Kärkkäinen 2006 Stancetaking in conversation From subjectivity to intersubjectivity

Stance taking in conversation:From subjectivity to intersubjectivity

ELISE KARKKAINEN

Abstract

In this paper I argue that stance in discourse is not the transparent linguis-tic packaging of internal states of knowledge, but rather emerges fromdialogic interaction between interlocutors. Thus, stance is more properlyviewed from an intersubjective vantage point, rather than being regardedas primarily a subjective dimension of language. I begin by outlining pat-terns of epistemic stance marking within the speech of single speakers andshow that these arise from the intersubjectivity between conversational co-participants. Then, I focus on stance taking as a joint activity between par-ticipants in story reception sequences and demonstrate that stances oftenemerge as a result of joint engagement in evaluative activity. Finally, I con-centrate on how the particular linguistic resources used for stance taking fitinto the intersubjective pattern by demonstrating syntactic, semantic, andprosodic resonances between contributions by dierent speakers.

Keywords: stance; subjectivity; intersubjectivity; I think; stance taking;resonance.

1. Introduction

It has become widely accepted within linguistic anthropology and conver-sation analysis that meanings are co-constructed and social in nature. Forexample, in their seminal article Goodwin and Goodwin (1992) investi-gate the collaborative production of assessments (or the evaluation ofpersons, events, objects, etc., being described in talk), during which par-ticipants can calibrate their separate evaluations of events in their phe-nomenal world and intricately demonstrate how their minds are in tunewith each other (Duranti and Goodwin 1992: 149, in their introductionto Goodwin and Goodwin). But within linguistics, even of the functional

18607330/06/00260699 Text & Talk 266 (2006), pp. 699731Online 18607349 DOI 10.1515/TEXT.2006.0296 Walter de Gruyter !"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"6759

4$'&/1'-:7'/6(+(;;?>;>==?@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

variety, there has until recently been a fairly resolute adherence to single-speaker contributions as loci of meaning. Similarly, in linguistic studieson what has variously been termed modality, evaluation, attitude, aect,subjectivity, or stance, the focus has been on utterances produced by sin-gle speakers. Many recent treatments regard stance from the point ofview of the individual speaker, i.e., as a static and isolated mental posi-tion or interior state of an individual speaker, who in expressing herstance draws from a more or less definite set of (primarily) grammaticaland lexical markers.However, in this paper I will show that stance is not only constructed by

grammatical or lexical means, but that the sequential occurrence of stancemarkers and the degree of syntactic, semantic, and/or prosodic parallel-ism or resonance across speakers is also a resource for stance taking.1

Relatedly, I argue that stance is not primarily situated within the minds ofindividual speakers, but rather emerges from dialogic interaction betweeninterlocutors in particular dialogic and sequential contexts. Thus, stance ismore properly viewed from an intersubjective vantage point, rather thanbeing regarded as a primarily subjective dimension of language.

2. Purpose

Stance in linguistics, then, has been treated as predominantly a matter ofthe expression of internal psychological states of an individual speaker.Epistemic modality, or epistemic stance, has commonly been conceivedof as manifesting some of the most apparently subjective marking lin-guistic resources. Among such linguistic resources, I think can be seen asa sort of prototype of subjective language, because it contains reference tothe speaker (the first-person pronoun) and a verb that denotes a private orinterior cognitive process. Kockelman (personal communication), how-ever, warns against automatically assuming that reference to the speakersmental states can be taken as the prototype of subjectivity. His pointis well taken; I think in my data of spoken American English from a vari-ety of interactional settings often appears in sequential contexts whereits referential meaning is quite vague and bleached (to the point of dele-tion of the first-person pronoun), and clearly acts as a discourse markerthat simply frames an upcoming stance or marks boundaries within aspeakers speech (see also Notes 9 and 14). Nonetheless, I think continuesto be considered prototypically subjective by many linguists (and Englishspeakers), who consider it to introduce an internal, private mental state ofthe speaker. However, in this paper my emphasis is on how epistemic andother markers are used by speakers in actual interactions (and what other

700 Elise Karkkainen

!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

Camille Debras

Camille Debras

Camille Debras

Camille Debras

Camille Debras

Camille Debras

Camille Debras

kinds of linguistic resources there are for stance taking in interaction),without assuming a priori a subjective or stance function for any lin-guistic marker. On the basis of a close sequential analysis of function, itappears that we have to remain open to the possibility that a linguisticmarker that looks highly subjective may actually not function to displaya stance at all in many contexts of use.My data come from Part I of the Santa Barbara Corpus of Spoken

American English (SBCSAE; Du Bois et al. 2000), which consists of 14conversations or 6 hours of digitized audio and transcriptions. My focushere is on informal everyday conversation. The data corpus was tran-scribed using the conventions in Du Bois et al. (1993; cf. Appendix). Thedata are transcribed into intonation units, or stretches of speech utteredunder a single intonation contour, such that each line represents one into-nation unit (Chafe 1979, 1987, 1994).To reiterate, my main aim in this paper is to show that stance is a pub-

lic action that is shaped by the talk and stances of other participants insequentially unfolding turns-at-talk. In what follows, I will begin (Section3) by reviewing some recent linguistic work on subjectivity and stance,which has established some pervasive linguistic and discourse patterns innatural discourse data. In Section 4, I briefly present the findings of anearlier, primarily quantitative, discourse-functional study of the most fre-quent manifestations and patterns of epistemic stance within singlespeakers contributions in naturally occurring American English data(Karkkainen 2003a). Then, in Section 5, using examples of I think I dem-onstrate that the pervasive patterns established in the speech of singlespeakers can be understood to arise from interaction and the intersubjec-tivity between conversational coparticipants. In Section 6, I turn towardmapping out some intersubjective, dialogic practices of stance taking,while also including instantiations of nonepistemic stance in the analysis.Finally, in Section 7, I return to the level of language and linguistic form,and show some of the particular linguistic resources deployed by Englishspeakers when they are involved in joint stance taking over longer conver-sational sequences.

3. Subjectivity and stance in linguistics

Let us start with a (linguistic) definition of subjectivity, which comes fromFinegan (1995). He claims that subjectivity is the

. . . expression of self and the representation of a speakers (or, more generally, alocutionary agents) perspective or point of view in discoursewhat has beencalled a speakers imprint. (Finegan 1995: 1)

Stance taking in conversation 701

!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

Camille Debras

Camille Debras

Camille Debras

Camille Debras

Subjectivity, then, refers to the phenomenon that the speaker with her at-titudes and beliefs is present in the utterances that she produces. In otherwords, rather than simply describing an event or presenting an objectivestatement of some event or state of aairs, the speaker represents an eventor state of aairs from a particular perspective. When the speakers eval-uative, aective, and epistemic perspective is reflected in her actual lan-guage choices, we commonly talk about the expressive, emotive, aective,or attitudinal function of language, as opposed to the referential, cogni-tive, or descriptive function.2 The notion of subjectivity has generally notbeen a very precise one in linguistic investigation, despite the fact that ithas been a topic of some interest for several decades, as evidenced byTraugotts work on subjectivity from the diachronic perspective (1989,1995, which shows that subjective meaning represents the last stage in se-mantic change) and Langackers approach to subjectivity within his theoryof cognitive grammar (1985 and later). The description of the grammaticalmarking of evidentiality and epistemic modality as important manifesta-tions of subjectivity have further been the object of intensive linguisticscrutiny for several decades in dierent languages of the world.3

Other terms have also been used for the formal-notional domain of thespeakers imprint and subjective perspective, and of late, stance is in-creasingly gaining ground as a (near) synonym for subjectivity. Kockel-man (2004) proposes the following definition for stance markers: anysigns that members of a community associate with a speakers personalcontribution to event-construal, where stances are to be understood aspossible kinds of such personal contributions. This definition is not unlikeour definition of subjectivity above, although, crucially, it brings in thespeakers community and therefore takes into account the larger culturaland ideological dimension. In this paper, my focus is not so much on try-ing to give a hard and fast definition to this (formal, notional, and cul-tural) domain of language structure and use (I feel that it is premature atthis stage, as I will explain below), but to look beyond a speaker-basedand largely individual notion of stance (and subjectivity), toward the pub-lic nature of displaying stances and the activity of joint stance taking indiscourse.There has recently been an upsurge in the interest shown by linguists

in subjectivity, and it is coming to be seen as a major organizing principlein much of language use. Most linguistic studies to date have treated sub-jectivity as a rather static category relating to the speaker; for example,Iwasaki (1993; cf. also Finegan 1995) in his treatment of subjectivityin Japanese examines the speaker as the center of deictic elements, thespeaker as the center of evaluation, attitude, and aect, and the speakeras the center of epistemological perspective (see also Langacker 1985


