This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
If the fans want to blame me then and if we have to go to court, then life,” adding: “If that means prison, for example. If that means complexity, much the better. If force is needed, them. If it lacks easy saleability, If we have to do it by being hip, and if we are soundly beaten, then
so be it. A JONES, Blackford so be it. But, believe me, th so be it. But Hearts have bee so be it. We must be competit so be it. Moreover, there hav so be it. It’s a parody of re so be it.
We can get back so be it,” he says. Of course so be it.” A spokesman for Gu so be it.” Mr Leeson said he so be it. As a result of this so be it. In this context, Mr so be it -- the most interest so be it.”
The agency has so be it,” the manager said.
In short, example 23 contains several phrases that can be shown to recur in the context of evaluative stance, even though they do not evaluate themselves. In other words, what makes this paragraph stance-heavy is the presence of a very real…about, if…so be it, to the point of, and the redundancy they represent, rather than the words concern, tragedy, and so on themselves. The investigation of this short text extract leads to a hypothesis: what distinguishes subjective (or stance-heavy) from objective (or stance-light) texts is not the quantity of explicitly evaluative lexical items in each, but the embedding or otherwise of those items in phraseologies, which frequently co-occur with stance. Those phraseologies can be identified intuitively, but intuition in this regard is unreliable, and an examination of many instances of the target phrase is required to corroborate the perception of its role in the text under investigation. A corpus can provide just such a set of instances. This argument also implies an alternative view
45
46 Susan Hunston
of the role of quantitative information in studies of stance. Whilst the counting of stance items, and the comparison between various corpora in this regard, remains a viable and useful activity, it is clear that of necessity it provides a partial view of stance. However, the quantitative, in the sense of giving information about ‘what is usually said’ or ‘how a phrase is typically used,’ has an important role to play in the qualitative activity of identifying stance elements.
8.
Conclusion: Quantitative and qualitative corpus research
In this paper, 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 qualitative approach – and raw figures should be treated as the starting point of investigation rather than its end point. 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 unknown. These markers are typically phrases rather than individual words. I have also tried to suggest that whereas computers are essentially used to count items, in the case of corpora, that counting can be used in the service of qualitative research as well as quantitative. 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.
Notes 1. As demonstrated in other contributions to this volume, especially the chapters by Du Bois, Keisanen, and Haddington, stance meaning is also distributed across larger interactional units. In addition to phraseologies, a broader qualitative analysis of stance likewise incorporates turns-at-talk, interactional sequences, and intertextuality. 2. Both these corpora are components of the Bank of English corpus, jointly owned by the University of Birmingham and Collins publishers. Other examples used in this paper are from the same corpus unless otherwise stated. 3. Channell notes that the phrase seems to be infrequent in American English – perhaps the essential pessimism it expresses is a purely British phenomenon. 4. These concordance lines are taken from the Bank of English corpus. 5. I am grateful to Robert Englebretson for disagreeing with me here, and so forcing me to examine the evidence more closely. 6. The use of tragedy to indicate a type of play is excluded here.
Using a corpus to investigate stance
References Becher, T. 1989. Academic Tribes and Territories: Intellectual Enquiry and the Culture of Disciplines. Milton Keynes: Open University Press. Becher, T. and Trowler, P. 2001. Academic Tribes and Territories. 2nd ed. Milton Keynes: Open University Press. Biber, D. 1988. Variation in Speech and Writing. Cambridge: Cambridge University Press. Biber, D., Johansson, S., Leech, G., Conrad, S. and Finegan, E. 1999. Longman Grammar of Spoken and Written English. London: Longman. Channell, J. 2000. “Corpus-based analysis of evaluative lexis.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, S. Hunston and G. Thompson (eds.), 38–55. Oxford: Oxford University Press. Charles, M. 2004. The Author’s Voice in Academic Writing with Reference to Theses in Politics and Materials Science. Ph.D. dissertation, Department of English, University of Birmingham. Conrad, S. and Biber, D. 2000. “Adverbial marking of stance in speech and writing.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, S. Hunston and G. Thompson (eds.), 57–73. Oxford: Oxford University Press. Francis, G. 1994. “Labelling discourse: An aspect of nominal-group lexical cohesion.” In Advances in Written Text Analysis, M. Coulthard (ed.), 83–101. London: Routledge. Francis, G., Hunston, S. and Manning, E. 1996. Collins Cobuild Grammar Patterns 1: Verbs. London: HarperCollins. Francis, G., Hunston, S. and Manning, E. 1998. Collins Cobuild Grammar Patterns 2: Nouns and Adjectives. London: HarperCollins. Granger, S. 1998. Learner English on Computer. London: Longman. Groom, N. 2004. “The scholarly, the sound and the thin: Evaluative adjectives and disciplinary values in academic book reviews.” Postgraduate seminar, University of Birmingham. Hunston, S. 2004. “Counting the uncountable: Problems of identifying evaluation in a text and in a corpus.” In Corpora and Discourse, A. Partington, J. Morley and L. Haarman (eds.), 157–188. Bern: Peter Lang. Hunston, S. and Francis, G. 1999. Pattern Grammar: A Corpus-Driven Approach to the Lexical Grammar of English. Amsterdam: John Benjamins. Hunston, S. and Thompson, G. (eds.). 2000. Evaluation in Text: Authorial Stance and the Construction of Discourse. Oxford: Oxford University Press. Lindquist, H. and Mair, C. (eds.). 2004. Corpus Approaches to Grammaticalization in English. Amsterdam: John Benjamins. Louw, B. 1993. “Irony in the text or insincerity in the writer? – The diagnostic potential of semantic prosodies.” In Text and Technology: In Honour of John Sinclair, M. Baker, G. Francis and E. Tognini Bonelli (eds.), 157–176. Amsterdam: John Benjamins. Maley, A. 1999. “Surviving the 20th century.” English Teaching Professional 10: 3–7. Matthiessen, C. 2006. “Systemic functional profiles of system and text: Investigations based on texts, text archives, and corpora.” In System and Corpus: Exploring Connections, G. Thompson and S. Hunston (eds.), 103–142. London: Equinox. Semino, E. and Short, M. 2004. Corpus Stylistics: Speech, Writing and Thought Presentation in a Corpus of English Writing. London: Routledge. Sinclair, J. 1991. Corpus Concordance Collocation. Oxford: Oxford University Press. Sinclair, J. 2004. Trust the Text: Language, Corpus and Discourse. London: Routledge.
47
48 Susan Hunston
Sinclair, J. and Coulthard, M. 1975. Towards an Analysis of Discourse: The English Used by Teachers and Pupils. Oxford: Oxford University Press. Teubert, W. 2004. “Language and corpus linguistics.” In Lexicology and Corpus Linguistics, M.A.K. Halliday, W. Teubert, C. Yallop and A. Čermáková (eds.), 73–112. London: Continuum. Tognini-Bonelli, E. 2001. Corpus Linguistics at Work. Amsterdam: John Benjamins.
Linking identity and dialect through stancetaking Barbara Johnstone
Carnegie Mellon University
1.
Introduction1
Geographic mobility and new patterns of social interaction associated with the globalizing new economy have resulted at the same time in dialect leveling (Milroy 2002; Trudgill 1986) and, in some places, increased popular attention to regional variation (Beal 1999; Dubois and Horvath 2002). One such place is the Pittsburgh, Pennsylvania (U.S.) area, where talk about local identity very often includes talk about the local dialect (Johnstone 2000a; Johnstone and Andrus 2005; Johnstone and Baumgardt 2004; Johnstone et al. 2002). When “authentic” Pittsburghers or Pittsburgh activities are described or parodied, local speech is almost invariably mentioned or performed. Souvenir vendors offer t-shirts and sweat shirts, refrigerator magnets, shot glasses, and coffee mugs decorated with lexical items thought to be local and other words spelled in such a way as to suggest local pronunciations. A dictionary of Pittsburghese, as the variety is locally known, has been continuously in print since 1982, and there is a copy in most middle-class homes; and Pittsburghers and others go online to contribute to lists of local expressions and discuss what the dialect means for the community. One way to attribute a quintessentially local identity to a person is to label him or her a yinzer, a word derived from the local variant of the second-person-plural pronoun, yinz. While much of this discourse about local speech arises in the context of nostalgia for the city’s working-class industrial past, discourse about Pittsburghese also enters into youthful identity work. Pittsburghers in their 20s and 30s, whether or not they make routine use of stereotypically local phonological or lexical variants, sometimes refer to themselves as yinzers, and they perform local identity by means of playful performances of local-sounding forms, both in faceto-face interaction and in more public fora. In a playful nod to the New Yorker,
50
Barbara Johnstone
a new Pittsburgh literary magazine was named the New Yinzer, 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.” Heightened dialect awareness, like that seen in Pittsburgh, arises through discursive practices that call attention to and normativize regional forms, at the same time refiguring their role in presentations and representations of local identity (Johnstone and Andrus 2005). 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. In the conversation, the Pittsburghers deploy several strategies for epistemic stancetaking: for making the implicit or explicit claim, that is, that they know enough, from the appropriate sources, to describe and evaluate the local dialect. 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, while others invoke the speaker’s own competence in the dialect. These include performances of local dialect forms and other allusions to local identity. 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. As a result, the other speaker, who at first distances herself from speakers of the local dialect, eventually revises her identity claim, recasting herself as a local dialect speaker who knows local forms directly rather than by hearsay. In other words, she links her identity as a Pittsburgher with competence in the local dialect because it is rhetorically useful in this interaction to do so. The micro-rhetorical interactional exigency that requires epistemic stancetaking drives the identity claim. Stancetaking is thus one of the mechanisms through which dialect and local identity become linked in discourse. 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. Variationist sociolinguists in the Labovian tradition are coming to see “identity” as a useful explanatory dimension in accounting for some patterns of linguistic variability and their role in language change. Analysts of discourse in interaction 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. Sociolinguists drawing on both traditions are beginning to explore how stancetaking can be accomplished through phonological, morphological, and lexical choices, and
Linking identity and dialect through stancetaking
how sets of such choices can accrete into stances that index culturally meaningful styles or identities (Eckert 2000; Kiesling 2005). This paper continues the work of exploring how connections between linguistic forms and social identities are shaped by interactional needs for stancetaking. With Eckert and Kiesling, I show that stancetaking and identity are intertwined, in this case because adducing and performing the social identity of a competent dialect speaker is a powerful resource for epistemic stancetaking, so much so that the need for epistemic stancetaking can actually drive dialect-identity claims. 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, and how variationists have adduced these concepts in accounting for patterns of variation across populations. I then show how dialect forms and regional identity can become linked through stancetaking, by virtue of the fact that performances or evocations of dialect competence can function as stancetaking moves. The text I use to illustrate this is an extended transcribed extract from a sociolinguistic interview. The interview was conducted in the course of a project that brings together analyses of regional patterns of phonological, morpho-lexical, and syntactic variation with analyses of the real-time, interactive design of talk in particular rhetorical situations, by particular individuals. 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.
2.
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.g., whether they are sure or unsure about it, happy or sad about it, surprised or not). Building on early work on epistemic and attitudinal stance (Biber and Finegan 1989, 1994; Conrad and Biber 2000), Hunston and Thompson (2000) operationalize stance as evaluation. Others use “stance” to talk about the marking and claiming of interpersonal relations in talk as well. For Ochs (1992), particular linguistic forms directly index evidential stances such as certainty, interpersonal stances such as friendliness or intensity, or social actions such as apologizing. Particular stances or social actions can then get linked indirectly to social identities such as gender categories (so that, for example, in a particular sociocultural milieu, a stance such as deference might become indexically linked with femaleness). For Du Bois
51
52
Barbara Johnstone
(this volume: 163), “Stance is a public act by a social actor, achieved dialogically through overt communicative means, of simultaneously evaluating objects, positioning subjects (self and others), and aligning with other subjects, 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. Alignment or disalignment with another social actor can be accomplished through membership categorization moves; thus, claims to social identity for oneself and ascriptions of identity to others fall under the rubric of stancetaking. Like Goffman’s (1959) “presentations of self,” “identities,” in the interactionalsociolinguistic tradition, are social categories to which speakers orient as they become relevant in the interaction at hand. Some identities are “macrolevel demographic categories,” while others are more situation-specific roles or “ethnographically emergent cultural positions” (Bucholtz and Hall 2005: 585). Research on social interaction that adduces identity as an explanatory tool almost always includes under “identities” culturally circulating, frequently adduced ways of categorizing groups of people that are often oriented to as being relevant outside of and prior to the interaction as well as inside it. “Identities” thus include ethnic, class, and gender categories, and categorizations in terms of attributes such as deviance vs. normalcy, tastes, and activities, as well as “discourse identities” such as speaker or audience member. (Perhaps because of a misplaced fear of being seen as locating social agency in the individual human, this body of literature pays little attention to the way in which identities can be associated with individually-embodied speakers (Johnstone 1996, 2000b).) But “identity work,” as it is described in some of this literature, includes interactional moves that could also be described as stancetaking. As Bucholtz and Hall point out, “identities may be linguistically indexed through…stances” (2005: 585), and a repeated stancetaking move or pattern of moves may emerge as an identity. Interactionists in the conversation analysis tradition take a similar approach. For them, 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 Widdicombe 1998: 3). “Casting into a category” can be accomplished through stancetaking. Conversation analysts stress the need to treat identity as interactional achievement with consequences for the structure of the talk, even if social actors think identities pre-exist interactions and sometimes predict how they will play out.
3.
Identity and stance in variationist sociolinguistics
Labovian-variationist sociolinguistics (Labov 1994, 2001) is primarily focused on uncovering the mechanisms of language change. Variability is an inevitable con-
Linking identity and dialect through stancetaking
comitant of change, so in order to understand how language change is likely to proceed, we need to understand how variability arises and what predicts its outcome. The classic Labovian account uses the correlational methods of quantitative sociology to model how “social facts” about speakers, such as socio-economic status and sex, account for patterns of variation in groups. In the research settings in which most early work was done, upper-class and female speakers consistently tended to use more of certain variants than lower-class and male speakers did, presumably because the variants in question were more “prestigious.” Variation within an individual’s speech was linked with self-consciousness, which gives rise in many cases to more careful speech, aimed at what speakers think of as more prestigious variants. From the beginning, however, in Labov’s own work as well as in others’, more particular questions were being asked about why these kinds of patterns might exist. Why, for example, 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, the answers that were proposed suggested that phonological variability could be a resource for indexing attitude and affiliation. In his groundbreaking study of Martha’s Vineyard, Labov (1963) showed that the speakers who identified most closely with traditional local ways of living were most likely to use the more conservative, older, less standard-sounding variants of certain phonological variables. This suggested that the more local-sounding forms might in fact be part of the process through which local identity was claimed. James and Lesley Milroy’s work in Belfast (J. Milroy 1992; L. Milroy 1987) explored the utility of social network theory for uncovering and explaining patterns of variation. People with relatively many and overlapping ties to their neighbors (such that they knew the same people in multiple roles, and the people they knew also knew each other) were relatively likely to talk in a more local-sounding, less standard-sounding way, because dense, multiplex social networks are effective enforcers of local norms. “Social class” was thus reconceptualized in terms of social activity and from the perspective of speakers’ everyday experience of language. Peter Trudgill’s (1972) equally influential study of sex-correlated differences in Norwich, England suggested that men may use more non-standard, local-sounding forms because such forms carry “covert prestige” as signals of working-class solidarity. At the same time, R. B. LePage and Andrée Tabouret-Keller (1985), concerned with describing the complicated linguistic situation in the Caribbean, were suggesting that choices among variants in a speaker’s repertoire could be thought of as “acts of identity,” and that speech communities and linguistic varieties could be seen as ways of labeling the fact that ways of identifying and, accordingly, of talking, sometimes become relatively consistent.
