common function of the phrase. . and it is negatively evaluated. Terminal. In this example. the public gets the wrong end of the stick. One example is the following sentence. The more typical use is illustrated by the following randomly-selected concordance lines.goin the dail. is ign at cross purposes -. in its non-golfing uses. par for the course indicates an attitudinal stance and also performs a further interpersonal function in implying solidarity – expressing a shared ac- ceptance that life is going to deal us blows that we simply have to weather. which comes from an email message: I’m afraid I may have given Martin the wrong end of the stick. Par for the course out here… (20) …was cancelled at the last minute: par for the course… (21) Disturbing dreams are par for the course in pregnancy… Channell also points out that in dialogue the phrase is used frequently by a second speaker to show sympathy for the first speaker. and is women somehow managed to get the wrong end of the stick.getting the wrong end of the stick -. Well it’s par for the course for today I think. So erm I can only think there’s something in the file that’s doing it. par for the course is used to comment on two aspects of a situation: it is non-remarkable. Women treatment. Examples from Channell’s paper (2000: 47) include: (17) …the vicious infighting that was par for the course… (18) The third was out of order. ‘He must have got the wrong end of the stick’ said o that is because they’ve got the wrong end of the stick and the calls back erm I think she’s got the wrong end of the stick and blo <ZF1> you’ve <ZF0> you’ve got the wrong end of the stick there. but sometimes they get the wrong end of the stick. which was par for the course… (19) A for sexuality and E for subtlety. in which speaker F is complaining about problems in printing a computer file: (22) F: …Soon as you hit ‘Okay’ that’s it. It can be shown that the phrase the wrong end of the stick is being used atypically here.

” Actua And she’d really got hold of the wrong end of the stick. There applicant obviously got hold of the wrong end of the stick.) It is true. becaus In most of these instances. your teacher will get the wrong impression. lastly. as in the cited email.g. while get hold of implies agency. for example. however. This argument was based first on an intuitive interpretation of the examples. This suggests that either speaker or hearer can be deemed to be responsible for the wrong impres- sion.’ Ex- amples include the gaffe-prone Prince had got the wrong end of the stick again. is the Thumbs Up control of the recording VCR solely by infrared remote control? A: You seem to have got the wrong end of the stick. agency is at least ambiguous. Following Louw’s suggestion that atypical usage can indicate insincerity. and It is. Frequently. with this phrase. there are no co-occurrences with give. an inter- esting asymmetry emerges. where the verb is get.” STIMAC then got hold of the wrong end of the stick and tho give wrong advice or get hold of the wrong end of the stick. the phrase the wrong end of the stick follows get or get hold of.’ Examination of more context in 25 examples (one-third of the total in the Bank of English) only partially resolves the issue. . however.38 Susan Hunston too many men have got the wrong end of the stick.5 The problem is that. not the giver of it. Q: And.) If the search is expanded to less meta- phoric phrases such as the wrong impression or the wrong idea. most instances of get the wrong end of the stick do have something in the surrounding co-text that clearly ascribes blame to the holder of ‘the stick. as not all speakers share that intuition. This negative evidence provides some support for the view that with this phrase the hearer is typically construed as responsible for the misunderstand- ing. however. get can be used either to imply agency (as in I’m going to get a drink) or with a meaning similar to ‘be given’ (as in it was a big surprise to get the Young Player award). In other words. The phrase the wrong impression.. In fact. into the whole text if necessary. that other examples have no such contributing evidence and so are potentially ambiguous (e. however. very easy for an individual to establish a limited mind-set and get the wrong end of the stick. of which 72 occur with the verb give and 33 with get. (It might be noted in passing that this search for evidence goes well beyond the confines of the standard 80-character concordance line. clear ascription of agency to the speaker never or very rarely occurs. I would argue that in each case responsibility for the misunderstanding is represented as lying at the door of the receiver of the message. so that a clear indication of the contrary. is markedly atypical. Thus. In the case of the wrong end of the stick. an example such as the public gets the wrong end of the stick could mean either ‘the public misun- derstands’ or ‘someone misleads the public. as in If you let sloppy errors slip by. They m 4 February 1995) has the wrong end of the stick as far him off. further evidence is necessary. however. occurs 140 times.

reproduced here: (23) The intention of this article is…to share a very real concern about our survival both as teachers and as human beings. important embodiments of stance. Sinclair uses the term semantic prosody to refer to instances where an extend- ed phrase. the media. then so be it. First. at risk. typically involving attitudinal meaning. and that it is cumulative. the pressures of. Often. indicate stance on the part of the writer. or “unit of meaning. however. 6. unless they are negated or attributed. is the stance to be found in the words and phrases of the paragraph: a very real concern about. This is partly what lends stance its subtlety in text: it is difficult to pin down. which is frequently associated with stance even though it does not appear to . If this seems unduly dramatic. such as dramatic. such as tragedy. interpreting as undesirable such concepts as living at a fast pace. there are words that indicate different kinds of stance. Third. however. or is it something that the reader brings to the paragraph with his/her experience of life. Using a corpus to investigate stance 39 using the phrase with give instead of get might indicate that what appears to be an apology for causing a misunderstanding is actually an accusation. the semantic prosody of a phrase is difficult to intuit and can be observed only by using corpus search techniques. Hunston (2004) discusses this problem in relation to the opening paragraph of an article about stress among English-language teachers. leaving less and less time for mature reflection and the exercise of indepen- dent choice. Our situation is already dramatic almost to the point of tragedy. Its purpose is to tell the reader that something is wrong. We are at risk – from the pressures of consumerism. Such phrases are. that it is difficult to be specific about where in the paragraph this stance is ar- ticulated. Our lives are lived at an increasingly accelerating pace. which. Where is stance located? Sinclair’s work highlights the fact that evaluative meaning does not occur in discrete items but can be identified across whole phrases. not having time for reflection and choice? There are in fact several different answers to this question. More particularly. technology. Hunston (2004) argues. there is phraseol- ogy. or units of meaning.  (Maley 1999) This paragraph is unapologetically opinionated and highly evaluative. and rampant trivialisation. there are words here.6 Second. tragedy. depending on context.” such as get the wrong end of the stick has a con- sistent pragmatic or discourse function.

The two possible meanings of the word are exploited to embody two competing stances: that of the imagined . Its appearance in example (23) indicates evaluative stance. both meanings appear. all of which. the writer ascribes to the reader the sense that the writer’s statements are unduly dramatic (unreliable and exaggerated). indicate a more general sense of misfortune and sorrow. among its other uses. it can attribute to a speech act the quality of ‘unreliable exaggera- tion’ or it can indicate a judgment of an action as ‘a significant activity with nega- tive effects. 6. (27) I am sometimes asked why I believe.40 Susan Hunston indicate stance in itself. Each of these examples will now be considered in greater detail (following Hunston 2004).’ In example (23). which separates the specific from the general. The example to be illustrated below is an increasingly accelerating pace. where considerable corpus investigation and interpretation is needed to make sense of the evidence. when there is so much evidence of trag- edy and evildoing. he asserts that the situation is indeed dramatic to the point of tragedy (significant. with negative effects). Apart from the count/non-count distinction. continuing: (24) Survivors of the tragedy have been angered by the report’s findings. First. Typical examples include: (25) …my explanation of recent inner-city history is filled with a deep sadness and sense of tragedy… (26) Many of the graves tell their own moving stories of a night of tragedy at sea.2 Dramatic Investigation of this word in the Bank of English corpus suggests that. One example recalls a shipping accident and recounts the results of a government inquiry. 6.1 Tragedy A search of the Bank of English corpus suggests that tragedy is typically used as a count noun and refers anaphorically or exophorically to an event assumed to be recognised by the reader. such as to the point of. except those refer- ring to a dramatic genre. A search for of tragedy yields non-count uses. second. Finally. there are examples that are more subtle still. there is little apparent variation in the use of the word.

in order of frequency. even to the point of looking ridiculous. the phrase is used to indicate that a normally acceptable attribute or behaviour is exaggerated in a particular instance so that it becomes much less acceptable. and that of the writer himself. in almost all instances of its use in the Bank of English. it can be shown to co- occur. 6. Using a corpus to investigate stance 41 Table 2. If . The stances are conveyed in each case by the word in context. More accurately.  Words preceding and following to the point of close to the point of being thin absurdity naïve self simple idiocy modest paranoia cautious invisibility familiar recklessness quiet farce misleading emaciation plain stupidity intense rudeness frank near brave obsession polite madness sincere lunacy charming insanity confident caution cool non shy making careful cruelty blunt vulgarity selfish impossibility curious blandness reader. while the concordance lines show specific examples. essures which saw him tense to the point of obsession over every to Restaurants are often basic to the point of being downright scruffy ll copy them slavishly. with what might simplistically be called negative evaluation.3 To the point of Although to the point of does not in itself indicate stance. Table 2 lists the words occurring to the left and right of to the point of.

so you will re ubject of trade. The two side can be close-mouthed almost to the point of curtness. . This example indicates that we cannot simply assume that every example of to the point of will indicate negative evaluation. It is possible. one minute laid-back to the point of torpor. in the context of committee chairmanship. to the point of gentleness. It can be deduced in this example that gentleness. Examination of a very wide context (the whole article in this case) can find no evidence that the writer considers invigoration to be a negatively-evaluated quality. that attempts to quantify negative stance by counting phrases such as to the point of. Apparent counter-examples include the following: (28) At 23. Hazel Dooney is fresh to the point of invigoration. however. The navigat we thought he was obsequious to the point of irritation. Unfailingly courteous. given by one of the nine members of his committee. will not be wholly successful. told the imp efers to his cinematographer to the point of throwing away necessary In not every case. The counter-example is a warning. and with just one exhibition behind her. but later r nfused. but in the absence of other evidence this must remain speculation. to the point of. because unfailingly courteous to the point of gentle- ness occurs as a concession countermanded by positive evaluation (fierce inde- pendence and skill).42 Susan Hunston were also low and restricted to the point of inadequacy. without the implications of the dominant use. does the word to the right of to the point of clearly in- dicate a negative evaluation. A more substantiated argument is that this writer simply uses the phrase in an unusual way. Lord Nolan also displayed a fierce independence and skill in both bringing together his disparate committee members… The second of these demonstrates how well the phrase to the point of does its job. However. is a bad thing. to suggest that the writer’s positive evaluation is in fact insincere. this use is so dominant that it can be used to argue that dramatic to the point of tragedy has three indications of stance (dramatic. On the other hand. Ten days ago articulate and high-handed to the point of arrogance. (29) This assessment. reflects a genuine warmth and respect for their chairman. however. the phrase to the point of provides the reader with a short cut to arriving at this interpretation. it is vague to the point of being coy. and that to the point of invigoration is a slip that reveals this. and tragedy) rather than two. without examining each instance in its context. the next a mass all enjoyed the joke almost to the point of seizure. following Louw (1993). No similar explanation can be found for the first example.

The notion of belonging to t 33. increasingly attractive) as well as with those indicating negative evaluation (increasingly difficult. Ge He died as he lived. however. With this phrase. can no longer live at the hot pace set by today’s younger. we cannot do something so simple as to look at the examples in a corpus and count the good and the bad. Come-on. For example. at breakneck pace. in fact. increas- ingly desperate). Rather. and suggesting that there is no dominant usage to be called on. or to an individual (at your own pace): . Looking at examples in the Bank of English. with the triplets and thei that life is lived at such a hectic pace and people are so self-cent population living at a hectic pace. in that a search of the Bank of English reveals no examples of it. str has learnt to live with the jittery pace of development and capital expect to live with his searing pace. In the same week that Cham a slow pace of life is not necessarily good: mortals who live life at a gentler pace.” recalled the woman execut we are to living life at a frantic pace. to a period of life (at a child’s pace). Using a corpus to investigate stance 43 6. the pace of life is not described as fast or slow as much as appropriate: to humanity (at a human pace). This suggests that increasingly is being used to indicate stance rather than information.” <hl> Missing. shows that the adverb co-occurs with words indicating positive evaluation (increasingly confident. it is very easy to live as They live their lives at a frenetic pace that allows contact with th said the family lived life at full pace. Doctor Pain’ is acceptably In many instances. Open the door to new roman to live with the STUS’s plodding pace. He was only 24. it is clear that although a fast pace of life is usually bad: They live life at an accelerated pace. We therefore have to examine the concept of accelerating pace. yet he was and live our lives at such a fast pace. most Ibizans preferring ju and live life at their own yawning pace. we have to use the examples to uncover the cultural presuppositions behind the phrase. It can also be pointed out that the phrase is an instance of tautology.4 An increasingly accelerating pace The phrase an increasingly accelerating pace is very different. with increasingly being redundant in terms of the ideational meaning of the phrase. a situation the two rambun setting lives at its own slow pace. ‘It’s slower but it’s much their lives organized at a snail’s pace. indicating that the phrase itself is rare.

In the only example with an agent – The administration sought a means to control the acceler- ating pace of Soviet arms acquisition – the agency is unsuccessful. In other words. the opinionated style of example (23) does not depend on its words.44 Susan Hunston nimals is living at the appropriate pace of its own biological clock could be like if lived at a human pace. Just the odd contact every It is at this pace. 7. but because a ‘good’ pace of life is one that is controlled. a normal living pace. If we now look at accelerating pace (not necessarily pace of life) we find mixed evidence – about half the examples indicate something bad: (30) …have seen events move towards war at an accelerating pace… (31) The administration sought a means to control the accelerating pace of Soviet arms acquisition and half indicate something good or neutral: (32) He was quickly swept up in the studio’s accelerating pace (33) …the accelerating pace of reform In all the examples. a pace they’ve been used t learn to live life at your own pace. and occasionally wear fals There’s time to live at a child’s pace rather than being ruled by An interpretation of this evidence is that what is highly valued is a pace of life that is under the control of the person living it. and an accelerating pace of anything is not controlled. but on its phraseology. whether that be fast and exciting or slow and manageable. that the elderly must func enjoy living life at your own pace. not because it is fast. he preferred s prefers to live and work at his own pace. It sug- gests that the evaluative character of a text – the amount and the type of evaluative stance that it expresses – depends not only on the presence of evaluative lexical items but also on the use of phrases that can be shown to resound with intertextu- al meaning. This person doesn’t care f But they live life at their own pace. In other words. Other examples can be found. though. In none of the examples – good or bad – is there an agent controlling the pace. the accelerating pace is always out of control.” Initially. if we put together the results of these two studies. An example that has been examined in detail above is the phrase to the point of. even in such a . A rhetoric of stance The above discussion of a number of instances from example 23 confirms the view that evaluative meanings are cumulative and occur across phrases in texts. an accelerating pace of life is bad.

Mr them. but intuition in this regard is un- reliable. The function of the whole is to indicate that an unfortunate side-effect of a rec- ommended action is considered acceptable. A corpus search for very real NOUN about shows that the nouns that occur in this sequence include: concern. One candidate is the phrase a very real concern about. Sample concordance lines are shown below. This argument also implies an alternative view . dismay. which frequently co-occur with stance. so be it -. Using a corpus to investigate stance 45 short paragraph (Hunston 2004). rather than the words concern. If it lacks easy saleability. and reserva- tions. occurrence. so be it. so be it.” adding: “If that means prison.the most interest If we have to do it by being hip.” Mr Leeson said he for example. But Hearts have bee If that means somebody gets booked. As a result of this much the better. It’s a parody of re If it means going into the red then so be it. If that means taking drugs. and an examination of many instances of the target phrase is required to corroborate the perception of its role in the text under investigation. fears.” the manager said. This again indicates a degree of redundancy. In this context. example 23 contains several phrases that can be shown to recur in the context of evaluative stance. the if clause indicates an unfortunate. Of course and if we have to go to court. but unavoidable. <p> We can get back <p> If the fans want to blame me then so be it. to the point of. Blackford If we have to battle this out in court so be it. A corpus can provide just such a set of instances. In each case. so be it. so be it. then so be it. her turn. th If it upsets their preparations then so be it. there hav if that means looking a bit cartoon-y so be it. but the embedding or otherwise of those items in phraseologies. Moreover. If she dies in the meantime. We must be competit disease. concerns. A JONES.” he says. and the redundancy they represent. believe me. in that the frame very real … about restricts the choice of noun to the extent that its meaning is almost wholly predictable (see also Tognini-Bonelli 2001: 118 on the meanings of real). then so be it. The investigation of this short text extract leads to a hypothesis: what dis- tinguishes subjective (or stance-heavy) from objective (or stance-light) texts is not the quantity of explicitly evaluative lexical items in each. so be it. tragedy. even though they do not evaluate themselves. misgivings. Those phraseologies can be identified intuitively. Another example is the sequence If…so be it. If force is needed. In short. so be it. and so on themselves. so be it. if…so be it.” <p> The agency has and if we are soundly beaten. In other words.” A spokesman for Gu life. But. what makes this paragraph stance-heavy is the presence of a very real…about. If that means complexity. The recurrence of the function of the sequence in turn means that the stance it represents can be taken as given.

. 2. Corpora present us with evidence for intertextuality and lead us to conclusions that the meaning of a single instance is dependent on the meanings of many other single instances. a broader qualitative analysis of stance likewise incorporates turns-at-talk. the quantitative. and the comparison between various corpora in this regard. The use of tragedy to indicate a type of play is excluded here. jointly owned by the University of Birmingham and Collins publishers. Whilst the counting of stance items. Conclusion: Quantitative and qualitative corpus research In this paper. especially the chapters by Du Bois. Channell notes that the phrase seems to be infrequent in American English – perhaps the essential pessimism it expresses is a purely British phenomenon. 6. and Haddington. remains a viable and useful activity. stance meaning is also distributed across larger interactional units. Other examples used in this paper are from the same corpus unless otherwise stated. However. As demonstrated in other contributions to this volume. 8. interactional sequences. and intertextuality. it is clear that of necessity it provides a partial view of stance.46 Susan Hunston of the role of quantitative information in studies of stance. These concordance lines are taken from the Bank of English corpus. in the case of corpora. These markers are typically phrases rather than individual words. that counting can be used in the service of qualitative research as well as quantitative. I have also tried to suggest that whereas computers are essentially used to count items. Keisanen. and so forcing me to examine the evidence more closely.’ has an important role to play in the qualitative activity of identifying stance elements. Notes 1. Both these corpora are components of the Bank of English corpus. 5. I am grateful to Robert Englebretson for disagreeing with me here. 4. I have also suggested that qualitative work using corpora can show typicality of use and in doing so can enable us to identify stance markers – particularly markers of evaluative stance – that were previously un- known. I have tried to show that corpora can be used effectively to quantify markers of stance – though this work must be complemented by a more qualita- tive approach – and raw figures should be treated as the starting point of investi- gation rather than its end point. In addition to phraseologies. in the sense of giving information about ‘what is usually said’ or ‘how a phrase is typically used. 3.

57–73.. S.” In Text and Technology: In Honour of John Sinclair. G. Corpus Approaches to Grammaticalization in English.” In Evalua- tion in Text: Authorial Stance and the Construction of Discourse.” Postgraduate seminar. Hunston.). 1999. 1994. Hunston. London: Equinox.D. Amsterdam: John Benjamins. Hunston. Bern: Peter Lang. 1989. and Mair. E. Oxford: Oxford University Press.). S. Oxford: Oxford University Press. G. Sinclair. D. M. Milton Keynes: Open University Press. and Manning. Collins Cobuild Grammar Patterns 2: Nouns and Adjectives. Department of English. “Labelling discourse: An aspect of nominal-group lexical cohesion. the sound and the thin: Evaluative adjectives and disciplinary values in academic book reviews. Conrad. 157–176. Collins Cobuild Grammar Patterns 1: Verbs. H.).” English Teaching Professional 10: 3–7. G. S. Evaluation in Text: Authorial Stance and the Con- struction of Discourse. Coulthard (ed. B. G. 2004. 2004. Morley and L.). 1998. Matthiessen. T. J. J. Tognini Bonelli (eds. J. “Corpus-based analysis of evaluative lexis.. Corpus and Discourse. University of Birmingham. P.” In System and Corpus: Exploring Connections.” In Evaluation in Text: Authorial Stance and the Construction of Discourse. E. “Systemic functional profiles of system and text: Investigations based on texts. C. The Author’s Voice in Academic Writing with Reference to Theses in Politics and Materials Science. S. G. and Finegan. London: HarperCollins. “Counting the uncountable: Problems of identifying evaluation in a text and in a corpus. and Thompson.). and Trowler. Baker. S. S. Francis. Using a corpus to investigate stance 47 References Becher. S. Francis and E. Corpus Stylistics: Speech. and Biber. and corpora. M. Francis. Johansson. London: Longman. Academic Tribes and Territories.. E. S.. Leech. 1996. S. 38–55. Ox- ford: Oxford University Press. Milton Keynes: Open University Press.” In Ad- vances in Written Text Analysis. Becher. Ph.). Variation in Speech and Writing. Hunston and G. Charles. Haarman (eds. G. Hunston. C. 2000. Hunston (eds. G. Learner English on Computer. 2000. M. 1991. Corpus Concordance Collocation. Semino. and Manning. 2006. London: HarperCollins. “Surviving the 20th century. T. Granger. S. G. Partington. 1988. (eds. (eds. “Adverbial marking of stance in speech and writing. Francis. University of Birmingham. E. Maley. 2nd ed. 103–142. 1999.). 2000. S. and Francis. Groom. D. Thomp- son and S. Cambridge: Cambridge University Press. D. Lindquist. 83–101. Pattern Grammar: A Corpus-Driven Approach to the Lexical Grammar of English. Conrad. . Amsterdam: John Benjamins. 2001. Sinclair. 2004. London: Routledge. Thomp- son (eds. text archives. “The scholarly. London: Routledge. Academic Tribes and Territories: Intellectual Enquiry and the Culture of Disci- plines. London: Routledge. A. Biber. Hunston. 157–188. N. Hunston and G. dissertation. A. 1999. J.. Louw. M. Channell. Longman Grammar of Spoken and Written English. 2004.” In Corpora and Discourse. and Short.). Biber. Writing and Thought Presentation in a Corpus of English Writing. Amsterdam: John Benjamins. Oxford: Oxford University Press. 1993. 2004. 2004. “Irony in the text or insincerity in the writer? – The diagnostic potential of se- mantic prosodies. Trust the Text: Language. 1998. Thompson (eds. London: Longman.

2001. and Coulthard. 1975. Amsterdam: John Benjamins. C.K. London: Con- tinuum. M. 73–112. Tognini-Bonelli.” In Lexicology and Corpus Linguistics. Yallop and A.48 Susan Hunston Sinclair. Čermáková (eds. 2004.A. Halliday. Corpus Linguistics at Work. J. Oxford: Oxford University Press. Teubert. W. M. Towards an Analysis of Discourse: The English Used by Teach- ers and Pupils. Teubert. W. E. “Language and corpus linguistics.). .

and they perform local identity by means of playful performances of local-sounding forms. and there is a copy in most middle-class homes. local speech is almost invariably mentioned or performed. Pennsylvania (U. Johnstone and Baumgardt 2004.S. Johnstone et al. Trudgill 1986) and. While much of this discourse about local speech arises in the context of nostalgia for the city’s working-class industrial past. A dictionary of Pittsburghese.) area. whether or not they make routine use of stereotypically local phonological or lexical variants. and coffee mugs decorated with lexical items thought to be local and other words spelled in such a way as to sug- gest local pronunciations. Johnstone and Andrus 2005. discourse about Pittsbur- ghese also enters into youthful identity work. shot glasses. a word derived from the local variant of the second-person-plural pro- noun. One way to attribute a quintessentially local identity to a person is to label him or her a yinzer. Dubois and Horvath 2002). In a playful nod to the New Yorker. . increased popular attention to re- gional variation (Beal 1999. has been continuously in print since 1982. 2002). Linking identity and dialect through stancetaking Barbara Johnstone Carnegie Mellon University 1. in some places. and Pittsburghers and others go online to contribute to lists of local expressions and discuss what the dialect means for the community. where talk about local identity very often includes talk about the local dialect (Johnstone 2000a. sometimes refer to themselves as yinzers. as the variety is locally known. both in face- to-face interaction and in more public fora. When “authentic” Pittsburghers or Pittsburgh activities are described or parodied. yinz. refrigerator magnets. Pittsburghers in their 20s and 30s. One such place is the Pittsburgh. Souvenir vendors offer t-shirts and sweat shirts. Introduction1 Geographic mobility and new patterns of social interaction associated with the globalizing new economy have resulted at the same time in dialect leveling (Mil- roy 2002.

the other speaker. while others invoke the speaker’s own competence in the dialect. Stancetaking is thus one of the mechanisms through which dialect and local identity become linked in dis- course. As a result. This paper explores the role of stancetaking in one instantiation of one such practice: a conversation about local speech between a sociolinguistic fieldworker and two lifelong residents of the Pittsburgh area. who at first distances herself from speakers of the local dialect. she links her identity as a Pittsburgher with competence in the local dialect because it is rhetorically use- ful in this interaction to do so. to describe and evaluate the local dialect. These include performances of local dialect forms and other allu- sions to local identity. that they know enough. Repeated engagement in metalinguistic talk in which claiming the social identity of a competent dialect speaker is useful for epistemic stancetaking serves to strengthen and stabilize the idea that being a Pittsburgher means being able to speak the local dialect. and . Variationist sociolinguists in the Labovian tradition are coming to see “iden- tity” as a useful explanatory dimension in accounting for some patterns of lin- guistic variability and their role in language change. at the same time refiguring their role in presentations and representations of local identity (Johnstone and Andrus 2005). like that seen in Pittsburgh.50 Barbara Johnstone a new Pittsburgh literary magazine was named the New Yinzer. recasting herself as a local dialect speaker who knows local forms directly rather than by hearsay. from the appropriate sources.” Heightened dialect awareness. In the conversation. In other words. and in 2003 a student conceptual artist produced (removable) stickers with brief definitions of local terms and affixed them to bus-stop shelters and mailboxes in a project she referred to as “guerilla linguistics. Analysts of discourse in in- teraction in the tradition of Gumperz (1982) and Ochs (1992) find “stance” a useful explanatory category in accounting for how particular linguistic choices in interaction accomplish particular social and rhetorical actions. morphological. This strategy proves particularly effective in the interaction in question: the speaker who represents herself as a competent speaker of the dialect and who can perform local dialect forms gets the floor more often and is oriented to as an expert on Pittsburgh speech. and lexical choices. arises through dis- cursive practices that call attention to and normativize regional forms. Sociolinguists drawing on both traditions are beginning to explore how stancetaking can be accomplished through phonological. The micro-rhetorical interactional exigency that requires epistemic stancetaking drives the identity claim. Some such epistemic stancetaking moves invoke external sources of authority such as published lists of local words and examples provided by other people represented as authentic dialect speakers. eventually revises her identity claim. the Pittsburghers deploy several strategies for epistemic stan- cetaking: for making the implicit or explicit claim. that is.

Particular stances or social actions can then get linked indirectly to social identities such as gen- der categories (so that. in a particular sociocultural milieu. Others use “stance” to talk about the marking and claiming of in- terpersonal relations in talk as well. by particular individuals. a stance such as deference might become indexically linked with femaleness). for example. surprised or not). and syntactic variation with analyses of the real-time. With Eckert and Kiesling. For Du Bois . happy or sad about it.g. Linking identity and dialect through stancetaking 51 how sets of such choices can accrete into stances that index culturally meaningful styles or identities (Eckert 2000. Kiesling 2005). by virtue of the fact that performances or evocations of dialect competence can function as stancetaking moves. morpho-lexical. Conrad and Biber 2000). Hunston and Thompson (2000) operationalize stance as evaluation. and how variationists have adduced these concepts in accounting for patterns of variation across popula- tions. interactive design of talk in particular rhetorical situa- tions. so much so that the need for epistemic stan- cetaking can actually drive dialect-identity claims. in this case because adducing and performing the social identity of a competent dialect speaker is a powerful re- source for epistemic stancetaking. The text I use to illus- trate this is an extended transcribed extract from a sociolinguistic interview. whether they are sure or unsure about it. I begin by reviewing how the terms stance and identity have been used in recent work by interactional sociolinguists and conversation analysts that aims to account for what goes on in particular interactions. Identity and stance Almost all work about social interaction that adduces the idea of stance as an explanatory tool includes under stancetaking the moment-by-moment choices speakers make that index their relationship to what they say (e. 1994. particular linguistic forms directly index evidential stances such as certainty. or social actions such as apologizing. I show that stancetaking and identity are intertwined. I describe how dialect identity – one’s positioning as a user or nonuser of the local dialect – emerges as a rhetorical resource for and through stancetaking in a conversation meant to probe Pittsburghers’ ideas about the local dialect and explore how these ideas arise. Building on early work on epistemic and attitudinal stance (Biber and Finegan 1989.. I then show how dialect forms and regional identity can become linked through stancetaking. For Ochs (1992). interpersonal stances such as friendliness or intensity. This paper continues the work of exploring how connections between linguistic forms and social identities are shaped by interactional needs for stancetaking. 2. The interview was conducted in the course of a project that brings together analyses of regional patterns of phonological.

” as it is described in some of this literature. and gender categories. Like Goffman’s (1959) “presentations of self. Research on social interaction that adduces identity as an explanatory tool almost always includes under “identities” culturally circulating. Variability is an inevitable con- . Interactionists in the conversation analysis tradition take a similar approach. this body of literature pays little attention to the way in which identities can be associated with indi- vidually-embodied speakers (Johnstone 1996. As Bucholtz and Hall point out. with respect to any salient dimension of the sociocultural field”. the “sociocultural field” consists of two social actors and an object to which both are oriented. even if social actors think identities pre-exist interactions and sometimes predict how they will play out.” while others are more situation-specific roles or “ethno- graphically emergent cultural positions” (Bucholtz and Hall 2005: 585). achieved dialogically through overt communicative means. of simultaneously evaluating objects. For them.” in the interactional- sociolinguistic tradition. Alignment or dis- alignment with another social actor can be accomplished through membership categorization moves. (Perhaps because of a misplaced fear of being seen as locating social agency in the individual human. thus. and aligning with other subjects. “Stance is a public act by a social actor. “Casting into a category” can be accomplished through stancetak- ing. “Identities” thus include ethnic. and a repeated stancetaking move or pattern of moves may emerge as an identity. Conversation analysts stress the need to treat identity as interactional achieve- ment with consequences for the structure of the talk. normalcy. frequently adduced ways of cat- egorizing groups of people that are often oriented to as being relevant outside of and prior to the interaction as well as inside it.) But “identity work. includes interactional moves that could also be described as stancetaking. 3. tastes.” “identities.52 Barbara Johnstone (this volume: 163). 2001) is primarily focused on uncovering the mechanisms of language change. as well as “discourse identities” such as speaker or audience member. class. 2000b). posi- tioning subjects (self and others). and activities. “identities may be linguistically indexed through…stances” (2005: 585). identity arises in interaction: “for a person to ‘have an identity’…is to be cast into a category with associated characteristics or features” (Antaki and Widdi- combe 1998: 3). claims to social identity for oneself and ascriptions of identity to others fall under the rubric of stancetaking. Identity and stance in variationist sociolinguistics Labovian-variationist sociolinguistics (Labov 1994. and categorizations in terms of attributes such as deviance vs. Some identities are “macrolevel de- mographic categories. are social categories to which speakers orient as they become relevant in the interaction at hand.

Labov (1963) showed that the speakers who identified most closely with tradi- tional local ways of living were most likely to use the more conservative. “Social class” was thus reconceptualized in terms of social activity and from the perspective of speak- ers’ everyday experience of language. so in order to understand how language change is likely to proceed. local-sounding forms because such forms carry “covert prestige” as signals of working-class solidarity. In the research settings in which most early work was done. however. From the beginning. in Labov’s own work as well as in others’. the answers that were proposed suggested that phonological variability could be a resource for in- dexing attitude and affiliation. Linking identity and dialect through stancetaking 53 comitant of change. for example. The classic Labovian account uses the correlational methods of quantita- tive sociology to model how “social facts” about speakers. because dense. James and Lesley Milroy’s work in Belfast (J. accordingly. might males use less “prestigious” forms than females? What might make a working-class speaker adopt more local-sounding forms? Although not framed in terms of stancetaking or identification. aimed at what speakers think of as more prestigious variants. such as socio-economic status and sex. LePage and Andrée Tabouret-Keller (1985). were suggesting that choices among variants in a speaker’s repertoire could be thought of as “acts of identity. . sometimes become relatively consistent. we need to understand how variability arises and what predicts its out- come. account for patterns of variation in groups. This suggested that the more local-sounding forms might in fact be part of the process through which local identity was claimed. Peter Trudgill’s (1972) equally influential study of sex-correlated differences in Norwich. L. older. Milroy 1987) explored the utility of social network theory for uncovering and explaining patterns of variation. less standard-sounding way. Why. of talking. multi- plex social networks are effective enforcers of local norms. presumably because the variants in question were more “prestigious. upper-class and female speakers consistently tended to use more of certain variants than lower-class and male speakers did.” and that speech communities and linguistic varieties could be seen as ways of labeling the fact that ways of identifying and. R. At the same time. less standard-sounding variants of certain phonological variables. and the people they knew also knew each other) were relatively likely to talk in a more local-sounding. more particular questions were being asked about why these kinds of patterns might exist. which gives rise in many cases to more careful speech.” Variation within an individual’s speech was linked with self-consciousness. Milroy 1992. B. concerned with describing the complicated linguistic situation in the Caribbean. People with relatively many and overlap- ping ties to their neighbors (such that they knew the same people in multiple roles. England suggested that men may use more non-standard. In his groundbreaking study of Martha’s Vineyard.

54 Barbara Johnstone Variationists doing quantitative.’” exploring how shifts from a speaker’s “na- tive” language or variety to and from one clearly associated with another group can accomplish shifts in stance. Speakers who do this are more likely than others to think highly of Texas as a place to live. a speaker self-consciously demonstrating the local pronunciation of an expression he thinks of as particularly local performs the role (Schilling-Estes does not use the term identity) of a stereotypical local person with a more vernacular accent than his own. For example. where sounding Southern is a resource more or less “natively” available to many people. Hazen (2000) finds that North Carolina speakers with ties to institu- tions and cultural characteristics from outside their county are more likely to shift toward standard-sounding pronunciation in formal speech (perhaps because they are more aware of the stigmatization of local speech). Guy Bai- ley and his colleagues (Bailey et al. Rose (2003) has found that among older people from the upper Midwest. or place of residence. Ben Rampton (1995. 1993) show that. finding that identity (vari- ously operationalized) sometimes correlates more closely with variation than do variables such as age. arguably as badges of local identity. while people with more lo- cal identities are linguistically more consistent.” but rather part of a speaker’s native repertoire (Coupland 2001. as in [ra:t] for right. for example. island).S. Stancetaking work can also be accomplished through shifts into and from ways of speaking that are not so clearly “other. For example. 1999) has studied “lan- guage crossing” or “styling the ‘other. depending on the speaker and the situation. moves toward more Southern- accented speech range. correlational work about speech commu- nities have started to add “identity” to the lists of socio-demographic variables they use to account for patterns of phonetic variability. Schilling-Estes suggests that such performances may be more patterned than variationists have assumed and may reveal facts about dialect awareness than can affect the course of change. Rampton (1995) describes how brief excursions into Punjabi by London adolescents of Anglo and Afro-Caribbean as well as In- dian and Pakistani descent can signal oppositional footing in interactions with adults or key a joking stance in playground interaction. in Texas. occupation. linguistic changes that diffuse from rural to urban settings typically involve the reassertion of tra- ditional speech norms. One such feature is the monophthongization of /ay/ before voiceless obstruents. Rampton 2003). interactional sociolinguists have explored how choices among forms associated with different languages and dialects can signal attitude and affiliation. In Texas. the pronunciation of /θ/ and /ð/ as [t] and [d] is not only correlated with rural residence but overtly commented on as a symbol of a “country” identity. Turning to stancetaking. Schilling-Estes (1998) shows that in Ocracoke (an Atlantic-coast U. from fairly automatic style shifts correlated with register to quite self-conscious rhetorical .

and others come to be. I exemplify how stancetaking. suggesting how correlations like the ones described by Bailey. in many contexts and media. the features get used repeatedly together. 2002. McCarthy 2004. dressing in a particular style. 2000) work in a Detroit high school suggests how the variationist and interactionist approaches to stance. Hazen. Pittsburghese in conversation: Identity as a stance resource Both linguists and non-linguist Pittsburghers associate a distinctive set of linguis- tic features with southwestern Pennsylvania (Brown 1982. For Eckert. is par- ticularly likely to be relevant for members of subordinate groups. by linguists and laypeople alike. Johnstone and Baumgardt 2004. just as is cruising in cars. fre- quently in connection with talk about local identity (Johnstone 1999b. as when “Southern Belle”-sounding speech is used in footing-shifts meant to manipulate men (Johnstone 1999a). and Pittsburghers talk about this dialect often. For example. not simply a reflection of already-existing differentiation. that of the recent immigrant. linked with a repeatable social identity. and phonology might be linked. or they may be linked with other sources of identity. Rose. When the identity in question is regional. the style that indexes it is often referred to as a dialect. identity. Johnstone et al. derived from a derogatory term for non-whites). Hankey 1965.” which. Kiesling (2005) shows that a particular set of morpho-phonological features that co-occur in the English of recent immigrants in Australia work together to project a face-saving epistemic/ interactional stance of “authoritative connection. and a repeatable style (locally called wogspeak) emerges. Particular linguistic features available in a speaker’s sociolinguistic environment can be used for stancetaking. they may have to do with local social categories (as with Eckert’s jocks and burnouts). or do- ing some things rather than others in school. Linking identity and dialect through stancetaking 55 moves. Eckert’s and Kiesling’s work shows how linguistic variants can become indexically linked to social identities through stancetaking. the choice between one variant and another is part of the semiotic activity in which social identities are created. and sets of co-occurring stancetaking features can come together as styles that index identities. Kiesling and Wis- . adopting certain variants of vowels is one way of adopting a stance toward and participat- ing in local activity. Andrus. identity. Because they work together as a stancetaking strategy. 1972. Kiesling claims. and linguistic variation are linked with ref- erence to how ideas about what counts as a dialect are negotiated and circulated in interaction. and Danielson 2006. In what follows. McElhinny 1999). Scott F. Johnstone. Penelope Eckert’s (1989. 4. 2004. Johnstone et al. The identities linked with linguistic styles may be ethnic (as in Kiesling’s wogspeak.

shows that it is now more common in the speech of older. with out realized as [a:t]. Simultaneous talk is linked with square brackets. Insiders. makes routine use of a number of phonological variables that make her sound local. Italics mark loud or otherwise stressed words. They offer examples of what Pittsburgh speech sounds like and argue about which forms are authentically local and how local forms should sound. in discursive contexts such as nostalgic talk by ex-Pittsburghers and identity work by newcomers. She fronts /u/ in words like move and sometimes vo- . Jen R. and. In the interview extract that we will examine in detail. equal signs indicate “latched” talk.. Johnstone and Baumgardt 2004). The dialect appears to be receding: research on one feature. seldom noticed or commented on local speech. and Danielson 2006. for example. BJ is the fieldworker. as it is locally known. co-workers and I have conducted over 100 sociolinguistic interviews in four Pittsburgh-area neigh- borhoods. pronounc- ing job as [jfb]. throughout the interviews and other research tasks. I have made notes about their accents in the right-hand col- umn of the transcript rather than using the IPA transcription or respelling in the text. To keep the transcript readable and avoid caricaturing Jen and Donna by means of eye dialect. Her pronunciation of /aw/ is sometimes monophthongized. the population that might be expected to use a receding feature longest (Kiesling and Wisnosky 2003. stylized way (Johnstone. The interview protocol elicits explicit talk about Pittsburghese. by which I mean close. working-class speakers. a woman in her 40s and her 13-year-old daughter talk about Pittsburgh speech in an explicitly nor- mative way. the mother. they use claims about and performances of their own speech as ways of establishing the authority to describe the dialect. for the calling to attention of linguistic difference that creates heightened awareness of regional varieties. My analytical method in this paper is discourse analysis. linking epistemic stance with dialect identity. As one phase in a larger study of Pittsburgh speech. McCarthy 2004). out-migration caused by the collapse of the steel industry and in-migration caused by the growth of the educational and health-care sectors provided increased opportunities. and Pittsburghers started to use certain local speech features to point to local identity in a more reflexive. Beginning in the 1960s. acquired early in life in face-to-face encounters.56 Barbara Johnstone nosky 2003). people often break into spontaneous performances of the dialect. Andrus. in daily contact with people who sounded the same as they did. systematic reading of a small amount of text (Johnstone 2002). Local-sounding talk. /aw/- monophthongization. The area thus lends itself to a study of how dialect awareness arises in a variety of metalinguistic genres of talk. however. She merges and rounds the low back vowels (LBV). once identified people as Pittsburghers only to occasional outsiders who noticed the dialect. As they do this.

] =What do you think it is? I mean is it-? 11 jr I think it’s the way we say words. She reduces the diphthong /ay/ to a more monophthongal form when it is followed by /l/ (which is vocalized). but in fact they sound very much the same in this segment as elsewhere. /aw/ in out (twice). I’ve heard of [Pittsburghese.] definitely. Yeah. [there’s] 3 bj [What.her husband’s job took them out of state and to many in moved.] 4 jr that store over on the Southside in um. As do most sociolinguistic interviews in the Labovian tradition (Labov 1984). rounded low back vowel (LBV) in job. [Station Square that has diphthongal /aw/ the] Pittsburghese shirts in Southside 5 dr [((Breath intake)) Yeah] 6 bj Uh huh= 7 jr =In fact. other states. vocal- izes /l/. I remember when my friend Karen moved out of state. have you ever heard of Pittsburghese? 2 jr Oh yes. this one included modules on topics meant to elicit a range of levels of self-consciousness. she does not monophthongize /aw/ but fronts /u/ and /o/. fronted /u/ with. I mean. This is another fairly common local-sounding variant. So. as when while is realized as [wa:u]. My summary of Jen and Donna’s accents is based on the whole interview. as well as unrecorded talk in other contexts. Linking identity and dialect through stancetaking 57 calizes /l/. Yeah. and merges and rounds the low back vowel. 12 dr [((laughs))] 13 bj [Yeah?] 14 jr I think it’s how we say “downtown” and um= diphthongal /aw/s in downtown 15 dr = “down[town”] diphthongal /aw/s 16 jr [“South]side” and. I remember sending her a couple Pittsburghese monophthongal  shirts for them. The topic of accent may well have made them self-conscious about their speech. “dropped g” in sending 8 dr Hm 9 jr Umm. diphthongal /aw/ in Southside 17 dr “Y’all” . Her daughter Donna has a less local accent.= 10 bj [Mm hmmm. (1) FH01 (“Jen R”) and FH02 (“Donna R”) Pittsburghese 1 bj Um.

” That’s. It’s funny. I’m really fronted /u/ in two.] constantly. um. There’re not local pronunciations of wash.that’s what they tell me. And like. 29 bj Huh! 30 jr “Younz” is more a Pittsburgh thing than [“y’all. /o/ less fronted than previously in over 45 jr [Mm hmm] 46 bj [Mm hmm] 47 jr Yeah. what other ones can you think [yHnz] of? 22 dr Just “y’all” and “yinz.] 43 jr [but] I’m just thinking. They say [it’s a Pittsburghese] 27 bj [And that’s a Pittsburgh thing?] 28 dr It. iron 19 dr “Yinz” [yHnz] 20 jr [Just the uh. He’s a year older than me.” You hear LBV not rounded “yinz” a lot.] But I hear it a lot from [them LBV not rounded when] I’m over there in lot.] 33 dr [I was thinking] Southern. Yeah “yinz.” vocalized /l/ in like “Yinz wanna do somethin’?” or [like] ((laughing)) you know older so rounded LBV in [I hear that.” is.58 Barbara Johnstone 18 jr “wash” and “iron” and different words and the way Pittsburgh “standard. know 38 bj [Mm hmm] 39 jr [We don’t use that. And then.”] [y~nz] 31 dr [yeah] 32 jr “Y’all”’s more like a Georgia. good friends with their son. fronted /o/ in both. do. [but ] 34 jr [Yeah] 35 dr they still say “y’all” to me.] 40 dr Yeah I never said you [used it] but 41 jr [Yeah. 25 jr “Y’all?” 26 dr Yeah. your dad and I don’t use that [too often. in lot 36 bj You do? In- 37 dr Yeah well like our neighbors like two doors down. ‘cause like both his parents say “yinz. 23 bj [“Y’all?”] 24 dr [Drives] me crazy.] 44 dr [No.like] 21 bj [“Yinz” is another one. like. [Uh huh. they say “y’all” to me. [Southern. .] What.] 42 bj You don’t use that. I know. that’s the most my friends always are saying “y’all” to me. he says “yinz” constantly.

I [say “iron”] iron = [arn] 56 dr [I do?] 57 jr for “iron. You start to say then.] 50 dr you’re around people so often. 54 bj Uh huh [(probably you-)] 55 jr [You s-] you do. but you don’t reali-. “Okay. Linking identity and dialect through stancetaking 59 48 dr And you pick up on it.you wouldn’t say you use any of the. “citation” Southside 72 dr “Southside” Monophthongal /aw/ . I don’t think so. “performance. 59 dr [“iron”] [arn] 60 jr [I] say “wash w. Oh. No.you don’t.” Southside.I started to say “yinz” diphthongal /aw/ to people. I don’t pronounce my words as clearly as-. you start.” in around ((skeptical. I [mean] [w~š] [wfš] [wfš] 61 dr [“wash”] [wfš] 62 jr I don’t. Pittsburghese things? 53 dr Not really. [once] 49 jr [Sure.” iron = [ay6rn] 58 bj Uh huh. ((laughing)) And they’re looking at me like. I know I do. amused voice)) 51 dr Mm hmm 52 bj So.wash” for “wash”.” monophthongal /aw/ and assimilated /θ/ in first. would. exaggerated diphthongal /aw/ and /θ/ in second. 63 dr “wash” [wfš] 64 jr or the accent’s [on] 65 bj [uh huh] 66 jr a different [part] 67 dr [“wash”] [wf] 68 bj Uh-huh 69 jr Yeah 70 dr “wash” [wfš] 71 jr “Southside” instead of “Southside.

in different cities.] 90 dr [And then she’s like] “Pennsylvania she’s like. diphthongal /aw/ right? in South 83 jr “You’re from Pennsylvania=”. she was from the Pittsburgh area.” And she’s like. I remember one time. fronted /u/ in two. how. we were in South Carolina diphthongal visiting my.” /aw/ monophthongal /aw/s in first.”] Pennsylvania 91 bj [How did she know?] How do you think she knew? 92 jr She. “You guys from Pennsylvania?” We’re like “Yeah. “citation” downtown 74 dr “downtown” diphthongal /aw/s 75 jr I know I. intonation. are you from the Pittsburgh area?” 86 dr Yeah. 89 jr [Yeah. “performance. right?” And we’re like. my uncle and my two cousins and my aunt.60 Barbara Johnstone 73 jr I say “Southside. and we were talking about how like the vocalized /l/ in South kind of moves slow. went to (s. moves. 79 bj [Mm hmm] 80 dr They’ll.you’re [definite-]” 85 jr [“Are you-]. exaggerated diphthongal /aw/s in second. she said she originally] .they’ll say “You’re from Pittsburgh. ] [God. /l/ not vocalized in Pennsylvania 84 dr “=Yeah. 76 bj Mm hmmm. vocalized /l/ in “Oh. She didn’t grow up in didn’t = [dι’6n] Pitt. diphthongal /aw/ in about. I mean. “We’re from Pittsburgh.” And question” she’s like “You guys wouldn’t happen to be from Pittsburgh. too.] 82 dr [When we were in] South Carolina.” downtown.use a lot of Pittsburghese. okay. She said. I know I kn. [you know?] uncle.” monophthongal “Downtown” instead of “downtown. ((laughing)) 87 bj Does that happen to you.some) store. or=? 88 dr =Yeah. they’ll immediately [say.[yeah. 77 dr I [probably do] and I don’t realize it fronted /u/ in do 78 jr [I know I do.] Well when I’ve been in [different] states.” 81 jr Yeah. it drives you crazy. I can [tell by your accent. monophthongal /aw/ in South. And we /aw/ in South. you’re.

Linking identity and dialect through stancetaking 61

93 dr [Yea-, I thought she was] from Ohio.
94 jr Yeah, [but she she was,]
95 bj [(Eastern)] Ohio?
96 jr Right [but she]
97 bj [This area?]
98 jr she was in Pittsburgh and then they moved to Ohio, fronted /u/ in
99 dr Yeah
100 jr and then from Ohio they moved to South Carolina. fronted /u/
in moved;
/aw/ in South
101 dr [She didn’t have the] accent either. She still had like a Pittsburgh
102 jr [She said that-] Mm hmm. Yeah. She didn’t [have] didn’t = [dι’6n]
103 bj [Mm hmm]
104 jr a Southern accent at all. She was, like a mom. [You know,] rounded LBV in
105 bj [Mm hmm]
106 jr I would say she was in her early 30s at least.
107 dr Yeah.
108 jr [Uhh maybe 40s.]
109 dr [ I don’t really remember]
110 jr I don’t know. I- I’m a bad judge of age. But, she was from, eh or glide-reduced /ay/
grew up in Pittsburgh for a while, moved to Ohio, and then was in I; vocalized /l/
in South- then they relocated to South Carolina. in while; fronted
/u/ in moved;
/aw/ in South
111 dr [yeah]
112 jr [That’s] how she knew. ‘Cause she even said that Pittsburgh
accent, when you’re not a-round it, when you do hear it, you diphthongal /aw/
really pick up on it fast. in around

Epistemic stancetaking and dialect identity come into play here in a number of
ways. At the beginning of the extract, Jen claims the authority to speak on the
topic of Pittsburghese with reference to knowing about Pittsburghese shirts and
sending one to a friend who has moved away. The shirts Jen is referring to, pro-
duced largely for the tourist and local-nostalgia markets, feature words spelled in
ways that suggest their “Pittsburgh” pronunciation; on the back, there may be a
dictionary-like list of words and phrases thought to be local. Epistemic stancetak-
ing is independent of dialect identity here. Referring to Pittsburghese shirts is a
way of arguing from external authority, a resource that is potentially available
whether or not one is a speaker of the dialect. Jen supports her epistemic claim
(Oh yes [I’ve heard of Pittsburghese], line 2) with reference to indirect, mediat-

62 Barbara Johnstone

ed knowledge about the dialect – she has seen it on t-shirts. She maintains this
relatively detached epistemic stance for another turn (line 9), Yeah, I’ve heard of
Pittsburghese, definitely. I’ve heard of locates the epistemic source elsewhere, in
what other people say. Then, however, in response to my question What do you
think it is?, Jen switches to a different mode of evidence, taking up my invitation
to adopt an epistemic stance rooted in personal authority (I think) and aligning
herself with other competent speakers of the dialect (we), I think it’s the way we
say words.
Jen then begins to list some of these (lines 14–18), downtown, Southside, wash,
iron. While continuing to locate the source of knowledge in her own competence
(I think), she disaligns somewhat from dialect speakers and returns to a more dis-
tanced mode of epistemic stancetaking that does not rely on competent-speaker
dialect identity. The citation forms she produces are not the local pronunciations;
anyone who has read or heard about Pittsburghese could produce them, whether
or not they knew how they sounded when pronounced by someone with a local
accent. Jen pronounces /aw/ as a diphthong in downtown and Southside, using the
less local-sounding variant. She also pronounces wash and iron in the standard
ways, rather than in the local-sounding variants [w~š] or [worš] and [arn]. Nor
do these citation forms fully reflect what Pittsburghers usually imagine is local
about “the way we say [these] words.” Downtown is typically spelled “dahntahn”
on artifacts like t-shirts, the spelling suggesting that the monophthongization of
/aw/ is to be attended to, whereas Southside is often spelled “souside,” with a diph-
thongal /aw/ but a deleted or assimilated /θ/.
In lines 71 and 73, Jen returns to two of these words, contrasting what she
represents as their standard pronunciation with the way she claims to say them. In
citing the “standard” forms, she exaggerates the diphthongs in both words and the
/θ/ in Southside. In her performance of her own pronunciation, she overdoes what
popular local spellings suggest are the local pronunciations, monophthongizing
the /aw/ in both words rather than just in downtown. Here, Jen claims an authori-
tative stance in two ways. In citing examples of Pittsburghese in their standard
pronunciations, she is doing something that either a speaker or a non-speaker
of the dialect could presumably do, assuming he or she had access to lists of lo-
cal forms like those on t-shirts or folk dictionaries. In this activity, authoritative
stance is independent of dialect identity. But Jen also performs the local pronun-
ciations, an activity that indexes the dialect identity of a competent speaker. She
also points to this competent-speaker identity repeatedly in more explicit claims
to be an actual user of the dialect, We don’t use that…I’m just thinking…your dad
and I don’t use that too often (lines 39–43), I know I [use the Pittsburghese things]
(line 55), I don’t pronounce my words as clearly…(line 62), I know I, I know I kn-
use a lot of Pittsburghese (line 75). Note how in this final extract Jen starts to say

Linking identity and dialect through stancetaking 63

she knows Pittsburghese, which could signal second-hand access to the dialect,
but revises the claim to I use a lot of Pittsburghese, explicitly claiming to speak it.
To summarize, Jen makes epistemic stancetaking moves throughout the con-
versation. Some of these involve displaying familiarity with external sources of
authority such as Pittsburghese shirts. Some of these moves involve performances
of this knowledge, in the form of citations of local forms in a standard-sounding
way. Other stancetaking moves involve direct claims to competent-speaker dia-
lect identity. Sometimes, as we saw above, performances of competent-speaker
dialect identity are embedded in these claims, I say [arn] for [ay6rn] (lines 55–57)
and I say [w~š] for [wfš] (line 60).
Donna, the 13-year-old, tries to participate in all these activities. At first, her
epistemic stancetaking is marked by moves that distance her from dialect speak-
ers and locate the source of her knowledge about the dialect in others. Invited,
like her mother, to talk about what she thinks Pittsburghese is, she talks about
what other people say it is. In line 17, she suggests an addition to the list Jen is
building, y’all, then, after there is no uptake from her mother or me, another in
line 19, yinz. I acknowledge this contribution in line 21 and encourage her to offer
more. She repeats y’all and yinz in line 22, then explicitly adduces the source of
her epistemic authority on the topic of local speech, my friends always are saying
“y’all” to me and They say it’s a Pittsburghese [thing] (line 26). When she continues
to be met with skepticism, she makes the same stancetaking move again: It- that’s
what they tell me (line 28). These stancetaking moves are not linked to compe-
tent-speaker identity – Donna would have access to this source of knowledge
whether or not she claimed to be a speaker of the dialect herself – but rather to
external authority. Her mother then rebuts Donna’s externally-based claim with
a dialect-performance move: arguing that y’all is not really “a Pittsburgh thing,”
she pronounces yinz not in the stereotypical version represented on t-shirt lists,
which would be [yHnz], but in an older, more traditional-sounding way, [y~nz].
She then supplements this with a more distanced move referring to presumably
widespread knowledge that does not require dialect identity, Y’all’s more like a
Georgia, Southern [thing] (line 32). Donna continues to argue that y’all is local, but
continues to disalign herself from the local way of talking, contrasting I with they
and positioning herself as the recipient of local speech rather than its initiator, I
was thinking Southern, but they still say “y’all” to me (lines 33–35). But her mother’s
competent-speaker knowledge apparently trumps Donna’s external knowledge:
Donna retreats to a discussion of yinz, which everyone in the interaction agrees is
local, Yeah “yinz.” You hear “yinz” a lot (line 35). Using you hear rather than I hear,
she aligns herself, if not with dialect speakers, at least with a group larger than
herself. She then provides an extended illustration of her claim to hear yinz a lot,
which includes a dialect performance (“Yinz wanna do somethin’?” line 37). Note

64 Barbara Johnstone

that this is not the same sort of dialect performance as Jen’s have been: Donna is
imitating other people, not making a claim about her own dialect identity. A per-
formance like this displays local knowledge (she knows how local speech sounds)
but falls short of a claim to being a speaker of the dialect herself.
As the conversation proceeds, Donna begins to supplement epistemic moves
that appeal to external sources with her mother’s interactionally more successful
mode of stancetaking by evoking a competent-speaker dialect identity. In her first
claim to actually being a dialect speaker, Donna frames her competence as an
unintentional and uncharacteristic consequence of being around dialect speak-
ers, .. you pick up on it. You start to say then once you’re around people so often, you
start- I started to say “yinz” to people. ((laughing)) And they’re looking at me like,
“okay” ((skeptical, amused voice)) (lines 48–50). In answer to my direct question,
however, she then explicitly disaligns herself from other dialect speakers, Not re-
ally. No, I don’t think [I use any of the Pittsburghese things] (line 53). Her mother
steers her toward alignment with dialect speakers, You do, but you don’t real[ize
it] (line 55).
Donna appears to take the hint. She begins to reframe her dialect identity in
such a way that it becomes useful in epistemic stancetaking, the way Jen’s dia-
lect identity is. One revealing segment begins in line 60, where Jen, listing and
performing local forms, makes and tries to illustrate a claim about how she says
wash. In an apparent performance error, she almost confuses the “correct” form
with the “Pittsburghese” form, so that the second time she says the word it sounds
like the standard [wfš] but is apparently meant to be an improved performance
of what Jen represents as the local pronunciation, [w~š]. Donna, who has just
claimed that she does not use Pittsburghese things (line 53), then starts to repeat
the word over and over in lines 61, 63, 67, and 70, in a low voice, apparently try-
ing to imitate the local pronunciation so as to contrast it with her own. But since
Donna picks as her target Jen’s second performance, which was actually the more
standard-sounding variant, Donna seems to conclude that her own pronuncia-
tion is in fact the local one. So after “trying out” Southside and downtown in a
similar manner, she explicitly claims the identity of a dialect speaker in line 77,
echoing her mother’s earlier wording, I probably do [use Pittsburghese] and I don’t
realize it.
Shifting identity in this way means that Donna can now adopt the epistemic
stance of an actual dialect speaker, which her mother has been drawing on, and
she does this in co-narrating the story about the family’s having been identified as
Pittsburghers by their accents. This begins as an explicit claim, co-constructed by
Jen and Donna, to the identity of a recognizable dialect speaker:

Linking identity and dialect through stancetaking 65

78 jr [I know I do.] Well when I’ve been in [different] states, in different cities,
79 bj [Mmhmm]
80 dr They’ll- they’ll say “You’re from Pittsburgh.”

Jen then claims she is also recognized by her accent, Yeah, they’ll immediately say
“You’re from Pennsylvania” (lines 81–83). The ensuing narrative, which involves
densely overlapped joint production by Jen and Donna, supports their now mutu-
al claim to competent-speaker dialect identity. It culminates with Donna’s voicing
of her family (we) and a woman they met in the South (she): And we’re like, “We’re
from Pittsburgh.” And she’s like, “Oh, okay. I can tell by your accent” (line 90).

5. Discussion

I began this paper by noting that, in Pittsburgh, local identity and local dialect are
often linked, and by asking how such links are forged. How does being a Pittsbur-
gher get associated in so much popular discourse with speaking “Pittsburghese”?
I have explored one way this can happen: if people are talking about local speech,
then it can be interactionally useful to claim the identity of a local dialect speaker,
because doing so provides one with resources for epistemic stancetaking.
This conversation illustrates how both dialect identity and epistemic stan-
cetaking arise in interaction, in response to particular prompts (such as my Have
you ever heard of Pittsburghese? What do you think it is?) and more general inter-
actional exigencies such as wanting to get the floor. The two are intertwined in a
particularly visible way here. Since both the interactional genre (the interview) and
the particular topic called for knowledge claims and displays of the authority to
make such claims, epistemic stancetaking was an interactional requirement. Since
the topic was local speech, claims about and performances of competent-speaker
dialect identity were a particularly useful way to make epistemic stancetaking
moves. To get a sense of how stancetaking works in the conversation, I explored
both explicit claims to epistemic authority and indirect claims to such authority
via citations of local words and sounds. To see how and when dialect identity
becomes relevant in the conversation, I described explicit moves that characterize
the participants as speakers of the dialect (I talk that way) or not (I don’t really use
Pittsburghese things), and indirect claims to local-speaker identity through per-
formances of the local accent. As we have seen, epistemic stancetaking moves in
this conversation are often scaffolded on allusions to and performances of dialect
identity. The micro-rhetorical exigencies that require stancetaking can be seen to
drive identification moves, as when Donna recasts her identity in order to assume
a more authoritative epistemic stance.

66 Barbara Johnstone

That the topic of local speech came up in this case is not surprising: I brought
it up, as a module in a sociolinguistic interview. But the topic comes up nowadays
in many ways. As I have shown in this paper, once the topic arises, dialect can
become linked with local identity via the interactional usefulness of representing
oneself as a speaker of the dialect. It should be stressed, however, that not every-
one has the same kind of access to this resource. There is an important sense in
which Jen has a stronger local accent than Donna does, in part for linguistic and
cognitive reasons that are not related to identification or stancetaking. Claim-
ing to be a speaker of the local dialect is not the same as being one in the sense
linguists usually have in mind; it does not require anything more than knowing a
few local-sounding words. A person like Jen, who can perform local pronuncia-
tions, may have interactional resources in certain contexts that Donna, who can
only say she speaks the dialect, lacks. Discursive activities like the one examined
here give the upper hand to more competent speakers of the local dialect. This
means that competent dialect speakers like Jen have the stancetaking advantage
in this conversation and ones like it. It would be an oversimplification, however,
to suppose that its usefulness in discursive activities like this will automatically
contribute to the maintenance of the dialect in the face of powerful homogenizing
pressures. That is a hypothesis that remains to be tested.


1. Work on this project was partially supported by the National Science Foundation, Award
BCS-0417657. Jennifer Andrus helped with the transcription, and other co-workers on the
Pittsburgh Speech and Society project, in particular Scott F. Kiesling, have helped by discussing
earlier iterations of this paper with me. Comments on a different version of this chapter by fel-
low participants in the 10th Biennial Rice University Linguistics Symposium, “Stancetaking in
Discourse,” as well as comments on this version by Robert Englebretson, have been most helpful.
I am especially grateful to “Jen,” “Donna,” and other members of their family for their generosity
in talking to me and providing other kinds of invaluable assistance with the project.


Antaki, C. and Widdicombe, S. (eds.). 1998. Identities in Talk. London: Sage.
Bailey, G., Wikle, T., Tillery, J. and Sand, L. 1993. “Some patterns of linguistic diffusion.” Lan-
guage Variation and Change 3: 359–390.
Beal, J.C. 1999. “‘Geordie nation’: Language and regional identity in the north-east of England.”
Paper presented at Methods X. St. John, Newfoundland, Canada, August 1–6.
Biber, D. and Finegan, E. 1989. “Styles of stance in English: Lexical and grammatical marking
of evidentiality and affect.” Text 9: 93–124.

Linking identity and dialect through stancetaking 67

Biber, D. and Finegan, E. (eds.). 1994. Dimensions of Register Variation. New York: Oxford
University Press.
Brown, C. 1982. A Search for Sound Change: A Look at The Lowering of Tense Vowels before
Liquids in the Pittsburgh Area. MA thesis, University of Pittsburgh.
Bucholtz, M. and Hall, K. 2005. “Identity and interaction: A sociocultural linguistic approach.”
Discourse Studies 7: 585–614.
Conrad, S. and Biber, D. 2000. “Adverbial marking of stance in speech and writing.” In Eval-
uation in Text: Authorial Stance and The Construction of Discourse, S. Hunston and G.
Thompson (eds.), 57–73. New York: Oxford University Press.
Coupland, N. 2001. “Dialect stylization in radio talk.” Language in Society 30: 345–375.
Dubois, S. and Horvath, B.M. 2002. “Sounding Cajun: The rhetorical use of dialect in speech
and writing.” American Speech 77: 264–287.
Eckert, P. 1989. Jocks and Burnouts: Social Categories and Identity in High School. New York and
London: Teachers College Press.
Eckert, P. 2000. Linguistic Variation as Social Practice. Oxford: Blackwell.
Goffman, E. 1959. The Presentation of Self in Everyday Life. Garden City, NY: Doubleday An-
chor Books.
Gumperz, J.J. 1982. Discourse Strategies. Cambridge: Cambridge University Press.
Hankey, C.T. 1965. “Miscellany: ‘Tiger,’ ‘tagger,’ and [ai] in western Pennsylvania; Diphthongal
variants of [e] and [æ] in western Pennsylvania.” American Speech 40: 226–229.
Hankey, C.T. 1972. “Notes on west Penn-Ohio phonology.” In Studies in Linguistics in Honor
of Raven I. McDavid, Jr., L.M. Davis (ed.), 49–61. University, AL: University of Alabama
Hazen, K. 2000. Identity and Ethnicity in the Rural South: A Sociolinguistic View Through Past
and Present Be. [Publications of the American Dialect Society 83] Durham, NC: Duke
University Press.
Hunston, S. and Thompson, G. (eds.). 2000. Evaluation in Text: Authorial Stance and the Con-
struction of Discourse. New York: Oxford University Press.
Johnstone, B. 1996. The Linguistic Individual: Self-expression in Language and Linguistics. New
York: Oxford University Press.
Johnstone, B. 1999a. “Uses of Southern speech by contemporary Texas women.” Journal of So-
ciolinguistics 3: 505–522. [Theme issue on Styling the ‘other,’ Ben Rampton (ed.)]
Johnstone, B. 1999b. “How to speak like a Pittsburgher: Representations of speech and the
study of linguistic variation.” Plenary address presented at the Sociedad Argentina de
Lingüística, Mar Del Plata, Argentina.
Johnstone, B. 2000a. “Representing American speech.” American Speech 75 (Diamond Anni-
versary Essays): 390–392.
Johnstone, B. 2000b. “The individual voice in language.” Annual Review of Anthropology 29:
Johnstone, B. 2002. Discourse Analysis. Malden, MA: Blackwell..
Johnstone, B., Andrus, J., Baumgardt, D., Schardt, A.M. and Kiesling, S.F. 2004. “Whose social
meaning? Pittsburgh monophthongal /aw/ in perception and production.” Paper present-
ed at New Ways of Analyzing Variation 33. Ann Arbor, MI, September 30–October 3.
Johnstone, B., Andrus, J., and Danielson, A.E. 2006. “Mobility, indexicality, and the enregister-
ment of ‘Pittsburghese’.” Journal of English Linguistics 34: 77–104.
Johnstone, B. and Baumgardt, D. 2004. “‘Pittsburghese’ online: Vernacular norming in conver-
sation.” American Speech 79: 115–145.

68 Barbara Johnstone

Johnstone, B., Bhasin, N. and Wittkofski, D. 2002. “‘Dahntahn Pittsburgh’: Monophthongal
/aw/ and representations of localness in southwestern Pennsylvania.” American Speech 77:
Kiesling, S.F. 2005. “Variation, stance, and style: Word-final -er, high rising tone, and ethnicity
in Australian English.” English World Wide 26: 1–44.
Kiesling, S.F. and Wisnosky, M. 2003. “Competing norms, heritage prestige, and the decline
of /aw/-monophthongization in Pittsburgh.” Poster presented at New Ways of Analyzing
Variation 32, Philadelphia, PA, October 9–12.
Labov, W. 1963. “The social motivation of a sound change.” Word 19: 237–309.
Labov, W. 1984. “Field methods of the project on linguistic change and variation.” In Language
in Use: Readings in Sociolinguistics. J. Baugh and J. Sherzer (eds.), 28–66. Englewood Cliffs,
NJ: Prentice-Hall.
Labov, W. 1994. Principles of Linguistic Change: Internal Factors. New York: Blackwell.
Labov, W. 2001. Principles of Linguistic Change: Social Factors. Malden, MA: Blackwell.
LePage, R.B. and Tabouret-Keller, A. 1985. Acts of Identity: Creole-Based Approaches to Lan-
guage and Ethnicity. Cambridge: Cambridge University Press.
McCarthy, C. 2004. “Language change in Pittsburgh: The decline of /aw/-monophthongization
and the Canadian shift.” Poster presented at New Ways of Analyzing Variation 33, Ann
Arbor, MI, September 30–October 3.
McElhinny, B. 1999. “More on the third dialect of English: Linguistic constraints on the use of
three phonological variables in Pittsburgh.” Language Variation and Change 11: 171–195.
Milroy, J. 1992. Linguistic Variation and Change. Oxford: Blackwell.
Milroy, L. 1987. Language and Social Networks, 2nd ed. Oxford: Basil Blackwell.
Milroy, L. (ed.). 2002. “Investigating change and variation through dialect contact.” [Theme is-
sue]. Journal of Sociolinguistics 6: 3–130.
Ochs, E. 1992. “Indexing gender.” In Rethinking Context: Language as an Interactive Phenom-
enon, A. Duranti and C. Goodwin (eds.), 335–358. New York: Cambridge University
Rampton, B. 1995. Crossing: Language and Ethnicity among Adolescents. London: Longman.
Rampton, B. (ed.). 1999. “Styling the ‘other.’” [Theme issue]. Journal of Sociolinguistics 3: 421–
Rampton, B. 2003. “Hegemony, social class and stylisation.” Pragmatics 13: 49–83.
Rose, M. 2003. “‘On de farm’: Sociolinguistic meaning in town and country.” Paper presented at
New Ways of Analyzing Variation 32, Philadelphia, PA, October 9–12.
Schilling-Estes, N. 1998. “Investigating ‘self-conscious’ speech: The performance register in Oc-
racoke English.” Language in Society 27: 53–83.
Trudgill, P. 1972. “Sex and covert prestige: Linguistic change in the urban dialect of Norwich.”
Language in Society 1: 179–195.
Trudgill, P. 1986. Dialects in Contact. Oxford: Basil Blackwell.

Grammatical resources for social purposes
Some aspects of stancetaking in colloquial
Indonesian conversation

Robert Englebretson
Rice University

1. Introduction1

This paper is predicated on two points of departure: First, that stancetaking is
a pervasive activity which speakers engage in through the use of language; sec-
ond, that grammar is motivated and shaped by language use. Assuming these two
propositions, then, we should expect grammar itself to be directly implicated in
stancetaking. In other words, if stancetaking is a frequent activity of language use,
and if frequent activities of language use themselves play a part in shaping lan-
guage form, then it follows that stancetaking must therefore also be understood
as having a role in shaping language form, and the grammatical resources of a
language should likewise reflect principles of stancetaking. This paper examines
three such resources in colloquial Indonesian, based on a corpus of spontaneous
conversational data. I demonstrate how Indonesian speakers use first-person-sin-
gular referring expressions, the -nya clitic, and verbal diathesis (voice) respec-
tively to manage and index three facets of stancetaking in everyday conversational
interaction: identity, epistemicity, and positioning. I further argue that traditional
descriptions and accounts have overlooked crucial aspects of the meaning and
function of these three grammatical resources. Adopting a view of grammar as
rooted in social interaction (in this case, specifically, the social and interactional
processes of stancetaking) provides a rich set of previously undocumented obser-
vations about their meaning and use.
Regarding the first point of departure mentioned above, the ubiquity of
stancetaking in its various forms has been noted for several decades. As Stubbs
points out:

in general. numerous linguists have indeed taken up this call. All sentences encode such a point of view.  (1986: 1) In other words. in addition to fulfilling their traditional. since Stubbs’s proposal. sociocultural.2 In the past 20 years. For example. which contribute directly to the social worlds speakers are constructing through stancetaking. es- pecially those working in corpus-based. Rather. every utterance enacts a stance. Language form has been argued to be determined largely by the cognitive and biological makeup of human beings. The present paper obviously cannot attempt a comprehensive and thorough analysis of the entirety of stancetaking in the language. I offer these observations and analyses in hopes that they will spark future research by other Indonesian scholars interested in investigating stancetaking more thoroughly. … and the description of the markers of such points of view and their meanings should therefore be a central topic for linguistics. Ford et al. There are undoubt- edly more than just these three resources for stancetaking in colloquial Indone- sian. as a first step toward this goal. most of this work has been focused on English. while there are numerous grammati- cal descriptions available for both formal and colloquial varieties of Indonesian. and the communicative contexts of language use (cf. cognitively-based referential functions of expressing and managing information. Givón 1979. they have specific interactional functions too. The present paper hopes to expand this view by offering some observations on the social/interac- tional nature of Indonesian grammar. if not solely) a cognitive object and referential system. they encode their point of view to- wards it …The expression of such speakers’ attitudes is pervasive in all uses of language. 2002. has likewise informed the work of numerous researchers over the past several decades. Wouk (1998. and a recognition of the centrality of stancetaking has by and large not found its way into field linguistics. nor have they addressed the lexico-grammatical means which speakers use to accomplish it. and aware- ness of this should inform the linguist’s work at all levels. The second point of departure for this paper.70 Robert Englebretson whenever speakers (or writers) say anything. the functionalist principle that grammar is motivated and shaped by language use. For example. For the three aspects of Indonesian grammar that I discuss in this paper. I show that. to roughly paraphrase. Cum- ming and Ono 1997. and interactional approaches to language. However. Ochs et al. nor into the writing of descriptive grammars of other languages. Yet. Langacker 1999. 2001) analyzes two clause-final particles used in expressing solidarity in social interaction. In- donesian descriptive and pedagogical grammars tend to reflect the general bias in descriptive/typological/field linguistics which conceives of grammar as (pri- marily. systemic-functional. and seeks to initiate a discussion of how Indonesian speakers take stances. none of these sources has addressed stancetaking per se. .