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

Camille Debras

Camille Debras

Camille Debras

Camille Debras

Camille Debras

Camille Debras

and many others). In recent discourse-functional and linguistic anthropo-logical studies, however, subjectivity (or aect) has been shown to influ-ence a wider range of aspects of language structure and use than has per-haps been thought (e.g., Ochs and Schieelin 1989; Hopper 1991; Iwasaki1993; Finegan 1995; Dahl 2000; Bybee and Hopper 2001; Scheibman2001, 2002). Research is beginning to show that not just grammatical cat-egories such as deictic terms, mood, modality, tense, and evidentials areindices of the speakers point of view or attitude, but that our everydaylanguage use is inherently subjective at many, if not most, levels. Indeed,Ochs and Schieelin (1989: 22) propose that aect permeates the entirelinguistic system, so that linguistic resources for expressing aective andepistemic stance include, not only the lexicon, but grammatical and syn-tactic structures (e.g., choice of pronouns, determiners, verb voice, tense/aspect, sentential adverbs, hedges, cleft constructions, diminutives, aug-mentatives, quantifiers, word order, and so on), phonological features(intonation, voice quality, sound repetition, sound symbolism), and dis-course structures (e.g., code-switching as instantiated by taboo words, di-alect, couplets, and repetition of own/others utterances; Ochs 1992: 412;Ochs and Schieelin 1989: 1214).In this vein, Biber et al. (1999: 859) also observe that conversation is

characterized by a focus on interpersonal interaction and by the convey-ing of subjective information. And Bybee and Hopper (2001: 7) state thatmost utterances are evaluative in the sense of either expressing a judg-ment or presenting the world from the perspective of the self or of the in-terlocutor. The authors go as far as to suggest that natural discourse ispreeminently subjective, concerned with the here-and-now world of thespeaker and the hearer. In a corpus-based analysis, Thompson and Hop-per (2001: 25) found that in American-English conversation the speakersdo not talk much about events or actions, but rather display their identi-ties, express feelings and attitudes, and check their views of the worldwith their community-mates.4 Thus, even though subjectivity has been anotion relating primarily to the speaker, research is beginning to showthat much of our everyday speech, based on speakers choices of lin-guistic forms (e.g., clause types, syntactic patterns, lexical choices), is cen-tered around the speech-act participants: not only the speaker but also thehearer (see Section 7 below). Indeed, Scheibman (2001: 77, 79) points outthat speakers situate their utterances in relation to other speech act partic-ipants, and she uses the term interactive subjectivity to characterize this:for example, the second-person-singular utterances (e.g., you can use thisfor your muns. Scheibman 2001: 77) reflect an interactive or empatheticsubjectivity because speakers frequently mediate direct assertions aboutother speech-act participants by modal elements (can).5


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

Camille Debras

Camille Debras

Camille Debras

Camille Debras

Camille Debras

Acknowledging this recent body of research but going one step further,I will argue for a more dialogic, dynamic, and emergent view of stance inwhat followsviewing it more as an intersubjective rather than subjec-tive characteristic of language.6 Here I draw on recent work by Du Bois(2000, 2002, 2004, forthcoming) who advocates the notion of stance asinvolving not only the subjective dimension, but also intersubjective en-gagement with other subjectivities: without intersubjectivity, subjectivityis inarticulate, incoherent, unformed (Du Bois 2004).7 Hunston andThompson (2000: 143) similarly point out that the expression of attitudeis not, as is often claimed, simply a personal matterthe speaker com-menting on the worldbut a truly interpersonal matter in that the basicreason for advancing an opinion is to elicit a response of solidarity fromthe addressee. We do not express our evaluations, attitudes or aectivestates in a vacuum; participants in discourse do not merely act, but inter-act. They achieve intersubjective understandings of the ongoing conver-sation as they display their own understanding (their subjectivities, ifyou like) in their sequentially next turns, while correcting or confirmingthose of their coparticipants (Heritage 1984; Nofsinger 1991; Hutchbyand Woott 1998). It is a central claim in conversation analysis thatparticipants update their intersubjective understandings on a turn-by-turn basis, or, in other words, intersubjectivity is sequentially constructed(Heritage 1984). Displaying stances is part and parcel of the interactionbetween participants who respond to prior turns and design their talk forthe current recipient(s). We therefore need contextual grounding in thedialogic and sequential context to arrive at a suciently enriched inter-pretation of stance, as is convincingly argued by Du Bois (forthcoming).An analysis of stance must also appreciate the nature of turns-at-talk asemergent, incrementally achieved, and subject to local interactional con-tingencies (cf. Scheglo 1996). It is also often the case that participantsdo not deal with the prior talk purely in its own terms, but rather theyaddress it in the way it is relevant for their own subsequent purposes(Goodwin and Goodwin 1987: 4). As a result, participants may constructstances by building on but also modifying the immediately co-presentstance of a dialogic partner (Du Bois 2000).

4. Patterns of epistemicity in natural discourse

Epistemic modality or epistemic stance has been established in earlier lin-guistic research to be a semantically coherent formal, often highly gram-maticized, domain (see, e.g., Lyons 1977; Perkins 1983; Palmer 1986;Coates 1990; Bybee et al. 1994; Biber et al. 1999). At the same time, it is


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

Camille Debras

Camille Debras

considered one of the most frequent manifestations of subjectivity in lan-guage (Lyons 1977; Palmer 1986). By epistemic stance I mean markingthe degree of commitment to what one is saying, or marking attitudes to-ward knowledge.8 This definition also includes evidential distinctions, orhow knowledge was obtained and what kind of evidence the speaker pro-vides for it (see also Chafe 1986 and Palmer 1986, who subsume eviden-tials under epistemic modality).9

The expression of epistemic stance is pervasive in everyday spoken in-teraction; in fact, speakers mark epistemic stance more frequently thanthey mark attitudes or evaluations, or express personal feelings or emo-tions (Biber et al. 1999; Thompson 2002). Ochs (1992: 419420), how-ever, claims that both aective and epistemic stances are central meaningcomponents of social acts and social identities and have an especiallyprivileged role in the constitution of social life. In Karkkainen (2003a), Iestablished that epistemic stance marking is, besides frequent, also highlyregular and routinized in terms of the linguistic forms used, as American-English speakers tend to use a limited set of high-frequency markersin everyday speech, whether these then straightforwardly index a clearspeaker stance or not.10 This finding is nicely corroborated by Prechtswork within the appraisal framework (Precht 2003: 240): her resultsalso suggest that English speakers tend to use the same small set of stancemarkers repeatedly (something like 150 out of more than 1,400 dierentstanced words available in English). The most common type of epistemicmarker in my American-English data was a cognitive verb with a first-person subject (without the complementizer that after it). A paradigm ex-ample of this is I think, by far the most common marker in my data. Asalready mentioned above, this epistemic marker is commonly held by lin-guists to be prototypically subjective.11 Other frequent markers in thedata studied were epistemic phrases (he or she said, I dont know, I guess,I thought), epistemic adverbs (maybe, probably, apparently, of course),and epistemic modal auxiliaries (would, must, might, could, will, may).Some recent studies have provided strong support for these findings.Thus, Biber et al. (1999) use a larger database but come up with thesame patterns for American English regarding the relative frequency ofepistemic markers, while Scheibman (2001, 2002) and Thompson (2002)note the same trends in smaller sets of conversational American English.Another consistent finding from my earlier study is that epistemic

stance marking predominantly occurs initially, i.e., before the actual issueor question at hand (Karkkainen 2003a).12 More specifically, epistemicstance markers as a class show great unity in the way they pattern in thedata in view of a relevant unit of social interaction, namely the intonationunit: they are most frequently placed at the beginnings of intonation


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

units. While some might be tempted to attribute such a pattern to cogni-tive processes (relating to the ease of information processing, a la Auer1996 for example),13 I aim to show that even this initial pattern of episte-mic stance marking has roots in the interaction between conversationalcoparticipants.

5. The emergent, contingent nature of displays of epistemic stance

In this section, I argue that epistemicity is a phenomenon that derivesfrom the inherently dialogic nature of speech, that talk is always directedto a particular recipient or recipients within the sequential context ofthe turn-by-turn unfolding talk (Bakhtin 1981 [1953]; Voloshinov 1973[1930]). In such a view, subjectivity is no longer regarded as a more orless static mental state of the speaker, but a dynamic concept constructedin the course of some action; i.e., subjectivity is an integral part of theinteraction between conversational coparticipants. The data show variousroles that one routinized stance marker, I think, plays in discourse, all ofwhich are interactionally contingent. Among these are qualifying sensitiveclaims, dealing with minor interactional trouble spots, and showing height-ened speaker commitment. It is worth pointing out that previous treat-ments of the interactional functions of I think have almost exclusivelyviewed this item within the framework of linguistic politeness; as a hedgeon the illocutionary force of utterances and showing respect for the neg-ative face wants of the recipients, or the need of every adult speaker toretain freedom from imposition (Hubler 1983; Holmes 1984; Brown andLevinson 1987; Coates 1990; Karkkainen 1991; Nikula 1996; Aijmer1997; Turnbull and Saxton 1997). While it is true that I think can beused for facework, this is by no means its only function in interaction, asI will show below (see also Notes 9 and 14).The following example of I think in intonation-unit-initial position il-

lustrates the interactional motivation behind initial marking of epistemicstance. Rebecca, a lawyer, is preparing a female witness, Rickie, to ap-pear in an upcoming trial. Here she is expressing her concern that someyoung single men on the jury panel may not be able to put themselves inthe position of a victim.