53
54
Barbara Johnstone
Variationists doing quantitative, correlational work about speech communities have started to add “identity” to the lists of socio-demographic variables they use to account for patterns of phonetic variability, finding that identity (variously operationalized) sometimes correlates more closely with variation than do variables such as age, occupation, or place of residence. For example, Guy Bailey and his colleagues (Bailey et al. 1993) show that, in Texas, linguistic changes that diffuse from rural to urban settings typically involve the reassertion of traditional speech norms, arguably as badges of local identity. One such feature is the monophthongization of /ay/ before voiceless obstruents, as in [ra:t] for right. Speakers who do this are more likely than others to think highly of Texas as a place to live. Hazen (2000) finds that North Carolina speakers with ties to institutions 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), while people with more local identities are linguistically more consistent. Rose (2003) has found that among older people from the upper Midwest, 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. Schilling-Estes (1998) shows that in Ocracoke (an Atlantic-coast U.S. island), 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. 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. Turning to stancetaking, interactional sociolinguists have explored how choices among forms associated with different languages and dialects can signal attitude and affiliation. For example, Ben Rampton (1995, 1999) has studied “language crossing” or “styling the ‘other,’” exploring how shifts from a speaker’s “native” language or variety to and from one clearly associated with another group can accomplish shifts in stance. Rampton (1995) describes how brief excursions into Punjabi by London adolescents of Anglo and Afro-Caribbean as well as Indian and Pakistani descent can signal oppositional footing in interactions with adults or key a joking stance in playground interaction. Stancetaking work can also be accomplished through shifts into and from ways of speaking that are not so clearly “other,” but rather part of a speaker’s native repertoire (Coupland 2001; Rampton 2003). In Texas, for example, where sounding Southern is a resource more or less “natively” available to many people, moves toward more Southernaccented speech range, depending on the speaker and the situation, from fairly automatic style shifts correlated with register to quite self-conscious rhetorical
Linking identity and dialect through stancetaking
moves, as when “Southern Belle”-sounding speech is used in footing-shifts meant to manipulate men (Johnstone 1999a). Penelope Eckert’s (1989, 2000) work in a Detroit high school suggests how the variationist and interactionist approaches to stance, identity, and phonology might be linked, suggesting how correlations like the ones described by Bailey, Hazen, Rose, and others come to be. For Eckert, the choice between one variant and another is part of the semiotic activity in which social identities are created, not simply a reflection of already-existing differentiation. For example, adopting certain variants of vowels is one way of adopting a stance toward and participating in local activity, just as is cruising in cars, dressing in a particular style, or doing some things rather than others in school. Scott F. 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,” which, Kiesling claims, is particularly likely to be relevant for members of subordinate groups. Because they work together as a stancetaking strategy, the features get used repeatedly together, and a repeatable style (locally called wogspeak) emerges, linked with a repeatable social identity, that of the recent immigrant. Eckert’s and Kiesling’s work shows how linguistic variants can become indexically linked to social identities through stancetaking. Particular linguistic features available in a speaker’s sociolinguistic environment can be used for stancetaking, and sets of co-occurring stancetaking features can come together as styles that index identities. The identities linked with linguistic styles may be ethnic (as in Kiesling’s wogspeak, derived from a derogatory term for non-whites); they may have to do with local social categories (as with Eckert’s jocks and burnouts), or they may be linked with other sources of identity. When the identity in question is regional, the style that indexes it is often referred to as a dialect, by linguists and laypeople alike. In what follows, I exemplify how stancetaking, identity, and linguistic variation are linked with reference to how ideas about what counts as a dialect are negotiated and circulated in interaction.
4.
Pittsburghese in conversation: Identity as a stance resource
Both linguists and non-linguist Pittsburghers associate a distinctive set of linguistic features with southwestern Pennsylvania (Brown 1982; Hankey 1965, 1972; Johnstone et al. 2002; Johnstone et al. 2004; McCarthy 2004; McElhinny 1999), and Pittsburghers talk about this dialect often, in many contexts and media, frequently in connection with talk about local identity (Johnstone 1999b; Johnstone, Andrus, and Danielson 2006; Johnstone and Baumgardt 2004; Kiesling and Wis-
55
56
Barbara Johnstone
nosky 2003). The dialect appears to be receding: research on one feature, /aw/monophthongization, shows that it is now more common in the speech of older, working-class speakers, the population that might be expected to use a receding feature longest (Kiesling and Wisnosky 2003; McCarthy 2004). Local-sounding talk, acquired early in life in face-to-face encounters, once identified people as Pittsburghers only to occasional outsiders who noticed the dialect. Insiders, in daily contact with people who sounded the same as they did, seldom noticed or commented on local speech. Beginning in the 1960s, however, 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, in discursive contexts such as nostalgic talk by ex-Pittsburghers and identity work by newcomers, for the calling to attention of linguistic difference that creates heightened awareness of regional varieties, and Pittsburghers started to use certain local speech features to point to local identity in a more reflexive, stylized way (Johnstone, Andrus, and Danielson 2006; Johnstone and Baumgardt 2004). The area thus lends itself to a study of how dialect awareness arises in a variety of metalinguistic genres of talk. As one phase in a larger study of Pittsburgh speech, co-workers and I have conducted over 100 sociolinguistic interviews in four Pittsburgh-area neighborhoods. The interview protocol elicits explicit talk about Pittsburghese, as it is locally known, and, throughout the interviews and other research tasks, people often break into spontaneous performances of the dialect. My analytical method in this paper is discourse analysis, by which I mean close, systematic reading of a small amount of text (Johnstone 2002). In the interview extract that we will examine in detail, a woman in her 40s and her 13-year-old daughter talk about Pittsburgh speech in an explicitly normative way. They offer examples of what Pittsburgh speech sounds like and argue about which forms are authentically local and how local forms should sound. As they do this, they use claims about and performances of their own speech as ways of establishing the authority to describe the dialect, linking epistemic stance with dialect identity. To keep the transcript readable and avoid caricaturing Jen and Donna by means of eye dialect, I have made notes about their accents in the right-hand column of the transcript rather than using the IPA transcription or respelling in the text. BJ is the fieldworker. Simultaneous talk is linked with square brackets; equal signs indicate “latched” talk. Italics mark loud or otherwise stressed words. Jen R., the mother, makes routine use of a number of phonological variables that make her sound local. Her pronunciation of /aw/ is sometimes monophthongized, with out realized as [a:t]. She merges and rounds the low back vowels (LBV), pronouncing job as [jfb], for example. She fronts /u/ in words like move and sometimes vo-
Linking identity and dialect through stancetaking
calizes /l/. She reduces the diphthong /ay/ to a more monophthongal form when it is followed by /l/ (which is vocalized), as when while is realized as [wa:u]. This is another fairly common local-sounding variant. Her daughter Donna has a less local accent; she does not monophthongize /aw/ but fronts /u/ and /o/, vocalizes /l/, and merges and rounds the low back vowel. As do most sociolinguistic interviews in the Labovian tradition (Labov 1984), this one included modules on topics meant to elicit a range of levels of self-consciousness. My summary of Jen and Donna’s accents is based on the whole interview, as well as unrecorded talk in other contexts. The topic of accent may well have made them self-conscious about their speech, but in fact they sound very much the same in this segment as elsewhere. (1) FH01 (“Jen R”) and FH02 (“Donna R”) Pittsburghese 1 2 3 4 5 6 7
8 9 10 11 12 13 14 15 16 17
bj jr bj jr
Um. So, have you ever heard of Pittsburghese? Oh yes. I mean, [there’s] [What- ] that store over on the Southside in um, [Station Square that has the] Pittsburghese shirts dr [((Breath intake)) Yeah] bj Uh huh= jr =In fact, I remember when my friend Karen moved out of state, with- her husband’s job took them out of state and to many other states. I remember sending her a couple Pittsburghese shirts for them.
diphthongal /aw/ in Southside
fronted /u/ in moved; monophthongal /aw/ in out (twice); rounded low back vowel (LBV) in job; “dropped g” in sending
dr Hm jr Umm. Yeah, I’ve heard of [Pittsburghese,] definitely. Yeah.= bj [Mm hmmm.] =What do you think it is? I mean is it-? jr I think it’s the way we say words. dr [((laughs))] bj [Yeah?] diphthongal /aw/s jr I think it’s how we say “downtown” and um= in downtown dr = “down[town”] diphthongal /aw/s diphthongal /aw/ jr [“South]side” and, in Southside dr “Y’all”
57
58
Barbara Johnstone
18
jr “wash” and “iron” and different words and the way Pittsburgh is. There’re
19 20 21
dr “Yinz” jr [Just the uh- like] bj [“Yinz” is another one.] What, what other ones can you think of? dr Just “y’all” and “yinz.” That’s, that’s the most my friends always are saying “y’all” to me. bj [“Y’all?”] dr [Drives] me crazy. jr “Y’all?” dr Yeah, they say “y’all” to me. They say [it’s a Pittsburghese] bj [And that’s a Pittsburgh thing?] dr It- that’s what they tell me. bj Huh! jr “Younz” is more a Pittsburgh thing than [“y’all.”] dr [yeah] jr “Y’all”’s more like a Georgia, [Southern.] dr [I was thinking] Southern, [but ] jr [Yeah] dr they still say “y’all” to me. And then. Yeah “yinz.” You hear “yinz” a lot. bj You do? Indr Yeah well like our neighbors like two doors down, I’m really good friends with their son. He’s a year older than me. And like, he says “yinz” constantly, ‘cause like both his parents say “yinz,” like “Yinz wanna do somethin’?” or [like] ((laughing)) you know so [I hear that.]
22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
38 39 40 41 42 43 44
45 46 47
bj jr dr jr bj jr
“standard,” not local pronunciations of wash, iron [yHnz] [yHnz]
[y~nz]
LBV not rounded in lot fronted /u/ in two, do; vocalized /l/ in older rounded LBV in constantly; fronted /o/ in both, know
[Mm hmm] [We don’t use that.] Yeah I never said you [used it] but [Yeah,] You don’t use that. [Uh huh.] [but] I’m just thinking, I know, um, like, your dad and I don’t use that [too often.] dr [No.] But I hear it a lot from [them LBV not rounded when] I’m over there in lot; /o/ less fronted than previously in over jr [Mm hmm] bj [Mm hmm] jr Yeah. It’s funny.
48 49 50
Linking identity and dialect through stancetaking
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71
dr And you pick up on it. You start to say then, [once] jr [Sure.] dr 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)) dr Mm hmm bj So, would- you don’t- you wouldn’t say you use any of the, Pittsburghese things? dr Not really. No, I don’t think so. bj Uh huh [(probably you-)] jr [You s-] you do, but you don’t reali-. Oh, I know I do. I [say “iron”] dr [I do?] jr for “iron.” bj Uh huh. dr [“iron”] jr [I] say “wash w- wash” for “wash”. I [mean] dr [“wash”] jr I don’t, I don’t pronounce my words as clearly as-, dr “wash” jr or the accent’s [on] bj [uh huh] jr a different [part] dr [“wash”] bj Uh-huh jr Yeah dr “wash” jr “Southside” instead of “Southside.”
72
dr “Southside”
51 52 53 54 55
diphthongal /aw/ in around
iron = [arn] iron = [ay6rn] [arn] [w~š] [wfš] [wfš] [wfš] [wfš]
[wf]
[wfš] monophthongal /aw/ and assimilated /θ/ in first, “performance,” Southside; exaggerated diphthongal /aw/ and /θ/ in second, “citation” Southside Monophthongal /aw/
59
60 Barbara Johnstone
73
jr I say “Southside.” “Downtown” instead of “downtown.”
74 75 76 77 78
dr jr bj dr jr
79 80 81 82
bj dr jr dr
83
jr
84 85
dr “=Yeah. you’re- you’re [definite-]” jr [“Are you-], are you from the Pittsburgh area?” dr Yeah. ((laughing)) bj Does that happen to you, too, or=? dr =Yeah. I mean, I remember one time, we were in South Carolina visiting my, my uncle and my two cousins and my aunt. And we went to (s- some) store, and we were talking about how like the South kind of moves slow, [you know?]
86 87 88
89 90
91 92
“downtown” I know I, I know I kn- use a lot of Pittsburghese. Mm hmmm. I [probably do] and I don’t realize it [I know I do.] Well when I’ve been in [different] states, in different cities, [Mm hmm] They’ll- they’ll say “You’re from Pittsburgh.” Yeah, they’ll immediately [say,] [When we were in] South Carolina, right? “You’re from Pennsylvania=”,
jr [Yeah, ] [God, it drives you crazy.] dr [And then she’s like] she’s like, “You guys from Pennsylvania?” We’re like “Yeah.” And she’s like “You guys wouldn’t happen to be from Pittsburgh, right?” And we’re like, “We’re from Pittsburgh.” And she’s like, “Oh, okay. I can [tell by your accent.”] bj [How did she know?] How do you think she knew? jr She, she was from the Pittsburgh area. She didn’t grow up in Pitt- [yeah. She said, she said she originally]
monophthongal /aw/ monophthongal /aw/s in first, “performance,” downtown; exaggerated diphthongal /aw/s in second, “citation” downtown diphthongal /aw/s
fronted /u/ in do
diphthongal /aw/ in South /l/ not vocalized in Pennsylvania
diphthongal /aw/ in South; vocalized /l/ in uncle; fronted /u/ in two, moves; diphthongal /aw/ in about, how; monophthongal /aw/ in South. “Pennsylvania question” intonation; vocalized /l/ in Pennsylvania didn’t = [dι’6n]
93 94 95 96 97 98
Linking identity and dialect through stancetaking
dr jr bj jr bj jr
[Yea-, I thought she was] from Ohio. Yeah, [but she she was,] [(Eastern)] Ohio? Right [but she] [This area?] she was in Pittsburgh and then they moved to Ohio,
99 dr Yeah 100 jr and then from Ohio they moved to South Carolina.
fronted /u/ in moved fronted /u/ in moved; monophthongal /aw/ in South
101 dr [She didn’t have the] accent either. She still had like a Pittsburgh accent. 102 jr [She said that-] Mm hmm. Yeah. She didn’t [have] didn’t = [dι’6n] 103 bj [Mm hmm] rounded LBV in 104 jr a Southern accent at all. She was, like a mom. [You know,] mom 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; monophthongal /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, produced 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 stancetaking 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-
61
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 distanced 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 diphthongal /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 authoritative 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 local 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 pronunciations, 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 knuse a lot of Pittsburghese (line 75). Note how in this final extract Jen starts to say
Linking identity and dialect through stancetaking
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 conversation. 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 dialect 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 speakers 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 competent-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
63
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 performance 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 speakers, .. 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 really. 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 dialect 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 trying 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 pronunciation 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:
78 79 80
Linking identity and dialect through stancetaking
jr bj dr
[I know I do.] Well when I’ve been in [different] states, in different cities, [Mmhmm] 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 mutual 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 Pittsburgher 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 stancetaking 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 interactional 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 performances 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.
65
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 everyone 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. Claiming 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 pronunciations, 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.
Notes 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 fellow 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.
References 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.” Language 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
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 Evaluation 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 Anchor 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 Press. 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 Construction 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 Sociolinguistics 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 Anniversary Essays): 390–392. Johnstone, B. 2000b. “The individual voice in language.” Annual Review of Anthropology 29: 405–424. 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 presented 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 enregisterment of ‘Pittsburghese’.” Journal of English Linguistics 34: 77–104. Johnstone, B. and Baumgardt, D. 2004. “‘Pittsburghese’ online: Vernacular norming in conversation.” American Speech 79: 115–145.
67
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: 148–166. 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 Language 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 issue]. Journal of Sociolinguistics 6: 3–130. Ochs, E. 1992. “Indexing gender.” In Rethinking Context: Language as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 335–358. New York: Cambridge University Press. 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– 556. 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 Ocracoke 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; second, 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 language 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-singular referring expressions, the -nya clitic, and verbal diathesis (voice) respectively 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 observations 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:
70 Robert Englebretson
whenever speakers (or writers) say anything, they encode their point of view towards it …The expression of such speakers’ attitudes is pervasive in all uses of language. All sentences encode such a point of view, … and the description of the markers of such points of view and their meanings should therefore be a central topic for linguistics. (1986: 1)
In other words, to roughly paraphrase, every utterance enacts a stance, and awareness of this should inform the linguist’s work at all levels.2 In the past 20 years, since Stubbs’s proposal, numerous linguists have indeed taken up this call, especially those working in corpus-based, systemic-functional, sociocultural, and interactional approaches to language. Yet, in general, most of this work has been focused on English, and a recognition of the centrality of stancetaking has by and large not found its way into field linguistics, nor into the writing of descriptive grammars of other languages. For example, while there are numerous grammatical descriptions available for both formal and colloquial varieties of Indonesian, none of these sources has addressed stancetaking per se, nor have they addressed the lexico-grammatical means which speakers use to accomplish it. Rather, Indonesian descriptive and pedagogical grammars tend to reflect the general bias in descriptive/typological/field linguistics which conceives of grammar as (primarily, if not solely) a cognitive object and referential system. The present paper hopes to expand this view by offering some observations on the social/interactional nature of Indonesian grammar, and seeks to initiate a discussion of how Indonesian speakers take stances. For the three aspects of Indonesian grammar that I discuss in this paper, I show that, in addition to fulfilling their traditional, cognitively-based referential functions of expressing and managing information, they have specific interactional functions too, which contribute directly to the social worlds speakers are constructing through stancetaking. There are undoubtedly more than just these three resources for stancetaking in colloquial Indonesian. For example, Wouk (1998, 2001) analyzes two clause-final particles used in expressing solidarity in social interaction. The present paper obviously cannot attempt a comprehensive and thorough analysis of the entirety of stancetaking in the language. However, as a first step toward this goal, I offer these observations and analyses in hopes that they will spark future research by other Indonesian scholars interested in investigating stancetaking more thoroughly. The second point of departure for this paper, the functionalist principle that grammar is motivated and shaped by language use, has likewise informed the work of numerous researchers over the past several decades. Language form has been argued to be determined largely by the cognitive and biological makeup of human beings, and the communicative contexts of language use (cf. Cumming and Ono 1997; Ford et al. 2002; Givón 1979; Langacker 1999; Ochs et al.