Various grammatical resources have been shown to emerge (Hopper 1987) out of broader cognitive. Conrad and Biber 2000). Hunston and Sinclair (2000) have proposed a “local grammar” of evaluative adjectives and nouns. In particular. One of the reasons why . I offer here a preliminary look at how three aspects of the grammar of one particular language are implicated in stancetaking. inter alia). The purpose of this paper is to explore some aspects of the grammar of stan- cetaking in a language other than English – in this case colloquial Indonesian – and to show ways in which traditional grammatical categories described for this language must also be understood as doing stance work. And going beyond the traditional systems of English grammar. and system- atically within specific languages. rather. The grammatical marking of stance. the Longman Grammar of Spoken and Written English (Biber et al. and style stances (cf. 1999) where chapter 12 is entitled. then focus- ing on stancetaking would seem a natural. worthwhile next step in the devel- opment of functionalist research paradigms concerning usage-based models of grammar. this volume). for example. Barlow and Kemmer 1999. Ono and Thompson 1995. the grammar of English modals has proven to be a rich area for the epistemic evaluation of propositions (cf. as a first step to initiating interest in a functional link between stancetaking and language form. the link between grammar and stancetaking has virtually gone un- noticed in descriptive field linguistics. Bybee and Hopper 2001. outside of English. discourse. 2002. While a thorough review of this literature lies well outside the scope of the pres- ent paper. like the one we find in. Researchers in both quantitative corpus linguistics and systemic-functional linguistics have worked to identify macro-level features of lexis and grammar that serve as markers of stance or evaluation. review and discussion in Thompson and Hunston 2000: 20–21). attitudinal. Tomasello 1998. both cross-linguistically. As alluded to above. as have epistemic phrases such as I think and I guess (cf. English adverbials have also been widely investigated as signaling various types of epistemic. Substantial work has yet to be done to flesh out the role of stancetaking in motivating and constraining grammar. Bybee 2006. and social-interactional aspects of language use – a view which has widely come to be known as usage-based approaches to grammar (cf. Since stancetaking is such a pervasive activity found in language use. The current paper does not attempt to address the proposed underlying causal nature of this relationship. Kärkkäinen 2003. 1989. Stancetaking in colloquial Indonesian conversation 71 1996. Yet. considerable English-centered research already exists that explores the connection between stancetaking and grammar. among many others). the relevance of this approach becomes clear in light of the ubiquity of stancetaking discussed above and outlined in the introduction to this volume. Biber and Finegan 1988. It would be unusual indeed to open to the table of contents of a reference grammar of another language and find a chapter devoted to stance.

we see that much of what has previously been ascribed to meaning and information coding can. when examined in real contextualized language-in-interaction. and verbal diathesis (voice). and subsequently glossed and coded them into a relational database. All examples presented in this paper come from these six speech events. (1993). . and transcribing and glossing the data. and are labeled with the title of the tran- script and the line-number(s) of the IU(s) cited in the example. After a brief discussion of the data used for this study. the fifth is a discussion about music and art by a university student group that meets monthly. But when grammar is also understood as simultaneously being firmly rooted in social interaction. choosing the seg- ments. The corpus comprises 36. also function at a local level of discourse to index stance. Appendices A and B of this paper outline the transcription and glossing conventions. represent three very different facets of stancetaking. in addition. Native- speaking research assistants and I transcribed the segments using a modified ver- sion of the Discourse Transcription system outlined in Du Bois et al. I will focus on first-person-singular referring expressions. 2.72 Robert Englebretson reference grammars have generally not included categories such as stancetaking is likely due to pervasive (mis)conceptions of grammar as being primarily a cogni- tively-oriented. and I (the foreign researcher) was not present during the recordings. and about the specific methodology used for recording the speech events. but. These six segments are part of larger speech events. these may function in stancetaking to in- dex identity. I have chosen these three grammatical resources specifically because they occur frequently in colloquial Indonesian. and the sixth is a radio call-in show about Islam. the -nya clitic. Englebretson (2003) provides further details about the six segments in the corpus. propositional system for information coding. Data The database for this study consists of a corpus of six transcribed audio recordings of naturally-occurring spontaneous Indonesian discourse. and clearly il- lustrate the kinds of gains that can be achieved by incorporating stancetaking into the realm of grammatical description. The speech events were not elicited. and positioning respectively. also be understood as social and interactional in nature.265 total words in 12. This paper seeks to flesh out three such resources in colloquial Indonesian – lin- guistic forms that have general grammatical functions. In conversational interaction. epistemicity. which make up the approximately 25 hours of spoken colloquial Indonesian I collected while con- ducting fieldwork in Yogyakarta (Central Java) in 1996. nearly four hours of running speech.972 Intonation Units (IUs). Four of the segments in the corpus are casual face-to-face conversations among friends.

and use it in their daily lives and interactions. The relevance of these basic grammatical facts about Indonesian will become clear as the discussion progresses throughout the paper. patient-trigger. and the NP arguments must be determined based on contextual and pragmatic factors. and most are students at various universities in the city of Yogyakarta. education. and East Timor (which. and noted its similarities with what other researchers have called Spoken Jakarta In- donesian (cf. While English and many other lan- . standardized variety of Indonesian promulgated by the government. For an over- view of colloquial Indonesian grammar see Ewing (2005). typically Javanese. was still part of the Republic of Indonesia). Before moving on to our discussion of the three aspects of stancetaking in Indonesian that form the core of this paper. the colloquial Indonesian cor- pus also includes basic ethnographic data about each of the speakers (with the exception of those appearing on the radio call-in show for whom this informa- tion was unavailable). applicative. Indone- sian is a primarily isolating language with sparse inflectional morphology. and for an overview of standard Indonesian see Sneddon (1996) among others. Wouk 1989. Person. and tense/aspect/mood categories are not indicated on verbs. Englebretson 2003). and others. 3. First-person-singular reference as indices of identity I will begin the discussion of grammatical resources used for specific stance pur- poses by focusing on the complex system of first-person referring expressions available to speakers of colloquial Indonesian. noun-phrase ellipsis is quite frequent. middle. Speakers range in age from late teens to late 20s. a few brief notes on Indonesian grammar are in order. I have characterized this language variety elsewhere as spoken colloquial Indonesian (cf. so that clauses in fact often consist of a verb alone. which. Verbal af- fixes consist chiefly of voice and valence marking: agent-trigger. There are significant differences between the grammar of colloquial Indonesian and the grammar of the formal. Typologically. Sulawesi. including Irian Jaya. at the time of recording. and media. reflexive. but the corpus comprises speakers from diverse regions of In- donesia as well. but report using Indonesian as the language of communication with each other most of the time – as is represented in the corpus. Sumatra. In conversation. 1999). has spread widely outside of Jakarta as well. The majority of the speakers (as with the majority of the Indonesian population in general) are from the island of Java. number. as noted by Poedjosoedarmo (1982: 142). Most speakers speak a local language as well. were educated in Indonesian. Stancetaking in colloquial Indonesian conversation 73 In addition to the audio files and transcripts. as is typical of educated urban residents of Indonesia. All speakers claim Indonesian as a native language.

This is often noted by Indonesian speakers in language atti- tude surveys. as cited by Errington in his work on social identity. For example.  (Errington 1998: 92) And indeed. But. pronouns are one means used for stancetaking. As compared with many local languages in Indonesia that have intricate lev- els of speech styles to index and maintain social relations (e. language shift. Indonesian is relatively restricted in the kinds of power and solidarity relations that it allows speakers to express. at least as normatively described. However. Javanese commonly alluded to Indonesian’s lack of stylistic elaboration as mak- ing it simple (I: sederhana). For example.. the varying forms available for first-person-singular reference are one resource that speakers can draw upon to evoke such things as social distance.). In colloquial Indonesian. self-expressive linguistic stance likewise reflects and creates locally-rel- evant aspects of social and personal identity. paradigmatically-unrelated. and mine). or toughness. colloquial Indonesian has several unique. the social-psychological work of Brown and Gilman (1960). in that it serves to indicate how the speaker is constructing the self to be perceived by others. bland (I: tawar). just as culturally-relevant body posture.74 Robert Englebretson guages have only one unique first-person-singular (1SG) pronominal form (e. in addition to this referential function. The gen- eral grammatical function of these expressions is. In response to my various queries about differences between their two languages. to create and maintain social relations and identity.g. of course. English I and its paradigmatically-related inflections me. the variation observed among these multiple 1SG forms as used by a single speaker in conversation illustrates a specific means of stancetak- ing. the speech levels and vocabulary registers of Javanese.g.. in Indonesian as in many languages. This type of stancetaking is self-expressive. as described by Errington 1988). The so-called t/v distinction (for the French pro- nouns tu and vous) illustrates that pronouns are far more than merely referring expressions in the structuralist sense – speakers use them to indicate social rela- . for example. a particular hand gesture to signal in-group social affilia- tion. 1SG forms. etc. can index social relations and attitudes (arms folded across the chest to index social distance from or indifference toward an interlocutor. The social- relational work of pronouns has long been noted for other languages. the following quote is typical of the attitudes of many bilingual Javanese speakers toward Indonesian. or plain (I: polos) in comparison with Javanese. and codeswitching in Java. as a means for the speaker to refer to the self in the ongoing discourse. my. which characterizes the “familiar” and “formal” distinction found in the second-person pronouns of many European languages as reflecting interpersonal solidarity or deference be- tween speaker and addressee. or physical stance. casualness. Indonesian does incorpo- rate a few interactionally crucial stylistic distinctions in its personal pronoun rep- ertoires.

these 1SG forms are characterized in terms Table 1. Many of the clauses in the corpus. From a more anthropologically-informed per- spective. therefore. in particular.  Total 1SG referring expressions in the corpus Form Total tokens aku (aku. and descriptive grammars. Because of the high prevalence of argument ellipsis in colloquial Indonesian conversation. and the fifth is self- reference through the use of one’s own name. gua. sya. and how pronouns function in the expression of social relations and attitudes. Table 1 lists each of the 1SG referring expressions found in the corpus. This table shows. The question of what motivates a speaker to either use or to ellipt a 1SG form is of great interest in its own right. In all discussion of these forms in the literature that I am aware of. Colloquial Indonesian speakers have at their disposal a range of personal pronouns. but lies outside the scope of the present paper.3 While grammars and text- books of Indonesian typically contrast these forms based on macro-level variables of region and formality. and tak) are 1SG pronouns. However. including textbooks. along with their overall token frequencies. ku-. in descending order of frequency. and are used by a single speaker within the same speech event. and -ku) 578 saya 222 gua/gue 83 tak 37 proper name 35 . what I wish to focus on for purposes of this paper is the rather unusual case (typologically speaking) of the multiplicity of forms that Indonesian speakers have available for 1SG reference. includ- ing second-person forms reminiscent of the t/v distinction widely discussed for other languages. how they function in stancetaking. a sixth op- tion also available to speakers is to use no form at all – so-called argument ellip- sis – thus leaving the identity of the unexpressed arguments to contextual infer- ence. even though they are clearly referring to the speaker. and. Mühlhäusler and Harré (1990) provide an overview of how pronouns in various languages relate to culturally-diverse concepts of the self.4 I am concerned here with the ways in which overt first-person-singular forms are used. focusing here only on the five overt forms shown in the preceding table. The first four (aku. Stancetaking in colloquial Indonesian conversation 75 tionships of affiliation or distance. I will leave the issue of colloquial Indonesian argument expression versus ellipsis for future research. evidence from the corpus demonstrates that these also function at a local level of discourse. dictionaries. have no overt 1SG form. the five forms that speak- ers in the corpus regularly use in referring to themselves. pedagogical materials.

and.(a first-person pro- clitic on verbs) and -ku (a possessive enclitic and object of a preposition). Each is invariant for grammatical categories such as case or possession. me. A wel- come exception is Sneddon (2002) who provides a quantitative analysis of 1SG . no paradigmatic forms of these pronouns exist. In sum. and ethnicity as follows. reflecting the way these pronouns are generally treated in the Indonesianist literature. including its paradigmatically-related clitic forms ku. intuition-based claims of analysts and native speakers.76 Robert Englebretson of the traditional macro-level sociolinguistic variables of formality. used among intimates. The second-most frequent 1SG pronoun in the corpus is saya. by me” Readers who are not familiar with Indonesian grammar may wonder whether any of these forms are inflectionally related to each other. gua. which is generally characterized as a Javanese loan. It is considered more formal than aku. other than the clitic forms of aku. Chinese. and the most appropriate form for public interaction and among people who are not so- cially intimate. the remaining two pronouns (gua/gue and tak) are characterized based on the regional/ethnic/linguistic background of the speaker. According to the impressions of native-speaking Indonesian consultants. as I will demonstrate shortly. It is typically associated with young urban residents of Jakarta and/or speakers of Chinese-Malay background. familiar first person pronoun” tak “(Javanese) I. The answer is negative. The following entries from Echols and Shadily’s (1989) Indonesian-Eng- lish dictionary summarize the 1SG pronoun forms found in my corpus. or by a person of higher social status to a person of lower social status. my” gua “(Jakartan. The least-frequent 1SG pronoun in the corpus is tak. and the use of self-refer- ence by one’s own name is characterized in terms of the gender and marital status of the speaker. Such definitions and general characterizations have tended to be based on impressionistic. aku “I (familiar. do not fully account for their use in conversa- tional interaction. region. intimate)” saya “I. two of these forms (aku and saya) are traditionally characterized by differences in for- mality. and its variant pronunciation gue. The final entry in the table indicates that there are 35 instances in the corpus of speakers referring to them- selves by use of their own name. There has been an overall lack of empirical research into the actual distribution of these forms in naturally-occurring language-in-use. The most frequent of these pronouns observed in the corpus is aku. which is prescriptively the all-pur- pose Indonesian 1SG pronoun. Aku is generally characterized as familiar and informal. The next-most-frequent form in the corpus. the use of one’s own name is generally restricted to young unmarried women – and this observation holds true for all 35 instances in the corpus. was borrowed into Indonesian through Hokkien Chinese. Colloquial) I.

These empirical findings seem to provide robust support for the general claims that the macro-level variables of formality and in- timacy strongly determine the use of saya or aku respectively. if 1SG pro- nominal forms are indeed sensitive to formality level. as I will demonstrate in the remainder of this section. However. (3) in the call-in show. based on the following evidence. Tanya-Jawab represents a far more formal register than any of the other five speech events in the corpus. then. As mentioned in Section 2. five of the speech events in my corpus consist of face-to-face conversation among people who know each other well. radio-broadcast telephone interaction. In contrast. there are no in- stances whatsoever of the familiar aku nor of the regionally/ethnically marked gua/gue or tak. On all three of these dimensions. face-to-face conversations. Results from my own corpus support this as well. while the sixth speech event (which I have entitled Tanya-Jawab ‘question and answer’) is an entirely different genre of spoken language – a phone-in radio show in which a radio host and a Muslim cleric answer callers’ questions about Islam. These numbers provide strong evidence that speakers treat saya as the 1SG pronominal form most appropriate for formal public interaction. . participant roles are more balanced. This table shows a striking difference between 1SG reference in the Tanya- Jawab speech event as compared with the aggregate use of 1SG reference in the other five. Sneddon’s quantitative findings do bear out the claims that 1SG pronouns vary based on formality level. (2) the callers are not socially acquainted with the ra- dio host or the cleric. above. with 98 instances of this form and only one instance of any other (self-reference by means of the speaker’s own proper name). This speech event is different from the other five in a number of ways: (1) it is a public. participant roles and social relations are clearly delineated – the cleric is in a posi- tion of authority as expert. the overwhelming preference for aku found in the conversational speech events suggests that speakers treat aku as the preferred form for informal inter- action among social familiars. In this speech event. in the other five speech events. As shown in Table 2. while the other five are private. and the callers are positioning themselves as lacking the knowledge that the cleric provides. whereas in the five other speech events. and there is no overtly recognized authority or expert. the situation is far more complex than could be accounted for by a direct deterministic relationship between social variables and linguistic form. then we would expect to find significant differences between the radio call-in show as contrasted with the other five informal face-to-face conversations. Stancetaking in colloquial Indonesian conversation 77 forms (and other stylistically-stratified variables) in a corpus of various genres of spoken Indonesian among residents of Jakarta.5 In the radio call-in show. the data do indeed provide strong evidence for this claim. participants know each other well and interact with each other regularly. speakers overwhelmingly prefer saya. Thus.

originally from Eastern Indonesia. in the same speech event. top-down. again from Ari in the Pencuri speech event. Other material in parentheses indicates that no corresponding form is used in the Indonesian data. I have produced similar tables for all speakers in the corpus. every speaker in the corpus (aside from those in the Tanya-Jawab radio call-in show) regularly uses more than one 1SG form. in the same physical setting. and a student in Yogyakarta at the time of recording. Note that her repertoire of 1SG forms includes all five found in the corpus. Primary evidence for this claim comes from observ- ing 1SG usage by single speakers. during the same speech event. with the same interlocutors. there are no single-form speakers in the face-to-face conversa- tions. These findings suggest that the traditional. but is dynamic. a priori account of what motivates 1SG use is incomplete. For the benefit of non-Indonesianist readers and to facilitate the reading of this excerpt. there are no speakers who refer to themselves with only one 1SG form. . This excerpt is part of a larger narrative Ari is telling about a friend who had allegedly stolen money from her and then lied to cover up the theft. Following is a table illustrating the use of 1SG forms in the speech of Ari (speaker A) from the Pencuri speech event.  Comparison of 1SG reference in informal and formal speech events Form Face-to-face tokens “Tanya-Jawab” tokens Total aku 578 0 578 saya 124 98 222 gua/gue 83 0 83 tak 37 0 37 proper name 34 1 35 The data suggest that pronoun choice in conversational interaction is not pre- determined based on static social variables. takes place at the local level of discourse. and is used in stancetaking to index the speaker’s construction and expression of identities. As an actual example of the use of 1SG forms in real-time conversational interaction.7 seconds) of a conversational narrative. Table 3 shows a single speaker using all five of the 1SG forms to refer to her- self. female. with the same interlocutors. Not all speakers use all five forms in such striking numbers as does this particular speaker. 21 years of age. and as mentioned above. In the five face-to-face conversations in the corpus. I have included the actual Indonesian 1SG forms in parentheses in the English free translation. but it is an empirical fact that every speaker in the face-to-face conversational data uses two or more.78 Robert Englebretson Table 2. Ari is one of the more loquacious speakers in the corpus. I offer the following short excerpt of 18 intonation units (lasting 17. but is necessary in the English free translation.

feeling just:now prt (My) feeling just then was that. 2sg ask 3sg D: You asked her? 212 A: . 3sg at this A: She was at.. rel hold wallet 3sg prt she was the one who had taken (my) wallet. 1sg prt A: I_(aku) was um. 215 . . 213 bingung lho kok... Dia di ini.. 211 D: Kamu tanya dia. Perasaan tadi tuh. Di depan pintu rumah Ari. 216 [terus gua bilang-in]. 210 . you know. Stancetaking in colloquial Indonesian conversation 79 Table 3. yeah M: [Yeah]. 214 . yang pegang dompet dia kan. 217 M: [Yah]. Aku kan.  1SG forms in the speech of a single speaker Form Tokens aku 81 saya 40 proper name 25 gua/gue 8 tak 5 (1) Pencuri IU 209–227 209 A: . next 1sg say-app [So then I_(gua) said to her]. um... confused prt prt (I) was confused.. at front door house Ari in front of the door to my_(Ari) house.

. Kamu. 219 . hunh-unh. a total of 17. weren’t you?” 223 Nggak tuh.. neg run like:that prt [she] didn’t run away. name-pos prt Rifka A: Her name was Rifka. gua in 216.. Rif! 220 . money 1sg rel inside wallet prt my_(saya) money that was inside (my) wallet.7 seconds of talk. neg know (I) don’t know (anything about it). aku in IU 212. Note that in this brief excerpt of only 18 IUs of conversation. next Ari enter again search to inside So then I_(Ari) went in again and looked for it inside. I will now turn to a discussion of two possible hypotheses to explain the 1SG variation seen in this brief excerpt and in the corpus more broadly: (1) distribution of 1SG forms could be grammatically conditioned.. 221 tadi nge-lihat-in. . 222 duit saya yang dalam= dompet ya. She uses Ari (her own name) in IUs 210 and 225.80 Robert Englebretson 218 A: . just:now at-see-app were just now looking at. neg prt “No.. Terus Ari masuk lagi cari ke dalam. Eh Rif. hey Rif “Hey... and saya to refer to herself in reported speech in IU 222.. 226 dia itu ngak- -- 3sg that neg She didn’t – 227 nggak lari gitu lho.. the speaker seamlessly switches among four of the five 1SG forms found in the corpus. Nama-nya kan Rifka. 224 nggak tahu.” 225 . 2sg You. and (2) distribution of 1SG forms could be based .

Both of these hypotheses fail to fully account for what we actually observe speakers doing. as illustrated in the following three examples. neg ex money 1sg I don’t have any money. these very tokens are actually attested elsewhere in the corpus.’ These four forms. the 1SG form saya in IU 222 duit saya ‘my money’ would be per- fectly acceptable with the substitution of any of the other 1SG forms found in this short excerpt: duit gua (using gua/gue). illustrate that each of these four 1SG forms (saya. These three examples. some 1SG forms in the corpus would occur in grammatical environments where other 1SG forms do not occur. following classic structuralist reasoning. ‘My_(aku) money doesn’t exist. money -1sg neg ex I don’t have any money. and I will therefore suggest a third ap- proach which more thoroughly accounts for the observed distribution in 1SG forms. along with tak. In fact. are all fully attested in the corpus in other grammatical roles too: as both the trigger and the non- trigger argument in transitive clauses.. speakers are actively using 1SG forms in stancetaking in order to construct and project aspects of their social and personal identities. (with the possible exception of tak. In other words.. namely. then. (2) Pencuri IU 2403 Duit -ku nggak ada. gua. pull money Ari all (She) pulled out all my_(Ari) money. ‘My_gua money doesn’t exist. to which I will return shortly). along with IU 222 in excerpt (1). The first hypothesis to consider is that 1SG forms may be motivated and con- strained by grammar. aku. we would expect to find 1SG forms occurring in some sort of contrastive distribution. and self-reference with proper name) all occur in the same grammatical environment: as a 1SG possessor in a possessive NP headed by duit ‘money. Stancetaking in colloquial Indonesian conversation 81 on macro-level social variables. (lit. above.’6) (4) Pencuri IU 266 Cabut duit Ari semua. and as the single argument of intransitive . there appear to be no such grammatical restrictions. or duit Ari (using the speaker’s own name) are all acceptable and natural. (lit.’) (3) Dingdong IU 1002 Nggak ada duit gua. For example. However. duitku (using the enclitic form of aku). If a speaker’s choice of 1SG form were determined by gram- matical factors.

Thus. there are unmarried women of the same age in the corpus who never use this form. I have al- ready discussed this to some extent above. who uses all five 1SG forms). The hypothesis that the variation of 1SG forms is grammati- cally conditioned receives minimal support from tak. that saya and aku are strongly correlated with formal- ity. and it tends to strongly collocate with certain verbs of cognition. and claims not to speak Javanese (suggesting that tak is not simply Javanese code-mixing. However.82 Robert Englebretson clauses. For example. this speaker has never lived in Jakarta at all (thus suggesting that calling gua Jakartan is an oversimplification). and the speakers who do use their names as 1SG reference also use other 1SG forms as well (e. where a single speaker freely uses both saya and aku (along with the other 1SG forms) in the same speech event. and female. based on a comparison of 1SG usage in a radio call-in show with 1SG usage found in everyday conversa- tion among social familiars. unmarried. cf. The social variables hypothesis does. while this collocation does suggest a restrict- ed grammatical profile for tak.). and even within the same short span of discourse as seen in excerpt (1). These distributional facts suggest that the variation among 1SG forms is not grammatically conditioned.’ suggesting that tak pikir ‘I think’ is a grammaticized epistemic phrase (cf. Rauniomaa. However. Ari in Table 3. However. and the three speakers in the corpus who use this form are in fact young. it never occurs in the corpus as a possessor in a possessive NP. It also cannot account for the distribution observed in Table 3. The regional account of gua and tak is similarly problematic. while aku is the preferred form in the conversa- tional data. a regional account of gua and tak does not account for much of what speakers are actually doing in conversational interaction. along with the other three 1SG forms.g. appear to receive strong sup- port from 1SG reference by means of proper name. unmarried women. however. this cannot account for the fact that there are still 124 tokens of saya in the informal conversational data. it nonetheless overlaps with the other 1SG forms in other environments. Saya is the only 1SG form found in the call-in show (except for one instance of self-reference by proper name). on similar constructions in Finnish). this volume on English I think and I guess respectively. this volume. Ari uses eight tokens of gua (which is claimed to be colloquial and Jakartan) and five tokens of tak (which is claimed to be Javanese). and have shown. So for speakers like Ari who . as shown in Table 2. and no support whatsoever from the other four 1SG forms. this is seen as a form used by young.. Traditionally. as shown in Table 3. The second hypothesis to consider is that 1SG use is linked to macro-level social and stylistic variables such as formality and regional background. The one form which does appear to have some grammatical restriction is tak. However. Kärkkäinen 2003. Eighteen out of the 37 tokens of tak in the corpus co-occur with the verb pikir ‘think.

An analysis of 1SG usage based on social variables fails to account at all for the observation that every speaker in the conversational data uses two or more forms.) or stable attributes of the particular speech situation (e. or even causes particular [language] behaviours” (Benwell and Stokoe 2006: 26). etc. In other words. the social variables hypothesis still does not explain what would condition her to use her name at some points in the conversa- tion. a performative approach would ask: What is it that a speaker is accomplishing by using this particular 1SG form at this particular moment in the interaction?. speakers may use lan- guage to evoke social or personal identity categories in order to achieve certain goals within talk. we understand identity categories (such as gender. As shown in Table 3. The fact that 1SG usage varies. If. regional origin of speakers. in traditional variationist work. she is speaking with the same interlocutors (housemates and friends whom she has known for nearly two years). identity categories are emergent. or formality of speech events) from a performative point of view – as something speakers ‘do. casualness. These shortcomings of the social variables hypothesis reflect larger. formality. and her ethnic/ regional/language background also does not change throughout the interaction. one which stands a much better chance of accounting for how speakers use these forms in everyday talk than do either of the hypotheses examined above.. Stancetaking in colloquial Indonesian conversation 83 do use their names as 1SG reference. age.’ rather than as something speakers ‘are’ – then we can begin to ap- proach the variation among Indonesian 1SG forms from a new perspective. prob- lematic assumptions present in much of traditional variationist sociolinguistic research regarding the nature of identity categories in general. regional origin. a single speaker may use all five of the 1SG forms. a priori characteristics of the speaker (age. in the same physical context (in the common room of their boarding house). In other words.. the variables of formality and regional origin for Ari are remaining con- stant.) In this view. gender.e. Variationist so- ciolinguistics has tended to assume identity “as a pre-discursive construct that correlates with. while the macro-level sociolinguistic variables re- main constant. Rather than seeking macro-level variables to explain the distribution of 1SG forms in the data. the par- ticular variant a speaker uses in any given instant of talk falls out probabilistically from these pre-existing social and stylistic categories. on the other hand. yet her use of 1SG form varies. etc. socially constructed practices . Rather than being relatively stable attributes that pre-exist the current discourse. What kind of stance is the speaker evoking by using this form at this time? According to a performative conception of identity. despite the fact that she is using them all in the very same speech event. language background. social categories of identity are treated as relatively stable.g. and to use different 1SG pronouns at other points in the talk. i. suggest that there is an additional level of detail not captured by the social variables hypothesis.

they are shifting their stance: they are using this form to construct social distance or formality at a particular point in the ongoing interaction. But what about the 124 tokens of saya in the conversational data? If speakers are actively constructing these interactions as casual and intimate. Bucholtz and Hall 2005 for similar discussion). to show deference. with 578 tokens in these five speech events – nearly five times more frequent than the use of saya in these same speech events. with its connotations of formality and distance? I suggest that when speakers use saya in informal conversation. above. In this clause. aku is overwhelmingly the most frequent form. On the other hand. with 98 tokens (plus one 1SG refer- ence by use of proper name). as to how to package the utterance for 1SG self-reference. (To return to the analogy of physical stance. among other lexical and grammatical items. and to index social distance among participants. and in excerpt (1). 1989) I shall hereafter refer to as constructed dialogue). in the casual conversational data. and building the social intimacy among speakers typical of this type of interaction. I shall now offer some qualitative observations of how 1SG forms are used in the performance of identity. As an example. following Tannen (1986. con- sider the use of saya in IU 222 of excerpt (1). the speak- er has several potential options at any given moment in talk. I will begin by dis- cussing the use of aku and saya from this perspective. specific types of identity categories. one way that a speaker “performs” identity is through the packaging of utterances in ways that reflect and index specific categories.) A crucial observation which lends support to this idea is that most of the tokens of saya come from instances of reported speech or thought (which. even when a different pronoun would have been used in the original context of utterance. which starts in IU 216 with the quotative verb bilang ‘say. then why are these same speakers simultaneously using saya. relax- ing the prescribed norms of public language use. Accordingly. to take a formal stance: to construct a public social and personal identity.84 Robert Englebretson of language use (cf. Bearing in mind a performative conception of identity. I would contend that speakers here are using aku as one means of taking a casual stance: actively constructing their personal identities as informal. I suggest that speakers here are actively using saya. As demonstrated in Table 3. The many types of identities in discourse include “local identity categories and transitory inter- actional positions” and “temporary and interactionally specific stances and par- ticipant roles” (Bucholtz and Hall 2005: 592). Saya is the only 1SG pronoun used in the call-in radio show. The distribution in Table 2 has already demonstrated that there is a strong correlation between these two pronouns and the types of speech events in which they occur. I suggest that these vari- ous forms serve to index. or evoke.’ Ari is reporting the words she alleg- . the use of saya in informal conversation could be compared with the physical act of tensing the body or temporarily adopting a more rigid posture. as suggested by the high frequency of aku.

one in which the speaker uses the more intimate aku. aku and other 1SG forms are commonly found in reported speech and thought as well. The full narrative will be presented and explicated for other purposes in Section 5 below. Based on these observations.. Maya that perfv nonvol-pull very I: Maya got really attracted. Interestingly. this is cheap! 2753 aku pingin handuk tiga ribu itu. For this reason. 1sg want towel three thousand that I_(aku) want this towel for 3. one would have expected aku here. Stancetaking in colloquial Indonesian conversation 85 edly had spoken to her friend. pre- sumably to reflect a more distant stance than the informal aku. since aku would most certainly have been the form she would have used with her friend in the original context. The following excerpt comes from a narrative about a woman named Maya who is nearly tricked into buying shoddy towels at an outdoor market. For several speak- ers in the corpus. Maya itu udah ter-tarik banget.” . While saya is the norm in constructed dialogue in the corpus. 2755 harus‑nya enam ribu lebih to. but for the present discussion. oh cheap this “Oh. the speaker is using saya.000 (Rupiah). “Hey. the relevant aspect of this example is the use of aku in the constructed dialogue in IU 2753. it is illuminating to compare the example of saya in IU 222 of excerpt (1) (as just dis- cussed) with another instance. if 1SG usage were conditioned by casualness and intimacy. I suggest that speakers are using saya to take a stance on reported speech or thought: to imply social distance from it. should‑stm six thousand more prt It should be 6. saya is only found in contexts of constructed dialogue – either the alleged words of others. weren’t you?”.000 (Rupiah) more. (5) Pencuri IU 2751–2757 2751 I: . it is by no means the only form of 1SG reference in this context. The form aku also would have been in parallel with the second-person familiar pronoun kamu with which she addresses her friend in the constructed dialogue found in IU 220. However. 2754 Padahal gede tebal itu. 2752 Ah murah ini. Rif! You were just now looking at my_(saya) money that was inside (my) wallet. in:fact big thick that It’s big and thick. or in the speakers own reported utterance as in this example.

000 (Rupiah). While a more thorough study of constructed dialogue in Indonesian con- versation is clearly warranted to assess this conjecture. as contrasted with Ari’s constructed dialogue in IU 222 of excerpt (1) above. The subsequent four IUs construct Maya’s inner dialogue. and (for whatever reason) as more distant from the speaker. The IU of interest for the present discussion is IU 2753. 2757 . the use in constructed dialogue of the less-intimate form saya versus the more-intimate form aku may be one means that Indonesian speakers have of marking it as less vivid.” What is notewor- thy about the constructed dialogue in IU 2753 of this excerpt. On the other hand.. is that Maya’s reported ‘thoughts’ use the more intimate aku. the theory of “territory of information”. as lower in epistemic value. In other words.86 Robert Englebretson 2756 dia kan. you know. In IU 2751. mikir‑nya ah lumayan @mahasiswa @baru @gitu ya. in informal colloquial Indonesian conversation (whether in the context of constructed dia- logue or not). Clearly this observation warrants further in-depth investiga- tion. If one difference between saya and aku in constructed dialogue has to do with the distance/intimacy or degree of vividness which the speaker is cre- ating. but I offer it here along with the presented examples to illustrate one resource which colloquial Indonesian speakers have of taking a stance. I would like to suggest that the use of saya versus aku in these contexts may be similar to the effect that many languages achieve by a distinction between indirect and direct speech. thought‑pos prt reasonable student new thus prt Her thoughts were reasonable for a new student. the use of aku in constructed dialogue reflects a conception of it as more vivid. In sum. 3sg prt he was (like). where Maya is reported to think: “I_(aku) want this towel for 3. then this observed use of saya for ‘speech’ and aku for ‘thoughts’ is indeed motivated by the speaker’s social world being constructed through stancetaking. including her thoughts about the towels and their price. and are thus constructed with greater intimacy. perhaps ‘thoughts’ are being presented as more internalized to the speaker. and IUs 2756–2757 explicitly frame this speech as having been constructed dialogue “her thoughts…”. as higher in epistemic value. the use of saya or aku is one means of stylistic elaboration avail- . and thus is constructed with greater distance (cf. as articulated by Kamio 1997). while ‘speech’ is more observable and less internalized. and as less distant. while Ari’s reported ‘speech’ uses the more distant saya. the speaker (Indra) provides an abstract of this portion of her narra- tive. in which she summarizes what happens to Maya at the market: “Maya got re- ally attracted” (by the towels she saw ).

but as a preliminary step toward this goal. to evoke some type of temporarily relevant identity in the ongo- ing discourse. and flirting with women). gua draws upon common language attitudes and stereotypes of- ten associated with speakers who use this form. Ari presents herself as blunt and outspoken. they are indexing these stereotypes. crucially. a regional/ethnic characterization of this pronoun does not accurately account for the observed data. For example. drinking at a local bar. that this form occurs to frame the rather blunt accusation Ari reports having made of her friend (IUs 219–222): “Hey. with 83 tokens in the informal conversational data. all expressed the view that this speaker sounded ‘tough’ or ‘rude’ or ‘macho. weren’t you?” In this short excerpt. Rif! You were just now looking at my money that was inside my wallet.” or appropriating the speech styles . 19 years of age.9%). Responses of several Indonesian consultants to excerpts from this corpus where gua is used bolster my claim that these stereotypes are related to ideologies of being “tough” and “outspoken. Stancetaking in colloquial Indonesian conversation 87 able to speakers. as discussed earlier. as they use language to construct their social worlds through stancetaking. The third-most frequent 1SG form in the corpus is gua/gue. I offer the case of the most prolific user of gua in the corpus: speaker L from the speech event entitled Dingdong (the onomato- poeic name of a pinball arcade game discussed in the speech event). and had lived in Jakarta from age 10 until age 13. I suggest that one of the ways in which speaker L is able to construct an identity of being ‘tough’ and ‘outspoken’ is through his use of gua. This pronoun is generally associated with colloquial Indonesian spoken in Jakarta and/or by speakers of Chinese ethnicity.” I readily acknowledge the need for both more in-depth language attitude surveys and qualitative sequential analysis of conversation to assess this hypothesis. When I played portions of this recording separately to three Indonesian consul- tants. consider Ari’s use of gua in IU 216 of excerpt (1) above: “So then I_(gua) said to her…” Note. Rather than indexing regional af- filiation per se. Rampton 2005 for work on “crossing. yet she has never lived in Jakarta and is not ethnically Chinese. Just as with saya and aku. whether or not they have any affiliation with Jakarta at all (cf. speaker L produces 63 of them (75. In other words. He uses only two 1SG forms in the data: 63 tokens of gua and seven tokens of aku. constructing her identity as someone who is ‘tough’ and unafraid to confront her friend whom she believes is a thief. This speaker is male. Table 3 shows that Ari uses eight tokens of gua in the corpus. When speakers use gua.’ and one consultant was offended by the topics of this spontaneous conversation (playing video games. I would like to suggest that her use of gua contributes to this locally-relevant construction of identity. I suggest that speakers use gua in stancetaking. of the 83 gua tokens in the data. Yet. As a first approximation of what such a temporarily-relevant iden- tity might be.

much of self-reference is closely linked to self-expressive stancetaking. the linguistic stance evoked by the use of one 1SG form over another creates similar transitory identities. and the variation among these forms does not simply fall out from social variables. This type of linguistic stancetaking is self-expressive and reflexive. I readily acknowledge that substantial work remains to be done on this topic. in addition to their referential function of referring to the speaker. The kind of stancetaking discussed in the present section. 1SG forms do far more than simply refer to the speaker. on the other hand. this volume. the linguistic expression of stance may likewise entail the display of some currently-relevant part of the speaker’s identity. Epistemic -nya constructions The previous section has illustrated how Indonesian speakers may use first-per- son singular pronouns to construct aspects of their personal and social identities. Following Wu. and implying facets of their social and personal identities. Johnstone. In this case. involves epistemic stance toward something external to the speaker – in this case toward the current utterance itself. Rather. To summarize this section. this type of epistemic stance refers to “a speaker’s indication of how he or she knows . on the use of stancetaking to bolster specific local identities). I have shown that 1SG self-reference in collo- quial Indonesian is an interactional practice available to speakers for creating. I have offered here a first step toward the goal of moving Indonesian grammatical description beyond traditional formal and referential boundaries. partly because there are so few tokens of these forms in the corpus (37 tokens of tak and 35 of proper name). as with the other three pronouns discussed here. both quanti- tative and ethnographic. The 1SG forms of Indonesian are a general grammatical resource which. their stance functions can best be accounted for by observing the kinds of identity categories speakers evoke at a local level of conversation. through body posture or gesture or other means. 4. and is managed locally in the ongoing discursive construction of identity. In this section. are simultaneously employed in the construction and maintenance of stance. I leave this to future research.88 Robert Englebretson of other groups for specific social purposes. such as social distance with saya or tough- ness with gua. maintaining. and toward a greater recognition of the ways in which grammatical resources are used in the service of stancetaking. Just as a speaker’s physical stance reflects and indexes temporary social categories. I have not offered an account here of how tak or proper names are used in stancetaking. Just as the act of physically taking a stance may often entail the outward presentation of some facet of the self. with the suggestion that.

outside of my own work on -nya (see Engle- bretson 2003: Chap. if any. .. this clitic frequently occurs on lexemes related to cognition. including the agent. is commenting on. and are used for epistemic stancetaking regarding the current ut- terance. In addition to its general. 5 for a more thorough discussion than is possible in this brief section). These stance-related functions have generally gone unnoticed in descrip- tive grammars of Indonesian. In this example. as illustrated by the following three examples. Grammars of formal varieties of Indonesian typically gloss this as a third-person-singular possessive marker. Nama-nya kan Rifka. I will illustrate how Indonesian speakers use the -nya clitic to do just this. referential functions as a marker of third-person possession. (2) assessment of interactional relevance: the degree of value a speaker places on the utterance. (6) Pencuri IU 218 A: . well-described. defi- niteness. Yet. which tend to focus only on the referential and grammatical aspects of this clitic.and patient-trigger verbal prefixes as well as pronominal clitics. usually regarding its role in the ongoing interaction. or (3) affect: the speaker’s mental or emotional attitude toward the proposition expressed by the rest of the clause. This frequency is greater than the occurrences of all other grammatical morphemes in the corpus.. In this section. or modality. name-pos prt Rifka Her name was Rifka. the -nya clitic occurs on the noun nama ‘name’ to indicate a third-singular possessor ‘her name. or is taking an affective or other position toward the person or matter being addressed” (2004: 3). it has received relatively little analysis. these constructions function as adverbials. with little. while the second again indexes a third-per- son possessor as in the previous example. and nominalization. In such a context. utterance. The first is a definite marker.’ The following example contains two tokens of -nya. These three referential uses of -nya are found in the colloquial corpus too. with 1. Stancetaking in colloquial Indonesian conversation 89 about. discussion of its use in stan- cetaking. and also recognize its use as a marker of definiteness and as a nominalizer. The -nya clitic is the most prevalent bound morpheme found in the corpus.570 tokens in the six speech events. The kinds of epistemic stances implicated by -nya constructions include: (1) evidentiality: the source of knowledge of the current utterance. suggesting that -nya deserves recognition as the most prevalent bound morpheme found in colloquial Indonesian in general.

this noun was mentioned previously in the discourse context. IU 834 of the next example illustrates a third grammatical function of ‑nya. from back this 627 pencopet-nya buka kalung-nya. on the verb minum ‘drink’ in IU 834.’ The three functions of -nya we have observed so far – third-person-singular possession. Rather. which then becomes the subject of the clause in IU 835. and nominalization – are the typical grammatical func- tions of this morpheme discussed in descriptive and pedagogical grammars of Indonesian. There are three in- stances of -nya in this example. (8) Dingdong IU 833–835 833 . The second instance of -nya in IU 627 marks third-person possession on the noun kalung ‘necklace. (Their) drinking even surpassed martinis. Yet. the thief undid her (Agnes’s) necklace. The first marks the noun pencopet ‘thief ’ as definite. speaker L is telling his friends about a group of women he had seen the previous evening at a bar. they said... and are not nominalizing a verb to be a clausal argument. ‘(Their) drinking even surpassed martinis. Minum-nya..’ just as in example (6). they said. is a nominalizer. these instances of -nya are being used in stancetaking. are not marking third-person-singular pos- session. are not indicating definiteness. . and plays a central role in the ongoing narrative. that of nominalization. this account of -nya does not fully address the range of use ob- served in colloquial Indonesian. pokok-nya bukan minum bir aja. in IU 599 (not shown here). definiteness. me-lebih-i= Martini kata-nya. and is commenting on his shock about the quantity of alcohol and types of drinks they were consuming. It marks this word as a noun. In this excerpt. The second. 580 out of 1. at-more-app martini word-stm the thing is..570 tokens.90 Robert Englebretson (7) Pencuri IU 626–627 626 D: dari belakang ini.. The first and third are stance markers (STM) to which we will return shortly. IU 627 contains two instances of the -nya clitic. main-stm neg drink beer just 834 . drink-nom 835 . More than one-third of the instances of -nya in the corpus. (they) weren’t just drinking beer. pickpocket-def open necklace-pos From behind.

Grammatically. indicating that it originated from the words of others. and knowledge from visual perception (example 11). That is. Dia kan kata-nya mau n-yari saudara-nya 3sg prt word-stm want at-look:for relative-pos 987 di sini. This serves as an epistemic evaluation of the interactional relevance of the utterance for the discourse at hand. It is the fact that these women were “not just drinking beer. the source of knowledge of the speaker’s utterance. or modality. return to example (8) above.. consists of -nya on pokok (a word roughly translated as ‘main. The first. Stancetaking in colloquial Indonesian conversation 91 enabling a speaker to comment on some aspect of the current utterance: how the speaker knows about it (evidentiality). these two tokens of -nya illustrate that when this clitic occurs on lexemes related to utterance. at here He said he wanted to visit his relatives here. One stance-related function of -nya constructions is the marking of eviden- tiality. each illustrating a type of evidentiality: knowledge based on the words of another (example 9). or how the speaker feels about it (emotional attitude). how the speaker evaluates it in light of the current conversation (assessment of interactional relevance). cognition.’ or ‘basic’). these expressions are adverbi- als.7 To illustrate two of the stance-marking functions of -nya constructions. one in each IU. in example (8). The remaining two -nya tokens in this example are stance markers. With this word. In sum. The remainder of this section will take up each of the three stance functions of -nya constructions in turn. as seen from their syntactic freedom to occur at numerous positions in the clause. speaker L knows what the women were drink- ing because this is something they had told him. knowledge from general inference (example 10). in IU 833. the source of knowledge for this utterance.’ Here. it forms an adverbial which marks the speaker’s stance toward the rest of the clause. occurs in IU 835 on the noun kata ‘word.. L highlights the rest of the clause in terms of his assessment of its relevance for the interaction – as something especially important for the interlocutors to pay attention to. Note that there are three instances of -nya in this ex- ample.’ ‘fundamental. The token found in IU 834 has already been discussed as a nominalizer. changing the verb minum ‘drink’ into a noun which functions as the subject of the clause. (9) Pencuri IU 986–987 986 . . The second instance of ‑nya as a stance marker. Following are three more examples. -nya marks evi- dentiality.” which provides essential background for the rest of his unfolding narrative about the antics of this group of friends.

the way they were dressed. and indicates her evidential source for this claim as based on how they looked.-an. Pokok-nya. ke-lihat-an-nya orang kaya itu lho nonvol-see-nom-stm person rich that prt 1250 dari <opo>? from what 1251 . IU 1249 of the following example illustrates a third type of evidential source. There are 59 instances of katanya (‘word’-nya) in the corpus.. all of which mark evidentiality based on another’s speech. The speaker knows that the person in question wanted to visit his brother. dead like-stm watch Mega My watch seems to have stopped. In IU 986 of example (9). notices that her watch is showing the same time as it had given earlier. Kayaknya (‘like/as’-nya) occurs 76 times in the corpus. In IU 1249. Mega. kata ‘word’ is suffixed with -nya.92 Robert Englebretson This example is parallel to the token already discussed in IU 835 of example (8). The thing is. because this was something told to her by a friend. making it into an adverbial expressing a general evidential source. changing the meaning .. As in this example. (11) Pencuri IU 1247–1251 1247 Trus ada ibu-ibu yang keren itu lho next ex mother-mother rel well:dressed that prt 1248 . The verb lihat ‘see’ is affixed with the stative/nonvolitional circumfix ke. In this example. but indicates that the speaker is making the state- ment based on general inference. they looked like rich people – because of. attire-pos And then there were some well-dressed women. this expression does not specify a particular evidential source. and infers that it therefore must have stopped. (10) Pencuri IU 2299 mati kayak-nya jam Mega. in this case visual perception. again indicating that the speaker’s source of knowledge is the words of another. what’s-it?. the colloquial preposition kayak ‘like/as’ is suffixed with the clit- ic -nya. the speaker states that the women must be rich. Dandan-nya. The following example illustrates -nya marking evidentiality based on general inference. The speaker. main-stm 1249 ..

(12) Pencuri IU 994 inti-nya dia malam itu nggak bisa n-emu-i gist-stm 3sg night that neg can at-meet-app saudara-nya itu. indicating the speaker’s assessment of interactional relevance. In sum. In this example. as essential background for understanding why they were being targeted by pickpockets. Stancetaking in colloquial Indonesian conversation 93 of the verb to ‘visible. used by a speaker to frame the utterance as a specific. we have seen three types of evidential sources indicated by -nya constructions: evi- dence from the words of another. in IU 1248 of example (11). As discussed previously regarding IU 833 in example (8) the word pokok (‘main’/‘fundamental’/‘basic’) is suffixed with -nya. Other similar expressions found in the corpus include misalnya ‘for example’ (‘example’ -nya). sibling-pos that The gist of it was that he couldn’t meet his brother that night. masalahnya ‘the problem is’ (‘problem’-nya) which indicates that the current utterance is somehow prob- . the speaker is concluding a narrative told to her by another per- son.8 The remaining -nya token. in IU 1248. Note that example (11) contains two additional tokens of -nya.’ which is then suffixed with the evidential -nya. She uses the noun inti ‘gist/nucleus/core’ suffixed with ‑nya to indicate her evaluation of this utterance as being the gist or summary of the surrounding nar- rative. suggesting that this is a highly grammaticized way of highlighting the main point of the ongoing interaction. which is the second type of epistemic stance we will be addressing in this sec- tion. the speaker uses the -nya construction here as an assessment of interactional relevance – indicating the degree of value of the current utterance regarding its contribution to the ongoing discourse.’ to indicate the evidential source as one’s own thoughts and feelings. and evidence from direct (visual) perception. is a stance marker. Here. The corpus contains 53 similar in- stances of pokoknya (‘main’-nya). As in the previous example. the speaker is highlighting the fact that the women looked rich. This clitic also commonly occurs on the verbs pikir ‘think’ and rasa ‘feel. evidence from general inference. The occur- rence of -nya in IU 1251 is a possessive. Following are two more examples of -nya constructions used as a stance marker of interactional relevance. indicating the speaker’s atti- tude toward the rest of the clause as being central to the understanding of the ongoing interaction. usually hypothetical example of the overall general point of the discourse.

Speaker L uses the noun untung ‘fortune/luck’ with . the verb takut ‘fear’ is suffixed with -nya. The third type of stance indicated by -nya constructions relates to affect – the speaker’s emotional attitude toward the rest of the proposition in the utterance.’ which have been illustrated in the previous examples. who thwarted an attempted robbery on a bus because she felt someone trying to undo her necklace. 3pl bring object sharp (We’re) afraid they might bring sharp objects. fear-stm that 546 mereka bawa benda tajam. The following example is similar. as well as pokoknya ‘the thing is’ and intinya ‘the gist of it is. She is afraid that the pickpockets might ‘bring sharp objects’ with them on the bus. (15) Pencuri IU 697 L: Untung-nya Si Agnes ke-rasa ya? fortune-stm prt Agnes nonvol-feel prt It was fortunate that Agnes felt it. The following example illustrates another type of mental/emotional attitude.. The following three examples are illustrative. and would hurt anyone who tries to thwart their criminal activities. namely the speaker evaluating the rest of the utterance as being lucky or favorable. indicating the speaker’s emotional attitude toward the proposition in the rest of the utterance. right? This clause serves as the coda to a story about one of the speakers and her friend Agnes. (13) Pencuri IU 545–546 545 A: Takut-nya tu.. Takut-nya lupa ya? fear-stm forget prt Are you afraid you’ll forget? As in the previous example. framing the proposition in the rest of the clause ‘(you’ll) forget’ as being something to poten- tially fear.94 Robert Englebretson lematic in light of the ongoing interaction. In this example. Takutnya (‘fear’-nya) in the first IU thus frames the next IU in terms of the speaker’s emotional attitude. the verb takut ‘fear’ is suffixed with -nya. (14) Blewah IU 13 .

Stancetaking in colloquial Indonesian conversation 95

the -nya suffix to indicate her emotional attitude toward the rest of the clause as
being favorable; it was ‘fortunate’ that ‘Agnes felt it.’ In sum, I have shown here
three examples of -nya being used as part of a construction to indicate affect – the
speaker’s emotional attitude toward the rest of the proposition. Examples (13)
and (14) indicate negative affect (specifically ‘fear’), while (15) illustrates positive
affect (the occurrence of an event being evaluated as ‘fortunate’).
To conclude, the current section has offered an overview of the ways in which
the -nya clitic is implicated in epistemic stance in colloquial Indonesian. Return-
ing to the definition which began this section, this type of stancetaking refers to “a
speaker’s indication of how he or she knows about, is commenting on, or is taking
an affective or other position toward the person or matter being addressed” (Wu
2004: 3). The three categories made explicit in Wu’s definition align closely with
the three main stance functions outlined here for -nya constructions. First, as an
evidential marker, -nya constructions indicate the speaker’s source of knowledge
for, or “how the speaker knows about” the utterance. Second, when marking the
interactional relevance of an utterance, -nya constructions serve to indicate how
the speaker “is commenting on” the utterance – assessing its value, usually to-
ward its contribution to the overall talk at hand. Thirdly, -nya constructions may
be used when the speaker “is taking an affective or other position toward” the
proposition in the rest of the clause, in which case they may indicate the speaker’s
emotional or mental attitude toward the event, e.g., as negative, as favorable, or
as something to be feared. I have presented numerous examples in this section
to illustrate the scope and variety of stancetaking in which the -nya clitic par-
ticipates. Yet, despite its pervasiveness in everyday language use, with the excep-
tion of Englebretson (2003) the stance-related functions of this clitic have been
virtually ignored in previous literature. Descriptive and pedagogical grammars of
Indonesian have tended to focus on the more general referential and grammatical
functions of -nya, such as third-person possession, definiteness, and nominaliza-
tion. In this section, I have shown that what has previously been described as
primarily referential and grammatical in nature is actually doing substantial work
in stancetaking. These findings offer another step toward the incorporation and
recognition of the centrality of stance in a description of Indonesian grammar.

5. Voice as a means of positioning

As a third illustration of general facets of Indonesian grammar that are used in
stancetaking, I conclude this paper with a discussion of the Indonesian grammat-
ical voice system of agent-trigger and patient-trigger clauses. I will demonstrate
that in addition to the syntactic, semantic, pragmatic, and discourse-structur-

96 Robert Englebretson

ing features attributed to these constructions in previous literature, grammatical
voice may also serve as a resource for speakers to position themselves and others
(cf. Harré and van Langenhove 1999) in a moral landscape of intentionality and
responsibility. Self-expressive stancetaking was discussed in Section 3, epistemic
stancetaking was discussed in Section 4, and the type of stancetaking at issue in
the present section concerns the stances that speakers take about moral agency
and responsibility – either of themselves or of the people whom they are talking
As with many Western Austronesian languages, the verb morphology of In-
donesian offers an array of voice and valence inflectional prefixes. Verbs may be
marked by the agent-trigger prefix meN-, (where N represents a nasal consonant
whose pronunciation is determined by the place of articulation and voicing of
the initial phoneme of the stem), which is often rendered as N- in colloquial In-
donesian; the patient-trigger prefix di-; other valence-marking prefixes such as
ber- (often glossed as a reflexive or middle-voice), ter- (whose meaning includes
nonvolitional, adversative, abilitative, and others), and ke- (also often glossed as
nonvolitional); or, as is commonly the case in colloquial Indonesian, no prefix at
all. The two prefixes I will focus on here are the agent-trigger meN-/N- and the
patient-trigger di-. Considerable debate exists in the Indonesianist literature as to
whether these prefixes represent active and passive respectively, antipassive and
ergative respectively, or something else (cf. Cumming and Wouk 1987; inter alia).
This debate is not relevant to the discussion here, so I have adopted the more
neutral terms agent-trigger and patient-trigger (cf. Cumming 1991; Englebretson
2003; Wouk 1989; inter alia). The following two examples give clause level illus-
trations of these prefixes as found in the corpus.

(16) Pencuri IU 753
Lagi mem-perbaik-i sepeda motor kan,
prog at-improve-app bike motor prt
(He) was fixing his motorcycle.

The verb in this example contains the agent-trigger prefix mem-, indicating that
the agent-argument of the verb memperbaiki ‘repair’ (in this case an unexpressed
3SG reference clear from the discourse context) is the subject of the clause, and
the patient-argument sepeda motor ‘motorcycle’ is the object.9 As suggested by
the free translation, Indonesian agent-trigger clauses often correspond to English
active voice. Following is an example of a patient-trigger clause.

Stancetaking in colloquial Indonesian conversation 97

(17) Pencuri IU 1263
Aku di-tahan sama temen-ku samping-ku.
1sg pt-restrain with friend-1sg side-1sg
My friend next to me held me back.
(lit., I was held back by my friend next to me.)

In this example, the verb tahan ‘restrain/hold back’ is prefixed with the patient-
trigger di-, indicating that the patient-argument aku (one of the 1SG pronouns
discussed in Section 3) is the subject of the clause. The agent-argument temenku
‘my friend’ is preceded by the preposition sama ‘with/by,’ which marks it as an
oblique.10 While it is true that Indonesian patient-trigger clauses appear structur-
ally similar to English passives, note that there are substantial functional differ-
ences between how each of these clause types are used in each language; therefore
a direct English translation of patient-trigger clauses can be awkward if not im-
possible, and is often not the most accurate choice. I have included a literal Eng-
lish translation in many examples, in addition to the idiomatic free translation, in
order to remind the non-Indonesianist reader of the structure of the Indonesian
clauses under discussion.
Substantial literature exists on the functions of agent- and patient-trigger
clauses in Indonesian. (See Kaswanti Purwo 1988 and Wouk 1989 for broadly
functionalist perspectives and overviews of relevant research.) General explana-
tions for their use tend to focus on syntactic factors (e.g., only the trigger ar-
gument may participate in relativization and act as the pivot in certain types of
clause combination); clusters of related semantic factors such as animacy, refer-
entiality, and individuation of the arguments, and the event structure and implied
aspect of the verb; and constellations of discourse and pragmatic factors such as
topicality, foregrounding and backgrounding, discourse transitivity, and narrative
versus non-narrative clause types. In the current section, I claim that an impor-
tant aspect of Indonesian voice has so far been overlooked. Namely, in addition
to previous explanations such as those just mentioned, the Indonesian voice sys-
tem has profound social implications for constructing moral agency, and for how
speakers position themselves and others. Following Duranti,
Agency is here understood as the property of those entities (i) that have some
degree of control over their own behavior, (ii) whose actions in the world affect
other entities’ (and sometimes their own), and (iii) whose actions are the object
of evaluation (e.g., in terms of their responsibility for a given outcome).
 (2004: 453)

This tripartite approach to human agency (intentional control of one’s actions,
ability to affect others, and moral accountability) is fully relevant with regard

98 Robert Englebretson

to the Indonesian voice system. The ready availability of agent-trigger and pa-
tient-trigger morphology facilitates the positioning of individuals as intentional
or non-intentional, as agentive or patientive, and as morally responsible and ac-
countable, or as morally not responsible and unaccountable. When viewed from
this perspective, the Indonesian voice system is far more than a means of express-
ing semantic meaning and discourse structure, as it takes on a central role in the
social worlds which speakers are constructing through talk.
In support of this claim, I offer one particularly rich stretch of talk from the
corpus, a conversational narrative which is overtly concerned with aspects of
agency and responsibility and the lack thereof. I will show how a speaker uses
agent- and patient-trigger clauses in stancetaking, throughout this excerpt, to po-
sition a pair of friends in terms of moral agency and responsibility. Because of the
lengthy nature of this excerpt, I cannot reproduce it here in its entirety. Rather, I
will summarize the excerpt, and then discuss the clauses which contain either an
agent-trigger or patient-trigger verb form.
The excerpt begins approximately 38 minutes and 15 seconds into the Pencuri
segment. Earlier, one of the speakers had commented that she had recently read
in the newspaper about thieves at the Beringharjo market (the large, traditional
Javanese market in Yogyakarta) who have been victimizing people by means of
Gendam. Gendam refers to a reputed form of black magic that combines hypnosis
along with mystical and supernatural elements. Practitioners of Gendam allegedly
can use their powers to victimize unsuspecting people in public places, by causing
them to temporarily lose full consciousness of their surroundings (e.g., the victim
gives valuables to a stranger, believing that person to be a family member), to mis-
take a worthless object for one of value (e.g., the victim buys what s/he believes to
be a watch, and discovers only after returning home that it is actually a rock), or
to believe that a poorly-made item is of exceptional quality (i.e., the victim spends
a large sum of money for something that turns out to be worth far less). One of
the important details about Gendam is that its influence only operates within a
small area around the practitioner. As a consequence, if one is shopping at the
market with a friend and notices that the friend has fallen victim to Gendam, the
unaffected person must get the victim outside of the boundaries of the Gendam
in order to break its power. The relevant excerpt for the current section consists of
speaker I (Indra) recounting an event that happened to two friends of hers named
Maya and Erika. I will briefly summarize Indra’s narrative. Maya and Erika are
shopping at the market when Maya comes under the influence of Gendam. Maya
sees some towels for sale, believes them to be big, thick, and of high quality, and
is getting ready to buy the towels. Erika notices that the towels are actually junk,
and realizes Maya is being influenced by Gendam. A struggle ensues between the
two women, as Erika tries to pull Maya away from the towels and out of the reach

Stancetaking in colloquial Indonesian conversation 99

of the Gendam. Eventually, Erika succeeds, and Maya comes far enough away
that she is no longer under its power. At this point Maya becomes fully conscious,
looks again at the towels, and realizes that they are not at all good or worth the
high price.
For purposes of the present discussion, what is especially interesting about
Indra’s narrative is her use of agent- and patient-trigger clause types. While Maya
is under the power of Gendam, Indra uses only patient-trigger clauses to refer to
her. In contrast, the only agent-trigger clauses found in this part of the narrative
are those with Erika as the agentive subject. After Maya has become fully con-
scious and is no longer affected by Gendam, she once again becomes the subject
of agent-trigger clauses too. In other words, in this narrative, Maya’s being under
the influence of Gendam correlates with an avoidance of placing her as an agen-
tive subject, but this restriction in grammatical roles is lifted once she is again ful-
ly conscious. I would suggest that the distribution of clause types in this narrative
is no accident. Rather, Indra is actively constructing a social world in which she
is using patient-trigger clauses to position Maya as a victim of Gendam, who is
thus not in control, not able to affect others, and not morally accountable for her
actions. At the same time, Indra is actively using agent-trigger clauses to position
Erika (the woman who was not influenced by Gendam) as intentional, agentive,
and responsible for removing Maya from the influence. Following is a discussion
of the specific relevant clauses from Indra’s narrative.
After Maya has been introduced into the narrative as Indra’s friend and a fellow
student of English, Indra summarizes Maya’s appraisal of the towels at the market.
These first seven IUs have already appeared above in example (5) (regarding the
use of the informal aku 1SG pronoun in constructed dialogue), and are repeated
here for convenience as example (18). These IUs serve as a general summary of the
initiating events and an orientation to Maya’s predicament at the market:

(18) Pencuri IU 2751–2757
2751 I: .. Maya itu udah ter-tarik banget.
Maya that perfv nonvol-pull very
I: Maya got really attracted.
2752 Ah murah ini,
oh cheap this
“Oh, this is cheap!
2753 aku pingin handuk tiga ribu itu.
1sg want towel three thousand that
I want this towel for 3,000 (Rupiah).

100 Robert Englebretson

2754 Padahal gede tebal itu,
in:fact big thick that
It’s big and thick.
2755 harus‑nya enam ribu lebih to.
should‑stm six thousand more prt
It should be 6,000 (Rupiah) more.”
2756 dia kan,
3sg prt
She was (like),
2757 .. mikir‑nya ah lumayan @mahasiswa @baru @gitu ya.
thought‑pos prt reasonable student new thus prt
Her thoughts were reasonable for a new student, you know.

In IU 2751, Indra describes Maya as having been “very attracted” by something.
She uses the nonvolitional ter- prefix here. Maya is not presented here as an active
agent, but neither is she being construed yet as fully patientive. IUs 2752–2755
present Maya’s inner dialogue, and also the object of her interest – the towels that
she thinks are cheap, big and thick, and worth the price. In IUs 2756–2757, Indra
assesses Maya’s thoughts as being “reasonable for a new student” – presumably
because as a new student, Maya is likely short on money and in search of a good
bargain. Indra’s first introduction of Erika into the narrative follows:

(19) Pencuri IU 2760–2761
2760 .. Terus untung-nya Erika-nya itu,
next fortune-stm Erika-def that
2761 udah ng-erti itu lho.
perfv at-understand that prt
And then fortunately Erika understood (what was going on).

Here Indra is introducing Erika into the narrative. Significantly, she does so by
means of the verb mengerti ‘understand.’ This verb is relevant for two reasons.
First, it has the agent-trigger prefix, thereby construing Erika as an agent. Secondly,
the meaning of the verb itself implies that Erika is cognizant and aware of Maya’s
situation. The next segment of the narrative presents the struggle between the two
women, as the Gendam is pulling Maya in, while Erika is trying to pull her away.

11 prt Maya that 2sg that pt-pull:away-pull:away trunc Kept pulling Maya away.. This brief excerpt contains three instances of the di. kalau di-tarik. but Erika-def that at-persist But Erika kept on persisting. and 2769) occur on the verb tarik ‘pull. pergi jauh go far 2771 berapa meter dari situ. but it is clear in the overall context that Maya is the subject argument. 2765 (H) itu udah ng-otot banget.agent-trigger prefix.’ and all take the patientive Maya as the subject.. pt-pull-pull (Maya) got pulled in and pulled in. 2769 .. . the subject of these clauses. All three of the patient-trigger pre- fixes (IUs 2764. 2766 Si Maya itu kamu itu di-singkir-singkir kee-. that perfv at-persist very But (Erika) kept really persisting. who is being persistent in co- ercing Maya away from the towels and away from the influence of Gendam.) Both of the agent-trigger prefixes (IUs 2765 and 2768) occur on the verb otot ‘persist. Maya was pulled away. 2766. if pt-pull 2770 . (The actual NP is ellipted in two of these instances. Stancetaking in colloquial Indonesian conversation 101 (20) Pencuri IU 2764–2771 2764 di-tarik-tarik.’ indicating it is Erika. The following example contains one more instance of Maya as the subject in a patient- trigger clause. (lit.) 2767 Sampai ber-antem sendiri ber-dua itu. and two instances of the N. a few meters from there.patient-trigger prefix. far away. until mid-quarrel self mid-two that Until the two of them started fighting with each other. as well as speaker L’s co-construction of the upshot of the narrative. 2768 (H) Tapi Erika-nya itu ng-otot. namely that Erika had to take her outside of the boundaries of the Gendam. how:many meter from there She was pulled.

after they had passed the boundary of a few meters.. as Erika is addressing her: (22) Pencuri IU 2792–2795 2792 . (lit.102 Robert Englebretson (21) Pencuri IU 2779–2782 2779 I: . 2780 L: [Pokok-nya di luar batas. Langsung kamu ini gini-gini. In IU 2792. 2782 .’ this clause again construes Maya as the subject argument in a patient-trigger clause.patient-trigger prefix on the verb omong ‘speak’ and the oblique agent phrase sama Erika ‘by Erika. en- couraging her to keep coming with her. di luar batas tu]. yeah. at outside limit prt Outside the boundary]. 2781 I: [Sampai itu ho-o.” 2793 pokok-nya di-omong-in sama Erika itu.) 2794 Terus udah lewat batas berapa meter itu. In the following example. Maya is once again presented as the subject argument in a patient-trigger clause. main-stm pt-talk-app by Erika that Erika said to her. until that aff I: [To there. Again.. . main-stm at outside limit L: [The important thing is. 2795 . Maya is presented in a patient-trigger clause in IU 2779 as being pulled. Maya-def that I: (Maya) was pulled until Maya started getting angry. Jadi sadari itu lho Maya-nya become conscious that prt Maya-def Maya became conscious.. Because of the di.. Udah di-ta= rik sampai marah-marah perfv pt-pull until angry-angry Maya-nya itu. next perfv pass limit how:many meter that Then. outside of the boundary. (she) was told by Erika. direct 2sg this like:this-like:this “You just keep going like this. Indra is reporting the words that Erika allegedly said to Maya..

the distribution of agent. yeah yeah. Stancetaking in colloquial Indonesian conversation 103 In IU 2794. in contrast to the agent-trigger clauses she uses with Erika. Indra shifts her positioning through the use of agent-trigger clauses. in:reality towel-def that bad In reality the towel was bad. and not in fact what she had thought they were when she was under the influence of Gendam. as Maya is no longer portrayed as the victim of Gendam. and is responsible for the actions of seeing and looking. at-look again to that-def She looked at it again. Indra (the narrator) positions her as patientive and not responsible for her actions. 2798 n-engok lagi ke itu-nya. When Maya is under the influence of Gen- dam. And when Maya ‘sees’ and ‘looks. in:reality bad at-see again prt In reality it was bad when she looked at it again. Indra reports that they passed the boundaries of the Gendam. the verb lihat ‘see’ in IU 2797. 2799 . through the use of patient-trigger clauses for Maya. Once Maya has come outside the bound- aries of Gendam. (23) Pencuri IU 2796–2799 2796 (H) O ya ya.” 2797 ternyata jelek nge-lihat lagi to. and IU 2795 provides the resolution of the narrative – when Maya jadi sadar ‘becomes conscious. in part. as shown in the following example. oh yes yes “Oh.and patient-trigger clauses in the unfold- ing of this conversational narrative illustrates how speakers may use the general grammatical resource of voice in Indonesian to position themselves and others in terms of agency and responsibility. Similar to Duranti’s (1990.’ she recognizes that the towels were “in reality” bad. In sum. as Indra begins to construe Maya as the subject in agent-trigger clauses. This indicates that Maya is once again intentional and agentive. Indra achieves this positioning. are both marked as agent-trigger. Ternyata handuk-nya itu jelek. implying that Maya is once again able to control her actions and may now be held accountable for them..’ This marks a shift in positioning. 1994) analysis of erga- . and the verb tengok ‘look’ in IU 2798. and this shift in positioning is also indicated by a shift in grammar. who is not under the influence of Gendam. Then. The first IU of this example presents Maya’s reported speech: two acknowledgement tokens to presumably indicate that she is now aware of her situation.

more tra- ditional approaches to voice in Indonesian. I addressed the complex system of Indonesian first-person-sin- gular referring expressions. to actively construct their social worlds. and it is not surprising. as a grammatical resource which Indonesian speakers may use for specific stance purposes. the narrative analyzed here suggests that Indonesian voice may have similar social implications. in which grammatical relations (specifically ergative) allow speakers to assign blame or responsibility. In Section 3. this “careful choice of words” must also be understood to encom- pass the choice of agent. 1999) entitled The grammatical marking of stance.104 Robert Englebretson tivity in Samoan public discourse. I argued that these forms are not simply a determin- istic reflection of a priori social variables such as regional origin of the speaker . if so. how? ) but also from the point of view of the type of persons and the type of world that speakers build through their typi- cally unconscious but nevertheless careful choice of words. or discourse factors. 6. this paper has offered a discussion of three facets of Indonesian grammar that speakers use in constructing their social worlds through stancetaking. However. How- ever. which highlight semantic. for instance. Many English grammarians and linguists have been engaged in such an enterprise over the course of the past several decades. The current section in no way seeks to discredit or minimize other. As a first step toward such a goal. Conceiving of voice as a means of positioning the self or oth- ers is not contradictory to approaching it from these other angles as well. is the agent of this event expressed and. placing central emphasis on the social nature of this phenomenon provides a potentially fruitful line of inquiry into how speakers use grammar in stancetak- ing. interest in stancetaking and grammar has not seemed to have had much influence on field linguistics or descriptive grammars of other languages.  (2004: 466) In Indonesian.g. As Duranti notes about agency. and suggested that one goal of the descriptive gram- marian should be to account for how speakers use the grammatical resources of a language to carry out the activity of stancetaking. Conclusion I began this paper by noting the pervasiveness of stancetaking as an activity real- ized through language.or patient-trigger clause. to find a chapter in the Longman Grammar of Spoken and Written English (Biber et al. syntactic. a possible direction for future research is to expand our horizon of theoretical and empirical research to include an understanding of these phenomena not only from the point of view of the type of information that is being encoded (e. I then tied this in with the functionalist view of a usage- based approach to grammar.

The epistemic functions of -nya constructions fall into three categories: evidentiality (how a speaker knows about the current utterance). Thirdly. forms which have previously only been characterized as part of a referen- . This kind of self-expressive stancetaking enables speakers to highlight facets of the self. By having put these three areas on the table. What I offer here is a first glimpse of the fruitful nature of consciously bringing stancetaking into the realm of grammati- cal description. nor with stancetaking through conversa- tional sequences and dialogic interaction. dem- onstrating how the speaker uses grammatical voice to attribute agency. 2001). each is highly frequent and productive in colloquial Indonesian. I hope to have shown that when we approach grammar with this perspective in mind. as relevant in the ongoing conversation. assessment of interactional relevance (how a speaker values the current utterance in light of the ongoing interaction). each represents a particular kind of stancetaking: self-expressive stance for the 1SG reference forms. I demonstrated that much of its work in everyday interaction is to provide a means by which a speaker can indicate epistemic stance – specifically toward the talk at hand. and affect (the speaker’s emotional attitude toward the proposition expressed in the current utterance). I hope to encour- age other linguists to take seriously the question of how grammatical resources are functioning in the service of stancetaking in other languages. which has traditionally been characterized solely in terms of syntax. semantics. responsi- bility. and a broader-level stance in terms of moral positioning for the agent-trigger and patient-trigger clause types. Sec- ond. and discourse-pragmatics. speakers actively use these vari- ous forms to construct and reflect locally-relevant aspects of social and personal identity such as distance. these only begin to scratch the surface of how Indonesian grammar is involved in stancetaking. or nominalization. In Section 4. is additionally a social resource for stancetaking that speakers can use in order to position themselves and others in terms of moral agency and responsibility. or toughness. I have focused on these three aspects of Indonesian grammar for several rea- sons. While this has traditionally been characterized as a marker of possession. epistemic stance for the -nya constructions. definiteness. nor with modals and adverbs. I have not dealt at all with clause-final particles (cf. these three grammatical resources are formally quite different from each other: pronoun choice. I argueed that grammatical voice. and the lack thereof to the protagonists in the narrative. and voice constructions. casualness. In this paper. but rather. Yet. Section 5 explored the social nature of Indonesian agent-trigger and patient- trigger clauses. a clitic. First. Stancetaking in colloquial Indonesian conversation 105 or formality level of the speech event. Wouk 1998. I il- lustrated this approach with reference to a specific conversational narrative. I took up the issue of the Indonesian -nya clitic and its partici- pation in epistemic constructions.