(1) SBCSAE 0008 Tell the Jury That Rebecca (Re); Rickie (Ri)1 Re: . . . (H) And,2 especially,3 . . I have some . . young single men,


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

Camille Debras

4 . . on my jury panel.5 Ri: [Mhm],6 Re: [(H)] And,7 . . . % I% --8 . . my . . worry is that they dont . . relate to what a woman

feels,9 . . whe[n so]mething like that 21 Ri: Mhm.22 Re: (H) But if,23 . . um,24 . . . a man . . were to be exposed to,25 they would . . . laugh,26 . . or,27 . . you know,28 be disgusted,

Rebeccas overall turn design is responsive to the contingencies arisingfrom approaching a highly sensitive topic in the presence of an actual vic-tim, Rickie, who generally only provides continuers but does not takea full speaking turn. This is reflected in Rebeccas generally incrementalturn design. While she is producing happening on line 9, she launchesinto oering a reason why young men would not feel empathy: this wordis uttered with a discovery intonation, i.e., it is faster and somewhatlouder and higher in pitch than the turn-so-far, and her speech remainsgenerally faster all through lines 1120. The reason oered in line 11, be-cause their experience would be totally dierent., strongly projects furtherexplication of the reason, as the referent of their experience is not clear.Rebecca indeed goes on to clarify it, if a man exposes himself, but withoutactually providing the crucial information to a man (rather than a wom-an) before she retracts, which, a man would never do that. (this utterancein itself involving some hesitation and self-repair). This again requires


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

further explication, which Rebecca produces in lines 19 and 20, that exhi-bitionists do not choose men as their victims in the first place but morevulnerable people instead. We have good reason to suspect that Rebeccadesigns this utterance on the fly, to avoid bluntly referring to the fact thatRickie belongs to the category of vulnerable people (and therefore endedup as a victim), as that would constitute an evaluation of her. Rickie her-self was in fact seen in the talk immediately preceding this segment to bequite vulnerable: she expressed extreme emotional anxiety over her expe-rience, was very upset and cried, and can still be heard sniing through-out this extract. While already producing the utterance, Rebecca thenprefaces the potentially face-threatening evaluation vulnerable people byI think. There is some evidence that she redesigns the utterance here: thewhole turn in lines 1119 is said in fluent rapid speech, but in line 20 thereare two micropauses (indicated by two sets of dots) both before and afterI think, indicating some hesitation before the upcoming troublesome item.Such examples can be taken as evidence for what Goodwin (1979: 104)has claimed to be a pervasive feature of conversational interaction: aspeaker in natural conversation has the capacity to modify the emergingmeaning of his sentence as he is producing it in accord with the character-istics of its current recipient. Marking epistemic stance before the issue athand can then in itself be seen as interactionally contingent: as partici-pants mark their subsequent talk as sensitive by displaying uncertainty,in this case I think modifying words, phrases or clause fragments, this pre-pares the recipients to align themselves to the upcoming contentious orevaluatively loaded message and to design their own subsequent turn-at-talk accordingly.In Karkkainen (2003a), I demonstrated that instances of I think clearly

arise from the contingencies of the immediate speakerrecipient interac-tion in certain recurrent sequence types: in second-pair parts of adjacencypairs (e.g., answers, second assessments and opinions, weak agreements),but also in more free-ranging sequential environments where participantsare doing more strategic work, often dealing with issues of face.14 Twofurther examples are presented here to illustrate this. In Example (2), in aconversation with his wife Jamie and two friends, Harold uses I think in adisaligning second assessment to introduce a dierent slant to why child-rens bones heal quickly. The conversational topic of childrens bones wasinspired by the specific case, discussed earlier in the recording, of Har-olds nephew, whose broken leg had healed very quickly.

(2) SBCSAE 0002 Lambada Pete (P); Harold (H); Jamie (J)1 P: Thats like,


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

2 . . that he was being hauled aroundin a little wagon [2and stu2].

3 H: [2Right2].4 . . . [3He ^healed very `quickly3].5 J: [3 kids ^bones,6 just like3] . . [4grow4][5`back5] really ^fast (Hx).7 P: [4`M4] [5hm=5].8 H: [5Yeah5].9! theyre really `soft to ^start with.10 J: Theyre `made of ^rubber.11 . . . 12 H: `Thats why b-,13 . . little ^kids usually dont break their legs ^anyway.

In lines 5 and 6 of the extract, Jamie makes a first assessment that kidsbones grow back really fast. Part of her utterance is overlapped by whatHarold himself says in line 4 (apparently to Pete), yet Harold provides apositive acknowledgment before Jamies utterance has even come to com-pletion (Yeah. in line 8). But he immediately produces a second assess-ment that introduces the real state of aairs: I think theyre really softto start with. After Jamies humorous upgraded acceptance of and agree-ment with this assessment, which is latched on almost immediately in line10 and 11, Harold resumes his main line of argument, that children donot break their legs in the first place because their bones are so soft. It ispossible to argue that I think simply expresses that the current speakerorients to the assessment oered in the prior turn as not quite accurateand that he is about to bring in a slightly dierent take on it. The secondassessment simultaneously projects more talk to account for this new per-spective, which indeed follows in lines 12 and 13. The prosodic design ofthe turn contributes to the more to come interpretation: Couper-Kuhlenand Selting (1996) argue that high onsets typically accompany topic ini-tiations and thereby project a longer turn-at-talk. On the other hand,Ford (2002) discusses analogous turn trajectories or turn constructionformats like denial plus account/correction, where she oers strong evi-dence that the second component is projected after the first. Finally,further support for the speaker orienting toward the first assessment asdisputable can be obtained from Heritage and Raymond (2005). The au-thors claim that speakers who find themselves producing a responsive,rather than first, assessment often upgrade their (even similarly positive)assessment, by using various techniques like confirmation agreementtoken, oh-prefacing, tag questions, and negative interrogatives to updatetheir claimed epistemic access to, and/or rights to assess, a referent


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

(Heritage and Raymond 2005: 23). The authors further argue (2005: 16)that oering a first assessment carries an implied claim that the speakerhas primary epistemic or moral rights to evaluate some matter, and inusing some of the above means, second speakers may wish to defeat theimplication that their rights in the matter are secondary. Even though Ithink in Example (2) does not appear to be a technique used to actuallyusurp primary rights to assess (and is not one of the techniques discussedby Heritage and Raymond), it clearly displays that the status of the priorassessment is disputable, as also observed by Heritage and Raymond(2005: 33) of a particular example that they give.In the following excerpt of a conversation, where three senior citizens

discuss their various ailments, Doris is doing more demanding interac-tional work. She shows heightened commitment when she produces sev-eral strong claims about herself as a pill taker, prefaced by several in-stances of I think.

(3) SBCSAE 0011 This Retirement Bit Doris (D); Ang (A); Sam (S)1 D: . . (H) Then in the afternoon,2 take the capsule and one . . one Lazik.3 A: Mhm.4 D: (H) (Hx) (H)5 S: . . She did [that yesterday].6 D: [I cant do it].7 My stomach . . . gives me trouble,8 . . . I cramp,9 . . . (H)10 A: (H) [Well what] --11 D: [X] (Hx) --12 A: . . . (H) Are are you % eating Tums,13 . . . for [calcium]?14 D: [No].15 . . [2Im not eating Tums.16 A: [2@@@@@2]17 D: (H)2] I have- --18 . . . Oh I did take,19 . . I did s- --20 . . . call out last night,21 and say l-,22 . . % make a note for potassium though.23 A: (H) Do you have one of those little things that

(H) has a compartment in it for [each day].


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

24 S: [No].25 D: . . . No.26 . . . Heaven sakes.27 . . @@@28 D: . . . (H) not a very good ^pill taker,29 Im re- --30! . . I ^think Im [^resenting,31 A: [Im not êither,32 but I have] --33 D: (H) Ìm ^re]senting this [2^medicine2].34 S: [2(COUGH)2]35! D: And Î think its `contributing to my ^problems.36 I ^really ^do.37! . . . (1) (H) Î think that . . the . . ^cardazam ìs,38! . .Î think that the . . d- ^diarrhetic ìs,40 (H) % . . . (Hx)

From line 30 onwards, Doris produces two rather strong primary stressesin each utterance, her speech becomes distinctly faster and louder com-pared to her immediately preceding talk, and she rushes through whatshe has to say in lines 2836 without pausing or yielding the floor (Angelatries in vain to break into the conversation in lines 31 and 32). In priordiscourse especially one coparticipant has oered premature advice: An-gelas questions on lines 12, 13, and 23 can be heard as giving advice onhow to act in order to reduce the harmful eects of the medicine and tomake pill-taking a regular (and potentially thus less harmful) activity.Angela further oers a display of aliation very likely to be followed byadvice on lines 31 and 32, Im not êither, but I have --, in eect tryingto interrupt Doriss troubles-telling sequence (see Jeerson and Lee 1981for troubles talk). When Doris thus displays heightened commitment toher claim that she resents her medicine and that it is actually causingher problems, she is responding to the inappropriate alignment of thecoparticipants.As these examples show, subjectivity, in this case stance taking through

the epistemic item I think, is inherently interactively organized. Displaysof subjectivity are engendered by the local contingencies of what happensbetween the coparticipants in prior or present talk, and stance displaysconform to and manifest aspects of that interaction (of which recipientdesign is one). Thus, stance taking marked with I think is a dynamic inter-active activity, an interactional practice engaged in by coparticipants inconversation, rather than a way of framing an isolated thought or posi-tion of an individual speaker.