Stancetaking in colloquial Indonesian conversation
1996; Ono and Thompson 1995; Tomasello 1998, 2002; among many others). Various grammatical resources have been shown to emerge (Hopper 1987) out of broader cognitive, discourse, and social-interactional aspects of language use – a view which has widely come to be known as usage-based approaches to grammar (cf. Barlow and Kemmer 1999; Bybee 2006; Bybee and Hopper 2001; inter alia). While a thorough review of this literature lies well outside the scope of the present paper, the relevance of this approach becomes clear in light of the ubiquity of stancetaking discussed above and outlined in the introduction to this volume. Since stancetaking is such a pervasive activity found in language use, then focusing on stancetaking would seem a natural, worthwhile next step in the development of functionalist research paradigms concerning usage-based models of grammar. Substantial work has yet to be done to flesh out the role of stancetaking in motivating and constraining grammar, both cross-linguistically, and systematically within specific languages. The current paper does not attempt to address the proposed underlying causal nature of this relationship; rather, I offer here a preliminary look at how three aspects of the grammar of one particular language are implicated in stancetaking, as a first step to initiating interest in a functional link between stancetaking and language form. As alluded to above, considerable English-centered research already exists that explores the connection between stancetaking and grammar. 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. In particular, the grammar of English modals has proven to be a rich area for the epistemic evaluation of propositions (cf. review and discussion in Thompson and Hunston 2000: 20–21), as have epistemic phrases such as I think and I guess (cf. Kärkkäinen 2003, this volume). English adverbials have also been widely investigated as signaling various types of epistemic, attitudinal, and style stances (cf. Biber and Finegan 1988, 1989; Conrad and Biber 2000). And going beyond the traditional systems of English grammar, Hunston and Sinclair (2000) have proposed a “local grammar” of evaluative adjectives and nouns. Yet, outside of English, the link between grammar and stancetaking has virtually gone unnoticed in descriptive field linguistics. 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, like the one we find in, for example, the Longman Grammar of Spoken and Written English (Biber et al. 1999) where chapter 12 is entitled, The grammatical marking of stance. The purpose of this paper is to explore some aspects of the grammar of stancetaking 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. One of the reasons why
71
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 cognitively-oriented, propositional system for information coding. But when grammar is also understood as simultaneously being firmly rooted in social interaction, we see that much of what has previously been ascribed to meaning and information coding can, in addition, also be understood as social and interactional in nature. This paper seeks to flesh out three such resources in colloquial Indonesian – linguistic forms that have general grammatical functions, but, when examined in real contextualized language-in-interaction, also function at a local level of discourse to index stance. After a brief discussion of the data used for this study, I will focus on first-person-singular referring expressions, the -nya clitic, and verbal diathesis (voice). In conversational interaction, these may function in stancetaking to index identity, epistemicity, and positioning respectively. I have chosen these three grammatical resources specifically because they occur frequently in colloquial Indonesian, represent three very different facets of stancetaking, and clearly illustrate the kinds of gains that can be achieved by incorporating stancetaking into the realm of grammatical description.
2.
Data
The database for this study consists of a corpus of six transcribed audio recordings of naturally-occurring spontaneous Indonesian discourse. The corpus comprises 36,265 total words in 12,972 Intonation Units (IUs), nearly four hours of running speech. These six segments are part of larger speech events, which make up the approximately 25 hours of spoken colloquial Indonesian I collected while conducting fieldwork in Yogyakarta (Central Java) in 1996. The speech events were not elicited, and I (the foreign researcher) was not present during the recordings. Four of the segments in the corpus are casual face-to-face conversations among friends, the fifth is a discussion about music and art by a university student group that meets monthly, and the sixth is a radio call-in show about Islam. Nativespeaking research assistants and I transcribed the segments using a modified version of the Discourse Transcription system outlined in Du Bois et al. (1993), and subsequently glossed and coded them into a relational database. Englebretson (2003) provides further details about the six segments in the corpus, and about the specific methodology used for recording the speech events, choosing the segments, and transcribing and glossing the data. Appendices A and B of this paper outline the transcription and glossing conventions. All examples presented in this paper come from these six speech events, and are labeled with the title of the transcript and the line-number(s) of the IU(s) cited in the example.
Stancetaking in colloquial Indonesian conversation
In addition to the audio files and transcripts, the colloquial Indonesian corpus 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 information was unavailable). Speakers range in age from late teens to late 20s, and most are students at various universities in the city of Yogyakarta. The majority of the speakers (as with the majority of the Indonesian population in general) are from the island of Java; but the corpus comprises speakers from diverse regions of Indonesia as well, including Irian Jaya, Sumatra, Sulawesi, and East Timor (which, at the time of recording, was still part of the Republic of Indonesia). All speakers claim Indonesian as a native language, were educated in Indonesian, and use it in their daily lives and interactions, as is typical of educated urban residents of Indonesia. Most speakers speak a local language as well, typically Javanese, but report using Indonesian as the language of communication with each other most of the time – as is represented in the corpus. I have characterized this language variety elsewhere as spoken colloquial Indonesian (cf. Englebretson 2003), and noted its similarities with what other researchers have called Spoken Jakarta Indonesian (cf. Wouk 1989, 1999), which, as noted by Poedjosoedarmo (1982: 142), has spread widely outside of Jakarta as well. Before moving on to our discussion of the three aspects of stancetaking in Indonesian that form the core of this paper, a few brief notes on Indonesian grammar are in order. There are significant differences between the grammar of colloquial Indonesian and the grammar of the formal, standardized variety of Indonesian promulgated by the government, education, and media. For an overview of colloquial Indonesian grammar see Ewing (2005), and for an overview of standard Indonesian see Sneddon (1996) among others. Typologically, Indonesian is a primarily isolating language with sparse inflectional morphology. Person, number, and tense/aspect/mood categories are not indicated on verbs. Verbal affixes consist chiefly of voice and valence marking: agent-trigger, patient-trigger, middle, reflexive, applicative, and others. In conversation, noun-phrase ellipsis is quite frequent, so that clauses in fact often consist of a verb alone, and the NP arguments must be determined based on contextual and pragmatic factors. The relevance of these basic grammatical facts about Indonesian will become clear as the discussion progresses throughout the paper.
3.
First-person-singular reference as indices of identity
I will begin the discussion of grammatical resources used for specific stance purposes by focusing on the complex system of first-person referring expressions available to speakers of colloquial Indonesian. While English and many other lan-
73
74
Robert Englebretson
guages have only one unique first-person-singular (1SG) pronominal form (e.g., English I and its paradigmatically-related inflections me, my, and mine), colloquial Indonesian has several unique, paradigmatically-unrelated, 1SG forms. The general grammatical function of these expressions is, of course, as a means for the speaker to refer to the self in the ongoing discourse. However, in addition to this referential function, the variation observed among these multiple 1SG forms as used by a single speaker in conversation illustrates a specific means of stancetaking. This type of stancetaking is self-expressive, in that it serves to indicate how the speaker is constructing the self to be perceived by others. For example, just as culturally-relevant body posture, or physical stance, can index social relations and attitudes (arms folded across the chest to index social distance from or indifference toward an interlocutor; a particular hand gesture to signal in-group social affiliation; etc.), self-expressive linguistic stance likewise reflects and creates locally-relevant aspects of social and personal identity. In colloquial Indonesian, the varying forms available for first-person-singular reference are one resource that speakers can draw upon to evoke such things as social distance, casualness, or toughness. As compared with many local languages in Indonesia that have intricate levels of speech styles to index and maintain social relations (e.g., the speech levels and vocabulary registers of Javanese, as described by Errington 1988), Indonesian is relatively restricted in the kinds of power and solidarity relations that it allows speakers to express. This is often noted by Indonesian speakers in language attitude surveys. For example, the following quote is typical of the attitudes of many bilingual Javanese speakers toward Indonesian, as cited by Errington in his work on social identity, language shift, and codeswitching in Java. In response to my various queries about differences between their two languages, Javanese commonly alluded to Indonesian’s lack of stylistic elaboration as making it simple (I: sederhana), bland (I: tawar), or plain (I: polos) in comparison with Javanese. But, at least as normatively described, Indonesian does incorporate a few interactionally crucial stylistic distinctions in its personal pronoun rep(Errington 1998: 92) ertoires.
And indeed, in Indonesian as in many languages, pronouns are one means used for stancetaking, to create and maintain social relations and identity. The socialrelational work of pronouns has long been noted for other languages, for example, the social-psychological work of Brown and Gilman (1960), which characterizes the “familiar” and “formal” distinction found in the second-person pronouns of many European languages as reflecting interpersonal solidarity or deference between speaker and addressee. The so-called t/v distinction (for the French pronouns tu and vous) illustrates that pronouns are far more than merely referring expressions in the structuralist sense – speakers use them to indicate social rela-
Stancetaking in colloquial Indonesian conversation
tionships of affiliation or distance. From a more anthropologically-informed perspective, Mühlhäusler and Harré (1990) provide an overview of how pronouns in various languages relate to culturally-diverse concepts of the self, and how pronouns function in the expression of social relations and attitudes. Colloquial Indonesian speakers have at their disposal a range of personal pronouns, including second-person forms reminiscent of the t/v distinction widely discussed for other languages. However, 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.3 While grammars and textbooks of Indonesian typically contrast these forms based on macro-level variables of region and formality, evidence from the corpus demonstrates that these also function at a local level of discourse, and are used by a single speaker within the same speech event. Table 1 lists each of the 1SG referring expressions found in the corpus, along with their overall token frequencies. This table shows, in descending order of frequency, the five forms that speakers in the corpus regularly use in referring to themselves. Because of the high prevalence of argument ellipsis in colloquial Indonesian conversation, a sixth option also available to speakers is to use no form at all – so-called argument ellipsis – thus leaving the identity of the unexpressed arguments to contextual inference. Many of the clauses in the corpus, therefore, have no overt 1SG form, even though they are clearly referring to the speaker. The question of what motivates a speaker to either use or to ellipt a 1SG form is of great interest in its own right, but lies outside the scope of the present paper.4 I am concerned here with the ways in which overt first-person-singular forms are used, and, in particular, how they function in stancetaking. I will leave the issue of colloquial Indonesian argument expression versus ellipsis for future research, focusing here only on the five overt forms shown in the preceding table. The first four (aku, sya, gua, and tak) are 1SG pronouns, and the fifth is selfreference through the use of one’s own name. In all discussion of these forms in the literature that I am aware of, including textbooks, dictionaries, pedagogical materials, and descriptive grammars, these 1SG forms are characterized in terms Table 1. Total 1SG referring expressions in the corpus Form
Total tokens
aku (aku, ku-, and -ku) saya gua/gue tak proper name
578 222 83 37 35
75
76
Robert Englebretson
of the traditional macro-level sociolinguistic variables of formality, region, and ethnicity as follows. The most frequent of these pronouns observed in the corpus is aku, including its paradigmatically-related clitic forms ku- (a first-person proclitic on verbs) and -ku (a possessive enclitic and object of a preposition). Aku is generally characterized as familiar and informal, used among intimates, or by a person of higher social status to a person of lower social status. The second-most frequent 1SG pronoun in the corpus is saya, which is prescriptively the all-purpose Indonesian 1SG pronoun. It is considered more formal than aku, and the most appropriate form for public interaction and among people who are not socially intimate. The next-most-frequent form in the corpus, gua, and its variant pronunciation gue, was borrowed into Indonesian through Hokkien Chinese. It is typically associated with young urban residents of Jakarta and/or speakers of Chinese-Malay background. The least-frequent 1SG pronoun in the corpus is tak, which is generally characterized as a Javanese loan. The final entry in the table indicates that there are 35 instances in the corpus of speakers referring to themselves by use of their own name. According to the impressions of native-speaking Indonesian consultants, 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. The following entries from Echols and Shadily’s (1989) Indonesian-English dictionary summarize the 1SG pronoun forms found in my corpus, reflecting the way these pronouns are generally treated in the Indonesianist literature.
aku saya gua tak
“I (familiar, intimate)” “I, me, my” “(Jakartan, Chinese, Colloquial) I, familiar first person pronoun” “(Javanese) I, by me”
Readers who are not familiar with Indonesian grammar may wonder whether any of these forms are inflectionally related to each other. The answer is negative. Each is invariant for grammatical categories such as case or possession; other than the clitic forms of aku, no paradigmatic forms of these pronouns exist. In sum, two of these forms (aku and saya) are traditionally characterized by differences in formality, the remaining two pronouns (gua/gue and tak) are characterized based on the regional/ethnic/linguistic background of the speaker, and the use of self-reference 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, intuition-based claims of analysts and native speakers, and, as I will demonstrate shortly, do not fully account for their use in conversational interaction. There has been an overall lack of empirical research into the actual distribution of these forms in naturally-occurring language-in-use. A welcome exception is Sneddon (2002) who provides a quantitative analysis of 1SG
Stancetaking in colloquial Indonesian conversation
forms (and other stylistically-stratified variables) in a corpus of various genres of spoken Indonesian among residents of Jakarta. Sneddon’s quantitative findings do bear out the claims that 1SG pronouns vary based on formality level. Results from my own corpus support this as well, based on the following evidence. As mentioned in Section 2, above, five of the speech events in my corpus consist of face-to-face conversation among people who know each other 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. This speech event is different from the other five in a number of ways: (1) it is a public, radio-broadcast telephone interaction, while the other five are private, face-to-face conversations; (2) the callers are not socially acquainted with the radio host or the cleric, whereas in the five other speech events, participants know each other well and interact with each other regularly; (3) in the call-in show, participant roles and social relations are clearly delineated – the cleric is in a position of authority as expert, and the callers are positioning themselves as lacking the knowledge that the cleric provides; in the other five speech events, participant roles are more balanced, and there is no overtly recognized authority or expert. On all three of these dimensions, then, Tanya-Jawab represents a far more formal register than any of the other five speech events in the corpus. Thus, if 1SG pronominal forms are indeed sensitive to formality level, 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. As shown in Table 2, the data do indeed provide strong evidence for this claim. This table shows a striking difference between 1SG reference in the TanyaJawab speech event as compared with the aggregate use of 1SG reference in the other five.5 In the radio call-in show, speakers overwhelmingly prefer saya, with 98 instances of this form and only one instance of any other (self-reference by means of the speaker’s own proper name). In this speech event, there are no instances whatsoever of the familiar aku nor of the regionally/ethnically marked gua/gue or tak. These numbers provide strong evidence that speakers treat saya as the 1SG pronominal form most appropriate for formal public interaction. In contrast, the overwhelming preference for aku found in the conversational speech events suggests that speakers treat aku as the preferred form for informal interaction among social familiars. These empirical findings seem to provide robust support for the general claims that the macro-level variables of formality and intimacy strongly determine the use of saya or aku respectively. However, as I will demonstrate in the remainder of this section, the situation is far more complex than could be accounted for by a direct deterministic relationship between social variables and linguistic form.
77
78
Robert Englebretson
Table 2. Comparison of 1SG reference in informal and formal speech events Form
Face-to-face tokens
“Tanya-Jawab” tokens
Total
aku saya gua/gue tak proper name
578 124 83 37 34
0 98 0 0 1
578 222 83 37 35
The data suggest that pronoun choice in conversational interaction is not predetermined based on static social variables, but is dynamic, takes place at the local level of discourse, and is used in stancetaking to index the speaker’s construction and expression of identities. Primary evidence for this claim comes from observing 1SG usage by single speakers, with the same interlocutors, in the same speech event. In the five face-to-face conversations in the corpus, there are no speakers who refer to themselves with only one 1SG form; every speaker in the corpus (aside from those in the Tanya-Jawab radio call-in show) regularly uses more than one 1SG form. Following is a table illustrating the use of 1SG forms in the speech of Ari (speaker A) from the Pencuri speech event. Ari is one of the more loquacious speakers in the corpus, 21 years of age, female, originally from Eastern Indonesia, and a student in Yogyakarta at the time of recording. Note that her repertoire of 1SG forms includes all five found in the corpus. Table 3 shows a single speaker using all five of the 1SG forms to refer to herself, with the same interlocutors, in the same physical setting, during the same speech event. I have produced similar tables for all speakers in the corpus, and as mentioned above, there are no single-form speakers in the face-to-face conversations. Not all speakers use all five forms in such striking numbers as does this particular speaker, but it is an empirical fact that every speaker in the face-to-face conversational data uses two or more. These findings suggest that the traditional, top-down, a priori account of what motivates 1SG use is incomplete. As an actual example of the use of 1SG forms in real-time conversational interaction, I offer the following short excerpt of 18 intonation units (lasting 17.7 seconds) of a conversational narrative, again from Ari in the Pencuri speech event. 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. For the benefit of non-Indonesianist readers and to facilitate the reading of this excerpt, I have included the actual Indonesian 1SG forms in parentheses in the English free translation. Other material in parentheses indicates that no corresponding form is used in the Indonesian data, but is necessary in the English free translation.