I am especially grateful to the individuals in Yogyakarta.106 Robert Englebretson tial. of course. a perspective that can offer fresh insight into the very motivations behind grammar itself. Notes 1. Due to the uncontrolled nature of the data. Secondly. both of whom have provided invaluable encouragement and suggestions at various stages of this project. for presenting this maxim: ‘Every utterance enacts a stance. and is by no means restricted to 1SG: it occurs with all persons of pronouns. and especially to the three research assistants with whom I worked through the data. a direct statistical correlation for each of the 1SG reference forms across all six speech events is not feasible. solely my fault and responsibility.or third-person reference. A reviewer has raised the provocative question as to the role of non-self-referral in the en- actment of stance. There are actually two plausible structural analyses of this clause. I will not be discussing second. 2. All shortcomings of this paper. the categorical nature of the 1SG forms produced in the radio call-in show as com- pared with the other five speech events is nonetheless striking. In sum. 4. who allowed themselves to be recorded for the col- loquial Indonesian corpus. and any misunderstandings of Indonesian language or culture are. and their social back- grounds. the pronoun gua ‘I’ would be the subject of the clause. Ewing 2005). the confounding variables are too numerous to allow reliable conclusions to be drawn. followed by the subject gua ‘I. where nggak ada duit is the predicate.’ 3. this view presupposes that arguments of a clause are normatively expressed – which may not in fact be true for colloquial Indonesian at all (cf. I wish to credit Jack Du Bois. and helped to clarify several details. argument ellipsis is widespread in this language. The comments by the two anonymous reviewers were likewise valuable.’ According to this interpretation. the topics. I would like to thank the participants in the 10th Biennial Rice Linguistics symposium for comments and discussion on an early draft. nor the first-person-plural inclu- sive and exclusive pronouns. and the overall focus construction has the pragmatic force of placing strong emphasis on the 1SG . propositional system for expressing and managing information can be seen from a new perspective. a discussion of which is far afield of the current paper.’ A literal English translation of this clause would then be: ‘My money doesn’t exist. cross-linguistic project on argument ellipsis. and are not limited to self-reference or individual identity in the same way as the different 1SG pronominal forms seem to be. duit gua is a possessive NP consisting of the head noun duit ‘money’ followed by the 1SG form gua as a possessor: ‘my money. However. As with all my work on Indonesian. which the author is beginning to address in the context of a larger. I would also like to thank Susanna Cumming and Sandy Thompson. This is indeed an important issue. as well as with full noun-phrases.’ Another possible analysis of this clause would be as a type of so-called focus construction. 6. Because the speakers vary con- siderably among themselves in terms of the amount of talk. In the one intended by this example. First. considering this a colloquial Indonesian “zero 1SG form” in opposition to the overt pronominal forms is problematic for at least two reasons. in several unpublished conference talks. they crosscut all pronominal and full forms. The use or ellipsis of arguments does indeed have social and interactional implications – but whatever these implications are. 5.

2001. G. 1999. and.L. Usage-Based Models of Language. (eds. D.” Discourse Studies 7 (4–5): 585–614. -nya is said to mark third-person-sin- gular possession. Prescriptively. R. I will refrain from commenting further on this here.). See Englebretson (2003: Chap.). Biber. 8. Johansson. London: Longman. Brown. 11. and examples to support the analysis of these constructions as adverbials. Benwell. however. and Gilman.L. MA: MIT Press. and Finegan. Conrad. K. T. 1988. Bybee. Stanford: CSLI.. I see no need to choose one or the other structural interpretation here. 9. I am aware that the terms subject and object are debatable and problematic in the context of Indonesian grammar. Amsterdam: John Benjamins. and Finegan. “The pronouns of power and solidarity. D. and Hall. 10. J.). they both ex- ist here simultaneously. and Finegan. 253–276. and Stokoe. Bybee. “Styles of stance in English: Lexical and grammatical marking of evidentiality and affect. The extended use of -nya to mark possession in all persons and numbers is another well-known difference between formal and colloquial varieties of Indonesian. 1989. and because the debate over Indonesian grammatical relations is not central or relevant for the overall point of this section. and Kemmer. S.” Both analyses are plausible. 2006. S. just as real-life utterances are often uncertain and multifaceted for interlocutors themselves in realtime con- versation. E. Bucholtz.” In Style in Language.J. “Adverbial stance types in English..” Text 9 (1): 93–124. 2006. Frequency and the Emergence of Linguistic Structure.” Language 82 (4): 711–733. Biber. A.’ The use of sama ‘with’ here is typical of colloquial Indonesian. 5) for further discussion of the syntactic characteristics and distribution of epistemic -nya constructions. Edinburgh: Edinburgh University Press. to facili- tate the easy understanding of these examples by non-Indonesianist readers. References Barlow. I am aware of the disfluencies and ambiguities in this IU. so as not to distract from the discussion at hand. which is generally con- sidered a nonstandard use of this clitic. the preposition marking the oblique argument in a pa- tient-trigger clause is oleh ‘by. (eds.. I don’t have any money. . E. Biber. “From usage to grammar: The mind’s response to repetition. I am presenting this example as an instance of gua as the possessor in a possessive NP. E. E. Leech. I am using these terms here solely for the sake of convenience. Cambridge. This example shows -nya marking third-person-plural possession. D. and the difficulty in glossing it in a more accurate way. 1960. Stancetaking in colloquial Indonesian conversation 107 subject “(As for me). Sebeok (ed. J. For purposes of exposition. 1999.” Discourse Processes 11 (1): 1–34. S. “Identity and interaction: A sociocultural linguistic approach. The Longman Grammar of Spoken and Written English. and Hopper. 7. M. M. In formal varieties of Indonesian. 2005. P. although I recognize the other structural possibility as well. Discourse and Identity. B.

Cumming. Vol. Cambridge. T. R. J.). Kärkkäinen. Searching for Structure: The Problem of Complementation in Colloquial Indonesian Conversation. Epistemic Stance in English Conversation: A Description of Its Interactional Functions. L. An Indonesian-English Dictionary. P. Oxford: Oxford University Press. Amsterdam: John Benjamins. S. S. A. NJ: Lawrence Erl- baum.” In Evaluation in Text: Au- thorial Stance and The Construction of Discourse. J. 1993. London: Routledge. M. R. 1988. 1990. 1990. Givón. On Understanding Grammar. S. 1997. Vol. Thompson (eds. 1989. S.A. and Cumming. Wolff and James T. S. Shibatani (ed.108 Robert Englebretson Conrad. 1. Ithaca: Cornell University Press. Philadelphia: University of Pennsylvania Press. Paolino. Dixon (ed. New York: Mou- ton de Gruyter.J. 1988.). Oxford: Oxford University Press. 2. D. 1999.C. J. NJ: Lawrence Erlbaum. B. Functional Change: The Case of Malay Constituent Order. “Emergent grammar. C. Hunston. Hillsdale.D. Berkeley: University of California Press. Fox. “Adverbial marking of stance in speech and writing. 1998. T. 56–73.W.W. Lampert (eds. London: Sage. Du Bois. Amsterdam: North-Holland. Hopper. 195–241. Tomasello (ed. Edwards and M.A. Duranti. Territory of Information. 2003. Adelaar and N. A. New York: Mouton de Gruyter. Amsterdam: John Benjamins.” In Passive and Voice.E. Cumming. 1999. J. D. Echols. “Agency in language. Structure and Style in Javanese: A Semiotic View of Linguistic Etiquette. 2000. “Discourse and grammar. Cumming. “Social interaction and grammar.” In The Austronesian Languages of Asia and Mada- gascar.” In Talking Data: Transcription and Coding Methods for Language Research. Amsterdam: John Benjamins. Errington. E. Schuetze-Coburn.” BLS 13: 139–157.). J..A.J. R. P.). van Dijk (ed. R. Positioning Theory: Moral Contexts of Intentional Ac- tion. S. 74–101. 45–89. 2003.). S. MA: Blackwell. Ewing. J. 2004. 1997.). Malden. “Colloquial Indonesian. Shifting Languages: Interaction and Identity in Javanese Indonesia. Englebretson.” In The New Psychology of Language. S. 3rd ed. (revised and edited by John U. 271–297. 112–137. T. Thomp- son (eds. Himmelmann (eds. and Ono. S. 119–143. and Thompson. and Harré. New York: Academic Press. Mahwah. Grammar and Conceptualization.. and Biber. From Grammar to Politics: Linguistic Anthropology in a Western Samoan Vil- lage. Kamio. Malden. 2002. 1994. Langacker. Duranti (ed. Duranti.). .M.). Pronouns and People: The Linguistic Construction of Social and Personal Identity. 2000. A.M.” In Evalua- tion in Text: Authorial Stance and the Construction of Discourse. A. S.” American Ethnologist 17 (4): 646–666. Hunston and G. R.W. 2005.. A. and Sinclair.” In Discourse Studies: A Multidisci- plinary Introduction. Mühlhäusler. F.J. M. MA: Blackwell. “Is there ‘discourse ergativity’ in Austronesian languages?” In Studies in Ergativity. New York: Cambridge University Press. Hunston and G. and Shadily. Kaswanti Purwo. 451–473. 1987. MA: Blackwell. Collins in cooperation with Hassan Shadily). and van Langenhove. A. 227–258. B. “Voice in Indonesian: A discourse study.” In A Companion to Linguistic Anthropology. 1979. “Outline of discourse transcription. Errington. Harré. H. and Wouk. Ford. “Politics and grammar: Agency in Samoan political discourse. Duranti. “A local grammar of evaluation. 1991. Amsterdam: John Benjamins.). 1987. M. with a Focus on ‘I think’.A.

” Multilingua 17 (4): 379–406. 2005. 311–332. MA: St. D.). 1996. NJ: Lawrence Erlbaum. F. S.” In Direct and Indirect Speech. Thompson. The New Psychology of Language: Cognitive and Functional Ap- proaches to Language Structure. Tannen. Tomasello. The Impact of Discourse on Grammar: Verb Morphology in Spoken Jakarta In- donesian. 1993) Each transcript line represents a single intonation unit. Department of Linguistics. “Solidarity in Indonesian conversation: The discourse marker ya. January 8–11. “Dialect contact and koineization in Jakarta. Sneddon. 1989. S. Sneddon. E. Wouk. 2002.). 213–271. and Thompson. Stancetaking in colloquial Indonesian conversation 109 Ochs. Stance in Talk: A Conversation Analysis of Mandarin Final Particles. M. Series D. and are followed by a colon. Javanese Influence on Indonesian. (ed.D. Wu. Rampton. Wouk. Northampton. 7.. Stubbs. 1989. “Solidarity in Indonesian conversation: The discourse marker kan.W. “Variation in informal Jakartan Indonesian: A quantitative study. J. Dialogue. and Imagery in Conversational Discourse. Vol.A. 1995. and Thompson. No. 2nd ed. 2. Oxford: Oxford University Press. Mahwa. Appendix A: Summary of transcription conventions (Adapted from Du Bois et al. G.).” In Evaluation in Text: Au- thorial Stance and the Construction of Discourse. D. R. E. Speaker labels appear in uppercase. M. Ph. Jerome Pub. S. New York: Cambridge University Press. “Introducing constructed dialogue in Greek and American conversational and literary narrative. 1998. Wouk. Poedjosoedarmo. Materials in Languages of Indone- sia. Schegloff.N. Talking Voices: Repetition. F. Tomasello. “What can conversation tell us about syntax?” In Alterna- tive Linguistics: Descriptive and Theoretical Modes. (eds.-J. Amsterdam: John Benjamins. F. University of California. Hunston and G. and Hunston. Indonesian: A Comprehensive Grammar. M. 1996. Mahwa. NJ: Lawrence Erlbaum. S. Vol.). (ed. “Evaluation: An introduction. Wouk. Indonesia. 1982. dissertation. Los An- geles. Amsterdam: John Benjamins. “‘A matter of prolonged field work’: Notes towards a modal grammar of Eng- lish. Crossing: Language and Ethnicity among Adolescents. 2000. 1986. New York: Routledge. Canberra. 2002. . New York: Cambridge University Press.).” Language Sciences 21 (1): 61–86. 1–27. 1. F. Ono.N. Interaction and Grammar. B.A. The New Psychology of Language: Cognitive and Functional Ap- proaches to Language Structure. 1999.” Journal of Pragmatics 33 (2): 171–191. Thompson (eds). New York: Mouton de Gruyter. Coulmas (ed. 2004. J. F. Davis (ed. S.A.” Paper presented at the Ninth International Conference on Austronesian Linguistics. 2001.” Applied Linguistics 7 (1): 1–25. P. 1986. Canberra: Pacific Linguistics. 1998. Tannen. T.

3pl first-person plural. third-person plural aff affirmative backchannel app applicative suffix at agent-trigger prefix def definite ex existential mid middle voice neg negative morpheme nom nominalizer nonvol nonvolitional perfv perfective aspect pos possessive prog progressive aspect prt discourse particle pt patient-trigger rel relativizer stm stance marker (epistemic -nya) . Long pause (TSK) Tongue click (H) In-breath Appendix B: Abbreviations and glosses 1sg.110 Robert Englebretson Simultaneous speech is indicated by square brackets [ ] (not aligned because of glossing dif- ficulty). Material in angle-brackets < > is codeswitching. or Arabic. second-.. 2sg. and third-person singular 1pl. English.. Continuing intonation contour ? Appeal intonation contour ‑‑ Truncated intonation unit ‑ Truncated word @ One pulse of laughter % Glottal stop = Prosodic lengthening . Final intonation contour . usually Javanese. 3sg first-. .. Short pause .

in the conversations used for this study.1 Introduction to the study Generality of meaning has been of interest to scholars working primarily in the areas of semantics. they are not categories based on reported properties shared by multiple entities or ideas in the world. These are expressions of position and attitude that are rel- evant both to individual speakers (their subjective uses) and to relational activities among participants (their intersubjective uses). The present study examines stance-related functions of generalizations in English conversations. things. activities. Van Leeuwan (1995) suggests that generalization and underspecification of agency and social actions are important considerations in critical analyses of texts. and rela- tions. and language change. however. These findings suggest that generality of expression also has influences on social and expressive activities in conversational discourse. Instead participants tend to generalize (index classes of experience) . included with orientation and attitude. To date. generalizations based on these classes are not inductively constructed or construed – that is. Introduction1 1. attitudes. generalization and generality of meaning in discourse have not received a great deal of attention. Subjective and intersubjective uses of generalizations in English conversations Joanne Scheibman Old Dominion University 1. Additionally. The expressions designate types – as opposed to tokens – of people. Berman (2005) lists generality of reference as one of three dimensions of discourse stance. For example. These utterances index and evalu- ate classes in American English conversations. reactions. even though there is evidence that speakers’ uses of general meanings have discursive and pragmatic functions. language acquisition. And Zhang (1998) points out that general expres- sions rely on Grice’s maxims for their resolvability.

as a continuum – with linguistic features at one end categorized as personal and intimate. the con- versational generalizations are “personalized” in a different sense: they convey subjective meanings (primarily evaluation) in interactional contexts. generalizations. and impersonal (e. affective).g. deontic. those with specific agent subjects such as I). and stance In a description of a crosslinguistic project investigating the development of stance over time based on spoken and written narratives. and the use and interpreta- tion of these utterances is keenly sensitive to interactive contexts.. places. their uses in conversations highlight collaborative and interactive aspects of stancetaking.g.g. and times mentioned in the text. Berman (2005: 107) character- izes discourse stance as having three related dimensions: orientation (relations be- tween sender. and recipient). and gen- erality (of reference and quantification). Indeed. even though utterances exhibiting the formal properties of personal and specific expressions as described by Berman are not included in this study (e. or inclusive.. generic (e. participants’ classifications of people and phenomena for purposes of predication have expressive and social consequences.g. demonstrate solidarity with one another. worse things could happen to the poor guy. animate societal discourses. and all these teachers are com- ing in and saying. 1. Moreover. and authorize opinions – activities which. As analyses offered in this paper suggest. attitude (epistemic.. The category generality “concerns how relatively general or specific reference is to people. in the sense that they typically index commonly held beliefs.” and in these studies textual features are coded according to three values: personal and specific (e. The functions and properties of these generalizations also overlap with Berman’s orientation and attitude dimensions of discourse stance. The specific types of generalizations examined here are utterances with gram- matical subjects that refer to classes or groups (e. Berman operationalizes referential generality.112 Joanne Scheibman based on personal and social expectations and beliefs. Because construction and interpretation of meaning and refer- ence in conversation occur interactively.. People/We/You tend to think). because these utterances are naturally broadening. I/my parents think. one . in many cases. The set of utterances defined as generalizations in the current study includes those in Berman’s generic group and also contains expressions with lexical noun- phrase subjects that refer to general classes. and those at the other end viewed as impersonal and distant. It’s well known) (Berman 2005: 108).g. generalizations in these conversations are used by participants to evaluate.2 Generality. For example. the French are basically a northern people). then. text.. My/This boy’s father made me/him apologize).

2005). aspects of individual stancetaking (evaluation. then. (2) which gives them broadening. psychological. a message or topic. Stance in discourse is a relational notion. Regarding the broadening properties of generalizations. something else – prototypically. don’t only position themselves with respect to propositional material. since these activities presuppose a response to some linguistic. a central type of subjective expression and stancetaking. Reilly et al. One might argue that all dis- cursive and social practices reflect stancetaking. or inclusive. expressiveness) are framed as subjective. Berman 2005. Their unre- markable interpretation and use in conversations. or a po- sitioning relative to. beliefs).2 Depending on what and whom interactants position themselves relative to (e. or sociocultural (relevant to general beliefs of people as members of com- munities). Stancetaking is a reaction to. the subjective and intersubjective functions of generalizations are typically overlapping or simultaneous. functions in discourse. or social reference point (Langacker 1993). stances can be collaboratively constructed and jointly held. In the organization of this study. While it is not unusual for studies of stance to focus on the use of linguistic features by individual speakers and writers to represent events and entities in the world (e. Biber and Finegan 1989. Speakers and groups of speakers. this paper is informed by the understanding that speakers also express stance by allying themselves with (or sometimes separating themselves from) one another. to varying degrees they are underspecified. because these utterances are referentially general. Generalizations in English conversations 113 of the consistent findings of the analyses presented here is that in interactive con- texts. discourse participants. utterances. In the conversational extracts. how- ever. however. objects. the generalizations function evaluatively in their conversational contexts. suggest that partici- pants elaborate the meanings of these expressions by making reference to shared knowledge and to interactive contexts. In these ways. and to sociocultural beliefs. they also situate themselves in relationship to one another. Berman suggests that stance “reflects a key facet of human discourse in general: the fact that any state of affairs in the worlds of fact or fantasy can be described in multiple ways” (2005: 109). It is also the case that construction of an indexical class by a speaker is an expressive act. to expectation.. aspects of intersubjectivity in discourse.g. The predicates occurring with general subjects in these utterances are often evaluative. Generalizations have two related characteristics that contribute to their social and expressive force in English conversations: (1) they make reference to (and as- sess) general classes. interactive (relevant to local discourse activi- ties). for analytical purposes stance might be classified as individual (construed as relevant to a speaker’s position in discourse).g. general assertions and evaluations are likely to be com- .. In these conversations. however. groups of people. while interactive activities and the sharing of beliefs contributing to collaborative expression of evaluation and attitude are discussed as intersubjective.

example 8). speakers may intro- duce people. and situations in general terms in order to orient participants to their stories (Labov and Waletzky 1967).114 Joanne Scheibman mon ground among interactants (Clark 1996). in these conversations for a speaker’s stance toward a generally construed class (of people. and I-- and I think I’m starting. the subject noun phrase the people I work with delineates a group of people who are characterized by the speaker as sometimes being really stupid. (1) I don’t know. in particular. relations.) to reflect evaluations that are shared by other discourse partici- . the evaluative stances expressed by speakers using these utterances with general subjects tend to reflect meanings and expectations that are presupposed by oth- er participants in the conversations. For example. Furthermore. Additionally. It is not unusual. In structuring conversational narratives. and its interpretation by interactants. because of the inclu- siveness of these utterances. for example. places. occur principally in relation to the speaker’s expressed stance toward her coworkers. a speaker broadens the domain of her assertion. It can also be the case that by expand- ing a reference class subsumed by an evaluation using a generalization. the indexical category denoted by the people I work with is a common one in mainstream American English-speaking communities. evoking this discursive class. to figure them out a little bit more. an intersubjective activity that contributes to the collaborative construction of stance in conversations. The speaker’s generalizing her opinion to the entire group of people she works with – as opposed to simply selecting the relevant individuals – is functional in the conversation. it invites participants to fill in and elaborate the meaning of this expression to include. and as such. 1. the people I work with. in (1). Since it is unlikely that the speaker has assessed the intelli- gence of all of the employees of the medical laboratory where she works. etc. the familiar and sometimes familial attitudes people hold toward this unique group of people they see on a daily basis – an interpretation that is compatible with the speaker’s evalu- ation of the group in the generalization in (1). then. thus implicitly augmenting its evidential weight or appeal in the interaction (see. sometimes are really stupid.3 An example of a conversational generalization In the conversational extract in (1). generalizations can be used as expressions of solidar- ity.

Section 6 presents summaries and conclusions. Generalizations in English conversations 115 pants. In conversational contexts. What is paradoxical about generalizations found in the conversational data is that because these utterances make reference to general classes of phenom- ena. because in Western cultural discourses the mak- ing of general statements about the world is an activity that can be elevated or deprecated. and Section 5 illus- trates two ways in which generalizations can augment speaker stance in interac- tive discourse. Operationalizing the term based on its multiple uses has been one of the most challenging aspects of this project.. enriching of meanings of these expressions for participants) and interaction (e. generalization itself is subject to social valuation.. The next section surveys popular and scholarly uses of the term generaliza- tion itself. the best things are German) are often treated as matter-of-fact descriptions of situations in the world.. subjective construction and their intersubjective treatment as statements of norms – inform the analyses offered below. That is. Sections 4 and 5 present analyses of generalizations from a conversational corpus. Generalizations 2. demonstrating solidarity by acknowledging shared stance and attitude). In spite of this appearance. gen- eralization is viewed negatively. And finally. This indexing of jointly held beliefs by these general utterances has conse- quences for both interpretation (e. leading to a working characterization of the concept to be applied to generalizations in the conversational data. utterances containing general subjects (e. In some cases. In high school and university writing classes in the United States. Furthermore. Section 3 describes and illustrates the formal properties of generalizations in the database used for the study.g. then.g. generalizations in conversational usages are necessarily selective in the sense that what speakers choose to generalize about is contingent on expressive and interpersonal exigencies of both the local con- text and larger culture. students are routinely directed to provide specific examples to support their generalizations. which might be defined as ‘unsubstan- . for example. These two characterizations of generalizing utterances – their local. these utterances can be used by interactants as expressions of stance and solidarity – to evaluate or to show alliance with other participants by confirming adherence to particular so- cietal discourses. there is a sense in which they can be viewed by interactants as proposition- ally powerful. Section 4 discusses generalizations that function as shared evaluations.1 Popular and scholarly characterizations of generalization The meaning of generalization varies by context and discipline.g. 2.

Processes of generalization figure prominently in linguistics research. in statements of principles..3 What unites the meanings of the processes of generalization in these language studies is that in each case there is a weakening of formal and semantic specificity of individual linguistic expressions based on experiences of use. and acquisition of grammar as generalizations over specific usages. which contributes to the development or change of linguistic patterns. On the other hand. bolding in original). The generalizations that have value in Western cultures tend to be those that are overtly inductive. such as those contributing to the system of argument structure in English (Goldberg et al. general statements based on implicit or insufficient data tend to be viewed unfavorably (e. children are assumed to be learning generaliza- tions (abstract patterns of form-meaning correspondences) when acquiring the semantics and grammar of their first language. of context or process).. And in language acquisition research. individual grammaticizing elements have been analyzed as undergoing processes of generalization of meaning (reduc- tion in semantic specificity) accompanied by generalization (expansion) of the item’s contexts of use (e. is that they refer to the construction or highlighting of a class (e. the prominence and longevity . tend to be descriptions of cognitive processes relevant to language use and change. change.g. warrant the label bigotry.’ In contemporary public discourse in the United States and else- where.. however. but only in degree of specificity” (1987: 58. Bybee and Pagliuca 1985. analysts discuss technical aspects of the formation. weaken- ing of lexical meaning of grammaticizing elements... In functional and cognitive usage-based approaches to morphosyntax. those conclusions that have explicit links to specific facts or cases..g. which differ from other sym- bolic structures not in kind. The generalizations illustrated for linguistics. a theoretical principle. an opinion) (Bouacha 1995). What these different treatments of general- ization share.g. and in some contexts. 1994).g. that is. 2004). with a concomitant widening of applicability (e. In contrast.g. The generalizations of scientific theorists are analytical con- clusions based on particular types of data or texts. generalizations over the activities and attributes of (usually marginalized) groups of people are also viewed negatively. a grammatical process. positive and authoritative treatment of general- izations is apparent in Western academic and scientific disciplines in which gen- eralizations have an influential role in construction of theories (e. stereotypes).g. Clearly the generalizations presented in this survey are of different types and have distinct goals. In studies of language change. often accompanied by an analytical focus on the attenuation in specificity or salience of individual elements that are in a relation to the class (e. Kafura 1998: 26). for example. on the other hand.116 Joanne Scheibman tiated opinions. Ronald Langacker characterizes grammatical patterns as “schematic symbolic units. Bybee et al.

In academic argumen- tation and theory-building. they are socially situated.. behavior.g. and he did a blood test) or oblique predicates (e. the common details that inform the construction of speakers’ generalizations are often inexplicit and covert.g. I found that a relatively small percentage of tokens in a conversational corpus could be described as referentially specific. 2. define the term in a similar way.g. These more specific utterances are most frequently expressions with third-person- singular subjects with human referents occurring with past-tense material predi- cates (e. While it is the case that conversational interactants can (and do) engage with one another at various levels of specificity. despite the attention such expressions enjoy in linguistic theories and practices (Hopper 1997. Kafura states that “[g]eneralization identifies commonalities among a set of entities. itself. However.g. Because generalizations are common linguistic practices in discourse. opening birthday pres- ents). as illustrated in this section. Van Leeuwan suggests that generalization is “an important issue in critical discourse analysis.g.g. For example. my test was Friday). not all processes of generaliza- tion designate classes that are based on shared properties of multiple entities (e. More frequent in these data are utterances that convey general meanings. which tend to highlight membership criteria and relationships among category members. conversational interaction). for a critique of this view)... formal treatments of generalization. For example. as texts which are mainly concerned with legitimizing or de-legitimizing actions and reactions tend to move high up on the generaliza- . In contrast to treatments of categorization.. They tend to cluster in descriptive narratives and in settings in which participants are involved in activities in which objects are focal (e. such as those used in software and database design. the maligned generalizations of the composition student). Scheibman 2002. the process of generalization focuses on the forma- tion of the class. The commonality may be of attributes. or both” (1998: 26). Analogous to the way membership in classi- cal categories has been delineated based on shared properties of members (see Lakoff 1987..2 Generality of expression in English conversations Expression of general meaning is ubiquitous in English conversations. or category. But in non-scientific regis- ters and genres (e. variation in generality and specificity in the representation of social groups and actions in discourse can be ideologically con- sequential. in previous analyses of English conversation.. the authority of generalizations is explicitly tied to common findings of individual studies and analyses.4 Generalization. is a type of categorization. Generalizations in English conversations 117 of scientific principles over time relative to the individual data analyses that sup- port them). Silverstein 1976). then. e.

for this study it was necessary to select for analysis an utterance type that consistently conveyed generality of meaning in the conversational data. . Cauliflower does well here) are quite common in naturally-occurring discourse (1995: 47). suggesting that the clearest difference is that strict generics (e. beliefs. are classified as general subjects (Section 3. 3.g. and times” (2005: 109). and aspectless. generics. The analyses in the present study do not take into account differences between general and ge- neric expressions. Based on cross- linguistic studies on the development of stance. and attitudes that interactants generalize over in collaborative contexts are both sensitive to and influence interaction. Lyons (1977: 193–194) states that generic propositions are tenseless. the classes indexed by conversational generalizations might be said to be “empty” categories in the sense that membership is often im- plicit and covert. to time. systematic semantic comparison is difficult. whereas general referring expressions can occur in sentences that express time-bound propositions.. Generalizations. are expressions that lack specific reference (e. The data presented here suggest that speakers use these general- izations to evaluate and strengthen stance.1 Generality of meaning. as delimited here. Rather. all subject noun phrases in the database that designate classes. including the so-called generic pronouns you and they.. as distinct from general. Data 3. event) and the formal presence of a speaking subject – both of these being properties of generics as well (Bouacha 1995).118 Joanne Scheibman tion scale. Beavers build dams) are primarily found in scholarly metalanguage. timeless. Because of this formal and functional heterogeneity (illustrated be- low for English). Lyons (1977: 194) also notes that the status of “generic. Bouacha agrees that it is not easy to distinguish generalizations from generics. whereas generalizations (e.3).g. Berman points out that the same text “may be both specific and general in reference to persons. or of whole social practices” (1995: 99). and to create intersubjective ties both by generalizing experience and attitude and by ratifying others’ points of view while mutually adhering to societal discourses. Because texts typically do not display multiple representations of the same actions and events at different levels of specificity. reference” is philosophically controversial. With respect to generalizations in English conversations. place.. including only the names of episodes. places.g. In contrast to formal or academic generalizations. and generalizations Van Leeuwen (1995) notes that determining the linguistic properties that con- tribute to generality of expression can be challenging. the expe- riences.

g. all the teachers were telling me) or verbs in the habitual present (e. and referentiality) (Scheibman 2002).g.. analyses in this paper are restricted to conversational utterances with grammatical subjects that designate classes (Section 3. Utterances were tagged for a variety of structural and functional properties for a previous study (e. 2000. time. Adverbials of time (e. expressions with general subjects tend also to contain generalizing predicates (Section 3.. Therefore. As it turns out. for example.g. (class: religious group) everybody in the world has those problems (class: hyperbolic subject) .2 Sources of data The database of utterances from which the generalizations were drawn consists of portions of 10 conversations.. so you go and you practice on the weekends). events. totaling approximately 90 minutes. General construal and expression of events can be expressed aspectu- ally. grow- ing up around here you would know better) also index classes of expectation and experience. (2) the English are just the same though. the prices are so much lower there than here.. subject animacy.4). plural noun phrases regularly denote classes (e.6 (class: nationality) your people . Generalizations in English conversations 119 English speakers generalize about people.. they play salsa?). and place (e.g.g..g. place. in conversation. 2003). For example. tense. as illustrated in (2). and attitudes. and five are unpublished transcripts.3). tortured him to death. Five of the conversations are from the Santa Barbara Corpus of Spoken American English (Du Bois et al. 3. with progressive constructions (e. they come in with that attitude). a variety of grammatical and lexical constructions can ex- press generality of meaning. Because of this struc- tural and semantic heterogeneity of generalizing types in English. manner (e. he threatened little kids). reactions.5 There are 153 utterances with general subjects in the corpus representing the language of 19 adult speakers of American English. verb type. A coding field was added specifying the general class indexed by the grammatical subject. Du Bois et al..g..

Chafe (1994) suggests that English subjects are often starting points for additional information (MacWhinney 1977).g. institutional entities (e. and generic they is often used in these conversations to characterize people or institutions construed as outsiders (e.  Generalizations by subject type (n = 153) Subject type Examples they 74 (48%) so they don’t know what the hell they’re doing and in high school you know they teach about it you 31 (20%) you can’t kill peas growing up around here you would know better Lexical noun phrase 30 (20%) the ballroom people don’t do it that way (plural) those guys’ll be all over you Lexical noun phrase 16 (11%) the ceviche’s full of vegetables (singular) the press will pick it up and smear you for it one 2 (1%) one is more tolerant of feminist views one’s certainly more sensitive to them . everybody in the world).. vacuum cleaner bags.g.g.120 Joanne Scheibman 3. the women).. classes indexed by subject noun phrases are evaluated and elaborated in accompanying predicates.g.8 The classes referred to by lexi- cal noun-phrase subjects in the database (and anaphoric they subjects as well) are diverse and vary in their conventionality.g. the generalizations examined in this study are utterances with grammatical subjects that designate classes. you don’t admit you’re a feminist). which makes utterances with general subjects a useful choice for these analyses. and hyperbolic subjects (e.. Table 1 summarizes the generalizations in the database by subject type. They include. and Langacker (1993) points out that topics (and subjects) can serve as reference points for further talk. Table 1.. While it is the case that speakers evoke classes in object or oblique positions (e. In this study. the press). resulting in full utterances that convey generality of meaning.. Generic you utterances tend to universalize experi- ence in conversations (e.3 General subjects As stated above. The three major subjects that refer to general classes are generic they and you forms and lexical noun phrases (plural and singular). concrete items and groups of people expressed with plural nouns (e. English subjects have interactive and cognitive importance. He knows Cats..g. you can’t really tell when they blush).7 I’m not dancing with guys). for example.

Sharon uses a general subject and a past-progressive predicate to present a series of similar events. All the teachers were telling me. as opposed to individual accounts. only 12 percent (18/153) of the predicates occurring with general subjects in the database are past tense. (3) (Raging Bureaucracy SBC0004 670. the speaker is not emphasizing individual tellings of specific teachers at different times. 82 percent of the present-tense utterances (89/109) are simple present. Many of these utterances are normative statements.41)9 1 → SHARON: All the [teachers were] telling me. This means that there are relatively few predicates in the group that might designate referentially specific events (e. Table 2 displays the generalizations by tense and modality.65-674.).10 2 SHANE: [Could be worse]. Notably. Rather. 5 your kids are gonna take advantage of you . This analysis is supported by repetition of the teachers’ advice in the quoted material in lines 3 and 4. In the utterance. in (3). expressing stative and habitual meanings.g. I talked to a guy that’s thirty-four in my class. In contrast to the high frequency of present tense in these utterances. and the most frequent modals in these utterances are can and can’t (69% or 9 out of 13 tokens). Sharon’s use of all and a progressive construction highlights her general reaction to her fellow teachers’ advising her (that all of them were doing it).4 Predicates occurring with general subjects The formal properties of predicates occurring in utterances with general subjects also convey general meanings. Seventeen percent of the predicates occurring with general subjects (26/153) contain a modal auxiliary. The examples in Table 2 illustrate how these predicates express processes and states as non-specific. and not to the details of each encounter.. Generalizations in English conversations 121 3. line 1. the past-tense predicates occurring with general subjects tend to reflect construal of events as summaries. Half of these occur with generic you subjects. 71 percent of these predicates (109/153) are present tense. As is the case. while only 12 percent are past tense. For example. 3 SHARON: You’re too nice. 4 you’re too nice. Additionally. however.

Generalizations as shared evaluations 4. 4. and that even lexical items like ex- pensive and tall are not solely informative. preference). such as higher frequency of present-tense stative and imperfective meanings.1 Subjectivity. evalu- ation. and Section 5 illustrates utterance types that are used to strengthen speakers’ claims. to varying degrees. evaluation. Subjective expressions are those which. are grammatically. or pragmatically anchored to the speaker.g. lexically.putting out people’s eyes 3. Verhagen (1995: 116) also suggests that subjective expres- sion is the most usual mode of language use. and lower frequency of past-tense predicates. The distribution of these utterances suggests that the predicates co-occurring with general subjects exhibit properties that also express generality of meaning. The rest of this paper presents subjective and intersubjective functions of gen- eralizations in the conversational data.5 Summary Because of the empirical challenges relevant to determining correspondences be- tween linguistic structure and generality of meaning in English conversational ut- terances. research based on conversational data has noted a variety of associations between linguis- . and generalization Linguistic subjectivity is the general phenomenon that refers to the expression of self and the speaker’s point of view in discourse (e. mental states.. a decision was made to focus on tokens containing general subjects. Relevant to the current study. affect.  Generalizations by tense and modality (n = 153) Examples Present 109 (71%) people have the information they need (stative) everything breaks on my kitchen floor (habitual) Modal 26 (17%) you can go up there you can’t constantly be torn Past 18 (12%) all the teachers were telling me they were pull. Section 4 discusses generalizations that function as evaluations in the conversations.122 Joanne Scheibman Table 2. Benveniste ob- serves that “[l]anguage is marked so deeply by the expression of subjectivity that one might ask if it could still function and be called language if it were constructed otherwise” (1971: 225).

Kärkkäinen 2003. Generalizations in English conversations 123 tic structure and expression of point of view (e. one of the participants. 3 but.g. The generalizations illustrated in this section are evaluations that fulfill both subjective and intersubjective functions. It is offered by participant O in response to Z’s narrative of her gray hair dilemma. is the evaluative generalization in focus. 15 O: <@ right. 8 Z: (CLEARS THROAT) 9 → O: the whole thing is pretty ghastly. Z. 13 I guess not. in conveying opinions. 7 S: you do it the way you do.5) no it’s. 11 you know there isn’t much alternative.. 12 Z: no. N. and because they bother her. mocks Z about her worry. evaluations invariably reproduce cultural norms and values: “Every act of evaluation expresses a communal value-system. One of the participants. Preceding this episode. she pulls them out whenever she sees them. The utterance in line 9.4) let’s face it that. as characterized by Thompson and Hunston.(1.. or the expression of attitude. Thompson and Mulac 1991). 4. 10 …(1.2 Construction of an evaluative class Example (4) is part of a longer conversation among four women friends. For example.g.. A common type of linguistic subjectivity is evaluation. adjectives such as ugly) and grammatically (e. a 32-year-old woman. and every act of evaluation goes toward building up that value-system” (Thompson and Hunston 2000: 6). Thompson and Hunston (2000) observe that evaluation is conveyed both lexically (e. 2002. . characterizing her as wanting to remain a teenager (line 1). the use of the pro- gressive in some contexts). 14 N: die young..g. the whole thing is pretty ghastly. was complaining to the group that she keeps finding gray hairs. Tao 2001. they also play a role in participant interaction. 5 S: well? 6 O: the whole thing. Scheibman 2000. (4) 1 N: she wants to be a teenager.. 4 . 2 Z: no I don’t want to be a teenager. Not only are evaluations expressive components of speakers’ activities.

or intersubjectively. brazilians). ceviche). the whole thing is a general construct based on Z’s gray-hair story and others like it in the culture.. the whole thing. The utterance comes on the heels of Z’s defense of herself from N’s teenager accusation (lines 2–4). it also functions inclusively. . This attitude toward aging is common ground among the participants (Clark 1996). Even though the phrase itself has little explicit content. pretty ghastly. is propositionally empty (the head noun is thing) and hyperbolic (the whole thing). parents of elementary school students in the speaker’s class. As noted in Section 3. It linguistically indexes a popular attitude toward aging shared by the members of this group (that it is undesirable but inev- itable). highly insti- tutionalized classes are referential groups such as nationalities (e. in the interaction in two ways. O’s generalization expresses a shared cultural norm. giving the utterance strong expressive force. For example. might be glossed as ‘the experience of aging. general classes referred to with lexical noun phrases in the conversational data vary in their conventionality.. Not only does the generalization express O’s own stance. substitute teachers). the whole thing.’ and in- cluded in the meaning are shared attitudes toward getting older. or common objects and phenomena (e. the whole thing is pretty ghastly. Less conventional are generalizations based on groups that are indexed locally in discourse whose interpretation depends. And at a more global level. O’s statement sup- ports Z’s gray-hair narrative in the interaction by meeting it with her own nega- tive evaluation of aging.g. not only on lexical specification.g. as demonstrated by subsequent ratification and con- tinuation of the topic in lines 12–20. professions (e. allowing O to show solidarity with Z at this point in the conversation (a positive politeness strategy). 20 that sounds good.g. An example of a highly indexical construction is illustrated by the generaliza- tion in (4).. especially when combined with the evaluative complement. and one that unifies their individual stances. but also on participants’ sociocultural expectations and knowl- edge of discourse proceedings (e. It is also the case that O’s generalization has interpersonal function in the conversation. The referent of the noun-phrase subject. In summary. 18 yeah. guys in their twenties in the Bay Area).3..124 Joanne Scheibman 16 [exactly @>] 17 Z: [yeah really].g. The lexical noun phrase. 19 N: boy. understanding of this indexical subject by the interactants is apparent.

Kärkkäinen. and another refers to the shar- ing of beliefs and attitudes among participants as members of communities. Traugott and Dasher (2002: 22) suggest that “intersubjectivity is most usefully thought of in parallel with subjectivity: as the explicit. coded expression of SP/W’s [speaker/writer’s] attention to the image of ‘self ’ of AD/R [addressee/reader] in a social or an epistemic sense. or interactively. focuses on the lo- cal activities of participants in interactive contexts. in situations of language use expressions of speakers’ evaluations and perceptions are fundamentally sensitive to the interactional exigencies of the context. Keisanen.g. hedges. then. in honorification”. social (e. politeness markers). Traugott and Dasher also characterize as intersubjective those utterances for which the speaker pays particular attention to the addressee as a speech-act participant (e. politeness) and the beliefs and values com- ponents. cultural. Schiffrin (1990) suggests that intersubjectivity is interactively constructed among participants (e. and in the speaker. intersubjectivity refers to shar- ing of sociocultural expectations. In a study of evi- dential expression based on spoken and written Dutch and German. and she notes that it plays a role in the sharing of background knowledge in conversational interac- tion – activities that make reference to norms and expectations. In conversation. despite the ana- lytical convenience of labeling linguistic activities as either subjective or intersub- jective. for example. Generalizations in English conversations 125 4. Generalizations can be interpersonally inclusive .3 Local and global aspects of intersubjectivity in discourse Analysis of the preceding example suggests that generalizations have social uses in friendly English conversations: they mark solidarity between participants. These two functions are aspects of intersubjectivity. In fact. in negotiating turns and topics). both of these intersubjectivities are accomplished simultaneously. Nuyts (2001) uses intersubjectivity in a similar way – to refer to people’s sharing of beliefs and conclusions with the speaker. the intention of influencing the other in some way” (Benveniste 1971: 209). it is also intersubjective. intersubjective. One characterization of intersubjectivity in discourse. For Benveniste (1971) not only is language inherently subjec- tive. Indeed. however. In a study of the cultural standing of opinions in discourse. Benveniste suggests that the process of com- munication itself is “only a mere pragmatic consequence” of the dialogic relation- ship between ‘I’ and ‘you’ (1971: 225). and Haddington). In these treatments. as demonstrated in subsequent analyses in this paper as well as in other chapters in this volume (especially those by Du Bois.g. The former might be characterized as interpersonally. Strauss (2004) calls these interpersonal elements of the context. He writes that “every utterance assum[es] a speaker and a hearer. and they can be sites for the sharing of cultural beliefs.. the latter makes reference to more global sharing of expectations.g...

There are two generalizations in this conversational extract that contribute to interactants’ construction of shared stance in discussing this topic: the jointly produced generalization in lines 2–3 (I would think that a person would be more/you mean more cosmopolitan). On the other hand. 7 C: X 8 A: other things than= . Specifically the group is trying to figure out why Vanda is conservative and opinionated.. because generalizations often make reference to societal norms. three participants are discussing a coworker of speaker B. 4 B: yeah. These subject types permit global assessments of people and their behaviors. the interactants suggest that she should have more sophisticated views about feminism than she does given that she has been exposed to a range of cultural experiences due to having lived in different countries (line 1). demonstrate the participants’ local construction of a shared at- titude in the conversation. Even so. The collaboration of interactants in articulating this generalization. (5) 1 B: . The joint production of this generalization is accomplished in two ways: the turns of the two participants are overlapping. Both of the generalizations in this episode (lines 2–3 and lines 6 and 8) contain generic subjects (a person. along with speaker B’s ratification of C’s completion of her utterance (line 4). At the same time. and speaker C completes speaker B’s comple- ment (Ono and Thompson 1995). The . 6 → some people are more open to. 4. which is how she was characterized previously in the discourse. than others.126 Joanne Scheibman in that they demonstrate participants’ solidarity or empathy with one another’s stance. some people) co-occurring with relational predicates.. they are involved in the reproduction of belief systems. The evaluative generalization in lines 2–3 is collaboratively produced. she’d been exposed to enough that I would’ve -- 2 → I would think [that a person would be more]. Speaker A’s generalization in lines 6 and 8 is a contribution to the group’s explanation for Vanda’s attitudes.4 Joint construction of stance using generalizations In the conversation in (5). The participants agree that Vanda’s having lived in Vienna until she was 16 could have contributed to her conservative outlook (Vi- enna is awfully conservative). and the generalization offered by speaker A in lines 6 and 8 (some people are more open to other things than others). 3 C: [you mean more cosmopolitan]. 5 A: but that’s probably her personality. who is a woman named Vanda.

The use of non-personal they is so conventional in English conversations (e. what began as speaker B’s subjective stance becomes shared and elaborated by the group. the referent of they was not previously established in the discourse. line 2) is an utterance con- taining a generic they subject occurring with a habitual predicate (they come in with that attitude).5 Evaluating others with they The evaluations discussed here are utterances with non-personal. they is commonly used to evaluate groups that participants do not belong to. social groups both in public discourse and in conversational interaction (e.g. The shared belief informing the gen- eralization in lines 2–3 is the expectation that when one lives in different coun- tries one acquires a certain sophistication. In this humorous conversa- tion. 4. they refers to an unnamed but culturally shared conception of an unspeci- fied institutional entity who is being ridiculed by the participants for engaging in genetic engineering of chickens for profit. Because third-person pronouns make reference to non-speech-act par- ticipants (Quirk et al. or generic. Hongladarom 2002 for Thai. to refer to institutional entities) that it can occur without its having been referentially introduced (Biber et al.. 3 You mean so that they can go right from that to Chicken McNuggets? The evaluative generalization in the next example (7. The antecedent for the pronoun occurred a minute before this point in the conversation when one of the participants. they subjects. they is often used by speakers of English and other languages to exclude. In English conversations. in the sense that notions of what it means to be more cosmopolitan or more open are subjectively and intersubjectively construed. tells how a sub . For example in (6). The commonsense view underlying the generalization in lines 6 and 8 (including line 5) is that people’s attributes can be divided into those that have to do with cultural experiences and those given by personality. (6) (Conceptual Pesticides SBC0003 295.58) 1 ROY: they’re trying to breed like a forty foot long tube chicken? 2 MARILYN: No= @@.. Skarżyńska 2002 for Polish). 1985: 3354). or other. The general class designated by non-personal they in this utterance is substitute teachers. 1999: 331). and often these assessments express disapproval or derision. In this way. These generalizations are offered by participants as they collab- oratively construct an explanation for Vanda’s character – a person unknown to two of the participants in the conversation.15-305. Generalizations in English conversations 127 predicate-adjective constructions found in both generalizations are evaluative.g. Sharon.

Even so. Subsequently. creating solidarity among themselves. important aspects of the meaning of this class depend on shared stance. In this example. they here indexes a class of people with social and institutional relevance who are viewed derisively by the conversational participants. attitude. 5 SHARON: @@ 6 CAROLYN: Now. which prevented the child from receiving her lunches. The simple form of come in the generalization in line 2 expresses habitual meaning that generalizes over supposed specific instances..128 Joanne Scheibman at the public school where she works lost a student’s free-lunch application. Carolyn acts out her evaluation of a too-eager substitute teacher walking into a classroom.13) 1 CAROLYN: You know. 3 and they go. For example. but incompetent nonetheless. as it is referred to in this conversation. 4 ((THUMP)) <Q I’ve always wanted to teach math. the class is not simply a referential set of teachers who cover classes on a tempo- rary basis when “real” teachers are ill or on leave. they don’t give a shit about anything). In the excerpt in (7). evaluation and elaboration of meaning is accomplished by all of the participants who share the cultural stance that views substitute teach- ers as other. in fact). according to her. Carolyn.6 Summary The examples in this section illustrate situations in which participants’ evaluative stances expressed with generalizations become jointly held in the conversations. Rather. the participants be- gin to condemn substitute teachers as a group for their general incompetence (they don’t care. initiates a second diatribe about beginning substitute teachers. 2 → they come in with that . 4. (7) (Raging Bureaucracy SBC0004 379. This othering of substitute teachers continues after this point in the conversation with other participants laughing and joining in the perfor- mance. are more enthusiastic than the experienced group (too enthusiastic.05-385. who. One of the participants. 7 what are we on Q>? substitute teachers is a relatively conventionalized class in mainstream Amer- ican English-speaking cultures. It is notable that the subjective and intersubjective uses of these utterances are not distinct from one another. while speaker O’s construction and eval- uation of a highly indexical class in (4) (the whole thing is pretty ghastly) reflects . thereby intensifying the speaker’s expression by portraying the behavior she’s sati- rizing as repetitive. That is.

It is offered to show solidarity with speaker Z’s evalua- tion of aging. In this way. The two generalization types discussed in this section augment speakers’ stances in these two aforementioned ways. Propositionally speaking.. Similarly. Tannen 1984). And finally. generalizations exhibit broadening properties. In the excerpt in (5). If a speaker’s claim can be considered true for a larger class of entities or situations. speakers enhance their positions in conversations by appealing directly to common experiences and beliefs of other participants. typically serves to make that principle theoretically more powerful. are used to mock institutional groups who are viewed as outsiders. Generalizations in English conversations 129 her individual stance. This broadening can apply to propositional material or to participant interaction. then there is an assumption that the claim is stronger than it would be. expanding the reach of an assertion in conversation using a generalization can potentially increase the expressive power or author- ity of that assertion.. broadening the domain of applicability of a sci- entific principle. a sign of recognition as well as amusement (Coates 1996). Augmenting stance with generalizations 5. As illustrated in (9) below. generalizations with non-personal they subjects. In both of these extracts.g. the indexicality of the subject noun phrase co-occurring with an evaluative predicate expresses a common cultural view that invites. for example. if the assertion were based only on one or perhaps two specific instances.g. for example. and at a more global level of intersubjective activity. Generalizations can also be interactively broadening in the sense that par- ticipants tend to share beliefs expressed with these utterances (e. the relationship between in- dividual and interactive stance is illustrated by the linguistic co-construction of a general evaluation based on shared beliefs. and both of these aspects can be used to enhance speakers’ stances in conversa- tions. which suggests that interactants share the speaker’s evaluation. In this way. speakers offer satirical gener- alizations that evoke laughter from the other participants. The analysis of the generalization in . the use of this generalization simultaneously demonstrates intersubjective functions. the whole thing is pretty ghastly). high involvement. motivated by positive politeness (e. illustrated in (6) and (7). on the other hand. a speaker can explicitly invite other participants to agree with her generalization by using a generic you pronoun. shared interpretation by the participants.1 Propositional and interactive broadening As discussed in the introduction. or perhaps marks. generalizations can strengthen a participant’s stance by appearing to expand the reference class on which a particular claim is based. 5.

Miles is ani- mating a message that is authored by these guys about these women. Hill and Irvine 1993). The former is a case in which the broadening properties of the generalization are applied to the propositional content of the speaker’s argument.130 Joanne Scheibman (8) suggests that this utterance is used by the speaker to reinforce the substance of his argument by evoking a general group as evidence for his claim. In the language of Goffman’s (1981) participation framework. men in their twenties. yeah. At the same time. Miles’s discursive construction of the demographic. The full generalization is in lines 4–5 (I mean these guys say these women don’t care if they use rubbers or not). A component of Miles’s explanation for this alarming state of affairs is that women who are having sex with guys in their twenties are not insisting that these men use condoms. whereas the latter generalizations are interactively expansive in the sense that they invite corroboration from conversational participants. 5 say these women don’t care if they use rubbers or not. such as right. 6 this is what guy=s have been telling me. I mean these guys say these women don’t care if they use rubbers or not (Fox 2001. po- tentially strengthens his stance in this conversation by broadening the reference class on which his claim is based. and .. 4 → I mean these guy=s. and the antecedent for these guys appears in lines 1–2 (guys in their twenties). in their twenties. (8) (Lambada SBC0002 696.2 Use of a general class to strengthen a claim Example (8) is part of a longer conversation in which the speaker. and neither of these general groups (men in their twenties and women they have sex with) are individual agents whose experiences might be contested. The other participants in the conversation do not question Miles’s general- ization. which are used by the speaker to strengthen her stance by appealing to the experiences and beliefs of the other participants.. he also mitigates having direct knowledge of the women who are not insisting on using condoms during sex with the men by using a reported speech construction in lines 5–6. 2 . In contrast. they provide supportive uptake. stupid. 5. the excerpt in (9) illustrates generalizations with generic you subjects. 3 why. In fact.49-70594) 1 and just talking with guy=s. is incred- ulous that HIV/AIDS cases in the San Francisco Bay Area had been increasing dramatically despite the availability of information on preventing transmission of the disease. Miles.

is implicit reference to an addressee. specified. . 3 [[I tol=d you=]] 4 JOANNE: [[Isn’t that-- 5 he’s not]] hopeless. The generalizations in focus are those in Joanne’s final turn (lines 14–20).590-850. 7–8.g. On the one hand. (9) (Deadly Diseases SBC0015 835. somebody like that. I mean the guy is great. generic you pronouns formally make reference to speech-act partici- pants. the general argument proposed by Miles – that young women in the Bay Area are not requiring their male sex partners to use condoms – is accepted. I told you. Another participant. But you have to at one point let go. 13) overlapping with Joanne’s utterances (lines 1. just torn to pieces by. Jamie. Generalizations in English conversations 131 participate with the speaker in elaborating his claim. Lenore and Ken contribute turns (lines 2–3. Miles provides evidence for his generalization.. it seems to retain second-person refer- ence in that “the speaker is appealing to the hearer’s experience of life in general. in which she described her brother’s down-and-out life before he kicked a disruptive drug habit. In the opening line in this excerpt. The episode in (9) occurs subsequent to a narrative by the primary speaker. or else of some specific situation” (Quirk et al. and jointly maintained by the participants.3 Appealing to other participants Unlike lexical noun-phrase and they subjects. see). In this interaction. Joanne is referring to her brother as he is now since his successful recovery. You can’t constantly be torn. Built into the speaker’s generalization using this form. then. then. Joanne. adds her support by mentioning that her sisters don’t use condoms either. 4–6.11 Subsequent to the excerpt reproduced in (8). He tells the group that a woman whom he had planned to ask out on a date was telling him that her sister just won’t use rubbers. 5. 1985: 354). which index third-person classes and groups. you know. this pronoun is used by speakers as an informal version of the generic pronoun one. On the other.745) 1 JOANNE: I mean [the guy is] grea=t. After this utterance. 2 LENORE: [I told you]. 12) in which they remind her that they had tried to reas- sure Joanne during the time of her brother’s difficulties that he would be fine (e. The ex- ample in this section illustrates a frequent context of generic you generalizations in the database – one in which speakers shift from using first-person-singular expressions to you forms to augment their stance by appealing to the experiences and beliefs of the other participants.

line 21). Joanne’s shift from I to generic you in this conversation may in part be motivated by Lenore and Ken’s repeated reminders that they had told Joanne all along that her brother would rebound.132 Joanne Scheibman 6 [Well I had given] him up [for dead]. the speaker’s generalizations. then. expressed in lines 6 and 12 (I’d given him up for dead). somebody like that. make tacit reference to discourses of drug and alcohol abuse that advise against codependency. which contribute to their having normative function. Joanne’s generic you utterances may justify her previous distancing from her brother by generalizing her reac- tions for Lenore and Ken. Specifically. 7 LENORE: [I told you]. 9 JOANNE: [Oh=. from uncomfortable topics).g. 14 JOANNE: But you have to -- 15 you have to at one point let go. This suggests that generic you utterances fulfill the ritualized intersubjec- . these general statements expressed with you generalizations in the conversational data often prompt speaker changes (illustrated in example (9). to the beliefs of other participants. In generalizing her individual stance. 16 you can’t constantly be torn. 8 KEN: [I] always tol=d her. 11 and she’s]-- 12 JOANNE: I’d given him up] [for dea-]-- 13 LENORE: [see]. These reminders mark a contrast between Joanne’s friends’ expressions of confi- dence in her brother’s situation and her own lack of confidence. Joanne’s shift to generic you. 19 you know. Notice that in line 14 Joanne switches from using I to you. Abney (1996) finds that these shifts can be used as distanc- ing techniques (e. Another use of shifts to generic you is to generalize the speaker’s experience in order to build empathy with other participants. too. and by extension. But you have to at one point let go. 21 LENORE: (TSK) So your mother’s happy now.. may be viewed as an iconic distancing from a situation that might hold conflict for her. 10 KEN: [it’s gonna. 17 just to=rn to pie=ces. 18 by. Based on spoken and written narratives. just torn to pieces by. you can’t constantly be torn. This illustrates the broadening quality of the generalizations. she implicitly ap- peals to societal norms. you know. And finally. 20 somebody like that. Notice. that these three you utterances all contain the modal elements can’t and have to.

and sociocultural stances. In (8).g. Similar to the examples discussed in Section 4.. the subjective uses of generalizations are tied to their intersubjective uses in several ways. Notably. the whole thing). Ad- ditionally. For example. typi- cally evaluation. Summaries and conclusions The set of conversational generalizations examined in this study are utterances that designate classes of experiences or entities for purposes of predication. Analyses offered in Sections 4 and 5 suggest that these expres- sions are not referential descriptions of the world. the generalizations with they subjects illustrated in Section 4 convey speakers’ evaluative stances as they function to create ingroup solidarity by othering outsiders. and are used by participants in the conversations to express individual. Generalizations in English conversations 133 tive function of codas in conversational narratives – expressions conveying “gen- eral observations that are timeless in character” (Labov and Fanshel 1977: 109). the interactive success of Miles’s having indexed a general class to support his argument can be gauged by the subsequent uptake and specification of the generalization by participants. particularly when the class is not a conventionalized one (e.g. these utterances dem- onstrate both subjective and intersubjective uses. with her stance at that point in time. Rather. or agree. Joanne’s you generalizations are normative statements that make reference to societal discourses. on the other hand. or inclusive. because generalizations can have a broadening. demonstrations of solidarity) and more globally in the maintenance of cultural norms through tacit sharing of societal discourses. In (9). interactive. the speaker’s stance in this episode is jointly maintained. they participate in intersubjective activities at an interper- sonal level (e. the speaker.4 Summary The examples in this section illustrate two ways in which the broadening charac- teristics of generalizations can potentially strengthen a speaker’s stance in conver- sations. Joanne. as was suggested for the evaluative generalizations in Section 4. The generalizations in these data frequently function as evaluations and can also be used as rhetorical devices to strengthen speakers’ stances. 5. perhaps to invite her friends to understand. In (9). uses you generaliza- . 6. function in these conversations. Additionally. the speaker generalizes motivations for her past reactions to her brother using generic you constructions. The generalizations are subjective in several ways. politeness. Selecting and indexing a general class as a starting point for evaluation or assertion is invariably an ex- pressive act..

on the other hand. such as software and database de- sign. The strong relational aspects of stancetaking suggest that the reference point notion might also be a useful way to characterize participant stances in discourse. illustrate ways in which links between expression of individual and social attitudes are reinforced in conversa- tional interactions. and scientific authori- ties. Sally McConnell-Ginet observes that “[t]he reproduc- tion of meaning refers to our dependence. they contribute to the construction and reproduction of cultural belief systems. And in friendly conversations among intimates a common consequence is the construction of jointly held stances. and attitudes toward aging. the generalizations discussed in this paper allude to common societal discourses such as: the sexual responsibilities of young women in heterosexual relationships. individuals’ expressions of stance using generalizations have interactive conse- quences. . Indeed. 4. Generalizations in these conversations. appropriate ways to engage with people with drug dependencies. 2. In definitions of generalization used in formal systems. Langacker describes the reference point phenomenon as the cognitive ability “to invoke the conception of one entity [reference point] for purposes of establishing mental contact with an- other [target]” (1993: 5). then.134 Joanne Scheibman tions to broaden her stance in an appeal to the beliefs of her friends. in producing meanings. These analyses support the idea that in English conversations. “Abstraction aims at simplifying the description of an entity.” both socially and linguistically. reduction in specificity of an entity or description is called abstraction. Because my primary objective is to examine gen- eralizations in English conversations (for which differences between abstraction and generaliza- tion are not specifically relevant). because these utterances are often statements of norms. Tomasello (2000) presents compelling evidence from studies us- ing nonce words suggesting that young children do not generalize verbs as a syntactic category as they do nouns. are expressions that link indi- vidual subjectivities to the group. In the guys in their twenties example in (8) and in the extract about Vanda in (5). Moreover. And conversations are important sites in which participants engage with “one another’s experience. 3. while generalization looks for common properties among these abstractions” (Kafura 1998: 27). Language-acquisition researchers have also found that generalizations are sensitive to local patterns of use. Participants’ uses of these generalizations. on previous meanings or interpretations. I will not pursue this technical distinction here. work. then. I am grateful to Robert Englebretson for many valuable comments on an earlier version of this paper. For example. participants collaborate with speakers by specifying and elaborating their generalizations. to our dependence in particular on one another’s experience with the linguistic forms being used” (1998: 199). Notes 1.

1996. J. L. 1994.. and Finegan. CD. CD. S. Conrad. generalizations and utterances that figure prominently in discussion in the text are bolded. 8.A.. D. Excerpts from the Santa Barbara Corpus of Spoken American English are cited. E. Translated by Mary Elizabeth Meek. G. Schuetze-Coburn. S. Lampert (eds. 9. Oxford: Blackwell. 1999. Hillsdale. J. Coral Gables. J. Meyer. Using Language.W. Clark. W.A. London: Longman. Transcripts of conversational excerpts used in this paper have been edited for readability.D. 2001. Santa Barbara Corpus of Spoken American English. Cats in this conversation refers to a brand of tractor made by Caterpillar. In the conversational extracts.H. Santa Barbara Corpus of Spoken American English.. which were previously addressed. Johansson.” Journal of Linguistic Anthropology 11: 167–192. S. FL: University of Miami Press. Discourse.” In Talking Data: Transcription and Coding in Discourse Research.” Journal of Pragmatics 37: 105–124.W. 59– 83. 10. Benveniste. Aspect. 2003. Two general subject types. and Mo- dality in the Languages Of The World.. Cambridge: Cambridge University Press. 11. C. Ed- wards and M. D. B. 6. Meyer.” Le Français dans le Monde 34: 47–52. 1996. A key to transcription symbols is provided in the Appendix. Part 1. H. 2000. “Outline of discourse transcription. Consciousness. Berlin: Mouton de Gruyter. Bybee. Du Bois. 1989. Leech. W. were excluded from the study: (1) we subjects that designate classes (Scheibman 2004). W. Coates. J.. S. Philadelphia: Linguistic Data Consortium. 1995. R. and Thompson.” Text 9: 93–124. Fisiak (ed. References Abney. Chafe. Bybee. R. 1985. 7.).A. Bouacha. S. Philadelphia: Linguistic Data Consortium. 1996. J. “Pronoun shift in oral folklore. personal experience and literary narratives. Berman.. Du Bois. 45–89.W. and Thompson. W.. E. Biber. responsibility. C.. These contributions are not reproduced in (8). The Evolution of Grammar: Tense. 2005. Biber. Du Bois. while those from unpublished transcripts do not have an accompanying reference. J. W. and Pagliuca. 1993.. Women Talk: Conversation between Women Friends. Chafe. J. 1994. Generalizations in English conversations 135 5.. E. and Finegan. Problems in General Linguistics. University of Pennsylvania.A. Fox. “Styles of stance in English: Lexical and grammatical marking of evidentiality and affect. Part 2.A. Chicago: University of Chicago Press. “Evidentiality: Authority. Chafe.). Chicago: University of Chicago Press. “Cross-linguistic comparison and the development of gram- matical meaning. and entitlement in English conversa- tion.A. 1971. University of Pennsylvania. . and Pagliuca. “Linguistique théorique et recherche en didactique. “Introduction: Developing discourse stance in different text types and lan- guages. D. and Paolino. Longman Grammar of Spoken and Written English. Perkins. Cumming. and (2) utterances with impersonal it and that subjects (Scheibman 2002 for it and that subjects of relational clauses). J. or ‘What’s up with you?’” SECOL Review 20: 203–226. S. NJ: Lawrence Erlbaum. Historical Word Formation. M. and Time.” In Historical Semantics.

London: Longman. Helm (ed. 1987. Scheibman. Davis (ed. Epistemic Stance in English Conversation: A Description of its Interactional Functions. Greenbaum. Hopper. A. McConnell-Ginet. Reilly. 1998.” Journal of Pragmatics 32: 105–124. 321–339. Responsibility and Evidence in Oral Discourse. D. MacWhinney. 2nd ed. J. Fire. London: Routledge.” In Us and Others: Social Identities Across Languages. A Comprehensive Grammar of the English Language. “Subjectivity as an evidential dimension in epistemic modal expressions.A. M. G. Haiman and S. 2004. Scheibman. Bybee. E.W. P. 1997. Thompson (eds. 1995. Foundations of Cognitive Grammar. J.W. B. Langacker. 231–247. Kärkkäinen.H. J. Lyons. I: Theoretical Prerequisites. A. 1993. Nuyts. and Dangerous Things: What Categories Reveal about the Mind. Duszak (ed. In Essays on Language Function and Language Type. “Learning argument structure gen- eralizations.” Language 53: 152–168. N.). J. G. S. Philadelphia: University of Pennsyl- vania Press. Casenhiser. T. Object-Oriented Software Design and Construction with C++. Amsterdam: John Benjamins. “Starting points. Point of View and Grammar: Structural Patterns of Subjectivity in American English Conversation. Achard and S. R. “Inclusive and exclusive patterning of the English first person plural: Evi- dence from conversation. Quirk. 12–44.” Jour- nal of Pragmatics 33: 383–400. J. Cambridge: Cambridge University Press. Labov. Women. J. 1967. 1977. R. and Thompson.. Vol.. 198– 210. Stan- ford.” In The Feminist Critique of Language: A Reader. NJ: Prentice Hall. W. Culture and Mind. Amsterdam: John Benjamins. D. A. and Fanshel.” Cognitive Linguistics 15: 289–316. 2002.” In Forms of Talk. Stanford. Chicago: University of Chicago Press. Lakoff. J. 1977. S.” Journal of Pragmatics 37: 185–208. “Narrative analysis: Oral versions of personal experience. “The sexual (re)production of meaning: A discourse-based ap- proach..E..J. P. Leech. 124–159.). W.. R. “What can conversation tell us about syntax?” In Alterna- tive Linguistics: Descriptive and Theoretical Modes. J. With a Focus On ‘I think’.” In Essays on the Verbal and Visual Arts.). 2005. and Irvine. 1993. “Acquiring perspective in English: The devel- opment of stance. “Discourse about them: Construction of ethnic identities in Thai print. Semantics. J. 213–271.W. 2004. 2003. Amsterdam: John Benjamins. “I dunno . Cambridge: Cambridge University Press. 2000. CA: CSLI Publications. Goldberg. J. Cameron (ed.A. S. Labov. D.F. Scheibman. 1987. Kafura. Kemmer (eds. Amsterdam: John Benjamins. CA: Stanford University Press. 1977. 1985. New York: Academic Press. and Sethuraman. 2002. Zamoraa. and Svartvik. Discourses and Cultures. Vol.. “When ‘grammar’ and discourse clash: The problem of source conflicts”. and McGivern. Hongladarom. D. Amsterdam: John Benjamins. . A usage-based account of the phonological reduction of don’t in American English conversation..). 1.). (eds. 377–396. 1981. Therapeutic Discourse: Psychotherapy as Conversation. Ono. and Waletzky. Hill. E. Langacker.).). Seattle: University of Washing- ton Press. J. J. D. “Footing.136 Joanne Scheibman Goffman. R.” Cognitive Linguistics 4: 1–38. 1998. “Reference-point constructions. 2001.” In Language. Upper Saddle River. J.

Ann Arbor: University of Michigan Press. Thompson (eds. Cambridge: Cambridge University Press.” In Corpus Linguistics in North America: Selections from the 1999 Symposium. 2002. 11–55. Tomasello.” In Us and Others: Social Iden- tities across Languages. 103–128. Oxford: Oxford University Press.” Language in Society 33: 161– 194. D. Amster- dam: John Benjamins. M. Cambridge: Cambridge University Press. and cultural description. syntax. 2001. E. Selby (eds. Qiao. E.).H.” Se- miotica 80: 121–151. 116–114.). linguistic categories. and communication. 1991. A. 2000. Simpson and J. M. “Representing social action. Regularity in Semantic Change. Stein and S. “We and they in Polish political discourse. “The principle of intersubjectivity in communication and conversation. D. and Dasher. Volume II: Fo- cus on Types of Grammatical Markers. Thompson. “Shifters. van Leeuwan. and Hunston. Hunston and G.” Discourse and Society 6: 81–106. Silverstein. Albuquerque: University of New Mexico Press. Traugott. 313–329. 1993) Speech overlap [ ] or [[ ]] Truncated intonation unit -- Transitional Continuity Final . 1–27. NJ: Ablex. Continuing . H.” In Meaning in Anthropology. Wright (eds. R.). Conversational Style: Analyzing Talk among Friends. 1998. 1976. 2004.” In Approaches to Grammaticalization. 1995. Amsterdam: John Benjamins. Swales (eds.). and Mulac. K.” In Subjectivity and Sub- jectivization: Linguistic Perspectives. S. Heine (eds. “Subjectification. Strauss. R. “Fuzziness – vagueness – generality – ambiguity. D. Skarżyńska.). Zhang. K. “Cultural standing in expression of opinion. Appeal ? Lengthening = Medium pause … Long pause …(N) . 1990. S.B. 249–264. “A quantitative perspective on the grammaticization of epistemic parentheticals in English. 2000.A. Verhagen. 1984. Thompson. “Evaluation: An introduction. A. Appendix: Transcription symbols (Du Bois et al. Generalizations in English conversations 137 Schiffrin. 2002. Duszak (ed. C. “Discovering the usual with corpora: The case of remember. Tao. Traugott and B. Basso and H. A. T. “Do young children have adult syntactic competence?” Cognition 74: 209–253.). Discourses and Cultures.A. S. Tannen. G.” Journal of Pragmatics 29: 13–31. 1995.” In Evaluation in Text: Authorial Stance and the Construction of Discourse. Norwood.

138 Joanne Scheibman Laughter @ Laugh quality <@ @> Quotation quality <Q Q> Vocal noises () Indecipherable syllable X Researcher’s comment (( )) .

what role language and interaction play in the process. If we are to achieve any general understanding of stance. how we do it. it is necessary to go beyond merely cataloguing their contents or classifying their types. To frame a theory of stance means to provide a general account of the mode of production of any stance and of its interpretation in a context of interaction. it will be necessary to seek the foundational principles which underlie the act of taking a stance and negotiating its meaning. Yet very little is understood at pres- ent about stance: what it is. Because the diversity of observable stances extends in principle without limit. Realizing such a goal will require us to define a research agenda capable of bring- ing together multiple coordinated lines of inquiry drawn from a range of disci- plines concerned with the use of language. Setting the problem in this way brings into play several aspects of language in interaction. These are the questions which will inform this paper. As one step in this direction. to calibrate alignment between stancetakers. interac- tion. As we seek the theoretical resources needed to account for the achievement of stance. and what role the act of taking a stance fulfills in the broader play of social life. and to invoke presupposed systems of sociocultural value. Stance can be approached as a linguistically articulated form of social action whose meaning is to be construed within the broader scope of language. as I seek to develop a language for describing the phenomenon of stance and clarifying the role it plays in the larger contexts of language and interaction. Introduction1 One of the most important things we do with words is take a stance. Stance has the power to assign value to objects of interest. we find ourselves faced with a . and sociocultural value. to position social actors with re- spect to those objects. Du Bois University of California. this paper presents a preliminary sketch of some of the theoretical resources and analytical tools that are likely to be required for such an account. Santa Barbara 1. The stance triangle John W.