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

6. Stance taking as an interactive activity

In this section, we will broaden the scope of analysis from an individ-ual speakers stance, toward stance taking as a joint interactive activitybetween conversational coparticipants. Instead of starting from pre-established and routinized linguistic manifestations of stance such as Ithink, the focus here is on the participants engagement and participationin evaluative activity in some recurrent sequential slots or environmentsin talk-in-interaction. Here my work has anities to the work done byDu Bois (2000, 2004, forthcoming) on stance taking as a dialogic jointactivity between conversational coparticipants, but my work is slightlymore influenced by conversation analysis, notably its attention to closesequential analysis and temporality of interaction. I will show here that astance only emerges as a result of joint engagement in evaluative activity.I propose that one such activity environment that frequently involvesjoint negotiation of stance is at points where conversational stories cometo completion, i.e., in story reception sequences. This position is muchlike what Labov and Waletzky (1967) refer to as a coda in conversa-tional narrative. Within conversation analysis, Jeerson (1978) discussesthis environment as one where exit from the story is negotiated betweenteller and recipients, typically showing the tellers orientation to what arecipient makes of the story, and recipients in turn displaying their ap-preciation. Scheglo (1984: 44) also points out that there is a structuralplace for story recipients to display their appreciation of the story com-pletion and their understanding of the story, and that these are oftenlinked tasks.Some terminological points need to be clarified here. I use the terms

assessment, opinion, evaluative utterance/turn, and the stance en-coded therein somewhat interchangeably. But where assessments havebeen defined by Goodwin and Goodwin (1992: 155) as actions that areproduced by single speakers (the same goes for opinions and evaluativeutterances), I seek to use the notion of stance (which may be, but is notexclusively, manifested in assessments, opinions, or other types of evalua-tive turns) as an intersubjective notion rather than pertaining to singleindividuals.15 The focus in this section is precisely on stance taking as asequential activity, rather than an individual action, or in the terminologyof Goodwin and Goodwin, on assessment activity rather than assess-ment. I will therefore include in the analysis any turn or action thatforms part of the stance negotiations (e.g., a simple Yeah.), even thoughit may not include any overt or inherent displays of stance like evidentialsor epistemic items, or any assessment tokens like evaluative adjectives ornouns.


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

In what follows, I will illustrate, by way of two examples, what kind ofinteractional practices are at play when participants negotiate a sharedstance toward aspects of the prior discourse, typically some protagonist(s)in a conversational story. Generally, the coparticipants do join in andindicate their agreement and alignment with each other, in addition tovarying degrees of disagreement and nonalignment.In the following example, two cousins are talking about a relative, Li-

sabeth. Alina has just told Lenore about Lisabeths recent encounterwith Alinas mother, during which Lisabeth had complained that Alinasmother now has a whole new life since she became a widow and does notseem to need her any more. The italicized items in Examples (4) and (5)are instances of epistemic stance.

(4) SBCSAE 0006 Cuz Alina (A); Lenore (L)1 A: Mom said I do.2 . . @@@ @ @3 (H) .7 L: [3(H)3]8 A: [4Moms go-4] --9! L: [4`Maybe shes just4] kind of ^dense.10! `Hunh.11 A: (H) . . `Well,12 . . she `wants everything on her [^terms].13 L: [(H)]14 A: [2You `know2].15 L: [2`Is she `vicious or ^dense2].16 A: . . `Shes a .17 (H)18 L: @[@@]@19 A: [So],20 L: @ .21 [@@@@]22 A: [@@@ .23 (H) So],24 t- Mom said,25 you know,26 % she goes,27 when can I see you.


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

Alina has established a clear negative stance toward Lisabeth during thetelling, portraying Lisabeth as the primary person responsible for the lackof communication between her and Alinas mother: she never calls her.The mothers rather unsympathetic reply to Lisabeth on line 1, reportedby Alina with a curt prosody and clear falling intonation, and Alinassubsequent malicious and very loud laughter can be interpreted by Le-nore to invite evaluation and appreciation of the story. Lenore indeedprojects that the story is now being brought to possible completion andproduces an assessment of the protagonist, Poor Lisabeth. Maybe shesjust kind of dense. Hunh. She also happens to start the assessment simulta-neously while Alina still continues to talk and to act out Lisabeths imag-inary reply to the mother (lines 3, 5, and 6), which makes it apparent thatLenores projection of closure misfired. Lenore displays some uncertainty(maybe, kind of, hunh,), however, as to whether this is the right kind ofconclusion to draw (that Lisabeth just does not get it that Alinasmother can do without her support). She also produces a conventionaltoken of sympathy toward the protagonist, poor Lisabeth. In all, the waythis turn is designed can be heard as the story recipient siding with aprotagonist that was just presented in a negative light (cf. Scheglo1984: 47), the other protagonist being, if not the teller herself, very closeto her, namely Alinas mother. Yet, even though Alina is not done withthe story, she joins in the assessment activity: she momentarily stops thetelling to engage in joint evaluation of the assessed protagonist (this isthen in eect not simply a case of the recipient producing an understand-ing upon story completion that is unacceptable for the teller, cf. Scheglo1984: 45). That Alina interrupts her story and oers a second assessmentor opinion . . Well, . . she wants everything on her terms. You know. is ofcourse also made particularly relevant by the design of Lenores first as-sessment, which ends in an interrogative tag Hunh. (cf. Pomerantz 1984:61).Now, as the negative stance toward the protagonist that Alina has

incorporated into the story was not received with an unequivocal andequally negative appreciation by Lenore (who had shown some sympathytoward Lisabeth), Alinas subsequent evaluation is not in full agreementwith Lenores assessment. Alina starts with Well, a potential disagree-ment preface (cf. Pomerantz 1984, even though well may here also markthe fact that Alina was interrupted in her storytelling), and produces anassessment with rather high contrastive stress on ^terms, that now por-trays Lisabeth as a domineering, rather than stupid, person. It is possiblethat the significance of this turn is not immediately obvious to Lenore (cf.Scheglo 1984 on ambiguity), who in turn confronts Alina by asking: Isshe vicious or dense. We may argue that being simply dense is perhaps


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

socially more acceptable than being vicious in the sense of domineering,and it is therefore rather important to establish which of the two Alinathinks applies to Lisabeth. Another interpretation is that Lenore therebypursues her original assessment of Lisabeth (as kind of dense, lines 9 and10) and again sides with the wrong protagonist. Alinas subsequent turn,Shes a dope., does not really choose either one of Lenores assessmentsignals as such. As it is roughly synonymous with dense, however, Alinacan be seen to align with and even escalate the stance in Lenores initialassessment on lines 9 and 10. But she dierentiates her stance rather pre-cisely by choosing a third assessment signal instead of one oered by Le-nore. Her pitch also falls markedly on ^dope. These choices appear to bedoing something in addition to simply oering an answer: they indicatethat Alina wishes to resume the telling of the story (notice So, on line19). We may argue that Lenore interprets the curtness of Alinas turn assignaling that she wishes to continue: Lenore indeed laughs and makes ametacomment, Explains that, which indicates that she, too, is ready tomove on. Even this turn design still gets a symmetrical matching contri-bution from Alina: laughter followed by Exactly. These turns achieve ashared stance, a working consensus of the characteristics of Lisabeth,and the assessment sequence is brought to a close. Alina can now con-tinue her story starting at line 23 with a conjunction so.What becomes clear from the example above is that involving ones

coparticipant in assessment activity is interactively important, and thatcoparticipants do indeed join in, even at points where storytellers havenot yet completed their story. Thus, it is not so much that once a conver-sational story is told, with a more or less clear stance becoming obviousin its course, that the recipients are able to oer the right kind of under-standing or appreciation of the story; they may display uncertainty aboutits import or upshot, and try to involve the storyteller in the assessmentactivity. On the other hand, the storytellers themselves may attempt toguide the interpretation of their story, and may not simply let the recipi-ents initial assessment pass. This example also demonstrates the impor-tance of establishing a provisional consensus regarding the kind of stancethat is being jointly constructed (e.g., Lenores question Is she vicious ordense and Lenores reply Shes a dope). In other words, participants dis-play an orientation toward reaching a common understanding and ashared stance (for the actual linguistic design of such turns, see Section7). And finally, such activities are usually brought to some sort of comple-tion before relevant next actions can follow (even though speakers mayoccasionally abandon their stances for lack of recipient uptake).The next example shows some similar stance negotiation patterns.

Sharon has recently been hired as a teacher in a bilingual (Spanish


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

English) classroom in Houston. She has just told Carolyn and Kathy, hersisters, about her generally very negative experiences: despite frequent at-tempts, she has been unsuccessful in securing a free lunch ticket for one ofher pupils, because the school bureaucracy prevents her from finding outwhether the girls application form has been filed in the first place. Thestory is another example of troubles talk, and in answer to yet anotherpiece of advice oered by Kathy before the example segment starts (. .You gotta go in and talk to the principal about this.), Sharon finallypresents the situation as a case of raging bureaucracy that she has noway of tackling.