Stancetaking in colloquial Indonesian conversation
Table 3. 1SG forms in the speech of a single speaker Form
Tokens
aku saya proper name gua/gue tak
81 40 25 8 5
(1) Pencuri IU 209–227 209 A: ... Dia di ini. 3sg at this A: She was at, um, 210 .. Di depan pintu rumah Ari. at front door house Ari in front of the door to my_(Ari) house. 211 D: Kamu tanya dia. 2sg ask 3sg D: You asked her? 212 A: .. Aku kan, 1sg prt A: I_(aku) was um, 213 bingung lho kok. confused prt prt (I) was confused, you know. 214 ... Perasaan tadi tuh, feeling just:now prt (My) feeling just then was that, 215 .. yang pegang dompet dia kan, rel hold wallet 3sg prt she was the one who had taken (my) wallet. 216 [terus gua bilang-in], next 1sg say-app [So then I_(gua) said to her], 217 M: [Yah]. yeah M: [Yeah].
79
80 Robert Englebretson
218 A: ... Nama-nya kan Rifka. name-pos prt Rifka A: Her name was Rifka. 219 ... Eh Rif, hey Rif “Hey, Rif! 220 .. Kamu, 2sg You, 221 tadi nge-lihat-in, just:now at-see-app were just now looking at, 222 duit saya yang dalam= dompet ya. money 1sg rel inside wallet prt my_(saya) money that was inside (my) wallet, weren’t you?” 223 Nggak tuh, neg prt “No, hunh-unh. 224 nggak tahu. neg know (I) don’t know (anything about it).” 225 ... Terus Ari masuk lagi cari ke dalam, ... next Ari enter again search to inside So then I_(Ari) went in again and looked for it inside. 226 227
dia itu ngak- -3sg that neg She didn’t – nggak lari gitu lho. neg run like:that prt [she] didn’t run away.
Note that in this brief excerpt of only 18 IUs of conversation, a total of 17.7 seconds of talk, the speaker seamlessly switches among four of the five 1SG forms found in the corpus. She uses Ari (her own name) in IUs 210 and 225, aku in IU 212, gua in 216, and saya to refer to herself in reported speech in IU 222. 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, and (2) distribution of 1SG forms could be based
Stancetaking in colloquial Indonesian conversation
on macro-level social variables. Both of these hypotheses fail to fully account for what we actually observe speakers doing, and I will therefore suggest a third approach which more thoroughly accounts for the observed distribution in 1SG forms; namely, speakers are actively using 1SG forms in stancetaking in order to construct and project aspects of their social and personal identities. The first hypothesis to consider is that 1SG forms may be motivated and constrained by grammar. If a speaker’s choice of 1SG form were determined by grammatical factors, then, following classic structuralist reasoning, we would expect to find 1SG forms occurring in some sort of contrastive distribution. In other words, some 1SG forms in the corpus would occur in grammatical environments where other 1SG forms do not occur. However, (with the possible exception of tak, to which I will return shortly), there appear to be no such grammatical restrictions. For example, the 1SG form saya in IU 222 duit saya ‘my money’ would be perfectly acceptable with the substitution of any of the other 1SG forms found in this short excerpt: duit gua (using gua/gue), duitku (using the enclitic form of aku), or duit Ari (using the speaker’s own name) are all acceptable and natural. In fact, these very tokens are actually attested elsewhere in the corpus, as illustrated in the following three examples. (2) Pencuri IU 2403 Duit -ku nggak ada. money -1sg neg ex I don’t have any money. (lit., ‘My_(aku) money doesn’t exist.’) (3) Dingdong IU 1002 Nggak ada duit gua. neg ex money 1sg I don’t have any money. (lit., ‘My_gua money doesn’t exist.’6) (4) Pencuri IU 266 Cabut duit Ari semua. pull money Ari all (She) pulled out all my_(Ari) money.
These three examples, along with IU 222 in excerpt (1), above, illustrate that each of these four 1SG forms (saya, aku, gua, 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.’ These four forms, along with tak, are all fully attested in the corpus in other grammatical roles too: as both the trigger and the nontrigger argument in transitive clauses, and as the single argument of intransitive
81
82
Robert Englebretson
clauses. These distributional facts suggest that the variation among 1SG forms is not grammatically conditioned. The one form which does appear to have some grammatical restriction is tak; it never occurs in the corpus as a possessor in a possessive NP, and it tends to strongly collocate with certain verbs of cognition. Eighteen out of the 37 tokens of tak in the corpus co-occur with the verb pikir ‘think,’ suggesting that tak pikir ‘I think’ is a grammaticized epistemic phrase (cf. Kärkkäinen 2003, this volume on English I think and I guess respectively; cf. Rauniomaa, this volume, on similar constructions in Finnish). However, while this collocation does suggest a restricted grammatical profile for tak, it nonetheless overlaps with the other 1SG forms in other environments. The hypothesis that the variation of 1SG forms is grammatically conditioned receives minimal support from tak, and no support whatsoever from the other four 1SG forms. The second hypothesis to consider is that 1SG use is linked to macro-level social and stylistic variables such as formality and regional background. I have already discussed this to some extent above, and have shown, based on a comparison of 1SG usage in a radio call-in show with 1SG usage found in everyday conversation among social familiars, that saya and aku are strongly correlated with formality. Saya is the only 1SG form found in the call-in show (except for one instance of self-reference by proper name), while aku is the preferred form in the conversational data. However, this cannot account for the fact that there are still 124 tokens of saya in the informal conversational data, as shown in Table 2. It also cannot account for the distribution observed in Table 3, where a single speaker freely uses both saya and aku (along with the other 1SG forms) in the same speech event, and even within the same short span of discourse as seen in excerpt (1). The regional account of gua and tak is similarly problematic. For example, as shown in Table 3, 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), along with the other three 1SG forms. However, this speaker has never lived in Jakarta at all (thus suggesting that calling gua Jakartan is an oversimplification), and claims not to speak Javanese (suggesting that tak is not simply Javanese code-mixing.). Thus, a regional account of gua and tak does not account for much of what speakers are actually doing in conversational interaction. The social variables hypothesis does, however, appear to receive strong support from 1SG reference by means of proper name. Traditionally, this is seen as a form used by young, unmarried women; and the three speakers in the corpus who use this form are in fact young, unmarried, and female. However, there are unmarried women of the same age in the corpus who never use this form, and the speakers who do use their names as 1SG reference also use other 1SG forms as well (e.g., Ari in Table 3, who uses all five 1SG forms). So for speakers like Ari who
Stancetaking in colloquial Indonesian conversation
do use their names as 1SG reference, the social variables hypothesis still does not explain what would condition her to use her name at some points in the conversation, and to use different 1SG pronouns at other points in the talk. 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. As shown in Table 3, a single speaker may use all five of the 1SG forms, despite the fact that she is using them all in the very same speech event. In other words, the variables of formality and regional origin for Ari are remaining constant, yet her use of 1SG form varies; she is speaking with the same interlocutors (housemates and friends whom she has known for nearly two years), in the same physical context (in the common room of their boarding house), and her ethnic/ regional/language background also does not change throughout the interaction. The fact that 1SG usage varies, while the macro-level sociolinguistic variables remain constant, suggest that there is an additional level of detail not captured by the social variables hypothesis. These shortcomings of the social variables hypothesis reflect larger, problematic assumptions present in much of traditional variationist sociolinguistic research regarding the nature of identity categories in general. Variationist sociolinguistics has tended to assume identity “as a pre-discursive construct that correlates with, or even causes particular [language] behaviours” (Benwell and Stokoe 2006: 26). In other words, in traditional variationist work, social categories of identity are treated as relatively stable, a priori characteristics of the speaker (age, gender, regional origin, language background, etc.) or stable attributes of the particular speech situation (e.g., formality, casualness, etc.) In this view, the particular variant a speaker uses in any given instant of talk falls out probabilistically from these pre-existing social and stylistic categories. If, on the other hand, we understand identity categories (such as gender, age, regional origin of speakers, or formality of speech events) from a performative point of view – as something speakers ‘do,’ rather than as something speakers ‘are’ – then we can begin to approach the variation among Indonesian 1SG forms from a new perspective, 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. Rather than seeking macro-level variables to explain the distribution of 1SG forms in the data, 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?, i.e., What kind of stance is the speaker evoking by using this form at this time? According to a performative conception of identity, speakers may use language to evoke social or personal identity categories in order to achieve certain goals within talk. Rather than being relatively stable attributes that pre-exist the current discourse, identity categories are emergent, socially constructed practices
83
84
Robert Englebretson
of language use (cf. Bucholtz and Hall 2005 for similar discussion). The many types of identities in discourse include “local identity categories and transitory interactional positions” and “temporary and interactionally specific stances and participant roles” (Bucholtz and Hall 2005: 592). Accordingly, one way that a speaker “performs” identity is through the packaging of utterances in ways that reflect and index specific categories. As demonstrated in Table 3, and in excerpt (1), the speaker has several potential options at any given moment in talk, as to how to package the utterance for 1SG self-reference. I shall now offer some qualitative observations of how 1SG forms are used in the performance of identity. I suggest that these various forms serve to index, or evoke, specific types of identity categories. Bearing in mind a performative conception of identity, I will begin by discussing the use of aku and saya from this perspective. 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. Saya is the only 1SG pronoun used in the call-in radio show, with 98 tokens (plus one 1SG reference by use of proper name). I suggest that speakers here are actively using saya, among other lexical and grammatical items, to take a formal stance: to construct a public social and personal identity, to show deference, and to index social distance among participants. On the other hand, in the casual conversational data, aku is overwhelmingly the most frequent form, with 578 tokens in these five speech events – nearly five times more frequent than the use of saya in these same speech events. I would contend that speakers here are using aku as one means of taking a casual stance: actively constructing their personal identities as informal, relaxing the prescribed norms of public language use, and building the social intimacy among speakers typical of this type of interaction. But what about the 124 tokens of saya in the conversational data? If speakers are actively constructing these interactions as casual and intimate, as suggested by the high frequency of aku, then why are these same speakers simultaneously using saya, with its connotations of formality and distance? I suggest that when speakers use saya in informal conversation, they are shifting their stance: they are using this form to construct social distance or formality at a particular point in the ongoing interaction. (To return to the analogy of physical stance, 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.) 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, following Tannen (1986, 1989) I shall hereafter refer to as constructed dialogue), even when a different pronoun would have been used in the original context of utterance. As an example, consider the use of saya in IU 222 of excerpt (1), above. In this clause, which starts in IU 216 with the quotative verb bilang ‘say,’ Ari is reporting the words she alleg-
Stancetaking in colloquial Indonesian conversation
edly had spoken to her friend, “Hey, Rif! You were just now looking at my_(saya) money that was inside (my) wallet, weren’t you?”. Interestingly, if 1SG usage were conditioned by casualness and intimacy, one would have expected aku here, since aku would most certainly have been the form she would have used with her friend in the original 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, the speaker is using saya, presumably to reflect a more distant stance than the informal aku. For several speakers in the corpus, saya is only found in contexts of constructed dialogue – either the alleged words of others, or in the speakers own reported utterance as in this example. Based on these observations, I suggest that speakers are using saya to take a stance on reported speech or thought: to imply social distance from it. While saya is the norm in constructed dialogue in the corpus, it is by no means the only form of 1SG reference in this context; aku and other 1SG forms are commonly found in reported speech and thought as well. For this reason, it is illuminating to compare the example of saya in IU 222 of excerpt (1) (as just discussed) with another instance, one in which the speaker uses the more intimate aku. The following excerpt comes from a narrative about a woman named Maya who is nearly tricked into buying shoddy towels at an outdoor market. The full narrative will be presented and explicated for other purposes in Section 5 below, but for the present discussion, the relevant aspect of this example is the use of aku in the constructed dialogue in IU 2753. (5) 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_(aku) want this towel for 3,000 (Rupiah). 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.”
85
86 Robert Englebretson
2756 dia kan, 3sg prt he 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, the speaker (Indra) provides an abstract of this portion of her narrative, in which she summarizes what happens to Maya at the market: “Maya got really attracted” (by the towels she saw ). The subsequent four IUs construct Maya’s inner dialogue, including her thoughts about the towels and their price, and IUs 2756–2757 explicitly frame this speech as having been constructed dialogue “her thoughts…”. The IU of interest for the present discussion is IU 2753, where Maya is reported to think: “I_(aku) want this towel for 3,000 (Rupiah).” What is noteworthy about the constructed dialogue in IU 2753 of this excerpt, as contrasted with Ari’s constructed dialogue in IU 222 of excerpt (1) above, is that Maya’s reported ‘thoughts’ use the more intimate aku, while Ari’s reported ‘speech’ uses the more distant saya. 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 creating, 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; perhaps ‘thoughts’ are being presented as more internalized to the speaker, and are thus constructed with greater intimacy, while ‘speech’ is more observable and less internalized, and thus is constructed with greater distance (cf. the theory of “territory of information”, as articulated by Kamio 1997). While a more thorough study of constructed dialogue in Indonesian conversation is clearly warranted to assess this conjecture, 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. In other words, 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, as lower in epistemic value, and (for whatever reason) as more distant from the speaker. On the other hand, the use of aku in constructed dialogue reflects a conception of it as more vivid, as higher in epistemic value, and as less distant. Clearly this observation warrants further in-depth investigation, but I offer it here along with the presented examples to illustrate one resource which colloquial Indonesian speakers have of taking a stance. In sum, in informal colloquial Indonesian conversation (whether in the context of constructed dialogue or not), the use of saya or aku is one means of stylistic elaboration avail-
Stancetaking in colloquial Indonesian conversation
able to speakers, as they use language to construct their social worlds through stancetaking. The third-most frequent 1SG form in the corpus is gua/gue, with 83 tokens in the informal conversational data. This pronoun is generally associated with colloquial Indonesian spoken in Jakarta and/or by speakers of Chinese ethnicity. Yet, as discussed earlier, a regional/ethnic characterization of this pronoun does not accurately account for the observed data. For example, Table 3 shows that Ari uses eight tokens of gua in the corpus, yet she has never lived in Jakarta and is not ethnically Chinese. Just as with saya and aku, I suggest that speakers use gua in stancetaking, to evoke some type of temporarily relevant identity in the ongoing discourse. As a first approximation of what such a temporarily-relevant identity might be, consider Ari’s use of gua in IU 216 of excerpt (1) above: “So then I_(gua) said to her…” Note, crucially, that this form occurs to frame the rather blunt accusation Ari reports having made of her friend (IUs 219–222): “Hey, Rif! You were just now looking at my money that was inside my wallet, weren’t you?” In this short excerpt, Ari presents herself as blunt and outspoken, constructing her identity as someone who is ‘tough’ and unafraid to confront her friend whom she believes is a thief. I would like to suggest that her use of gua contributes to this locally-relevant construction of identity. Rather than indexing regional affiliation per se, gua draws upon common language attitudes and stereotypes often associated with speakers who use this form. 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.” I readily acknowledge the need for both more in-depth language attitude surveys and qualitative sequential analysis of conversation to assess this hypothesis, but as a preliminary step toward this goal, I offer the case of the most prolific user of gua in the corpus: speaker L from the speech event entitled Dingdong (the onomatopoeic name of a pinball arcade game discussed in the speech event). This speaker is male, 19 years of age, and had lived in Jakarta from age 10 until age 13. He uses only two 1SG forms in the data: 63 tokens of gua and seven tokens of aku. In other words, of the 83 gua tokens in the data, speaker L produces 63 of them (75.9%). When I played portions of this recording separately to three Indonesian consultants, all expressed the view that this speaker sounded ‘tough’ or ‘rude’ or ‘macho,’ and one consultant was offended by the topics of this spontaneous conversation (playing video games, drinking at a local bar, and flirting with women). 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. When speakers use gua, they are indexing these stereotypes, whether or not they have any affiliation with Jakarta at all (cf. Rampton 2005 for work on “crossing,” or appropriating the speech styles
87
88
Robert Englebretson
of other groups for specific social purposes; Johnstone, this volume, on the use of stancetaking to bolster specific local identities). To summarize this section, I have shown that 1SG self-reference in colloquial Indonesian is an interactional practice available to speakers for creating, maintaining, and implying facets of their social and personal identities. I readily acknowledge that substantial work remains to be done on this topic, both quantitative and ethnographic. I have not offered an account here of how tak or proper names are used in stancetaking, partly because there are so few tokens of these forms in the corpus (37 tokens of tak and 35 of proper name); I leave this to future research, with the suggestion that, as with the other three pronouns discussed here, their stance functions can best be accounted for by observing the kinds of identity categories speakers evoke at a local level of conversation. In this section, I have offered here a first step toward the goal of moving Indonesian grammatical description beyond traditional formal and referential boundaries, and toward a greater recognition of the ways in which grammatical resources are used in the service of stancetaking. In this case, 1SG forms do far more than simply refer to the speaker, and the variation among these forms does not simply fall out from social variables. Rather, much of self-reference is closely linked to self-expressive stancetaking, and is managed locally in the ongoing discursive construction of identity. Just as a speaker’s physical stance reflects and indexes temporary social categories, the linguistic stance evoked by the use of one 1SG form over another creates similar transitory identities, such as social distance with saya or toughness with gua. The 1SG forms of Indonesian are a general grammatical resource which, in addition to their referential function of referring to the speaker, are simultaneously employed in the construction and maintenance of stance.