140 John W. The significance of the necessary link between the intersubjective. Clancy 1999. Traugott 1989) and the discourse-functional organization of language use (Maynard 1993. intersub- jectivity has long been recognized as an issue for the social sciences. for the theory of stance. In this light. which will be seen to carry significant implications for the interaction at hand. the words of those who have spoken be- fore – whether immediately within the current exchange of stance utterances. at a more general level. Voloshinov [1929]1973). and. dialogic syntax details the process of mapping resonances between juxtaposed utterances in discourse. when the formal method of dialogic syntax is applied to conversational interaction. Vygotsky 1986). Dialogic syntax looks at what hap- pens when speakers build their utterances by selectively reproducing elements of a prior speaker’s utterance. Silverstein 2001. Schutz 1962. where it has influenced especially those approaches which locate language and cognition within a sociocultural and/or sociocognitive perspective (Bruner 1986. it turns up a remarkable number of stance pairs characterized by a recurring functional-interactional configuration. and further engage with. stance represents a promising testing ground to ex- plore the potential of a more explicit dialogic method in the context of conversa- tional interaction. Toma- sello 1999. Scheibman 2002). the analogy implied by their structural parallelism triggers a series of interpretative and inter- actional consequences. in which the stance utterance of a subsequent speaker is constituted as bearing a close analogy to the stance utter- ance of a prior speaker. the social actors who jointly enact stance. While subjectivity has been getting the greater share of attention late- ly. Injecting dialogicality into the analysis of stance leads naturally to a concern with intersubjectivity. intersubjectiv- ity. As it happens. and objective will become evident . Here I draw on my recent formulation of dialogic syntax as a specifically dialogic method (Du Bois 2001). and the mediating frameworks of linguistic structure and sociocultural value they invoke in doing so. Lucy 1993. As an analytical practice. being recognized for its role in the cognitive organization of language structure (Langacker 1991. While the idea of dialogicality has proved theoretically evocative (Bakhtin [1934]1981. Ochs 1988. which I understand as the relation between one actor’s sub- jectivity and another’s. Though its conceptual origins lie in philosophy. subjective. it cannot be said to have translated successfully into a well-defined program for substantive research on observed instances of dialogic interaction (but for some hopeful possibilities see Linell 1998). or more remotely along the horizons of language and prior text as projected by the community of discourse. As stances build on each other dialogically. Dialogicality makes its presence felt to the extent that a stancetaker’s words derive from. Hanks 1990. Du Bois complex web of interconnections linking stance with dialogicality. It should be noted that intersubjectivity presupposes subjectivity. intersubjectivity is no less indis- pensable as a piece of the larger stance puzzle.

as locally relevant values are activated to frame the significance of participant actions. arguing that making sense of intersubjectiv- ity in interaction depends on a dialogic understanding of language use. The act of taking a stance necessarily invokes an evaluation at one level or another. Moreover. Lucy 1993). Sociocultural value is mo- bilized and deployed through stance processes. I go on to explore the role of intersubjectivity and other sociocognitive relations in the elaboration of stance. whether by assertion or inference. the very act of tak- ing a stance becomes fair game to serve as a target for the next speaker’s stance. I begin the paper with a brief presentation of sev- eral different kinds of stance phenomena. I introduce the stance triangle as a way of representing the components of the stance act. insofar as they serve to ground the sociocognitive aspects of stancetaking in dialogic interaction. since lan- guage is reflexive (Haviland 1996. and participants routinely monitor who is responsible (Hill and Irvine 1993) for any given stance. value can be focused and directed at a precise target. To complete the picture. The stance triangle 141 once we recognize them as fundamental sociocognitive relations which organize language use. In this paper. In its guise as a representation of the fundamental structure of the stance act. Stance is realized. in the usual case. Moreover. I then address the problem of contextu- alizing stance. the stance triangle attempts to shed light on the realization. there is groundwork to be laid. in many cases the cur- rent stance act resonates both formally and functionally with a stance taken in prior discourse. This sets the stage for the next level of theorizing stance. articulating their multiplex interrelations. and to the sociocul- tural frames that mediate the consequences of their actions. whose lives are impacted by the stances they and others take. Stance both derives from and has consequences for social actors. This in turn implicates those dimensions of sociocul- tural value which are referenced by the evaluative act. Via specific acts of stancetaking. But before we get to the stance triangle. I work to bring these elements together in a unified framework for stance. and more importantly. The result is a fundamentally dialogic perspective on stance which sees the stance act . Thus the value of any stance utterance tends to be shaped by its framing through the collaborative acts of co-participants in dialogic interaction. by a linguistic act which is at the same time a social act. The model I propose is articulated in terms of a set of triangular rela- tions among the components of stance. and consequences of stance in interaction. We begin to appreciate why stance should come to wield both subtlety and power in the dynamics of social life. interpretation. Intersubjectivity and its companions play a prominent role in the present approach to stance. we need to look at how the dialogic and intersubjec- tive dimensions of stance relate to the stancetakers’ actions. articulating a set of questions designed to identify some of the key aspects of the indexical context of interaction which are needed for interpreting stance.

what we can do is look at a few of the types that have played a leading role in previous discussions of kinds of stance or stance-related catego- ries. it will soon become clear that the actual stance taken cannot be fully interpreted without reference to its larger dialogic and sequential context. after all. For now we can think of these as different kinds or types of stance acts.46-318. Thompson and Hunston 2000). 2000. and additional important work on stance. Du Bois et al. Pomerantz 1984). as analyzed in conversation analysis (Goodwin and Goodwin 1992. A closely related concept is that of assessment. is so special about stance. Work on the related notion of appraisal has been pursued from the perspective of systemic functional grammar (Martin 2000). 2003b. Evaluation has received considerable attention in recent years (Conrad and Biber 2000. Kinds of stance One way to begin thinking about stance is to look at some likely exemplars. 3 (1) (Conceptual Pesticides SBC003: 317.440-194. Labov and Waletzky 1967. drawn from the Santa Barbara Corpus of Spoken American English (Du Bois et al. (3) (Appease the Monster SBC013: 1117. 2002. That’s horrible. Berman 2005. but in the end we will have to consider alternative formulations of the issue of stance diversity. Lemke 1998. Chafe 1994. point of view. I summarize the main contributions of the stance triangle.341) LANCE.12-1118. 2003):2. . @that’s @nasty. Du Bois as shaped by the complex interplay of collaborative acts by dialogic co-partici- pants. Kockelman 2004. By way of conclusion. Goodwin 2006.12) KEVIN. although for expository purposes we begin with ab- breviated examples of stance utterances viewed in isolation. Macken-Horarik and Martin 2003. By the same token. Linde 1997. (TSK) That’s ideal. While the range of proposals in the literature is too broad to survey here.16) PETE. and related notions has been developed by a number of scholars (Berman et al. Perhaps the most salient and widely recognized form of stancetaking is evalu- ation. based on what has been recognized as stance (or a stance-related category) by one ana- lytical tradition or another. Hunston and Thompson 2000. Hunston and Sinclair 2000. 2. and pose the question of what.142 John W. Shoaps 2004). Consider the following three examples of evaluation. (2) (Runway Heading SBC022: 193. Kärk- käinen 2003a.

or amazed. either glad or so glad. Such utterances have often been described as indexing affective stance (Besnier 1993. The next set of examples presents a somewhat different pattern of stance: (4) (Runway Heading SBC022: 612. as the act of situating a social actor with respect to responsibility for stance and for invoking sociocultural value. Thompson and Mulac 1991). Haviland 1991.4 As speakers position themselves affectively. Haviland 1991. for example. whether with respect to an affective (glad) or an epistemic (know) state. for example. I’m just ama:zed. Shoaps 2002).970) DAN. (6) (Lambada SBC002: 665. but also along an epistemic scale (Clift 2006. (5) (Hey Cutie-Pie SBC028: 52. I kno:w.99) KENDRA. the stance predicates horrible.565-588. so glad. The general concept which subsumes both affective and epistemic stance acts of the sort illustrated here is positioning (Davies and Harré 1990.895-613.160) LANCE. Maynard 1993. Ochs 1996. (TSK) I’m so glad. is followed by an affective predicate. (8) (Risk SBC024: 588. Speakers may position themselves not only along an affective scale. Kärkkäinen 2003b.. In the examples presented here for both affective and epistemic positioning.50-185.355) JEFF. presenting them- selves as knowledgeable or ignorant: (7) (Appease the Monster SBC013: 185. taken from three different conversations. while the stance predicate (adjective or verb) specifies the nature of the stancetaker’s position. or both (amazed).35) MILES. the speaker who is taking the stance is indexed via a first-person pronoun in syntactic subject role (I). and nasty are used to evaluate something. ideal. Du Bois 2002a). the first person pronoun I. Positioning can be defined. evaluation can be defined as the process whereby a stancetaker orients to an object of stance and characterizes it as having some specific quality or value.330-53. . provisionally.79-667. positioning the speaker as glad. The evaluative target may be called the object of stance (for reasons that will become clear in due course). The stance triangle 143 In these examples. In general terms. . In these stance utterances. Heritage and Raymond 2005. they choose a position along an affective scale – as. I’m glad. The thing evaluated – the specific target toward which the evaluation is oriented – is referred to in each case by the demonstrative pronoun that. indexing the stancetak- er. I don’t know.

I agree with you. Just as often. More commonly. typically the person she is addressing. Heritage 2002. most seem to involve one of three well-differ- entiated kinds of stance function. But what is being achieved thereby is different enough to warrant recognition as a distinct type of stance. The addressee’s role is usually left implicit. but can be made explicit on occasion: (11) (LSAC 1396-01)6 LESLIE.391-128. namely evaluation (that’s horrible). The difference lies in the prag- matic-interactional configuration that it enacts. it is certainly not the usual way. while these are sometimes viewed as different types of stance . and alignment (I agree). I totally agree. But the question remains as to how to interpret this vari- ability. Here a first-person pronoun in syntactic subject position is followed by a stance predicate (the verb agree). These examples make it clear that stance utterances can vary in significant and systematic ways. and more? The problem is. positioning (I’m glad). and by implication between two stancetakers.183) MELISSA. affective stance (I’m glad). But do these stance functions really represent separate categories of stance – or simply different facets of the speaker’s stance. The general term for this kind of stancetaking is alignment (Du Bois 2002a. Du Bois In the next set of examples. more broadly conceived? What about other candidates that have been proposed as stance types – epistemic stance (I know). agreement normally implies agreement with some- one. (10) (Doesn’t Work in This Household SBC019: 127. Heritage and Raymond 2005). The role of implicit stance alignment will prove to be especially important in the management of intersubjectivity. participants allow their alignment to remain implicit. Alignment can be defined provisionally as the act of calibrating the relationship between two stances.144 John W. the structural pattern is superficially similar to the one just presented: (9) (LSAC 1296-02)5 PAT. speakers show alignment by stance markers like yes or no.7 which in conversation is usually the person being addressed. By uttering I agree the speaker defines her stance in relation to that of another party. or gestures like a nod or a headshake. As this example illustrates. Although a stance verb like agree may be the most transparent way to display alignment with another speaker. I agree. Of the cases examined so far. inviting the listener to infer it based on comparing the relevant stances. or any number of other forms that index some degree of alignment.

In other words: Can a single stance combine aspects of more than one of these stance functions? This question raises the possibility of an alter- native approach which recognizes a more complex picture of the stance act. let us begin this discussion where it has so often begun before in the study of language – not with the utterance. with a little imagination. but simply as different facets of a single unified stance act. Building a more unified understanding of stance has the advantage of avoiding a limitless proliferation of stance types. More seriously. which have specific content and are located in a particu- lar dialogic and sequential context. In part. Contextualizing the emerging stance If stance is an act. What can the decontextu- alized sentence reveal – and what can it not? Consider the following sentences: (12) The Caribbean is incredible. This is the issue we begin to explore in the next section. The stance triangle 145 acts. for example. no one has yet resolved the question of how many such types should be recognized – whether there will turn out to be three. and who are they agreeing with? These questions point to some important issues about stance. The answer to some of these questions may hinge in part on the realization that. we should expect to locate it in utterances. 3. in stances involving alignment (9–11). it seems equally essential to ask: What is the speaker agreeing about. or many more subtypes of stance. we need to learn more about how speakers realize stances and how hearers interpret their situated meanings. something seems to be missing. For the sake of ar- gument. seen as encompassing multiple facets at once. five. be hypothesized as an abstract linguistic structure detached from any mooring in a specific context of use. in the analysis presented so far. this question con- cerns how participants contextualize the stance utterance in order to interpret it. The choice between proliferation of stance types and integration of stance facets will be returned to below. however. but the sentence. . The sentence can. as the stance emerges across successive utterances through processes of dialogic action. Nor have the criteria for deciding the question been established. In stances which were characterized as positioning (4–8). Speakers do not just perform generic stance types. there is the matter of whether the various forms of stance occur sepa- rately or together. Such an approach would seek to interpret the diversity of stances not as distinct types of stance. they per- form specific stance acts. the question arises: What is the speaker positioning himself about? Similarly. (13) It was really great. To fully understand what stance is being taken on any given occasion.

In the following utterances by three different speakers. Each of these sentences contains at least one apparently evaluative word: incred- ible. typically by supplying each transcribed utterance or turn with a label indicating the identity of the speaker. Du Bois et al.075-212. But stance is more than the context-free connotations of words or sentences. Any utterance carries cues for its own indexical contextualization (Gumperz 1992. The missing ingredients can only be found by contextualizing the utterance. nasty. Contextualization cues (or indexical signs) work by pointing beyond the utterance to its presupposed conditions of use. In fact. participants normally care who says what.730) JOANNE. the identity of the speaker is indicated as usual at the beginning of each turn. but imply value judgments regarding some refer- ent. These words are obviously not neutral descriptions of external reality. . Silverstein 1976) helps us to identify those aspects of context which must become known in order to arrive at a successful interpretation of the stance at hand. and monitor it ac- cordingly. There are at least three things we need to know about a given occasion of stancetaking.146 John W. Silverstein 1976). great. defined as the situated realization of language in use. great. Jefferson 2004). Du Bois (14) I would love to go. Participant awareness of the attributability of utterances is routinely represented in most systems for transcribing discourse (Du Bois 1991. and love (and perhaps others). and so on). the first question is: What’s missing? In pragmatic terms this translates to: What are the indexical absences? Tracing out the salient indexical meanings (Jakobson [1957]1990. the evaluative meaning often comes through even in an arbitrary listing of individual words (incredible.1 Who is the stancetaker? In conversation. horrible. beyond what may be overtly present in the words and structures of the stance sentence itself: (1) Who is the stancetaker? (2) What is the object of stance? (3) What stance is the stancetaker responding to? Each question points to one component of the process of interpreting stance. stance re- mains incomplete. Thus in situating the understand- ing of any stance. ideal. The evaluative connotations of such words are evident even from sentences taken in isolation. the– the Caribbean is incre:dible. 3. Peirce [1885]1933. 1993. (15) (Deadly Diseases SBC015: 210. In the grammarian’s standard presentation of the isolated sentence.

just attributing speakership in this way doesn’t reveal much unless the speaker’s identity carries some significant associations for us.05-1326. but what they are speaking about. I would love to go:. A key component of the context of any utter- ance is the speaker who is responsible for it. which they may retain as socially salient. that Marilyn is responsible for the statement that something was really great. what sort of relationship they have displayed up to now relative to co-present others. in part because of the dialogic connections that arise between stances. participants derive memorable information about each other from stances taken. in addition to what the speaker is saying right now. Participants remember interactionally salient information about co-participants.2 What is the object of stance? To make sense of a given stance we need to know not only who is speaking. or other identities may be. ethnic.85) MARILYN. But in most conversational settings. with the option of introducing it into future processes of stance interpretation. and so on. The stance triangle 147 (16) (Conceptual Pesticides SBC003: 1326. where the speaker displays a desire to go. Consider the following examples: .070) KEN. Knowing which social actor is responsible for a specific stance utterance in the past can make a significant difference in the interpretation of a current stance utterance. and so may factor into their stance in- terpretation. what accent. Among other things. gender. whether they appear entitled to their claimed identities. what their displayed regional. it was really great. what is claimed to be incredible or great. we are told that someone named Joanne8 is responsible for the claim that a certain place is incredible. One important difference between attested utterances like those in (15–17) and hypothetical sentences like those abstracted in (12–14) is that a real utterance is always framed by its context of use. 3. and so on. if known. and that Ken is responsible for a display of affect or preference about going somewhere. details of their life story. some or all of the following: what the speaker has said previously (whether on this occasion or some other).025-188. To be sure. regardless of whether they happen to know the speaker’s name. voice quality. we need to know the referen- tial object or target toward which the stance is being directed – for example. and intonation they are speaking with. (17) (Deadly Diseases SBC015: 187. participants may draw on a range of biographical associations for the current speaker. Moreover. In the representations of utterances given here.

differ slightly in their evaluative predicates (is great vs. If two people each evaluate something as great.. Stance is a property of utterances. you know. Compare the following two stance utterances. but they will be great in very different ways. A crucial part of interpreting any stance utterance is identifying the object of stance.05-1326. demonstrative noun phrase.730) JOANNE. it was really great. That stuff is great. which both employ the word in- credible in the evaluative predicate.18-1326.82) CAROLYN. (19) (Conceptual Pesticides SBC003: 1326. cannot be regarded as the same. It’s not just pronouns that need to be resolved by contextualizing the stance utterance.82) CAROLYN. it).60-170. What we want to know. (21) (Raging Bureaucracy SBC004: 15. but apply it to rather different stance objects: (22) (Deadly Diseases SBC015: 210.85) MARILYN. But these small differences don’t really speak to the main question. and utterances are inherently embedded in their dia- logic contexts. as part of the process of referential grounding (Hanks 1990). Du Bois (18) (Raging Bureaucracy SBC004: 17. (H) The male athletes were incredible. S– really relaxing weekend. if we are to decide what stance is being taken. the– the Caribbean is incre:dible. was really great) and in the noun phrases which designate their respective objects of stance (that stuff vs. (0.67-18.42-18. The respective stances of Marilyn and Carolyn. it was really great. Often the immediate prior discourse provides sufficient contextualization to resolve the reference of a pro- noun.148 John W. have they taken the same stance? The cited utterances. A weekend and a drink of hibiscus cooler may both be great. (23) (Cuz SBC006: 168. hibiscus cooler. drawn from two different conversations.30) ALINA.85) MARILYN. or other referring form. . once referentially grounded.075-212. not of sentences. thus establishing the identity of the relevant object of stance: (20) (Conceptual Pesticides SBC003: 1324. but also the meaning of content words and other elements as well. is what it and that stuff refer to. You know what’s good is .2) That stuff is great.

Why this stance is being taken.420-188. why just now. Despite the fact that as sentences these are grammatically complete. and so on). if you will. as is the case here. they agree with someone about something. as if carrying its meaning complete within itself. (H) I don’t know how many Americans have been to Nicaragua. for its interpretation it must still indexically incorporate a prior stance content. This contextualization remains essential to the interpretation of the current stance even if the relevant prior stance occurred quite a bit earlier. but specifically to going to Nicaragua. Here. But this illusion is soon dispelled by examination of the prior discourse: (24) (Deadly Diseases SBC015: 164. The stance object that Ken is evaluating is thus to go to Nicaragua. In sum. Consider the examples of alignment given previously (I agree. the contextualized stance utterance takes its interpretation in part from the prior stance of a dialogic co-participant. as stances they are pragmatically incomplete. for participants and analysts alike. Nor- mally the relevant stance content will be locatable in the prior discourse. abbreviated as to go. tranquility. I would love to go:. identifying the object of stance – what the evaluation is about – is an essential part of the process of stance interpretation. 3. In I would love to go (17). People don’t agree in the abstract. agility. and so on) or to athletes (valued for the active display of strength. specify- ing what specific stance is being agreed with: . but we remain on uncertain ground until we know what prior stance the current stance is being formulated in response to – its counterstance. In such cases. talent. I agree with you). The stance triangle 149 The evaluative content of the message conveyed by an evaluative word like incred- ible may shift according to whether it applies to a vacation destination (valued for passive attributes like visual beauty. in Langacker’s (1987) terms – the dimen- sion of alignment almost exclusively. the utterance may initially appear more self-con- tained.3 What stance is the stancetaker responding to? Knowing the identity of the stancetaker and the object of stance is a good start. including the relevant object of stance. why in these terms – to answer these questions we need to monitor the dialogic and sequential shape of the ongoing exchange of stance and counterstance. While a sentence like I agree with you (11) foregrounds – or profiles. the phrase love to go refers not to going in the abstract. ((39 LINES OMITTED)) 41 KEN.070) 1 JOANNE. nor to going some- where at random.

31 I should go downstairs.11 To generalize.150 John W. to find the proposition spelled out overtly.12 Only by referencing the relevant prior stance. 2 PAT. The result can be informally paraphrased9 as I agree with you that we are considered white collar (because we’re social workers). Du Bois (25) (LSAC 1396-01) 1 CORA.. Yeah. whose rel- evant context is as follows: (27) (Doesn’t Work in This Household SBC019: 98. Leslie’s I agree with you creates a convergent alignment with her addressee which immediately entails or implicates an endorsement of the addressee’s stance. I agree (with you that it’s) no more or no less than any other school. Take it downstairs. 3 I agree. consider Melissa’s utterance of I totally agree (10). Similarly in the following example: (26) (LSAC 1296-02) 1 KIM.. When Pat says I agree in line 3 (and again in line 4). 4 I agree. if need be. it is not possible to tell from the sentence alone what is being agreed with. Here. ((28 LINES OMITTED)) 30 MELISSA. No more or no less than any other school is the way I see it. ((2 LINES OMITTED)) 4 LESLIE. and can be paraphrased as. yielding a composite stance that is roughly paraphrasable as I totally agree (that) I should go downstairs.652-129.. I totally agree. Her stance is quite specific. Melissa is not agreeing in the abstract. Lest there be any doubt as to what proposition she is agreeing to in line 30. The requirement to specify the particular content for the stance of agreement is so strong that participants will go back 29 lines in the conversation.138) 1 JAN. I agree with you). but specifical- ly agreeing with her mother Jan’s directive that she should take her homework downstairs.10 Finally. locatable ana- . I totally agree. Melissa makes it explicit in her next utterance (line 31). In saying I totally agree. I agree with you. she is doing more than just “being agreeable” (or even “doing agreement”). given a decontex- tualized sentence which apparently expresses simple agreement (I agree. we are considered white collar because we’re social  workers.

if limited to the sentential level of referring-and-predicating functions (Silverstein 1976. prosody. Cer- tainly there are additional questions that will need to be posed as we continue our efforts to tease out the recurrent features of the stance act. But it is important to point out here that even the questions in- formally posed so far frequently require participants (and analysts) to go beyond what is explicit in the words of the stance utterance itself. 12–14) in isolation. but is grounded ultimately in the systematic knowledge which participants control regarding what can be expected to be present in any stance. We have been considering how the contextualization of actual stance utter- ances (e. The claim is that in each case. Still the questions must be answered. Their relevance to stance does not depend solely on the presence of explicit words. these questions about stance can be linked to notions of stance subject. The constant relevance of the general components of stance influences what we expect to know about any act of stancetaking. even if the sentences are fully grammatical and fully meaningful – at least as meaningful as they can be. what we want to describe is the participant’s interpretive process. whose overt theoretical positioning may otherwise appear so disparate as . The stance triangle 151 phorically in the dialogic context. the necessary ingredients start to fall into place for a full pragmatic and interactional interpreta- tion. 15–27) contributes to the interpretation of stance. Yet once we identify the relevant contextual features. In asking.. and thereby shapes its spe- cific interpretation. gestures. can the meaning of the present agreeing stance be understood. In contrast. to the interpretation of any act of stancetaking. in some formulation or other. connections which will be further developed below. or the best formulations of them. This interpretive inquiry is akin to that which has motivated scholars as diverse as Geertz (1973) and Sacks (1992). In more general terms. In constructing a logic of stance interpretation. We have considered three questions about the context of stance which are likely to be relevant. what is the stance about. however important these may be. Who is the stancetaker.g. or necessarily the main ones..g. little or nothing of the relevant contextualizing information can be gleaned from in- specting idealized sentences (e. But we are far from claiming that these are the only questions. 2001). or is only to be found distributed across multiple utter- ances by different speakers within extended sequences of dialogic exchange. or other communicative elements. certain well-defined items of information are actively sought out by participants in response to the projectable structure of stance. This holds true whether the information is directly expressed in the stance utterance. and alignment. and some of these will come up below. and what stance is the stancetaker responding to? we are seeking to fill in some of the blanks that must be filled if we want to understand what the stance now being tak- en actually is. which we track by close observation of their own interpretive actions in stance-rich environments. stance object.

12-445. the subjective dimension is registered overtly through several discrete linguistic elements. Subjectivity and positioning In light of this understanding of the processes of stance contextualization. don’t like. The personal pronoun I points directly to the speaking subject (Benveniste 1971. including personal pronouns (I) and affec- tive verbs (like. love. we are now ready to re-examine the issue of positioning. inasmuch as the act of positioning regularly invokes a dimension of speaker subjectivity. We cannot overlook the influence of even so basic interpretive processes as these if we want to understand how stances come to be understood. the specificity of stance is grounded in its dialogicality. In these stance utterances. is the one taking the stance. Voloshinov [1929]1973) to arrive at a sufficiently enriched inter- pretation of the utterance (Recanati 1989.152 John W. Only when the dialogic context is taken into account does stance become complete. Yet it cannot be overlooked that subjective predi- cates which are transitive (like. Subjectivity and positioning go hand in hand. and so on) regularly specify also the object to which the subject is orienting affectively. The affective stance predicate indexes spe- cific aspects of the subject’s feelings. Schegloff 1996. In sum. Hanks 1990.91) JAMIE. Sperber and Wilson 1995). Silverstein 1976) in dialogic and sequential context (Linell 1998. 4. We need contextual grounding (Gumperz 1992.14 The transitive clause I like this song (28). Ducrot 1972. This example compactly illustrates the co-existence of subjective (I) and objective (this song) elements within a unified stance. Culioli 1990. positioning the speaker subjectively along some scale of affective value. Similarly. uttered in the auditory presence of a song playing on the stereo. directly specifies the object of stance – the entity being oriented to. hate. (29) (This Retirement Bit SBC011: 444. in these cases at least. I like this song. I don’t like those. respectively). what is required is an orientation to a specific object of the speaking subject’s . don’t like). Ultimately. the coherent expression of speaker subjectivity requires both a subject and an object of stance (specified as I and those. Du Bois to be incommensurable. which we consider in light of speaker subjectivity. 1984)13 who. Con- sider the following: (28) (Lambada SBC002: 865. Sacks 1992. in I don’t like those (29).36) SAM. to articulate this kind of subjectiv- ity.

355) JEFF. but it also commits the stancetaker to a certain evaluation of the object. In these examples. with respect to a scale of affective value (glad). In contrasting Lance’s I’m glad with Jeff ’s I’m so glad. First. speaking subjects position themselves subjectively – and that seems to be the end of it. and so on. we label the words which overtly express or index the stance subject and the stance object. in analyzing the various discrete components of the utterance. we label the verb or other stance predicate according to the kind of stance action it performs. we find it essential to represent the identity of the speaker. not liking. what is immediately obvious as a difference is the presence of the intensifying adverb so in the latter case. because this tells us who is the speaking subject (the stancetaker). liking. (TSK) I’m so glad. While some stance utterances evidently perform combinations of functions (for example. the predicate (like. While this may be true as far as it goes. (33) (Hey Cutie-Pie SBC028: 52.160) LANCE.895-613. It is true that there are stance utterances which overtly po- sition the speaking subject without explicitly including any reference to a stance object. don’t like) obviously serves to position the sub- ject. evaluation of a stance object combined with positioning of a stance subject). loving. hating. We might take this to indicate that Jeff is positioning himself as claiming a more intense subjective experience along the scale of gladness than Lance is claiming. repeated here for convenience: (32) (Runway Heading SBC022: 612. The stance triangle 153 stance. I like this song (31) Speaker Stance Subject Positions/Evaluates Stance Object SAM. to focus exclusively on the subjective side of the equation is to . In what appear to be simple one-place predicates (Thompson and Hopper 2001). Consider examples (4) and (5). In recognition of this apparent dual stance function. I don’t like those This kind of stance diagram is useful in making explicit several aspects of our analysis. Third. such as desiring. combined with specification of a particular intentional relation (Searle 1983). Up to now we have analyzed these cases as simply reflecting the speaker’s subjec- tive self-positioning.330-53. The following stance diagrams represent the relations of evaluation and posi- tioning between the stance subject and stance object: (30) Speaker Stance Subject Positions/Evaluates Stance Object JAMIE. the representations of evalua- tion and positioning are combined in a single column. we might still ask whether it is possible for a stance utterance to express only pure subjectivity. Second. I’m glad.

. his trainer: (34) (Runway Heading SBC022: 607. I’m glad.6) 4 JEFF. Y:es:. Such appears to be the case in examples (32)–(33) above.355) 1 JEFF. This explains why two stances in which similar or identical words are used may still differ substantially with respect to what they are stances about. Here the stance content of the utterance I’m glad emerges from successive contri- butions by Randy and then Lance. Are you guys having fun? 2 JILL. Jeff is talking on the phone to Jill about her friend who is visiting her. N:o significant problems.15 In the second example.154 John W. If subjectivity requires orientation to an object. The difference between Jeff ’s and Lance’s stances may hinge primarily on what they are glad about – the object of stance – but this is left unmentioned within the stance utterance itself. the full meaning of any subjective stance must remain mysterious until we locate the object. and asks: (35) (Hey Cutie-Pie SBC028: 49. In the first I’m glad example. their actual significance becomes clear. (TSK) I’m so glad.505-613. with the cumulative result being paraphrasable as something like I’m glad (that there are) no significant problems to talk about. But can an object be part of the stance if it is not part of the sentence? The answer may hinge on whether orientation to an object is taken to be a necessary part of the process of constituting subjectivity. But that’s not to say that the participants are not orienting to a stance object. The evidence we have seen points to a positive conclusion: displays of subjectivity always make relevant the relation between a stance subject and a stance object.160) 1 RANDY. Lance. .985-53. Here the stance object is. If the stance object is not overtly specified within the immediate stance utterance. to talk about. participants will feel that something is missing. 2 (0. is being debriefed after a work session by Randy. even if this requires us to search the discourse context to find it.5) 3 #to #dea– — 4 . Du Bois leave out a key variable: the object of stance. once again. what the speaker is affectively orienting to. The claim is that it is a regular feature of subjectivity to orient to an object. an apprentice air traffic controller.0) 6 LANCE. 5 (2. Once we take into account the sequential context in which the stances developed in the first place. 3 (0.

Clearly. Hm. The following diagrams represent the relevant relationship between stance subject and stance object: (36) Stance Positions/ Stance # Speaker Subject Evaluates Object 6 LANCE.79-667.75-667.0) 697 PETE. People are not glad in general. The stance triangle 155 Quite a different stance emerges from this sequential contextualization. I’m just ama:zed. Miles’ stance amounts to something like I’m just amazed (that) there’re a lot of women out there who (apparently) don’t believe in using condoms. I’m just ama:zed. the stance act of affective self-positioning (as . 698a (0. but is instead based on differences in the dialogic sequence through which the stance emerged.16 The main difference in meaning between the last two examples turns out to have little to do with the presence of the intensifier so. Summing across the full discourse context. To take into account only the current stance utterance (I’m glad) would be to miss out on what the stance predicate glad is pointing to.35) 69817 MILES. but glad that (something). 697a (1. Obviously it makes a difference whether the complement of glad is to be understood as you guys are having fun or there are no significant problems to talk about. I ’m glad {there are no significant problems} (37) Stance Positions/ Stance # Speaker Subject Evaluates Object 4 JEFF. the specific entity or state of affairs toward which the speaker expresses their subjec- tive stance. But the sequential context makes it clear that Miles is not just amazed. para- phrasable as I’m so glad you guys are having fun. the next stance utterance appears at first to be a simple display of amazement: (38) (Lambada SBC002: 665. Even self-positioning presupposes an object. namely. Cause there’re a lot of women out there who apparently don’t believe in using condoms.35) 696 MILES. he is amazed about something: (39) (Lambada SBC002: 660.6) 698 MILES. I ’m so glad {you guys are having fun} Along the same lines.

Kendra and Dan each speak no more than three words.13-921.970) DAN. At least this is what the words taken literally and in isolation seem to mean. How far off the mark they are becomes evident once we take into account the dialogic con- text: (42) (Appease the Monster SBC013: 919. . Are we not attacking each other until we get rid of the striped guy? DAN. this should be as relevant to epistemic subjectivity as to affective subjectivity. The conclusion: Subjectiv- ity takes an object.565-588. I kno:w. But the stances they achieve through their slender utterances are more complex than what is immediately evident in the stance utterances themselves. I don’t know.. repeated here for convenience: (40) (Appease the Monster SBC013: 920. Even in its absence. Kendra positions herself as knowing. yet manage to produce ap- parently complete stance utterances thereby. . Consider examples (7)–(8). the stance object remains relevant and hence may trigger a search for it in the prior discourse. (43) (Risk SBC024: 585. Those are good spatula[s]. I know {those are good spatulas} .630-588. [I] kno:w... But such interpretations in isolation are meaningless. Du Bois glad or amazed) is incomplete until we include the object of stance – what the speaker is glad or amazed about. On the face of it. Although it may be merely implicit in the cur- rent stance utterance. as shown in the following representations: (44) Stance Positions/ Stance # Speaker Subject Evaluates Object 2 KENDRA.970) JENNIFER.156 John W. the stance object is an indispensable component of even a subjective stance. I don’t know.23 -921. while Dan positions him- self as not knowing.13) WENDY. Their respective stances emerge only from the larger dialogic sequence.13) KENDRA. If subjectivity must have its object. KENDRA. . (41) (Risk SBC024: 588.

. is already there in the immediate prior discourse – often in the utter- ance of a dialogic partner. if not present in the sen- tence itself. The stance triangle 157 (45) Stance Positions/ Stance # Speaker Subject Evaluates Object 2 DAN. sentence. Likewise. is a big part of what it means for linguistic ac- tion to be dialogic. as realized in the con- text of conversational interaction. as successive stance utterances – stance and counterstance – are deployed in response to each other. the stance which culminates in a short and apparently simple utterance like I’m glad or I know cannot be a matter of subjectivity in isolation. The link that is constituted between subject and object creates a vec- tor of subjectivity. Stance utterances like I know and I don’t know are designed to incorporate their dialogic antecedents. clause. It takes Wendy’s and Kendra’s utterances together for Kendra’s stance to emerge. or even turn. subjectivity proves meaningful only when subject and object are defined in relation to each other. the stance act is not necessarily complete within a single intonation unit. Second. the prior stance must be incorporated ana- phorically into the interpretation of the overall emergent stance which culminates in the current stance utterance. Generally the precise specification of what they know. Third. Rather. The subject-object link is often achieved dialogically. and so on). In the end. a stance verb like know often points to a dialogically prior stance. Despite initial appearances. the requirement for inclusion of a stance subject (I) is intimately connected to the requirement for inclusion of a stance object (the state of affairs that the speaker is glad about. it necessarily combines a subjective and an objective component. This has important consequences for our understanding of subjectivity. it takes Jennifer and Dan working together to articulate Dan’s emerging stance. Rather. articulating a precise indexical relation to it. This kind of co-action. solipsistic state of the individual psyche. whether affective or epistemic. Despite popular conceptions of subjectivity as a purely internal. First. through which they gain the interpretative specificity they need to be complete. we see from the evidence of stancetaking that the presence of a subjective element in no way precludes the presence of an objective element as well. I don’t know {if we are not attacking each other until we get rid of the striped guy}18 The point is that people do not normally present themselves as knowing (or not knowing) in the abstract. In constituting subjectivity. Once our analysis systematically incorporates the dialogic co-participants’ contributions to the emerging stance. through separate but coordinated contributions by several co-actors. informed about. several things start to become clear. they know (or don’t know) particular things.

In many such cases. and to attribute each element to a distinct portion of the discourse. We need to ask what else may be needed to complete the picture of the stance in ques- tion. and the prior stance being responded to – make for a good start toward filling in the information that we need to interpret any given stance utter- ance. no stance stands alone. The three questions about stance contextualization that were introduced at the outset – regarding the identity of the stancetaker. The concrete localization of the overt words or other meaningful ele- ments which ground the various components of stance is important for assessing the compositional contribution of the several evaluative words and constructions in a stance utterance. but dependent on the prior context. the ob- ject of stance. however. But still more is needed if we wish to grasp how the contextually recovered stance components combine with the explicit components of the stance utterance itself to yield the stance interpretation. This predictable pattern remains relevant even if one of the elements is not explicitly verbalized within the current stance utterance. whether as a specific word or phrase in the current or prior stance utterance. To get to the next level of understanding. In the end. It is in fact routine to find complex stance culminations. it is possible to distinguish the object-centered element from the subject-centered element.158 John W. Du Bois Though we have not yet arrived at a full understanding of what is needed to contextualize and interpret stance. it is necessary to monitor the developmental history of the emergence of the stance. a key issue that remains to be addressed concerns the role of intersubjectivity. but is to be located anaphorically in the discourse of a prior speaker. as its content is enriched via the structure of expectations regarding the necessary recurrent components in the general model of stance. the place of the intersubjec- tive in relation to other elements of stance will prove essential in arriving at a more complete picture of the interpretive matrix within which stance is dialogi- cally realized. . As with the subjective and objective elements of stance. precipitated in the end by a seemingly simplex stance utterance. we have come far enough to draw some preliminary conclusions. It is the stance utterance with its dialogic context that is the relevant unit for stance interpreta- tion. we have identified a regular pattern whereby a single stance utterance culminates in the bringing together of an object-centered act of evaluation and a subject-centered act of positioning. But the full significance of these bits of information must remain obscure until we understand better why we should need them in the first place. In order to arrive at a situated interpretation for any particular stance utter- ance. So far.

The stance triangle 159 5. intersubjec- tivity rises to focal prominence. the shared stance object becomes the cornerstone of the dialogic construction of intersubjectivity. I don’t like those. But even after identifying the missing stance object. Consider the following exchange. Some hint of what remains to be incorporated is suggested by a simple observation. we will have to expand our view to encompass more than one subjectivity – to bring into focus the sociocognitive relations that arise between two subjectivities.2) 3 ANGELA. see Hobson 1993. In the dialogic realization of stance. Analogical relations are established between the juxtaposed stances (I don’t like those : I don’t either). Tomasello 1999. we still only have part of the story. we will have a real opportunity to witness the dialogic emergence of intersubjectivity.) As we will see. I don’t either. 2 (0. 2005. A similar . at least etymologically.12-446. Moreover. Kidwell and Zimmerman 2006. A positioning utterance like I’m glad foregrounds its subjectivity via overt cues such as first-person pronouns. Moore and Dunham 1995. the subjective stance pre- sumes an orientation to an object. The foregrounding of this dialogic relation potentially invites inferences based on the comparison. But to observe intersub- jectivity in action. affective predicates. What are we to make of intersubjectivity? How does it enter into the dialogic realization and interpretation of stance? As the word itself suggests. whether overtly present in the stance utterance (I don’t like those) or not (I’m amazed). Through this kind of sequential juxtaposition of evaluative stances. most evidently in the case of positioning. and other elements that index salient aspects of the speaking subject. which provides a larger window of context for example (29): (46) (This Retirement Bit SBC011: 444. (For the related notion of joint attention. Intersubjectivity and alignment We have seen how subjectivity figures in stance.30) 1 SAM. Indeed this object-orientation may extend across multiple stance acts by different speakers. the subjective orientation to a stance object may be shared among more than one participant. This gives rise to what I call the shared stance object. where what is positioned is typically the speaking subject – the stan- cetaker. intersubjec- tivity presupposes subjectivity. Tomasello et al. When we learn to see how one speaker’s subjectivity reacts to another’s subjectivity. when the subjective stances of two participants collide within a dialogic exchange.

which are specified as parallel within the diagraph. but is simply one tool to be used in conjunction with oth- ers.6) 3 MARY. Note that the diagraph is not intended to capture everything about the utterances represented. I don’t know if she would either.19 and columns correspond to the specific elements which resonate between different utterances. diagraphs display the parallelism of elements using a representation like the following (abstracted from example 47): (48) 1 ALICE: I do n’t know if she ’d do it .) The two most immediately relevant dimensions of the diagraph are. foregrounding their relationships to each other. A word is in order regarding the organization and significance of diagraph analy- sis.‘across’ + graph ‘mapping’. While the diagraph was developed for independent purposes – primarily for ana- lyzing the structural relations that characterize dialogic syntax – it can be equally effective in bringing out the similarities and contrasts between stances in dialogic sequence. As such. Du Bois dialogic relation arises between the two participants’ stances in the next example as well: (47) (A Tree’s Life SBC007: 581. It is useful to have a way to represent the implicit structure of the stance parallels in such cases. but rather should remain implicitly linked to it for purposes of interpretation. The recognition of a central role for participant identity in resonance relations is part of what makes dialogic syntax dialogic. A further element that is always included in the diagraph is speaker labels. 3 MARY: I do n’t know if she would either. I don’t know if she’d do it. the diagraph representation is not supposed to replace the original transcription with all its de- tail. 2 (0. The diagraph is designed to represent the mapping of structured resonances across utterances. (Diagraph.99) 1 ALICE. One tool that has proven useful for representing dia- logic relations between stances is what I have called the diagraph (Du Bois 2001). The diagraph focuses primarily on those elements which display significant resonance. means liter- ally ‘mapping across’. in a way that perspicuously captures the potential for participant inferencing about stance. from dia.32-585.160 John W. The connection between the two representations of the same discourse excerpt is made explicit via such notational devices as line numbers which cross-reference . in simplistic terms: rows correspond to utterances. which serve to index the identity of the participants who enact the dialogic resonance in question. Foregrounding the relevant structural relations. the diagraph is formatted so that elements that resonate with each other are aligned vertically in columns. To display the resonance icon- ically.

and this is no mere adornment. along with their parallels. Mary’s stance follow is virtually required to include . In this light. what does the diagraph in (48) tell us? One speaker takes a stance. One way to look at stance is to ask how it is constituted as an action within an interpretive framing erected by the ongoing dialogic activity. Schegloff 1996. Silverstein 1984. the two stances are also fundamentally different. the word either in this con- struction serves to index a specific intersubjective relation between two speakers engaged in dialogic interaction. see Du Bois 2001. the evidence from many similar cases makes it clear that either cannot normally be omitted from the second stance utterance without causing pragmatic anomaly (Du Bois 2004). (For more detailed discussion and exemplification of diagraph analysis. despite their similar stance content. The second stance utterance ends with the word either. The stance triangle 161 each line of the diagraph with the corresponding line of the transcription. see Blanche-Ben- veniste et al. If Mary had responded to Alice’s utterance with a lexically identical utterance – just I don’t know if she’d do it – the effect would likely be perceived as somewhat strange. the diagraph in (48) can be interpreted in terms of who leads and who follows. while the sec- ond (Mary’s in line 3) is a stance follow. The diagraph in (48) shows how Alice responds to Mary’s I don’t know if she’d do it with a very similar utterance: I don’t know if she would either. As similar as these two stance utterances are. The first stance (Alice’s in line 1) can be characterized as a stance lead. suggesting that there is a general principle involved. Why is the presence of the word ei- ther so crucial here? The same issue arises in I don’t either in (46). 1981. Linell 1998. Harris 1952. the participants mark their stances differently. the intersubjective relationship between one’s own resonating stance and that of the prior speaker must ordinarily be acknowledged indexically – if one wishes to avoid being judged interactionally incompetent. In both cases. There is no other explana- tion for the virtually obligatory presence of either in such sentences than its role in indexing the intersubjective relationship between two stances in dialogic juxta- position. the differences turn out to be impor- tant because of what they tell us about the dialogic relations that are established through the sequential realization of stances. While space precludes full exploration of the de- tailed workings of this pattern here. whenever an interactionally salient dialogic resonance arises between two stances. in more or less the same way. For related work on structural parallelism in discourse. then a second speaker takes a seemingly equivalent stance. Tannen 1987. But a closer examination of the diagraph makes it clear that.) Now. In general terms. in part because of the absence of the word either. 1991. there is a limit to their convergence. Jakobson 1966. Though subtle. Johnstone 1994.21 The importance of this contrast becomes clear when we note that.20 The strangeness cannot be explained away as a problem with an “echoic” utterance: saying just I don’t know if she would would be pragmatically aberrant as well.

which serves an intersubjective indexical function here. Thus two participants in dialogic interaction should be understood as engaging in the alignment process when they converge to varying degrees. or as often happens. a negative pole (disaligned). by the same token. so that stance alignment can be relatively positive or negative – or. In contrast to common usage which forces a binary choice between a positive pole (referred to as aligned) vs. By recognizing the variability of scalar align- ment we can take into account the fact that stances are aligned by subtle degrees. pragmatically if not grammatically. from including the intersubjective indexical either. the approach I favor treats alignment as continuously variable in principle. convergent or divergent to some degree. we have been working to lay the groundwork for the next stage of theorizing stance. ambiguous between the two. In developing new ways of analyzing these and other elements. as well as the sociocognitive relations of objective. 6. To the question posed at the outset (in Section 2) as to whether . or tri-act. divergent. positioning. Alignment is in play whether the direction is convergent. as well as modal (if) in effect. subjective. while the subjective positioning in (47) is epistemic (don’t know). The stance triangle We have been assembling an analytic toolkit of interconnected concepts and methods designed to shed light on the various elements and processes of stance. the use of either in the cited examples should be seen as part of an act of alignment that serves to calibrate the intersubjective relationship implicit in the stances of en- gaged co-participants. words like either (or too. more precisely speaking. when they diverge to varying degrees. Example (46) is like (47) in this respect: the intersubjective use of either is pragmatically required. and. The subjec- tive positioning in (46) is affective (don’t like). as discussed below) can be said to function as intersubjective alignment markers. Du Bois the word either. But the need to index the intersubjective relationship between a stance follow and a prior stance lead remains the same. In this light. On the level of action.162 John W. There is an important connection between intersubjectivity and the stance act of alignment that is visible in these examples. Now it is time to bring all the elements together to forge a unified framework for stance. Alice’s stance lead is virtually precluded. Key components include the concepts of evaluation. and intersubjective intentionality. and alignment. represents a point along a continuous scale or range of values. In such cases. In con- trast. stance is to be understood as three acts in one – a triune act. Note that alignment. The picture we are moving toward is one in which stance is seen as a single unified act encompassing several triplet sets of distinct components and process- es. as I use the term.

Figure 1. adopting the first-person point of view of the stancetaker as speak- ing subject. The following definition sums it up: (49) Stance is a public act by a social actor. Rather than three separate types of stance. Alternatively. and aligning with other subjects. achieved dialogically through overt communicative means. yet the three are yoked together through their integration in the dialogic stance act. the view from the stance triangle suggests that they are simply different aspects of a single stance act. unified stance act. and thereby align with you. the stancetaker (1) evaluates an object. In tak- ing a stance.  The stance triangle . (2) positions a subject (usually the self). The stance act thus creates three kinds of stance consequences at once. positioning subjects (self and others). we can informally gloss this definition as follows: (50) I evaluate something. and thereby position myself. Each subsidiary act is distinguishable from the others by virtue of its own distinctive consequences. and (3) aligns with other subjects. we interpret them as subsidiary acts of a single overarching. positioning. with respect to any salient dimension of the sociocultural field. of simultaneously evaluating objects. The stance triangle 163 evaluation. and alignment represent three different types of stance.

While the stance triangle com- prises the three subsidiary acts of evaluating. each of the three stance act vectors is relational and directed. wheth- er the alignment is convergent or divergent. The three sides of the triangle represent vectors of directed action that organize the stance relations among these entities. The stance triangle shows how a stance utterance that specifies only one of the three vectors can allow participants to draw inferences about the others. and again the very same three stance acts for the second subject. and aligning. The stance triangle is unusual in that it depicts three stance acts for the first subject. Rather. as long as the rest of the triangle is known. This makes for a total of six arrowheads. as in the expected one-to-one cor- respondence found in conventional triangular models. the second from the second subject. For each vector of directed action in the diagram. positioning. Because there are two social actors represented in the stance tri- angle – the first and second stance subjects – there are two tokens of each action vector type. namely the first subject. The clearest way to represent the stance model I am proposing is in the form of a triangle (Figure 1). these are not distributed evenly among the three sides. Vectors of alignment may originate in either the first or second subject and be directed toward the other subject. positioning. two of the three sides represent evaluative vectors directed from one of the two stance subjects toward the single shared stance object. The three nodes of the stance triangle represent the three key entities in the stance act. Crucially for the analysis. all three of the three-in-one subsidiary acts remain relevant to stance interpretation even if only one or two of them are overtly expressed in the linguistic form of the stance utterance. the second time around.22 And yet the acts are different. it is possible to draw inferences regarding any unspecified portion of the stance triangle. corresponding to the three acts of evaluation. along with some initial suggestions as to how it can expand the analytic reach of the stance concept. Sig- nificantly. stancetakers position themselves. I will further sketch out a preliminary version of this framework. the second subject. The third side of the triangle (the vertical line on the left) represents alignment between the two subjects. The first evaluative vector originates from the first subject. linking two nodes of the triangle. stancetakers define alignment with each other. an arrowhead points in the direction of action’s object or target. Concomitant to evaluating a shared stance object. In the rest of this section. Depending on the circumstances. Concomitant to positioning themselves. For example. Du Bois The beginnings of a framework for analyzing stance are implicit in these defini- tions. and alignment. she positions herself as taking . The stance triangle provides the basis for understanding the causal and infer- ential linkage that may arise between the various subsidiary acts.164 John W. and the (shared) stance object. if Melissa agrees with Jan. doubled by the co-presence of two subjects.

alignment. and the shared stance object (in Sam’s utterance. Attending to the structured interrelations among the acts and entities which comprise stance allows participants. While the stance acts of evaluation and positioning are more or less evi- dent from direct inspection of the conversational example. The three entities at the nodes of the stance triangle are more or less transpar- ently represented in this example – the first stance subject (Sam’s I). I do n’t either. Angela marks her contribution as a stance follow to Sam’s stance lead. I don’t like those. with strong implications for inferencing regarding participants’ positioning. considering it now in light of the stance triangle: (51) (This Retirement Bit SBC011: 444. the stance triangle pos- its a model of the components of stance and of the organization of the relations between them. and ana- lysts. In sum.12-446. including the evaluation Jan has per- formed in her own prior stance. I don’t either.30) 1 SAM. The stance triangle 165 the same stance (roughly speaking) as Jan. To assess these claims for the theoretical significance of the stance triangle. Our understanding of stance is enhanced. 3 ANGELA. those. . the second stance subject (Angela’s I). I do n’t like those . and evaluation. We begin by taking a second look at an example introduced earlier (46). to see the alignment clearly it will be useful to create a diagraph: (52) 1 SAM. Sam’s stance predicate (don’t) like serves both to position the entity expressed by its syntactic subject (I) and to evaluate the entity expressed by its syntactic object (those). this allows us to infer her convergent alignment with the previous speaker. or a deletion. we need to see it in action – to test its utility in the analysis of actual instances of stance in interaction. 2 (0. what some would call a zero. by taking seriously the interrelations among components of the stance act as speci- fied in the stance triangle model. if Melissa expresses an evaluation that is effectively the same as Jan’s previous evaluation. What I am proposing is that the structure of dialogic action repre- sented in the stance triangle offers a framework for analyzing the realization and interpretation of stance. representing the under- standing that Angela is referring implicitly to the same referent as Sam’s those). deploying the word either in its intersubjective alignment function. to draw inferences by triangulating from the explicit components of stance to the implicit. I suggest.2) 3 ANGELA. Conversely. in Angela’s.

Ken is talking specifically about going to Nicaragua. There is always more to stance in dialogic interaction than can be captured in any labeled diagram as simple as this one. the verb specifies both the evaluation of the object and the positioning of the subject.23 (53) Stance Positions/ Stance # Speaker Subject Evaluates Object Aligns 1 SAM. so the two labels are combined in a single column. Angela’s use of the word either indexes alignment.. Note that this stance diagraph representa- tion is only intended as an informal aid to visualizing those elements in the stance utterance which correspond most directly to the relevant stance triangle entities and actions. Nevertheless. (The first line of this example was analyzed above in (17) and (24). we can incorporate labels specifying which entities and actions are present in the stance utterance. if we really want to specify how alignment is achieved in this kind of utterance. Having parsed out the various component acts and enti- ties of a stance exchange via the stance diagraph.540-198. and how they are expressed (or implied) in it. I would love to go:.3) 3 LENORE. positioning. 2 (0. and alignment. The column heading thus marks the alignment function of the word either accordingly. The stance diagraph serves as a useful intermediate stage in the analysis leading to the stance triangle. Yeah. I1 don’t like those 3 ANGELA. the stance diagraph helps to make visible key aspects of the mapping between forms which resonate across utterances. For example. we would have to acknowledge that it’s not just using the word either that does it.166 John W. termed a stance diagraph.) As is clear from the prior discourse. but just as important. taking account of the fact that Angela’s stance utterance is a stance follow which builds dialogi- cally off of Sam’s prior stance lead. Now look at what develops next. the differences that constitute what I have called the stance differential. but also the resonance generated through the act of repro- ducing words and structures of the prior speaker. The next example provides a further test case involving the three subsidiary acts of evaluation. as in the following dia- gram.625) 1 KEN. . in the subsequent exchange of stances: (54) (Deadly Diseases SBC015: 186. in these data. especially in foregrounding the stance act of alignment. This brings out the similarities. . Du Bois To display the analysis in terms of the stance triangle more precisely. one can in principle then map this analysis onto the stance triangle. as well as the stance differential. I2 don’t {like} {those} either As for the three stance actions.

or at least an ambivalence toward. as the relevant prior stance. After another sub- stantial pause. First. Yeah? 6 (0. suggesting a questioning of. To create a more explicit representation labeling the relevant stance entities and actions. the analogy between the two stance utterances (I would love to go : I want to go too) is unmistakable – suf- ficiently transparent to be recognizable across a few intervening utterances. primarily by intonation but also by other factors such as voice quality and sequential placement. The crucial resonance relations in this exchange are highlighted in the following diagraph of Ken and Joanne’s utterances: (55) 1 KEN. Lenore’s more definite yeah. and I would love to go : I want to go too (lines 1 and 7). displaying her convergence with Ken’s stance in line 1. Joanne goes on in line 7 to commit to a sort of stance follow. Even though the words are identical. The stance triangle 167 4 (1. Uttered with a final falling intonation. Al- though it might appear that Joanne is saying more or less “the same thing” as Ken. but two resonance pair- ings in particular are of immediate interest: the two yeah’s (lines 3 and 5). a presumable downgrade of the level of commitment to the proposition at issue. to which Ken and Lenore are both by now on-record adherents. a strong marker of convergent alignment. There are interesting aspects of resonance to be found in various of the possible pairings of stance utterances that could be made here. choosing want instead. consider Lenore’s yeah in line 3.0) 5 JOANNE. 7 JOANNE.25 This sheds light on her lack of commitment as implied by her use of a questioning intonation for what could have been.24 Joanne’s display of a subtle stance differential in line 5 effectively raises some question as to whether she will fully align herself with the prior salient stance. After a one-second pause – a saliently long pause for a lively conversation like this – Joanne utters the same word as Lenore. with different intonation. I want to go too. we recast (55) as a stance diagraph: . but with a differ- ent intonation. I would love to go . a stance differential is created nonetheless. as the diagraph above suggests.9) 7 I want to go too. the differences which constitute the stance differential are actually quite crucial. this amounts to an intersubjective alignment marker. While not realized in immediately successive utterances. Joanne uses a rising appeal contour. Where Lenore had a terminative intonation contour (realized as a final fall) on Yeah. But she withholds the word love.

Yet the argument can be made that the stance triangle applies even in such less-than-transparent cases. from the age of about one year. as in this case. While this question deserves a more extended response than can be presented here. It will be an important task for future research to show how the stance triangle extends naturally to incorporate such observations. 2005. given certain well- defined conditions.. of stance? The answer may hinge on whether . thereby articulating an intersubjective relation between them. I1 would love to go {to Nicaragua} 7 JOANNE. using the appropriate linguistic index: too for positive utterances.. In the present analysis of stance. It has been claimed that all meaningful use of human language.27 Although the role of intersubjectivity in language usually remains implicit or is subtly expressed (e. binding the subjectivities of dialogic co-actors together. One further point about the stance triangle calls for comment here. if partial. Among the more promising candidates were evaluation. the shared stance object obviously plays a critical role. and thereby to achieve an insightful. for example toward a word’s referent (Tomasello 1999.g. and Joanne a stance follow.168 John W. the indexical function of too and either becomes a valuable diagnostic for the individual speaker’s obligatory engagement with dialogically constructed intersubjectivity. which is not omissible in this context without pragmatic anomaly.26 the intersubjectivity involved must be registered obligato- rily. see also Hobson 1993). To- masello et al. sometimes it is realized overtly in the tangible form of a specific word. an individual stance vector constituting the subject-ob- ject evaluative relation).g. a word of commentary may give some idea of what kind of answer will be required. But what about cases which don’t seem to involve a shared stance object? This would appear to present a challenge for the present analysis. I2 want to go {to Nicaragua} too This diagram shows clearly how Ken performs a stance lead. either for negative utterances. At the outset (Section 2) we considered whether we could set up a distinction between different types of stance. then. As this and earlier examples attest. But should we consider these as distinct types. via prosody or sequential placement). The evidence from many such cases makes it clear that when a stance follow is juxtaposed to a dialogically resonant prior stance lead. it is usually possible to break the triangle down into its component vectors (e. and alignment. Joanne explicitly registers the difference between her follow and his lead via the word too. analysis of stance. or merely different facets. Du Bois (56) Stance Positions/ Stance # Speaker Subject Evaluates Object Aligns 1 KEN. positioning. presupposes shared orientations. In the meantime. in cases where it may not be obvious that the full stance triangle is in play.

Discussion The conception of stance we need is one capable of being situated within a larg- er mediating framework for linguistic action and interpretation. 2002b) the stance triangle has come to be used in a diverse and informative body of work on stance in in- teraction (e. but it must not be forgotten that there is a larger interpretive apparatus that subtends this geometric metaphor. through which social actors simultaneously evaluate objects. Haddington. The framework I have proposed is encapsulated in a triangle. To come to terms with the complexity of the stance act and its interpretive frame. or tri-act. The evidence we have seen suggests that they are not so much alternatives as different facets of a more general phenomenon of stance. 2005. Key to this configuration is a set of three entities (first subject.g. The stance triangle has proved well suited to framing effective research questions. alignment). which I have elabo- rated in some detail in this paper. Keisanen. and align with other subjects. as a single unified act encompassing three subsidiary acts – in effect. gesture. without imposing predetermined limitations on what the researcher can undertake to investigate about stance. 2006. in complementary distribution. I have argued for a par- ticular configuration of actors and actions as the defining feature of stance. I define stance as a public act by a social actor. and shared responsibility between conversational co-participants. we need a way to represent how stance works. 7. To do this. and about how the multiplex meanings and consequences of stance play out in the public space of interaction. a triune act. This is what the stance triangle tries to represent. interactional collaboration. and other symbolic forms). We need this if we want to be able to frame questions about how discourse participants achieve stance. stance object) and a set of three actions (evaluation. see also the papers by Kärkkäinen. The analy- sis of stance in terms of these elements lays the basic foundations on which the . positioning. The stance triangle 169 they regularly occur separately. with respect to any salient dimension of value in the sociocultural field. achieved dialogical- ly through overt communicative means (language. Since its first introduction (Du Bois 2002a. position subjects (themselves and others). or whether they rather co-occur as different aspects of a single act of stancetaking. in its fundamental structure. Kärkkäinen 2006.. Haddington 2004. and Rauniomaa in this volume). I have argued that stance can be analyzed. second subject. it is necessary to articulate a systematic approach to understanding what stance is. Takanashi 2004. which is itself grounded in the dialogic dimensions of sociocognitive relations. The stance triangle is built on certain ba- sic assumptions about what is needed to constitute a stance.

as the polarity of co-action cycles rapidly via the dynamics of dialogic exchange. based on an analogy which is partly structural and partly functional. this tri- angle is meant to be used. positioning subjects (subjective). The stance triangle provides a framework for understanding the sociocognitive relations (objective.28 In the world of theory-making. and tries to clarify how these relations are constituted through the stance acts of evaluating objects (objective). of first subject and second subject. Crucially. The parallelism of two of the three stance vectors allows us to analyze the phenomenon of alignment. by . Especially in cases involving what I have called the stance differential. constituted. and map one subject-object vector to another to constitute the intersubjective relation.170 John W. Most triangle diagrams try to maximize the contrast between three terms selected to display the most important and sharply differentiated abstract concepts in the theory. subjective. But there is more to the stance triangle than this. Unlike many triangles which remain in the realm of the theoretical. the stance triangle specifies the effective relationships between entities and acts through vectors of stance action. The idea is that its general architecture underlies the ac- tual practices of realizing stances and negotiating their significance in particular events of language use. Appropriately deployed. Here. two of the three points of the triangle are used to represent what amounts to the same thing twice – two stance subjects. sociocognitive relations link the subject and object of stance relations. On another level. the triangle provides leverage to analyze the fine calibration of convergent and divergent alignment in the stancetakers’ posi- tioning of themselves relative to others. Du Bois stance triangle is built. This parallelism may in turn invite analogical inferences (Itkonen 2005. Schutz 1962) of dialogic stance partners who alternately fill the roles of speaker and hear- er. What consequences flow from the triangular model of stance? Here I will point to just a few of the implications that arise. there are many triangles. Jakobson 1966. In addition. Silverstein 1984). This is not done to flout the principle of economy. the stance triangle can clarify the array of entities and sociocognitive relations that are activated. this makes it relevant to describing what is happening in actual instances of stance in interaction. and aligning with other subjects (intersubjective). and brought into relation to each other by a particular stance action. but because it displays the interchangeability of perspectives (Mead 1934. on the other hand. The model shows how stance can be analyzed in terms of a set of triangular relations which link entities via vectors of dialogic co-action and intersubjectivity. the stance triangle is available to participants as a resource for organizing their evaluative actions on any specific occasion. intersubjective) that are present in all dialogic interaction. Why one more? But this is not your average triangle. To the extent that it articulates a predictable framework for action. Looked at from the analyst’s per- spective.

Stance must be intelligible within that framework. Stance is best understood in terms of the general structure of the evaluative. The expectation of a well-defined framework for stance remains in play even if the framework must in part be dynamically constituted by the participants themselves in the very act of taking a stance. and other forms of symbolic action – to arrive at a dialogic achievement of stance in the public arena. but while the stance triangle as a theoretical object is general by design. To realize stance dialogically means to invoke a shared framework for co-action with others. Stance is an activity built for two (or more). Par- ticipants use their knowledge of the elements. Taking a stance cannot be reduced to a matter of private opinion or attitude. and aligning processes that organize the enactment of stance. it is also intended to frame the concrete analysis of specific stances. invoking relevant components of the stance frame as we both shape and respond to the multiplex consequences which flow from our actions. As we maneuver within the constantly shifting field of stances. as it comes into existence in its natural environment of dialogic in- teraction. the stance triangle provides a way to apprehend the regularity of stance inferencing that brings systematicity to the consequences of stance. including principles of dialogic . gesture. The stance triangle 171 allowing us to triangulate between one participant’s observable (if partial) stance actions and those of their dialogic partner. actions. it unfolds within a recognized framework for interpreting action. Our exceptional agility at managing the dialogic play of stance and counterstance is underlain by an implicit awareness of the structure of the activity system that frames and enables the achievement of stance. rather than as a catalog of the contents of stance. This is what the stance triangle tries to achieve: a general level of analysis that can be applied in principle to any instance of stance. but something you do – something you take.29 Stance always combines elements of general- ity and specificity. Using the language of Wittgenstein (1953) we might say: There are no private stances. or even – as important as these may be – of the sociocultural value categories that are referenced by stance. Through joint and sev- eral acts we engage in the activity of stance. positioning. Stance can be imagined as a kind of language game in Wittgenstein’s sense. The general principles governing the stance framework. not a property of interior psyche. and vectors of stance – as described here by the stance triangle – to project the multiplex consequences of their own and their partners’ unfolding stances. Conclusions Stance is not something you have. we find that even our own stance must be enacted collaboratively. We deploy overt communicative means – speech. 8. which is to say.

we can go on to draw detailed inferences regarding aspects of the stance differ- ential between the two participants. But beyond this. as these are activated across the paired stance utterances. constitutes an ongoing challenge for any theory of stance. This will be fruitful as long as we remember that dialogic processes must be real for participants if they are to found the claims about stance which are implicit in the stance triangle. including convergences and divergences of epistemology. as resonance across stances shapes the sociocognitive alignment between speakers. The tension between generality and particularity. and thus helps frame intersubjectivity. Stance utterances gain added levels of significance through their juxtaposition with other stance utterances. The diagraph representation depicts the precise resonance relations between two stance utter- ances. Each stance is already specific with respect to. any stance realization is capable of extending its particularity further by entering into unique relationships with other stance realizations. This sets up dialogic juxtapositions of stances which may foreground the structural analogy between pairs of successive stances. The potential is extremely rich for analytical and theoretical cross-fertilization between the meth- ods of the stance triangle and the diagraph. Du Bois co-action and the intersubjective organization of sociocognitive relations. The stance triangle – including its component vec- tors – provides a general account of the framing processes which adapt to both the particularities of the individual stance act and its unique configuration of dia- logic-sequential development. for example. while remain- ing constant and general. within particular con- figurations of dialogic exchange. the objects it evaluates. Part of the meaning of any stance derives from the manner of its realization in interaction. including the dialogic dimensions of the form it takes. Yet the resources available to participants for interpreting such a diversity of stances are sufficient to the task. From a dialogic perspective. is necessarily reflected in the interpretation and realization of stance. including the present one. Given an explicit set of mapping relations between counterpart elements across paired stance utterances. Much of the dialogic quality of stance comes from the way a present stance may resonate with a prior stance. allowing us to probe in finer detail the specific reciprocal relations between sub-portions of the utterances mapped in the diagraph. impinge on the realization and interpretation of individual stances. and specifically the diagraph. . the participants it indexes. yielding a combinatorial explosion of resonance relations. or simply to embrace it. To resolve this tension. and the dimensions of sociocultural value it invokes. affect. The method used here to analyze resonance across successive stance utteranc- es is dialogic syntax (Du Bois 2001).172 John W. characteristic of all culturally framed social practice. no stance stands alone. and other regular concerns of stancetakers.

the smallest unit of so- cial action. comes with ownership. We are left with a view of the stance act as. perhaps. where it inherits the peculiar contingencies of all dialogic action. Still. who they stand with. a way of conceiving and analyzing the nature of linguistic and social action? What is the ultimate import of stance? The short answer: Stance is an act of evaluation owned by a social actor. we might wish for something simpler – a take-home message that we can reference to keep a clear focus on the central meaning of stance. an identity. We make it our business to know where the other players stand. Through its open-ended social circulation. And we care about the state of the game. This charac- terization. presupposed systems of sociocultural value. unites three key aspects of social life: act. both influencing and being influenced by the co-actions of others. a history. Responsibility. The act of stance is performed in the public space of dialogic interaction. In the dialogic shop of stances. Despite its apparent dependency as a local symbolic action delimited by the uptake of other individuals. Such a fundamental status will be justified if stance can be shown to . Because stance responsibility is of interest to all within the stancetaking community of discourse. there’s a rule: If you take it. Responsibil- ity for the stance act is serious business. Stance always invokes. you own it. As players in the language-game of stance. in what condition the players leave the turf – and what all of this implies for the environs of sociocultural value in which we all must live. What is it that sets stance apart as a research agenda. stance may be among the most broadly consequential of social actions in its cumulative effects – an act whose force can be measured by its ef- fective penetration into virtually all domains of sociocultural life. the question of who took which stance is perennially salient. the second element in this capsule cameo of stance. There is no avoiding the rapid ramification of meaning and consequence as stances emerge from the flux of dialogic interac- tion. In sum. and where they’re headed. Social actors are accountable for how they manage and indeed reshape the systems of social value on which we all depend. responsibility. who plays it well and fairly. while by no means complete. we’ve all got some skin in the game. Third. The stance triangle 173 Stance is undeniably complex. is remembered over time. stance (and meta-stance narratives) can only expand its role in the broader calculus of social meaning. ownership of stance is the glue that binds the stance act together with actor responsibility and sociocultural value. too: how it is played. so that all is linked to a social actor with a name. explicitly or implicitly. while at the same time contributing to the enactment and reproduction of those sys- tems. value is what stance is all about – literally. with potentially profound consequences for the relationships of stance actors with their dialogically co-responsible part- ners and with their expanding networks of social relations along wider horizons of time and space. and counts as negotiable coin in the currency of reported discourse. and value.

For full transcription details. Adam Jaworski. for which I am grateful. Pentti Haddington. Volkswagen Foundation conference on Rheto- ric Culture. Patricia Clancy. Sweden (2004). it must come from striving to represent at once the unity.174 John W. Susan Hunston. Cornelia Ilie. For example. Note that some of the transcriptions have been simplified for the sake of clarity. This is what the stance triangle aims to represent: the minimum structure of stance as dialogic action. early versions of my ideas about stance were presented in a series of colloquia during this period at the Linguistics Department and the group on Language. Oulu. 3. LangNet Symposium. In depicting the co-participants’ joint evaluative ori- entation to a shared stance object. square brackets for overlapping speech have been omitted when the cited example doesn’t include the other half of the pair of overlapping utterances. Mainz (2002). Joanne Scheibman. In addition. Michael Silverstein. Berkeley (2003). I am especially thankful for comments on earlier versions of this work from Mira Ariel. a lecture series on stance and dialogic syntax. Du Bois bind together the minimum structures necessary to attain the force of social ac- tion. rather than to simply demand mute agreement to the normative assumptions of social reality as the conventional prerequisite to communication. I have benefited greatly from discussion with all these audiences. My re- search on ‘too’ and ‘either’ in the Santa Barbara Corpus was presented at the Rice Symposium and in fuller form at the International Computer Archive of Modern and Medieval English (ICAME). of stance as it emerges in dialogic interaction. Oulu (2003). Södertörn. Robert Englebretson. Finland (2003). Geoffrey Raymond. finally. including at the meetings of the American An- thropological Association. But the evidence from stance in in- teraction is clear: convergence and divergence of evaluative alignment are equally at home in the dialogic engagement of co-participants. Kroeber Anthropological Society. and the ambivalence. Gene Lerner. My understanding of stance as presented here has developed during the course of a series of presentations at conferences and symposia. If the stance triangle is to have analytic value in the end. Fulbright lecture. John Haviland. as described in the following note. Santa Barbara. Rachel Giora. Most of the examples in this paper are taken from conversations in the Santa Barbara Cor- pus of Spoken American English (Parts 1 and 2). John Lucy. Interaction. Mary Bucholtz. Robin Shoaps. and an anonymous reviewer for the present volume. The source citation gives the title of the dis- . Hiroko Takanashi. the stance triangle proposes a framework for understanding the dialogic realization of intersubjectivity in a way that is capable of embracing both convergence and contestation. on Stancetaking in Discourse. To be sure. New Orleans (2002). it may seem coun- terintuitive to locate contestation within intersubjectivity. See the Appendix for transcription conventions. at Rice University. William Hanks. the original sources may be consulted. Elise Kärkkäinen. and. Houston (2004). Amy Kyratzis. Mirka Rauniomaa. Sandra Thompson. and Social Organization (LISO) at the University of California. 2. Tiina Keisanen. Johannes Gutenberg University. University of Verona (Du Bois 2004). Södertörn University. culminating in a presentation on “The Intersubjectivity of Interaction” at the 10th Rice Symposium. Notes 1.

Alternatively.linguistics. recorded under my direction by researchers from UC Santa Barbara. 2000. 6. or on the internet. 5. nevertheless. The final portion of the citation for each excerpt consists of two numbers.. a salient epistemic dimension as well.. With this information it is possible for interested readers to listen to the appropriate portion of the relevant audio file (e. The following is a key to the original labels in the LSAC corpus for the speakers cited here: in LSAC 1396-01. The challenge of precisely and perspicuously describing stances that emerge in dialogic interaction is a serious one. 2003). by the word that in I agree with that. This is because to paraphrase is to attempt to give a monologic approximation of a stance that may have taken two people to create dialogically in the first place. A summary paraphrase cannot do justice to the specific quality of the actual stance as it emerges dialogically in discourse. The names of speakers are arbitrarily assigned pseudonyms.edu/research/sbcorpus. The stance triangle 175 course in italics (e. The content of Kim’s stance is constituted as no more or no less than any other school. 8. Further information on contents of and access to the Santa Barbara Corpus of Spoken American English is available at http://www.WAV) by accessing the Santa Bar- bara Corpus as published on CD and DVD (Du Bois et al. the general point about alignment should be sufficiently clear. Names used for speakers are pseudonyms. 10. In explicating the interpretation of a stance in its dialogic context. Lambada).html. KIM=<1564> and PAT=<1565>. Conversations in the LSAC were recorded all over the United States using expertise and methodology developed for the Santa Barbara Corpus of Spoken American English and corpus design concepts developed by Longman in their work helping to build the British National Corpus. The source of this example. as denoted. in addition to its affective di- mension. so the framing phrase is interpreted as outside the scope of Pat’s agreeing move. which pertains narrowly to Kim’s framing of the proposition. cited as LSAC. . for example. representing respectively the start time and end time in seconds. 7. How this works in detail on a structural level is a rich topic for further analysis. but unfortunately this cannot be clarified without expending a lot more time and space than is warranted for the present example. 4. followed by the identifying number of the discourse in the Santa Barbara Corpus of Spoken American English (e. I have sometimes ven- tured a paraphrase of the “content” of the stance as it has emerged from the successive con- tributions of several participants. 11. 9. CORA=<1828> and LESLIE=<1829>. SBC002). This is a parenthetical framing of self-positioning. is the Longman Spoken American Corpus. not part of the stance proposition as such. The problem is that the paraphrase usually comes across as too explicit or too literal in character.g. leaving the previous speaker’s metalinguistic framing or self-positioning out of the calculus of stance alignment. agreement can be articulated in relation to the proposition expressed by a co- participant – the stance content of a prior utterance.) This is to be distinguished from the final portion of the utterance (is the way I see it).ucsb.g. which deserves more attention than we can give it here.g. Du Bois et al. (This phrasing is admittedly rather opaque in itself. in LSAC 1296-02.. To be sure. The LSAC was developed under con- tract to Longman publishers. SBC002. The LSAC is a five million word corpus of spoken American English conversation. Alignment markers like the verb agree typically pick out just the stance content or stance proposition. Melissa’s use of the modal should introduces a subtle shift which slyly mitigates the directive force of Jan’s original unmitigated imperative. The predicate amazed can be understood as incorporating.