(5) SBCSAE 0004 Raging Bureaucracy Sharon (S); Kathy (K); Carolyn (C); Environment (E)1 S: What kind of fucking @law [2is that2].2 K: [2@@@@@@2]3 S: (H) that youre gonna tell me,4 that,5 you know,6 this kid [might have] to wait,7 C: [(H)]8 S: another month and a half,9 to e- have any lunch,10 (H) Because you cant access,11 . . you know,12 . . her form?13 C: . . . (TSK)14 S: 15! C: [(H) Theyre just `giving--16! K: [Unbelievable].17! C: I `think],18! it `sounds,19! like,20! to `me,21! theyre `giving you a lot of ^shit for no ^@reason.22 S: (H) . . `Well they ^really are ^picking on the `fact that Im

^new,23 like,24 . . y- --25 uh,26 [Its really ^annoying].27 C: [The `fact that youre ^new],28 `uh,


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

29 . . ùh,30 S: [That I] --31 C: [` I would],32 . . Ì would go ^further than `that.33 S: Ìm not ^certified.34 . . Ànd,35 >E: ((SIREN_STOPS))36 C: ^Yeah,37 [@@]38 S: [its like],39 the [2teachers2],40 C: [2and2],41 S: Coop this last week,42 C: ,43 S: (H) First theyre like,44 . . first I only had fifteen kids.45 right?

Throughout the story, Sharon had presented herself as a rather activeagent. Carolyn makes a concluding assessment on lines 1521: Theyrejust giving-- I think, it sounds, like, to me, theyre giving you a lot of shitfor no @reason., thereby presenting her sister somewhat as a victim ofsome unwarranted acts of ill will on the part of the school personnel.However, her laugh token on @reason can be interpreted in dierentways: it can be heard as projecting that there is more to it and that Caro-lyn probably has a certain reason in mind, but possibly also that she doesnot have a clear idea or that she is unwilling to suggest at this point thereason, or that she is a little embarrassed by her formulation (which in ef-fect excludes all reasons). Sharon then oers a second less than fully ali-ating assessment: according to her these are not unprovoked acts but canbe accounted for by the fact that she is a new teacher at the school. In line27, Carolyn in turn disaliates with Sharons assessment by repeatingwhat to her is the problematic part in it, The fact that youre new, and byadding two distinctive uh tokens, which strongly project that she is notgoing to accept being new as an adequate reason and that some third ac-counting factor will follow. But she only ends up hinting at what it couldbe by saying I would go further than that, which again strongly projectselaboration. Sharon herself is thereby prompted to immediately oer thereal reason: Im not certified. Carolyn accepts this with an emphatic^Yeah, and some laughter (lines 36 and 37).In line 41, Sharon then continues to further relate her negative experi-

ences at the school, now specifically to Kathy or Coop (her nickname),


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

while Carolyn can be heard to say and, . The full signifi-cance of this truncated utterance cannot be deciphered with certainty,but one interpretation is that Carolyn was going to convey that it is in-deed not possible for a noncertified teacher to take an aggressive role inthe face of an injustice. In all, Carolyn has now led Sharon to the heartof the matter, to recognize that the problem lies not in being a new teach-er but a noncertified one, and therefore in a situation with less power overher working conditions. Crucially, the two coparticipants have reached amutual stance toward this issue.It is significant that the two participants jointly formulated the conclu-

sion that the underlying reason for Sharons maltreatment is the fact thatshe is not a qualified teacher. This had not emerged from Sharons earliertroubles-telling story in any way, nor was it explicit in Carolyns first as-sessment in lines 1521. Rather it emerged in the course of the jointstance-taking sequence that the participants were engaged in. Story recep-tion is thus not a straightforward and simple issue for participants, but theymay work out together an understanding of the gist or upshot of the story(for conversational storytelling as a collaborative and interactive activity,see Lerner 1992 and Mandelbaum 1993; for a case where the problematicstance taken by a story recipient leads to story expansions, see Ford 2004).As we saw in the examples in this section, stance is very often estab-

lished and negotiated as an interactional practice. The articulation anddierentiation of stance is a joint activity between discourse participants,i.e., it essentially involves the recipients coparticipation. Assessments areone type of action that involve taking up positions and making evalua-tions, and they indeed often come, not only in pairs, but in longer strings,with each speaker constructing a stance by building on, modifying, andaligning or disaligning with the immediately co-present stance of a dia-logic partner (see Du Bois 2000).

7. Linguistic practices of stance taking

When we start our analysis by looking at the actions and interactionsthat participants are engaged in doing, rather than by identifying pre-established linguistic forms allegedly displaying subjectivity, we becomeaware of linguistic resources over and above the lexicon, grammar, linearsyntax, or even discourse. In this section, finally, I will highlight someof the linguistic practices that participants in interaction resort to whenjointly taking positions and negotiating stances.I will present two stance negotiations from the examples above in dia-

grammatic form, by mapping the relevant intonation units produced by


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

dierent speakers onto each other (roughly) according to syntactic struc-ture (see diagraphs within the theory of dialogic syntax outlined in DuBois 2001 and forthcoming). Such a representation makes visible the per-vasiveness of (at least) two frequent dialogic practices of stance taking.The first of these is the initial framing of stance in individual speakerscontributions, which was already discussed above in connection with Ithink. The second is not a discrete lexical item or syntactic device, butrather the high degree of parallelism at various levels of linguistic struc-ture across speakers and speaking turns, in other words the resonance ofsyntactic structures, semantic meanings, and prosodic devices used.16

Resonance has been defined by Du Bois (2003) as the activation of po-tential anity across utterances, between comparable linguistic elementsat any level.A diagraph puts into focus not only the initial framing of the relevant

conversational turns, but also the parallelism and resonance across dif-ferent turns by dierent speakers. As can be seen in Figure 1, epistemicstance markers tend to occur initially, i.e., before the actual issue or ques-tion at hand, which is expressed later in the turn. For example, Carolynends up aligning several epistemic markers as separate intonation units, Ithink, it sounds, like, and to me, and the actual assessment is pushed untillater in the turn. Their abundance and separate encoding (that they

C: They re just `giving--I `think,it `sounds,like,to `me,

they re `giving you alot of ^shit

for no ^@reason.

S: `Well they ^really are ^picking on the `fact that Im ^new,C: The `fact that youre

^new,`I would go ^further than `that,

S: `I m not ^certified.

Initial framing ofstance beforeassessment:I think, it sounds,like, to me,well

Syntactic resonance:parallel syntactic structure: subject pronoun, verb tenses intensifiers (a lot of, really) extracting and repeating

object from prior turn

Semantic resonance: synonymous predicate

verbs (give a lot ofshit & pick on)

Prosodic resonance: e.g., stress, intonation-

unit-final intonation

Figure 1. Diagraph of Example (5) Raging Bureaucracy


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

appear as multiple as opposed to a single long intonation unit togetherwith the actual proposition) stems from the self-repair that Carolyn isengaged in, because she initially overlapped with Kathy (on line 16 inExample [5]). As we saw in Section 4 on patterns of epistemicity in natu-ral discourse, it is much more usual for epistemic stance markers to pat-tern intonation-unit-initially (Karkkainen 2003a). We also saw in Section5 that initial framing in itself can be seen as interactionally motivated: es-tablishing stance before the upcoming utterance helps recipients to alignthemselves to what is coming.The emphasis in much linguistic research has been on these initial and

more recognized stance markers, but, as this diagraph illustrates, there isa lot of regularity to be observed in the patterns of stance taking beyondthat. Notably, as Du Bois (2000) has observed, in conversational interac-tion one speaker often constructs a stance based partly on the immedi-ately co-present stance of a coparticipant.