4.
Epistemic -nya constructions
The previous section has illustrated how Indonesian speakers may use first-person singular pronouns to construct aspects of their personal and social identities. This type of linguistic stancetaking is self-expressive and reflexive. Just as the act of physically taking a stance may often entail the outward presentation of some facet of the self, through body posture or gesture or other means, the linguistic expression of stance may likewise entail the display of some currently-relevant part of the speaker’s identity. The kind of stancetaking discussed in the present section, on the other hand, involves epistemic stance toward something external to the speaker – in this case toward the current utterance itself. Following Wu, this type of epistemic stance refers to “a speaker’s indication of how he or she knows
Stancetaking in colloquial Indonesian conversation
about, is commenting on, or is taking an affective or other position toward the person or matter being addressed” (2004: 3). In this section, I will illustrate how Indonesian speakers use the -nya clitic to do just this. In addition to its general, well-described, referential functions as a marker of third-person possession, definiteness, and nominalization, this clitic frequently occurs on lexemes related to cognition, utterance, or modality. In such a context, these constructions function as adverbials, and are used for epistemic stancetaking regarding the current utterance. The kinds of epistemic stances implicated by -nya constructions include: (1) evidentiality: the source of knowledge of the current utterance; (2) assessment of interactional relevance: the degree of value a speaker places on the utterance, 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. These stance-related functions have generally gone unnoticed in descriptive grammars of Indonesian, which tend to focus only on the referential and grammatical aspects of this clitic, with little, if any, discussion of its use in stancetaking. The -nya clitic is the most prevalent bound morpheme found in the corpus, with 1,570 tokens in the six speech events. This frequency is greater than the occurrences of all other grammatical morphemes in the corpus, including the agent- and patient-trigger verbal prefixes as well as pronominal clitics, suggesting that -nya deserves recognition as the most prevalent bound morpheme found in colloquial Indonesian in general. Yet, outside of my own work on -nya (see Englebretson 2003: Chap. 5 for a more thorough discussion than is possible in this brief section), it has received relatively little analysis. Grammars of formal varieties of Indonesian typically gloss this as a third-person-singular possessive marker, and also recognize its use as a marker of definiteness and as a nominalizer. These three referential uses of -nya are found in the colloquial corpus too, as illustrated by the following three examples. (6) Pencuri IU 218 A: ... Nama-nya kan Rifka. name-pos prt Rifka Her name was Rifka.
In this example, the -nya clitic occurs on the noun nama ‘name’ to indicate a third-singular possessor ‘her name.’ The following example contains two tokens of -nya. The first is a definite marker, while the second again indexes a third-person possessor as in the previous example.
89
90 Robert Englebretson
(7) Pencuri IU 626–627 626 D: dari belakang ini, from back this 627 pencopet-nya buka kalung-nya. pickpocket-def open necklace-pos From behind, the thief undid her (Agnes’s) necklace.
IU 627 contains two instances of the -nya clitic. The first marks the noun pencopet ‘thief ’ as definite; this noun was mentioned previously in the discourse context, in IU 599 (not shown here), and plays a central role in the ongoing narrative. The second instance of -nya in IU 627 marks third-person possession on the noun kalung ‘necklace,’ just as in example (6). IU 834 of the next example illustrates a third grammatical function of ‑nya, that of nominalization. (8) Dingdong IU 833–835 833 ... pokok-nya bukan minum bir aja. main-stm neg drink beer just 834 ... Minum-nya, drink-nom 835 .. me-lebih-i= Martini kata-nya. at-more-app martini word-stm the thing is, (they) weren’t just drinking beer. (Their) drinking even surpassed martinis, they said.
In this excerpt, speaker L is telling his friends about a group of women he had seen the previous evening at a bar, and is commenting on his shock about the quantity of alcohol and types of drinks they were consuming. There are three instances of -nya in this example. The first and third are stance markers (STM) to which we will return shortly. The second, on the verb minum ‘drink’ in IU 834, is a nominalizer. It marks this word as a noun, which then becomes the subject of the clause in IU 835, ‘(Their) drinking even surpassed martinis, they said.’ The three functions of -nya we have observed so far – third-person-singular possession, definiteness, and nominalization – are the typical grammatical functions of this morpheme discussed in descriptive and pedagogical grammars of Indonesian. Yet, this account of -nya does not fully address the range of use observed in colloquial Indonesian. More than one-third of the instances of -nya in the corpus, 580 out of 1,570 tokens, are not marking third-person-singular possession, are not indicating definiteness, and are not nominalizing a verb to be a clausal argument. Rather, these instances of -nya are being used in stancetaking,
Stancetaking in colloquial Indonesian conversation
enabling a speaker to comment on some aspect of the current utterance: how the speaker knows about it (evidentiality); how the speaker evaluates it in light of the current conversation (assessment of interactional relevance); or how the speaker feels about it (emotional attitude). Grammatically, these expressions are adverbials, as seen from their syntactic freedom to occur at numerous positions in the clause.7 To illustrate two of the stance-marking functions of -nya constructions, return to example (8) above. Note that there are three instances of -nya in this example, one in each IU. 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. The remaining two -nya tokens in this example are stance markers. The first, in IU 833, consists of -nya on pokok (a word roughly translated as ‘main,’ ‘fundamental,’ or ‘basic’). This serves as an epistemic evaluation of the interactional relevance of the utterance for the discourse at hand. With this 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. It is the fact that these women were “not just drinking beer,” which provides essential background for the rest of his unfolding narrative about the antics of this group of friends. The second instance of ‑nya as a stance marker, in example (8), occurs in IU 835 on the noun kata ‘word.’ Here, -nya marks evidentiality, the source of knowledge for this utterance, indicating that it originated from the words of others. That is, speaker L knows what the women were drinking because this is something they had told him. In sum, these two tokens of -nya illustrate that when this clitic occurs on lexemes related to utterance, cognition, or modality, it forms an adverbial which marks the speaker’s stance toward the rest of the clause. The remainder of this section will take up each of the three stance functions of -nya constructions in turn. One stance-related function of -nya constructions is the marking of evidentiality, the source of knowledge of the speaker’s utterance. Following are three more examples, each illustrating a type of evidentiality: knowledge based on the words of another (example 9), knowledge from general inference (example 10), and knowledge from visual perception (example 11). (9) Pencuri IU 986–987 986 ... Dia kan kata-nya mau n-yari saudara-nya 3sg prt word-stm want at-look:for relative-pos 987 di sini. at here He said he wanted to visit his relatives here.
91
92
Robert Englebretson
This example is parallel to the token already discussed in IU 835 of example (8). In IU 986 of example (9), kata ‘word’ is suffixed with -nya, again indicating that the speaker’s source of knowledge is the words of another. The speaker knows that the person in question wanted to visit his brother, because this was something told to her by a friend. There are 59 instances of katanya (‘word’-nya) in the corpus, all of which mark evidentiality based on another’s speech. The following example illustrates -nya marking evidentiality based on general inference. (10) Pencuri IU 2299 mati kayak-nya jam Mega. dead like-stm watch Mega My watch seems to have stopped.
In this example, the colloquial preposition kayak ‘like/as’ is suffixed with the clitic -nya, making it into an adverbial expressing a general evidential source. The speaker, Mega, notices that her watch is showing the same time as it had given earlier, and infers that it therefore must have stopped. Kayaknya (‘like/as’-nya) occurs 76 times in the corpus. As in this example, this expression does not specify a particular evidential source, but indicates that the speaker is making the statement based on general inference. IU 1249 of the following example illustrates a third type of evidential source, in this case visual perception. (11) Pencuri IU 1247–1251 1247 Trus ada ibu-ibu yang keren itu lho next ex mother-mother rel well:dressed that prt 1248 .. Pokok-nya, main-stm 1249 .. ke-lihat-an-nya orang kaya itu lho nonvol-see-nom-stm person rich that prt 1250 dari ? from what 1251 .. Dandan-nya, attire-pos And then there were some well-dressed women. The thing is, they looked like rich people – because of, what’s-it?, the way they were dressed.
In IU 1249, the speaker states that the women must be rich, and indicates her evidential source for this claim as based on how they looked. The verb lihat ‘see’ is affixed with the stative/nonvolitional circumfix ke- -an, changing the meaning
Stancetaking in colloquial Indonesian conversation
of the verb to ‘visible,’ which is then suffixed with the evidential -nya. In sum, we have seen three types of evidential sources indicated by -nya constructions: evidence from the words of another, evidence from general inference, and evidence from direct (visual) perception. This clitic also commonly occurs on the verbs pikir ‘think’ and rasa ‘feel,’ to indicate the evidential source as one’s own thoughts and feelings. Note that example (11) contains two additional tokens of -nya. The occurrence of -nya in IU 1251 is a possessive.8 The remaining -nya token, in IU 1248, is a stance marker, indicating the speaker’s assessment of interactional relevance, which is the second type of epistemic stance we will be addressing in this section. As discussed previously regarding IU 833 in example (8) the word pokok (‘main’/‘fundamental’/‘basic’) is suffixed with -nya, indicating the speaker’s attitude toward the rest of the clause as being central to the understanding of the ongoing interaction. Here, in IU 1248 of example (11), the speaker is highlighting the fact that the women looked rich, as essential background for understanding why they were being targeted by pickpockets. The corpus contains 53 similar instances of pokoknya (‘main’-nya), suggesting that this is a highly grammaticized way of highlighting the main point of the ongoing interaction. Following are two more examples of -nya constructions used as a stance marker of interactional relevance. (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. sibling-pos that The gist of it was that he couldn’t meet his brother that night.
In this example, the speaker is concluding a narrative told to her by another person. 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 narrative. As in the previous example, 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. Other similar expressions found in the corpus include misalnya ‘for example’ (‘example’ -nya), used by a speaker to frame the utterance as a specific, usually hypothetical example of the overall general point of the discourse; masalahnya ‘the problem is’ (‘problem’-nya) which indicates that the current utterance is somehow prob-
93
94 Robert Englebretson
lematic in light of the ongoing interaction; as well as pokoknya ‘the thing is’ and intinya ‘the gist of it is,’ which have been illustrated in the previous examples. 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. The following three examples are illustrative. (13) Pencuri IU 545–546 545 A: Takut-nya tu, fear-stm that 546 mereka bawa benda tajam, 3pl bring object sharp (We’re) afraid they might bring sharp objects.
In this example, the verb takut ‘fear’ is suffixed with -nya, indicating the speaker’s emotional attitude toward the proposition in the rest of the utterance. She is afraid that the pickpockets might ‘bring sharp objects’ with them on the bus, and would hurt anyone who tries to thwart their criminal activities. Takutnya (‘fear’-nya) in the first IU thus frames the next IU in terms of the speaker’s emotional attitude. The following example is similar. (14) Blewah IU 13 ... Takut-nya lupa ya? fear-stm forget prt Are you afraid you’ll forget?
As in the previous example, the verb takut ‘fear’ is suffixed with -nya, framing the proposition in the rest of the clause ‘(you’ll) forget’ as being something to potentially fear. The following example illustrates another type of mental/emotional attitude, namely the speaker evaluating the rest of the utterance as being lucky or favorable. (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, right?
This clause serves as the coda to a story about one of the speakers and her friend Agnes, who thwarted an attempted robbery on a bus because she felt someone trying to undo her necklace. Speaker L uses the noun untung ‘fortune/luck’ with
Stancetaking in colloquial Indonesian conversation
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. Returning 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 toward 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 participates. Yet, despite its pervasiveness in everyday language use, with the exception 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 nominalization. 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 grammatical voice system of agent-trigger and patient-trigger clauses. I will demonstrate that in addition to the syntactic, semantic, pragmatic, and discourse-structur-
95
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 about. As with many Western Austronesian languages, the verb morphology of Indonesian 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 Indonesian; 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 illustrations 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
(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 patienttrigger 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 structurally similar to English passives, note that there are substantial functional differences 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 impossible, and is often not the most accurate choice. I have included a literal English 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 explanations for their use tend to focus on syntactic factors (e.g., only the trigger argument may participate in relativization and act as the pivot in certain types of clause combination); clusters of related semantic factors such as animacy, referentiality, 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 important aspect of Indonesian voice has so far been overlooked. Namely, in addition to previous explanations such as those just mentioned, the Indonesian voice system 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
97
98 Robert Englebretson
to the Indonesian voice system. The ready availability of agent-trigger and patient-trigger morphology facilitates the positioning of individuals as intentional or non-intentional, as agentive or patientive, and as morally responsible and accountable, or as morally not responsible and unaccountable. When viewed from this perspective, the Indonesian voice system is far more than a means of expressing 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 position 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 mistake 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
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 conscious 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 agentive subject, but this restriction in grammatical roles is lifted once she is again fully 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).
99
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.
Stancetaking in colloquial Indonesian conversation 101
(20) Pencuri IU 2764–2771 2764 di-tarik-tarik, pt-pull-pull (Maya) got pulled in and pulled in. 2765 (H) itu udah ng-otot banget. that perfv at-persist very But (Erika) kept really persisting. 2766 Si Maya itu kamu itu di-singkir-singkir kee-,11 prt Maya that 2sg that pt-pull:away-pull:away trunc Kept pulling Maya away. (lit., Maya was pulled away.) 2767 Sampai ber-antem sendiri ber-dua itu. until mid-quarrel self mid-two that Until the two of them started fighting with each other. 2768 (H) Tapi Erika-nya itu ng-otot, but Erika-def that at-persist But Erika kept on persisting. 2769 .. kalau di-tarik, if pt-pull 2770 .. pergi jauh go far 2771 berapa meter dari situ, how:many meter from there She was pulled, far away, a few meters from there.
This brief excerpt contains three instances of the di- patient-trigger prefix, and two instances of the N- agent-trigger prefix. All three of the patient-trigger prefixes (IUs 2764, 2766, and 2769) occur on the verb tarik ‘pull,’ and all take the patientive Maya as the subject. (The actual NP is ellipted in two of these instances, but it is clear in the overall context that Maya is the subject argument.) Both of the agent-trigger prefixes (IUs 2765 and 2768) occur on the verb otot ‘persist,’ indicating it is Erika, the subject of these clauses, who is being persistent in coercing Maya away from the towels and away from the influence of Gendam. The following example contains one more instance of Maya as the subject in a patienttrigger clause, as well as speaker L’s co-construction of the upshot of the narrative, namely that Erika had to take her outside of the boundaries of the Gendam.
102 Robert Englebretson
(21) Pencuri IU 2779–2782 2779 I: .. Udah di-ta= rik sampai marah-marah perfv pt-pull until angry-angry Maya-nya itu. Maya-def that I: (Maya) was pulled until Maya started getting angry. 2780 L: [Pokok-nya di luar batas, main-stm at outside limit L: [The important thing is, outside of the boundary. 2781 I: [Sampai itu ho-o. until that aff I: [To there, yeah. 2782 .. di luar batas tu], at outside limit prt Outside the boundary].
Again, Maya is presented in a patient-trigger clause in IU 2779 as being pulled. In the following example, Maya is once again presented as the subject argument in a patient-trigger clause, as Erika is addressing her: (22) Pencuri IU 2792–2795 2792 .. Langsung kamu ini gini-gini, direct 2sg this like:this-like:this “You just keep going like this.” 2793 pokok-nya di-omong-in sama Erika itu, main-stm pt-talk-app by Erika that Erika said to her. (lit., (she) was told by Erika.) 2794 Terus udah lewat batas berapa meter itu, next perfv pass limit how:many meter that Then, after they had passed the boundary of a few meters, 2795 .. Jadi sadari itu lho Maya-nya become conscious that prt Maya-def Maya became conscious.
In IU 2792, Indra is reporting the words that Erika allegedly said to Maya, encouraging her to keep coming with her. Because of the di- patient-trigger prefix on the verb omong ‘speak’ and the oblique agent phrase sama Erika ‘by Erika,’ this clause again construes Maya as the subject argument in a patient-trigger clause.
Stancetaking in colloquial Indonesian conversation 103
In IU 2794, Indra reports that they passed the boundaries of the Gendam, and IU 2795 provides the resolution of the narrative – when Maya jadi sadar ‘becomes conscious.’ This marks a shift in positioning, as Maya is no longer portrayed as the victim of Gendam, and this shift in positioning is also indicated by a shift in grammar, as Indra begins to construe Maya as the subject in agent-trigger clauses, as shown in the following example. (23) Pencuri IU 2796–2799 2796 (H) O ya ya, oh yes yes “Oh, yeah yeah.” 2797 ternyata jelek nge-lihat lagi to, in:reality bad at-see again prt In reality it was bad when she looked at it again. 2798 n-engok lagi ke itu-nya. at-look again to that-def She looked at it again, 2799 .. Ternyata handuk-nya itu jelek. in:reality towel-def that bad In reality the towel was bad.