More precisely. To- masello et al. While the complex case of multivocality is very interesting to consider from the perspective of the stance triangle. This number represents the intonation unit number (or line number). which in turn endorses Jeff ’s own setting of the question in line 1. Jeff ’s affective stance utterance in line 4 builds off of Jill’s prior yes in line 2. Where information about the intonation unit is unavailable. see Haddington 2005. we show only the head of the arrow for the self-positioning vector. see Heri- tage (2002) and Heritage and Raymond (2005). But these issues are largely orthogonal to the present discus- sion. the phenomenon of other-questioning is surprisingly commonplace. 13. the timing of pauses. This can be considered an instance of other-positioning (line 1). and what could be considered a lengthy delay on Lance’s part in coming in with I’m glad. This is not to suggest that an epistemic paraphrase tells the whole story of Dan’s I don’t know. intonation units represent the most salient and productive unit for dialogic mapping in conversation. 19. who don’t know if she’d do it). this must await a separate treatment. There are additional interesting issues here having to do with the incremental realization of Randy’s turn. 2005). as it reflects back on the stance subject. This reflection can be seen as being triggered by a sort of “blowback” from the . in the philosophical sense (Searle 1983). it is intonation units (Chafe 1993. e. 18. In other words. and Haddington this volume. 16. Given that this often seems to provide the best analysis for. Du Bois et al. Note that the self-positioning act defines a vector which emanates from a subject and re- flects back on that same subject. as well as from char- acterization of the multivocality of certain kinds of utterances (Bakhtin [1934] 1981. In most of the examples presented in this paper. How this stance develops over time and across speakers is of some interest. with varying degrees of impositive force. To explicitly represent both source and target of the reflexive vector of self-positioning would involve a circular arrow originating in the subject and reflect- ing back on itself. Rather than display such an arrow in this simple diagram of the stance trian- gle. see note 16) that Jennifer presents him with. counting from the beginning of the published transcription. they are intentional predicates. 1993) that generally define the rows (or “strands”) of a diagraph. which may have as much (or more) to do with an act of demurral that hedges and blurs his response to (and responsibility for) the candidate stance (other-positioning. I have developed the concept of other-positioning to account for cases in which the first subject (speaking subject) proposes a candidate stance for the second subject (addressee). 14. but not you.g.) 17. ordinary questions. 21. Goffman 1981. the most viable alternative for identifying the rows of a diagraph would generally be the clause. the stance subject can be taken to be more or less equivalent to the speaking subject. (For application of this concept to questioning by interviewers in television news formats. Based on my research (Du Bois 2001). But that’s another story. Agha 2005). One can try to imagine special circumstances in which the pronoun I is given heavy con- trastive stress (suggesting that it is only I. see discussion in conversation analysis of the interactional negotiation of turn status as epistemically “first” (=lead) or “second” (=follow) (Heritage and Raymond 2005).176 John W. but this would be quite unusual. 15. Du Bois 12. as is clear from the sophisticated analyses of the sujet de l’énonciation by the cited authors. 20. cf. My terms here are modeled on the notion of gaze follow(ing) (Tomasello 1999: 62–67. 22. For a rich treatment of agreement from the perspective of conversation analysis. For a related notion. which is parallel to the phenomenon of stance follow in interesting ways. But things are not always so simple.

know. Bakhtin. The stance diagraph incorporates aspects of the stance diagram (see discussion of examples (30)–(31)). But that is a long story. but a fuller analysis of these events must be reserved for another occasion. E. Bakhtin.) 23. stance object) and functions (evaluation. 26. Meek (trans). as social actors position themselves and evaluate entities with respect to specific values along any socially salient dimension of the sociocultural field. positioning.” Journal of Linguistic Anthropology 15: 38–59. enregisterment. and thus is not represented in Figure 1. E. Benveniste. How stancetaking processes both invoke and construct the associated systems of sociocultural value is a critical issue which we can only point to in this paper. 27. Coral Gables.M. Conversely. but one which represents a prime topic for further research. While too and either should be recognized as powerful diagnostics for intersubjectivity when their conditions of use are applicable. alignment). While it is beyond the scope of this article to lay out all the conditions governing the use of too and either. The stance triangle 177 subject’s act of evaluating the stance object. both words can be used to mark referential semantic relations as well (although in spoken discourse this objective function is much rarer than the (inter-)subjective function presented here (Du Bois 2004)). While the inclusion of the shared stance object – seemingly a specialized property of cer- tain kinds of stance exchanges – might seem to limit the general applicability of the stance triangle. there are many intersubjective contexts in which these particular indexical forms do not appear.” In Problems in General Linguistics. such as the labeling of stance roles (stance subject. 223–230. The passage continues with several further developments regarding this stance negotiation. : Yeah?). The stance triangle bears an important relation to the systems of sociocultural value that stances invoke and reproduce. “Subjectivity in language. including diagraph analysis.M. 2005. from a dialogic perspective it can be argued that a shared orientation to a stance ob- ject is a general property. and want (Du Bois 2004). such as the inclu- sion of multiple lines representing dialogically resonating utterances. [1934]1981. 24. . References Agha. and must be reserved for another occasion. 28. not only of stance acts but of the use of language in general (see the discussion at the end of Section 6). combined with aspects of the diagraph. 25. footing. 1971. While the dialogic resonance in lines 5 and 7 between Lenore and Joanne is interesting in that the intersubjective stance differential is subtly realized through an audible contrast located primarily in the intonational difference (Yeah. it will not be diagrammed here for rea- sons of space. 29. think. Aside from their high frequency use in marking intersubjective pragmatic relations. FL: University of Miami Press. “Voice. TX: University of Texas Press. one key factor can be mentioned here: the use of subjective intentional stance predicates such as like. the analysis of intersubjectivity in language must draw on a wide variety of additional tools. The Dialogic Imagination: Four Essays by M. (Other-positioning requires a significantly more elaborate notation. with vertical alignment of elements to iconically display which are resonating with which. M. A. love. Austin. This is not to say that too and either always function as diagnostics for intersubjectivity.

University of California. N. W.178 John W.).). S. Part 2.. 2006. Schuetze-Coburn. Department of Linguistics. 1973. “Prosodic and functional units of language. Du Bois. “Stance and consequence. Santa Barbara. Cambridge: Cambridge University Press. J. R. Possible Worlds. O. Oxford: Oxford University Press. Dire et ne pas dire.W. W. Ed- wards and M. C. Berman. 1999.W. J. Hunston and G. Clift. 45–89. Du Bois. Ducrot.” In Talking Data: Transcription and Coding in Discourse Research. S.” Jo- hannes Gutenberg University.). July 4–7.. 1990. Ducrot.A. 161–181. W. B. 1991.D.L. Bilger. Bruner. “Outline of discourse transcription.. Le français parlé: Études grammaticales. 2005. Besnier.W.” In Talking Data: Transcription and Coding in Discourse Research. Discourse. “Dialogic rhetoric: Stance and intersubjectivity in the syntax of engage- ment.H.L. 2000.” In Responsibility and Evi- dence in Oral Discourse. H. Chafe. Irvine (eds. 1991.. Philadelphia: Linguistic Data Consortium. Cumming.” Pragmatics 1: 71–106. “Introduction: Developing discourse stance in different text types and lan- guages. Le dire et le dit. “Transcription design principles for spoken discourse research. S.” Journal of So- ciolinguistics 10: 569–595. 56–73.” Paper presented at Volkswagen Foundation conference on “Rhetoric Culture. 1993.A. S. “Indexing stance: Reported speech as an interactional evidential.A. R. J. New York: Basic Books. J. and Strömqvist. Du Bois.W.. 2002b. 2002. Mainz. and Martey. Cambridge. Thompson. Blanche-Benveniste. M.” Written Language and Literacy 5: 255–290. Du Bois.D. K. Davies. 2001. The Interpretation of Cultures. 2000. J. “Positioning: Conversation and the production of selves. NJ: Lawrence Erlbaum. Pour une linguistique de l’énonciation. Hill and J. November 20–24. Ragnarsdóttir. Santa Barbara Cor- pus of Spoken American English..” Journal for the Theory of Social Behavior 20: 43–63. Paris: Hermann. C. New Orleans. and Thompson. Du Bois. Hillsdale. J. S.A. “Searching for intersubjectivity: ‘Too’ and ‘either’ in stance alignment. O.L. C. Conrad. and Paolino. Consciousness. “The socialization of affect in Japanese mother-child conversation.. 2002a.W.M. Part 1. CD.A. LA. J. D. Du Bois. 1994. Clancy.” In Evalua- tion in Text: Authorial Stance and the Construction of Discourse. Chafe.. 1986. Du Bois. Geertz. W. J. Chafe. Lampert (eds. Meyer. D. Du Bois. 1993.L.. “Adverbial marking of stance in speech and writing. “Reported speech and affect on Nukulaelae atoll. J. Du Bois Berman. J.” Journal of Pragmatics 37: 105–124.). and Biber. 1972... Actual Minds. University of Verona. 1990. CD. and van den Eynde. Thomp- son (eds. “Discourse stance. J. Culioli. Lampert (eds. R. Paris: Ophrys. S. S. NJ: Erlbaum. Paris: Éditions du Centre National de la Recherche Scientifique.” Pa- per presented at 25th Conference of the International Computer Archive of Modern and Medieval English (ICAME). Paris: Minuit. MA: Harvard University Press. Chafe. P. Santa Barbara Corpus of Spo- ken American English. Hillsdale.W. Meyer. Ms. 1984. C. 1993. 2003.A. 2004.” Paper presented at Annual Meetings of the American Anthropological Association. J. Towards a Dialogic Syntax. 33–43. Chicago: University of Chicago Press.W.W. Edwards and M. Philadelphia: Linguistic Data Consortium. R.T. and Time: The Flow and Displacement of Conscious Experience in Speaking and Writing. . and Harré. N. A. Rouget.” Journal of Pragmatics 31: 1397–1421. C.

147–189.T. and Sinclair.H. B. G. G. Johnstone. Hunston. Hill. 271–323. “Assessments and the construction of context.” In On Language: Roman Jakobson. Goodwin (eds.). J. R. Thompson (eds. Goodwin (eds. Duranti and C. E.” Text and Talk 26: 281–328. 87–97. Haddington. 2006. “The organization of gaze and assessments as resources for stance taking. R. C. Duranti and C. J. Hillsdale. 2005. status. B. 386–392. J.). 1993. 1981. 196–224. MA: Harvard Univer- sity Press. Autism and the development of mind. J. 1996. “‘Sure. Cambridge: Cambridge University Press. 2000. G. Waugh and M.). [1957]1990. P. Harris. and Goodwin. 1993. W. Ford. and exclusion. 2005.A. “Glossary of transcript symbols with an introduction. 1994. Levinson (eds. University of Oulu: Oulu University Press. Analogy as Structure and Process. “The terms of agreement: Indexing epistemic authority and subordination in talk-in-interaction. Repetition in Discourse: Interdisciplinary Perspectives. disser- tation. Amsterdam: John Benjamins.P. Jefferson.J. E.” In The Language of Turn and Sequence. C. 13–31. A. 1990. 2006. P. Monville-Burston (eds. “Stance taking in news interviews. Haddington.” Social Psychology Quarterly 68: 15–38. J. sure’: Evidence and affect.A. Heritage. “Contextualization and understanding. 74–101.). The Hague: Mouton de Gruyter.” Language 42: 398–429. The hidden life of girls: Games of stance. Haddington. “Footing.” In Rethinking Linguistic Rela- tivity.). “Grammatical parallelism and its Russian facet. and Raymond. Gumperz.” In Conversation Analysis: Studies from the First Generation. Haviland. Oxford: Oxford University Press. Thompson (eds. Evaluation in Text: Authorial Stance and the Construction of Discourse. 2004. 1952. 1992. Amsterdam: John Benjamins. A. S. 229–252.H.B. and Thompson. Heritage. Ph. R. L. P. J. and relativity.).B. Jakobson. Referential Practice: Language and Lived Space Among the Maya. English Department.” In Rethinking Context: Language as an Interactive Phenomenon. Hunston and G. M. NJ: Erlbaum. M. Vol.C.M. S.). Gumperz and S. NJ: Ablex.S. 1966. “A local grammar of evaluation. G.” In Re- thinking Context: Language as an Interactive Phenomenon. 2002. Z. Lerner (ed.E. “Oh-prefaced responses to assessments: A method of modifying agree- ment/disagreement. 2004.” SKY Journal of Linguistics 17: 101– 142.” Language 28: 1–30. 2000. Goodwin.” In Selected Writings. transpositions. Philadelphia: University of Pennsylvania Press. Hobson. Cambridge. and Irvine. J.H.). (ed. Norwood. Hunston. “Projections. Haviland. Jakobson. Goodwin. Evidence and Responsibility in Oral Discourse. (eds. The stance triangle 179 Goffman. The Intersubjectivity of Stance Taking in Talk-in-Interaction. Jakobson. 1991. Chicago: University of Chicago Press. Hanks.J. 3: Poetry of Grammar and Grammar of Poetry.H. “Discourse analysis.). Fox and S. R. Cambridge: Cambridge University Press. 1981. 1992. Cambridge: Cambridge University Press. Oxford: Oxford University Press.” In Evaluation in Text: Authorial Stance and the Construction of Discourse. S.F. J.R. Itkonen. Oxford: Oxford University Press.” In Forms of Talk.D. Cambridge: Cambridge University Press. “Shifters and verbal categories.” Text 9: 27–68. Oxford: Blackwell. J. . 2005. “Poetry of grammar and grammar of poetry.

1987. 1996. 142–175. Du Bois Kärkkäinen. 1997. Hunston and G. Kidwell. 3. CA: Stanford University Press. (ed.L. 12–44. Lemke. D.). 1984. New York: Addison Wesley Longman. Atkinson and J. 1988. A.” In The Construction of Professional Discourse.” Text & Talk 26: 699–731. M. Oxford: Oxford University Press. Self. B. Cambridge: Cambridge University Press. “Agreeing and disagreeing with assessments: Some features found in pre- ferred/dispreferred turn shapes. with a Focus on ‘I think’. J. 2006. Amsterdam: John Benjamins. 47–65.K. Kärkkäinen. 1993. Vol. Negotiating heteroglossia: Social perspectives on evaluation (Special Issue). Ono and H.180 John W. “Narrative analysis: Oral versions of personal experience.).” Functions of Language 5: 33–56. 407–437.” In Essays on the Verbal and Visual Arts. Text 23. Tao (eds. Labov. Stanford. and Zimmerman. Emotion. Image. Lucy. Hartshorne and P. Hillsdale. Culture and Language Development: Language Acquisition and Language Social- ization in a Samoan Village. “Resources for attitudinal meaning: Evaluative orientations in text semantics. P. MA: Harvard University Press. J. Linell and B. M. G.C. E. (eds. Moore. “‘Is she vicious or dense?’: Dialogic practices of stance taking in conver- sation. E.).” Journal of Linguistic Anthropology 14: 127–150. P. 1: Theoretical Prerequisites. E. Santa Barbara: University of California.. Cambridge: Cambridge University Press. J. 1993. J. J. Maynard.A. S. J. Ochs. “Three kinds of signs.).). Seattle: University of Washing- ton Press. Vol. Thompson (eds.). 2003. Interaction. C.W. Amsterdam: John Benjamins. and Voice in The Japanese Lan- guage. and Contexts in Dialogical Perspectives. Macken-Horarik. Ochs. “Stance and subjectivity. and Symbol: The Cognitive Basis of Grammar. 1995. E. W. E. Langacker.” In Rethinking Linguistic Relativ- ity.R. Linell. Martin. Cambridge: Cambridge University Press. 2004. T.H. R. 2000. C. Kockelman.” Communication Monographs 73: 1–28. 57–101. 2003b. C. Gunnarsson. Helm (ed. Mead. 1998. and Dunham. Peirce. Reflexive Language: Reported Speech and Metapragmatics. J. 1967. Joint attention: Its origins and role in development.). 1998. S. and Waletzky. Mind. Epistemic Stance in English Conversation: A Description of its Interac- tional Functions. T. Discourse Modality: Subjectivity. Weiss (eds. Kärkkäinen. Berlin: Mouton de Gruyter. Concept. “Linguistic resources for socializing humanity. 2003a.R.). 1991.S.).-L. “Stance taking in conversation: From subjectivity to intersubjectivity. Nordberg (eds.” In Structures of Social Action. Nakayama. R.H. 1934. 151–172. Approaching Dialogue: Talk.” In Evaluation in Text: Authorial Stance and the Construction of Discourse. Santa Barbara. “‘Observability’ in the interactions of very young children. Pomerantz. Linde. Foundations of Cognitive Grammar.W.M.J. J. “Beyond exchange: APPRAISAL systems in English.” In Santa Barbara Papers in Linguistics 12: Recent Studies in Empirical Approaches to Language. “Evaluation as linguistic structure and social practice. C.J. Langacker. Cambridge. [1885]1933.” In Collected Papers of Charles Sanders Peirce. and Martin. Levinson (eds. P. NJ: Erlbaum. P. Chicago: University of Chicago Press. Amsterdam: John Benjamins. . 2006. Cambridge: Cambridge University Press. Heri- tage (eds. and Society. Gumperz and S.

repetition. K. Thompson (eds. 1992.” In Meaning. “The limits of awareness. “The pragmatics of what is said.” American Journal of Sociology 102: 161–216. Oxford: Black- well. MA: Harvard Uni- versity Press. Sperber. M.D. F. 1976. Washington. Call. S. Vol. M.” In Meaning in An- thropology. J. 1–27. 2002. J. and Wilson. V. Sacks. Amsterdam: John Benjamins.H.. R. dissertation. Cambridge: Cambridge University Press.C. Intentionality: An Essay in the Philosophy of Mind. Thompson. Carpenter.” Journal of Linguistic Anthropology 12: 34–71. 2002. Tannen. The Interactional Co-Construction of Play in Japanese Conversation. Heine (eds.A. A. J. Oxford: Oxford University Press.L.). “Confirming allusions: Toward an empirical account of action. R. M. Schegloff. Marxism and the Philosophy of Language. and Hunston. “Shifters.. S. 2: Focus on Types of Grammatical Markers. Basso and H.). 2000.. D. E. G. The Hague: Nijhoff. Tomasello. 2001. D.).R. “On the rise of epistemic meanings in English: An example of subjectifica- tion in semantic change. Oxford: Blackwell. 2001. Tomasello.” In Approaches to Grammaticalization.A.” In Frequency and the Emergence of Linguistic Structure. 1991. H. Form.A.). Cambridge. 2004. Takanashi. S. “Evaluation: An introduction. Shoaps. P. Point of View and Grammar: Structural Patterns of Subjectivity in American English Conversation. Linguis- tics Department. Morality in Grammar and Discourse: Stance-taking and the Negotiation of Moral Personhood in Sakapultek (Mayan) Wedding Counsels. H. Selby (eds. T. Duranti (ed. Behne. 1984. Shoaps. 1989. Lectures on Conversation. Scheibman. 2005. S. Linguistics Department. Ph. Schiffrin (ed. Santa Barbara. M. and argument structure: Evidence from conversation. and Use in Con- text: Linguistic Applications. [1929]1973. 382–401. Hopper (eds. . The stance triangle 181 Recanati. dissertation. and cohesive structure in the time course of dyadic conversation..” In Evaluation in Text: Authorial Stance and the Construction of Discourse. Voloshinov. Thompson. Silverstein.). 27–60.” Language 65: 31–55. Traugott. Amsterdam: John Benjamins. J. A. University of California. and Hopper. Oxford: Blackwell. Thompson.A. clause structure. M. D. Silverstein. 1987. Relevance: Communication and Cognition. MA: Har- vard University Press. Traugott and B. Santa Barbara.J. “Repetition in conversation: Towards a poetics of talk. DC: Georgetown University Press. The Problem of Social Reality.” Behavioral and Brain Sciences 28: 675–735. “Transitivity.J.N. “On the pragmatic poetry of prose: Parallelism. 1989. D. Ph. Bybee and P.). “A quantitative perspective on the grammaticization of epistemic parentheticals in English. The Cultural Origins of Human Cognition. 1999.C. A. “‘Pray earnestly’: The textual construction of personal involvement in Pen- tecostal prayer and song. E. Hunston and G. “Understanding and sharing intentions: The origins of cultural cognition. and cultural description.” In Linguistic Anthropology: A Reader.A. 313–329. University of California. 11–55. 1995. Albuquerque: School of American Re- search. Cambridge. and Mulac.D. Collected Papers I. Schutz. 1983. E.” Mind and Language 4: 295–328. and Moll. Amsterdam: John Benjamins. M. Silverstein.” Language 63: 574– 605. Searle. H. verbal categories. 1962. 1996. 2004.

182 John W. Thought and Language.ucsb. although there have been a number of significant updates in my more recent transcrip- tion practice. (1992).edu/projects/transcription/representing).2) pause duration in seconds hold (micro-pause) . 1986. less than 150 milliseconds lag (prosodic lengthening) : colon marks slowing of local tempo breath (H) audible inhalation laugh @ one symbol for each pulse of laughter unintelligible ### one symbol per unintelligible syllable uncertain hearing #you’re #kidding transcribed words are uncertain pseudograph ~Jill name change to preserve anonymity . timed (1. (For further details see http://www. Wittgenstein.. 1953. linguistics. The most relevant symbols are given below. L.S. MA: MIT Press. L. semicolon follows name in CAPS simultaneous speech [ ] brackets show overlap start and end pause. Cambridge. Appendix: Transcription conventions The transcription symbols and conventions used in this paper are largely as in Du Bois et al. New York: MacMillan. Du Bois Vygotsky. Philosophical Investigations. Meaning Symbol Comments intonation unit line one new line for each intonation unit speaker/turn attribution JILL.

but worked up in a way that is suitable for what is being done in interaction (see e. Here. In what follows. This is the case with another frequent expression. an epistemic/evidential fragment that is very frequent in American English (but not in British English). see also Rauniomaa this volume for es- sentially similar findings concerning the Finnish minun mielestä ‘in my opinion’ and minusta ‘I think’). while it may also do less routine-like recipient design and face-work at points of trouble in interaction (Kärkkäinen 2003. I draw from evidence provided by recent linguistic work that has established epistemic/ evidential stance marking to be a highly regular and routinized phenomenon in terms of the linguistic forms used: I guess belongs to this group of frequent mark- ers. which in everyday American English has been shown in first-position turns to routinely frame an upcoming stanced turn or longer opinion sequence. of evaluations as not carried around ready-made by participants. I think. this volume) of stance as a social act that is achieved dialogically and publicly. namely I guess. Introduction1 My overall starting-point in this paper is an epistemic stance marker that has been strangely neglected in linguistic research so far. I established that American English speakers use a lim- ited set of high-frequency markers in everyday speech for their expression of epis- temic/evidential stance. I will view I guess. The role of I guess in conversational stancetaking Elise Kärkkäinen University of Oulu 1. in an overall theo- retical framework proposed by Du Bois (2001. 2003.2 It is therefore conceivable that many of these frequent stanced items have specialized into some routine function(s) in the interactional organization of conversation. In Kärkkäinen (2003).. through our engagement with our co-participants. or in second-position turns to project that a different speaker perspective will follow.g. . This view is not unlike the view held in discursive social psychology.

e. sound repetition. It is significant that when we look beyond the initial frame. In this paper. but the following: grammatical and syntactic structures such as choice of pronouns. I argue that I guess. or cases where stance is not explicitly inscribed but seems nevertheless to be evoked (Martin 2003. voice quality. What. as opposed to stand-alone evidentials which inherently index a stance). or a (strong) claim. Ochs and Schieffelin (1989: 22) have alerted us to the possibility that affect (or stance) may permeate the entire linguistic system.’ i. couplets. quantifiers. dialect. hedges. I here mean a fragment of speech that provides a perspective or stance toward the action that is produced in the associated utter- ance (see Section 3. see also Kärkkäinen 2006). markers of epistemic modality and evaluative lexis. see also Clift 2006 for interactional evidentials which are dependent on sequential position to index stance.2). that inherently involves taking a stance or a posi- tion. the rest of the utterance typically also contains explicit stance-indicating or stance-evoking material. and repetition of own/other’s utterances (Ochs and Schieffelin 1989: 12-14.’ ‘com- ing to a realization. it . then. sentential adverbs. Ochs 1992: 412). is the contribution of I guess in the conversational actions and sequences that it appears in. if those actions need not be indexed or framed by I guess to already count as stanced actions? The conversational data at hand suggest that I guess brings in a special meaning of ‘making an inferential discovery. But if we view the action containing/framed by I guess as constituting one relevant action within a longer sequence of stancetaking actions between co-participants. In this paper. Voloshinov [1930]1973. I guess. an opinion. cleft constructions. Students of language have further acknowledged the necessity and difficulty of identifying evoked or implied evalu- ation. a subjective marker par excellence (but see Section 3. augmentatives. tense/aspect. such as evidential markers. I indeed view stance essentially as emerging from dia- logic interaction between interlocutors (Holquist 1990. and sound symbolism. verb voice. The authors argue that the linguistic resources for expressing affective and epistemic stance include. determiners. diminutives. phonological features such as intonation. The speaker thereby frequently produces an action such as an assessment. I draw on conver- sation analysis as a methodological tool to examine in detail the conversational actions that I guess frames.. and how those actions are designed linguistically and prosodically. and discourse structures such as code-switching as instantiated by taboo words. By frame. Indeed. not only the lexicon. and I am primarily interested in the practices and pro- cesses of (the activity of) stancetaking between co-participants.2 for a more detailed definition of frame). this marker displays and makes public a (reasoning) process of the speaker at a particular point in interaction.’ or ‘drawing a conclusion.184 Elise Kärkkäinen Potter 1998). and word or- der. functions as an intersubjective stance frame that organizes the stancetaking activity between conversational co-participants in a surprisingly consistent fashion.

some 50 sequential environments of I guess were analyzed in detail. A frequent resource used thereby for taking stances is the syntactic. If some of the most frequent “traditional” stanced items in a language. and ultimately to the larger activity that the participants may be seen to orient to in a given situation. We clearly need to also look beyond these frequent and routinized stance markers (such as markers of epistemic modality and evi- dentiality. Thus. attention will be paid to the multiple ways that stances are taken in turns-at-talk containing this initial stance frame. Data and transcription My database consists of 54 occurrences of I guess gleaned from 17 different con- versations of about 15–25 minutes each. as proposed by Du Bois (2001.. and in extended turns to insert a stance relevant for the understanding and appreciation of a telling. . rather than themselves indexing an interactionally strong and salient stance on many occasions. I wish to shed some light on the practices and processes of stancetaking beyond individual speakers’ contributions. As we will see below. from Parts I and II of the Santa Barbara Corpus of Spoken American English (Du Bois et al. 2. repetition of own/other’s utterances mentioned above). semantic. across speakers and speaking turns (cf. turn out to essentially only act as stance frames. see also Anward 2000). to the turns-at-talk coming before and after the current turn. in responsive actions for displaying and projecting that the speaker will give up on a stance and/or adopt a new one. but also in prior and subsequent turns. I will therefore expand the scope of examination from the actions and turns containing I guess. my aim in this paper is to go deeper into the more in- teractional resources of stancetaking. and affect markers) that linguists have largely occupied themselves with. Du Bois et al. Potter and Hepburn 2003): in sequence-ini- tiating actions for displaying the discovery of a speaker’s stance. The role of I guess in conversational stancetaking 185 becomes pertinent to view I guess as not only indexing inference or conclusion. Indeed. 2003). such as epistemic I think and I guess.e. evaluative lexis. One ex- ample is taken from a corpus of British and American television news interviews (Haddington 2003). it is essential that we acquire some breadth in our descriptions of the manifestations of stance in naturally-occurring discourse and language use. I guess can be viewed as a discursive practice used for perform- ing a practical task in discourse (cf. In other words. to display and project an upcoming stanced action and organize the stancetaking activity between participants. 2000. i. and prosodic resonance between contributions by different speakers. but in effect organizing the stancetaking activity between the co-participants.

Ford et al. to alert the analyst to the fact that it is frequently possible to assign the pause to some other speaker instead. All TCUs are also directional. . The marking of pauses is generally done in accordance with the conversation ana- lytical methodology: pauses appear on a separate line rather than at the beginning of a certain speaker’s IU. or produced piece-by-piece rather than as one long chunk (Schegloff 1996: 55). of course. (1999). Thus. It is advantageous to transcribe data into IUs. and Scheibman (2001. This finding gains strong support from some other recent studies based on spoken language corpora.1 Frequency of I guess in spoken English Several studies have pointed out that I guess is a very frequent phrase in everyday spoken American English. The way we as analysts transcribe our data. I have modified the Du Bois et al. The very frequency of I guess makes it a worthy object of such study. 3. 3.186 Elise Kärkkäinen The data are transcribed according to the discourse transcription conventions presented by Du Bois et al. because the very visual representation. I know. I guess from a linguistic point of view Recent research in functional and interactional linguistics has acknowledged the need to view certain recurrent lexico-syntactic structures and discourse patterns from a new vantage point. the construction of a turn is frequently incremental. I feel (like). such as I think. personalized stance markers that make reference to the speaker. I guess. Keisanen this vol- ume). In Kärkkäinen (2003). already affects our perception of how speech and interactional actions are produced and emerge through time. even though not quite as frequent as I think or I don’t know. in order to better account for their use in actual ev- eryday and institutional interactions (see e. namely those of Biber et al. However.g. Thompson (2002).. and I found. 2002). available to participants as they unfold. (1993. were by far the most frequent epistemic markers. with the intonation unit (IU) as the basic unit of transcription. 2002. even though most studies to date only deal with it in passing. and not only once they have been produced and completed (Lerner 1996: 307). see the Appendix for a list of symbols used). incremental. immediately alerts us to the emergent. and directional nature of turns-at-talk and the turn-construction units (TCUs) that turns are made of. with each IU lined up on a separate line. and I will here give a brief description of this collocation as a linguistic phenomenon. (1993) convention in one respect.

the statistics provided in the Biber et al. Here. suggest.4 million words of conversational American and British English. I guess was also highly frequent. I confine myself to American English usage. or 100 face-to-face and telephone conversations of 5–20 minutes in length (totaling 205. Thompson (2002: 138) also observes that the most common complement-tak- ing predicates (CTPs) in her data are the following.15 % I guess4 20 tokens or 0.01 % We can see here the same overall order of frequency between the three phrases as in the above American English data sets. (1999) grammar show a dramatic drop in frequency after the epistemic verb guess: the next verbs.608 words I don’t know 309 tokens or 0. For comparison. see. As this search yielded 20 occurrences of I guess. This can be taken as evidence for my claim of the relatively small size of the set of frequent stance markers. find. I conducted a search of the three most common epistemic phrases in the conversational data of the International Corpus of English–Great Britain (ICE-GB) database. the number of I guess per million words in these data would very roughly amount to 100. feel. I think 729 tokens or 0. along with the two other epistemic phrases I think and I don’t know. (1999) database.35 % of 205.5 . the great majority of which occur with first-person subjects as fixed formulas: think/thought 139 know/knew 51 see/saw 17 guess 17 remember 15 In Scheibman’s (2001. The role of I guess in conversational stancetaking 187 Biber et al. the most common verbs controlling that-clauses are the following: think 2.000 per million words3 say 1.250 per million words know 750 per million words guess 500 per million words (in American English only) In fact. and show only occur less than a hundred times per million words of talk. (1999: 667–669) found that in a large database of some 6. however. which is considerably less than the 500 in the Biber et al. 2002) data. believe.608 words). it is not entirely unheard of in everyday British English. But we may conclude that even though this item is much more frequent in American English.

and I guess (see Thompson 2002 for a summary of studies that hold this view).. I guess is commonly regarded as a conventionalized.g. Complement clauses have commonly been regarded as subordinate to the main clause. sedimented pragmatic expression.2 The fragment I guess as a stance frame As becomes obvious from many recent linguistic studies. be interesting). that the actual degree of subjectivity and referentiality of frequent epistemic fragments like the ones above can be quite low.g. (1999: 667–669). Thompson further claims that the most frequent of these phrases become relatively fixed epistemic formulas (e. notably with first-person subjects. tell. We may also ask how much referential meaning is left in the predicate and ultimately in the whole collocation. but the entire collocation takes on pragmatic function with a concomitant reduction in specificity of the pronominal referent. I think) in her data. however.. while Stenström (as cited in Helt 1997) observes of I think. I thought. I (don’t) know. you know. I mean.e.. I remember. Traugott (1995: 38) further claims of the English I think that it is becoming more subjective both in function (toward a fixed phrase indicat- ing speaker’s epistemic attitude) and also in the overwhelming selection of the first-person-subject form.. Scheibman (2001: 70– 71.188 Elise Kärkkäinen 3. and as was mentioned above in Section 3. make sure.e. i.7 It is worth pointing out. and I know/knew – epis- temic/evidential/evaluative fragments. and you see that they fill various slots along a semantic continuum that represents a strong relationship to their literal meanings at one end. 76) also found that there are highly frequent formulaic collocations of first- person-singular subjects and especially verbs of cognition (I guess.e. as adverbials. i. this finding is corroborated in Biber et al. and be- have much like epistemic morphemes in other languages. which may eventually become eroded and leave only a discourse particle (presumably think). I don’t know. I think/I don’t think/I thought/I didn’t think). I mean. Scheibman (2001) claims that there is a weakness or generality of referential- ity in the subject of expressions like you know. Thompson (2002) goes on to argue that we might term the frequent CTPs – I think. while the less frequent ones show more diversity of form (e. rather than main clauses obtaining a com- plement clause. They note. a stance marker. i.. . I guess. however. and I think: the speaker is not specifically referring to himself in these usages. 1991b) claim I think and I guess to have grammaticized into epistemic phrases that act in the same way as adverbs. Schif- frin (1987) argues that some cases of discourse markers like y’know are less ref- erentially meaningful than others (1987: 319). and a minimal relationship to literal meaning at the other. to an epistemic phrase like I think.6 Thompson and Mulac (1991a. In a later study. which frequently precedes complement clauses in English (Scheibman 2002: 32). that it is not clear what the grammatical status of an epistemic phrase is.1.

to be discussed in more depth in Section 5). Thompson (2002: 141) proposes a view of grammar as reusable fragments. to guess the answer). it may have a range of prosodic contours. the assessments. even though it still carries remnants of its referen- tial or propositional meaning in many contexts of use. in Kärkkäinen (2003: 175–179).8 Indeed. being done in the associated utterance.e. everybody’s getting uh. and proposals. that consti- tutes the actual action of the turn. and Kendra has asked for a toothpick) WENDY: . after Kevin was discovered to have lettuce on his tooth. 146) that what has been termed the main clause in much linguistic research simply serves as a stance frame for the clause that it occurs with: such epistemic/evidential/ evaluative frames or fragments provide a certain type of perspective or stance toward the actions. The role of I guess in conversational stancetaking 189 Lenk further points out that the (short) lexical items that may figure as discourse markers. it commonly appears in initial position of an utterance. i. often also have a (separate) propositional meaning (cf. here I guess. KEN: I guess we a=re. Thompson draws our attention to the fact that Ken’s aligning agreement to Wen- dy’s summary of the previous turns is expressed in the complement clause we are. it is syntactically detachable from sentences. claims. The collocation I guess. presented in Thompson (2002: 132): (1) (at a birthday party. i.. as practices of turn construction. This is very much the view that I also adopt here. counterclaims. mostly appears with a prag- matic meaning in my data. Let us then look at the following example. however. i.. and either has no mean- ing or only a vague meaning (see Schiffrin 1987 for criteria for discourse markers.e.. In effect.. and in extended or multi-unit turns. tooth obsessed. to be used as turns or parts of turns. I established the discourse-marker status of I think in everyday spoken English. and in the present paper we will note clear evidence for a similar status of I guess: it may operate at both a local and a global level in discourse. everyone has jokingly commented on it. . I guess will be examined in connection with different kinds of conversational actions done in different sequential environments: in initiating and responsive actions.e. or to other segments within the discourse” (1998: 52).. and that it is not the complement-taking phrase. and in what follows I will examine what exactly the role and function of the frame I guess is in turn construction and more specifically in stancetaking activity between participants. that are used to “signal the sequential and ideational relationship of the two utterances between which they occur. She ends up suggesting (2002: 142.

which in semantic orientation actually expresses points on a continuum from speaker uncertainty to relative certainty (even though it appears in very similar sequential environments.” He further claims (1986: 266) that belief is a mode of “knowing” in which concern for evidence is downgraded. Here. which are followed by second actions or second-pair parts. This. it differs from I think. assertions. the immediately prior turns. I think and I guess have still retained vestiges of their earli- er lexical meaning in the grammaticized form: the two still have a difference in meaning. so that I think is a stronger assertion of belief than I guess. It turns out that I guess is actually used as an evidential rather than an epistemic marker. as will become clearer from many of the examples that I will now turn to.9 4. as- sertions. and questions. or is even present in the actual physical or wider social environment where talk takes place.3 A note on the semantic meaning of I guess Few studies actually deal with the semantics of this stance marker. Such a stance. Practices of stancetaking: Beyond the stance frame I guess 4. and hence less commitment to a proposition than think does” (1991a: 325). and is based on evidence gleaned from a just prior turn or turns. whether this is actually asserted or otherwise inferable (Du Bois this volume. opinions. or from the wider social context. or inferable from. such as (first) assessments. ‘Evaluation’ in turn means charac- . The following examples show that I guess marks actions in which some kind of change in the speaker’s current state of knowledge or aware- ness or orientation has (just) taken place (cf. opinions. see Kärkkäinen 2003).1 I guess in sequence-initiating actions: Discovering a stance I will here examine cases in which I guess frames sequence-initiating actions.190 Elise Kärkkäinen 3. see also Martin 2003 for evoked evaluation). from the actual environment where talk takes place. which is presented in. is “traceable to the difference between think and guess as verbs: guess implies an assertion based on little or no evidence. and answers. notes that I guess is a marker of “belief. as any act of stancetaking. as epistemic phrases. and speakers believe things because other people do or simply because they want to believe them. But on a closer analysis. however. necessarily invokes an evaluation at some lev- el. Heritage 1984 for the change-of-state token oh). such as second assessments. The stance displayed frequently arises from an inference made by the speaker. the au- thors argue. Thompson and Mulac (1991a: 325) observe in passing that. I guess in my data frequently appears in contexts where the speakers do show an orientation toward evidence. Chafe (1986).

So they’re `color-cycling right ^now. The role of I guess in conversational stancetaking 191 terizing the stance object as having some specific quality or value (Du Bois this volume.e. Does it ^zoom `later? 3 [I `guess it ^must]..or herself. The semantic meaning of must ‘induction’ (Chafe 1986) of course strengthens this interpretation. together). the actual question. who may legitimately speak at the end of the first turn-construction unit. typically a question: (2) (Wonderful Abstract Notions SBCSAE 0017 <00:19:44>) 1 MICHAEL: . The topic of prior discourse has been Harold’s nephew Thomas. that sooner or later the picture will zoom in on something. actions involv- ing I guess propose something as a possible stance object that the co-participants can subsequently position themselves towards. and participate in. finally.. who is “awesome” and incredibly fast.. And. yet the second part of such a turn format may also contain other types of predicates (or even no predicate at all: Is @Yoyo @Ma Chi- nese? .10 Such turns are also regularly responded to by the recipient. and simultaneously makes public his or her inferential process and discovery of some state of affairs. or the speaker mak- ing a discovery based on inference. i. 4 JIM: [(H) Yeah. Michael seems to reason and infer.. even though the current speaker also continues to speak. Displaying a sudden change in the speaker’s state of knowledge or aware- ness is frequently witnessed in first-position turn formats of the following type in which the speaker displays on-line that his or her immediately preceding ut- terance was somehow wrong or mistaken and no longer relevant. 5 `I’ve got] ^zooms on here `too. is provided by the following example (3). on the basis of how videos are usually designed. . A further example of a change in the speaker’s awareness. Prior to this extract the partici- pants have been discussing and assessing him from their respective angles. The boy’s parents had been inspired by a famous young tap dancer. a three-year-old who is now learning to tap dance. While producing the question in line 2. SBCSAE 0019). I guess with a name like Yoyo=. 2 .. such actions frequently initiate stancetaking activity between the conversational co-participants (see also Good- win and Goodwin 2002 for assessment activity as something that the participants orient to. and in effect. cancels out the need for the action performed in it. Notice how interactive this turn design actually is: the speaker engages in interaction with him. see also Goodwin and Goodwin 2002: 154 for assessments as evaluations of persons and events being discussed within talk).

Harold also produces the last two words with some laughter. Their subsequent turns display that Harold has indeed drawn their attention to something that they had not thought of or fully realized during the discussion so far.on line 4. I mean he has a broken leg.-. he produces the action of simultaneously (in passing) announcing a piece of news and evaluating it (doing okay).192 Elise Kärkkäinen (3) (Lambada SBCSAE 0002 <00:00:54>) 1 HAROLD: I’m sure `Thomas is all ^over it. very likely talking about the nephew as well. 10 PETE: But it was his ^leg? Harold brings the prior topic to a close on line 1 by offering a concluding evalu- ative claim or assessment concerning the little nephew’s enthusiastic attitude to- ward tap dancing. Then several participants self-select to speak almost simul- taneously. 6 HAROLD: I `guess that `means his broken `leg is [3@doing @^okay3].7) 3 JAMIE: Prob[ably XX][2XXX2] -- 4 HAROLD: [<F><HI> I mean `he][2has a bro-2</HI><F/>] -- 5 MILES: [2XXXX could have2] ^see=n `him. Pete and Jamie decline this invitation and take up the topical import of Harold’s talk instead. which is marked prosodically by Harold’s lowered volume and a long pause after this turn. even though they clearly had some prior knowledge of the boy’s broken leg.. This is overlapped by Jamie (line 3. On line 6. even though we can now be fairly certain that Harold intended to announce it as news on line 4. 9 JAMIE: [4<HI> Oh yeah= </HI>4]. as will become obvious a little later. The broken leg had not been topicalized in the conversation so far. There is clearly a topic closure here (cf. 8 I was `imagining [4he had `broke an ^arm4] or something. 2 (3. largely inau- dible) and Miles (line 5. a turn design that highlights the unexpected nature of the state of affairs. wishing that he could have seen him tap dance). he refers to the nephew. which he is about to put on the . Goodwin and Goodwin 1992: 169). 7 PETE: [3I was ^wonder3]ing about `that. The role of I guess – or in this case I guess that means – here is to project and frame the speaker’s recent inference or realization. but Harold’s new beginning is clearly signaled by a rather loud and high-pitched I mean `he has a bro. thereby inviting laughter from the other participants (Jefferson 1979). On line 6 Harold redesigns his turn (rather than simply restarting and repeating what he had set out to say) and produces an inference I `guess that `means his broken`leg is @doing @^okay. which already incorporates as a presupposition that the nephew has in fact broken a leg.

. 8 MILES: [I ^guess the] [2`ones below are `women2]. 3 (1. In the next example... In example (3) the co-participants join in to further evaluate the nephew’s recovery. 19 PETE: [Right].. and a long sequence follows during which the participants discuss and assess how quickly the nephew had healed and how children’s bones in general “grow back really fast”. by introducing such a stance. We can therefore say that stance and evalua- tion take the center stage here. 9 PETE: [2@@@@@2] 10 JAMIE: [2@@@(H)2] 11 MILES: [3Is `that why they look so ^different3]? 12 JAMIE: [3@@ (H)3] 13 HAROLD: . i%. the In- donesian masks that Harold and Jamie have hanging on their wall. 2 .8) 6 JAMIE: @@@ 7 HAROLD: [We `need a ^verb]. 14 MILES: I mean. The role of I guess in conversational stancetaking 193 table.7) 4 HAROLD: ^Yep. and thereby affect the original action trajectory and sequential organization of talk.. namely first assessments. the evidence is clearly seen in the environment.-- 16 (1. the speaker invites and involves the co-participants in stancetaking in the same way as has been shown for one specific subtype of stanced actions. 5 (0. 22 (1. `there. 18 [I guess]. I guess signals that this particular stance is such that it arises from an almost on-line reasoning process of the speaker. 20 ..1) 17 Well you see their eyeballs. the broken leg. (4) (Lambada SBCSAE 0002 <00:07:05>) 1 MILES: . the speaker digresses from his original trajectory: rather than simply self-repairing and re-producing his original news announcement. <HI> `Those two top ^ma=sks </HI>. Yeah [2the other2] -- 21 MILES: [2In addition2] to a mustache and a beard. 15 . ^Uh-oh.0) . which make second assessments conditionally relevant (Pomerantz 1984). he proceeds to evaluate the newsworthy item. When producing such an inference. What is more.

the actual question. Miles introduces the two top masks as the actual referential focus of talk on line 1. We can see essentially a similar pattern in the following excerpt from insti- tutional data. he displays some trouble in continuing his talk. Miles himself realizes soon enough that the women in the masks have “a mustache and a beard”. he appears to be struck by a difference between them and offers as a reason that the two lower ones portray women rather than men. continuing at the height where the previous intonation unit left off. in line 11. In sum. even biases them toward a certain stance.194 Elise Kärkkäinen 23 But the thing is. The same example is also discussed by Haddington (this volume) with respect to the turn-design features of the inter- viewer’s and the interviewee’s turn. He also explicitly invites his co-participants to take a stance by asking them to ratify the reason he has just proposed on line 11: the question/request already presupposes as a given that they indeed look different. But before he initiates talk on that exact referent. and only in lines 23–24 does he proceed to make the point that one of the two top figures looks like a performer in a show that he has recently seen. the figure portrayed in one of the top masks. which is said in a kind of latching prosody. The two utterances are clearly designed to be produced together as one chunk: the noticing in line 8 (which ends in a continuing intona- tion) acts as a necessary background stance for the more explicit stancetaking ac- tion. what we can say about the role of I guess here is that it again appears in a context. the insertion of such an inference takes place in a sequen- tial position where the speaker rather abruptly changes the trajectory of his talk. In line 8 Miles then produces a noticing. In what follows. both Jamie and Harold display that there is some trouble involved in Miles’s proposed stance: Jamie laughs on line 12. in an overall high pitch. by a mock alarm cry uh-oh. in a side-sequence. and Harold displays. that there was something quite problematic in the prior turn. that they do not portray women at all but male figures. where the speaker makes public an inference (a noticing) that was based on some evidence provided by the immediate environ- ment. . At the same time. namely from an American television news interview recorded in the aftermath of the 9/11 terrorist attacks. as can be heard from the long pauses and the fact that Harold humorously prompts him to continue (lines 4 and 7). The inference is also part of a turn design that generally invites others to take a stance on the matter. and engages in a brief comparison of the top masks and the ones hanging below them. After this side sequence he resumes the original topic. 24 [that second one looks like the guy] who was in [2one of the2] [3Oba Oba3] skits.

and displays to the interviewee that the action that follows is not fully coherent with the main trajec- tory of the question. could you=. 6 IE: . as he re-directs it after having al- ready initiated the question Brian.. IE: Brian Jenkins (001 / 5 / 3:49) 1 IR: (0.<PAR> well I `guess the `public. Larry King Live. 5 `Could you ^bomb him out. this inserted stance is also produced in a parenthetical prosody. a phrasal break and subsequent insertion of a new unit (a word. who are represented as favoring the rather extreme action of simply bombing the ter- rorist leader out from wherever he is hiding (a rather widespread public opinion established on some rapid opinion polls).6)(TSK) ^Brian. which projects an upcoming stance.. The role of I guess in conversational stancetaking 195 (5) CNN.-- 8 I don’t know that you could ^bo=mb him out. For example. however. He subsequently inserts an utter- ance which incorporates the stance of an ‘other. I guess he’s. something else is going on. Well here signals an orientation shift within the speaker’s extended turn (Schiffrin 1987: 102–127). 2001: Bomb him out IR: Larry King. (TSK)(H)U=h. or a whole sentence) in the emergent turn is often caused by the fact that the speaker has not yet secured the recipient’s gaze or needs to otherwise negotiate a state of mutual focus with. Simultaneously with the onset of . The IR engages in looking at the IE only until the end of line 2. 4 would `look at this `simply </PAR>. Goodwin (1981: 127) gives the following example that contains the insertion of I guess at precisely the point where the gaze of the recipient (marked by X) arrives: ↓ GARY: He’s a policeman in Bellview and he :. and coordinate his or her actions with those of.’ namely of the general public. in this case a third-party stance arrived at via an on-line realization of the speaker (see Haddington this volume for a fuller description of such “position- ing” activity). Sep 12. Iconically. The interviewer (IR) designs his turn on-line. namely in slightly lower pitch and faster tempo. the recipient. 3 . 2 could `you=. Goodwin (1981: 127) suggests that such modifications of the turn construc- tion are regularly done by participants to “coordinate their production with the actions of a recipient”. a phrase. [ X__________ In example (5) above.. It is further framed by I guess. 7 I d.

This stance arises from a reasoning process of the speaker. or as an ‘offer’ or a ‘presentation’ of the discourse (or stance) object to the recipient (Kendon 2004: 123). on line 2. especially the downward gaze and the (parenthetical) prosodic production. while doing so. it is possible to argue that the withdrawal of gaze works towards ensuring that the speaker can keep the turn during the production of the insert and subsequently continue with the projected trajectory of the turn (cf. rather than the actual stance of the current speaker himself. not unlike . He also turns his palm upward. this turn design invites the IE to take a stance on the matter and. a realization of the IR inserted in the larger question sequence.” By inserting the I guess-framed digression. I guess in initiating actions like the ones above indexes a just discov- ered stance. Only after the insertion is completed does he turn his gaze toward the IE again. which is often based on evidence or some other stimulus gleaned from just prior turn(s) or even the ongoing one. The difference between I guess and surely in the degree of change in awareness is then traceable to the semantic meanings of guess and sure. In sum. together contribute to its interpretation as an interactionally engendered ‘remembering’ (Heritage 2005). This evidential form is said to index information that is new or unexpected to a speaker. while the stimulus for this is derived most likely from the verbal message but also from visual or sensory evidence.196 Elise Kärkkäinen the parenthetical digression starting with well I guess. the use of which is claimed by Downing (2001) to be triggered by a psychological event of “coming into awareness” at the very moment of speaking. or from the wider social context. We can find hints as to how the IR wishes to present himself in view of the matter at hand: he had started out by saying could you=. he withdraws his gaze from the IE and looks down. surely. A special flavor of I guess. The above insertion again causes talk to diverge from its original trajectory. On the other hand. which would have constituted a valid question presented in the IR’s “own name. Lerner 1991). The embodied production of the insertion. and to now asking the question “on behalf of ” the general public rather than himself. from the actual environment where talk takes place. We may conclude that I guess acts as a frame in a turn that is designed to convey a third-party stance inserted on-line. the IR switches from this to proposing another stance that he may or may not himself share. take into account the ostensibly simple solu- tion favored by the general public (but not necessarily shared by the IR). a sudden change in the speaker’s epistemological state of knowledge or awareness or orientation. very likely resulting in could you bomb him out. while the palm remains in its upward position. Finally. I guess here comes quite close to another marker of evidentiality and stance in British English data. or via inference (Downing 2001: 277). possibly in acknowledge- ment that he had already brought up the results of the opinion poll earlier in the interview.

in respon- sive actions to some other actions) in conversational sequences that often involve some degree of disagreement and disaffiliation between the participants. and in doing so to take into account the stance already implicit or explicit in the de- sign of the ongoing turn. as one in an adjacency pair sequence). as the actions containing this frame often invite them to also take a stance in subsequent discourse. i. even though she has no prior knowledge of the incident. to align with another speaker and. when he had witnessed the feeding of a fish.. The role of I guess in conversational stancetaking 197 that of surely.. an Oscar. In the next example. these nevertheless derive from and surface in the intersubjective to and fro of dialogue. is taking an active role at points in the telling (Joanne and Ken are a couple).or herself (Lerner 1991. initiated and closed by the speaker him. Here. in what follows. stances gain in importance to such an extent that they impinge on the sequential organization of talk and in some way change or temporarily withhold its original (action) trajectory.2 I guess in responsive actions: Giving up and redefining a stance Let us now turn to cases where I guess appears in second position (i. . 4. in a pet store: a goldfish was put in the tank and the Oscar started chasing it. Duvallon and Routarinne 2005. is that the stance that “surfaces” in the associated utterance often has a humorous and surprised quality.e. as one of the two recipients. Local 1992.e. Note that even though many researchers use the terms ‘alignment’ and ‘affiliation’ almost interchangeably. The I guess-framed ut- terance often constitutes a clear side sequence within the current speaker’s ongo- ing sequence. while an affiliat- ing action is used to refer to one that establishes and increases social solidarity between the social actors involved. to dis- play affiliation and a convergent stance. withdraw. Ken is telling a story about his childhood. and sometimes comes close to sarcasm. by aligning action is simply meant one that is produced within the trajectory of the sequence-so-far (e. What is essential is that even though we see evidence of some private pro- cesses of inference in the turns framed with I guess. I guess can further be seen to or- ganize the stancetaking activity of the co-participants.g. cf. We may also argue that at such points in interaction. also Jefferson 1972 for subsidiary side sequences initiated by another speaker). frequently. I guess similarly acts to display and project – on the basis of just prior discourse – that the current speaker wishes to modify. Joanne. and redefine his or her origi- nal stance at this point.

. Well it was go -- 11 went ^tail `first. which is produced with a drop in pitch on tail that indicates some resignation at the realization of this fact. 7 JOANNE: [^Fin `first]? 8 KEN: it was `going. For Joanne. and actively pursues that in the discourse that follows. Ken starts to disagree on lines 5–6 (^No. her point having been that the snake would always eat the goldfish by the head. 12 JOANNE: `That’s the [^problem. Mandelbaum 1989: 118). <MRC> ^half-way ^into his ^mou[=th </MRC>]. but he is interrupted by Joanne. There is a danger from line 4 onwards that Joanne may halt or derail the storytell- ing (cf..) and again in line 8 (it was `going. line 7 and ^Mouth` first?. it was `going. line 8). 19 (H) you got this ^Oscar there. On lines 10–11. this of course proves . went ^tail `first. His turn design conveys that he comes to this conclusion pretty much at the moment of speaking: he self-repairs and restarts with a simple past tense in line 11. She therefore assumes that the problem in Ken’s story was that the Oscar was eating the fish by the fin (which for her equals tail. It `went the wrong ^way. 15 KEN: <A> `I don’t `know </A>. 4 JOANNE: [It `went the] wrong ^way. who proposes further candidate alternatives for how the fish was going (^Fin `first?. 2 (H) But the `goldfish got ^s=tuck. as she has been talking about the two interchangeably) rather than by the head. it is of some consequence which way the fish was going: this story may in actual fact confirm her just prior story about a snake that was likewise fed with a gold- fish. 9 JOANNE: . 13 KEN: [I `guess that’s the ^wrong way.198 Elise Kärkkäinen (6) (Deadly Diseases SBCSAE 0015 <00:24:14>) 1 KEN: and `then= uh. She hastens to propose a reason why the goldfish might have been stuck in the Oscar’s mouth. Ken displays recognition that the fish actually went in tail first. 16 (H)] ^Anyway. 14 JOANNE: (H) `That’s the <HI> ^problem </HI>]. 5 KEN: ^No. For Joanne. 20 `swimming around [in the ^tank.).. 3 . 6 it [was `going].. ^Mouth `first? 10 KEN: .. (TSK)`So. 17 JOANNE: [2(H)2] 18 KEN: [2(H)2] .

In fact. Jefferson 1972: 316–320 for resumption of the ongoing sequence. in the form of two diagraphs. cf. There is some evidence for this in how Ken designs his turn. the speaker is giving up his original stance and adopting a new one. to make visible the frequent modification of stances that takes place between dis- course participants: a stance is often a “product of an immediately prior act of stancetaking toward a shared stance object” (Du Bois 2003). this particular syntactic chunk or collocation strongly projects that Joanne is about to produce an evalua- tive turn. Diagraphs map the relevant intonation units produced by different speakers onto each other (rough- ly) according to syntactic structure. I will present the turns above in which stances are being ne- gotiated and a (more) convergent stance is achieved. even though he misprojects what type of evaluative item will follow (the problem). in order to retain his role as the main storyteller and to continue with his story. in light of what was established in just prior discourse. line 4). The role of I guess in this sequence is to display that. Ken indeed displays an understanding of Joanne’s turn-so-far as an evaluative one. After an audible inbreath he contin- ues the story with ^Anyway.’ Ken finally tags on I don’t know in a very accelerated form. In the following. and produces the wrong way instead. Lenk 1998: 71–78 on anyway closing general conversational digressions and resuming the earlier conversational topic). the participants have slightly shifted their stance object here: Joanne orients toward gaining wider evidence for her larger point (that eating a fish by the tail can be a problem not just for snakes but also for predatory fish). and she rather enthusiastically displays this by producing `That’s the ^problem. I `guess that’s the ^wrong way. . (cf. Ford and Thompson 1996: 169. whether this is of a formulaic type (That’s the problem/point) or simply ends by repeating Joanne’s earlier evaluative item the wrong way (cf. Such a diagrammatic form of presentation has been proposed by Du Bois (2001). <A> `I don’t `know </A>. He hastens to agree already upon hearing Joanne’s that’s the (line 12). because here as well. its use here is not far removed from that witnessed in initiating actions (see Section 4. the stimulus or evidence that occasions the adjustment of the speaker’s original stance derives from the immediately preceding turns (which again display the speaker’s on-line recognition of some relevant facts). with the pitch going up and the pitch range wid- ening during the second production of <HI> ^problem </HI> (partly also be- cause she is overlapped by Ken). This turn indicates that Ken is giving up on what for him is possibly a relatively small detail in the bigger scheme of things.1). thereby marking that this side sequence in the story has come to a completion (cf. Scheibman 2000). Ken finally concedes on lines 13 and 15 that Joanne may be right. The role of I guess in conversational stancetaking 199 her earlier point about the snake. In this sense. while Ken still orients to the more local ‘which way is the wrong way.

we can see that Ken’s actual turns-at-talk are largely built on Joanne’s previous talk and the stance displayed therein. Ken resonates with the syntactic structure and evaluative terms in Joanne’s prior turns: he picks up the reference term that and the predicate ’s from Joanne’s immediately previous turn. Ken produces an affiliating second evaluation on line 13. . where he engages in self-repair: the design of this utterance displays his initial recognition of the fact that the fish may indeed have gone in the wrong way. the participants first establish which way the fish was going in the Oscar’s mouth. 13 KEN: I `guess that ’s the ^wrong way. Diagraph 2 (from example 6): 4 JOANNE: It `went the wrong ^way. 5 KEN: ^No. Here. slightly shifting the stance object here. even though. also in the past tense. The next diagraph presents how this is done. as we have seen. He also produces the final item. The past tense went signals finality compared to the past progressive was going and resonates with Joanne’s original wording on line 4. 6 KEN: it was `going. ^tail first. Diagraph 1 (from example 6): 4 JOANNE: It `went the wrong ^way. 7 JOANNE: ^Fin `first? 8 KEN: it was `going. Joanne’s initial anticipatory judgment of the fish going in ‘the wrong way’ was expressed on line 4. the speakers resonate both with their own and with each others’ talk at several levels of linguistic structure. Mandelbaum 1993: 263 for negotiations about versions of “re- ality”). Their versions of reality now match factually (cf.200 Elise Kärkkäinen In Diagraph 1.’ Joanne produces an explicit ‘I- told-you-so’ concluding evaluation on line 12. in resonance with both the syntactic form and prosodic realization (in terms of accent) of Joanne’s candidate items in lines 7 and 9. 12 JOANNE: `That ’s the ^problem. 9 JOANNE: ^Mouth `first? 10 KEN: Well it was go -- 11 went ^tail `first. but what the implications of this are for their initial disagreement need to be made more explicit. Joanne completes the syntactic structure initiated by Ken (lines 6–7 and 8–9). This format is broken by Ken himself on line 11. After the participants have established together that the fish had in fact entered the Oscar’s mouth ‘tail first. In doing so.

disagrees on such a framing and puts in a correction in lines 6–7. ^imitation `cologne.he was. the story consociate. The role of I guess in conversational stancetaking 201 and also recycles her earlier evaluative element. 10 I <HI>`guess </HI> he was `trying to like. or doomed to fail. an evoked stance. 4 KEVIN: he was trying to `sell us ^cologne [2f-2]– 5 KENDRA: [2(COUGH)2] [3(COUGH)3] 6 WENDY: [2No2]. and can be heard to put on a voice quality of “slight despera- tion. `place where they ^would `sell. 16 `be[cause] </VOX>. 14 but he `said.. here it is apparent that for her there is now some trouble involved (cf. The following is another instance of storytelling.. the wrong way. 7 [3he `wasn’t3] trying to ^sell [4us `co4]lo[5gne5]. Some guy came out and he was trying to sell us cologne (lines 1 and 4). Kevin produces an assertion that frames the story and the incident as having been mainly about selling. their interpretations of what exactly the pro- tagonist was doing or trying to do differ. He further signals his convergent stance by framing the turn with I guess. Kevin and Wendy (a couple) are jointly telling about an incident that they had witnessed together recently in the parking lot of a mall. namely that try- ing to sell cologne to them is something rather strange. 12 like. 11 ^lure us to a .. 8 KEVIN: [4`Well it-4]– 9 [5No5]=. or about trying to sell. however. Lerner 1992). funny.11 But Wendy. This turn design contains an implied evaluation. He produces a laughing particle in the middle. 2 @[=] 3 WENDY: [<HI> Oh </HI>]. No he wasn’t try- . 15 <VOX> it’s `not ^imitation. restarts several times as if to further emphasize his upcoming point. (7) (Appease the Monster SBCSAE 0013 <00:12:33>) 1 KEVIN: . We can find evidence for this in Kevin’s production of this turn. Wendy explicitly disaffiliates with Kevin.. 13 .” He is thereby orienting the other participants toward the upcoming story in a certain way. Even though in the immediately preceding story preface sequence Wendy had aligned herself as a story consociate. Some `guy came `out and he.

’ It seems enough of a concession to Kevin.. Lerner 1992 on story consociates correcting facts). Diagraph 3 (from example 7): 4 K: he was trying to `sell us ^cologne 7 W: he was n’t trying to ^sell us cologne. sell. it is rather that their fundamental perceptions differ of what actually happened in the parking lot. place where they ^would sell. who then as- sumes the role of the main storyteller and continues the story.displays that he may be about to start with a further disagree- ment. ^lure us to a . with an emphasis on the contentious item..202 Elise Kärkkäinen ing to ^sell us cologne. It is not the case that Wendy is disagreeing on the facts of the story (cf. and in addition. `Well it. that the guy was going to sell something later. bringing in some new information. that further specify how he had perceived the original incident. 10 K: I guess he was trying to like. a place and imitation cologne. We can see from the diagraph that on line 7 Wendy is building her divergent stance on Kevin’s just expressed stance. I guess he was try- ing to like. 11 ^lure us to a . but he then retracts and produces an explicit No=. The form ^would sell even escalates the content in trying to sell. It has a concessive flavor and could be generally glossed as ‘In light of what you just said there is enough evidence for me to modify my stance.-. by resonating with their syntactic form (he was trying to X us). She does so with the negation and with just slightly shifting the stress from cologne to sell. in effect agreeing with Wendy’s previous disaffiliating turn. in other words Kevin does not altogether withdraw the idea of selling. like. Kevin immediately starts to redesign his turn. thus displaying that Kevin still maintains his original position on this point. He offers a new and more precise understanding of the past events. Upon hearing the emphasized word sell. 13 imitation cologne. 12 like. Kevin in turn builds his now more converging stance on his own and Wendy’s immediately preceding turn. . and their versions of “reality” are being negotiated here (Mandelbaum 1993: 263). In what follows. place where they ^would sell. Kevin’s use of I guess is highly interactive in that it arises out of a need to retract and redesign a turn to display a new and modified stance that converges more with the previous speaker’s stance. imitation cologne. he produces a new claim that takes back some but not all of what he had said initially.

in answers to requests for information. Here.. ^Small ^percentage. in which the function of this frame comes close to the seman- tic meaning of (weak) “belief ” as proposed by Chafe (1986) and Thompson and Mulac (1991a). ^Caucasians. Here.g.. 7 . the speaker may simply answer I ^guess or preface his or her answer with I ^guess.. ^Basically they’re . Lastly.” (8) (Car Sales SBCSAE. At the risk of simplifying somewhat. On occasion.. there is a typical reactive response to evaluative first actions. 6 . I guess so. First. The role of I guess in conversational stancetaking 203 In sum. ^Spanish `speaking `people go ^in there? 3 D: <YWN> U=h.. I will briefly mention two other types of second-position actions framed by I guess. in re- sponse to some evidence or other material presented in just prior turns. unpublished) 1 G: . B: <@> Yeah I guess you do </@>. where I guess marks the participant’s striving toward a conver- gent stance (e. I guess is brought in when the current speaker signals that. a (sud- den) change in his or her state of knowledge or awareness has occurred and he or she wishes to align and affiliate with the previous speaker and to express a (more) convergent stance. whereas I guess basically affiliates with the prior speaker. or a new speaker perspective (Kärkkäinen 2003). I guess appears simply to express that the actual answer to be offered is probably true. and the function again comes close to the semantic meaning of “belief. the difference between I think and I guess in second position is that I think tends to project or frame dis- aligning or disaffiliative second actions. Even though very similar. SBCSAE 0005). I guess projects an inserted stance that is often based on an infer- .. while the other participants align as recipients. namely a different slant or take on a given issue. I guess occurs in conversational sequences in which one main speak- er produces a multi-unit turn or tells a story. 4. Yeah. 2 . it also takes on a sarcastic tone or conveys joking reluctance in agreements with denigrating prior turns (A: I get a little ahead of myself. (H) They have `many uh.. Second. 5 I ^guess‑ </YWN>.3 I guess in extended turns: Inserting a stance In many cases.. 4 u=h. the two are not inter- changeable but have each specialized in a certain type of interactional function. compliance).

8 KEN: . 15 [six or eight inch]es. 5 . [they are these really. Such an aside often works toward increasing un- derstanding or appreciation of the telling or story.they’re big]. Here.or herself as not possessing or relating first-hand knowledge.</X>. The following extract comes just before the one in example (6) and starts the orientation sequence of the story (the actual telling starts on line 39). you know like ^goldfish. 28 you know. 19 . 20 They eat like small [^fish. 25 get the `brunt of ^everything. 24 <A> And I guess #the </A> goldfish or ^g=uppies. 9 They’re like [2um. 4 (H)= And they are these like. 17 [2(H)2] 18 KEN: [2(H)2] and they’ve got... 3 (H) they had this fish in there called] an Oscar. I guess sometimes comes very close to an actual hearsay evidential used to mark that the current speaker positions him. 26 . (H) Poor ^guys. (9) (Deadly Diseases SBCSAE 0015 <00:23:47>) 1 KEN: [Anyway. 29 they’re like the ^cattle of the @`animal @^world. 16 JOANNE: [big eyes]. they `only eat like other <HI> ^fish </HI>.-- 14 I don’t know. 10 JOANNE: [2Yeah. They’re these] big gnarly suckers. 11 I remember what they look2] like. 12 KEN: . 21 JOANNE: [(SNIFF)=] 22 KEN: And they eat like]... 27 (H)(Hx) They’re `like the.204 Elise Kärkkäinen ence by the teller. the stance may constitute a side sequence or an aside from the main story line and its trajectory.. about feeding the Oscar with goldfish.. Here is an example from the same story as in example (6). 23 . (TSK) (H) They’re like m-2] -- 13 f.. 7 yeah th.. 2 I remember they had a. . 6 JOANNE: [<X> I remember.

(TSK) And they dropped in the goldfish. 33 Yeah. 38 in every pet shop (Hx)]. He first asserts that Oscars eat small fish. Scheibman this volume). and then establishes that goldfish are among them.. The discourse markers you know and like as well as rather strong primary accent on ^goldfish can be taken as displays of his awareness that in just prior discourse Joanne has told a story about a snake that was fed with goldfish (“this is the type of fish that we talked about before”). Throughout the orientation sequence Ken casts a very negative stance toward the Oscars by using negatively evaluative lexemes (big gnarly suckers). it is generally very low in pitch. 34 they’re feeders. and further evaluates their lot in the animal world (lines 27–30). On lines 24–25. He is struck by the likeness of what he is about to tell to what Joanne had just told before. 32 f=eeds. Poor ^guys. 35 KEN: (H) Anyway. Ken makes a generalizing evaluative claim about the hard des- tiny of goldfish. he yet again evokes a negative stance toward them through only. namely with a “discovery” prosody: it is latched on with and without a break and with a very fast tempo at the beginning of the IU. but they are also predatory fish. Ken then stops to evaluate the goldfish some more: he offers a token of sym- pathy toward them. very likely to intensify interest in the upcoming story about the dramatic feeding of the Oscar. at least to express their validation or ratification but possibly also a convergent . (cf. 37 JOANNE: They’re nothing but tanks of feeders. Pomerantz 1986). get the `brunt of ^everything. Ken wishes to involve the story recipients in taking a stance on the same object. 40 when I was a kid. It is possible and even likely that by engaging in such escalat- ing stancetaking activity that gradually also becomes more specific. In line 19. the goldfish. The utterance can now be heard as really making an inference on the basis of the previous story and the present story-so-far. and the extreme case formulation ^everything re- ceives a rather strong primary stress (cf. he engages in a kind of incremental turn design and escalating stancetaking activity. From here onwards. 39 KEN: (H) I remember w-] watching them feed the Oscar once. not only are the Os- cars big and ugly. 31 JOANNE: [Yeah.. The role of I guess in conversational stancetaking 205 30 [They’re br=ed only to f=eed other things. 36 (H) anyway so. and this is reflected in the way this generalization is produced. 41 . <A> And I guess #the </A> goldfish or ^g=uppies.

of course. shoer. 11 they have . . The inserted stance.. The next example briefly illustrates a case in which no visible uptake follows from the recipient.. 15 (H) <A> And I guess like in `Minnesota it’s real </A> <LO> ^we=t </LO> . the recipients have not yet heard Ken’s actual story. (10) (Actual Blacksmithing SBCSAE 0001 <00:11:18>) 1 LYNNE: (H) Anyway. and Ken continues his story (Anyway.12 Such initial and final marking or bracket- ing of discourse sequences. 8 LYNNE: That do this? 9 (H) This summer I met one.. ferrier girl=s you know? 7 LENORE: Uhhuh.. 14 they’ve had him for years and years and years. the generalizing inference. strengthens the interpretation of I guess as a prospective discourse marker for which more support will be given in Section 5. The recency here arises from the speaker recognizing the similarity of import of the prior story as well as the present story-so-far (or really upcoming story).-- 3 . the girl. 4 you know. The role of I guess in this conversational story is to display and project that the speaker is about to digress from the main story line and insert a stanced comment of some kind.. Joanne expresses a strong convergent stance at the earliest possible point on lines 31–34.. 12 that comes to their house. but in which the turn design is nevertheless a very interactive one. 13 all the time. very many. 2 I was gonna tell you about that %. This is a story told by a young girl who is herself learning equine science and in this story expresses her appreciation of a girl ferrier in Minnesota. 10 (H) Jorgensen’s. The speaker opens this digression or side sequence with I guess (or in this case And I guess) and closes it with anyway (cf. 5 there isn’t mary. Once Ken has elaborated his point some more. again Jefferson 1972 for side sequences initiated by a recipient and Lenk 1998: 71–78 for anyway after general conversational digressions). But the recipients may not have been able to do so at line 26 (notice the short pause). what is more. one . 6 ...206 Elise Kärkkäinen stance. as ‘getting the brunt of everything’ is still rather vague in reference and. anyway so). is of the same “recently discov- ered” quality as we have seen in earlier sections of this paper.