Words and structures and other linguistic resources invoked by the first speakerare reused by the second, whether the second speakers stance is parallel, opposed,or simply orthogonal to the firsts. (Du Bois 2000: 3)

Such recycling across speakers and turns may result in considerable struc-tural parallelism in language as manifested in the repetition of words,phrases, syntactic structures, or prosodic patterns (Du Bois 2001; cf. alsoTannen 1987, 1989; Anward 2004). Specifically, we can see this in thediagraph of Example (5): even though in their first turns Carolyn andSharon take slightly dierent positions and focus on slightly dierent is-sues (i.e., being harassed for no apparent reason versus for being a newteacher), their utterances resonate both syntactically and lexically. Theclauses they are giving you a lot of shit and they really are picking on Xhave parallel syntactic structure: they have identical and co-referentialsubject pronouns, the verbs are in the present-perfect tense, and bothpredicates contain an intensifier.17 The two utterances also resonate se-mantically, as the predicates are synonymous. The dierentiation of stan-ces comes at the level of the adverbial clause, for no reason, versus objectof the verb, the fact that Im new. What is more, Carolyns second disa-ligning turn is directly built on the structure and meaning conveyed inSharons previous turn, by extracting a syntactic segment, the object Thefact that youre new, out of it. Anward observes of this phenomenon,where the recycled element occupies the rheme position in an utterance(as it does in Sharons turn):

Repeating or retaining another participants rheme, in contrast to doing a varia-tion on it, may thus indicate not only continuity of a contribution but also troubleof some sort. (Anward 2004: 37)


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

Anward (2004: 34) further notes that there are particular combinations ofrecycled elements, notably combinations of contour, stance markers, andresponse items, that are used to negotiate stance and establish intersub-jectivity, a statement that lends strong support to the claim made in Sec-tions 6 and 7 of the present paper.Another pattern in naturally occurring discourse, which we saw in all

of the examples, is the tendency for speakers to use first-person state-ments. Carolyns choice to make a first-person statement (I would gofurther than that., projecting elaboration) instead of a second-person one(e.g., saying directly Youre not certified.) enables Sharon herself to actu-ally produce the real reason, which she indeed does by similarly using afirst-person statement (Im not certified.). Bybee and Hoppers (2001: 7)claim that most utterances in natural conversation express a judgmentor present the world from the perspective of the speaker or of the interloc-utor has been substantiated in several studies. For instance, Thompsonand Hopper (2001: 37) found that the favored (in terms of frequency)grammatical constructions in conversational English are in fact intransi-tive verbal clauses (I forgot); copular clauses (it was confidential ); andepistemic/evidential clauses (I dont think its workable). Scheibman(2001: 86) also uses American-English discourse corpora to show thatthe most common constructions in conversation are those subjectpredicate combinations that permit speakers to personalize their contri-butions, index attitude and situation, evaluate, and negotiate empatheti-cally with other participants. In the light of such findings, then, suchhigh-frequency material is very likely to resonate across speakers as well.We can also use Example (5) to make some tentative observations of

the prosodic features that resonate across speakers. The two noun phrasesthe `fact that Im ^new, and The `fact that youre ^new, are produced al-most identically in terms of stress and final intonation contour (con-tinuing intonation). The same goes for the next two utterances or theI-statements in `I would go ^further than `that. and `Im not ^certified.The diagraph of Example (4) further illustrates the syntactic, semantic,

and prosodic resonance between speakers, as well as the tendency to placestance markers initially in utterances (see Figure 2).In this diagraph we again see repetition across speakers of a syntactic

frame, in this case she is X or is she X. Also dense and dope are semanti-cally parallel and resonate, even though, as we saw above, the very choiceof a dope instead of dense by Alina has the eect of dierentiating herstance from Lenores. While it could be argued that such dierentiationis a display of the storytellers claim to primary rights to assess (cf. Heri-tage and Raymond 2005), it is nevertheless a striking instance of recy-cling with dierance (Anward 2004), i.e., of a more general pattern of


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

the participants performing varied operations in especially rhematic recy-cling to create a dierence to the prior contribution and its rheme: oneexample mentioned by Anward (2004: 44) is finding a semantically re-lated expression for an expression used so far.What is also striking in the example above is the degree of phonologi-

cal recycling and sound similarities involved in the choice of lexical items:dense and dope are not just similar semantically, but they also show allit-eration and contain the same number of syllables (even though they rep-resent dierent parts of speech); the same holds for Explains that andExactly.18 In addition, we have seen that laughter as a nonverbal deviceis used almost symmetrically by the two speakers here. These nonverbaland prosodic cues may be argued to actually contribute toward a conver-gent stance between speakers, as opposed to the dierentiation of stancesbrought about by their choice of dierent lexical items (dense and dope;see also Szczepek 2001 for types of prosodic orientation or how onespeaker responds prosodically to another speakers prosody).By oering these preliminary observations about similarities in forms

and meanings across dialogic turns, I do not mean to suggest that reso-nance and parallelism are a necessary feature of stance-taking sequencesspecifically. Stance-taking sequences are simply one frequent environmentwhere such resonances are readily observable, and therefore worthy ofour serious attention. Nor do I wish to indicate that resonating structuresonly construct agreement and aliation; on the contrary, as we sawabove, some fine-tuned dierentiation of stances can be achieved bymeans of nearly parallel and resonating structures. As Tannen (1987:

L: `Maybe she `s just kind of ^dense.A: `Well she `wants everything on her own

^terms.L: `is she `vicious or ^dense.A: `She `s a ^dope.L: .A: .

Initial framingof stance:maybe, well

Syntactic resonance: parallel syntactic frame parallel verb tenses

Semantic resonance: dense & dopeProsodic resonance: e.g., stress, intonation-unit-final

intonation phonological recycling and

sound similarities (dense &dope, explains that & exactly)

Figure 2. Diagraph of Example (4) Cuz


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

599) puts it, the metamessage of involvement communicated by repeti-tion can gainsay a message of disagreement communicated by the wordsspoken, and, on the other hand, meanings may disagree even if the wordsspoken are identical. In all, this kind of work is still in its infancy; none-theless, I hope to have pointed out the utility of resonance as a new sitefor investigations of stance taking in conversation.

8. Conclusion

In this paper, I first outlined some patterns of epistemic stance markingwithin the speech of single speakers and showed that these arise from theintersubjectivity between conversational coparticipants. Even though ithas not been customary in linguistic research to view subjectivity as aninteractional notion (in large part because of the long-standing focus onsingle-speakers contributions, the term subjectivity of course reflectingthis bias), it is clearly necessary to regard it as such. I argue that it is pref-erable to delimit the term subjectivity to refer to a formal (i.e., lexically,phonologically, morphologically, or syntactically marked) quality thathas to do with the speakers personal contribution, and which is perhapsmost useful for capturing the referential meanings of the relevant linguis-tic resources. When our focus of inquiry is the description of the interac-tional functions of linguistic resources for displaying subjectivity (as Ibelieve it should be), with less emphasis on the degree of subjectivity thatthey express, we are bound to find that what appear to us as subjectiveelements of language emerge first and foremost intersubjectively. In thissense, intersubjectivity as discussed in this paper is a quality inherent tothe interactional process of stance taking (as an instance of interaction),regardless of the formal characteristics of the linguistic resources used. Itis not advisable to limit our scrutiny only to what have been recognizedas subjective elements in language, because these often reflect the analy-sts cultural and linguistic bias, and furthermore, as I have shown, otherless discrete factors may clearly play a role in the stance-taking activity.In the latter part of this paper, I focused on the interactive nature of

stance taking. A careful analysis of the interaction between participantsin one type of activity environment, story completion, presented furtherevidence of the fact that stance taking is an interactive activity engagedin by coparticipants in conversation, rather than being an isolated mentalactivity of an individual speaker. When making a concluding assessmentof the preceding story, speakers display an orientation toward involvingrecipients in the assessment activity, and recipients generally join in to ne-gotiate a shared stance toward, and an understanding of, the story or


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

some aspect of the story. But such joint stance taking is not only a formof participation in the reception of a story, it is often simultaneously a ne-gotiation of some underlying social norm or value implicit in the story.Because assessments involve taking up a position, and assessors can beheld responsible for the positions they state (Goodwin and Goodwin1992), participants in interaction have motivation to make sure that thestances they advocate, and the understandings that they jointly reach,are such that they can be held accountable for them. Therefore it is under-standable that at least a working agreement is often negotiated before thetopic is closed.Only when stance is approached by students of language as something

jointly oriented to by the coparticipants can we hope to observe linguisticpatterns of stance taking that go beyond specific, discrete grammatical orlexical devices analyzable in single-speakers contributions. As I demon-strated above, constructions that are considered prototypically subjec-tive in linguistic theory, such as I think, are really intersubjective in na-ture. Furthermore, as we have seen, in natural discourse many patternsof stance taking may show a phonological, syntactic, and/or semanticresonance between the contributions of dierent speakers. Such globalpatterns have not been examined in any detail by linguists so far, perhapsfor the simple reason that they cannot be observed without attention to alarger context than single-speaker contributions.

Appendix: Transcription conventions (from Du Bois et al. 1993)

UnitsIntonation unit {carriage return}Truncated intonation unit --Truncated word -

Transitional continuityFinal .Continuing ,Appeal (seeking a validating response from listener) ?

SpeakersSpeech overlap [ ](numbers inside brackets index overlaps)

Accent and lengtheningPrimary accent (prominent pitch movement ^carrying intonational meaning)Secondary accent `Lengthening


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

PauseLong and medium (length indicated in seconds) . . . (1)Short (brief break in speech rhythm) . .

Vocal noisese.g., (TSK), (SNIFF), (YAWN), (DRINK)Glottal stop %Exhalation (Hx)Inhalation (H)Laughter (one pulse) @

QualityPiano: soft Higher pitch level Lowered pitch level Voice of another

Transcribers perspectiveUncertain hearing Analyst comment ((WORDS))

Notes

1. This paper is based on work done in a research project entitled Interactional practicesand linguistic resources of stance-taking in spoken English and has been financed bythe Academy of Finland (grants 00381 and 53671). Some parts of this paper havebeen published in Karkkainen (2003b) while some material is included in Karkkainenet al. (forthcoming). I have had valuable comments on an early version of this paperfrom Sandy Thompson and Jack Du Bois, to both of whom I owe sincere thanks. Iam also very grateful to Robin Shoaps and Paul Kockelman for many challengingquestions and insightful comments on earlier versions of this paper. Many thanks alsogo to three anonymous reviewers of Text & Talk, as well as three other reviewersduring an earlier developmental stage of this article. Any remaining inconsistencies,ambiguities, or mistakes are solely my own.