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. Then, the verb lihat ‘see’ in IU 2797, and the verb tengok ‘look’ in IU 2798, are both marked as agent-trigger. This indicates that Maya is once again intentional and agentive, and is responsible for the actions of seeing and looking. And when Maya ‘sees’ and ‘looks,’ she recognizes that the towels were “in reality” bad, and not in fact what she had thought they were when she was under the influence of Gendam. In sum, the distribution of agent- and patient-trigger clauses in the unfolding 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. When Maya is under the influence of Gendam, Indra (the narrator) positions her as patientive and not responsible for her actions. Indra achieves this positioning, in part, through the use of patient-trigger clauses for Maya, in contrast to the agent-trigger clauses she uses with Erika, who is not under the influence of Gendam. Once Maya has come outside the boundaries of Gendam, Indra shifts her positioning through the use of agent-trigger clauses, implying that Maya is once again able to control her actions and may now be held accountable for them. Similar to Duranti’s (1990, 1994) analysis of erga-
104 Robert Englebretson
tivity in Samoan public discourse, in which grammatical relations (specifically ergative) allow speakers to assign blame or responsibility, the narrative analyzed here suggests that Indonesian voice may have similar social implications. The current section in no way seeks to discredit or minimize other, more traditional approaches to voice in Indonesian, which highlight semantic, syntactic, or discourse factors. Conceiving of voice as a means of positioning the self or others is not contradictory to approaching it from these other angles as well. However, placing central emphasis on the social nature of this phenomenon provides a potentially fruitful line of inquiry into how speakers use grammar in stancetaking, to actively construct their social worlds. As Duranti notes about agency, 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.g. is the agent of this event expressed and, 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 typically unconscious but nevertheless careful choice of words. (2004: 466)
In Indonesian, this “careful choice of words” must also be understood to encompass the choice of agent- or patient-trigger clause, as a grammatical resource which Indonesian speakers may use for specific stance purposes.
6.
Conclusion
I began this paper by noting the pervasiveness of stancetaking as an activity realized through language. I then tied this in with the functionalist view of a usagebased approach to grammar, and suggested that one goal of the descriptive grammarian should be to account for how speakers use the grammatical resources of a language to carry out the activity of stancetaking. Many English grammarians and linguists have been engaged in such an enterprise over the course of the past several decades, and it is not surprising, for instance, to find a chapter in the Longman Grammar of Spoken and Written English (Biber et al. 1999) entitled The grammatical marking of stance. However, interest in stancetaking and grammar has not seemed to have had much influence on field linguistics or descriptive grammars of other languages. As a first step toward such a goal, this paper has offered a discussion of three facets of Indonesian grammar that speakers use in constructing their social worlds through stancetaking. In Section 3, I addressed the complex system of Indonesian first-person-singular referring expressions. I argued that these forms are not simply a deterministic reflection of a priori social variables such as regional origin of the speaker
Stancetaking in colloquial Indonesian conversation 105
or formality level of the speech event, but rather, speakers actively use these various forms to construct and reflect locally-relevant aspects of social and personal identity such as distance, casualness, or toughness. This kind of self-expressive stancetaking enables speakers to highlight facets of the self, as relevant in the ongoing conversation. In Section 4, I took up the issue of the Indonesian -nya clitic and its participation in epistemic constructions. While this has traditionally been characterized as a marker of possession, definiteness, or nominalization, 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. The epistemic functions of -nya constructions fall into three categories: evidentiality (how a speaker knows about the current utterance), assessment of interactional relevance (how a speaker values the current utterance in light of the ongoing interaction), and affect (the speaker’s emotional attitude toward the proposition expressed in the current utterance). Section 5 explored the social nature of Indonesian agent-trigger and patienttrigger clauses. I argueed that grammatical voice, which has traditionally been characterized solely in terms of syntax, semantics, and discourse-pragmatics, 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. I illustrated this approach with reference to a specific conversational narrative, demonstrating how the speaker uses grammatical voice to attribute agency, responsibility, and the lack thereof to the protagonists in the narrative. I have focused on these three aspects of Indonesian grammar for several reasons. First, each is highly frequent and productive in colloquial Indonesian. Second, these three grammatical resources are formally quite different from each other: pronoun choice, a clitic, and voice constructions. Thirdly, each represents a particular kind of stancetaking: self-expressive stance for the 1SG reference forms, epistemic stance for the -nya constructions, and a broader-level stance in terms of moral positioning for the agent-trigger and patient-trigger clause types. Yet, these only begin to scratch the surface of how Indonesian grammar is involved in stancetaking. I have not dealt at all with clause-final particles (cf. Wouk 1998, 2001), nor with modals and adverbs, nor with stancetaking through conversational sequences and dialogic interaction. What I offer here is a first glimpse of the fruitful nature of consciously bringing stancetaking into the realm of grammatical description. By having put these three areas on the table, I hope to encourage other linguists to take seriously the question of how grammatical resources are functioning in the service of stancetaking in other languages. In this paper, I hope to have shown that when we approach grammar with this perspective in mind, forms which have previously only been characterized as part of a referen-
106 Robert Englebretson
tial, propositional system for expressing and managing information can be seen from a new perspective, a perspective that can offer fresh insight into the very motivations behind grammar itself.
Notes 1. I would like to thank the participants in the 10th Biennial Rice Linguistics symposium for comments and discussion on an early draft. The comments by the two anonymous reviewers were likewise valuable, and helped to clarify several details. I would also like to thank Susanna Cumming and Sandy Thompson, both of whom have provided invaluable encouragement and suggestions at various stages of this project. As with all my work on Indonesian, I am especially grateful to the individuals in Yogyakarta, who allowed themselves to be recorded for the colloquial Indonesian corpus, and especially to the three research assistants with whom I worked through the data. All shortcomings of this paper, and any misunderstandings of Indonesian language or culture are, of course, solely my fault and responsibility. 2. I wish to credit Jack Du Bois, in several unpublished conference talks, for presenting this maxim: ‘Every utterance enacts a stance.’ 3. I will not be discussing second- or third-person reference, nor the first-person-plural inclusive and exclusive pronouns. 4. A reviewer has raised the provocative question as to the role of non-self-referral in the enactment of stance. This is indeed an important issue, which the author is beginning to address in the context of a larger, cross-linguistic project on argument ellipsis, a discussion of which is far afield of the current paper. In sum, considering this a colloquial Indonesian “zero 1SG form” in opposition to the overt pronominal forms is problematic for at least two reasons. First, this view presupposes that arguments of a clause are normatively expressed – which may not in fact be true for colloquial Indonesian at all (cf. Ewing 2005). Secondly, argument ellipsis is widespread in this language, and is by no means restricted to 1SG: it occurs with all persons of pronouns, as well as with full noun-phrases. The use or ellipsis of arguments does indeed have social and interactional implications – but whatever these implications are, they crosscut all pronominal and full forms, and are not limited to self-reference or individual identity in the same way as the different 1SG pronominal forms seem to be. 5. Due to the uncontrolled nature of the data, a direct statistical correlation for each of the 1SG reference forms across all six speech events is not feasible. Because the speakers vary considerably among themselves in terms of the amount of talk, the topics, and their social backgrounds, the confounding variables are too numerous to allow reliable conclusions to be drawn. However, the categorical nature of the 1SG forms produced in the radio call-in show as compared with the other five speech events is nonetheless striking. 6. There are actually two plausible structural analyses of this clause. In the one intended by this example, duit gua is a possessive NP consisting of the head noun duit ‘money’ followed by the 1SG form gua as a possessor: ‘my money.’ A literal English translation of this clause would then be: ‘My money doesn’t exist.’ Another possible analysis of this clause would be as a type of so-called focus construction, where nggak ada duit is the predicate, followed by the subject gua ‘I.’ According to this interpretation, the pronoun gua ‘I’ would be the subject of the clause, and the overall focus construction has the pragmatic force of placing strong emphasis on the 1SG
Stancetaking in colloquial Indonesian conversation 107
subject “(As for me), I don’t have any money.” Both analyses are plausible, and, just as real-life utterances are often uncertain and multifaceted for interlocutors themselves in realtime conversation, I see no need to choose one or the other structural interpretation here; they both exist here simultaneously. For purposes of exposition, I am presenting this example as an instance of gua as the possessor in a possessive NP, although I recognize the other structural possibility as well. 7. See Englebretson (2003: Chap. 5) for further discussion of the syntactic characteristics and distribution of epistemic -nya constructions, and examples to support the analysis of these constructions as adverbials. 8. This example shows -nya marking third-person-plural possession, which is generally considered a nonstandard use of this clitic. Prescriptively, -nya is said to mark third-person-singular possession. 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. 9. I am aware that the terms subject and object are debatable and problematic in the context of Indonesian grammar. I am using these terms here solely for the sake of convenience, to facilitate the easy understanding of these examples by non-Indonesianist readers, and because the debate over Indonesian grammatical relations is not central or relevant for the overall point of this section. 10. In formal varieties of Indonesian, the preposition marking the oblique argument in a patient-trigger clause is oleh ‘by.’ The use of sama ‘with’ here is typical of colloquial Indonesian. 11. I am aware of the disfluencies and ambiguities in this IU, and the difficulty in glossing it in a more accurate way. I will refrain from commenting further on this here, however, so as not to distract from the discussion at hand.
References Barlow, M. and Kemmer, S. (eds.). 1999. Usage-Based Models of Language. Stanford: CSLI. Benwell, B. and Stokoe, E. 2006. Discourse and Identity. Edinburgh: Edinburgh University Press. Biber, D. and Finegan, E. 1988. “Adverbial stance types in English.” Discourse Processes 11 (1): 1–34. Biber, D. and Finegan, E. 1989. “Styles of stance in English: Lexical and grammatical marking of evidentiality and affect.” Text 9 (1): 93–124. Biber, D., Johansson, S., Leech, G., Conrad, S. and Finegan, E. 1999. The Longman Grammar of Spoken and Written English. London: Longman. Brown, R. and Gilman, A. 1960. “The pronouns of power and solidarity.” In Style in Language, T. Sebeok (ed.), 253–276. Cambridge, MA: MIT Press. Bucholtz, M. and Hall, K. 2005. “Identity and interaction: A sociocultural linguistic approach.” Discourse Studies 7 (4–5): 585–614. Bybee, J.L. and Hopper, P.J. (eds.). 2001. Frequency and the Emergence of Linguistic Structure. Amsterdam: John Benjamins. Bybee, J.L. 2006. “From usage to grammar: The mind’s response to repetition.” Language 82 (4): 711–733.
108 Robert Englebretson
Conrad, S. and Biber, D. 2000. “Adverbial marking of stance in speech and writing.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, S. Hunston and G. Thompson (eds.), 56–73. Oxford: Oxford University Press. Cumming, S. 1991. Functional Change: The Case of Malay Constituent Order. New York: Mouton de Gruyter. Cumming, S. and Ono, T. 1997. “Discourse and grammar.” In Discourse Studies: A Multidisciplinary Introduction, Vol. 1, T.A. van Dijk (ed.), 112–137. London: Sage. Cumming, S. and Wouk, F. 1987. “Is there ‘discourse ergativity’ in Austronesian languages?” In Studies in Ergativity, R.M.W. Dixon (ed.), 271–297. Amsterdam: North-Holland. Du Bois, J.W., Schuetze-Coburn, S., Paolino, D. and Cumming, S. 1993. “Outline of discourse transcription.” In Talking Data: Transcription and Coding Methods for Language Research, J.A. Edwards and M.D. Lampert (eds.), 45–89. Hillsdale, NJ: Lawrence Erlbaum. Duranti, A. 1990. “Politics and grammar: Agency in Samoan political discourse.” American Ethnologist 17 (4): 646–666. Duranti, A. 1994. From Grammar to Politics: Linguistic Anthropology in a Western Samoan Village. Berkeley: University of California Press. Duranti, A. 2004. “Agency in language.” In A Companion to Linguistic Anthropology, A. Duranti (ed.), 451–473. Malden, MA: Blackwell. Echols, J.M. and Shadily, H. 1989. An Indonesian-English Dictionary, 3rd ed. (revised and edited by John U. Wolff and James T. Collins in cooperation with Hassan Shadily). Ithaca: Cornell University Press. Englebretson, R. 2003. Searching for Structure: The Problem of Complementation in Colloquial Indonesian Conversation. Amsterdam: John Benjamins. Errington, J.J. 1988. Structure and Style in Javanese: A Semiotic View of Linguistic Etiquette. Philadelphia: University of Pennsylvania Press. Errington, J.J. 1998. Shifting Languages: Interaction and Identity in Javanese Indonesia. New York: Cambridge University Press. Ewing, M.C. 2005. “Colloquial Indonesian.” In The Austronesian Languages of Asia and Madagascar, A. Adelaar and N. Himmelmann (eds.), 227–258. London: Routledge. Ford, C.E., Fox, B.A. and Thompson, S.A. 2002. “Social interaction and grammar.” In The New Psychology of Language. Vol. 2, M. Tomasello (ed.), 119–143. Mahwah, NJ: Lawrence Erlbaum. Givón, T. 1979. On Understanding Grammar. New York: Academic Press. Harré, R. and van Langenhove, L. 1999. Positioning Theory: Moral Contexts of Intentional Action. Malden, MA: Blackwell. Hopper, P.J. 1987. “Emergent grammar.” BLS 13: 139–157. Hunston, S. and Sinclair, J. 2000. “A local grammar of evaluation.” In Evaluation in Text: Authorial Stance and The Construction of Discourse, S. Hunston and G. Thompson (eds.), 74–101. Oxford: Oxford University Press. Kamio, A. 1997. Territory of Information. Amsterdam: John Benjamins. Kärkkäinen, E. 2003. Epistemic Stance in English Conversation: A Description of Its Interactional Functions, with a Focus on ‘I think’. Amsterdam: John Benjamins. Kaswanti Purwo, B. 1988. “Voice in Indonesian: A discourse study.” In Passive and Voice, M. Shibatani (ed.), 195–241. Amsterdam: John Benjamins. Langacker, R.W. 1999. Grammar and Conceptualization. New York: Mouton de Gruyter. Mühlhäusler, P. and Harré, R. 1990. Pronouns and People: The Linguistic Construction of Social and Personal Identity. Cambridge, MA: Blackwell.
Stancetaking in colloquial Indonesian conversation 109
Ochs, E., Schegloff, E.A. and Thompson, S.A. (eds.). 1996. Interaction and Grammar. New York: Cambridge University Press. Ono, T. and Thompson, S.A. 1995. “What can conversation tell us about syntax?” In Alternative Linguistics: Descriptive and Theoretical Modes, P.W. Davis (ed.), 213–271. Amsterdam: John Benjamins. Poedjosoedarmo, S. 1982. Javanese Influence on Indonesian. Materials in Languages of Indonesia, No. 7. Series D. Canberra: Pacific Linguistics. Rampton, B. 2005. Crossing: Language and Ethnicity among Adolescents, 2nd ed. Northampton, MA: St. Jerome Pub. Sneddon, J.N. 1996. Indonesian: A Comprehensive Grammar. New York: Routledge. Sneddon, J.N. 2002. “Variation in informal Jakartan Indonesian: A quantitative study.” Paper presented at the Ninth International Conference on Austronesian Linguistics, Canberra, January 8–11. Stubbs, M. 1986. “‘A matter of prolonged field work’: Notes towards a modal grammar of English.” Applied Linguistics 7 (1): 1–25. Tannen, D. 1986. “Introducing constructed dialogue in Greek and American conversational and literary narrative.” In Direct and Indirect Speech, F. Coulmas (ed.), 311–332. New York: Mouton de Gruyter. Tannen, D. 1989. Talking Voices: Repetition, Dialogue, and Imagery in Conversational Discourse. New York: Cambridge University Press. Thompson, G. and Hunston, S. 2000. “Evaluation: An introduction.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, S. Hunston and G. Thompson (eds), 1–27. Oxford: Oxford University Press. Tomasello, M. (ed.). 1998. The New Psychology of Language: Cognitive and Functional Approaches to Language Structure, Vol. 1. Mahwa, NJ: Lawrence Erlbaum. Tomasello, M. (ed.). 2002. The New Psychology of Language: Cognitive and Functional Approaches to Language Structure, Vol. 2. Mahwa, NJ: Lawrence Erlbaum. Wouk, F. 1989. The Impact of Discourse on Grammar: Verb Morphology in Spoken Jakarta Indonesian. Ph.D. dissertation, Department of Linguistics, University of California, Los Angeles. Wouk, F. 1998. “Solidarity in Indonesian conversation: The discourse marker kan.” Multilingua 17 (4): 379–406. Wouk, F. 1999. “Dialect contact and koineization in Jakarta, Indonesia.” Language Sciences 21 (1): 61–86. Wouk, F. 2001. “Solidarity in Indonesian conversation: The discourse marker ya.” Journal of Pragmatics 33 (2): 171–191. Wu, R.-J. 2004. Stance in Talk: A Conversation Analysis of Mandarin Final Particles. Amsterdam: John Benjamins.