The role of I guess in conversational stancetaking 207 16 (0. 18 . The reference to Minnesota is explained by the fact that Lynne has spent most of the summer there with her boyfriend and his fam- ily..e. uh. as it turns out much later in the story.7) and ` stuff you know? 17 So like. 19 for not coming off. 21 is so much softer. Yet the inserted assessment in lines 15-16 acts as a preface to a longer assessment sequence that is consequential for the understand- ing of the story about the ferrier girl: the particular challenges that ferriers face in Minnesota is that the horses’ hooves are too wet (because of the wetness of the ground). 72 she’s going to an actual . The story is told in Montana. 69 and. Again. the Jorgensens. in this f=errier... is incredibly strong and competent after only nine months of ferrier college. I guess is here used as a stance frame to project that such an assessment and in fact a longer sequence elaborating it will follow.. (TSK) they always have this one horseshoer. Again by way of summary. and lines 15–16 are clearly pragmatically incomplete.. i. The assessment. speakers mark through I guess that they are diverging from the trajectory of the story or telling. I guess is in large part doing organiza- tional work in the telling. 70 (H) this girl’s been . they really have to watch their shoes. These involve taking an explicit stance toward some- .. even though this time the resumption of the main story line about the girl ferrier comes considerably later after 46 lines: 67 LYNNE: (H) But anyway. is designed to continue even after the high rising intonation at the end of the IU on line 16: there is no pause at all after this turn- construction unit. to insert a stanced utterance or even a longer sequence of such utterances. 20 because the hoof wall. 93–94 for anyway af- ter digressions supplying relevant background information and but anyway used for continuing the narration). at the beginning of a side sequence (see also example 11 below). The protagonist of Lynne’s story is a girl ferrier. as it comes at a transition point from the main story line. produced some- what faster than prior speech. 68 . ferrier college . this digression is closed with anyway (see Lenk 1998: 65–68.. while they are producing multi-unit turns such as conversational stories. who. 71 .. requiring elaboration (Ford and Thompson 1996).

we have also seen evidence in the examples above of a similarly discourse- organizing function of I guess. Kärkkäinen 2003. I guess is one of the very frequent stanced items in spoken American English. that is.1 above. 5. or serves as an explanation for the recipients to better appreciate the story. but is often something that the narrator him. and a summary of . and is also potentially disputable. or display only minimal ratification (or ratification through embodied action only). However. it is often also enough for recipients to align as pas- sive recipients of an extended turn such as a conversational story. we can say that I guess again organizes interaction in that it is used to draw the recipients’ atten- tion to the current speaker’s stance. And finally.2) that I guess is one of a group of epistemic/evidential/evaluative reusable fragments that speakers may use as practices of turn construction (Thompson 2002). I will relate the present findings to this body of research: this fragment or token very largely fulfils the criteria proposed by Schiffrin in her seminal work (1987: 328) for a linguistic item to count as a discourse marker (cf. Lenk 1998 for a view of discourse markers as primarily marking global coherence relations). stancetaking assumes a central role in interaction. In all. I guess as a discourse marker In this paper the view has been adopted (see Section 3. possibly also to invite them to appreciate and ratify this stance.or herself just inferred from the story-so- far. Such a view of grammar is a novel one and emphasizes grammar as consisting of a “collection of crystal- lizations of linguistic routines” used in the service of social interaction (Ford et al. Scheibman 2000). In this section.13 I guess may simultaneously mark that the teller does not possess first-hand knowledge or was not the actual ex- periencer of an event. even though not quite as frequent as I think and I don’t know. arises from the story as a conclusion. 2003: 120). this view fits in rather nicely with that proposed in a long body of linguistic research that views tokens like I guess under the category of ‘discourse markers.208 Elise Kärkkäinen thing that is consequential for the story. but that the telling at this point is based on hearsay or on some indirect evidence. an essentially similar list by Brinton represented in Jucker and Ziv 1998: 3. Similarly to first and second position I guesses. However. as it may signal that an upcoming sequence is not fully coherent with the main flow of the telling or talk (cf. it is worth noting that the I guess-framed stanced utterances or sequences deviate from the original trajectory of talk.’ As was established in Section 3. a shift in the epistemological stance of the speaker is thus again evident. While the latter have been shown in recent research to have specialized into some rather routine discourse-organizing func- tions in conversation (Beach and Metzger 1997.

some of the observations that I will make below of its discourse marker status are not directly observable in the example segments above. as it has largely lost all overt marking of a syntactic connection to them. <Actual Blacksmithing SBCSAE 0001 00:12:46>. But on other occasions. primary. or (I) guess you were right (1993: 38). I guess appears to retain some of its referential. I guess is syntactically detachable from sentences. Schiffrin’s (1987) second criterion for discourse markers is that they have to be commonly used in initial position of an utterance. Fourth. and not on all possible environments of use of this marker. because the focus in the present article has been on the role of I guess in the stancetaking activity between conversational co-participants. (5)–(7). it seems to make only a very general or vague reference to the actual speaker: it primar- ily frames and projects an upcoming stancetaking action and simultaneously the onset of a digression.g. I do not present such occurrences here). In many contexts of use. or no stress on guess (see examples (3). and (11) for second- ary stress. First. The role of I guess in conversational stancetaking 209 criteria by different scholars in Helt 1997: 16–17). (6). (4) and (8) for primary stress. It also has some discourse and sentence mobility. namely the complementizer that. examples (5). a side sequence. in example (4). a separate IU. However. where the side sequence extends over some 50 IUs). I guess has a range of prosodic contours. In the . This criterion is also clearly filled by I guess. Doris: … Guess I’ll put it o=n. but also a side sequence of several units (as in example (10). I guess operates at both local and global levels of discourse. and (9)).g. only one such instance was found in the present database.. Givón indeed observes of “epistemic quantifiers” that “the conventionalized subject pronoun is so specific to particular verbs. Third. the speaker bases his inference on visual evidence of the masks on the wall (and stresses ^guess). lexical meaning and brings in a particular kind of evidential stance. according to Schiffrin (1987) a discourse marker either has to have no meaning. that it is often dropped in rapid speech. it may also at times be en- coded as an independent intonational entity. Fifth. However. a vague meaning.and sentence-initially (as in all the examples of this paper). or to be reflexive (of the language. because it can project an inserted stanced digression not only one utterance long (e. e. and may appear (even though rarely) in the middle of an emergent syntactic sentence (for lack of space. while at times it is produced in a very accelerated and phonologically reduced form (ex- amples (9) and (10)) or with parenthetical prosody (example (5)). of the speaker). because even though it tends strongly to occur intonation-unit-initially and at the same time clause. and (9) for an unstressed case). which was established im- mediately above. or in example (10). it’s not on.. the teller does not have first-hand knowledge or experience about Minnesota.” as in (I) think she’s there. as it is produced with second- ary.

4 into `another ^car. 8 <A> (H) I `guess that was `like another ^stop. 7 . it seems was originally a marker of induction but may sometimes be used as a hearsay evidential). 11 . we cannot say that I guess has no meaning at all but a la- tent (vague) meaning.. Again the aside contains the speaker’s speculation about what may have caused the man to come back to her car. in this `seat right ^here. possibly. . that’s when he `came ... the victim of an exhibitionist. e.. The following two examples further display the difference of I guess and I think as discourse markers. I guess acts much like a hearsay evidential such as people say or I’m told. 12 ^sat . is engaged in the delicate issue of preparing Rickie. In example (11). and the special nature of each will become more apparent. 10 . I guess inserts the speaker’s conjecture as to what happened otherwise on the train when the defendant moved closer to her: Rickie infers that it was easier for him to move because people were moving about anyway as they were getting off the train (or.. 5 REBECCA: ^Oh[=]. and ^the=n. 2 REBECCA: [He `went through ^these `doors]? 3 RICKIE: into `it unhunh].g. Rebecca. Of course we can argue that there is some semantic meaning to I guess. to appear as a witness in court. Here. and `then. and the role of I guess is to mark it as such. because there were now fewer people). We can see that lines 8–9 are inserted in a faster tempo to bring in information that is subsidiary to the main activity of establishing the exact location of the de- fendant. Let us compare this example to an almost similar one that came a little earlier in the same conversa- tion. and displays the speaker’s fine-tuned positioning toward the knowledge that she is offering (see Chafe 1986 on how some evidentials may be “borrowed” from one function or mode of knowing to another. 6 RICKIE: [(H)] And `then he came `back ^again. and ^s=at b=y me. as speakers do not simply choose I think or some other discourse marker in its place – in other words. where we have I think instead of I guess. an attorney. 9 and more `people were getting ^off </A>..210 Elise Kärkkäinen latter cases. Rickie is relating to Rebecca where exactly on the train the defendant was when the incident took place: where he got in and where he was seated (they have the layout of the train in front of them).. (11) (Tell the Jury That SBCSAE 0008 <00:08:11>) 1 RICKIE: He went ^through <X> th[ere </X>.

and what its role is in the stancetaking activity between conversational co-par- ticipants. . when . as Rebecca’s question on lines 1–3 was really meant to establish where on the train. which may appear in rather diverse sequential positions in interac- tion. 6. where was he when you first saw him .. 11 . The role of I guess in conversational stancetaking 211 (12) (Tell the Jury That SBCSAE 0008 <00:06:53>)14 1 REBECCA: A=nd. let me see </VOX>. responsive actions..9) 10 Okay. The difference between the two markers is in the potentiality inherent in I guess to display an inference. I have argued that I guess. Rickie goes on to provide more information about whether the defendant got in the same car or not on line 7. often on the basis of available evidence. He got on another car. that is. 3 . By inserting I think (and or something) she simply marks part of the subsidiary information as uncertain. the observed commonalities can be summarized as follows. I think marks speaker uncertainty and orients more toward the speaker’s memory as the basis of uncertainty. 4 RICKIE: `He got on the ^trai=n `o=n like. 2 . 8 . By contrast. My aim was to examine what kind of stance is expressed by this marker. 5 (THROAT) I ^think ^Twelfth Street or something? 6 in . rather than where in Oakland... a reusable evidential (and not really epis- temic) fragment. Conclusion In this paper. namely sequence- initiating actions.. and extended tellings or multi-unit turns... The speaker therefore orients toward making a conclusion rather than simply of- fering uncertain information. 9 (3. the defendant had been when Rickie first laid eyes on him. come in. to display that the speaker is realizing or inferring something at the moment of speaking. when he came in. This is similarly a case of subsidiary information inserted in a lengthy answering turn. is frequently employed in American English conversation as a stance frame. Despite the different sequential positions established. (H) ^Oakland? 7 (H)<VOX> And like in the.

or from the wider social context.e. single-speaker vantage point: – The form and decontextualized meaning of I guess is that of a subjective evi- dential marker. I guess indexes a reasoning or inferential process of the speaker. with no overt evi- dence in sight). or is based on. – Within these actions and turns-at-talk. – I guess projects a certain type of stance to be expressed in the utterance to fol- low. The evidence in question is provided by prior discourse or by visual or sensory evidence. to display that the current speaker gives up or modifies his or her stance (and not only aligns but also often af- filiates with the recipient). – Depending on the sequential position. I guess-framed turns and actions may function to invite others to take a stance. we can say that stance impinges on the projected trajectory of talk and . or alerts recipients to an inserted stance at crucial points in a telling and invites them to join in the stancetaking activity. and to draw the recipients’ attention to the current speaker’s inserted stance (to increase and facilitate their understanding and appreciation of a telling). nor is it easy to determine without recourse to the actual context of use. From an intersubjective perspective: – The stance projected by I guess is not carried around ready-made but “surfac- es” in the sequentially constructed intersubjective weave of dialogue. I guess takes into account these interactional contingencies and displays the recency of the speaker’s inference and displayed stance. the speaker invites others to also take a stance..” “coming to a realization.” i. which involves a shift in the speaker’s epistemological stance. It arises from. Yet the actual referential meaning of this evidential is not very strong in many instances of use. – Finally.” or “drawing a conclusion. and also via inference (i.212 Elise Kärkkäinen From a subjective. It is used to signal that by making public a “just discovered” subjective stance.. from the actual physical en- vironment. – Such a stance can variously be said to involve “making an inference. evidence or some stimulus gleaned by the speaker (sto- ryteller) from the ongoing or just prior turn(s). I guess can be viewed as an organizer of the stancetaking activity between co-participants. I guess appears at points in the interaction at which stance and stan- cetaking is highlighted in favor of proceeding with business as usual. in that sense. gives up on or adopts a new stance.e. The relatively high frequency of I guess in conversational data serves as further proof for this organizing function: this fragment can be fully regarded as a discourse marker that signals an upcom- ing stanced digression or side sequence to follow.

discrete grammatical or lexical devices analyzable in single speakers’ contributions. it seems that I guess not so much provides a per- spective toward the actions. I guess signals. claims. projects. unambiguous “stance” anymore. The role of I guess in conversational stancetaking 213 thus organizes interaction on a par with other interactional organizations such as coherence. If. that stance marking can be of a very routin- ized nature linguistically. as it signals and displays a just discovered stance to the co-participants. As I hope to have shown in the above analysis. we need even more spread in our linguistic description. sequential organization. is based on work done in my research project entitled Interactional Practices and Linguistic Re- sources of Stance Taking in Spoken English (2002–2006). turn-taking. assessments. The project is greatly indebted to our collaborator John Du Bois. and prosodic resonances between contributions by different speakers (especially when I guess occurs in second position). An action. Finally. In light of my findings. however. This paper. who has provided us with a view of stance as achieved out in the social world and in interactions with other social agents. example (2) above: I guess that means his bro- ken leg is @doing @okay versus That means his broken leg is doing okay).. and Mirka Rauniomaa. In other words. such as an assessment or a claim. in that speakers only use a rather small set of inher- ently stanced words with some frequency. My heartfelt thanks are also due to Robert Englebretson for making the 10th Biennial Rice University Linguistics Symposium happen. and proposals) being done in the associated utterance. on one hand. Notes 1. Thompson (2002) argues that fragments like I guess pro- vide a certain type of perspective or stance toward the actions (i. and repair organization. semantic. The frame I guess is an essential part of the action being ac- complished in the turn. As we saw in Section 3. Tiina Keisanen. as also those by Pentti Haddington. and makes more explicit the kind of stancetaking action that is being done in the current turn. however. if we approach stance as something jointly oriented to by the co-participants. we turn our focus to stan- cetaking as a process and an activity oriented to by participants.e. which are also to be seen as resources for stancetaking and therefore deserve our full attention. counterclaims.2. I have demonstrated some syntactic. as well as for . some linguistic practices of stan- cetaking go beyond specific. we can see that some of these frequent markers then develop routine functions as organizers of such stancetaking activity and do not in themselves express a clear. preference organiza- tion. and has been financed by the Academy of Finland (grants 00381 and 53671). We have seen. but it would then constitute an essentially different type of assessment or claim (cf. can be well accomplished without I guess.

very specific subset of these options. Thanks also to all the participants of the Rice symposium for many helpful comments on my paper. This finding is nicely corroborated by Precht’s work within the “appraisal” framework (cf. emphases mine) 3.214 Elise Kärkkäinen many invaluable comments on earlier versions of this paper.’ mai tea ‘I don’t know’. For Swedish. 6. and commitments. minun mielestä ‘in my opinion’ and minusta ‘I think. and it is in many parts not ideally suited for close interactional analysis. I guess is often translatable with just a particle to other languages. and to the two anonymous referees for making many important points and helpful suggestions for revisions.e. But see for example Quirk et al.115 and 1. to form a computer-searchable syntactically tagged and parsed database. creer ‘be- lieve’ and pensar ‘think. I myself am of course responsible for what use I have made of the comments and suggestions I have received. is shaped by culture and custom – we are socialized to use particular stance markers in particular ways. 8.. Dahl (2000) reports on the high frequency of first-person pronouns (but also second-person pronouns and generic pronouns) clustering with mental verbs like tro ‘believe. First-person subjects typically co-occur with mental verbs also in colloquial Finnish (mä luulen ‘I believe/think’) and Estonian (ma arvan ‘I think. 4. By contrast. and yet we use only about 150 words for ninety percent of our stance expression [. with the Finnish kai or the Swedish väl.]. you know (954). The frequency of the verb (or noun) guess turned out to be only very little higher than that for I guess (23).. as also for all the remaining inadequacies.e. Similar observations about the frequency and formulaic nature of such explicitly subjective collocations have further been made of many other languages. namely I don’t think (118). 2003): The resources of language enable a virtually unlimited number of ways in which we could express ourselves.400 different stanced words in English. my results suggest that we are culturally ‘pro- grammed’ to use a very limited. Keevallik 2003)..g. Thus. Upon closer inspection these high figures are mainly accounted for by the inclusion of other highly grammaticized discourse markers (in addition to I think and I don’t know).’ This could be taken as further evidence for its discourse marker status.’ etc.’ minnas ‘remember. The ICE-GB data are essentially collected for corpus linguistics purposes. Although we have a myriad of options for expressing our emotions.713 respectively) were considerably higher than the ones for I think and I don’t know (729 and 309). 5. the total frequencies for the verbs think and know (1. This extremely high frequency is said to be largely due to the use of the clause I think. we tend to use the same small set of stance markers repeatedly. both meaning ‘probably.’ only displays a first-person subject but no verb (Rauniomaa this volume). i. e. 2.’ tänka ‘think.’ in spoken data. attitudes. . 7. with less attention to transcrip- tion detail or quality of sound. My previous analysis found more than 1.. Weber and Bentivoglio (1991) discuss essentially similar discourse patterns of the Spanish verbs of cognition. and I know (224). i. Martin 2000.. maybe. and in a recent study Karlsson (2003) discusses the interactional uses of the epistemic stance marker jag tycker/tycker jag ‘I think’ in spoken Swedish data. I would argue.’ tycka ‘think. (1985: 1112–1113) who hold the opposite view.  (Precht 2003: 240. consid- ering “comment clauses” like I think to be subordinate to the rest of the sentence. however. Our expression of stance. even though one of the frequent Finnish epistemic stance markers.

13. This way of framing the upcoming story is very similar to a more explicitly evaluative device for prefacing stories. preci- sion. Nichols (eds. 2001. (1999) include under epistemic stance not just markers of certainty. Chafe (1986). D.W.” Paper pre- sented at the Euroconference on Interactional Linguistics. T. whereas Biber et al. “‘Surely you knew!’ Surely as a marker of evidentiality and stance. September 11.W.. Norwood. S.A. and Conrad. and Metzger. Department of Linguistics. 14.. Leech. “Stance and consequence in interaction.” Items like problem or a wonderful/terrible thing in a story preface offer a framework for interpretation to the recipients. University of California. J.R.. 261–272. Cumming. Where evidentiality fits in with epistemicity and which one is considered the superordinate category varies from one researcher to the next.” Functions of Language 8 (2): 251–282. Downing. 1993. subsumes epistemic mo- dality under evidentiality. and (10).. namely by what Goodwin (1996) has called “prospective indexicals. in fact for three of my ex- amples (6). The role of I guess in conversational stancetaking 215 9.” In Talking data: Transcription and Coding in Discourse Research. 2000. (9). “How types of lexical resources emerge from turn-construction. “Indexing stance: Reported speech as an interactional evidential.” Subjectivity in this case comes out of and presupposes intersubjectivity. 1999. The Longman Grammar of Spoken and Written English. Biber. NJ: Ablex Publishing Corporation.” Functions of Language 7 (1): 37–77. Dahl. Schuetze-Coburn. 2001. S. W. 2006. Edwards and M. 2003. Towards a Dialogic Syntax. W.” Paper given at the Langnet Sym- posium. 2000. Beach. 45–87.W. “An outline of dis- course transcription... Du Bois. A. 11. I thank Robert Englebretson for drawing my attention to this. I thank the anonymous reviewer for pointing out this obvious similarity to me. S. in fact. Ms. Belgium. and Paolino. London: Longman. Chafe.” Human Communica- tion Research 23: 560–585. actuality. 12. R. 10. Finegan. J. Lampert (eds. “Egophoricity in discourse and syntax. Hillsdale. “Claiming insufficient knowledge. University of Oulu. J.). September 16– 21. . “Evidentiality in English conversation and academic writing. 1986. I thank John Du Bois for the observation that it is possible to describe the use of I guess as “I’m discovering something about my own subjectivity as I speak. or the speakers engaging with other subjectivi- ties in conversational dialogue (see also Du Bois this volume).” Journal of So- ciolinguistics 10 (5): 569–595. Clift. W. NJ: Lawrence Erlbaum. S.” In Evidentiality: The Linguistic Coding of Epistemology. J. G. D. Spa. References Anward. Johansson. J. Chafe and J. Ö. Du Bois. Santa Barbara. who do not yet have access to the story but who are expected to respond to it in an appropriate way upon its completion. 1997. or discusses evidentiality as coding both the speaker’s attitude toward the reliability of knowledge and his or her source of knowledge or mode of knowing. and limitation but also of source of knowledge and of the perspective from which the information is given. This example is analyzed in more detail in Kärkkäinen (2003).). Du Bois. E.

2003. J. 1–12. 1990. Kärkkäinen. Meyer. Amsterdam: John Benjamins.” In Struc- tures of Social Action: Studies in Conversation Analysis. “A technique for inviting laughter and its subsequent acceptance/declina- tion. Selting (eds. J. Interaction between Speakers and Hearers. CD.). G. J.). H. 2002. S.” In Discourse Markers: Descrip- tions and Theory. Cambridge: Cambridge University Press.E. Potter (eds.. ”Parenthesis as a resource in the grammar of conversa- tion. Philadelphia: Linguistic Data Consortium.A. E. M. Thompson (eds. Ziv (eds. Psathas (ed.” In Re- thinking Context: Language as an Interactive Phenomenon. 2003. “Transparent vision. 2005.). C.” In Conversation and Cognition.). Department of English. E. Duvallon. C. M. 1979. Holquist. B. and Thompson. 1992. W. “Assessment and the construction of context. Dialogism. Vol. C. J. Cambridge: Cambridge University Press. A.. Haddington. Kärkkäinen. Chafe. CD. 1996. “Interactional units in conversation: syntactic. and pragmatic resources for the management of turns... Vol. International Corpus of English–Great Britain (ICE-GB). 2006. Ochs. University College London. New York: Academic Press. Discourse Marker and Stance Adverbial Variation in Spoken American English: A Corpus-Based Analysis. 45–74. dissertation. 2. Jucker. 1996. 1-2.A. S. A.W. A Function-Based Introduction. O. 1998. 1981.. C. 2000. Schegloff and S. N. Heritage.W. 2005. 1993.A. 134–184. J. P. M. “Social interaction and grammar. Survey of English Usage. Santa Barbara Corpus of Spo- ken American English. intonational. A.216 Elise Kärkkäinen Du Bois. S.L.M.” In Interaction and grammar. Goodwin. “Side sequences. 147–190.). “Discourse markers: Introduction. Northern Arizona University. 119–143. Bakhtin and His World. and Martey. 370–404. and Thompson. Atkinson and J. W. 1997. 294–338. NY: Irvington Publishers.” In Syntax and Lexis in Conversation.). London: Routledge. NJ: Lawrence Erlbaum. Thompson. and Goodwin. te Molder and J. T. E. Epistemic Stance in English Conversation: A Description of Its Interactional Functions. Mahwah. Du Bois. Amsterdam: John Benjamins. Part 1.” Text & Talk 26 (6): 699–731. A. C. with a Focus on ‘I think’. Schegloff and S. English Grammar. Jefferson.). Goodwin (eds. Sudnow (ed. and Routarinne. Y.” In Studies in Social Interaction. Tomasello (ed). “A change-of-state token and aspects of its sequential placement. Oulu Corpus of US / British Television Interviews from 1999–2004. S. Heritage (eds. Givón. D. C. Fox. G. 2003. S. Part 2. G.A. Jefferson. Thompson (eds. Ms. Amsterdam: John Benjamins. E. Philadelphia: Linguistic Data Consortium.” In Interaction and Grammar. 1984. Jucker and Y.). Cambridge: Cambridge University Press.L. Cambridge: Cambridge University Press. Goodwin. Meyer. M. New York. University of Oulu. Studies on the use of linguistic resources in talk- in-interaction. .). 79–96. Ford.A. “Cognition in discourse. Hakulinen and M. and Thompson. 299–345. 184–202. Chafe. Ph. Conversational Organization.. Helt. C. New York: Free Press. and Ziv. Department of English. Heritage.” In The New Psy- chology of Language: Cognitive and Functional Approaches to Language Structure.. Amsterdam: John Benjamins.H.A. Goodwin. E. “Stance taking in conversation: From subjectivity to intersubjectivity. 1972.” In Everyday language: Studies in ethnomethodology.D. Duranti and C. Ochs. E. Cambridge: Cambridge University Press. Santa Barbara Cor- pus of Spoken American English. Ford.

“Interactional uses for an epistemic marker: The case of ‘jag tycker’/‘tycker jag’ in Swedish. A. Potter. 1989. J. A usage-based account of the phonological reduction of don’t in American English conversation. London: Longman. Cambridge: Cambridge University Press.” In Structures of Social Action.” Text 9 (1): 7–25.” Language in Society 20 (3): 441– 458. Ochs. 2003. E.” Text 13 (2): 247–266. G. “Continuing and restarting.R. Schegloff and S.” In Evaluation in Text.” In Rethinking Linguistic Relativ- ity. Special issue: Negotiating Heteroglossia: Social Perspec- tives on Evaluation.A. “Language has a heart. B. J. J. Heritage (eds.” In The Contextualization of Language. Precht. Hunston and G. Estonian Finite Verb Forms in Conversation.). Lerner. Authorial Stance and the Construction of Discourse. New York: Oxford University Press. 52–133. “On the syntax of sentences-in-progress. 273–296. Lenk. U. “Discursive social psychology: From attitudes to evaluations. R. E. “Introduction. Studies in Conversation Analysis. Cambridge: Cambridge University Press.. G.” Melbourne Papers in Linguistics and Applied Linguistics 1: 5–23. Keevallik. 2003.). Cam- bridge: Cambridge University Press. Martin. 2004. 1991. “Extreme case formulations. J. J. Thompson (eds. 142–175. and Svartvik. Auer and A. Scheibman. J.. 1996. 1984.” In Interaction and Grammar. Quirk.” European Review of Social Psychology 9: 233–266. J. A.” Qualitative Sociology 15 (3): 247–271. 2003. “Stance moods in spoken English: Evidentiality and affect in British and American conversation. and Schieffelin. 1985.). “‘I’m a bit concerned’ – early actions and psychological con- structions in a child protection helpline. A. “Finding ‘face’ in the preference structures of talk-in-interaction.). J. 1992. S. Potter.. “Agreeing and disagreeing with assessments: Some features of preferred/ dispreferred turn shapes.” Social Psy- chological Quarterly 59 (4): 303–321. Pomerantz. L. “Linguistic resources for socializing humanity. 1992.. “Beyond exchange: APPRAISAL systems in English. P. Kendon. Gesture: Visible Action as Utterance.” Human Studies 9: 219–230. E. Local. Lerner. 1998. “Assigning responsibility in conversational storytelling. Lerner.R. J. Leech.” Research on Language and Social Interaction 36 (3): 197–240. “Interpersonal activities in conversational storytelling. Acta Universitatis Upsaliensis. S. E. A Comprehensive Grammar of the English Language.” Western Journal of Speech Communication 53: 114–126. J. 2003.” Text. Greenbaum. 23 (2): 171–181. G. 1993. Cambridge: Cambridge University Press. Martin. Gumperz and S. 1986. Ochs. The role of I guess in conversational stancetaking 217 Karlsson. 2000. 1998. “Assisted storytelling: Deploying shared knowledge as a practical matter. di Luzio (eds. Tübingen: Gunter Narr. . G. and Hepburn. 1989. 2003. Special issue: Negotiating Heteroglossia: Social Perspectives on Evaluation 23 (2): 239–257. Pomerantz. S.” Text. E. “I dunno . Levinson (eds. Thompson (eds. Ochs. Atkinson and J. 1992. Schegloff.M. Marking Discourse Coherence. Amsterdam: John Benjamins. Mandelbaum.” Journal of Pragmatics 32: 105–124.). A. 2000. 1996. Mandelbaum. Studia Uralica Upsaliensia 34. From Interaction to Grammar. J. “Turn organization: On the intersection of grammar and interaction. 407–437. Functions of Discourse Markers in Spoken English. 57–101. K.

” In Marxism and the Philosophy of Language. MA: Harvard University Press. Hop- per (eds. Appendix: Symbols used in transcription (From Du Bois et al. The Evidence from Romance. “Subjectification in grammaticalisation. 61–89. A. “Verbal interaction. Chap. 83–98. “Local patterns of subjectivity in person and verb type in American English conversation. 2001.” Studies in Language 26 (1): 125–163. Cambridge: Cambridge University Press. S.A. “‘Object complements’ and conversation: Towards a realistic account. Continuing . 1993) Units Intonation unit {carriage return} Truncated intonation unit -- Truncated word - Transitional continuity Final . P. 194–213.218 Elise Kärkkäinen Scheibman. Stein and S. 1987. A. Cambridge: Cambridge University Press. Vol. Traugott and B. Schiffrin.). “The discourse conditions for the use of the complemen- tizer that in conversational English. “Verbs of cognition in spoken Spanish: A discourse pro- file. Weber. D.” In Discourse–Pragmatics and Verb. S. 1991. and Bentivoglio.). Appeal (seeking a validating response from listener) ? Speakers Speech overlap (numbers inside brackets index overlaps) [ ] Accent and lengthening Primary accent (prominent pitch movement carrying intonational meaning) ^ . Fleischman and L. Thompson. Amsterdam: John Benjamins. 1991a. Thompson. D. Cambridge. [1930] 1973. Wright (eds. “A quantitative perspective on the grammaticization of epistemic parentheticals in English. 1995.” In Subjectivity and Subjectivisa- tion: Linguistic Perspectives. Discourse Markers.A. Bybee and P. Scheibman. V. and Mulac.A. Waugh (eds. E. Thompson. J. J. 313–329.” In Frequency and the Emergence of Linguistic Structure. 2002. Traugott. Point of View and Grammar: Structural Patterns of Subjectivity in American English Conversation.). Voloshinov. London: Routledge. 31–54. J. 3. S.). 1991b. Amsterdam: John Benjamins.N. E. Heine (eds.C. 2002. E. S.” Journal of Pragmatics 15: 237–251. 1–2. Amsterdam: John Benjamins.” In Approaches to Grammaticalization.C. and Mulac.

The role of I guess in conversational stancetaking 219 Secondary accent ` Unaccented Lengthening = Pause Long and medium (length indicated in seconds) …(1. (SNIFF).g. (YAWN)..g. @two @words) Laughter during speech (+6 words) @ (e. Vocal noises e.0) Short (brief break in speech rhythm) ..g. <@> words </@>) Quality Special voice quality <VOX>words</VOX> Forte: loud <F>words</F> Higher pitch level <HI>words</HI> Lowered pitch level <LO>words</LO> Parenthetical prosody <PAR>words</PAR> Allegro: rapid speech <A>words</A> Marcato: each word distinct and emphasized <MRC>words</MRC> Yawning <YWN>words</YWN> Transcriber’s perspective Uncertain hearing <X> words </X> Uncertain #word Indecipherable syllable X . (TSK). (DRINK) Glottal stop % Exhalation (Hx) Inhalation (H) Laughter (one pulse) @ Laughter during speech (1–5 words) @ (e.


have drawn analysts’ attention before but mainly ended up in side comments and footnotes.e. affec- tive stances. Indeed. particularly in the context of assessments. Finland 1. a closer look at the expressions is in order. however. Selting and Couper-Kuhlen 2001). participants’ joint construction of evaluations. Kärkkäinen 2006. inner loca- tive. Stance markers in spoken Finnish Minun mielestä and minusta in assessments Mirka Rauniomaa University of Oulu. I argue that their use stems from participants’ need to communicate how the explicitly evaluative action that they are performing fits in the ongoing interaction. Ford et al. I will also briefly introduce the data used in this study (Section 2).g. other translations into English include ‘in my opinion.. Before proceeding to a discussion of stancetaking.’ and ‘it seems to me.. etc. attitudes. which consists simply of the first-person-singular pronoun minä and the elative case ending -stA (the capital A reflects Finnish vowel harmony: depending on the vowels in the stem. i. in dialogic interaction (Du Bois this volume.. e. which frequently occur in almost any Finnish interaction. i.’ Similar meanings can be attributed to the expres- sion minusta. In what follows. The expression minun mielestä consists of the first-person-singular pronoun minä in its genitive case and the noun mieli ‘mind’ in its elative.’ ‘to my mind. this volume). Introduction1 The expressions minun mielestä and minusta. I will discuss minun mielestä and minusta in terms of their morphosyntactic form and with relation to previous analyses. The literal translation of minun mielestä would thus be ‘from my mind’ (or ‘out of my mind’).e. The purpose of this paper is to contribute to the understanding of the two expressions by focusing solely on their locally situ- ated use. The central theme pursued in this paper is the notion of stancetaking. the analysis is guided by the idea that social interaction organizes linguistic phenomena. 2002. 1996. as posited in interactional linguistics (see. Ochs et al. the ending takes either the back vowel a or . case.

Lampinen 1990: 84–85. Nikula 1996: 115–121. for ob- servations on the diverse distribution of the expressions in spoken and written academic discourse). I refrain from the predetermined categories of politeness and carefully examine minun mielestä and minusta in the specific sequential contexts in which they occur. The current paper builds on this foundation to bring in the intersubjective aspects of stancetaking. subjective stances. 139–140. minun mielestä and minusta have commonly been de- scribed as hedges that are used to obscure the import of an utterance in order to address issues of politeness. more importantly.222 Mirka Rauniomaa the front vowel ä). to the extent that minun mielestä may come close to sounding like minusta. but I will not explore the phonetic realizations of the expressions in any detail here. the expressions de- limit the generalizability of an utterance by framing it not as a fact but as an opin- ion that the co-participants may disagree with (Luukka 1992a: 108.e. Aalto (1997: 65) points out that whether or not such a delimitation in effect strengthens or weakens the speaker’s commitment to a proposition is entirely context-dependent. consequently. Literally minusta would translate as ‘from me’ (or ‘out of me’). It is worth pointing out. that the structural similarity of the phrases and the possible phonetic reduction of minun mielestä suggest that minusta can be regarded as a contraction of minun mielestä. 1989: 118–119.. an expression that is more or less equivalent to the Finnish minun mielestä and minusta (see also . drawing on Prince et al. they consider the two expressions to mark speakers’ individual. Luuk- ka 1992a: 91–92. and because of data limitations. however. the present paper parallels Kärkkäinen’s (this volume) exploration of I guess and. 1989: 118–123. In this way. In previous literature. More spe- cifically. saving the speaker’s or a co-participant’s face (Aalto 1997: 58–61. In this way. A comprehensive analysis would certainly be useful. 1992b: 373. Hakulinen et al. Nonetheless. i. 1982: 85).’ Both expressions are often reduced phonetically. the most appropriate translation is ‘I think. whenever the analysis permits. her (Kärkkäinen 2003) study on I think. only touch upon such individual expressions as mi- nun mielestä and minusta. previous studies provide extensive overviews of linguistic modality in general and. What is more. As an example. 1992b: 372– 373). 135–136). 108. I keep the two expressions separate to encourage further investigations into their pos- sibly differentiated use (see Luukka 1992a: 118–119. I do not attempt to distinguish between them at this point. For this reason. and minusta may take an even more reduced form. Aalto argues that a turn-initial minun mielestä strengthens the expression of opinion and the speaker’s commitment to it when it is placed in a turn in which the speaker gives grounds for his/her dissenting opinion. 1992b: 363–373. they have been considered to represent a particular type of hedge that mitigates speakers’ commitment to the truth of a proposition (Hakulinen et al. Overall.

I will provide more evidence of the contingent nature of the stance markers by presenting a special case in Section 6. As access to knowledge plays a major part in some cases in the present data. Section 5. As Section 5 will show. before a discussion of the findings in the conclusion (Section 7). I will then give a brief overview of the stance markers’ positioning within an intonation unit and in the larger context of a sequence (Section 4). After a presentation of the data in Section 2. The bulk of the paper. In addition to a free English translation. I will discuss stancetaking and assessments in some detail in Section 3. Stance markers in spoken Finnish 223 Karlsson 2003 for a discussion of the Swedish jag tycker/tycker jag ‘I think’). Nonetheless. They amount to approximately two and a half hours. of casual conversation: there are 2 face-to-face conversations. ranging in length between 1 and 14 minutes. 1992. presents the three inter- actional functions of the stance markers that can be found in the data: projecting disagreement in a second assessment. Du Bois et al. and marking transition to a first assessment within an extended turn. but for the present study. I think marks topical and other boundaries. brings in a personalized speaker perspective. This paper proposes that the routinization is also evident at a sequential level. each lasting for about an hour. marking transition to a first assessment across turns. Data The data are drawn from the corpus of conversational Finnish maintained by the Department of Finnish Language and Literature at the University of Helsinki (Keskusteluntutkimuksen arkisto).000 words. 2. I have re-transcribed the relevant parts according to the intonation-unit-based discourse transcription system (Du Bois et al. or around 34. minun mielestä and minusta can be seen to serve similar interactional functions. and 12 telephone conversations. 1993). displays on-line planning. The remarks that have been made on minun mielestä and minusta in previ- ous literature suggest that the two items offer a routinized way for speakers to put their stamp on an utterance. signals turn comple- tion or pursues a response from the recipients. the starting point of this paper is not primarily on epistemicity per se but simply on answering the question why minun mielestä and minusta are used in the context of assessments. Kärk- käinen (2003) identifies several local purposes for the use of I think: depending on its sequential positioning. minun mielestä and minusta could be further cat- egorized as epistemic stance markers. I have . The term stance marker best captures this nature of the expressions: it implies that minun mielestä and minusta frame an utterance or a turn as relevant in terms of stance. The data were originally transcribed using the transcription conventions developed by Gail Jefferson (as laid out in Seppänen 1997).

it-gen cat_name-gen mind-ela it be:3sg always just front-ine ‘In his opi.in Perttu’s opinion it is always in his way. but be:3sg-q-cli these now you-gen mind-ela then good-pl ‘But are these ones good in your opinion. mi- nun mielestä ~ mun/miun mielest(ä). It should be pointed out that the speakers in these data come mainly from southern and eastern Finland. I use the full forms minun mielestä2 and minusta. In referring to the stance markers in general. there is only one case of a second-person reference (example 3) and one case of a third-person reference. Examples (1) and (2) typify the expressions minun mielestä and minusta. (1) sg067_a2 Lawyer Issues <T:00:03:35> R: Se on mu-n miele-stä tosi hyvä kirja. minusta ~ must(a)/miust(a)) and possibly of some other words in the transcripts.g.224 Mirka Rauniomaa added a morpheme-by-morpheme gloss for each line of transcript. 67 occurrences of the stance markers can be found in the data.g. but I deem it important that the evident dialect differences be reflected in the transcripts. e. respectively. This explains the varied spellings of the stance markers under examination (e. which represent two distinct dialect areas. such cases are rare in comparison with the first-person-singular forms. (3) sg151_a Chat about Something <T:00:34:53> I2: Mut on-k-s nää nyt siu-n miele-st sit hyvä-t.’ (4) sg067_a2 Lawyer Issues <T:00:02:34> R: Se-n mie- Pertu-n miele-stä se on aina just ede-ssä.’ .’ (2) sg151_a Chat about Something <T:00:37:33> I2: Mut ol-i miu-sta ne hyvä-n-näköse-t. Keys to rel- evant transcription and glossing symbols are provided in Appendices A and B. it be:3sg I-gen mind-ela really good book ‘In my opinion it is a really good book. In the present data. with another personal pronoun or with a proper name.. but be-pst:3sg I-ela they good-gen-looking-pl ‘But I think they were good-looking. in which the speaker refers to a cat called Perttu (example 4). minusta be- ing represented by a total of 25 cases. All in all. Minun mielestä is somewhat more frequent with a total of 42 cases.’ Although it is possible to use mielestä or simply the elative case ending -stA with other person references.

which gradually rises out of their contributions. in spoken Finnish. certain action types can be considered to be particularly rele- . This observation corresponds with Helasvuo’s (2001a: 32) finding that first-person subjects typically co-occur with mental verbs.. to talk from their particular point of view (see Bybee and Hopper 2001). Kärkkäinen 2003. that even this personalization. Du Bois emphasizes the significance of the stance differential. Scheibman 2001 for English. To put it differently. speakers leave in it traces of their attitudes. their specific utterances are affected by the fact that one precedes. Stancetaking through assessments It is a well-established observation that when producing an utterance. It has to be added. Stance markers in spoken Finnish 225 Examples (3) and (4) represent individual cases. i. Almost any utterance helps to locate the speaker’s perspective or point of view among an infinite set of possible ones. and aligning with other subjects. emotions. can also be seen as indices of the strong tendency in interactive discourse for speakers to personalize what they are saying. The Finnish minun mielestä and minusta. as adverbial noun phrases in- cluding an explicit reference to cognition and/or the speaker. follows. and so on. or verbs of cognition. the stance-marker elements mielestä and -stA are overwhelmingly used with first-person reference in the cur- rent data. i. however.e. Stancetaking is a social activity that is not restricted to any specific action type. I exam- ine the linguistic items that are the focus here. The interactive nature of stancetaking will be discussed next. or subjectivity. toward whatever they are talking about (see Thompson and Hunston 2000 for a review). Weber and Bentivoglio 1991 for Spanish).. Similar collocations between first-person referents and stance- inferring expressions in subject-verb combinations have also been observed in other languages (see Dahl 2000 for Swedish. Following this line of study. minun mielestä and minusta. within their local sequential contexts and consider them to arise contingently from those contexts. What is new is the shift in focus from individual stances to stancetaking. a subtle difference in the participants’ align- ment with each other. Within this view of stance.e. The interactional nature of stancetaking is well spelled out in Du Bois’s (this volume) definition of stance as being composed of three interlinked elements: evaluating an object. minun mielestä and minusta are here regarded to acquire their status as stance markers only in and through talk-in-interaction. or overlaps the other. 3. positioning the self. Keevallik 2003: 74–100 for Estonian. from single-speaker contributions to the joint construction of stance at both linguistic and interac- tional levels (see Wu 2004: 3–19 for a review). can only be constructed in interaction with other participants. even if participants fully agree on a matter. In other words. Nonetheless.

4. In other words. accomplishing orientational work. minun mielestä and minusta frequently occur within assessments. a pattern that can be presented as follows (slightly modified from Goodwin and Goodwin 1987: 22): [third-person pronoun] + [copula] + [adverbial intensifier] + [assessment term] As is evident in this syntactic pattern. The frequent IU-initial positioning of stance markers thus reflects the fact that they are an integral part of the utterance that they occur in. In the case of the English I think. Table shows that there is a tendency for both stance markers. I will make a few remarks on the stance markers’ possible positions in the sequence of actions. as the following sections will show. I will examine where the stance markers are placed within the intonation unit in which they oc- cur. and there is no need for an explicit reference to the speaker. and that they provide a frame for the whole upcoming clause or in some cases for a longer segment of talk (see Hakulinen et al. especially minusta.. the predicate and its core arguments) is typically produced within one intonation unit in Finnish. to be placed at the start of an intonation unit. assessments include an evaluation (assess- ment term) of a particular referent (third-person pronoun) and simultaneously mark the speaker’s position in relation to that referent and the co-participants.e. Firstly. The four possible positions that the stance markers can take in the intonation unit (IU) are presented in Table 1.. Helasvuo (2001b: 145–149) has found that the clause core (i. 1989: 136).226 Mirka Rauniomaa vant for stancetaking: assessments (as defined by Goodwin and Goodwin 1987: 9) are explicitly evaluative and as such clearly contribute to stancetaking.e. i. I will here discuss two levels of positioning. e. stance markers orient participants to treat . Secondly. with a special focus on stance markers within assessments. What is more. by choice of tense in the copula (Goodwin and Goodwin 1992: 165). Nevertheless.. It is therefore important to consider the immediate context of the stance markers minun mielestä and minusta. assessments typically follow more or less the same syntactic form as in English (Tainio 1996: 85). Kärkkäinen argues that its IU-initial positioning “helps recipients to align themselves to what is coming” (2003: 115). whether the stance markers occur in initiating or responding turns. Positions of minun mielestä and minusta The way in which a single utterance is constructed is consequential to its interpre- tation. Assessments clearly rep- resent the speaker’s perspective. In Finnish. assessments make relevant a type-connected uptake (Pomerantz 1984) and thus facilitate participants’ alignment with each other.g.

3 These include the negation element ei. with given in- formation preceding new information in a clause and especially with pronominal subjects preceding the verb (Helasvuo 2001b: 76–78). This suggests that IU-medial stance markers orient participants to what follows in much the same way as IU-initial ones. The two cases of minun mielestä in which the stance marker is placed in a separate intonation unit refer back to and cast new light on the speaker’s previous utterance. 1974: 722). or information delivery in a wider sense (Heritage 1984). In the present data. When the stance markers are positioned somewhere in the middle of an in- tonation unit. and which is syntactically tied to that question. niin ‘so’ and mutta ‘but’ are included within this group (see Hakulinen et al. 1989 for a discussion of discourse particles in Finnish). it is common for both stance mark- ers to occur in assessments: 20 cases of minun mielestä (48%) and 9 of minusta . Motivation for such use of the stance markers is also inextrica- bly tied to the ongoing interaction. It is striking that in all the cases in which a stance marker is placed IU-medially in an assessment. and the favoured word order is SVO. Additionally. either in the same intonation unit or in the subsequent one. and assessments. both stance mark- ers can be in turns that initiate new sequences or respond to a previous action. As for their possible positions in a somewhat larger context. there is only one case of an IU-final stance marker in the data. which typically come in pairs of first and second assess- ments (Pomerantz 1984). As Table 1 reveals. as example (6) will later demonstrate.’ which is produced as an answer to a previous speaker’s question. what follows as possibly requiring a response and in any case relevant to their stancetaking.4 the verb on ‘is. or a combination of these. and minusta in the intonation unit (assessments in parentheses) Initial* Medial Final Separate Total minun mielestä 24 (13) 15 (6) 1 (0) 2 (1) 42 (20) minusta 20 (5) 5 (4) 25 (9) Total 44 (18) 20 (10) 1 (0) 2 (1) 67 (29) * Cases in which a stance marker is preceded by a discourse particle such as no ‘well’. the stance marker can be considered to be especially relevant in making clear the current turn’s relation to the previous one (see Sacks et al. only a limited set of items can typically be placed in front. Their occurrence at that particular slot can be explained by the syntactic constraints of spoken Finnish: the negation element is commonly placed at the start of an intonation unit. Two types of action that can take place in either of these positions figure promi- nently in the data: statements. It occurs in the utterance no ei mun mielestä ‘well not in my opinion.  Position of stance markers minun mielestä.’ pronouns. the evalu- ative element comes only after the stance marker. Stance markers in spoken Finnish 227 Table 1. if an IU-initial position coincides with a turn-initial one.

e. the stance markers can be seen to perform three different functions: projecting disagreement in a second assessment. when there are clearly evaluative elements either within the intonation unit itself or in an adjacent intonation unit within the same turn. and marking transition to a first assessment within an extended turn. They do not in themselves reveal speakers’ stances but point out how an utterance is to be un- derstood there and then. Hakulinen et al. however. 5.. An interesting question can then be raised: What is it that makes stance markers relevant in assessments? Gener- ally put. Kaarina reports on what someone has said about people who come from Savo. Among the assessments in the present data. this function is relatively rare. Minun mielestä and minusta orienting participants to stancetaking Section 4 has shown that in the present data minun mielestä and minusta are of- ten either embedded in or followed by an assessment. the area in eastern Finland that Reija is from. i. This is the case especially when the stance markers occur within assessments.1 Projecting disagreement in a second assessment The few remarks that have been made on minun mielestä and minusta in previous research mainly deal with cases in which these stance markers project different types of disagreement (Aalto 1997: 65. Kaarina and Reija are discussing the characteristics of people coming from different parts of Finland. . More specifically. In example (5). marking transition to a first assessment across turns. I present it here first because of its clear trajectory: a speaker makes a first assessment. 1989: 136). 5.228 Mirka Rauniomaa (36%) are produced within an utterance accomplishing such an action. and an assessment clearly contributes to stancetaking already on its own. These observations quite clearly indicate that the stance markers minun mielestä and minusta are made relevant by the ongoing interaction. they orient participants to the stancetaking. That is to say. the stance markers are frequently used in a context that can also otherwise be identified as being about stancetaking. and a recipient makes a disagreeing second assessment that includes a stance marker. I will examine such cases in detail in the following sections.

.3) 3 et niinku. apparently it prt ‘E..sano vaan et ‘then he j.’ and realized as a that-clause on line 4.view like.’ 4 (H) et ne on helv[eti-n <F>ka</F>]teellis-i-i ihmis-i-ä.’ 5 R: [(SNIFF)] 6 .4) 7 <HI>Mu-n miele-st se</HI> on paska-puhet-[2ta2].’ 8 K: [2Se2] sano et se on niinku (H) -- it say:pst:3sg that it be:3sg prt ‘He said that it is like (H) --’ 9 E- ilmeisesti se nii[3nku3].(0. I-gen mind-ela it be:3sg shit-talk-ptv ‘In my opinion it is bullshit.apparently it like.just said that. Nonetheless. that they be:3sg hell-gen jealous-pl-ptv person-pl-ptv ‘(H) that they are damn jealous people..’ 2 . somehow come-3sg into.8) 13 R: Nii joo...’ 12 . sit se v.. Stance markers in spoken Finnish 229 (5) sg067_a1 Lawyer Issues <T:00:09:11> 1 K: Sit se v- sano vaan et. prt prt ‘Oh yeah. that prt ‘that like.’ The report does not convey the speaker’s own stance toward the matter at hand.that-ine small-town-ine ‘.’ 10 R: [3(SNIFF)3] 11 K: . jotenki tule-e esiin tollase-s pikku-kaupungi-s.. et ne on helvetin kateellisii ihmisiä ‘that they are damn jealous people. the report is stance-inducing in .. then it say:pst:3sg just that ‘Then he j.(0. somehow comes up in a small town like that.just said that.(0.’5 Embedded in Kaarina’s report is an assessment made by a third party: it is intro- duced with a reporting clause on line 1.

The stance marker does special work in projecting disagreement: it shows the speaker’s sensitivity toward the fact that. which is here referred to with the pronoun se ‘it. In other words... by the recipi- ent (see Heritage 1984).4-second pause on line 6. 1989: 136). The first syllable of kateellisii ‘jealous’ is produced with relatively loud voice.. Had Reija not used the stance marker on line 7. The fact that Kaarina from line 8 onward defends the reported speaker by giving an account for the third-par- ty assessment suggests that she may sympathize with both Reija’s and the reported speaker’s stances. making it hard for Kaarina to differ on the issue. as reported in earlier remarks on the stance markers (e. on the other hand. Starting the disagree- ing second assessment with a stance marker. minun mielestä indeed has a softening or modifying effect. which may be interpreted as a means of stall- ing disagreement (Pomerantz 1984: 65). i. a stance has already been reported on.e. rather than fully-fledged second assessments (Tainio 1993: 153–154. the assessment would have sounded uncompromising and left Kaarina with little opportunity to continue: especially because Reija represents the group of people that has been evaluated. another stance is yet to be taken. which highlights the point of the utterance and the relevance of an uptake.e. Kaarina’s turn accomplishes two actions that Reija can respond to. or particle and finite verb combinations.g. which in this case would not be necessary were the assess- ment displaying agreement (see Aalto 1997: 65. Reija does not respond immediately after Kaarina’s turn nor in overlap with it. Firstly. Hakulinen et al. Reija deals with both by producing an assessment of the reported assessment. A response from Reija is particularly relevant here as she represents the group of people that has been assessed.’ There are several features in Reija’s turn that project disagreement. On line 7. a plain assessment by her would have borne a sense of authority and finality. as reported in Tainio 1996: 108). Aalto 1997: 65. and. instead there is a 0. . 1989: 136). Thirdly. it can be understood to be an atypical and possibly dispreferred response (see Pomerantz 1984 for a discussion of preference in assessments). Here. a display of the telling’s impact. This suggests that as soon as the utterance is recognized to consist of something other than such a particle. Reija leaves Kaarina more room to manoeuvre between the reported stance and the one that Reija takes. i..e. i. in Finnish.’ The predicate noun paskapuhetta ‘shit talk’ or ‘bullshit’ marks negative evaluation of and strong disagreement with the third-party assessment.230 Mirka Rauniomaa that it makes relevant an uptake. the stance marker acknowledges the existence of several possible stances and locates the speaker’s stance among them. Hakulinen et al. mun mielest se on pas- kapuhetta ‘in my opinion it is bullshit. the report as such and the assessment within it. in Reija’s producing mun mielest. Secondly. first assessments are usually responded to by particles. on the one hand. the stance marker mun mielest projects contrast by explicitly marking the start of the speaker’s viewpoint.

However. Kaarina and Reija talk about Joni.) (6) sg067_a1 Lawyer Issues <T:00:16:33> 1 K: Se on jotenki semmonen vähän ällö.. Their status is further decreased by the fact that in the case presented as example (5) a second assessment is produced to disagree with a re- ported speaker. Reija has thus proved to be the more informed participant on this topic by the time the participants get to the segment presented here..’ . I-gen mind-ela ‘in my opinion. actually when it-ill acquaint-3sg better ‘actually when Ø gets to know him better. In my opinion he has been damn disgusting helveti-n ällö mut. hell-gen disgusting but for a long time but.’ 6 ((RATTLE)) 7 R: oikeestaan ku sii-hen tutustu-u paremmin. Stance markers in spoken Finnish 231 The amount of disagreements that are framed with the stance markers is lim- ited in these data. Here.B.6 (N. The main focus is on the stance marker on line 5.’ 2 <P><%>mu-n mie[le-stä</%></P>]. Kaarina has then asked several questions about him. but I will also briefly com- ment on the one on line 2. with whom they are both acquainted. the stance marker mun mielestä proj- ects disagreement with a co-participant. Kaarina has introduced Joni into their talk a little earlier by asking Reija to confirm a piece of information about his past whereabouts.’ 3 [((KNOCK))] ((KNOCK)) 4 R: (H) Mut se-ki on %m= -- but it-cli be:3sg ‘(H) But it/he is also m= --’ 5 .little disgusting ‘He is somehow a bit disgusting. and Reija has provided answers as well as some extra information about his social life. Reija’s first intonation unit is divided into two misaligned lines due to limitations of space. %Se on mu-n miele-stä ol-lu pitkään it be:3sg I-gen mind-ela be-pcp long ‘. which Kaarina has marked as news to her. there are also cases in which the assessment sequence is more straightforward: in example (6). it be:3sg somehow such a.

unlike in Kaarina’s turn. Reija’s assessment is an upgrade of Kaarina’s assessment in three senses: Firstly.’ On line 1. the planning par- ticle jotenki ‘somehow’ (see ISK 2004: 821–822) and the demonstrative adjective semmonen ‘such’ (see Erringer 1996. The first as- sessment already in itself invites a subsequent contribution. Reija adds a timespan into Joni’s disgustingness by using the present perfect form on ollu ‘has been’ and the adverbial pitkään ‘for a long time. as cited in Helasvuo 2004: 1327). but the added stance marker puts even more emphasis on the relevance of a response from a co-partici- pant who clearly has better access to the assessable. Kaarina makes an assessment of Joni: se on jotenki semmonen vähän ällö ‘he is somehow a bit disgusting. and on line 5. Reija’s mun mielestä responds directly . 2 mun mielestä. Reija redesigns her utterance into a second assessment that seemingly agrees with Kaarina’s. the intonation unit is truncated.’ There is some material. i. which seem to signal hesitation and to give Kaarina time to plan her speech.’ Kaarina’s and Reija’s turns can be mapped onto a diagraph. On line 2. On line 4. Reija inten- sifies the evaluative adjective ällö ‘disgusting’ with the amplified adverbial helvetin ‘damn’ rather than the toned-down vähän ‘a bit.e.. However.’ Secondly.’ The stance marker mun mielestä in Reija’s utterance on line 5 attributes the upgrades to the current speaker and highlights the emerging stance differential. Diagraph 1. 5 R: Se on mun mielestä ollu pitkään helvetin ällö mut. Reija starts an utterance with the particle mut ‘but. Diagraph 1 shows how Reija recycles elements from Kaarina’s turn.’ which projects contrast (see Sorjonen 1989: 176). it seems evident that Kaarina uses mun mielestä to mark her assessment as a possibly disputable one. Kaa- rina then adds the stance marker mun mielestä as a separate intonation unit and in softer speech. there are no signs of hesitation in Reija’s turn. This is one of the two instances in the data in which the stance marker minun mielestä is added after the proposition as an intonation unit of its own. se on mun mielestä ollu pitkään helvetin ällö mut ‘in my opinion he has been damn disgusting for a long time but. Taking into consideration that Reija has a moment earlier been identified as the more informed participant on the topic. which is used “to create a representation of the structure of resonance relations that arise between utterances” (Du Bois 2003a).232 Mirka Rauniomaa 8 ni se ällöys vähene-e. Recycled and upgraded elements in (6) 1 K: Se on jotenkin semmonen vähän ällö. prt it disgustingness diminish-3sg ‘the disgustingness diminishes. Thirdly.

the stance markers are placed at the start of a disagreeing second assessment. Here.2 Marking transition to a first assessment across speakers’ turns The stance markers minun mielestä and minusta are also found in contexts in which a speaker makes an assessment about a topic that has not been evaluated or has only been implicitly evaluated in the talk so far. he does not seem so disgusting anymore. who calls certain forward-project- ing verbs. such as thought and wanted to. but no longer. In this light. It signals that Reija is not going to take up the position that is already avail- able but will evoke another one. ‘first verbs’). More often. between speakers).e. and one in which there is a transition within a speaker’s extended turn. Reija goes on to specify the new state and argues on lines 7 and 8 that once better acquainted with Joni.’ to imply that anyone. As in example (5). Reija’s assessment can be seen as a report of her once. In sum. Stance markers in spoken Finnish 233 to Kaarina’s use of the stance marker by heightening the resonance between the participants’ turns and consequently draws attention to the differences in their turns. The use of the present perfect on ollu ‘has been’ can now be interpreted as projecting Reija’s changed state (see Sacks 1992: 181–182. including Kaarina.. the stance markers are used at points of transition from non-evaluative or implicitly evaluative to explicitly evaluative talk. The contrast that was implied by mut ‘but’ at the start of line 4 is reinstated at the end of line 5. oikeestaan ku siihen tutustuu paremmin ‘actually when Ø gets to know him better.3. the two minun mielestä in example (6) orient the participants to the stancetaking: the first one by anticipating alignment and the second one by re- sponding to that anticipation. as in Section 5.e. Such cas- es form the majority of the data and can be divided into two equally-sized groups: one in which the transition takes place across speakers’ turns. When the shift takes place across turns (i. It is worth noting that Reija uses the zero- person construction on line 7. as in the next set of examples. however. may in the long run change their view of Joni (see Laitinen 1995 for a discussion of the zero-person construction in Finnish). they subtly project disagreement in a context in which an explicit evaluation has already been made. the present data do not include any agreeing second assess- ments that would be framed with minun mielestä or minusta.. Such framing does not seem to be necessary when participants’ stances converge: indeed. the second stance marker in ex- ample (6) shows the speakers’ sensitivity toward the existence of several possible stances. i. the stance markers can be used to signal the . 5. having taken a similar stance as Kaarina takes now.

(0.. In example (7).off-inf-ill ‘Ø starts coming off. Ira_1 explicates to her co-participants how one can boil water and breathe in the warm steam in order to clear one’s respiratory system when suffering from a cold.’ 8 E: [lähte-e <X>irki</X>]. I-ela it be:3sg awful-gen nasty be-inf somewhere ‘(H) I think it is awfully nasty to be in some (H). (7) sg151_b2 Chat about Something <T:00:08:47> 1 I1: Mut si-tä pitää niin kauan (H)..’ .X such a hot kettle. such-gen hot-gen kettle-gen ‘ä.8) 11 I2: (H) Mu-st se on hirvee-n inhottava ol-la jossain (H).to so long ‘But Ø has to keep on (H). hold-inf or breathe-inf and ‘holding or breathing it and.(0.’ 6 .234 Mirka Rauniomaa start of a new action and to prepare the participants for relevant next actions that further contribute to the stancetaking. always boil-inf new-acc water-acc that ‘always boil fresh water until. come-3sg loose ‘Ø comes loose. feel-3sg that here-abl prt ‘Ø feels that like out of here.. there-abl come-3sg such-pl-ptv lump-pl-ptv ‘these lumps come out of there. but it-acc have.’ 10 .’ 2 .3) 7 [rupe-e irto-o-n.3) 3 pitä-ä tai hengittä-ä ja..’ 5 tunte-e et tää-lt niinku. start-3sg come.’ 9 I1: sie-lt] tule-e semmos-i-i kökkö-j-ä..(0..’ 4 aina keittä-ä k_uut-ta vet-tä et.’ 12 ä- <A>semmose-n</A> kuuma-n katti[la-n X.

’ 15 I1: [2Mm2]. In- deed. On lines 11–12. Ira_2 makes an assessment that is on the topic but provides a slightly different aspect to it: must se on hirveen inhottava olla jossain ä. The stance marker must at the start of Ira_2’s utterance on line 11 addresses the issue: it frames the assessment as being made strictly from one particular point of view and thus pre- empts potential disagreement. but simply acceptance of.8-second pause follows on line 10. Ira_1 gives a more detailed description of how the advice should be carried out. It also provides Ira_2 a way in to participating in topical talk.’ As the participants’ focus has so far been on the positive effects of steam breathing. once Ira_1 has finished her overlapping utterance on line 9. who is evidently suffering from a cold. Stance markers in spoken Finnish 235 13 I1: [Mm]. as well as Emma’s quiet response token joo on line 16 do not mark strong alignment with. a 0. In example (8) the same participants are discussing the benefits of going to a tanning salon before leaving for holidays to a sunnier and warmer part of the . On lines 1–7 and 9. The use of the zero-person construction on lines 1–5 makes the actor’s role available to Emma without explicitly attributing it to her (see Laitinen 1995): mut sitä pitää niin kauan pitää tai hengittää ja aina keittää k_uutta vettä et tuntee et täält niinku ‘but Ø has to keep on holding or breathing it and always boil fresh water until Ø feels that like out of here. although she has not been the main recipient of her co-participants’ previous turns and the sequence was coming to an end.’ When Ira_1 pauses to search for a word on line 6. Emma’s coming in at this point marks a successful receipt of the advice and implies readiness to end the sequence. and Emma’s alignment as the advice recipient. 14 I2: al]ka-a hiki va[2lu-a ja2] [3kai3]kke-e.’ The topic of steam breathing has started with Ira_1’s giving advice to Emma.X such a hot kettle. 16 E: [3<P>Joo</P>3]. This interpretation is reinforced by the fact that Ira_1’s minimal response tokens on lines 13 and 15. prt ‘Yeah. it is potentially precarious to evaluate it negatively. start-3sg sweat run-inf and everything-ptv ‘sweat starts running and everything. lähtee <X>irki</X> ‘Ø comes loose’ (see Helasvuo 2004 for a discussion of co-constructions in Finnish).sem- mosen kuuman kattilan X ‘I think it is awfully nasty to be in some ä. Ira_2’s stance (see Sorjonen 2001: 204–205 for a discussion of joo). Emma comes in on line 8 with a syntactically fitted continuation of Ira_1’s utterance.

Emma displays a kind of realization with the high-pitched nii joo ‘oh yeah. pari kolme kerta-a käy-ä-än ennen [ku lähe-tä-än].. couple three time-ptv go-pass-4 before than leave-pass-4 ‘.’ 7 I1: [<X><P>ettei ihan pala</P></X> (Hx)]. she goes on to state that she and her travel partner will nevertheless go to a tanning salon before their next trip (in order to get a more protective tan). et nyt. Emma uses the opportunity to give more weight to the point that she has been making about the necessity of using the sun bed frequently enough. Ira_1 has been told that she still has a tan left from her previ- ous trip.. In any case.’ 4 E: [<X><HI>Nii joo</HI></X>]. we go two or three times before we leave.’ 3 . . prt prt ‘Oh yeah.’ The realization is most likely induced by the fact that whereas Emma has been arguing for the benefits of the sun bed when one has not had exposure to the sun for a while. Emma has been particularly active in reporting on an article that she has read on the issue and in arguing for the need to use the sun bed several times for it to have any significant effect.. I-ela it be:3sg also quite sensible-ptv ‘I also think it is quite sensible.’ 5 ni se on ihan -- so it be:3sg quite ‘so it is quite --’ 6 Miu-st se on [kans ihan järke]vä-ä. on lines 1–3. that now. Ira_1 has here im- plied that going to a tanning salon may prove useful even when one still has an old tan. but so that ‘But so that. In overlap with the end of Ira_1’s utterance.236 Mirka Rauniomaa world.’ A moment earlier. that now ‘. that-neg quite burn ‘so that Ø doesn’t burn completely. (8) sg151_a Chat about Something <T:00:17:45> 1 I1: Mut silleen et..’ 2 .

There are two cases in the data in which minun mielestä and minusta mark the transition to a parenthetical assessment that requires only minimal uptake by a recipient (see Duvallon and Routarinne 2001 for a discus- sion of parenthesis in Finnish). The stance markers subtly deal with the risk of disagreement. but employing a stance marker helps them to display an understand- ing of the sequential context in which they produce the assessment. Speakers manage such transitions without using minun mielestä or minusta. reprimand. to an evaluative summary that invites more substantial contributions from the co-participants. that partici- pants take in moving from non-evaluative or implicitly evaluative to explicitly evaluative talk across turns. or implicitly evaluative talk by one speaker to an explicit evalua- tion by another. In the following section. while kans recognizes an implicit evaluation in the previous speaker’s turn (because people usually engage themselves only in what they consider sensible). an uptake is clearly expected after the assessment. 5. In the following two examples. the focus will be on transi- tions within speakers’ own talk. Reija and Kaarina are both lawyers and talk about work-related issues several times during the conversation. . The completed assessment on line 6 is more in line with the general progress of the sequence: miust attributes the assessment to the current speaker. which Du Bois 2003b calls an ‘intersubjective alignment marker’).3 Marking the transition to a first assessment within an extended turn The third function that the stance markers have in the present data is to mark a transition from non-evaluative or implicitly evaluative talk to an explicit evalua- tion by the same speaker.’ She repairs the assessment by adding the stance marker miust and the alignment marker kans (the Finnish kans is similar to the English too. Kaarina reports on how some of her col- leagues treat witnesses in court. non-evaluative.’ It is worth noticing that Emma starts to make her assessment already on line 5 but cuts it off before producing the assessment term. In examples (7) and (8). Stance markers in spoken Finnish 237 On line 6. however. shifts from non-evalu- ative or implicitly evaluative to explicitly evaluative talk occur in extended turns when a speaker moves from some kind of a telling that has received only minimal response tokens from the recipients. More typically. Emma assesses Ira_1’s proposed line of action in view of what she has said earlier: miust se on kans ihan järkevää ‘I also think it is quite sensible. the stance markers signal a transition from earlier. ni se on ihan ‘so it is quite. etc. Here.

preliminary.’ Kaarina starts a complaint sequence on line 1 by stating that she and her col- leagues have grossly divergent views on a court procedure. so-cli it be:3sg-cli ‘So it is. she specifies .’ 2 (TSK) (H) koska mä e-n ymmärrä si-tä että.’ 4 [((BANG)) ((BANG))] 5 K: ja sit lue-ta-an si-lle esitutkintapöytä[2kirja-a2].investigation-acc ‘and then (they) read him/her the record of the preliminary investigation.’ 13 R: Niin-hän se on-ki5]. like. Se on mu-n miele-st ihan per[5see-stä.. In my opinion it is really bollocks... because I neg-1sg understand it-acc that ‘(TSK) (H) because I don’t understand it that.’ 8 R: [3(SNIFF)3] 9 [4Nii4]. that ask-pass-4 witness in ‘.. On line 2.’ 11 R: [5Väärin. (they) ask the witness in. and then read-pass-4 it-all record.’ 6 R: [2(SNIFF)2] 7 K: ja kysy-tä-än [3että u3]udista-tte-ko [4tämä-n4]. but we-ade be:3sg like awful-ptv fight-ptv it-ela that ‘But we have like awful fights over the fact that..238 Mirka Rauniomaa (9) sg067_a1 Lawyer Issues <T:00:34:33> 1 K: Mut mei-l on niinku hirvee-t <X>kähmä-ä</X> sii-tä ku. että [.that matter ‘A thing like that.’ 10 K: .of.. it be:3sg I-gen mind-ela quite ass-ela ‘. that .’ 3 . pyyde-tä-än] todistaja sisään.’ 12 K: Tommonen asia5]. and ask-pass-4 that renew-2pl-q this-acc ‘and ask (him/her) that do you confirm this. prt ‘Yeah. wrong ‘Wrong.

On line 11. More specifically. Stance markers in spoken Finnish 239 her own take on the matter. (10) sg151_a Chat about Something <T:00:12:45> 1 I1: <HI>Ties-i-k-s Ira</HI> muuten et me vara-tt-i-in se know-pst-q-cli first_name by. following the common practice in Finnish. and then goes on to describe the contested procedure on lines 3. the stance marker projects an ex- plicit evaluation that invites further contributions. here.. Kaarina can be seen to seek support from Reija.’ The stance marker mun mielest is positioned in a transition point between im- plicit and explicit evaluation. in example (10). Kaarina ends her complaint on line 10 with an “overt expression of moral indignation” (Drew 1998: 309–311) by making an assessment that wraps up her complaint: se on mun mielest ihan perseestä ‘in my opinion it is really bollocks. Lappee-n …(0. Kaarina has by this time said enough for Reija to be able to produce an overlapping assess- ment term that syntactically fits in with Kaarina’s ongoing utterance.e.’ .way that we reserve-pass-pst-4 it ‘By the way Ira did you know that we reserved the . Similar to example (9). who is able to view the matter from the same professional perspective as Kaarina. cf. and 7.. such as I was so angry (Drew 1998: 311). It helps Kaarina to construct her own stance in rela- tion to and in contrast with all the possible ones that she has referred to a moment earlier. Reija produces her contribution in slight overlap with Kaarina’s utterance (see Pomerantz 1984: 69). Reija’s partly overlapping response token nii on line 9 implies that Kaarina has secured her alignment: “nii agrees with the prior talk by claiming recognition of what the co-participant is talking about” (Sorjonen 2001: 247). Reija displays recognition of the import of Kaarina’s utterance and reciprocates an expression of indignation.the. place_name-gen sauna .. sauna … at Lappee. 5. but the co-participants’ uptake is somewhat delayed. not understanding. It is worth noting that in English the expression of indignation at the end of a complaint sequence usually takes the form of a first-person assessment. i. the assessment is personalized with the stance marker mun mielest. The example starts with Ira_1’s pointing out to Ira_2 and Emma that she and some others have reserved a sauna. the overlap starts on the assessment term perseestä ‘from the ass’ or ‘bollocks’ rather than on the adverbial intensifier ihan ‘really’ (see Tainio 1996: 89. As is typically the case with agreements to assessments. Goodwin and Goodwin 1987: 22–24 for American English).4) sauna.

’ 6 E: [2Jo2]o. prt ‘Yeah. the cost of renting the sauna.. neg be ‘No it isn’t. This time. i. Ira_1 provides yet another piece of information.. but that they mark as known information.4) 11 [Koska otta-a huomio-on] että. prt ‘Yeah.’ On lines 1–6.time ‘at what time.’ 3 Kuul-i-n Tu[pu-lta].’ 5 I2: J[2oo2].’ 13 <X>monelta</X>].’ 4 I1: [Tapanin]päivä-ks. hear-pst-1sg first_name-abl ‘I heard from Tupu.(0. prt ‘Yeah. Ira_1 repeatedly provides her co-participants with pieces of infor- mation that might be news to them.’ 10 .e.240 Mirka Rauniomaa 2 I2: Joo.’ 12 I2: [Ei oo. On line 8. five ten-ptv cost-3sg three hour-ptv ‘Three hours cost fifty marks. neg I-ela it be awful-gen bad ‘I don’t think it is awfully high. Boxing. however.’ 7 ((CLANK))[=] 8 I1: [Viis kym]ppi-i maksa-a kolme tunti-a. because take-3sg account-ill that ‘Because if Ø takes into account that. at.Day-trans ‘For Boxing Day.’ 9 ei miu-st se oo hirvee-n paha. Ira_1 does not wait for any response from Ira_2 or Emma but continues on line 9 to assess the piece of information she has just .what..

Ira_1 starts on line 11 to account for the assessment that she made on line 9. It implies that Ira_1 makes the assessment specifically in the role that she is holding here. but it is somewhat obscured by the overlap. However. at the very same moment Ira_2 comes in with a second assessment on line 12. but by us- ing them speakers can orient their recipients to what follows. i. Ira_1 moves on to a logical next action and uses an explicit means to do so. the stance marker miust is here placed at a point in which it marks one participant’s explicit shift to stancetaking and signals an upcoming opportu- nity for the other participants to contribute. ei oo ‘no it isn’t. 6. Ira_2’s turn implies strong agreement with Ira_1’s (see Tainio 1996: 97–99). I will discuss this issue in more detail in the following section.e. As no such uptake ensues.’ So once it has become clear that the participants share the information about the sauna. Example (11) will provide extra evidence: more or less the same assessment is made first without and then with a stance marker (see Schegloff 1996: 192–199.’ Consisting simply of the negated predicate verb that Ira_1 used in her assessment. In sum. After the segment presented here. the participants continue to display their agreement although overlapping remains a potential problem for a while. Examples (9) and (10) have shown that the stance markers can be used at transition points from non-evalu- ation or implicit evaluation to an explicit evaluation of a state of affairs. who encourages analysts to pay attention to ‘eventful nonoccurrence’). The co-participants display recogni- tion of these functions in their belated responses. It is true that participants can make such shifts without using stance markers. Evidence from a special case: Stance markers reflecting interactional contingencies Previous examples have shown that speakers use the stance markers minun mielestä and minusta to orient their co-participants to the stancetaking. Ira_1’s assessment is followed by a 0. Kaarina and Reija have compared two possible ways of putting forward payment orders. Before the selected segment. especially if the speaker’s ongoing action will require recipients’ response. Stance markers in spoken Finnish 241 given. Kaarina then moves on to a slightly . as the one who has reserved the sauna and provided the others information about its price.4-second pause on line 10. The stance marker miust signals the transition and the upcoming relevance of further con- tributions. which would allow for an uptake by Ira_2 or Emma. The stance markers are made relevant by a need to mark stancetaking explicitly at some interactional moment. ei miust se oo hirveen paha ‘I don’t think it is awfully high..