2. The division into descriptive versus expressive functions of language is by no meansunproblematic, and many linguists have come to believe, as Ochs (1986: 256) puts it,that all sentences expressed in context will have an aective component. In some con-texts the aect conveyed is one of distance and objectivity in expressing information,while in others the subjective and the personal is more overtly expressed.

3. For English, see, e.g., Palmer (1986), Chafe (1986), Biber and Finegan (1989), andBiber et al. (1999); for a typological overview, see Willett (1988); and for an overviewof linguistic work on evidentiality as indexing social meaning, see Fox (2001).

4. The authors focused on transitivity in conversational data and especially on what iscalled argument structure in linguistics, i.e., the grammar of the verb and its argu-ments. Their main finding was that clauses in English conversation are low in transitiv-ity (see also Section 7 of the present paper).

5. The frequency of first-person evaluative statements has also been noted by Dahl (2000).His data on Swedish conversation show a high frequency of egocentric expressions,


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

Camille Debras

which for him include first-person as well as second-person and generic pronouns thatcluster with mental-state verbs (tro believe, tycka think, tanka think, minnas re-member, etc.), as well as with verbs relating to external appearance (verka seem, seut look, appear) and with copular verbs (vara be, bli become). For Dahl, thesepatterns are evidence of the inherent nature of discourse to reflect the speakers pointof view.

6. Indeed, a similar stance was advocated by Benveniste (1971: 224225), who did nottreat subjectivity as prior to intersubjectivity, but rather the other way around:

Consciousness of self is only possible if it is experienced by contrast. I use I only when Iam speaking to someone who will be a you in my address. It is this condition of dia-logue that is constitutive of person, for it implies that reciprocally I becomes you inthe address of the one who in turn designates himself as I.

7. Du Bois (2004) also brings into focus a third dimension of stance taking, namelythe objective dimension, the stance object: that speakers orient to a shareable stanceobject provides further evidence that subjectivity cannot mean an escape to the mentalinterior.

8. Lyons, like many others, mentions the concept of truth in his definition of epistemicmodality:

Any utterance in which the speaker explicitly qualifies his commitment to the truth ofthe proposition expressed by the sentence he utters, whether this qualification is madeexplicit in the verbal component [. .] or in the prosodic or paralinguistic component, isan epistemically modal, or modalized, utterance. (Lyons 1977: 797)

Biber et al. (1999), however, define epistemic stance markers much more broadly aspresenting speaker comments on the status of information in a proposition.

9. My decision to consider hearsay evidentials such as he or she said as a subtype ofepistemic stance is based on the following. Besides conveying one kind of speaker per-spective toward knowledge, namely that the speaker has heard the information fromsomebody else, in my database hearsay evidentials also behave remarkably similar toprototypical epistemic markers such as I think: they are highly frequent, they com-monly occur in intonation-unit-initial position, and are often prosodically unmarked(unaccented or receiving only secondary stress, faster in tempo, reduced in phoneticsubstance to the point of deletion of the subject pronoun). Furthermore, my data pro-vide evidence that utterance verbs are used as epistemic phrases that (simply) frame thestance expressed in the actual direct speech that follows (see Holt 1996, 2000; Shoaps1999; and Clift 2000 for the claim that reported speech can indirectly express thereporting speakers stance toward some past state of aairs or toward the reportedspeaker). This is not unlike utterances that start with an epistemic marker such as Ithink, in which this marker often only frames a stance that is expressed in the rest ofthe utterance, rather than in itself expressing a clear stance in many contexts of use.

10. The decision to start out by the frequency of occurrence of a coherent semantic system,that of epistemic modality (and only subsequently analyzing the interactional use madeof it) is mainly due to my research history, which in turn reflects a common approachin pragmatic and discourse-functional linguistics. Ochs (1992: 417) also observes thistendency of linguistic studies to focus analysis on lexical and grammatical systemsthat index a single situational dimension, e.g., evidential systems that index epistemicstance, because the dimension in question is usually grammaticized or lexicalized incomplex and interesting ways. Not surprisingly, the latter half of my work (Karkkai-


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

nen 2003a) then concentrated on the interactional functions of only one marker, Ithink.

11. Robin Shoaps (personal communication) aptly points out that linguists are basingthe whole idea of subjectivity (rather than recognition of intersubjectivity) on folktheories of language and the mind.

12. The utterance-initial occurrence of stance markers has been commented on in passingin some other studies as well, such as Biber et al. (1999), Anward (2000), and Thomp-son (2002).

13. Auer (1996) discusses the pre-front field in spoken German, a grammatical positionsometimes (but not always) realized as prosodically integrated with the rest of theutterance, in which case it is analogous to intonation-unit-initial position. One of itsdiscourse functions is said to be expressing something like epistemic stance (Auer1996: 310312). Auer (1996: 312) says of the pre-front field that:

Above all, it is iconic: the frame is clearly separated from the framed structure, and itprecedes it. Cognitively, this ensures that the context in which the following utteranceis to be processed is available from the very start of the interpretation process, thusavoiding cognitively and/or interactionally demanding repair work.

14. One sequence type not discussed here is points of transition in discourse (topic shifts,shifts to metadiscourse or subsidiary information, shifts back to an earlier point or toa completely new point), where I think acts as a boundary marker for the benefit of therecipient (Karkkainen 2003a). In such cases its referential meaning is vague, to thepoint that the subject pronoun may sometimes even be omitted.

15. Note however that Goodwin and Goodwin (1992: 155) also state that assessments canbe organized as an interactive activity that not only includes multiple participants butpotentially also types of actions that are not themselves assessments.

16. Resonance of the various lexical, syntactic, and prosodic resources across speakingturns can be considered to be manifestations of the overall poetic parallelism indialogue that was originally observed by Jakobson (1981) and which has been furtherdeveloped and elaborated by Silverstein (1984), Tannen (1987, 1989), Norrick (1987),and the various contributions in Johnstone (1994).

17. Anward (2004: 41) notes of what he considers core turns in conversation that newturns are made from recycled old turns, in such a way that the overall format, theframe, of the old turn is kept, but a new expression is substituted for a part of the oldturn, normally its rheme and/or its stance marker.

18. Anward (2000) has pointed out that speakers may appropriate each others phonologi-cal sound material and resort to prosodic recycling and matching of sounds in theirturn design.

References

Aijmer, K. (1997). I thinkAn English modal particle. In Modality in the Germanic Lan-guages: Historical and Comparative Perspectives, T. Swan and O. J. Westvik (eds.), 147.Berlin/New York: Mouton de Gruyter.

Anward, J. (2000). How types of lexical resources emerge from turn-construction. Paper pre-sented at the Euroconference on Interactional Linguistics, Spa, Belgium, September.


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

(2004). Lexeme recycled. How categories emerge from interaction. Logos and Language.Journal of General Linguistics and Language Theory V (2): 3146.

Auer, P. (1996). The pre-front field in spoken German and its relevance as a grammaticali-zation position. Pragmatics 6 (3): 295322.

Bakhtin, M. (1981 [1953]). Discourse in the novel. In The Dialogic Imagination, M. Holquist(ed.), 259442. Austin: University of Texas Press.

Benveniste, E. (1971). Problems in General Linguistics. Coral Gables, FL: University ofMiami Press.

Biber, D. and Finegan, E. (1989). Styles of stance in English: Lexical and grammaticalmarking of evidentiality and aect. Text 9 (1): 93124.

Biber, D., Johansson, S., Leech, G., Finegan, E., and Conrad, S. (1999). The grammaticalmarking of stance, chap. 12. In The Longman Grammar of Spoken and Written English.London: Longman.

Brown, P. and Levinson, S. (1987). Politeness. Some Universals in Language Usage. Cam-bridge: Cambridge University Press.

Bybee, J. and Hopper, P. (eds.) (2001). Frequency and the Emergence of Linguistic Structure.Amsterdam: Benjamins.

Bybee, J., Perkins, R., and Pagliuca, W. (1994). The Evolution of Grammar: Tense, Aspectand Modality in the Languages of the World. Chicago: University of Chicago Press.

Chafe, W. (1979). The flow of thought and the flow of language. In Discourse and Syntax, T.Givon (ed.), 159181. New York: Academic Press.

(1986). Evidentiality in English conversation and academic writing. In Evidentiality: TheLinguistic Coding of Epistemology, W. Chafe and J. Nichols (eds.), 261272. Norwood,NJ: Ablex.

(1987). Cognitive constraints on information flow. In Coherence and Grounding in Dis-course, R. S. Tomlin (ed.), 2151. Amsterdam: Benjamins.

(1994). Discourse, Consciousness, and Time: The Flow and Displacement of ConsciousExperience in Speaking and Writing. Chicago: University of Chicago Press.

Clift, R. (2000). Stance-taking in reported speech. Essex Research Reports in Linguistics 32:131.