Appendix A: Summary of transcription conventions (Adapted from Du Bois et al. 1993) Each transcript line represents a single intonation unit. Speaker labels appear in uppercase, and are followed by a colon.
110 Robert Englebretson
Simultaneous speech is indicated by square brackets [ ] (not aligned because of glossing difficulty). Material in angle-brackets < > is codeswitching, usually Javanese, English, or Arabic. . , ? ‑‑ ‑ @ % = .. ... (TSK) (H)
Final intonation contour Continuing intonation contour Appeal intonation contour Truncated intonation unit Truncated word One pulse of laughter Glottal stop Prosodic lengthening Short pause Long pause Tongue click In-breath
Appendix B: Abbreviations and glosses 1sg, 2sg, 3sg 1pl, 3pl aff app at def ex mid neg nom nonvol perfv pos prog prt pt rel stm
first-, second-, and third-person singular first-person plural, third-person plural affirmative backchannel applicative suffix agent-trigger prefix definite existential middle voice negative morpheme nominalizer nonvolitional perfective aspect possessive progressive aspect discourse particle patient-trigger relativizer stance marker (epistemic -nya)
Subjective and intersubjective uses of generalizations in English conversations Joanne Scheibman
Old Dominion University
1.
Introduction1
1.1
Introduction to the study
Generality of meaning has been of interest to scholars working primarily in the areas of semantics, language acquisition, and language change. To date, however, generalization and generality of meaning in discourse have not received a great deal of attention, even though there is evidence that speakers’ uses of general meanings have discursive and pragmatic functions. For example, included with orientation and attitude, Berman (2005) lists generality of reference as one of three dimensions of discourse stance. Van Leeuwan (1995) suggests that generalization and underspecification of agency and social actions are important considerations in critical analyses of texts. And Zhang (1998) points out that general expressions rely on Grice’s maxims for their resolvability. These findings suggest that generality of expression also has influences on social and expressive activities in conversational discourse. The present study examines stance-related functions of generalizations in English conversations. These are expressions of position and attitude that are relevant both to individual speakers (their subjective uses) and to relational activities among participants (their intersubjective uses). These utterances index and evaluate classes in American English conversations. The expressions designate types – as opposed to tokens – of people, things, attitudes, reactions, activities, and relations. Additionally, in the conversations used for this study, generalizations based on these classes are not inductively constructed or construed – that is, they are not categories based on reported properties shared by multiple entities or ideas in the world. Instead participants tend to generalize (index classes of experience)
112 Joanne Scheibman
based on personal and social expectations and beliefs, and the use and interpretation of these utterances is keenly sensitive to interactive contexts. The specific types of generalizations examined here are utterances with grammatical subjects that refer to classes or groups (e.g., and all these teachers are coming in and saying; worse things could happen to the poor guy; the French are basically a northern people). Because construction and interpretation of meaning and reference in conversation occur interactively, participants’ classifications of people and phenomena for purposes of predication have expressive and social consequences. As analyses offered in this paper suggest, generalizations in these conversations are used by participants to evaluate, demonstrate solidarity with one another, and authorize opinions – activities which, in many cases, animate societal discourses. Moreover, because these utterances are naturally broadening, or inclusive, in the sense that they typically index commonly held beliefs, their uses in conversations highlight collaborative and interactive aspects of stancetaking.
1.2 Generality, generalizations, and stance In a description of a crosslinguistic project investigating the development of stance over time based on spoken and written narratives, Berman (2005: 107) characterizes discourse stance as having three related dimensions: orientation (relations between sender, text, and recipient), attitude (epistemic, deontic, affective), and generality (of reference and quantification). The category generality “concerns how relatively general or specific reference is to people, places, and times mentioned in the text,” and in these studies textual features are coded according to three values: personal and specific (e.g., I/my parents think, My/This boy’s father made me/him apologize), generic (e.g., People/We/You tend to think), and impersonal (e.g., It’s well known) (Berman 2005: 108). Berman operationalizes referential generality, then, as a continuum – with linguistic features at one end categorized as personal and intimate, and those at the other end viewed as impersonal and distant. The set of utterances defined as generalizations in the current study includes those in Berman’s generic group and also contains expressions with lexical nounphrase subjects that refer to general classes. The functions and properties of these generalizations also overlap with Berman’s orientation and attitude dimensions of discourse stance. For example, even though utterances exhibiting the formal properties of personal and specific expressions as described by Berman are not included in this study (e.g., those with specific agent subjects such as I), the conversational generalizations are “personalized” in a different sense: they convey subjective meanings (primarily evaluation) in interactional contexts. Indeed, one
Generalizations in English conversations 113
of the consistent findings of the analyses presented here is that in interactive contexts, the generalizations function evaluatively in their conversational contexts. Stance in discourse is a relational notion. Stancetaking is a reaction to, or a positioning relative to, something else – prototypically, a message or topic. 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). Speakers and groups of speakers, however, don’t only position themselves with respect to propositional material, they also situate themselves in relationship to one another, to expectation, and to sociocultural beliefs. 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.g., Biber and Finegan 1989; Berman 2005; Reilly et al. 2005), this paper is informed by the understanding that speakers also express stance by allying themselves with (or sometimes separating themselves from) one another. In these ways, stances can be collaboratively constructed and jointly held. One might argue that all discursive and social practices reflect stancetaking, since these activities presuppose a response to some linguistic, psychological, or social reference point (Langacker 1993).2 Depending on what and whom interactants position themselves relative to (e.g., utterances, objects, discourse participants, groups of people, beliefs), for analytical purposes stance might be classified as individual (construed as relevant to a speaker’s position in discourse), interactive (relevant to local discourse activities), or sociocultural (relevant to general beliefs of people as members of communities). In the organization of this study, aspects of individual stancetaking (evaluation, expressiveness) are framed as subjective, while interactive activities and the sharing of beliefs contributing to collaborative expression of evaluation and attitude are discussed as intersubjective. In the conversational extracts, however, the subjective and intersubjective functions of generalizations are typically overlapping or simultaneous. Generalizations have two related characteristics that contribute to their social and expressive force in English conversations: (1) they make reference to (and assess) general classes, (2) which gives them broadening, or inclusive, functions in discourse. The predicates occurring with general subjects in these utterances are often evaluative, a central type of subjective expression and stancetaking. It is also the case that construction of an indexical class by a speaker is an expressive act. Regarding the broadening properties of generalizations, because these utterances are referentially general, to varying degrees they are underspecified. Their unremarkable interpretation and use in conversations, however, suggest that participants elaborate the meanings of these expressions by making reference to shared knowledge and to interactive contexts, aspects of intersubjectivity in discourse. In these conversations, then, general assertions and evaluations are likely to be com-
114 Joanne Scheibman
mon ground among interactants (Clark 1996). Additionally, because of the inclusiveness of these utterances, generalizations can be used as expressions of solidarity, an intersubjective activity that contributes to the collaborative construction of stance in conversations.
1.3
An example of a conversational generalization
In the conversational extract in (1), 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. Since it is unlikely that the speaker has assessed the intelligence of all of the employees of the medical laboratory where she works, evoking this discursive class, and its interpretation by interactants, occur principally in relation to the speaker’s expressed stance toward her coworkers. (1)
I don’t know, the people I work with, sometimes are really stupid, and I-and I think I’m starting, to figure them out a little bit more,
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. In structuring conversational narratives, speakers may introduce people, places, and situations in general terms in order to orient participants to their stories (Labov and Waletzky 1967). It can also be the case that by expanding a reference class subsumed by an evaluation using a generalization, a speaker broadens the domain of her assertion, thus implicitly augmenting its evidential weight or appeal in the interaction (see, in particular, example 8). Furthermore, the evaluative stances expressed by speakers using these utterances with general subjects tend to reflect meanings and expectations that are presupposed by other participants in the conversations. For example, in (1), the indexical category denoted by the people I work with is a common one in mainstream American English-speaking communities, and as such, it invites participants to fill in and elaborate the meaning of this expression to include, for example, 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 evaluation of the group in the generalization in (1). It is not unusual, then, in these conversations for a speaker’s stance toward a generally construed class (of people, relations, etc.) to reflect evaluations that are shared by other discourse partici-
Generalizations in English conversations 115
pants. This indexing of jointly held beliefs by these general utterances has consequences for both interpretation (e.g., enriching of meanings of these expressions for participants) and interaction (e.g., demonstrating solidarity by acknowledging shared stance and attitude). What is paradoxical about generalizations found in the conversational data is that because these utterances make reference to general classes of phenomena, there is a sense in which they can be viewed by interactants as propositionally powerful. That is, utterances containing general subjects (e.g., the best things are German) are often treated as matter-of-fact descriptions of situations in the world. In spite of this appearance, 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 context and larger culture. In conversational contexts, then, 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 societal discourses. These two characterizations of generalizing utterances – their local, subjective construction and their intersubjective treatment as statements of norms – inform the analyses offered below. The next section surveys popular and scholarly uses of the term generalization itself, leading to a working characterization of the concept to be applied to generalizations in the conversational data. Section 3 describes and illustrates the formal properties of generalizations in the database used for the study. Sections 4 and 5 present analyses of generalizations from a conversational corpus. Section 4 discusses generalizations that function as shared evaluations, and Section 5 illustrates two ways in which generalizations can augment speaker stance in interactive discourse. And finally, Section 6 presents summaries and conclusions.
2.
Generalizations
2.1
Popular and scholarly characterizations of generalization
The meaning of generalization varies by context and discipline. Operationalizing the term based on its multiple uses has been one of the most challenging aspects of this project. Furthermore, because in Western cultural discourses the making of general statements about the world is an activity that can be elevated or deprecated, generalization itself is subject to social valuation. In some cases, generalization is viewed negatively. In high school and university writing classes in the United States, for example, students are routinely directed to provide specific examples to support their generalizations, which might be defined as ‘unsubstan-
116 Joanne Scheibman
tiated opinions.’ In contemporary public discourse in the United States and elsewhere, generalizations over the activities and attributes of (usually marginalized) groups of people are also viewed negatively, and in some contexts, warrant the label bigotry. On the other hand, positive and authoritative treatment of generalizations is apparent in Western academic and scientific disciplines in which generalizations have an influential role in construction of theories (e.g., in statements of principles, Kafura 1998: 26). The generalizations that have value in Western cultures tend to be those that are overtly inductive, that is, those conclusions that have explicit links to specific facts or cases. In contrast, general statements based on implicit or insufficient data tend to be viewed unfavorably (e.g., stereotypes). Processes of generalization figure prominently in linguistics research. In functional and cognitive usage-based approaches to morphosyntax, for example, analysts discuss technical aspects of the formation, change, and acquisition of grammar as generalizations over specific usages. Ronald Langacker characterizes grammatical patterns as “schematic symbolic units, which differ from other symbolic structures not in kind, but only in degree of specificity” (1987: 58, bolding in original). In studies of language change, individual grammaticizing elements have been analyzed as undergoing processes of generalization of meaning (reduction in semantic specificity) accompanied by generalization (expansion) of the item’s contexts of use (e.g., Bybee and Pagliuca 1985; Bybee et al. 1994). And in language acquisition research, children are assumed to be learning generalizations (abstract patterns of form-meaning correspondences) when acquiring the semantics and grammar of their first language, such as those contributing to the system of argument structure in English (Goldberg et al. 2004).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, with a concomitant widening of applicability (e.g., of context or process), which contributes to the development or change of linguistic patterns. Clearly the generalizations presented in this survey are of different types and have distinct goals. The generalizations of scientific theorists are analytical conclusions based on particular types of data or texts. The generalizations illustrated for linguistics, on the other hand, tend to be descriptions of cognitive processes relevant to language use and change. What these different treatments of generalization share, however, is that they refer to the construction or highlighting of a class (e.g., a theoretical principle, a grammatical process, an opinion) (Bouacha 1995), 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.g., weakening of lexical meaning of grammaticizing elements; the prominence and longevity
Generalizations in English conversations 117
of scientific principles over time relative to the individual data analyses that support them).4 Generalization, then, is a type of categorization. In contrast to treatments of categorization, which tend to highlight membership criteria and relationships among category members, the process of generalization focuses on the formation of the class, or category, itself. Analogous to the way membership in classical categories has been delineated based on shared properties of members (see Lakoff 1987, e.g., for a critique of this view), formal treatments of generalization, such as those used in software and database design, define the term in a similar way. For example, Kafura states that “[g]eneralization identifies commonalities among a set of entities. The commonality may be of attributes, behavior, or both” (1998: 26). However, as illustrated in this section, not all processes of generalization designate classes that are based on shared properties of multiple entities (e.g., the maligned generalizations of the composition student). In academic argumentation and theory-building, the authority of generalizations is explicitly tied to common findings of individual studies and analyses. But in non-scientific registers and genres (e.g., conversational interaction), the common details that inform the construction of speakers’ generalizations are often inexplicit and covert.
2.2 Generality of expression in English conversations Expression of general meaning is ubiquitous in English conversations. While it is the case that conversational interactants can (and do) engage with one another at various levels of specificity, in previous analyses of English conversation, I found that a relatively small percentage of tokens in a conversational corpus could be described as referentially specific, despite the attention such expressions enjoy in linguistic theories and practices (Hopper 1997; Scheibman 2002; Silverstein 1976). These more specific utterances are most frequently expressions with third-personsingular subjects with human referents occurring with past-tense material predicates (e.g., and he did a blood test) or oblique predicates (e.g., my test was Friday). They tend to cluster in descriptive narratives and in settings in which participants are involved in activities in which objects are focal (e.g., opening birthday presents). More frequent in these data are utterances that convey general meanings. Because generalizations are common linguistic practices in discourse, they are socially situated. For example, variation in generality and specificity in the representation of social groups and actions in discourse can be ideologically consequential. Van Leeuwan suggests that generalization is “an important issue in critical discourse analysis, as texts which are mainly concerned with legitimizing or de-legitimizing actions and reactions tend to move high up on the generaliza-
118 Joanne Scheibman
tion scale, including only the names of episodes, or of whole social practices” (1995: 99). With respect to generalizations in English conversations, the experiences, beliefs, and attitudes that interactants generalize over in collaborative contexts are both sensitive to and influence interaction. In contrast to formal or academic generalizations, the classes indexed by conversational generalizations might be said to be “empty” categories in the sense that membership is often implicit and covert. The data presented here suggest that speakers use these generalizations to evaluate and strengthen stance; and to create intersubjective ties both by generalizing experience and attitude and by ratifying others’ points of view while mutually adhering to societal discourses.
3.
Data
3.1
Generality of meaning, generics, and generalizations
Van Leeuwen (1995) notes that determining the linguistic properties that contribute to generality of expression can be challenging. Because texts typically do not display multiple representations of the same actions and events at different levels of specificity, systematic semantic comparison is difficult. Based on crosslinguistic studies on the development of stance, Berman points out that the same text “may be both specific and general in reference to persons, places, and times” (2005: 109). Because of this formal and functional heterogeneity (illustrated below for English), for this study it was necessary to select for analysis an utterance type that consistently conveyed generality of meaning in the conversational data. Generalizations, as delimited here, are expressions that lack specific reference (e.g., to time, place, event) and the formal presence of a speaking subject – both of these being properties of generics as well (Bouacha 1995). Lyons (1977: 193–194) states that generic propositions are tenseless, timeless, and aspectless, whereas general referring expressions can occur in sentences that express time-bound propositions. Lyons (1977: 194) also notes that the status of “generic, as distinct from general, reference” is philosophically controversial. Bouacha agrees that it is not easy to distinguish generalizations from generics, suggesting that the clearest difference is that strict generics (e.g., Beavers build dams) are primarily found in scholarly metalanguage, whereas generalizations (e.g., Cauliflower does well here) are quite common in naturally-occurring discourse (1995: 47). The analyses in the present study do not take into account differences between general and generic expressions. Rather, all subject noun phrases in the database that designate classes, including the so-called generic pronouns you and they, are classified as general subjects (Section 3.3).
Generalizations in English conversations 119
English speakers generalize about people, events, time, place, reactions, and attitudes. Therefore, a variety of grammatical and lexical constructions can express generality of meaning. For example, in conversation, plural noun phrases regularly denote classes (e.g., the prices are so much lower there than here; he threatened little kids). Adverbials of time (e.g., so you go and you practice on the weekends), manner (e.g., they come in with that attitude), and place (e.g., growing up around here you would know better) also index classes of expectation and experience. General construal and expression of events can be expressed aspectually, for example, with progressive constructions (e.g., all the teachers were telling me) or verbs in the habitual present (e.g., they play salsa?). Because of this structural and semantic heterogeneity of generalizing types in English, analyses in this paper are restricted to conversational utterances with grammatical subjects that designate classes (Section 3.3). As it turns out, expressions with general subjects tend also to contain generalizing predicates (Section 3.4).