’ 10 tiä-t-kö se teke-e se-n esitysehdotukse-nsa periaat[tee-ssa] know-2sg-q it make-3sg it-acc proposal-acc:poss principal-ine ‘you know he/she makes his/her proposal basically’ vaan sii-hen. those thing-pl prt ‘to such an extent.’ 5 R: [(SNIFF)] 6 K: .. look:imp:2sg because it be:3sg it-ade way-ade so that ‘(H) see because it is so that. it be:3sg hell-gen good when you come-2sg there ‘(TSK) (H) It is damn good that you will come there.(0.’ 2 R: [(Hx)] 3 K: ku su-lla on kuitenki se-n verran hallu-s because you-ade be:3sg anyway it-gen much possession-ine ‘because after all you master those things’ noi jutu-t ni. ne ei niinku -- they neg prt ‘.’ 11 R: [Joo]. these young-comp-pl notary-pl ‘These younger notaries. prt ‘Yeah.’ 9 ni ne ei. prt they neg ‘they don’t...’ 4 (H) kato ku se on si-llä [tava-l]la niin että. they don’t like --’ 7 .242 Mirka Rauniomaa different topic by showing her appreciation of the fact that Reija will soon start working at the same workplace with her.. (11) sg067_a1 Lawyer Issues <T:00:29:46> 1 K: (TSK) (H) Se on helveti-n hyvä ku sä [tuu-t sinne].8) 8 Nää nuore-mma-t notaari-t. just it-ill ‘just there.’ .

(0.. which Kaarina begins to elaborate.’ The assessment works as an introduction to the current topic. that how it prt be. Kaarina continues on line 3 to account for the assessment and to praise Reija even further: ku sulla on kuitenki sen verran hallus noi jutut ni ‘because after all you master those things to such an extent.on top of the application. Indeed.’ 15 R: [Nii]. prt ‘Yeah.. Reija responds to Kaarina’s positive assess- ment only with a somewhat forceful exhalation on line 2. re: Kaarina’s colleagues)) On line 1.. Kaarina makes an assessment. write-3sg just it-ill prt ‘. Stance markers in spoken Finnish 243 12 K: . while Kaarina continues her lengthy account. it-ill application-gen on. From line 4 onwards..top ‘. Kirjotta-a vaan sii-hen niinku. Reija occasionally displays acknowledgement or agree- ment.’ 14 että miten se niinku kannatta-i[s <X>an]-taa</X> ni. on top.’ ((17 lines omitted..5) .beneficial-cond give-inf prt ‘that how Ø should like give it. sii-hen hakemukse-n pääl- päälle. There are two features in Kaarina’s turn that indicate that no elaborate uptake is expected: Kaarina’s utter- ance on line 1 ends in a fairly level intonation that implies continuity. se on helvetin hyvä ku sä tuut sinne ‘it is damn good that you will come there.’ 13 .. immediately after making the assessment. Just writes it there like. and there is no noticeable pause before her next utterance on line 3. There is then a moment of heavy overlap before Kaarina comes back to her previous point: (11) 16 K: (TSK) (H) Ni -- prt ‘(TSK) (H) So --’ 17 .’ Perhaps partly because of the constraints on self-praise (see Pomerantz 1978). Kaarina provides an extended account that contrasts Reija’s competence with the incompetence of Kaarina’s current colleagues who are somewhat negligent in making their proposals so that others will have to check their work afterwards.

that prt ‘(H) that like.’ 27 K: mä e-n rupee [2tollas2]-t-en kanssa [3nyh3]rää-[4mä-än4]. I see-1sg it-acc first_name-acc ‘I see Heikki..’ 21 .’ 23 <X>mä nää-n se-n Heiki-n. because ‘Because..’ 26 R: [Nii].’ 24 si-l on</X> niin älyttömä-sti tö-i-tä.’ 28 R: [2(SNIFF)2] [3(SNIFF)3] 29 [4Eli su-l4]la pitää ol-la joku jo-hon so you-ade have. I-ade go-cond:3sg it-ill prt ‘It would take me like.to be-inf someone who-ill ‘So you have to have someone sä voi-t [5luot-taa5]. I neg-1sg start that. you can-2sg trust-inf you can trust. neg-1sg I start prt they-acc that. it-ade be:3sg so unreasonable-adv work-pl-ptv ‘he has such an unreasonable amount of work to do.’ 22 Mu-lla men-is sii-hen niinku. I-gen mind-ela be:3sg prt really good that you come-2sg ‘In my opinion it is like really good that you will come.--’ 19 <A>Mu-n miele-stä</A> on niinku tosi hyvä et sä tuu-t.’ 25 (H)[= et] niinku.kind-pl-gen with tinker-inf-ill ‘I won’t start tinkering with those.244 Mirka Rauniomaa 18 <X>Tosiaan m-</X> -- really ‘Really i.way tinker-inf:ill ‘.’ 20 Koska. prt ‘Yeah. I won’t start tinkering with them. e-n mä rupee niinku nii-t silleen nyhrä-ä.’ .

Although both intonation units are truncated. 19 K: mun mielestä on niinku tosi hyvä et sä tuut. which has been omitted here. hyvä ‘good.’ The overlap. Speaker’s modified repetition of an assessment in (11) 1 K: Se on helvetin hyvä ku sä tuut sinne.’ which is added to the second assessment to give Kaarina time to plan her speech (ISK 2004 includes niinku in the group of ‘planning particles’). The fact that Kaarina repeats the assessment confirms her stance.’ which is left out from line 19 possibly because the referred place has already become well-established. the particles ni ‘so’ and tosiaan ‘really’ succeed in projecting that what Kaarina will say next is tied to what she has said before (see Hakulinen 1997: 45 for a remark on ni). may be one reason for the trouble that Kaarina has in resuming talk on lines 16–18.’ How- ever.’ is used in both assessments. exactly ‘Exactly. and neither does the adverbial sinne ‘there. nor the particle niinku ‘like. which is placed at the start of the intonation unit. prt ‘Yeah. When Kaarina from line 20 onward accounts for the assessment . prt ‘(TSK) (H) Yeah. The connection becomes evident on line 19.’ 32 R: [6(TSK) (H) Nii6]. but at the same time the addition of mun mielestä implies a shift in the stancetaking.6 The actual subject is realized as a complement clause in both assessments. but the ad- verbial intensifier helvetin ‘damn’ on line 1 is replaced with the equally strong tosi ‘really’ on line 19. but in the second one the optional dummy subject se ‘it’ is omitted in front of the copula and the conjunction ku ‘when’ is replaced with et ‘that.’ Diagraph 2 shows the lexical and syntactic similarities and differences between the two as- sessments. The most interesting difference between Kaarina’s assessments on lines 1 and 19 is the inclusion of the stance marker mun mielestä.’ 31 Ni[6menomaan6]. mun mielestä on niinku tosi hyvä et sä tuut ‘in my opinion it is like really good that you will come. these changes do not affect the import of line 19 in comparison with line 1. when Kaarina makes a similar assessment as she did on line 1. Diagraph 2. Stance markers in spoken Finnish 245 30 K: [5Nii5]. The same assessment term.

e. she brings in a more personal aspect: rather than compar- ing Reija with her colleagues as she did from line 3 onward. To put it more generally. the evaluative element never precedes the stance marker. Reija can contribute to the stancetaking more freely. they are placed either at the start or. the stance markers can be placed in either initiating or responding positions. Kaarina now expli- cates how she would benefit from having Reija around.. and a one-sided praise turns into mutual positive evalu- ation. to a lesser extent. I argued that as stancetaking is not limited to such explicit means as evaluative. On line 29. In moving from an unmarked assessment to one that includes the stance marker minun mielestä.g. The stance marker mun mielestä works here to orient the participants to this change. The range of elements that can be placed before the stance marker in the intonation unit is very limited. e. without making direct reference to herself. I first showed that there is significant regularity in the occurrence of the stance markers.’ In this way. In Section 4.g. lexical and grammatical items or the evaluative . in assessments.. the participants of example (11) engage themselves quite differently in the stancetaking. in the middle of the intonation unit. minun mielestä and minusta reflect the con- tingencies that arise in interaction: framing an assessment with minun mielestä or minusta provides speakers a means of taking into consideration the sequen- tial environment in which they produce their utterance as well as any emerging changes in it. A somewhat restricted kind of stancetaking is gradually opened up. Reija displays agreement with the positive assessment that Kaarina has made of her. This variety of sequential posi- tions is made possible by the types of action that the stance markers frequently occur in. as her subsequent turn indicates. that is clearly marked as such with the particle eli ‘so’: eli sulla pitää olla joku johon sä voit luottaa ‘so you have to have someone you can trust. At a sequential level. assessments. I then focused on the stance markers as they occur in the context of assessments. 7. At a more local level. Once Kaarina has brought in the personal aspect. On lines 30–31. Conclusion The purpose of this paper was to contribute to the understanding of the Finnish stance markers minun mielestä and minusta. Rather than attempting at a comprehensive overview of minun mielestä and minusta across different action types.246 Mirka Rauniomaa that she has just made. Reija name- ly produces a formulation (Heritage and Watson 1979). Kaarina accepts Reija’s formulation and consequent- ly reinforces her own stance.

in which both the per- sonal pronoun minun and the first-person-singular possessive suffix -ni are used. the possessive suffix is not used at all. there are two cases of minun mielestä in which most items precede the stance marker. and marking transition to a first assessment within an extended turn. Sanja Starck. What is more. In the present data. it give-3sg really good-acc picture-acc I-gen mind-ela it-ela film-ela ‘It gives a really good idea in my opinion of the film. Liisa Raevaara. In such environments. Tiina Keisanen. I identified three orientational functions for minun mielestä and minusta: projecting disagreement in a second assessment. However.’ . They do so by acknowledging the existence of several possible stances and showing recognition of the sequential context in which the assessment is produced. the fullest form of the expression is minun mielestäni. I am also grateful to the two anonymous reviewers for examining my paper and suggesting improvements. Paunonen 1995). Robert Englebretson. However. participants’ use of a stance marker within an assessment is brought about by an interactional requirement. 2. Elise Kärkkäinen. the Finnish Graduate School in Language Studies. This study was made possible by fund- ing from the Academy of Finland (project number 53671) and Langnet. To be exact. that are used in spoken and written academic discourse. (a) sg067_a1 Lawyer Issues <T:00:03:46> R: Se anta-a hirmu hyvä-n kuvan mun miele-st sii-tä leffa-sta. marking transition to a first assessment across speakers’ turns. Stance markers in spoken Finnish 247 action of assessing. and only the object of the transitive clause that comprises the intonation unit follows the stance marker – examples (a) and (b). including mielestäni. 3. as cited in Paunonen 1995. studying specific linguistic items as they occur in actual talk-in-interaction con- tributes to an understanding of stancetaking as the participants’ joint endeavour. but the fact that speakers make similar assessments with and without stance markers argues for an appreciation of minun mielestä and minusta as an interac- tionally motivated means of orienting participants to stancetaking. they display an in- tersubjective understanding in a demonstrably subjective environment and thus facilitate recipients’ stance taking. as well as the participants of the 10th Biennial Rice Linguistics Symposium for the inspiring and encouraging atmosphere in which I was able to prepare this paper. I wish to thank John Du Bois. and in spoken Finnish more generally (Ikola 1992. Stancetaking can very well be managed without any explicit linguistic fram- ing. Pentti Haddington. the stance markers can be seen to deal with the potential riskiness of disagreeing or moving from non-evaluative/ implicitly evaluative to explicitly evaluative talk. Maarit Niemelä. Notes 1. That is to say. see Luukka (1992a: 119) for a discus- sion of the varied forms.

and Paolino.” In Kieli 13: Keskustelunanalyysin näkymiä [Perspectives in Conversation Analysis]. Santa Barbara. Licentiate’s Thesis. 4. Oulu. the stance markers offer a flexible resource for participants to address some interactionally relevant issue. and Routarinne. 4): 295–325. 4. 2003a. “Egocentricity in discourse and syntax. J. and Hopper. 1997.’ These two cases offer interesting evidence of how contingencies are managed in conversation: although relatively fixed as regards their position in the intonation unit.. September 10.” In Talking Data: Transcription and Coding in Discourse Research. their combination serves yet another interactional purpose. September 8. Schuetze-Coburn. 1992. Hopper (eds. J.248 Mirka Rauniomaa (b) sg151_a Chat about Something <T:00:40:07> I1: Mut si-llä-hän voi käy-dä ostaa mi-un miele-st nii-t. Discourse Tran- scription.. Du Bois.D. Department of Finnish. S. Hillsdale.’ However. Department of Linguistics. For a discussion of contingency.. 2001) has shown that they are in effect used in different sequential contexts. “Dialogic syntax: The syntax of engagement. Ed- wards and M. Bybee and P.). Sor- jonen (1999. Amsterdam: John Benjamins. Dahl.W. S.” Functions of Language 7(1): 37–77. “The stance differential. Vol. M. “Introduction to frequency and the emergence of linguistic structure. 2003b. I thank Elise Kärkkäinen for this observation.W. Santa Barbara Papers in Linguistics.A. E. Dialogic Syntax and Interaction Seminar. Univer- sity of California.W. University of Helsinki. 1998.” In Frequency and the Emergence of Linguistic Structure.). D. In this paper. “Complaints about transgressions and misconduct. “Outline of discourse transcription. 5.” Paper presented at Stance Taking. 1–24. J. 122–154. Du Bois. see Ford (2004). Oulu. . Drew. S. Cumming. Duvallon. and Paolino. 6. 45–89. O. J. As the translation of nii joo in example (5) implies. “Parenteesi keskustelun kieliopin voimavarana [Pa- renthesis as a grammatical resource in conversation]. NJ: Lawrence Erlbaum Associates. (eds. J.” Research on Language and Social Interaction 31 (3. Du Bois. 2000. D. Ö. J. which may or may not have been material at the moment they started speaking.” Paper presented at Langnet Seminar.and particle-like features of the nega- tion element ei. P. 2001. Du Bois. but it-ade-cli can go-inf buy:inf:ine I-gen mind-ela those-acc ‘But you can use it to buy in my opinion those. 1993. Bybee. Helsinki: Department of Finnish.). Mielipiteen vuorovaikutuksellinen rakentaminen syntyperäisen ja ei-syntype- räisen suomenpuhujan keskustelussa [Interactional Construction of Opinion in Conversa- tions between Native and Non-Native Speakers of Finnish].. University of Jyväskylä. P. I have translated the response particles nii and joo as ‘yeah. S. Cumming. 2001. See Helasvuo (2001b: 78) for a discussion of the verb. The change may partly be inspired by the use of tosiaan on line 18. References Aalto. J. Lampert (eds. Routarinne (eds.). S.W. Schuetze-Coburn. Halonen and S.

2003. “Assessments and the construction of context. and Watson. 2006. 32–55.. 2003. 176–196. Amsterdam: John Benjamins. Keevallik.” In Re- thinking Context: Language as an Interactive Phenomenon. A. O. Syntax in the Making: The Emergence of Syntactic Units in Finnish Con- versation. and Thompson. L. D.-L. V. 1997. Korhonen.” In Everyday Language: Studies in Ethnomethodology.A. (eds. M. Atkinson and J. Heinonen and I. Department of Finno-Ugric Languages. University of Helsinki. Porvoo: WSOY. 1992. 1989. V.C. M. “Vuorottelujäsennys [Turn organization]. S. B. M. Helasvuo. Koivisto. ISK 2004. 115–146. Ph. Goodwin (eds. J.-L. Yli-Vakkuri. Duranti and C.H. Amsterdam: John Benjamins. “Partikkelit ja muut kiteymät vuoroissa [Particles and other clusters in conversational turns].H. 1979. Cambridge: Cam- bridge University Press. A. R. and Goodwin. M. Hakulinen.). Ford. 2003. A. S. Hakulinen. and Goodwin. Alho (eds. 2004. The Language of Turn and Sequence. and Sorjonen. From Interaction to Grammar: Estonian Finite Verb Forms in Conversation.” Journal of Pragmatics 36 (8): 1315–1336. Couper-Kuhlen and M.E. Suojala. Tampere: Vas. Epistemic Stance in English Conversation: A Description of Its Interactional Functions.R.” In Kieli 4: Suomalaisen keskustelun keinoja I [Characteristics of Finnish Conversation I]. Nyman (eds. Karlsson. M. Fox.R. C. C. L. C. M.).A. J. Uni- versity of Colorado. 123–162. Master’s Thesis..).” In Yhteiskunta muuttuu – kieli muuttuu. Depart- ment of Finnish. 25–50. Kärkkäinen. Heritage (eds. Goodwin.).” In Studies in Interactional Linguistics. “Concurrent operations on talk: Notes on the interac- tive organization of assessments.).” Discourse Studies 6 (1): 27–52. J. “A change-of-state token and aspects of its sequential placement. 2001b. Hakulinen (ed.). Tainio (ed. E. Stance markers in spoken Finnish 249 Erringer.” In Struc- tures of Social Action.D. Goodwin. 1996. Uppsala. Heritage. E.-L. Hakulinen. E. New York: Irvington Publishers. Ikola. New York: Oxford University Press. Nykysuomen Seuran 10-vuotisjuhlakirja [Society Changes – Language Changes. G. tapaino.” Text and Talk 26 (6): 699–731. Heritage. 1984. Psathas (ed.M. T. Helsinki: Finnish Literature Society. “Stance-taking in conversation: From subjectivity to intersubjectivity.” IPrA Papers in Pragmatics 1 (1): 1–54.” In Keskustelunanalyysin pe- rusteet [Introduction to Conversation Analysis]. C. 299–345. The 10th Anniversary Volume of the Association of Contem- porary Finnish].).). Helasvuo. 1987. Länsimäki and A. Helasvuo. M. Amsterdam: John Benjamins. 2002. Tämmöinen and Tuommoinen in Finnish Conversations. Selting (eds. “Shared syntax: The grammar of co-constructions. M. 1992. with a Focus on ‘I think’.” Melbourne Papers in Linguistics and Applied Linguistics 3 (1): 5–23.-L. “Formulations as conversational objects.). Vilkuna. Kärkkäinen. “Interactional uses for an epistemic marker: The case of ‘jag tycker’/’tycker jag’ in Swedish. M. dissertation. 147–189. “Emerging syntax for interaction: Noun phrases and clauses as a syn- tactic resource for interaction. . 2004. “Omistusliitteiden harvinaistuminen [Rarefaction of possessive suffixes]. “Contingency and units in interaction. A. The Functions of Demonstrative Adjectives Semmoinen. Department of Linguistics. Ford. A. A. Cambridge: Cambridge University Press. 2001a. Iso suomen kielioppi [A Comprehensive Grammar of Finnish].

“Compliment responses: Notes on the co-operation of multiple con- straints.). Norwood. E. Suomen kielen seuran vuosikirja 32 [Sananjalka.. (eds.” In Keskustelunanalyy- sin perusteet [Introduction to Conversation Analysis].A. 1989. Heritage (eds.. J. Interpersonal and Contextual Features of Academic Texts]. “Nollapersoona [The zero person]. Luukka. Schegloff. M.). Oxford: Blackwell. interper- sonaalisia ja kontekstuaalisia piirteitä [Academic Metadiscourse. G.A.). 57–101. Nikula. 61–89. and Thompson.A. J. C. L. Schegloff. E.-L. Selting. 18–31. Schegloff. Akateemista metadiskurssia. Scheibman. “Agreeing and disagreeing with assessments: Some features of preferred/ dispreferred turn shapes. Turku: Finnish Language Society. “Varmuuden kahdet kasvot tieteellisessä tekstissä [The Janus-faced cer- tainty in academic texts]. J. 1992. 1996. L.). Interaction and Grammar. E. Licentiate’s Thesis. Yearbook of the Finnish Language Society 32].” In Studies in the Organization of Conversational Interaction. 1992a. 2001. “On hedging in physician-physician discourse. Tampere: Vastapaino. “Puhesuomen muuttuva omistusmuotojärjestelmä [Morphological chang- es in spoken Finnish possessive forms]. Sacks. E. S. Textual. 79–112.” In Frequency and the Emergence of Linguistic Structure.” Virittäjä 99: 337–357. 1978. M. T. “Confirming allusions: An empirical account of action. 1995. E. . M. R. 1974. Bybee and P.-L. L. 83–97.J.” Virittäjä 4: 501–531. Atkinson and J. 1984.” In Structures of Social Action: Studies in Conversation Analysis. M. Studies in Interactional Linguistics. Luukka.” Virit- täjä 103: 170–194. and Jefferson. “Suomen kielen kohteliaisuusstrategiat [Politeness strategies in Finnish].).-R.. Hop- per (eds.” Language 50 (4): 696–735. NJ: Ablex. Amsterdam: John Benjamins. Sorjonen. University of Helsinki. M. Jyväskylä: University of Jyväskylä. M.-L. A Study in Interlanguage Pragmatics. A. “Local patterns of subjectivity in person and verb type in American English conversation. A. Department of Finnish. Sorjonen. A. Amsterdam: John Benjamins. Hakanen (ed. Pomerantz. Schenkein (ed. Lampinen. 1990. 1993.A. 162–176. and Couper-Kuhlen. A. Kannanoton tulkinta keskustelussa [Interpreting an Assessment in Conversation]. Lectures on Conversation. Ochs.). Prince. Pomerantz. Jyväskylä: Uni- versity of Jyväskylä.” In Kieli 4: Suomalaisen keskustelun keinoja I [Characteristics of Finnish Conversation I]. Sorjonen. Tainio (ed. “Vuorovaikutus paperilla [Interaction on paper]. Frader. 1999. Tieteellisten tekstien tekstuaalisia. (eds.” In Sananjalka. 1997.250 Mirka Rauniomaa Laitinen. Cambridge: Cambridge University Press. University of Helsinki. 1995.).” Virittäjä 2: 361–379. Amsterdam: John Benjamins.M. “Dialogipartikkelien tehtävistä [On the use of response words]. Paunonen.-L. Sacks. Helsinki: Department of Finnish. di Pietro (ed. J. “Vuoronalkuiset konnektorit: mutta [Turn-initial connectors: mutta ‘but’]. 77–92.” American Journal of Sociology 102 (1): 161–216. New York: Academic Press. H. H.-R. E. A. Seppänen. H. “A simplest systematics for the organization of turn-taking for conversation. 2001.” In Lin- guistics and the Professions: Proceedings of the Second Annual Delaware Symposium on Lan- guage Studies. 1996. Pragmatic Force Modifiers. 2001. Tainio. Hakulinen (ed.).). and Bosk. E. 1982. 1996. Responding in Conversation: A Study of Response Particles in Finnish. J. 1992b. Cam- bridge: Cambridge University Press.

-J. A. Thompson (eds. [2 2] Speech overlap (indexed by numbers inside brackets) = Lengthening …(N) Long and medium pause (duration in parentheses. S. New York: Oxford University Press. S. Department of Finnish. marked on a separate line if not within an IU) . “Verbs of cognition in spoken Spanish: A discourse pro- file. G. 194–213.G. S. Short pause (brief break in speech rhythm) (TSK). “Kannanotoista arkikeskustelussa [On assessments in everyday conversation]. Du Bois et al. Amster- dam: John Benjamins. 2000. University of Helsinki. Thompson. Wu. Final intonation contour .). Stance markers in spoken Finnish 251 Tainio. London: Routledge. 1993) {carriage return} Intonation unit (IU) -- Truncated intonation unit - Truncated word . Hunston and G.R. 81–108. (SNIFF) Vocal noises (Hx) Exhalation (H) Inhalation <P> </P> Soft voice (piano) <A> </A> Rapid speech (allegro) <HI> </HI> Higher pitch level <%> </%> Creaky voice <X> </X> Uncertain hearing X Indecipherable syllable (( )) Researcher’s comment Appendix B: Symbols used in glossing 1sg first-person singular ending 2sg second-person singular ending 2pl second-person plural ending 3sg third-person singular ending . 2004.R.). “Evaluation: An introduction. E. Weber.). Appendix A: Symbols used in transcription (From Du Bois et al.” In Evaluation in Text: Authorial Stance and the Construction of Discourse. and Hunston. Stance in Talk: A Conversation Analysis of Mandarin Final Particles. Hakulinen (ed. R. 1996. and Bentivoglio. Continuing intonation contour [ ] . 1–27. L. 1992..” In Discourse-Pragmatics and the Verb. Waugh (eds.” In Kieli 10: Suomalaisen keskustelun keinoja II [Characteristics of Finnish Conversation II]. Fleischman and L. P. 1991.

’ ‘to’) imp imperative ine inessive (‘in’) inf infinitive neg negation pass passive pcp participle pl plural prt particle ptv partitive (‘part of ’) q question marker trans translative (‘into.’ ‘for’) .’ ‘from’) gen genitive ill illative (‘into.252 Mirka Rauniomaa 4 passive person ending abl ablative (‘from’) acc accusative ade adessive (‘at.’ ‘on’) adv adverbial all allative (‘to’) cli clitic comp comparative cond conditional ela elative (‘out of.

it follows the projected course of action of the prior turn or the sequence (cf. or in short. specifically challenging. see also Kärkkäinen 2003). The present study is underlined by a certain view of language. so- cially constructed activity (Du Bois this volume. I will here start from a position whereby language in interaction is chosen as the domain of study. which has been the focus of much past linguistic research on stance. negotiate their points of view and alignment with each other. Finland 1. or epistemic certainty (or doubt) toward some state of affairs. In order to widen the scope of research from explicitly evaluative or epistemic material in individual utterances. namely negative yes/no interrogatives and two types of tag questions. as doing disaligning work. evaluation. The absence of such alignment is considered as a display of disalignment. thereby suggesting that this stance or claim is problematic. Introduction1 The study of how people display affect. a view that is dominant especially among conversation analysts and interactional linguists: all aspects of language are considered to be sensitive to the social action that the . This study discusses three types of polar interrogatives. has gained more interest lately within linguistics and also in other related areas of research. and holding the recipient accountable for it. in their local contexts. Stancetaking as an interactional activity Challenging the prior speaker Tiina Keisanen University of Oulu. the study of stancetaking. More specifically. and stancetaking is seen as an intersubjective. Du Bois this volume). The challenging actions are constructed by displaying doubt toward a claim or a stance embedded in the prior turn. Within the conversation analytic framework used in the pres- ent study. stancetaking is conceptualized here through the social-interactional processes of alignment and disalignment between discourse participants (cf. this is taken to mean that when a subsequent turn aligns with a prior turn. Keevallik 2003: 29).

like . The challenges are not. ok. ok how’s it background conveys ‘It’s not back- ground. Before presenting the database in more detail. such as ok how’s it background in the example below. is among these few. heh h 5 TC: It’s background?= 6 SD: =yeah.254 Tiina Keisanen discourse participants are engaged in at a given moment (Ford et al. Schegloff 1995). can function as a stance marker. discourse participants are observed to use the wh- interrogative on line 7. and therefore. In other words. and other wh-interrogatives like it. In a certain sequential context. .= 7 TC: ok how’s it background. any linguistic or paralinguistic feature of language. Good- win 2002. and all other features of language acquire local meanings. Koshik’s (2003) study on challenging wh-interroga- tives. grammar.. I am supporting a view that in some sequential contexts. but also the linguistic practices involved in it are considered to be contingent products of interaction. e. not only is stance considered to be an interactional achievement. or a linguistic construction for that matter.’ which suggests that according to the speaker producing the wh-inter- rogative the grounds for doing the prior action do not exist.h most 8 people wouldn’t know[: 9 SD: [((sniff)) 10 TC: maybe what he meant by direct indirect 11 ordinary innovative or visionary. prosody. designed to make a re- sponse relevant.=because u. 12 SD: yeah. (1) (Koshik 2003: 60–61) 1 TC: Is it relevant? 2 *to what you’re saying? 3 (0. This is because the challenging wh-interrogatives are found to convey a negative assertion.g.8)* ((SD is looking down from * to *)) 4 SD: No it’s just background. that is. however. re- sponding is not treated as the relevant next action (Koshik 2003: 63). 2003. Social-interactional practices are thus taken to form the framework within which lexis. Thus. There are a small number of studies which have examined the features of chal- lenging actions in interaction. challenging wh-interrogatives function as requests for an account for the prior utterance. Moreover. I will briefly discuss prior interactional research on challenging as an interactional practice. the practices of stancetaking are examined in and through the sequential organization of interac- tion and the actions and activities people are engaged in. Consequently. for challenging “the basis for or right to do an action done by the prior utterance” (Koshik 2003: 51).

The database includes mostly face-to- face conversations in everyday settings. The data are drawn from 42 speech events. and requesting that the recipient accounts for this observed discrepancy between the participants. It is argued that in their local contexts the interrogative turns are observed to constitute a disaligning challenge in the unfolding course of action by calling into question a claim or a position embedded in the prior turn. Challenging the prior speaker 255 in Koshik’s study the wh-interrogatives used as challenges typically occur in an already established environment of disagreement (e. while Section 3 presents the main findings. I searched the corpus for utterances with both negation and inverted word order (that is. but also some task-oriented interactions such as workplace talk and service encounters. Additionally. Du Bois et al. and thereby relate to the social co-or- dination of knowledge. The selection of data was initially based on morphosyntactic criteria.e. Section 2 introduces the da- tabase used in the study. 2. either the auxiliary/operator or the main verb precedes . The interviewees are also observed to treat the interviewer’s turns as hostile assertions since they typically respond with disagreements and rebuttals. 2003. one where a verb. Du Bois and Englebretson 2004).g. after accusations) “where challenging is a sequentially appropriate next response” (2003: 52). The focus will be on explicating two sequential contexts of challenging negative yes/no inter- rogatives and tag questions in the current database. i. Data The data for this study come from the first three volumes of the Santa Barbara Corpus of Spoken American English (SBCSAE) (Du Bois et al. like the wh-challenges discussed in Koshik (2003). It is argued that the challenges discussed here are warranted by a discrepancy in information or knowledge between the participants. the interview- ers’ turns are rendered essentially as unanswerable assertions due to the fact that they prefer responses that invite the interviewees to assent to critical or negative evaluations of their own conduct or that of their allies. The study is organized in the following manner. 2000. Heritage (2002) offers some observations on “argumentative challenges” in news interviews. which total about 16 hours of transcribed speech. rather than being presented as unanswerable or hostile assertions. However.. Interviewers regularly construct such turns by combining negative yes/no interrogatives with critical question prefaces and/or other evaluative mate- rials. Section 4 concludes the study with a discussion of the findings. This paper adds to these prior stud- ies by examining challenging negative yes/no interrogatives as well as tag ques- tions in two different sequential contexts in everyday interactions..

for instance. the recipient is guided to choose between a yes or no response (Raymond 2003). Quirk et al.256 Tiina Keisanen the subject). concepts such as conduciveness. should be seen as empirical problems to be worked out from the specifics of the local context in which a turn is produced. rather than basing such judgments exclusively on form or intuition. The three constructions therefore share many (grammatical) characteristics. isn’t it. Koshik (2005) illustrates this by discussing two contrastive accounts of the epis- temic strength of negative yes/no interrogatives in interaction. tag questions (in this case posi- tive statements followed by negative tags.really lived in the house . inter- rogatives are often found to initiate “adjacency pair” sequences. a negative yes/no interrogative is . and relatedly. rather than being presented as “questions” to be answered. prior linguistic research has argued that the constructions embody differences in presupposition. On the other hand. interrogatives. see. according to Schegloff and Sacks (1973) consist of two relatively ordered. either negative yes/no interrogatives such as. it should be noted that in interactional research.e. which may be agreed or disagreed with. For example.. have you? (nega- tive declarative followed by a positive tag) or . in the news interview context negative yes/no interrogatives are presented as conducive. and consequently set up dif- ferent trajectories for the recipient’s response in terms of its polarity (this has been called “conduciveness” or “epistemic bias” in the literature. epistemic strength of a turn. As such.. type-connected turns in which a certain type of first pair part makes a choice within a set of re- sponse types “conditionally relevant” (Schegloff 1968). which.. (posi- tive declarative followed by a negative tag). It’s kind of smelly. r. all three interrogative constructions ex- amined here are subtypes of polar. These three constructions share the general characteristic with other interrogative forms in that they regularly make a response relevant when used in conversation (Thompson 1998). Koshik (2005) discusses an example from an everyday interaction (originally pre- sented in Schegloff [1995] 2007). Most such structures are polar interrogatives. The first comes from news interviews. during the winter. In other words. Sadock and Zwicky 1985). i.. since negative statements followed by positive tags were not discussed in Heritage’s study) were found to function more like “yes/no questions to be answered” in news interviews. Moreover. while tag questions are not. or yes/no. However. or tag question constructions such as. Givón 2001. In that case. Isn’t that an oil tank?. in news interviews nega- tive yes/no interrogatives are regularly used to implement hostile assertions of opinion. On the other hand. whose grammati- cal form constrains the recipient’s response to two alternatives. 1985. epistemically strong assertions. this should not be taken to imply that they are considered to be identical grammati- cally or pragmatically. Adjacency-pair sequences provide the discourse participants with a basic resource for the organization of action in social interaction. for example. On the other hand. As Heritage (2002) has observed. you haven’t .

9% 31 30. instead of examining the conduciveness or the relative epistemic strength between the chosen constructions in more detail in everyday interac- tions and/or in different sequential contexts.7% 21 20. However.4% 28 27. epistemicity is here used to refer to those interactional and linguistic means by which discourse participants display their certainty or doubt toward some state of affairs or a piece of information in their own turn.9% 22 21. Moreover. The category other includes turns in which negative yes/no interrogatives and tag questions carry out some less frequent interactional func- tions such as requesting or reporting. I will concentrate on negative yes/no inter- rogatives and tag questions that are used as challenges (see Keisanen 2006 for a discussion of the other data). Challenging the prior speaker 257 observed to weaken the epistemic strength of a prior tag question in the face of incipient disagreement. or they are produced during a monologue (e.  Tag questions and negative yes/no interrogatives in the database by function Tag questions Negative yes/no interrogatives Tokens Percentage Tokens Percentage Requesting confirmation or 31 44. The data consist of 15 tag questions and 21 negative yes/no interrogatives.g. The data are transcribed according to the transcription system outlined in Du Bois et al. or in the turns of others. (1993). Example (3) below presents one sequentially very similar case from the current database. I have slightly modified the original transcripts in order to Table 1. 171 instances of tag questions and negative yes/no interrogatives found in the SBCSAE form the database used in this study. In this paper.6% Other 11 15.4% Challenges 15 21.6% Total 69 100% 102 100% . Negative yes/no interrogatives were also found rather fre- quently to do assessing. Table 1 below provides informa- tion on the frequency of different constructions and their functional distribution in the database.. Requesting confirmation or agreement was found to be the most common functional category for both tag questions and negative yes/no interrogatives in the current database. in a lecture or a sermon) or as reported speech (typically in the midst of storytelling sequences).4% agreement Assessments 12 17. this paper concentrates on providing an analysis of the interactional and sequential organization of action at points in which negative yes/no interrogatives and tag questions are found to function as disaligning challenges in the unfolding course of action.

258 Tiina Keisanen

better demonstrate the contribution of prosody and timing for the accomplish-
ment of disaligning challenges. First, more prosodic detail is added especially on
the interrogative turns and on turns adjacent to them. Second, to be more in line
with the standard conversation analytic transcription practice, the placement of
many of the longer pauses has been changed from the beginning of an intonation
unit to a line of their own. A list of symbols used in the transcripts can be found
in the Appendix.

3. Doing challenging with negative yes/no interrogatives
and tag questions

This section discusses two sequential contexts of challenging tag questions and
negative yes/no interrogatives in the current database. Section 3.1 includes cases
of negative yes/no interrogatives and tag questions that are responsive to types of
telling sequences, while Section 3.2 concentrates on sequences where the con-
structions are used as responses to some initiating actions.

3.1 Challenging the trajectory of an extended sequence

The challenging negative yes/no interrogatives and tag questions discussed in this
section are responsive to types of personal accounts such as stories, tellings, and
the like. As part of such multi-unit accounts, participants typically make various
types of claims and evaluations, as well as design their tellings so as to seek af-
filiation or a certain type of stance display from the recipient(s) (Clift 2000; Ford
2004). The challenging negative yes/no interrogatives and tag questions that are
responsive to extended sequences draw attention to some (explicit or implicit)
claim made or a stance taken in the telling by exploiting a place in the sequence in
which it would be appropriate to align with the teller, but in which a response is
not actually prompted by the teller. By calling into question a claim or a position
embedded in the prior speaker’s turn, the interrogative speaker holds the prior
speaker accountable for this position or claim. Simultaneously, the interrogatives
constitute a display of disalignment toward the unfolding course of action. The
following examples illustrate and discuss this use of negative yes/no interrogatives
and tag questions in the database.
The first excerpt below has three participants, Kathy, Sharon and Carolyn,
who are sisters. Just prior to the excerpt, Sharon has been talking about her expe-
riences as a substitute teacher. She has been complaining about how a lot of the
children in her mixed class of third and fourth graders have trouble with the times

Challenging the prior speaker 259

table. Sharon blames their previous teachers for being bad teachers: they had not
taught them times tables in the second grade, which she thinks should have hap-
pened. Carolyn’s turn that begins on line 2 responds to Sharon’s complaint.

(2) (Raging Bureaucracy SBCSAE 0004 <T: 00.18.31>)
1 KATHY: The teachers have just8] been [9passing them along XXX9].
2 CAROLYN: [9Wait a minute,
3 neither did I=9].
4 SHARON: [Yeah].
5 CAROLYN: [I] didn’t know them either.
6 ... I didn’t know times tables till s- --
7 .. like fifth gra=de.
8 Only cause I was a dufus.
9 (H) But,
10 SHARON: ... [Well <X were X>] --
11 KATHY: [No,
12 <X It was] [2because you were a X> lazy butt2].
13 CAROLYN: [2I could not --
14 I= could not2] do fractions,
15 .. I could not do fractions.
16 .. And I mean with .. ^~Samuel as a `teacher,
17 who ^wants= to ^tell <X them X>,
18 .. (H) who `wants to go up <@ to this ^man and say @>,
19 <VOX `I have a ^problem VOX>.
20 (0.3)
22 KATHY: [<HI You] `didn’t HI> really ^have Mister Samuel,
23 ^did you?
24 (0.3)
25 CAROLYN: ^Oh=,
26 hell ^yeah.
27 [@]
28 KATHY: [^Oh],
29 I `thought you got ^out of being in his `cla[ss].
30 CAROLYN: [`Unh]-`unh=.
31 I ^had him,
32 .. my ^whole fourth-`grade year.

Carolyn’s telling of her troubles with learning math at school starts in overlap with
Kathy on line 2. By demonstrating through her own experiences that difficulties
with math are common in third and fourth grade, Carolyn’s turn downplays Sha-

260 Tiina Keisanen

ron’s complaint about the students and their teachers at Sharon’s school. In effect,
Carolyn provides evidence that Sharon’s expectations may be unrealistic. In con-
junction with this, Carolyn brings in her old teacher, Mr. Samuel, by saying, And I
mean with .. Samuel as a teacher, who wants= to tell <X them X>, who wants to go
up to this man and say, <VOX I have a problem VOX> (lines 16–19). Mr Samuel’s
violent teaching methods were talked about earlier; here, that is implicitly referred
to as one reason for why Carolyn did not learn certain things at the same time as
others. Carolyn’s extended turn ends with an emphatic demonstration of her fear
of Mr. Samuel, <VOX `I have a ^problem VOX>, (0.3) (GASP)= (lines 19–21),
which imitates the voice of a little girl, and which is added with a gasping sound
that imitates a small girl being scared. In other words, Carolyn designs these in-
tonation units by using direct reported speech and altered voice quality (the por-
tion of the transcript marked with VOX), thereby indexing this as the climax of
the telling (Clift 2000; Drew 1998). Clift (2000) observes that such climaxes seek
affiliation from the recipient, but in an implicit manner, that is, a response is not
actually solicited. Instead of affiliating with Carolyn’s telling as would be relevant,
however, Kathy disaligns with the projected course of action by producing a tag
question challenge2 <HI You `didn’t HI> really ^have Mister Samuel, ^did you?
(lines 22–23). It singles out something that Kathy considers to be discrepant in
the content of Carolyn’s telling, namely the claim that Mister Samuel taught her at
elementary school. In so doing, Kathy displays doubt toward the accuracy of this
claim, thereby holding Carolyn accountable for the discrepancy.
The tag question includes a component that displays Kathy’s affective stance:
her prosody is marked to show emotion, which can vernacularly be called surprise
or the like. This heightens the degree of doubt displayed with the tag question.
Figure 1 below shows the pitch trace for the tag question.
The main prosodic feature that displays affect is the overall high pitch during
the turn, and the extremely high pitch in the beginning of the turn. The high-ris-
ing final intonation also contributes to the affective quality of the tag question. In
sum, the display of affect provides further warrant for the challenge embedded in
the tag question turn. Moreover, both parts of the verb phrase didn’t and have on
line 22 receive accent. In conjunction with the adverbial really, they mark contrast
with the claim that Mr. Samuel was teaching Carolyn; this claim was embedded in
Carolyn’s turn And I mean with ..^Samuel as a `teacher (line 16).
Carolyn’s prosodically and lexically emphatic response ^Oh=, hell ^yeah
(lines 25–26) provides an example of a direct, categorical disagreement toward a
disaligning challenge carried out by a prior interrogative turn. It begins with turn
initial ^Oh=, which in responses to prior inquiries mark that the inquiry “was
unexpected, unlooked for or ‘out of left field’” (Heritage 1998: 294). Nevertheless,
this is not enough to close the sequence. Kathy’s third-position turn (lines 28–29)

Challenging the prior speaker 261







You didn't really have Mr Samuel did you

0 2.07667
Time (s)

Figure 1.  Pitch trace of You didn’t really have Mr ~Samuel, did you?

begins with a turn-initial oh-token. In this context, in third position, it indicates
that she has gone through some change in her state of knowledge or orientation
(Heritage 1984). It is followed by an account for why she brought up the issue
in the first place, I `thought you got ^out of being in his `class. However, no cor-
rection to Kathy’s inaccurate understanding of Carolyn’s prior teachers is offered
yet. Carolyn therefore begins such an account on line 32, and the original topic is
never returned to.
In example (3) below, Michael is telling Jim how he would like to be the per-
son whose voice a telephone company has used for the recorded messages one
hears when calling a wrong number. The motivation behind Michael’s wish is the
money that comes from the royalties from the use of the voice.

(3) (Wonderful Abstract Notions SBCSAE 0016 <T: 00.01.11>)
1 MICHAEL: You know I wish I was,
2 uh,
3 the person=,
4 .. whose voice they used in the .. telephone,
5 when it tells you,
6 (0.4) the number has been changed,
7 .. (H) and that I had uh,
8 ^copyright,
9 @@
10 get some royalties.
11 Like,
12 (H) that ^lady,
13 you know,

262 Tiina Keisanen

14 h- you ^hear,
15 (H) (0.4) <Q the ^number you have `reached,
16 da-duh da-duh [da-duh da-duh da-duh da-duh Q>].
17 JIM: [They ^got a `different ^woman],
18 ^didn’t they.
19 (0.3)
20 MICHAEL: ^Hunh?
21 JIM: ^Didn’t they get a `different ^woman?
22 (0.2)
23 JIM: `When she th- `tried to do ^that?
24 MICHAEL: [I don’t `know if they ^did].
25 JIM: [They,
26 they,
27 they `hired an]other ^voi[2ce2].
28 MICHAEL: [2I2] didn’t `hear any ^follow-up.
29 But,
30 JIM: That’s what I thought that they did any[ways].
31 MICHAEL: [Yeah]?
32 (0.4)
33 MICHAEL: Hunh.
34 (1.4)
35 MICHAEL: They certainly u=sed her a lot.

Michael moves into demonstrating what the recorded messages sound like on
lines 14–16, h- you ^hear, (H) (0.4) <VOX the ^number you have `reached, da-duh
da-duh da-duh da-duh da-duh da-duh VOX>. As in example (2), these intonation
units are designed by using direct reported speech and an altered voice quality. At
this juncture of the telling, it would therefore be appropriate for Jim to align with
Michael’s telling by affiliating with it (e.g., by laughing, offering an assessment,
or some other agreeing or appreciative turn). However, instead of this, Jim’s tag-
question turn, They `got a different ^woman, ^didn’t they (lines 17–18), offers an
observation of a discrepancy in the information content of Michael’s telling. This
observation provides him warrant to come in at this point in the telling. Accord-
ing to Jim, the phone company fired the woman that Michael is talking about, be-
cause she tried to get money from the use of her voice. Jim’s tag question therefore
undermines Michael’s story by pointing out that someone already tried, but failed
in carrying out the wishful scheme that Michael is laying out here. Simultaneous-
ly, the tag turn constitutes a display of doubt (to the degree of dismissal) toward
Jim’s knowledge state with respect one of the presupposed facts of the telling (i.e.,
that the lady is still working for the company).

Challenging the prior speaker 263

After a short pause (line 19), Michael responds to the tag question with a
repair initiation ^Hunh? (line 20). The repair initiation may be caused partly be-
cause of troubles in hearing due to the overlap on lines 16–17. On the other hand,
hunh belongs to the class of repair initiators that have been observed on occasion
to indicate troubles in the appropriateness, relevance, and/or the sequential fit
of the prior turn (Drew 1997). Thus, ^Hunh? may display that Michael does not
understand how the tag turn is connected to his telling, or further, that he does
not consider it to be related to his telling thereby pre-monitoring disagreement. In
the repair on lines 21–23, Jim first repeats the candidate claim, but reformulates it
as a negative yes/no interrogative, ^Didn’t they get a `different ^woman? (line 21),
and adds, `When she th- `tried to do ^that? (line 23), in order to explicate how the
interrogative constructions were connected to the telling. In this context, facing
potential disagreement, the reformulation of the tag question as a negative yes/no
interrogative can be seen to downgrade the epistemic strength of the tag question,
and therefore to indicate backdown from the challenge. Also the prosody reflects
a change in epistemic strength as it changes from a tag question produced with
final falling intonation, to a negative yes/no interrogative produced with rising-
query intonation.
In any case, Michael responds by claiming to have insufficient knowledge of
the situation Jim is referring to with I don’t `know if they ^did (line 24). Such a re-
sponse enables Michael to avoid confirming Jim’s inquiry on the factual level, and
more importantly, to reject the action carried out with the prior turn (see Beach
and Metzger 1997). Michael continues with an account, I didn’t `hear any ^fol-
low-up (line 28), which demonstrates that his lack of knowledge is not a matter of
just claiming not to know, but that this claim is based on not having any evidence
that would support Jim’s turn. Jim responds to this with an account of his prior
understanding, That’s what I thought that they did anyways (line 30). Even though
Jim uses past tense to display some change and/or uncertainty toward his position
(cf. example 2), he does not offer further evidence (e.g., the source of knowledge)
for his challenging turn. The matter is thus left somewhat unresolved.
The following excerpt contains another example of a tag question that is re-
sponsive to an extended sequence. Throughout the excerpt, Wendy and Marci are
involved in doing disaligning actions, which in part contributes to the interpre-
tation of the tag question on lines 20–22 as a challenge. The excerpt starts with
Marci’s complaint about the bad quality of the containers of the soda that they are

^cocktail. get so [warped]. They don’t `make (0. 27 KEN: [(THROAT)]= 28 (0...that bottled ^water.. 31 [it’s all] [2with2] -- 32 MARCI: [But they don’t] [2make2] `this kind at ^a=ll [3any3]more. 2 ..07>) 1 MARCI: . 12 MARCI: .X> -- 22 ... ^don’t they. 10 KENDRA: that’s why4] -- 11 WENDY: and complain !Marci. %= write a letter to em4]. 9 WENDY: [4%Just ..00. 13 is that they ^don’t make white `grape. 14 ^This stuff [is] ^g=ood. or the `sparkling ^water.. 21 <X do.. 5 KEN: [(H)]= 6 WENDY: (H) % [2I think you should2] [3write3] -- 7 KEN: [2<X Cause they’re X>2] [3X3] -- 8 KENDRA: [3Cause they’re3] [4cheap ca=ns. get so warped. .. 33 KENDRA: [2(Hx)2] 34 [3(SNIFF)3] 35 [4(THROAT)4] 36 WENDY: [4. 23 . Why do these ca=ns. I couldn’t find it last time I was in there. 3 Only the -- 4 . Only the Sam’s Club cans . 30 ... 18 . 17 or `something. Oh they4] don’t? 37 MARCI: . 15 >ENV: [((POP-TOP))] 16 MARCI: It’s like ^sparkling= `grape `juice .4) 29 WENDY: Every `time I’ve ^looked at <X the X> bot.264 Tiina Keisanen (4) (Appease the Monster SBCSAE 0013 <T: 00.3) 25 MARCI: A `regular ^grape? 26 I don’t [^know]. (H)= `What I’m gonna ^complain about.4) a ^regular -- 24 (0... you know. 19 . `remember that [white] -- 20 WENDY: [(H) <HI<F They] ^only make `that F>HI>. with ^Nutrasweet though..

Wendy disaligns with the course of ac- tion that Marci pursues.. The second part of the turn. and specifically with disjunctive actions in which the wh-cleft may be used to evoke the speaker’s first-hand epistemic experience of the matter (Kim 1995). Kendra. but cuts off his turn most likely due to the extended overlap first by Wendy and then Kendra. which further strengthens the negative character of the turn. then. with ^Nutrasweet though. and Wendy react to Marci’s complaint on lines 1–4. As in example (3). she moves on to ar- gue for the good quality of the drink with ^This stuff is ^g=ood. Wendy latches on with a turn that questions the quality or the healthiness of the drink (Nutrasweet is an artificial sweetener. This turn is produced in a somewhat sarcastic tone. I don’t ^know. I don’t ^know.(lines 18–19). Wh-cleft constructions have been associated with stance displays (Hopper 2000.X> -. (line 26). Marci asserts her right to complain by using the cleft construction. (lines 14. is that they ^don’t make white `grape (lines 12–13). thus implicating that warping is what can be expected from them. 16).. which here functions as a marker that restricts a claim made in previous discourse.(line 23).4) a ^regular -. %= write a letter to em. and carries nega- tive connotations toward the use of Nutrasweet. which is believed by some to be toxic). Wendy disaffiliates with the complaint by producing a rather dismissive turn. it is inter- rupted by Wendy’s tag question. A `regular ^grape (line 25) is a candidate offered for Wendy’s word search. in- stead of joining Marci’s evaluation of the drink. Ac- counts are one type of relevant second pair part after complaints (Schegloff 1988). Thus. Ken also starts to formulate an account with <X Cause they’re X> (line 7).^don’t they (lines 20–22). The main accent on line 20 is on the adverbial only. In other words. Challenging the prior speaker 265 Ken. (H) <HI<F They ^only make `that F>HI>. After that. ^cocktail. Kendra offers the cheapness of the soda as an account for the bad quality of the containers (line 8). Essentially. the challenge carried out with the turn makes Marci’s positively evaluative stance of the beverage an accountable action. and complain !Marci (lines 9. Wendy’s turn receives a counter from Marci in the form of a wh-cleft. Kim 1995). or waiting to align as a recipient to a story. %Just . you know. This functions to “facilitate the foregrounding of the speaker’s counterac- tive stance towards the preceding talk” (Kim 1995: 268). It’s like `sparkling= `grape `juice . by appealing to the potential unhealthiness of the drink. They don’t `make (0.. whatever activity is started. does not confirm . contains Marci’s response to Wendy’s claim that the company does not make regular soda any more. Marci’s response to the tag question on lines 25–26 consists of two parts. <X do. However. Marci may be working her way into a story or a reminiscence of a shared experience relating to the drink that would provide more evidence of its good quality.. . `remember that white -. Moreover. However.. her assessment of the drink does not get taken up immediately. However. 11). and so Marci continues the turn with . `What I’m gonna ^complain about.

(line 29). the interrogative is produced after the repair sequence has not successfully solved a problem one of the recipients observes in the telling. Koshik’s (2003) discussion on wh- interrogative challenges includes a similar sequence from a teacher-student tutor- ing session. because Marci’s disagreeing turn. In these cases. 6 LISA: [2XX2] 7 MARIE: and his ^legs were ^bleeding and stuff. Wendy does not get a chance to finish her turn. the negative yes/no interrogative or a tag question is preceded by a repair sequence. the repair sequence is employed as pre-disagreement (Schegloff et al. The following example presents a sequential variation to the disaligning nega- tive yes/no interrogatives and tag questions that are responsive to types of telling sequences. 12 cause he ^bleeds.that bottled ^water. which indicates that whether or not the company uses Nutrasweet is not really the issue. and maintains her positively evaluative position toward the drink. (5) (Judgmental on people SBCSAE 0036 <T: 00. In short. while the following negative interrogative carries out the fully formulated disagreement. Marci reas- serts her right to complain about the state of affairs. 8 and I had ^blood all on my `shirt. in her challenging turn. 10 LISA: [^Blood. or challenge. 4 KEVIN: [@@@@] 5 MARIE: And [2then I pick2]ed him up.05>) 1 MARIE: [the next time I took him] he’s all.06. Moreover. In other words. Kevin is Lisa’s brother. As ob- served by Pomerantz (1984a). Marie is telling Lisa and Kevin about taking her son Issac to the doctor. which concerns the same issue as the following in- terrogative. by pointing out a flaw in Wendy’s account and consequently. 9 and I was all <VOX ^oh my @^Go=d VOX> [@@@(H)]. in such cases. 2 (H) (SCREAM) 3 [I was all <VOX oh my] ^Go=d VOX>. Consequently. Lisa and Marie are close friends. Wendy moves on to provide first-hand evidence for her tag-question turn with Every `time I’ve ^looked at <X the X> bot. this typically takes place if other discourse par- ticipants display doubt toward the validity of other participants’ claims. In the extract below. The turn includes the emphatic at ^all=. 1977).%(Hx)] -- 11 MARIE: (H) ^Blood. However. But they don’t make `this kind at ^a=ll anymore (line 32) starts in overlap with hers. .266 Tiina Keisanen Wendy’s inquiry on the factual level. After the challenge is met with a non-committing and disaffiliative response. it enables her to avoid conceding to the challenge carried out with the tag-question turn.

4) my child’s `bleeding VOX>. a [little ^dr]ibble. ^Blood. 28 KEVIN: they use3]. However. 35 LISA: [Oh]. 25 I mean ^usual[2ly you know like when you like2] [3give ^bloo=d. This type of repair would be sufficient to repair a trouble in hearing or understanding.. 31 MARIE: [It had] -- 32 N_yeah no. the trouble lies elsewhere. 34 and I was [just all] <VOX `oh my ^Go=d. where the first part of the story was received with recipient appreciation. 21 Do you know what I mean. 33 it had . Marie therefore continues to expand the repair turn by complementing it with a further turn constructional unit. 30 . first Kevin (lines 15–16) and then Lisa (line 17) both move on to produce further . 26 MARIE: [2A ^bandaid2]? 27 LISA: [3But it wasn’t ^all= ov3]er. Marie repairs the trouble by first reproducing the word blood (line 11). Unlike earlier. 23 [mm]. this part of the story gives rise to a trouble source: on line 10 Lisa initiates a repair sequence by produc- ing a partial repeat of the prior turn. but as is pre-monitored by the pause on line 13.. 16 like [put a ^thing on it right `af]ter? 17 LISA: [But that ^mu=ch=]? 18 (0. Issac was crying hard after the shot. 36 MARIE: (0. It of- fers an explication for why Issac would be bleeding in the first place. 20 but it was just coming ^out. (H) That’s good (data not shown). 29 LISA: it was just like=. `gotten on my ^shirt. (TSK) (H) After they give him a ^shot? (line 14). Challenging the prior speaker 267 13 (0. According to Marie.5) 14 (TSK) (H) After they give him a ^shot? 15 KEVIN: But ^don’t they. and then supplying cause he ^bleeds (line 12) by way of an explanation for why she would have blood on her shirt.4) 19 MARIE: ^No.%(Hx) --. 22 it was like. in addition to which he was bleeding. 24 KEVIN: [<HI Well they] should have HI> put>. The telling that starts on line 1 follows an account of a no-problem visit to the doc- tor.

. She disagrees with Marie (line 32) and goes on to redo the climax of the story (lines 33–34. the negative yes/no interrogative is used as a request for an account for this discrepancy in information or knowledge between the par- ticipants. No. it begins with the negative item. Additionally. before Kevin finishes his turn. however. Marie’s response. .e. Lisa tries to reformulate a detail in Marie’s telling by producing. 36). No. the challenging negative yes/no interrogatives and tag questions dis- cussed above provide a resource for story recipients with which they can disalign with the tellers and the projected trajectory of action. However.268 Tiina Keisanen turns. which indicates disagreement. it was just like=. The sequential placement is the decisive factor in the interpretation of the interrogatives as disaligning chal- lenges: the interrogatives are triggered by the prior turn. Kevin’s turn. There are thus two challenges on the table. it was like. Marie does not back down any further. which questions Marie’s account of the procedure that her son underwent. which in slightly differing ways challenge Marie’s telling about her son’s excessive bleeding after receiving a shot.. which was interrupted by the other-initiated repair and challenges discussed above. In sum. Lisa latches on in overlap with her disagreeing turn. According to Kevin. But ^don’t they. 28). In so doing. and a back down. which is frequently used in the beginning of disagreeing turns (Pomerantz 1984b). Marie restates her position and disregards both Kevin and Lisa’s challenges to her claims. she only specifies its timeline by claiming that everything happened so fast (i. Marie does not change any of the facts of her telling. However. It also challenges Marie’s descrip- tion of the excessive bleeding. (lines 19–23). This intonation unit begins his account of how the procedure should have gone. (line 24). is slightly ambiva- lent as to whom she is responding. Kevin overlaps the last intonation unit of Marie’s response with <HI Well they should have HI> put>. But it wasn’t ^all= over. The remainder of her turn can be seen to pro- vide a slight correction. mm. The challenge is formulated as a generalization and grounded on knowledge that Kevin has acquired through his prior experiences with doctors and needles (as is evidenced by his turn on lines 24–25. In short. which cause some trouble for responding. the spot should be covered with a bandaid or the like immediately after the needle is pulled out. In any case. to her telling. It is followed by a negative yes/no interrog- ative. 29–30). but it was just coming out. it was just coming out) that the doctor did not have time to cover the spot before the blood was already all over her shirt. Kevin’s interrogative displays doubt toward Marie’s telling by suggesting that Marie presents an inaccurate or a misleading description of what actually happened. Do you know what I mean. In effect. but they do not consti- . But that ^mu=ch=? (line 17). a little ^dribble (lines 27. begins with the contrast marker but. like put a ^thing on it right `after? (lines 15–16).

(6) (Cuz SBCSAE 0006 <T:00. Negative yes/no interrogatives and tag questions that are used to challenge the prior turn by pointing out a discrepancy in its content are also found outside extended sequences. the discrepancy is used to call into question a claim or a position embedded in the prior turn. Alina and Lenore are talking about Alina’s husband. However. 4 (H) ^He’s been `grumpy for `three months. thereby calling this claim or posi- tion into question. Hector. `He’s been `having a `hell of a ^time. The recipients may counter the challenges with explicit disagreement (example 2). 7 it’s ^o[ver now. 3 .1. the interrogative speakers may provide accounts for their turns (examples 4 and 5). which were observed to challenge the ba- sis or right for doing the prior action. In the first example below. The following section discusses challenging negative yes/no interrogatives and tag questions that provide a disaligning response to some ini- tiating action. the challenges discussed in this section are constructed by pointing toward a discrepancy between the participants with respect to some claim or position embedded in the prior turn. The challenges are based on an observed discrepancy between the participants with respect to some detail in the telling. However. 2 . and moreover.. 5 (H) 6 LENORE: <HI `Well.34>) 1 ALINA: (TSK) `Poor little ^~Hector. the negative yes/no interrogatives and tag questions discussed in this section provide a disaligning turn to some initiating action. Challenging the prior speaker 269 tute a sequentially relevant next action projected by the current activity. or to back down from their claims (ex- ample 5).2 Challenging prior action Similarly to Section 3. . or avoid accountability by appealing to insufficient knowledge (examples 3 and 4). such as alignment as a story recipient who would affiliate with the telling. to hold the recipient accountable for these claims or stances. and are used to challenge the appropriateness or the relevance of doing the action completed in the prior turn. Additionally. Such responses provide evidence for the disaligning nature of the interrogatives discussed above. recipients were also seen to provide accounts (example 2). `Pulling his `hair out. More specifically. 3..20. Such challenges bear a resemblance to wh-question challenges discussed in Koshik (2003).

it’s ^over now. she tells Lenore that her husband Hector (who is not present) is going through a difficult period at work. The use of marked prosody adds affect to the turn. Alina moves to position Hector’s situation with respect to herself. her pros- ody is marked to display that the first three intonation units deliver ‘bad news’: the pitch level is low. while Lenore depicts it to be over). the turn begins in rather high pitch. 12 When I `called him on the ^phone? 13 . He was ^actually ^talkative `tonight. or other alignment with respect to its status as a complaint (Schegloff 1988). and the pitch range is wider. . The last part of the turn therefore moves more toward complaining than delivering news. an aligning response could consist of a co-complaint. 15 (H) Do you know how long it’s `been since he’s been able to k. However. 19 LENORE: [3!^Rosen3]blu[4m4]? In the beginning of the excerpt.. 14 For like `fifteen ^minutes tonight. an agreement. Le- nore’s turn does not provide such alignment. that Alina’s prior turn includes a claim or a position that requires correction. If Alina’s turn were taken up as a complaint (or alternatively.^talk to me on [the `phone]? 16 LENORE: [`Who does he ^work] `with [2`now2].3 Lenore’s tag question is designed by using high pitch and high-rising final intonation. `isn’t it? (lines 6–7.. The prosody also changes. specifically by making available a contras- tive formulation of the state of affairs at Hector’s work (Alina has described the situation as an ongoing one. and the pitch range rather narrow (Freese and Maynard 1998). Second. 17 ALINA: [2(H)2] 18 (TSK) [3!^Kevin3]. However. `Well. the voice quality creaky. it includes the negatively evaluative expressions hell of a time and grumpy. She points out with the tag question. Alina moves from characterizing her husband as poor little Hector (line 1) to grumpy (line 4). as an inform- ing). 11 . (H) `He ^tal=ked to me. Alina initiates a new topic after talking at length about a party she went to recently. 9). In other words.270 Tiina Keisanen 8 ALINA: [(H)] (SIGH) 9 LENORE: `isn’t it HI>]? 10 ALINA: <HI ^Yeah HI>. One turn design feature that points toward this is the ad- jective grumpy. On lines 1–4. which characterizes Hector’s behavior toward Alina. in the last intonation unit on line 4. First. Figure 2 presents the pitch trace for the tag question. Her turn includes a display of an evaluative stance toward the informing. and provides further war- rant for the challenge.

So.4) 5 KEVIN: Because [6it’s the ^only kind we’re `not6] [7`allergic to7]. it could be noted that the actual content of the challenging turn can be seen to display compassion or sympathy: Lenore’s tag question turn indicates that the situation is not as bad as Alina claims it to be.  Pitch trace of Well it’s over now. The others have also sung Happy Birthday to Kendra. Marci is Kendra’s mother. 3 WENDY: [5(THROAT)5] 4 (0. As can be seen from Alina’s response (lines 10–15). On the other hand. The participants are cel- ebrating Kendra’s birthday. 6 MARCI: [6<HI Don’t you ^like ice `creamHI>6]? 7 KENDRA: [7I don’t `like ice7] ^crea=m. a presupposition) in the prior turn. isn’t it? Lenore disaligns with the projected trajectory of the complaint sequence by calling into question a claim embedded in Alina’s complaint. . 2 but `why do you guys [5`always5] give me ^ice cream `cakes.e. Lenore challenges the appropriateness of doing a complaint in the first place. (7) (Appease the Monster SBCSAE 0013 <T: 00:18:24>) 1 KENDRA: It’s a ^beautiful `cake. by calling into question Alina’s formulation of the situation. It is done by targeting an em- bedded claim (i. The example below presents another case where the interrogative speaker challenges the appropriateness of the prior action. the challenge con- tributes to a complete change in the direction of talk: Alina moves on to provide a positively evaluative account of Hector and his behavior toward her. Challenging the prior speaker 271 Figure 2. to put it differently. which could be taken also as a posi- tive thing. and has bought the cake that has been brought in just prior to the excerpt..

thus taking off the sharpest edge from the complaint. which targets the basis of Kendra’s complaint. compliments the cake. Do you like frozen yogurt cakes? As a result of the talk that follows (see excerpt 8 below). even though the first intona- tion unit. ^frozen `yogurt? 10 (0. ^frozen `yogurt? (line 9). Kendra ignores Wendy’s attempts. disaligns with the complaint with a negative yes/no interrogative. she also provides an account for not liking ice cream by using reported speech (lines 18–19). 12 . I don’t `like ice ^crea=m (line 7). Wendy’s following turn. It is no- table that Wendy returns to the issue after some intervening talk by reproducing her question almost word-for-word.. . (line 1). or why she does not like ice cream. The interrogative turn also functions as a means for denying accountability for any mistake on Marci’s part for choosing a ‘wrong’ cake. It displays that Marci did not expect Kendra not to like the ice-cream cake that was just offered to her.3) 9 WENDY: Do you like . 13 do you wanna ^try <X it X> for `me? Kendra’s turn on lines 1–2 amounts to a complaint. she does not provide an account that would inform the others of what type of cake she does like. Further. the in- terrogative is produced in high pitch and with high-rising final intonation. On the level of form. for Kendra’s confirmation or disconfirmation. Kendra almost apologizes for having complained (lines 1–3). As Kendra’s mother. the negative yes/no interrogative is used to challenge Kendra’s complaint by calling into question her stance that she does not like ice cream.272 Tiina Keisanen 8 (0. on the level of content Kevin’s turn counters the complaint with rather bold sarcasm. this position is implicitly embedded in the complaint. It provides one candidate. Kendra’s response. which indicates some degree of affect being involved in the turn (cf. ^frozen `yogurt. and moves on to the blowing out of the candles (lines 11–13). Kevin’s response can be seen to provide an aligning response to the complaint as it offers an account (line 5). in turn. `Dad. Do you like . Don’t you ^like ice `cream? (line 6). It’s a ^beautiful `cake.6) 11 KENDRA: ^I shouldn’t `blow this `out. In short.. and it can be assumed that she chose the cake on the premise that Kendra does like ice cream. Marci. However. Both Kevin and Marci offer their response to Kendra’s complaint. However.. Marci should be familiar with Kendra’s likes and dislikes. can be seen to pursue such an account. as it displays that Kendra’s dislike of ice cream is news to her. asserts her stance explicitly. However. Selting 1994) (due to overlap it is unfortunately impossible to get an accurate pitch trace for the turn).

e.-- 6 WENDY: we’re all] [2.-- 2 `How do you -- 3 She just ^looks `pregnant? 4 [`Now]? 5 JAMIE: [She’s] ^pregnant. 14 KENDRA: saying. 4 WENDY: [<HI Well. 13 then5]. Ice [cream’s okay]. 15 WENDY: [(THROAT)] 16 KENDRA: [I hate] ice [2cream2]. 6 She’s ^totally `pregnant. 19 It makes me too cold. i. Challenging the prior speaker 273 (8) (Appease the Monster SBCSAE 0013 <T: 00. the negative yes/no interrogative presented a positive candidate understanding. 12 KEVIN: [4your4] [5fault. just HI>2] -- 7 MARCI: You’ve] [2never2] told us what you like and don’t like. Jamie has announced that the neighbor is pregnant with another child. 17 WENDY: [2@@2] 18 KENDRA: I hate ice cream. In the next example. 10 MARCI: you just go off to your3] roo[4=m4].19. Previously. 9 KEVIN: [3I guess that would be=3]. on the other hand. and Pete is their friend. the tag-question construction presents a candidate understanding of nega- tive polarity.. all of whom they find to be rather annoying. Harold and Jamie have been telling Pete at length about their neighbor and her children. In the excerpt below.. The conversation takes place between three participants. Harold returns to this piece of news. 5 MARCI: [I didn’t n. In the previous example. (9) (Lambada SBCSAE 0002 <T: 00:06:21>) 1 HAROLD: I `can’t t. 2 I mean I like -- 3 .. during the telling. the interrogative speaker displayed that she would have supposed the recipient to like ice cream. Harold and Jamie are a couple. 11 KENDRA: [4I think4] [5I walk around all5] the ti=me. . 8 [3dear.35>) 1 KENDRA: I don’t want to hur=t you=.

11 I `guess. 30 XX? 31 `Oh= ^Go=d.274 Tiina Keisanen 7 HAROLD: Oh.this just ^happened? (lines 10–13). Harold acknowledges this briefly with Oh (line 7). Harold there- by expresses doubt toward the accuracy of the news. 26 (0. 10 HAROLD: ^So=..this just ^happened? 14 Or. (1. We’re gonna have `babies ^crying (line . in turn.6) 25 JAMIE: `Yeah ^right. 18 HAROLD: [(GROAN)] 19 (1.thi. On lines 1–4.2) I `mean thi. she’s ^pregnant (lines 8–9)..VOX>.this. 28 <VOX ^shut up you `ki. He asks Jamie for the basis or the source of information for her announcement that their annoying neighbor is pregnant again. 15 JAMIE: We’re gonna have `babies ^crying.. Harold topicalizes the neighbor’s pregnancy that was first men- tioned some time ago.2) 13 HAROLD: I `mean thi. 9 she’s ^pregnant. It’s `not . It’s ^not . Harold is still doubtful as he continues to work toward creating an understanding of the state of affairs by presenting a request for confirmation ^So=. 12 (1. ^eating too `much. Jamie does not respond to Harold’s inquiry.thi. 16 (0. 8 JAMIE: .4) 17 JAMIE: [<P In the `middle of the ^night P>].this..4) 27 Probably be like. ^eating too `much. 21 `is it? 22 (1. 24 (0.0) 23 PETE: `Yeah but `now you’ll have ^both. Instead. However. continues to attend to the inadequacy of her response by excluding one of the most obvious counterarguments against pregnancy with . Jamie responds not by giving such an account.2) 20 HAROLD: Well it’s no ^worse than her ^screaming at em. 29 you `know. She’s ^totally `preg- nant (lines 5–6). Jamie. she moves on to complain about the situation by mentioning one negative outcome that the pregnancy will have on their lives. I `guess. but by stating simply that She’s ^pregnant.