Coates, J. (1990). Modal meaning: The semanticspragmatics interface. Journal of Seman-tics 7: 5363.

Couper-Kuhlen, E. and Selting, M. (1996). Towards an interactional perspective on prosodyand a prosodic perspective on interaction. In Prosody in Conversation. InteractionalStudies, E. Couper-Kuhlen and M. Selting (eds), 1156. Cambridge: Cambridge Univer-sity Press.

Dahl, O. (2000). Egophoricity in discourse and syntax. Functions of Language 7 (1): 3777.Du Bois, J. (2000). Taking a stance: Constituting the stance dierential in dialogic interac-tion. Paper presented at the Annual Meeting of the American Anthropological Associa-tion, San Francisco, November 18.

(2001). Towards a dialogic syntax. Unpublished manuscript. Department of Linguistics,University of California, Santa Barbara.

(2002). Stance and consequence. Paper given at the Annual Meeting of the American An-thropological Association, New Orleans, November 21.

(2003). Dynamic resonance. Lecture given at the University of Oulu, September.(2004). The intersubjectivity of interaction. Paper given at the Tenth Biennial Rice Uni-versity Symposium on Linguistics: Stancetaking in Discourse: Subjectivity in Interaction,Rice University, April.

(forthcoming). The stance triangle. In Stancetaking in Discourse: The Intersubjectivity ofInteraction, R. Englebretson (ed.). Amsterdam: Benjamins.


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

Du Bois, J., Chafe, W., Meyer, C., and Thompson, S. (2000). Santa Barbara Corpus ofSpoken American English, Part 1 [six hours of digitized audio and transcriptions on threeCD-ROMs]. University of Pennsylvania: Linguistic Data Consortium.

Du Bois, J., Schuetze-Coburn, S., Cumming, S., and Paolino, D. (1993). An outline of dis-course transcription. In Talking Data: Transcription and Coding in Discourse Research, J.Edwards and M. Lampert (eds.), 4587. Hillsdale, NJ: Lawrence Erlbaum.

Duranti, A. and Goodwin, C. (1992). Editors introduction. In Rethinking Context: Lan-guage as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 147150. Cam-bridge: Cambridge University Press.

Finegan, E. (1995). Subjectivity and subjectivisation: An introduction. In Subjectivity andSubjectivisation. Linguistic Perspectives, D. Stein and S. Wright (eds.), 115. Cambridge:Cambridge University Press.

Ford, C. (2002). Denial and the construction of conversational turns. In Complex Sentencesin Grammar and Discourse. Essays in honor of Sandra A. Thompson, J. Bybee (ed.), 6178.Philadelphia: Benjamins.

(2004). Contingency and units in interaction. Discourse Studies 6 (1): 2752.Fox, B. (2001). Evidentiality: Authority, responsibility, and entitlement in English conversa-tion. Journal of Linguistic Anthropology 11 (2): 167192.

Goodwin, C. (1979). The interactive construction of a sentence in natural conversation. InEveryday Language: Studies in Ethnomethodology, G. Psathas (ed.), 97121. New York:Erlbaum.

Goodwin, C. and Goodwin, M. H. (1987). Concurrent operations of talk: Notes on theinteractive organization of assessments. Papers in Pragmatics 1 (1): 154.

(1992). Assessment and the construction of context. In Rethinking Context: Languageas an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 147190. Cambridge:Cambridge University Press.

Heritage, J. (1984). Garfinkel and Ethnomethodology. Cambridge: Polity Press.Heritage, J. and Raymond, G. (2005). The terms of agreement: Indexing epistemic au-thority and subordination in talk-in-interaction. Social Psychology Quarterly 68 (1), 1538.

Holmes, J. (1984). Modifying illocutionary force. Journal of Pragmatics 8: 345365.Holt, E. (1996). Reporting on talk: The use of direct reported speech in conversation. Re-search on Language and Social Interaction 29 (3): 219245.

(2000). Reporting and reacting. Concurrent responses to reported speech. Research onLanguage and Social Interaction 33 (4): 425454.

Hopper, P. (1991). Dispersed verbal predicates in vernacular writing. BLS 17: 402413.Hubler, A. (1983). Understatements and Hedges in English. Amsterdam: Benjamins.Hunston, S. and Thompson, G. (eds.) (2000). Evaluation in Text. Authorial Stance and theConstruction of Discourse. New York: Oxford University Press.

Hutchby, I. and Woott, R. (1998). Conversation Analysis. Principles, Practices and Appli-cations. Cambridge: Polity Press.

Iwasaki, S. (1993). Subjectivity in Grammar and Discourse. Amsterdam: Benjamins.Jakobson, R. (1981). Poetry of grammar and grammar of poetry. In Roman Jakobson. Se-lected Writings, S. Rudy (ed.), 8697. The Hague: Mouton.

Jeerson, G. (1978). Sequential aspects of storytelling in conversation. In Studies in theOrganization of Conversational Interaction, J. N. Schenkein (ed.), 219248. New York:Academic Press.

Jeerson, G. and Lee, J. R. E. (1981). The rejection of advice: Managing the problematicconvergence of a troubles telling and a service encounter. Journal of Pragmatics 5 (5):399422.


!"#$%&'('#()#$(*)(+(,"-./(01-2/"3-'/-'(453'/"675(8,"-./(01-2/"3-'/-'(453'/"67594$'&/1'-:7'/6(+(;;?>;>==?

@#A1B#76(@7'/(+(CD=CD;=(=EC;(FG

Johnstone, B. (1994). Repetition in Discourse. Interdisciplinary Perspectives, vols. 1 and 2.Norwood, NJ: Ablex.

Karkkainen, E. (1991). Modality as a strategy in spoken interaction. A discourse analyticaland pragmatic approach with a particular focus on the English language. Unpublishedlicentiate thesis. Department of English, University of Oulu.

(2003a). Epistemic Stance in English Conversation. A Description of its Interactional Func-tions, with a Focus on I think. Amsterdam: Benjamins.

(2003b). Is she vicious or dense? Dialogic practices of stance taking in conversation. InSanta Barbara Papers in Linguistics. Recent Studies in Empirical Approaches to Language,vol. 12, T. Nakayama, T. Ono, and H. Tao (eds.), 4866. Department of Linguistics, UCSanta Barbara.

Karkkainen, E., Sorjonen, M., and Helasvuo, M. (forthcoming). Discourse structure. InLanguage Typology and Syntactic Description, vol. 2, T. Shopen (ed.). Cambridge Univer-sity Press.

Kockelman, P. (2004). Stance and subjectivity. Journal of Linguistic Anthropology 14 (2),127150.

Labov, W. and Waletzky, J. (1967). Narrative analysis: Oral versions of personal experience.In Essays on the Verbal and Visual Arts. Proceedings of the 1966 Annual Spring Meeting ofthe American Ethnological Society, J. Helm (ed.), 1244. Seattle: University of Washing-ton Press.

Langacker, R. (1985). Observations and speculations in subjectivity. In Iconicity in Syntax:Typological Studies in Language, vol. 6, J. Haiman (ed.), 109150. Amsterdam: Benjamins.

Lerner, G. (1992). Assisted storytelling: Deploying shared knowledge as a practical matter.Qualitative Sociology 15 (3): 247271.

Lyons, J. (1977). Semantics, vol. 2. Cambridge: Cambridge University Press.Mandelbaum, J. (1993). Assigning responsibility in conversational storytelling. Text 13 (2):247266.

Nikula, T. (1996). Pragmatic Force Modifiers: A Study in Interlanguage Pragmatics. StudiaPhilologica Jyvaskylaensia, University of Jyvaskyla.

Nofsinger, R. E. (1991). Everyday Conversation. Newbury Park, CA: Sage.Norrick, N. (1987). Functions of repetition in conversation. Text 7 (3): 245264.Ochs, E. (1986). From feelings to grammar: A Samoan case study. In Language Socializa-tion across Cultures, B. Schieelin and E. Ochs (eds.), 251272. Cambridge: CambridgeUniversity Press.

(1992). Linguistic resources for socializing humanity. In Rethinking Linguistic Relativ-ity, J. Gumperz and S. Levinson (eds.), 407437. Cambridge: Cambridge UniversityPress.

Ochs, E. and Schieelin, B. (1989). Language has a heart. Text 9 (1), 725.Palmer, F. (1986). Mood and Modality. Cambridge: Cambridge University Press.Perkins, M. (1983). Modal Expressions in English. London: Frances Pinter.Pomerantz, A. (1984). Agreeing and disagreeing with assessments: Some features ofpreferred/dispreferred turn shapes. In Structures of Social Action. Studies in ConversationAnalysis, J. M. Atkinson and J. Heritage (eds.), 57101. Cambridge: Cambridge Univer-sity Press.

Precht, K. (2003). Stance moods in spoken English: Evidentiality and aect in British andAmerican conversation. Text (Special issue: Negotiating Heteroglossia: Social Perspec-tives on Evaluation) 23 (2): 239257.

Scheglo, E. (1984). On questions and ambiguities in conversation. In Struct

Kärkkäinen 2006 Stancetaking in conversation From subjectivity to intersubjectivity

Documents