Jamie adds. the different constructions are not interchangeable. Pete takes a turn to counter the challenge with `Yeah but `now you’ll have ^both (line 23). As such. The challenge sets the complaint in a wider perspective by referring to an earlier part of the conversation where Jamie herself portrayed the neighbor as a person who yells and screams at her children. However. and the ac- tual content of the turn may be seen to include elements that show support or the like. Moreover. during which no one responds (line 16). Pete counters Harold’s challenge by challenging its grounds. Harold uses the tag-question turn to challenge the appropriateness of Jamie’s complaint by indi- cating and/or reminding Jamie that the situation is already bad. thereby upgrading the complaint. In other words. in the present section the challenging tag questions and negative yes/no interrogatives likewise point toward some dis- crepancy in information between the participants. Harold produces a tag-question challenge that targets the basis of the complaint. In other words. Well it’s no ^worse than her ^screaming at em. the negative yes/no interrogatives and tag questions offer a disaligning response to the initiat- ing action. In response to initiating ac- tions such as complaints. seconds Pete with `Yeah ^right (line 25). no worse sets up an opposition with the prior turn.1. similarly to Section 3. The turn indicates that the two scenarios presented by Jamie and Harold are equally bad. the tag question is marked with turn-initial well to show disalignment or some discontinuity with the prior turn (Pomerantz 1984b). As was the case in example (7). instead of Jamie responding to Harold’s challenge. <P In the `middle of the ^night P> (line 17). for example. During the latter part of Jamie’s complaint Harold groans (line 18). it can be noted that in the current example Harold’s challenge is directed at the projected activity (complaining). After a pause. It can be noted that even though both negative yes/no interrogatives and the two types of tag questions are found to do a similar type of work in the two sections above. `is it? (lines 20–21). It presents a contrasting point of view on the issue by suggesting that it is not only the noise that the children make. However. no matter how much the neighbor’s new baby will cry. in turn. To sum up. In excerpt (2). the challenging turn consists of a negative declarative followed . the long pause (line 19) pre-monitors that the complaint will not get the align- ing response it invites. and therefore equally correct. In addition to being a disaligning second-position turn. and that it cannot get much worse. which shows recipiency and acknowledges that the complaint requires uptake. the discrepancy is used to challenge the relevance or the appropriateness of doing the action in the prior turn. and the evaluative verb screaming explicates Harold’s negative stance toward the noise the mother makes when communicating with her children. Challenging the prior speaker 275 15). but also the noise that comes from their mother screaming at them that should be taken into consideration when thinking about the overall noisiness of the neighbors. Jamie.

prior stances can be contested and challenged.. excerpts 3 and 4). The linguistic and prosodic turn-design features also contribute to the con- struction of disalignment. rather than align with the projected course of action. the challenges are regularly designed to negate or contradict a claim or position embedded in the prior turn. More specifically. which help to identify the problematic portion of the prior turn. These are some further indications of the disaligning and dis- preferred nature of the interrogatives discussed here. So. Future work on this topic is necessary to elucidate the exact differences in more detail. In about half of the cases in the database. The examples in the sections above provide evidence that sometimes dis- course participants can prioritize holding prior speakers accountable for their claims and stances or positions the have taken. the negative yes/no interrogatives in ex- cerpts (5) and (7). only (example 4). The turns may also include lexical markers.. . challenge a claim (explicit or implied) or stance of negative polarity. the challenges are placed in overlap with prior talk (e. in addition to sequential positioning. However.g. even though the preference for agreement (Sacks [1973]1987) limits the occurrence of disagreement in social interaction. on the sequential level such challenges can function as a vehicle with which to exert influence on the course of the un- folding action. these can be seen to function as restrictive stance markers in their local context. the smooth transition from a turn to a next with no gap and no overlap (Sacks et al. as the interrogative constructions discussed here demonstrate. Moreover. especially of high pitch. excerpts 5 and 9). Recipients are also observed to rather frequently withdraw or delay their response to the challenging turns (e. the differences between the constructions are not well studied. On the other hand. First. and no worse (example 9). 1974) is regularly disturbed by the challenging negative yes/no interrogatives and tag questions. the placement of the interrogative turns in the stream of talk also contributes to their interpretation. Sequential positioning is the central factor in the interpretation of the nega- tive yes/no interrogatives and tag questions discussed here as challenges. while the interrogative design is used for requesting the recipient to account for this claimed discrepancy. In terms of linguistic practices of stan- cetaking. the use of marked prosody. can be used to add an affective stance to the challenging turns. such as different (example 3). and is used to challenge a claim (explicit or implied) or position of positive polarity.g.276 Tiina Keisanen by a positive tag. for example. even implicitly. and other differences are also likely to exist. Howev- er. On the other hand. really (example 2).

in addition to which the prosodic turn-design features can be used to add an affective component to the turn. On the other hand. On the one hand. stancetaking can be conceptu- alized as alignment or disalignment between discourse participants with respect to the projected course of action or of the sequence. In other words. the challenging actions are used to display doubt toward prior claims or stances taken. such as. discourse participants . but interlocking levels. as well as turn-design features of the turns in ques- tion. Delays. What is notable is that the negotiation of alignment does not surface in the interaction explicitly. That is. In the data here. and reformulations are found also in the sequences discussed here. due to the action of challenging carried out with the interrogatives. the patterns of stancetaking discussed here expand on two different. They indicate troubles in turn transition. However. This may cause trouble for the recipient in interpreting the relevance of an interrogative. the information conveyed in statements of opinion often is. inter- rogatives that are responsive to telling sequences can be seen as opportunistic in assuming that a stance or a claim was provided for uptake when in fact it was not. However. The interrogatives discussed here provide the discourse participants with a resource for doing disalignment. but as I hope to have demonstrated. when the features of the aligning or disaligning actions are examined in more detail. but also toward others’ turns. repair initiators. delay devices such as pauses and repair initiators are com- monly found in disagreeing turns that are produced in response to turns that in- vite agreement. The data discussed here provide evidence for stancetaking as an intersubjec- tive achievement between discourse participants. By singling out some (explicitly or implicitly expressed) claim or position in the prior turn. Such sequences draw attention to the fact that discourse participants can be held accountable for anything that they produce in interaction. and point to- ward an interpretation that the recipients regard the interrogative turns as indica- tive of disalignment. it is still available to analysis through the examination of sequence organization. it is observed that actions may embody epistemic as well as affective stance displays toward not only one’s own turn. Challenging the prior speaker 277 4. even implicitly. As demonstrated by Pomerantz (1984b) in her discussion on agreement in assessment sequences. on the level of social relations these sequences involve the issue of alignment between the dis- course participants. the dis- tribution of silence and talk. Conclusion The speakers of negative yes/no interrogatives and tag questions discussed here initiate a sequence that ostensibly concerns some piece of information. for example. in telling sequences the troubles in uptake may also lie in the fact that the interrogatives foreground something that was not originally presented as contestable or challengeable information. specifically to challenge the projected or ongoing course of action.

Pentti Haddington. despite the label chosen for the action (cf.” Journal of Pragmatics 28: 69–101. P. I would also like to thank the participants of the 10th Biennial Rice Linguistics Sympo- sium held March 31 – April 3. Mirka Rauniomaa. 1997. Koshik 2003). . “‘Open’ class repair initiators in response to sequential sources of troubles in conversation. T. 2004 at Rice University where the ideas for this paper were first presented. and Metzger.278 Tiina Keisanen can call the interactional positioning implicated in the voicing of such claims or positions into question. However. as well as in the other cases. This chapter is based on a section in my dissertation research (Keisanen 2006). as sisters the three women might be expected to more or less know and remember each other’s prior teachers. Similar to example (2). as an epistemically less strong (skeptical) newsmark- er. then. Drew. for example. in this case the fact that the discourse participants are siblings seems to contribute to the challenging nature of the tag question turn. W. rather than being used. While outside factors to the ongoing conversation are not favored within conversation anal- ysis. References Beach. In other words. “Claiming insufficient knowledge. On occasion. high rising final pitch can be used to make the turn appear less hostile and/or challenging. Moreover. However. which it might well be between less intimate participants. 3. the high rising final pitch might be used here to mitigate the degree of the hostility component that has been associated with chal- lenging actions (Heritage 2002). the main argument here. Acta Uni- versitatis Ouluensis is gratefully acknowledged for allowing the reprinting of the relevant ma- terials. 1997. the participants in this example are close relatives (cousins) based on which Lenore can be seen to claim access to a shared knowledge base and on which she can build an epistemically strong challenging turn. All the remaining inaccuracies are mine alone. John Du Bois. the tag question turn is produced from a strong epistemic position and thereby carries out a challenge (cf. Heritage 2002). University of Essex.” Human Communication Research 23 (4): 560–585. Notes 1. “Stance-taking in reported speech. R. Pomerantz 1990). Maarit Niemelä. Robert Englebretson. Based on this knowl- edge base that seems to be evoked and claimed as common (cf. 2. taking a stance can be treated as an accountable action. 2000.” Essex Research Reports in Linguistics 32. One possible future direction for studies on these interrogative forms might thus be to examine the import of the height and direction of the terminal pitch in order to determine whether. Clift. Elise Kärkkäinen. De- partment of Language and Linguistics. and the Academy of Finland (project number 53671) for the funding of the research. Sandra Thompson and two anonymous reviewers have provided constructive critique and insightful comments on the paper for which I am very grateful. for example. is that the interrogative turn displays doubt towards the prior speaker’s claim or position.

Koshik. 2005.. Patterns of Stance Taking: Negative Yes/No Interrogatives and Tag Questions in American English Conversation.” Current Anthropology 43 (S4): S19–S35. L. “Time in action. R. Du Bois. . “A change-of-state token and aspects of its sequential placement.” In The New Psy- chology of Language. Uppsala University.).). Amsterdam: John Benjamins. and Englebretson.). S. Part 3. 2000. “Complaints about transgressions and misconduct. University of Pennsylvania. Kim. and Martey. “Wh-clefts and left-dislocation in English conversation: Cases of topicaliza- tion. P. and Thompson. and Paolino. Beyond rhetorical questions: Assertive questions in everyday interaction. 1995. J. J. M. 2001. C. S.W. 2003.A. J.” Language in Society 27: 291–334. To- masello (ed. Thompson. 2000. I. Epistemic Stance in English Conversation: A Description of Its Interactional Functions. D.). Finland: Oulu University Press. C. W.D. “Prosodic features of bad news and good news in conversa- tion. 1993. S. Vol. P. Mahwah.A.” Discourse Studies 5 (1): 51–77. Niemeier and R. Syntax: An Introduction. 247–296. CD.. Hillsdale. NJ: Lawrence Erlbaum.. Meyer. J. 2003. Pomerantz. Downing and M. Santa Barbara Corpus of Spoken American English.” In Word Order in Discourse. Meyer.” Research on Language and Social Interaction 31 (3. Hopper. 2003.. B. C. 2002. with a Focus on ‘I think’. Ams- terdam: John Benjamins. 1984. Du Bois. Keevallik. Schuetze-Coburn..W. A. 2002. J. Estonian Finite Verb Forms in Conversation. Koshik.W. Challenging the prior speaker 279 Drew. University of Pennsylvania. T. M. P.J. 1998. and Thompson. Santa Barbara Corpus of Spoken American English. Berlin: Mouton de Gryter. Amster- dam: John Benjamins. “The limits of questioning: Negative interrogatives and hostile question con- tent. Kärkkäinen. J. “Contingency and units in interaction. “An outline of dis- course transcription. Freese. Fox. Pütz. N. DVD. dissertation. 45–87. Amsterdam: John Benjamins. Ph.). Chafe. Department of Finno-Ugric Languages. Cognitive and Functional Approaches to Language Structure. Dirven (eds. Philadelphia: Linguistic Data Consortium. Philadelphia: Linguistic Data Consortium. From Interaction to Grammar. S. “Social interaction and grammar. 119–144. W. Santa Barbara Cor- pus of Spoken American English.. Heritage. Heritage.” Journal of Pragmatics 34: 1427–1446. C. Atkinson and J.A.A. NJ: Lawrence Erlbaum. E. Heritage. 2004. Lampert (eds. Cumming. 1984a.” In Struc- tures of Social Action.” Discourse Studies 6 (1): 27–52.-h.D. M. Ford. 1998.” Journal of Pragmatics 8 (5. J. J.” In Talking Data: Transcription and Coding in Discourse Research.. J. and Maynard. 109–129. 2003. 2004.” Language in Society 27: 195–219. Du Bois. Part 1. Du Bois.. 6): 607–625. T. Ford.W. University of Pennsylvania. “Wh-questions used as challenges. C. Acquisition and Language Peda- gogy. Goodwin. Acta Universitatis Ouluensis B 71. 2006. 1998. K. Philadelphia: Linguistic Data Consortium. Noonan (eds. Oulu.J. Heritage (eds. “Giving a source or basis: The practice in conversation of telling ‘How I know’. Chafe. Part 2. 299–345. Cambridge: Cambridge University Press. S. Studies in Conversation Analysis. 2. “Grammatical constructions and their discourse origin: Prototype or family resemblance?” In Applied Cognitive Linguistics: Theory. Edwards and M. CD. Keisanen. “Oh-prefaced responses of inquiry. 2003. D. 4): 295–325. Givón. I. S.

Thompson. M.” American Sociological Review 68 (6): 939–967. E. Schegloff. 1968.A. Typology. Clause structure. 2003. and Jefferson. 1977. and Sacks. P.” Language 53 (2): 361–382. Cambridge: Cambridge University Press. Cambridge: Cambridge University Press. J. H. Button and J.” In Language Typol- ogy and Syntactic Description.” Research on Language and Social Interaction 28 (3): 185–211. G. H. Lee (eds.” In Case. 309–341. “A discourse explanation for the cross-linguistic differences in the gram- mar of interrogation and negation.M. E. 57–101. “The preference for self-correction in the or- ganization of repair in conversation. T.. 1993) Units Intonation unit {carriage return} Truncated intonation unit -- Truncated word % . “On the preferences for agreement and contiguity in sequences in con- versation..R. “Sequencing in conversational openings.). Greenbaum. Siewierska and J. Heritage (eds. 1985. E. G.A. H. and Zwicky. E. Song (eds.” In Structures of Social Action. 54–69.M.M. Schegloff.A.A. Raymond.A. J.E. and Grammar. “Speech act disctinctions in syntax. 1995. and Sacks. A. Sacks. 1988. “Discourse as an interactional achievement III: The omnirelevance of ac- tion. Shopen (ed. [1973]1987.” Communication Monographs 57 (3): 231– 235. E. Schegloff.). Quirk. Wootton (eds.A. Jefferson. Cam- bridge: Cambridge University Press.). 1990. London: Longman.). Sadock.A.” American Anthropologist 70 (6): 1075–1095. “Agreeing and disagreeing with assessments: Some features of preferred/ dispreferred turn shapes. A Comprehensive Grammar of the English Language. Sequence Organization in Interaction: A primer in conversation analysis I. and Svartvik. 155–196. 1973. Studies in Conversation Analysis.. “Conversation analytic claims. G. A. “A simplest systematics for the organization of turn-taking for conversation.” Semiotica 7 (4): 289–327. Schegloff. A. 1998. Clev- edon: Multilingual Matters. “Opening up closings. Schegloff. A.J.” In Erving Goffman.. Leech.A. “Goffman and the analysis of conversation. H.” In Talk and Social Organization. Selting. 1974. 1985. Schegloff. Volume 1. 1994. Drew and A.). S. Atkinson and J. [1995] 2007. Exploring the interaction order. E. “Grammar and social organization: Yes/no interrogatives and the structure of responding. Pomerantz. J.280 Tiina Keisanen Pomerantz.” Language 50 (4): 696–735. Sacks. “Emphatic speech style–with special focus on the prosodic signaling of height- ened emotive involvement in conversation. G. Appendix: Transcription conventions (from Du Bois et al. 89–135. 1984b. R. S. Schegloff. G.” Journal of Pragmatics 22 (4): 375–408. Amsterdam: John Benjamins. Cambridge: Polity Press. E.

g. <@ six words @>) Quality Special voice quality <VOX VOX> Forte: loud <F F> Higher pitch level <HI HI> Parenthetical prosody <PAR PAR> Transcriber’s perspective Uncertain hearing <X X> Researcher’s comment (( )) Indecipherable syllable X . Appeal ? Speakers Speech overlap (numbers inside brackets index overlaps [2two words2]) [ ] Accent and lengthening Primary accent ^ Secondary accent ` Unaccented Lengthening = Pause Long and medium (N) Short (brief break in speech rhythm) .g. (DRINK) Glottal stop % Exhalation (Hx) Inhalation (H) Laughter (one pulse) @ Laughter during speech (1–5 words) @ (e. Continuing .g... (YAWN). (SNIFF). Vocal noises e. @two @words) Laughter during speech (+6 words) @ (e. Challenging the prior speaker 281 Transitional continuity Final . (TSK).


Stancetaking is here understood as an intersubjective activity in which interlocutors collaboratively construct and take stances about a stance object. which combines Du Bois’ (this volume) theory of stance with conversation ana- lytic methods of investigating the sequential and interactional aspects of talk-in- interaction and the grammar and grammatical practices for doing actions in in- teraction (Schegloff 1996). and turn organizational features of positioning and alignment. it is also a forward-type in- tersubjective activity. 2005a). Positioning or setting up a position is an activity in which the interviewer designs a question so that answering it poses difficult problems for the interviewee. presumption. Introduction1 This paper investigates two interactional stancetaking activities called positioning and alignment in American and British news interviews. Finland 1. When interlocutors take stances. and broader sociocultural beliefs and so on. and . it is worth consid- ering the contextual background and the indexical properties of language. they place themselves with regard to others and stance objects. but also engage and align with previous interlocutors and their stances (Du Bois this volume). or a presupposition that puts the interviewee in a quandary relative to her institutional background. audience expecta- tions. a publicly taken prior stance. Haddington 2005a). It represents a function- al-interactional approach to studying stancetaking (Haddington 2004. by engaging with and building upon stances taken by their co-par- ticipants in the immediately prior talk (Du Bois this volume. Consequently. A positioning question encodes or evokes a preferred stance. Positioning and alignment as activities of stancetaking in news interviews Pentti Haddington University of Oulu. Although the propositional context of the co-participants’ stance is not of interest per se. I concentrate on describing the linguistic. because it establishes and creates an interactional context for the interviewee. sequential.

2001. because talk in news interviews is largely composed of multi-unit turns (cf. how they also adopt adversarial stances and exert pressure on their respondents by incorporating third-party statements. In fact. Schegloff (1996) suggests that a linguistic vantage point can be used for looking at how speaker turns are composed of individual TCUs. and have thus given little attention to the design of entire questioning turns in news interviews (cf. and the actions within those TCUs. ten Have 2001: 6). Heritage and Roth 1995). Ford (2002. 2004) shows that it is fruitful to consider how recurrent action combina- tions and linguistic formats in turns are utilized for constructing co-participation in interaction. presuppositions. Nevertheless. and accusations in their questions (Berg 2003. how TCUs are sequentially positioned in their turns. previous conversation analytic (CA) work on news interview interaction has investigated how interviewers use various practices and actions in order to question their guests in different ways. transition relevance places (TRPs) and speaker turns. However. question delivery struc- tures. many of these studies have only described the ways in which these practices and actions are accomplished within a TCU. and how all this affects the understanding of the turns (Schegloff 1996: 64). Heritage 2003.3 It also en- ables us to consider how the combinations of TCUs. how they are related to each other. The starting point for looking at positioning is the under- standing of questioning as a complex activity in which several – and not single – turn constructional units (TCUs) affect the understanding of the question and the trajectory of the ensuing talk (Heritage and Roth 1995). Schegloff ’s (1996) and Ford’s (2002. Similarly. Clayman and Heritage 2002: 126–131. 2004) work provide a use- ful starting point for considering how the interlocutors’ stancetaking is organized across TCUs. contribute to the interlocutors’ stancetaking. which goes beyond the examination of individual TCUs. This body of work has described how interviewers on the one hand design their questions as neutralistic (Clayman 1988. Although the examples of the interviewees’ answers in this paper could be seen to be evasive. Heritage and Roth 1995. it is also fruitful to consider the ways in which they engage and align with the question.284 Pentti Haddington how they together with other practices invoke particular situated interpretations (Haddington 2004). The notion of alignment is . Heritage and Roth 1995) and the recurrent linguistic practices in them. it is also possible to understand the underlying motivations behind some instances of interviewee resistance. particular topical agendas. is particularly relevant for the inves- tigation of news-interview interaction. Heritage and Greatbatch 1991). By recognizing the practices by which interviewers set up positions for their guests. and on the other hand.2 Clayman and Heritage (2002: 209) refer to the possibility that combinations of practices in questions may strongly favor particular answers over others. Clayman and Heritage 2002. but do not discuss this possibility in their examples. This approach.

it has largely looked at them separate from the question. When this happens. And as Du Bois (2001) shows. Rather. by recycling linguistic forms. alignment. speakers bring two utter- ances into close relation to one another. Stancetaking in news interviews 285 here understood somewhat differently from its use in CA.c. Interlocutors use language for aligning with each other and therefore. it aims to explicate the range of possible types of convergent and divergent positions that interactants can take relative to each other. and prosody to construct their stances. this paper supplements previous work by examining positioning and alignment from the following vantage points: . interviewees use various types of turn-design features. this pa- per focuses on some aspects of the interviewer’s turn that act as an impetus for an indirect answer. It is noteworthy that by recycling elements. p. In sum. the interviewee’s alignment activity can be seen as a back- ward-type intersubjective activity. and a contingent achievement by which she responds to the position in the question. Clayman and Heritage 2002). interactants can frequently dis- play subtle differences between their stances. an aligning ac- tion is understood to be an appropriate or preferred next action. Greatbatch 1986). as Du Bois (2001) further claims. an aligning action to a request is an acceptance. change of topical agendas. One central element of linguistic alignment. Although conversation analysis has de- scribed such “evasive” interviewee conduct.4 One of the most important claims in this paper is indeed that it is the design of the question and the stance it incorporates that strongly affects the interviewee’s subsequent turn and the stance therein. speakers do not necessarily aim to display agreement (Du Bois 2001). is very much a linguistic process in which interactants use morpho- syntax. Rather. By recycling linguistic forms. As we will see. The notion of alignment here is also not synonymous with agreement. and some of the interviewees’ ways of responding to and answer- ing difficult questions. and a disaligning action is a refusal (Heritage 1984: 269). and question reformulations (Clayman 1993. lexis. and actions to respond to difficult questions and to intersubjectively engage with them. “how I put my stance in rela- tion to your stance” (Du Bois. Thus the connections be- tween the “formal” structures of language can naturally affect the way in which individual stances are understood in the interactional context (Du Bois 2001). even ones that at first sight seem disparate. In comparison. interviewer responses to such eva- sive maneuvers (Clayman 2001: 238–298.e. In news interviews.. interlocutors use linguistic forms dialogically in order to negotiate their stances and create new meanings.). i. is that speakers frequently use and recycle the linguistic elements that their co- participants have used. This paradigmatic connection between two utterances can generate new local meanings. In CA. these forms engage. as it is outlined by Du Bois (this volume) and adopted here. which fulfills the expectations raised by the previous action. For example. practices.

The data fragments are transcribed in the Dis- course Transcription (DT) style in which one line represents one intonation unit (Du Bois et al. I conclude that the positioning and aligning activities show that stancetaking in news interviews is not as emergent or performed as it is in everyday talk. to align with a proposition or presupposition incorporated in the question. evoking and encoding preferred stances. actions. action combinations. It describes a patterned and reusable linguistic practice that interviewees use to manage their stancetaking in news interviews. 2. Moreover. The corpus contains approximately 20 hours of news interviews from both the United States and Britain. I also identify some recurrent practices by which interviewees align with these positions. 2. The database used in this paper was collected between October 1999 and March 2004. which helps the interviewee not only align with the question but also construct a stance of her own. . I discuss how interviewers set up difficult positions for the interviewees by asking hostile questions. More specifically. This paper supports prior findings (Du Bois this volume.286 Pentti Haddington 1. The transcription conventions are given in the Appendix. 1993). It looks at how the combinations of practices. inter- viewees frequently use a stance marker (with cognitive or communication verbs) and a recycled element of the interviewer’s question. 3. Thus. Practices of positioning and alignment in news interviews In the following sections. Finally. It shows how the co-participants intersubjectively engage in stancetaking. the interviewee’s turn in these answers is organized into a negative-positive struc- ture. but rather that the interviewers come to the interview situation with certain predetermined questions and agendas in mind and the interview- ees – although they naturally construct their stances vis-à-vis the interviewer and the question – design their answers based on their own objectives and attitudes. and particular turn-organizational features are used as resources and contrib- ute to stancetaking in news interviews. the findings in this paper support the claim that stancetaking is a contingent achievement occasioned by prior stances. Kärkkäinen 2003b) that by taking a stance an interlocutor (in this case the interviewee) displays an under- standing of a prior stance (in the interviewer’s question). and by incorporating presuppositions in the questioning turns.

5 Nevertheless. 10 taxation of aviation fuel for example. 14 IR: [S-] -- 15 IE: (H) But in ^relation to=.-- 23 <A>Now I would</A> -- 24 →4 I wouldn’t argue that it’s too cheap.something. 11 (H) is ^dealt with by `international ^treaty..4) 6 IE: I think that aviation. 21 →3 IR: (0) Do you think it’s too cheap. the Secretary of State for the Department for Transport. Newsnight.. if he thought air travel (H) was too cheap. IE: Alistair Darling (033 / 1 / 0:34) 1 →2 IR: Well earlier I spoke to the transport secretary Alistair Darling. Stancetaking in news interviews 287 2. 9 Although . 2003: Cheap air travel IR: Jeremy Paxman.. 8 meet the= `costs of the environmental damage it ^%causes. (1) BBC2. 16 aviation costs in the last few years. 20 and passed that on to the passengers. 17 the reason it’s come ^down. 28 that it causes. 13 (H) %uh [u]nilaterally.(1. <MRC>it ought to meet the costs of </MRC>the environmental 27 damage... 12 and we ^can’t do sen. the interviewee does not bluntly evade the question but manages to align and engage with it. 25 (H) I ^do think though. which is from BBC2’s Newsnight.1 Positioning by hostile questions and interviewees’ practices  of aligning with them In example (1) below. Dec 17. 18 (H) is because. 3 →1 . 26 . 2 (TSK)(H) I kicked off by asking him. 4 ((CUT TO ALISTAIR DARLING)) 5 .. 19 the industry has ^dramatically cut its ^costs.I. 7 ought to=. (GLOTTAL) the ^difficulty there is that. Jeremy Paxman inter- views Alistair Darling. (H)(TSK) No I do. It presents an interesting example of how the interviewer positions the interviewee by designing a somewhat hostile question that the interviewee cannot answer di- rectly. 22 IE: . . The question is about whether the British government should intervene in the tendency of falling flight prices.

At first. he only reports some reasons why flight prices have come down and. the interviewer simultaneously uses the yes/no question (arrow 1) together with the interviewee’s identity (arrow 2) in order to put the interviewee in a problem- atic situation. in effect. Thus.that projects the utterance No I don’t think and goes on to modify the way he frames his stance. the inter- viewee moves (lines 15–20) toward the original topical agenda in the question. the interviewee chooses to evade the question altogether (lines 6–13) by raising the issue of the environmental damages that are caused by aviation (cf. By fram- ing his stance in this way. the interviewee has to consider the interests of the airline companies. After this. If it was not. the interviewee begins to align with the question. Then (in line 21) the interviewer orients to the interviewee’s dodging action and repeats the question. Clayman and Heritage 2002: 238–298). It is likely that the inter- viewee cut off a denial that was going to be produced as the stance marker I don’t think. he at first has problems in formulating his response to the interviewer’s hostile repeated question. The repeated question therefore resurrects the position setup in the first question with the additional element of holding the interviewee accountable for not answering the question the first time. It could be that the change of the modal auxil- iary is closely connected to the change of the verb as well. the interviewee refrains from producing a direct denial . He does this by producing a self-repair and begins to produce an alternative modal auxiliary wouldn’t. He does not finish the already started utterance No I do. The reported question is designed so that both of the projected alternative an- swers (an agreement or disagreement) are potentially harmful for the interviewee. but then changed it to a different stance marker I wouldn’t argue. this example provides a rare instance of how editing is used for changing the meaning of the question in news interviews and naturally has implications for the following analysis. which prefers cheap flight prices. 2003: 316) in line 24. How- ever. instead of recycling the verb think. in terms of positioning.6 Nevertheless. the interviewee uses the communication verb argue (Biber et al. the design of the subsequent interaction suggests that the original question cannot have differed greatly from the reported question. Especially the use of the it-pronoun in the interviewer’s question in line 21 seems to refer back to the (unanswered) previous question. Consequently (arrow 4). Clayman 2001. However. as an elected politician. It is not possible to know whether the wording of the actual question was the same.-. which of course makes giving a negative answer problematic. Then in lines 22–23. does not answer the question.288 Pentti Haddington The interviewer’s first question in lines 2–3 is a reported yes/no question. answering yes would obviously be counter to the general public. Judg- ing from the truncations and the hesitation in the beginning of the interviewee’s answer. As a transport secretary.I.

this volume.7 Diagraph 1 (from example 1): Cheap air travel 21 IR: you {do} think it ’s too cheap 22 IE: No I do. Stancetaking in news interviews 289 to the question (which I don’t think it’s too cheap would have done). In Diagraph 1 below. I don’t know seems to project more talk to come. and rather claims that he would not engage in an argument about the issue. and these expressions tend to precede complement clauses (Kärkkäinen 2003a. Stance markers with a first-person subject pronoun combined with a cogni- tive verb (e. these markers are used as epistemic/eviden- tial/evaluative fragments in contexts where a stance is going to be produced by the speaker (Thompson 2002). Therefore. claim) are not only frequent in news interviews. but also play a significant role in interviewee alignment. guess. However. Rauniomaa this volume. the interviewee recycles the phrase it’s too cheap (arrow 4) from the interviewer’s turn. the self-initiated self-repair and the new stance marker wouldn’t argue seem to reflect the interviewer’s quandary with respect to the position he has been put in. these types of stance markers occur in a specific context in news interviews: they pre- cede a complement clause that is composed of language recycled from the inter- viewer’s question turn. the collocation between a first-person subject and a cognitive verb is strong in every- day talk. a denial produced by the utterance I don’t think it’s too cheap would communicate a more personalized and a more certain stance than the denial I wouldn’t argue that it’s too cheap seems to express. argue.. who shows that in news interviews. Scheibman 2001). As many studies on different languages have shown.g.I- 23 Now I would -- 24 I wouldn’t argue that it ’s too cheap The above diagraph clearly demonstrates Greatbatch’s (1986) and Clayman’s (1993. think. and by Schegloff (1996: 61–62) who points out that in everyday talk.g. the stance marker I don’t think is used for conveying and framing the speaker’s subsequent opinion. know. he must word his stance carefully in order to take into account the airline companies’ and the public’s possibly differ- ing opinions. we can see how the interviewee uses the stance marker I wouldn’t argue and the recycled phrase in order to align with the question. 2001) findings that interviewees frequently recycle a unit from the question in . As the above example and the examples below show. it would seem that speakers are not specifically referring to themselves when using these markers.. In the above example. In other words. and accept) or a communication verb (e. This view is also supported by Simon-Vandenbergen (2000). Karlsson 2003. Rather.

prefer) particular types of responses (Hutchby and Wooffitt 1998: 44). It refers to the idea that in interac- tion particular actions invite (i. However. is constructed so that it displays a more explicit en- gagement with the question than his previous answer (in lines 6–20). and so on. This pattern is frequently used by interviewees for aligning with a difficult posi- tion posed by the interviewer’s question. Sacks and Schegloff 1979).e. The clearest evidence for preference organization can be seen in adjacency pairs. Heritage and Roth 1995). 2. Preference does not refer to the co-participants’ internal or subjective preferences. the interviewee rather than just evading the question. By repeating and incorporating linguistic elements from the question the interviewees show that they are attending to the question and on the surface level answering it (Clayman and Heritage 2002: 247. but rather to the structural features of how a turn or an action is produced in its sequential context. And even though the interviewee’s turn. answers that seem to just shift the topical agenda or evade the question altogether often in fact intersubjectively engage and carefully align with the difficult position the question sets up. I show how the interviewer sets up a position for the interviewee by designing the turn so that it evokes a preferred stance. Therefore. in example (1) above. whereas dispreferred turns often incorporate different types of dispreference mark- . Hutchby and Wooffitt 1998. in fact. displays careful alignment with the question and.290 Pentti Haddington order to shift the topical agenda or to evade the question altogether. and at the same time displays fine-tuned and careful alignment with the position in the question and avoids the pitfall that the interviewer laid out before him. Second pair parts in adjacency pairs that are structured as preferred tend to be unmarked. By using this pattern. answers it. this practice has a clear function. Sacks [1973] 1987. For example. it is the way in which the stance is formulated in the question that affects the design of the answer. an alternative reading that considers the impact of the question on the answer is perhaps more accurate in example (1). The engagement and align- ment of the answer with the question can be perceived in the linguistic pattern in example (1) and Diagraph 1 (stance marker + recycled unit from the question). In other words. assessments prefer agreeing second assessments.. an invitation prefers an acceptance.2 Positioning by incorporating preferred stances in questions  and practices of aligning with them In the following. Thus. The notion of prefer- ence is one of the central analytic concepts in CA (cf. the interviewee in the above example locates the trouble source. by carefully designing the answer so that it engages with the difficult question. he basically gives the same answer as he gave before the interviewer’s repeated question. In other words.

After this. preference as it relates to stancetaking benefits from considering at least some of the extra-tex- tual and participant-specific aspects in the analyzed segment. This third-party stance and the design of the question evoke a preferred stance in the response that the interviewee clearly orients to and aligns with. In other words. 2001: Bomb him out IR: Larry King. the interviewer. The third-party stance in the question preface is based upon recent polls on whether the United States should attack bin Laden in Afghanistan. This example comes from CNN’s Larry King Live. as is shown below. 2 could you=. Stancetaking in news interviews 291 ers (cf.. However. 14 a ^qualitatively different `response. 11 uh. formulates a yes/no question to the interviewee Brian Jenkins. 7 I d. IE: Brian Jenkins (001 / 5 / 3:49) 1 IR: . well I guess the `public.(0. Larry King. In the following example. 16 two acts of ^terrorism.-- 8 →2 I don’t know that you could ^bo=mb him out. 13 is going to `call for.. 4 →1 would look at this `simply.. Since stancetaking is not an action but a larger activity (Haddington 2004) the whole idea of prefer- ence as it relates to stancetaking requires reconsideration. a stance preferring a particular subsequent stance is a different issue from one action pre- ferring a certain type of next action. 5 Could you ^bomb him out. Sacks [1973] 1987). 3 →1 . the ^magnitude of this operation. the topic. This is particularly clear in news interviews in which negatively formulated questions are “routinely treated as embodying a very strong preference for a yes answer” (Clayman and Heritage 2002: 209). the interviewer asks a question that is framed as a third-party stance. (TSK)(H)U=h.6)(TSK) ^Brian. 6 IE: . 10 I I ^think that. (2) CNN. . Hutchby and Wooffitt 1998: 44. 12 ... Sep 12. Preference is also often signaled with particular use of syntax. 15 than we have seen in the ^past `two=. and where a sequence takes place. The question of whether a stance incorporated in the first turn can evoke a preferred next stance has not yet received much attention. 9 I mean.. Sacks ([1973] 1987: 58) rightly claimed that features of preference organization can be examined without knowledge of the participants. Larry King Live.

in this case.8 In the above example. questions) that invite second-pair parts. This is because although the question TCU Could you ^bomb him out. not only does the marker here occur in first position (a question- ing turn). It has been shown that interviewers use third-party statements in order to raise controversial topics. which the interviewer starts in line 2. of which according to the polls at the time 94 percent favored a military response. he now also has to orient to the third-party stance given in the question preface. presents the question not so much as the interviewer’s question. The complement clause is preceded by the epistemic stance marker I guess in line 3.. These dispreference markers and faltering talk are indicative of a stance that is divergent from the public stance . a position one does not seek willingly. which is indicated by the pause before the interviewee’s answer. and therefore. projects a yes/no question. not only does the question favor an agreeing answer (Sacks [1973] 1987).. in line 5 is a yes/no question and makes relevant either an af- firmative or a negative answer. a subsequent negative answer would mean that the interviewee disagrees with a large portion of the public and the TV-viewers. In other words. The sudden transformation of the question’s design and the invoked third-party stance change the way in which the question can be understood. How- ever. he abandons it and instead produces a TCU (lines 3–4) that frames the newly-produced question TCU (in line 5) as a stance by a third-party.g. The way in which the interviewer designs the questioning turn as a whole to favor a particular answer would seem to set up a difficult position for the inter- viewee. The evidence for this can be found in the beginning of the interviewee’s turn. and that I guess arises from some information or evidence that has been established in prior discourse or is based on a private reasoning process of the speaker. but the cut-off IU and the use of the third-party stance also mark the interviewer’s sudden shift from asking a simple information-seeking question to asking a question on behalf of the audience.292 Pentti Haddington The TCU. but since these stances are attributed to third parties. the third-party stance is placed sequentially before the question (at arrow 1). but as a stance that the majority of the general public share. Heritage and Greatbatch 1991). The third-party is mentioned in the complement clause in lines 3–4. In other words. Kärkkäinen (this volume) claims that I guess often frames actions in first position (e. the combination of the two actions) also strengthens the preference conditions for an answer that agrees with the question. but the question TCU together with this particular third-party stance (i. and the self-repair (line 7). He goes through some trouble in formulating an answer to the question. in addition to the fact that the interviewee has to answer the question.e. Indeed. Heritage 2003. and thus have important interactional consequences for stancetaking. the interviewers still maintain a neutralistic stance toward the topical agenda and the guest (Clayman 1988. the hesitation marker U=h (line 6).

the interviewee orients to the preferred stance and shapes the answer so that it disagrees with it as little as possible. It treats the third-party stance as doubt- ful or even hypothetical. The question concerns the so-called “bin Laden tapes” and whether these tapes include covert messages for inactive terrorists to take action and. who served in the Clinton administration. In front of the audience (the public referred to in the question). should not be shown on televi- sion. therefore. In light of this. after the stance marker.10 It was recorded in Decem- ber 2002. subtly incorporating his own stance. the use of the that-clause and the subjunctive contributes to the aligning activity in an interesting way. As Diagraph 2 below shows. even though the reported third-party stance was actu- ally based on topical polls. the interviewee now orients to the preferred stance by a subtle expression of doubt about the public’s view. The question also shows how the interviewee carefully aligns with this rather complex positioning activity. After this.” In sum. and designing the ques- tion to contain a preferred answer. the interviewee answers the question by ut- tering I don’t know that you could ^bomb him out (arrow 2). Example (3) below comes from CNN’s Crossfire. which in this context would display epistemic uncertainty or in- sufficient knowledge about the issue (Haddington 2005b). this is much more convenient for the interviewee than an explicit disagreement or an evasive answer would be. After this. The interviewer is Paul Begala. Diagraph 2 (from example 2): Bomb him out 5 IR: {you} could bomb him out 7 IE: I do. because the recycled part of the TCU is a finite dependent that-clause in which the verb is in the subjunctive mood. Stancetaking in news interviews 293 reported in the question. the interviewee again recycles an element from the question. the interviewee produces an explanation by introducing alternative actions to “bombing. In comparison to a conditional clause (I don’t know if you could bomb him out). the interviewer positions the interviewee by invoking a third-party stance. and the interviewee is Frank Gaffney. but they are sometimes used to express unreal or hypothetical meanings (Biber et al. a couple of months after the terrorist attacks on New York and Wash- ington.9 The use of subjunctives is very rare in spoken English. In this example. which establishes a connection between the question and the answer. . who is the former Assistant Secretary of Defense and an expert in foreign and defense policy.-- 8 I do n’t know that you could bomb him out The interviewees’ formulation of the TCU in line 8 is interesting. 2003: 261).

20 →3 (0) Is there <MRC>slightest shred of evidence.. 17 →1 ^potentially in these tapes. 8 →1 . I don’t think this is a matter of ^cryptography. 5 . 23 →4 .. 13 →1 . 21 →3 that she was right</MRC>? 22 →4 IE: . ^first that he would whip up -- 11 →1 ..7) (H) who are ^here in this country.. 29 I think this is a ^question of whether.. 12 →1 uhm.. 3 aired by Al-Jazeera ^today? 4 .. he goes on to report a third-party stance. 15 →1 and probably more ..(H)(TSK) When... 30 <MRC>people have gotten instructions</MRC>. 16 →1 (H) that there were <MRC>secret coded messages</MRC>. the interviewer introduces the stance object “the new bin Laden tape” in lines 1–3.(1. 2001: The new bin Laden tape IR: Paul Begala.. So. 28 →5 . Dec 27. which sets up the topical agenda and gives background in- formation to the question. 27 `no.294 Pentti Haddington (3) CNN. . IE: Frank Gaffney (003 / 1 / 1:13) 1 IR: .. that’s. <A>Now they’ve had some of these tapes for eight weeks. 9 →1 because `she ^feared. 25 →4 . 26 →4 . 31 . 32 or perhaps ^elsewhere in the world.(0.. At the beginning of his turn. The above interviewer’s turn is again composed of two main parts: the question preface (in lines 1–19). Uh the ^new bin Laden tape.. in lines 6–17. 19 →2 We have the best cryptographers in the world</A>. 6 →1 ... `ominously. uhv.... 33 . so 'fa=r. the President’s national security adviser. 14 →1 but then ^second. 7 →1 ^told the `networks.. you might deduce from that. 2 (H) .0) (H) (TSK) Well ^nothing.... and the actual question (lines 20–21). has been blown up. Crossfire. he ^began sending these tapes out.. 24 →4 .. 18 →2 ...7) They ^shouldn’t run these. (H) and will be prepared to ^operate on them.. anti-American `views. 10 →1 .(0... after which.

The preference for the negative answer is also strengthened with the superlative construction. the fact that the footing shift is sequentially positioned before the actual question frames the question in a par- ticular light. These words could be perceived as negative polarity items (Horn 1989). the interviewer’s action of bringing up a stance by a third party (at arrows 1) already builds up a connection between the third party and the interviewee. when the answer is considered. in example (3). The connection between the third party and the interviewee is made explicit in the question (arrows 3) in which the interviewer requests the interviewee to respond and take a stance in relation to the third-party stance (note the use of the pronoun she).) (arrows 2). Second. After this. when this inter- view was recorded. The negative polarity items and the superlative construction together suggest that in spite of all the intelligence resources. First. In December 2001. which presupposes that the issue is about cryptography (We have the best cryptographers in the world. these two TCUs indeed are contextually relevant and affect the way in which the following yes/no interrogative is understood. similarly to example (2). the question that ends the interviewer’s turn in lines 20–21 contains a preference. the interviewer makes a footing shift (Goffman 1981) and produces an assessment. as we will see later. This is further emphasized by the interviewer’s voice quality: each word in this unit is distinct and emphasized (marcato voice quality) and the interviewer’s pitch . Even though the interviewer in these TCUs does not (and cannot) express explicit disagreement with the third party. The question also contains the adjective slight in the superlative form and the noun shred. It is designed as a yes/no-type interrogative which narrows down the possible relevant answers to an affirmative yes or a negative no. Even though the question is not hostile or adversarial. he frames the question by pro- viding additional background information that undermines the reliability and the correctness of the third-party stance. the President’s National Security Adviser. Condoleezza Rice does not have any evi- dence to support her argument that the interviewer had reported in lines 6–17. what Condoleezza Rice. These two TCUs are produced in distinctly rapid speech. Nevertheless. the combination of the several practices in the question set up a position for the interviewee. has said about the stance object (that the networks should not run these tapes. Stancetaking in news interviews 295 i.. the interviewer concludes the turn by formulating an interrogative in lines 20–21. which usually embody a prefer- ence for a no answer (Clayman and Heritage 2002: 211–212). which suggests that he slips them in between the question preface and the question proper. which together invoke the idea of smallness and insignificance. and that there are potentially secret coded messages in them). Finally. Furthermore.e. politicians across the political field (as well as the majority of the American public and the TV networks) were almost unanimous in their sup- port of the President and his administration.

the question prefers a negative answer. the sequential proximity of the footing shift and question TCU. He does this again by first producing the stance marker I don’t think. Only after this does the interviewee continue to respond to other parts of the question (cf. the interviewee re- sorts to dispreference markers – the one-second pause and the adverb well – that display the trouble he has in responding to the question and the stance therein. he denies the presupposition in the question and its relevance in relation to the issue at hand. In sum. . This precedes and strongly projects a resolution or an account to the . However. Diagraph 3 (from example 3): The new bin Laden tape (question preface) 19 IR: We have the best cryptographers {} 28 IE: I don’t think this is a matter of cryptography The interviewee produces the denial by first locating a trouble source in the ques- tion preface and by repeating it. He does this by denying the presupposition about the relevance of cryptography (arrow 5). in his turn. the noun cryptography. the cryptogra- phy/secret coded messages-issue (lines 28–33). the third-party statement. In spite of the difficult position.) (cf. The interviewee com- plies with this by eventually answering no in line 27.) and induction (you might deduce from that. has been blown up. the issue regarding Condoleezza Rice (lines 22–27) and. Chafe 1986). The interviewee is basically asked to display his disagree- ment with the President’s National Security Adviser. can).. As we saw above. He further distances himself from the stance he takes by using the generic pronoun you and the modal verb might (cf. which at that historical mo- ment could have displayed the interviewee in an unfavorable light.296 Pentti Haddington is considerably higher than in the other parts of his turn. before producing the no-answer. but rather addresses and responds to both parts of it: on the one hand. . so ‘fa=r. By doing so. the interview- ee orients to the question’s problematic preference structure and tacks through it by designing the answering turn in a very careful manner. and then providing his view of what the issue is really about. even though the interviewee pro- vides the preferred answer no. the interviewee does not evade the question. fn 4 in Clayman and Heritage 2002: 106–107).. the interviewee carefully de- signs the first two TCUs (at arrows 4) so that he avoids encoding them as personal stances and rather relies on evidential information (nothing.. In addition to this. on the other hand. Thus. the two preceding TCUs frame the answer as a depersonalized stance. and the linguistic design of the question together contribute to the positioning activity that the turn is doing. In addition to this. which is followed by a recycled lexical item in a slightly modified form (cryptography) from the ques- tion in line 28.

. In spite of this. the two examples in this section have shown different ways in which interviewers incorporate preferred stances in the questions. which the interviewee relies on. Ac- cording to Levinson (1983: 167). and some are easier to deny than others. I concentrate in greater detail on how interviewers use presuppositions for setting up positions for the interviewees and how interviewees deny them. these practices turn out to be handy for expressing an alternative stance. the interviewees do not just avoid answering the questions. but also uses the interactional space he has been allocated to design and present the stance in his own terms. In sum. presuppositions are not the only practice used for setting up a position for the interviewee. and thereby put their guests in difficult positions to answer them. for setting up a position for the interviewee to take a stance. which is incorporated in an ut- terance. It is worth noting that in the following questions. In other words. but rather compositions of “given” information. at the same time. a presupposition. Stancetaking in news interviews 297 denied issue (Ford 2002. In sum. I briefly go through examples in which presuppositions are used for setting up a position for the interviewee. imply that he orients to the preferred stance in the question and agrees with it (although minimally). 2. Haddington 2005b). which. However. and as we saw in example (3). the following examples show that the subsequent speaker can orient to a presupposition. 203–208) note. interviewers can and often do embody presuppositions in the questions.3 Using presuppositions for positioning and linguistic practices  for aligning with them As Clayman and Heritage (2002: 127. on the one hand. In the following section. but in fact intersubjectively align with the position set up in the question. presuppositions are pragmatic inferences that are based on linguistic structure and are sensitive to contextual factors. “built-in” elements. After that. is used in the interviewer’s question for rendering the question sensible and understandable for the interviewee and the overhearing audience. Some presuppositions are closer to the surface than others. which is then provided immediately after the denial. and “self-evident” assumptions on which speaker stances are built. Here a presupposition is understood to be a background assumption. rather the interviewers actually use a combination of various practices to construct a problematic position for the guest. can affect the next utterance and the stance in that utterance. the above prac- tices. the interviewee not only orients to the preferred stance evoked in the question. and on the other hand. In other words. In the following. Clayman and Heritage (2002) claim that in news interviews presuppositions differ in terms of their “embeddedness” in the question. Although presuppo- sitions are not “explicit” stances.

IE: Christopher Whitcomb (001 / 5 / 1:24) 1 IR: Christopher. 12 →2 ^How do you . 13 →2 a ^bin Laden or `someone like a bin Laden. 14 who’s hidden. 16 and who ^may be under ^cover of the `country he lives in. my aim is to show that even though the answers can be claimed to be somewhat evasive. the alleged leader of the Al-Qaida terrorist organiza- tion.(0.8) (TSK)(H) Well `I don’t know about ^taking him out Larry. .. 18 But I think we have a ^lot of `options. the interviewees use the same linguistic practice (stance marker + linguistic recycling) for aligning with the question. As in the examples above. 11 →2 ^How `do you `do that Christopher.. (4) CNN. All in all. Here. 26 in the ^pa=st</L>. 17 →3 IE: . 23 (H) uh `go with the ^justice system. 9 in a sense. 4 wl. 22 and because `we ^have to=. 2001: Rid them from this planet IR: Larry King. 6 →1 we will=. Larry King Live. The question concerns Osama bin Laden. 15 who `moves arou=nd. 20 (H) The ^other=. 2 if it ^gets to the `point. 21 because this is a ^law enforcement `operation. the interviewer asks a question from Christopher Whitcomb. 7 u=h.-- 5 →1 `we will ^take them out. 3 →1 as we have ^promised that. the answers intersubjectively orient to the position set up in the question both linguistically and in terms of turn design.. 10 →1 We are at ^war with them. 8 <MRC>rid them from this planet</MRC>. ^take out a.298 Pentti Haddington I consider the resources with which the interviewees deal with the presupposi- tions. Example (4) below comes from Larry King Live. and whether bin Laden should be killed or taken captive and then brought to justice in the United States. a special agent for the FBI. 19 And one of them `obviously is ^military. 25 (H) <L>Uh we’ve ^done.. Sep 12. 24 and we want to bring `ultimately these people to ^justice.

Sacks [1973] 1987). which can also be called a shifter (cf. Haddington 2005b). which indeed marks the question to be inquiring about ways in which bin Laden can be taken out. we will=. which is made relevant through the action of promising. 29 (H) with the ^cooperation of those countries of course.11 Moreover. ^take out a. 30 and bring people ^back to the United States. Heritage 2003. First. 28 where we go into `other ^countries. the question preface contains stances that are attributed to we (arrows 1). which provides a contextual background for the question. . a ^bin Laden or ‘someone like a bin Laden.’ and thus raises potential problems for the interviewee. the interviewee does not share the question’s stances. namely as we have ^promised that. reported in the question preface. which by implication includes the interviewee. and the actual question in lines 11–16.” Therefore. First of all. in line 3. Jakobson et al.. the interviewee provides an ac- count (in line 18) by proposing other alternatives for catching and dealing with Osama bin Laden. u=h¸<MRC>rid them from this planet</ MRC>. The fact that this stance is here reproduced as an interrogative renders it presupposed information. The question contains two presuppositions. The question contains two distinct units: a question preface in lines 2–10.. nor its presuppositions. indexes a category that can loosely be equated with “Americans. the interviewer invites the interviewee to think about measures for ‘taking out bin Laden. is also used as a resource for constructing the actual question (^How do you . By building on the assumption that the interviewee shares both the stances in the question preface and the presupposition in the question. 1995. but designs his answer so that it contains a diver- gent stance. i. in lines 6–8. at the very beginning of his answer (arrow 3) the interviewee displays his divergence from the presupposition by denying it (cf. Levinson 1983: 184). which are not hostile (Clayman and Heritage 2002. it implies that the interviewee belongs to a group that is assumed to share the stance. and not whether it is reasonable or pos- sible to take him out.e. The recycled elements can be seen in the following diagraph. 2002) but nevertheless set up a posi- tion for the interviewee. This presupposition is also reflected by the prosodic design of the question TCU: the question word ^How in lines 11 and 12 has primary emphasis. As we can see in the example. After this. the third-party stance.. Sacks 1992. This pronoun. arrows 2). The divergence is also signaled by the long pause and the dispreference marker well (cf. by asking ‘how bin Laden or someone is taken out’ the interviewer presupposes that ‘somehow it is possible to take bin Laden or someone out’ (cf. Stancetaking in news interviews 299 27 (H) `operations which are referred to as ^renditions. Silverstein 1995).

the interviewee attempts to align with the interviewer’s positioning question by producing a denial. This issue caused a lot of dispute and controversy in Britain in spring 2004. the beginning of the interviewee’s resolution (But I think we have a lot of options) does not answer the question in the terms the interviewer set out in the first place. 4 how many of them. 5 . 2004: speaking for the universities IR: Jeremy Paxman. since the index of the first-person subject I (in line 17) is by implication included in the first- person-plural we (lines 3. and 10) used earlier. As in example (3). 3 (H)(GLOTTAL) Uh. Newsnight. Tim Yeo. Jan 21. The focus is on the question in lines 24–25.300 Pentti Haddington Diagraph 4 (from example 4): Rid them from this planet 5 IR: we will take them out 12 How you {do} take {a bin Laden} out 17 IE: Well I do n’t know about taking him out Larry The diagraph shows that in line 17. In other words. the recycled phrase take them out locates the trouble-source and is used as a resource for producing a denial to the presupposition in the question. it can be claimed that these two pronouns resonate with each other. which is a repeated question occasioned by the interviewee’s evasive answer (data not shown). 6. and thus indicate the stance differential between the interviewer’s and the interviewee’s turns. in contrast to example (4). and a (cognitive) verb (know). an auxiliary verb with a negative marker (don’t). Here.. In example (5) below. but by new terms the interviewee himself chooses. Moreover. This example comes from BBC2’s Newsnight. And in line 18. Jeremy Paxman is interviewing the conservative MP. . and the problem that the interviewee has with the indexical we. IE: Tim Yeo (037 / 1 / 7:18) 1 IR: Can I ask you just one small factual point. 25 →1 How many of the Vice Chancellor [s s]upport [2you2]. about the introduction of higher tuition fees in British universities. (5) BBC2. the interviewer holds the interviewee accountable for not answering the question by asking the same question again. 2 of the eighty-nine Vice Cha=ncellors. actually support your position? ((18 LINES OMITTED)) 24 →1 IR: You keep on claiming to speak for the universities. the recycled forms are again preceded by a stance marker that contains a first-person subject (I). the interviewer holds the interviewee accountable for not answering the question by asking the same question again. However.

This question is in itself probably difficult to answer. which amounts to coercing the in- terviewee to provide a relevant answer to the question. This is fortified by the design of the question. but it also seems improbable that the interviewee knows exactly which support him and which do not. Again. there are also other aspects that set up a position for the interviewee. Moreover. because not only are the Vice Chancellors numerous. example 1. if any indeed do. 29 but [I do claim]-- 30 IR: [You’ve been doing no]thing [2^but2] the entire [3ev3]en 31 [4ing4]. First of all. Note also that the primary stress on ^for in line 28 further em- phasizes the interviewee’s divergence from the interviewer’s claim. the interviewer’s questions are hostile. However. the stance marker with a communication verb (I don’t claim) precedes and is used as a resource together with the recycled unit to engage with the question. 32 IE: [2Uh2]. because they explicitly hint at a possible discrepancy between the stance that the interviewee has adopted (in line 24) and the reality (line 25) (cf. In spite of the fact that the interviewer is asking the same question again. the interviewee denies the assertion (arrows 2) contained in the question in line 24. Heritage 2003: 81). above). All the linguis- tic and interactional elements in the question intersubjectively set up a position for the interviewee that he needs to take into account in his response. presupposes that at least some Vice Chancellors support the interviewee’s views. . The sequential progression of the questioning. 35 (H) An[d]. both the original one in lines 1–5 and the repeated one (arrows 1). the interviewer’s question. As we can see in the diagraph below. the interviewee uses the phrase claiming to speak for the universities from the interviewer’s question for constructing his response. Stancetaking in news interviews 301 26 IE: [(GLOTTAL)] 27 →2 [2I2] don’t claim to 28 →2 speak ^for the universities. 33 [3%3] 34 [4I4] do claim to be concerned about their future. in which the verb phrase keep on claiming to (in line 24) suggests that the interviewee’s claims that are now submitted to critical evaluation have been continual and repeated. which then projects an answer that should give their exact number. in itself puts the interviewee in a difficult position (cf. and at the same time to undermine the claim in the interviewer’s question preface.

there is a strong and recurrent interactional-linguistic pattern by which interviewees respond to difficult ques- tions. their evasive answers are contingent products and occasioned by the asserted or presupposed stances in the question. 47 IE: [I am going to] answer it. in (3). not only sequentially precedes. It is . above) the interviewer responds with a forceful counter-argument. IE: Tim Yeo (037 / 1 / 7:51) 46 IR: You’re not going to answer that question [n. Newsnight. The trouble source is located and foregrounded by recycling it in the denial: in examples (1) and (2). The recycled element can engage with any element in the question. Interviewees recycle language from the question in order to orient to and undermine a stance encoded or evoked in the question. and (5).302 Pentti Haddington Diagraph 5 (from example 5): Speaking for the universities 24 IR: You keep on claiming to speak for the universities 27–28 IE: I do n’t claim to speak ^for the universities However. the interviewee locates a trouble source in the question (an assertion or a presupposition) and denies it. In this case. but also strongly projects a next action (more evidence for this is given in Section 5). By using a stance marker combined with a recycled linguistic unit from the interviewer’s question turn. the inter- viewee continues to produce an account of what his actions really are about (line 29). Indeed. These phenomena are closely related to the interlocutors’ stancetaking and especially to interviewee alignment with the questions. Jan 21. this time (contrary to example 4. stands as firm evidence of the backward- type intersubjectivity in news interviews (Haddington 2004). an assertion. a proposition in the question. which is composed of the stance marker in association with a recycled unit from the question. Rather. About four seconds later. The denial. Despite the interviewer’s intervention. the question TCU (Diagraphs 4 and 5) or both (Diagraph 3). As we have seen in this and the previous section. (4). 48 IR: are you]. the interviewees do not just evade the question for the sake of it. a presupposition in the question. the interviewee uses the denial + account action combination as a strategy for evading the hostile question. a presupposition or another problematic position in the question. The recycled language in this context is used by the interviewees as a linguistic practice to both orient to and undermine a claim. 2004: speaking for the universities IR: Jeremy Paxman. therefore. either in the question preface (Dia- graphs 1 and 2). the inter- viewer explicitly voices the fact that interviewee is not answering the question: (6) BBC2. This practice.

however. contrary to everyday talk. 2003. The second ac- tion. This is not an accident. By using it. This structure recurs in turn-initial position (Haddington 2005b). denials do not occur alone. In the following section. 2.4 The neg + pos pattern as a resource for alignment Interviewees frequently use the stance marker + recycled language. but also for denying a presupposition or a position set up in the question. this action combination is a robust and coherent dis- course structure in everyday talk. discussed in the previous section. which suggests that it is rare in everyday talk and indicates that it has a special function in news inter- views. Ford (2002) claims that in everyday talk a denial is actually only one part of a combination of two actions. 2000. According to Ford (2002). which follows the denial. the interviewees can design their turn so that rather than just bluntly evading the presuppositions or the preferred stances encoded and evoked in the questions. Du Bois et al. I show that interviewees use a particular turn-constructional format for aligning with preferred. is a correction or an account by the same speaker that gives an alternative interpretation to what has been denied (Ford 2002: 62). As was noted. this action combination is used exclusively by the interviewees and. this pattern is used for displaying engagement with the ques- tion. is frequently produced with the help of a linguistic pattern called the neg + pos pattern. This happens in news interviews as well. In my data. and presuppositional stances. The bolded parts picture the linguistic pattern. The following diagraphs depict the moments in which the interviewees first produce the denying actions and then proceed to give an account or a resolution for the denial. as a resource for aligning with the interviewer’s difficult question. Du Bois and Englebretson 2004) yielded only one example.12 Diagraph 6 (from example 1): Bomb him out 7 IE: I d- 8 I do n’t know that you could bomb him out . because in that sequential position it has a clear function in the organiza- tion of the answering turn and thereby also in the organization of the interviewee’s stance. A search of this pattern in the Santa Barbara Corpus of Spoken American English (Du Bois et al. evoked. Stancetaking in news interviews 303 an extremely efficient and productive practice for the interviewees to avoid the problematic position established for them in the question but at the same time is an efficient resource for constituting and organizing a relevant stance in their an- swer. I turn to a more detailed linguistic analysis of what happens in the second parts of the action combination. they actually align with them. However.

which is followed by a recycled linguistic element that identifies the issue that is being denied. but not always.I- 23 Now I would -- 24 I wouldn’t argue that it’s too cheap. The second part doing the account. The pattern can be schema- tized in diagraph form as: “I” + negative particle + cognitive / communication verb + recycled language “I” + verb + predication .304 Pentti Haddington 9 I mean 10 I I think that 12 the magnitude of this operation Diagraph 7 (from example 2): Rid them from this planet 17 IE: Well I don’t know about taking him out Larry 18 But I think we have a lot of options Diagraph 8 (from example 3): The new bin Laden tape 28 IE: I don’t think this is a matter of cryptography 29 I think this is a question of whether Diagraph 9 (from example 4): Cheap air travel 22 IE: No I do. 25 I do think though. Quite often. which is followed by the account. 26 it ought to meet the costs {} Diagraph 10 (from example 5): Speaking for the universities 4 IE: I do n’t claim to speak for the universities 6 but I do claim 9 I do claim to be concerned about their future The neg + pos pattern not only plays an important role in the organization of interviewees’ multi-unit responses.13 The first part doing the denial contains a first-person-singular pronoun (or sometimes some other pronoun) and a ne- gated cognitive or communication verb. but also organizes the interviewee’s intersub- jective stancetaking. then contains a first-person-singular pronoun with a cognitive or com- munication verb. It is composed of two parts. which always occur in different intonation units. these two parts also follow each other in consecutive intonation units.

This suggests that this practice in this interactional context can only be used if the unit that is recycled does not contain negative markers. 2 after ^writing that `manifesto. IE: Tony Blair (036 / 1 / 3:34) 1 IR: At `what ^point. 7 →2 <A>I ^don’t `accept that we have broken the promise. Well I. which. the interviewee recycles the structure of the phrase keep the promise (arrow 1) in the interviewer’s question but changes it into broken the promise (arrow 2). It is . influence the design of the recycled lin- guistic elements in the complement. until ^after the next general `%election. Consider the following deviant case: (7) BBC2. Newsnight. In the above example. 10 . One reason for this might be that the denial that projects the account al- ways contains a negative marker and that a positively formulated utterance in this context is easier to recycle.. only occur in this interactional context. 5 →1 (H) that you couldn’t keep the promise? 6 IE: (TSK)(H) . of course. 9 because the `new system doesn’t come into ^effect. Jan 19.. and that the stance marker in the denial is a “compulsory” part of the combination and can. of course. However. 3 in which <MRC>you requested our votes</MRC>. the interviewees deny an assertion or a presupposition in the interviewer’s ques- tion and then provide the next relevant action. 8 (0) as I said</A>. in cases in which the recycled linguistic unit comes from a negatively formulated TCU. communicates an opposite meaning com- pared to the original phrase. but it is a recurrent and routinized pattern in political news interviews and reproducible in this interactional context. in fact. Stancetaking in news interviews 305 It is worth noting that the original linguistic unit that the interviewee recycles is rarely negatively formulated. By using this particular pattern..14 The neg + pos pattern does not. the recycled element can be slightly modified. Consider the following diagraph: Diagraph 11 (from example 7) 5 IR: that you could n’t keep the promise 7 IE: I don’t accept that we have broken the promise This change shows that the denial + account action combination has a strong internal organization. 2004: Keep a promise IR: Jeremy Paxman. an account for the denial. 4 did you `realise.

e. This is particularly true in news interviews in which denial alone would be perceived as inadequate. and latching onto previous unit). Schegloff 1982). in some cases (see the pitch curve in Figures 2 and 5). it is possible that when an interviewee produces a denial. As Ford (2002) claims. Following Ford (2002). In example (1). These phonetic features further suggest that the first part projects a move toward the second part. a denial alone strongly proj- ects a resolution or an explanation component. Further evidence for the connectedness of the two actions can be heard in how this pattern is designed prosodically..306 Pentti Haddington worth noting that even though this pattern occurs in environments in which the interviewee does not produce a preferred answer to the question. Con- sequently. the interviewees minimize the interactional space at the TRP in which the inter- . As was noted above. Consequently.. i. asserted. Therefore. which also projects further talk from the speaker. The interviewer can use these features as a resource for understanding what the interviewee is doing in terms of turn projection. the interviewee produces a rush-through (Local 1992. Thus. they also serve an interactional function. i. In other words. in order to keep the turn. this action combination and the linguistic pattern in it could be perceived as a turn-constructional format. because inter- viewees are expected to support and give grounds to their statements. In all examples except (1) (see Figures 2–5). but in fact engage with it. the second part of the action combination is latched onto the first part. Although these features display the close relationship between the two TCUs within the interviewee’s turn. a pattern in which two ac- tions are closely connected.e. undermines it and goes on to propose another way of looking at the issue. rising intonation. a resolution or a counterstance. This actually happens in example (5) in line 30. or presupposed stance in the question. the end of the first part is produced with distinctly rising-continuing intonation. that the interviewee is going to produce more talk beyond the next TRP. which also projects more talk to come (see Figure 1). she aligns with the question. By producing the action combination pro- sodically as they do (no pause. as the acoustic measurements suggest. which helps him get to the next TCU. the interviewer can intervene after the denial and challenge or disagree with the interviewee’s denial. these two actions are not just individual actions. the denial projects a move toward the second part. but they organize a larger intersubjective stancetaking activity in which the interviewee responds to an evoked. when the interviewer says [You’ve been doing no]thing [2^but2] the entire [3ev3]en[4ing4]. the interviewee produces a clearly audible inbreath. the interviewees do not just bluntly avoid answering the question. this pattern shows that the interviewee orients to the position (or some part of it) set up in the question. interviewers have the right to interrupt or challenge the interviewee if they feel that she is not answering the question. In addition to this.

the interviewee is taking a stance based on an assertion or presupposition in the question turn. as we can see in example (5). However. in the second part. In other words. Rather.. Another interesting point that arises in these examples is that there is a ten- dency in how this pattern is produced prosodically. the questioners end up using a form that answerers can more readily agree with. and thereby aligning with it. this turn-internal orga- nization of the two actions is very strong. In addition to this. In all of the examples above. Stancetaking in news interviews 307 viewer has an opportunity to intervene. i. As we have seen.  Prosody in example (1): Cheap air travel . the maximum pitch in the first part is produced on the negative modal or aux- iliary verb (either don’t or wouldn’t). As was claimed above. an interesting question arises: What motivates the use of this pat- tern in news interviews? Sacks ([1973] 1987) noted that questions tend to prefer agreement. the primary emphasis tends to be on the first (cognitive or com- munication) verb or an accompanying modal. These similarities in the prosodic design of this action combination act as evidence of intersubjective stancetaking. Furthermore. the whole issue of preference is turned upside-down. the interviewee continues and finishes the second part of the action combination. in news interviews the interviewers rarely modify their questions for the interviewees.e. In other words. He also noted that if answerers treat questions as problematic and show this by using various dispreference markers. Moreover. the Figure 1. the practices in the action combination and the action combination itself are contingent achieve- ments made relevant because of the position set up in the question. the unit that is recycled from the interviewer’s question also receives strong emphasis. which shows that despite the interviewer’s intervention. this linguistic pattern almost never occurs in everyday talk. questioners tend to modify the ques- tion so that it contains the opposite preference. Therefore.

308 Pentti Haddington Figure 2.  Prosody in example (2): Bomb him out Figure 3.  Prosody in example (4): Rid them from this planet .  Prosody in example (3): The new bin Laden tape Figure 4.

formulate. 3. Stancetaking in news interviews 309 Figure 5. they can reduce the severe consequences that a strict disagreement might bring with it. Since the interviewees are forced to answer the questions. the above linguistic pattern is a useful resource for the interviewees to align with the question. defend. Against this back- ground. One purpose of this question-answer activity is to give information. and negotiate their stances.g. It has also shown that a linguistic analysis combined with the interactional analysis of . news interviews are the venues for politicians and other experts to pub- licly convey. they can produce a stance that diverges from but still does not totally disagree with an assertion or a presupposition in the question. Because of the pressure. Discussion and conclusions News interviews are publicly broadcast interactions in which the journalist’s pri- mary task is to ask questions of a public figure. In other words. but indeed orient to and engage with each other’s stances. it shows that the methodological approach to study stancetaking with data from real interactional situations is able to provide new and interesting find- ings about the ways in which speakers take stances. Scannell 1991: 4). This paper makes the following claims about stancetaking in news interviews. and thus engage in intersubjective stancetaking. First of all. and discuss and debate topical and often controversial issues. It also shows that speakers do not (just) express their subjective stances.  Prosody in example (5): Speaking for the universities interviewers can deliberately position their guests by designing their questions so that they prefer a problematic stance. The public figure is then expect- ed to provide appropriate answers to the questions (e. they have to align with the problematic stance in that question im- mediately. express opinions.. By designing the answer in this way.

One practice for denying an element in the question is the above-mentioned de- nial + account action combination in which interviewees use the stance marker (usually personal pronoun + negative particle + cognitive/communication verb) together with recycled language. In some cases. because they establish a fairly loose agenda and project a broader answer than. The positioning activity is essentially a one-directional and forward-type inter- subjective activity. recent research on stancetaking. for example. the turn- type pre-allocation. Potter 1998). By setting up a position the interviewers attempt to constrain and delimit the possibilities for the interviewee to construct a responsive stance. Second. the inter- viewer may be seen to be positioning the interviewee. but rather work together in interaction in order to reach some sort of a joint and negotiated stance (Du Bois this volume. for example in producing the denial + account action combination. the special (institutional) turn-taking organization in which sequences of questions and answers follow each other. There are also some general observations that can be drawn from the analysis above. 2003: 69). the fact that interviewers produce questions and inter- viewees produce answers means that only interviewers do the positioning and the interviewees the aligning. First. this linguistic pattern is a routin- ized and reusable linguistic pattern in this interactional context. but the interviewee does not treat the question as problematic. Heritage 2002. The interviewees indeed can decode pre- ferred stances and presuppositions in the questions and orient to and deny them. however.e. which have a fairly clear preference structure (cf. Of- ten (but not always) simple information-seeking wh-questions position the inter- viewees only minimally. i. As was shown above. if at all. It is also important to notice that positioning becomes clearly visible only if the interviewee treats the questions as doing positioning. the long multi-unit turns – which provide for the possibility that individual TCUs within a turn begin to resonate with each other – produce action combinations and combinations of linguistic practices that become relevant for the production of these activities.310 Pentti Haddington the data can provide new findings about stancetaking. negative interrogatives or tag ques- tions. attitudes and evaluation sug- gests that participants in everyday conversation do not converse with particular pre-determined stances in their heads. Second. Kärkkäinen 2003b. First of all. these activities are made possible by three special features of news in- terview interaction. the findings in this paper suggest that stancetaking is different in news interviews. The linguistic analysis es- pecially shows that participants in news interviews rely on recurrent linguistic formats and patterns in specific interactional situations. As was shown. Third. this paper has described some practices by which interviewers set up positions for the interviewees and how the interviewees align with these positions. Third. It is worth noting. that not all questions position the interviewees. However.. interviewers .

it is better not just to assume that interviewees evade questions. what their background is. On the contrary.e. and . The linguistic pattern discussed above stands as good evidence of this. tell news for the audience. or contextual. and what kind of answers can be expected from them. create entertain- ment.’ ‘what you say. and previously stated stances. but rather to talk about an aligning activity by which interviewees attempt to find a place to carefully word their stances that not only take into account the evoked and presupposed stanc- es. although the design of the interviewee’s stance is obviously contingent upon the design of the question. views that they have very likely had before they came to the program. Stancetaking in news interviews 311 use various practices and actions – ranging from the use of particular pronouns to different types of action combinations – for putting the interviewees between a rock and a hard place. report opinions and political viewpoints. As Heritage (1985) has shown. background. and for expressing the stances that better represent their views of the topical matter at hand. stancetaking in news interviews is perhaps even more about ‘who you are. In other words. Not only are evasive or agenda-shifting ac- tions often occasioned by the interviewers’ questions and therefore intersubjec- tive.’ than it is in everyday talk.. i. We could therefore call into question whether the stances in news interviews are as emergent as they are in everyday talk. political. The way in which these practices. the fact that in- terviewers sometimes reformulate interviewee answers in news interviews shows that news interviews are produced for an overhearing audience. It seems that stancetaking in news interviews relies on some types of prior beliefs. but also reflect the speaker’s own identity. It would seem that the stancetaking activities of positioning and alignment are also targeted for the audience in that they produce and distribute information. and cultural values that are bound to the chosen topical agendas and the viewpoints assumed by the participants. for example. They use their knowledge of who the interviewees are. it seems that the interviewees do not assume and accept the posi- tions and stances put on the table by their host but firmly state and support their own stances. The stancetaking activities described above also have interesting implications for the audience. but they are also contingent achievements occasioned by the evoked and preferred stances and presuppositions in the questions. actions.’ and ‘how you say it. and the com- binations of them function in the question turns suggests that interviewers come to the interview situation with particular questions and agendas in mind. for constructing their questions so that the answers the interviewees would prefer are very hard to implement. The idea of positioning as an intersubjective activity also forces us to recon- sider the idea of interviewee evasion. Consequently. aims. who or what they represent. The interviewees use this pattern for vitiating preferred stances and presup- positions.

i. the possibility for co-participants to real- ize and understand what is under way and to project the possible completion point of a TCU (Sacks et al. I am greatly indebted to the following persons for help. and Mirka Rauniomaa.. comments and encouragement: John Du Bois. 14.. Nuolijärvi (1994) discusses how the interviewer’s turn in fact creates the foundation and thereby acts as the sequential impetus for the interviewee’s non- answers. Turn constructional unit (TCU) is the basic unit out of which speakers set out to construct talk (Sacks et al. 1974). Heritage 1985). In institutionalized interaction.g. phrasal. TRPs refer to the ends of TCUs. do not apply to news interview interaction in the same way (see for example Clayman and Heritage 2002. I also want to thank the anonymous reviewer for helpful comments. together with the analysis of examples. These linguistic characteristics of turn construction provide for “projectability of a turn. Arja Piirainen-Marsh. Elise Kärkkäinen. and Clayman and Heritage (2002: 188–237) address the issue occasionally. the allocation of turns is pre-determined. that there is a possibility for transition between speakers at the end of a TCU (Sacks et al. Tiina Keisanen. Greatbatch 1988. or clausal. Therefore. these ac- tivities are characteristic of news interviews. and in which different rules of “current speaker selects next” or “self-selection” apply at each TRP. However. in news interviews. It seems. Robert Englebretson. TRPs are more “neutral” than in everyday talk (cf. we can only speculate about the significance of news interviews as popular media and their function in and impact on society and culture as a producer and bearer of particular beliefs and value-systems. Maarit Niemelä. Schegloff 1996: fn. In everyday conversation. However. 2.. 1974).” i. but are not likely to occur in other forms of question-driven institutional interaction (e.e. however. Although there are only scant mentions of how the interviewer’s question acts as an impe- tus for the interviewee’s evasiveness. the turn-taking rules that apply to everyday talk (Sacks et al. 4. but only in passing. interviewees are entitled to and even assumed to give answers that contain more than one TCU. Naturally. I alone am responsible for any errors and inadequacies that re- main. doctor-patient interac- tion). 3. but the fact that news interviews are publicly broadcast programs and aired for an audience. it would seem likely that on some level these situated activities may have an impact on how the individuals in the audience form their opinions about the topics in news interviews. This research has been partly funded by a grant from the Academy of Finland (research project 53671). sentential. 1974). in which one speaker is basically allotted one TCU. each speaker gets the right to construct a single TCU to a possible completion.312 Pentti Haddington thereby also (attempt to) influence and shape public opinions. therefore. 1974). This is a consequence of the ‘multi-unit- ness’ of turns and the pre-allocated turn types in news interviews.e. Notes 1. One TCU can constitute a recognizably complete turn and can char- acteristically be lexical. . For example. In fact. Although it is quite possible that one rationale behind producing these activities is the entertainment value of ne