3.2 Sources of data The database of utterances from which the generalizations were drawn consists of portions of 10 conversations, totaling approximately 90 minutes. Five of the conversations are from the Santa Barbara Corpus of Spoken American English (Du Bois et al. 2000; Du Bois et al. 2003), and five are unpublished transcripts.5 There are 153 utterances with general subjects in the corpus representing the language of 19 adult speakers of American English. Utterances were tagged for a variety of structural and functional properties for a previous study (e.g., tense, verb type, subject animacy, and referentiality) (Scheibman 2002). A coding field was added specifying the general class indexed by the grammatical subject, as illustrated in (2). (2) the English are just the same though,6 (class: nationality) your people .. tortured him to death, (class: religious group) everybody in the world has those problems (class: hyperbolic subject)
120 Joanne Scheibman
3.3 General subjects As stated above, the generalizations examined in this study are utterances with grammatical subjects that designate classes. While it is the case that speakers evoke classes in object or oblique positions (e.g., He knows Cats;7 I’m not dancing with guys), English subjects have interactive and cognitive importance, which makes utterances with general subjects a useful choice for these analyses. Chafe (1994) suggests that English subjects are often starting points for additional information (MacWhinney 1977), and Langacker (1993) points out that topics (and subjects) can serve as reference points for further talk. In this study, classes indexed by subject noun phrases are evaluated and elaborated in accompanying predicates, resulting in full utterances that convey generality of meaning. Table 1 summarizes the generalizations in the database by subject type. The three major subjects that refer to general classes are generic they and you forms and lexical noun phrases (plural and singular).8 The classes referred to by lexical noun-phrase subjects in the database (and anaphoric they subjects as well) are diverse and vary in their conventionality. They include, for example, concrete items and groups of people expressed with plural nouns (e.g., vacuum cleaner bags, the women), institutional entities (e.g., the press), and hyperbolic subjects (e.g., everybody in the world). Generic you utterances tend to universalize experience in conversations (e.g., you don’t admit you’re a feminist), and generic they is often used in these conversations to characterize people or institutions construed as outsiders (e.g., you can’t really tell when they blush). Table 1. Generalizations by subject type (n = 153) Subject type
Examples
they
74 (48%)
you
31 (20%)
Lexical noun phrase 30 (20%) (plural) Lexical noun phrase 16 (11%) (singular) one 2 (1%)
so they don’t know what the hell they’re doing and in high school you know they teach about it you can’t kill peas growing up around here you would know better the ballroom people don’t do it that way those guys’ll be all over you the ceviche’s full of vegetables the press will pick it up and smear you for it one is more tolerant of feminist views one’s certainly more sensitive to them
Generalizations in English conversations 121
3.4 Predicates occurring with general subjects The formal properties of predicates occurring in utterances with general subjects also convey general meanings. Table 2 displays the generalizations by tense and modality. Notably, 71 percent of these predicates (109/153) are present tense, while only 12 percent are past tense. Additionally, 82 percent of the present-tense utterances (89/109) are simple present, expressing stative and habitual meanings. The examples in Table 2 illustrate how these predicates express processes and states as non-specific. Seventeen percent of the predicates occurring with general subjects (26/153) contain a modal auxiliary. Half of these occur with generic you subjects, and the most frequent modals in these utterances are can and can’t (69% or 9 out of 13 tokens). Many of these utterances are normative statements. In contrast to the high frequency of present tense in these utterances, only 12 percent (18/153) of the predicates occurring with general subjects in the database are past tense. This means that there are relatively few predicates in the group that might designate referentially specific events (e.g., I talked to a guy that’s thirty-four in my class,). As is the case, however, the past-tense predicates occurring with general subjects tend to reflect construal of events as summaries, as opposed to individual accounts. For example, in (3), line 1, Sharon uses a general subject and a past-progressive predicate to present a series of similar events. In the utterance, All the teachers were telling me, the speaker is not emphasizing individual tellings of specific teachers at different times. Rather, 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), and not to the details of each encounter. This analysis is supported by repetition of the teachers’ advice in the quoted material in lines 3 and 4. (3)
(Raging Bureaucracy SBC0004 670.65-674.41)9 1 → SHARON: All the [teachers were] telling me,10 2 SHANE: [Could be worse]. 3 SHARON: You’re too nice, 4 you’re too nice, 5 your kids are gonna take advantage of you
122 Joanne Scheibman
Table 2. Generalizations by tense and modality (n = 153) Examples Present
109 (71%)
Modal
26 (17%)
Past
18 (12%)
people have the information they need (stative) everything breaks on my kitchen floor (habitual) you can go up there you can’t constantly be torn all the teachers were telling me they were pull- putting out people’s eyes
3.5 Summary Because of the empirical challenges relevant to determining correspondences between linguistic structure and generality of meaning in English conversational utterances, a decision was made to focus on tokens containing general subjects. The distribution of these utterances suggests that the predicates co-occurring with general subjects exhibit properties that also express generality of meaning, such as higher frequency of present-tense stative and imperfective meanings, and lower frequency of past-tense predicates. The rest of this paper presents subjective and intersubjective functions of generalizations in the conversational data. Section 4 discusses generalizations that function as evaluations in the conversations, and Section 5 illustrates utterance types that are used to strengthen speakers’ claims.
4.
Generalizations as shared evaluations
4.1
Subjectivity, evaluation, 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.g., mental states, affect, evaluation, preference). Subjective expressions are those which, to varying degrees, are grammatically, lexically, or pragmatically anchored to the speaker. Benveniste observes 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). Verhagen (1995: 116) also suggests that subjective expression is the most usual mode of language use, and that even lexical items like expensive and tall are not solely informative. Relevant to the current study, research based on conversational data has noted a variety of associations between linguis-
Generalizations in English conversations 123
tic structure and expression of point of view (e.g., Kärkkäinen 2003; Scheibman 2000, 2002; Tao 2001; Thompson and Mulac 1991). A common type of linguistic subjectivity is evaluation, or the expression of attitude. Thompson and Hunston (2000) observe that evaluation is conveyed both lexically (e.g., adjectives such as ugly) and grammatically (e.g., the use of the progressive in some contexts). Not only are evaluations expressive components of speakers’ activities, they also play a role in participant interaction. For example, in conveying opinions, evaluations invariably reproduce cultural norms and values: “Every act of evaluation expresses a communal value-system, and every act of evaluation goes toward building up that value-system” (Thompson and Hunston 2000: 6). The generalizations illustrated in this section are evaluations that fulfill both subjective and intersubjective functions, as characterized by Thompson and Hunston.
4.2 Construction of an evaluative class Example (4) is part of a longer conversation among four women friends. Preceding this episode, one of the participants, Z, a 32-year-old woman, was complaining to the group that she keeps finding gray hairs, and because they bother her, she pulls them out whenever she sees them. One of the participants, N, mocks Z about her worry, characterizing her as wanting to remain a teenager (line 1). The utterance in line 9, the whole thing is pretty ghastly, is the evaluative generalization in focus. It is offered by participant O in response to Z’s narrative of her gray hair dilemma. (4)
1 N: 2 Z: 3 4 5 S: 6 O: 7 S: 8 Z: 9 → O: 10 11 12 Z: 13 14 N: 15 O:
she wants to be a teenager. no I don’t want to be a teenager, but, ...(1.5) no it’s, well? the whole thing, you do it the way you do. (CLEARS THROAT) the whole thing is pretty ghastly. …(1.4) let’s face it that, you know there isn’t much alternative. no, I guess not. die young. ] 17 Z: [yeah really]. 18 yeah. 19 N: boy, 20 that sounds good.
As noted in Section 3.3, general classes referred to with lexical noun phrases in the conversational data vary in their conventionality. For example, highly institutionalized classes are referential groups such as nationalities (e.g., brazilians), professions (e.g., substitute teachers), or common objects and phenomena (e.g., ceviche). Less conventional are generalizations based on groups that are indexed locally in discourse whose interpretation depends, not only on lexical specification, but also on participants’ sociocultural expectations and knowledge of discourse proceedings (e.g., parents of elementary school students in the speaker’s class, guys in their twenties in the Bay Area). An example of a highly indexical construction is illustrated by the generalization in (4), the whole thing is pretty ghastly. The lexical noun phrase, the whole thing, is propositionally empty (the head noun is thing) and hyperbolic (the whole thing), giving the utterance strong expressive force, especially when combined with the evaluative complement, pretty ghastly. The referent of the noun-phrase subject, the whole thing, might be glossed as ‘the experience of aging,’ and included in the meaning are shared attitudes toward getting older. Even though the phrase itself has little explicit content, understanding of this indexical subject by the interactants is apparent, as demonstrated by subsequent ratification and continuation of the topic in lines 12–20. It is also the case that O’s generalization has interpersonal function in the conversation. The utterance comes on the heels of Z’s defense of herself from N’s teenager accusation (lines 2–4), allowing O to show solidarity with Z at this point in the conversation (a positive politeness strategy). In summary, the whole thing is a general construct based on Z’s gray-hair story and others like it in the culture. It linguistically indexes a popular attitude toward aging shared by the members of this group (that it is undesirable but inevitable). Not only does the generalization express O’s own stance, it also functions inclusively, or intersubjectively, in the interaction in two ways. O’s statement supports Z’s gray-hair narrative in the interaction by meeting it with her own negative evaluation of aging. And at a more global level, O’s generalization expresses a shared cultural norm. This attitude toward aging is common ground among the participants (Clark 1996), and one that unifies their individual stances.
Generalizations in English conversations 125
4.3 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, and they can be sites for the sharing of cultural beliefs. These two functions are aspects of intersubjectivity. For Benveniste (1971) not only is language inherently subjective, it is also intersubjective. He writes that “every utterance assum[es] a speaker and a hearer, and in the speaker, the intention of influencing the other in some way” (Benveniste 1971: 209). In fact, Benveniste suggests that the process of communication itself is “only a mere pragmatic consequence” of the dialogic relationship between ‘I’ and ‘you’ (1971: 225). 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, for example, in honorification”. 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.g., hedges, politeness markers). Indeed, despite the analytical convenience of labeling linguistic activities as either subjective or intersubjective, as demonstrated in subsequent analyses in this paper as well as in other chapters in this volume (especially those by Du Bois, Kärkkäinen, Keisanen, and Haddington), in situations of language use expressions of speakers’ evaluations and perceptions are fundamentally sensitive to the interactional exigencies of the context. Schiffrin (1990) suggests that intersubjectivity is interactively constructed among participants (e.g., in negotiating turns and topics), and she notes that it plays a role in the sharing of background knowledge in conversational interaction – activities that make reference to norms and expectations. In a study of evidential expression based on spoken and written Dutch and German, Nuyts (2001) uses intersubjectivity in a similar way – to refer to people’s sharing of beliefs and conclusions with the speaker. In these treatments, intersubjectivity refers to sharing of sociocultural expectations. One characterization of intersubjectivity in discourse, then, focuses on the local activities of participants in interactive contexts, and another refers to the sharing of beliefs and attitudes among participants as members of communities. The former might be characterized as interpersonally, or interactively, intersubjective; the latter makes reference to more global sharing of expectations. In a study of the cultural standing of opinions in discourse, Strauss (2004) calls these interpersonal elements of the context, social (e.g., politeness) and the beliefs and values components, cultural. In conversation, however, both of these intersubjectivities are accomplished simultaneously. Generalizations can be interpersonally inclusive
126 Joanne Scheibman
in that they demonstrate participants’ solidarity or empathy with one another’s stance. At the same time, because generalizations often make reference to societal norms, they are involved in the reproduction of belief systems.
4.4 Joint construction of stance using generalizations In the conversation in (5), three participants are discussing a coworker of speaker B, who is a woman named Vanda. Specifically the group is trying to figure out why Vanda is conservative and opinionated, which is how she was characterized previously in the discourse. The participants agree that Vanda’s having lived in Vienna until she was 16 could have contributed to her conservative outlook (Vienna is awfully conservative). Even so, 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). 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), and the generalization offered by speaker A in lines 6 and 8 (some people are more open to other things than others). (5)
1 B: .. she’d been exposed to enough that I would’ve -2 → I would think [that a person would be more], 3 C: [you mean more cosmopolitan]. 4 B: yeah. 5 A: but that’s probably her personality, 6 → some people are more open to, 7 C: X 8 A: other things than= .. than others.
The evaluative generalization in lines 2–3 is collaboratively produced. The joint production of this generalization is accomplished in two ways: the turns of the two participants are overlapping, and speaker C completes speaker B’s complement (Ono and Thompson 1995). The collaboration of interactants in articulating this generalization, along with speaker B’s ratification of C’s completion of her utterance (line 4), demonstrate the participants’ local construction of a shared attitude in the conversation. On the other hand, Speaker A’s generalization in lines 6 and 8 is a contribution to the group’s explanation for Vanda’s attitudes. Both of the generalizations in this episode (lines 2–3 and lines 6 and 8) contain generic subjects (a person, some people) co-occurring with relational predicates. These subject types permit global assessments of people and their behaviors. The
Generalizations in English conversations 127
predicate-adjective constructions found in both generalizations are evaluative, in the sense that notions of what it means to be more cosmopolitan or more open are subjectively and intersubjectively construed. The shared belief informing the generalization in lines 2–3 is the expectation that when one lives in different countries one acquires a certain sophistication. 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. These generalizations are offered by participants as they collaboratively construct an explanation for Vanda’s character – a person unknown to two of the participants in the conversation. In this way, what began as speaker B’s subjective stance becomes shared and elaborated by the group.
4.5 Evaluating others with they The evaluations discussed here are utterances with non-personal, or generic, they subjects. Because third-person pronouns make reference to non-speech-act participants (Quirk et al. 1985: 3354), they is often used by speakers of English and other languages to exclude, or other, social groups both in public discourse and in conversational interaction (e.g., Hongladarom 2002 for Thai; Skarżyńska 2002 for Polish). In English conversations, they is commonly used to evaluate groups that participants do not belong to, and often these assessments express disapproval or derision. The use of non-personal they is so conventional in English conversations (e.g., to refer to institutional entities) that it can occur without its having been referentially introduced (Biber et al. 1999: 331). For example in (6), the referent of they was not previously established in the discourse. In this humorous conversation, they refers to an unnamed but culturally shared conception of an unspecified institutional entity who is being ridiculed by the participants for engaging in genetic engineering of chickens for profit. (6) (Conceptual Pesticides SBC0003 295.15-305.58) 1 ROY: they’re trying to breed like a forty foot long tube chicken? 2 MARILYN: No= @@. 3 You mean so that they can go right from that to Chicken McNuggets?
The evaluative generalization in the next example (7, line 2) is an utterance containing a generic they subject occurring with a habitual predicate (they come in with that attitude). The general class designated by non-personal they in this utterance is substitute teachers. The antecedent for the pronoun occurred a minute before this point in the conversation when one of the participants, Sharon, tells how a sub
128 Joanne Scheibman
at the public school where she works lost a student’s free-lunch application, which prevented the child from receiving her lunches. Subsequently, the participants begin to condemn substitute teachers as a group for their general incompetence (they don’t care; they don’t give a shit about anything). One of the participants, Carolyn, initiates a second diatribe about beginning substitute teachers, who, according to her, are more enthusiastic than the experienced group (too enthusiastic, in fact), but incompetent nonetheless. In the excerpt in (7), Carolyn acts out her evaluation of a too-eager substitute teacher walking into a classroom. (7)
(Raging Bureaucracy SBC0004 379.05-385.13) 1 CAROLYN: You know, 2 → they come in with that .. attitude, 3 and they go, 4 ((THUMP)) ?
substitute teachers is a relatively conventionalized class in mainstream American English-speaking cultures. Even so, as it is referred to in this conversation, the class is not simply a referential set of teachers who cover classes on a temporary basis when “real” teachers are ill or on leave. Rather, important aspects of the meaning of this class depend on shared stance. That is, they here indexes a class of people with social and institutional relevance who are viewed derisively by the conversational participants. The simple form of come in the generalization in line 2 expresses habitual meaning that generalizes over supposed specific instances, thereby intensifying the speaker’s expression by portraying the behavior she’s satirizing as repetitive. This othering of substitute teachers continues after this point in the conversation with other participants laughing and joining in the performance. In this example, evaluation and elaboration of meaning is accomplished by all of the participants who share the cultural stance that views substitute teachers as other, creating solidarity among themselves.
4.6 Summary The examples in this section illustrate situations in which participants’ evaluative stances expressed with generalizations become jointly held in the conversations. It is notable that the subjective and intersubjective uses of these utterances are not distinct from one another. For example, while speaker O’s construction and evaluation of a highly indexical class in (4) (the whole thing is pretty ghastly) reflects
Generalizations in English conversations 129
her individual stance, the use of this generalization simultaneously demonstrates intersubjective functions. It is offered to show solidarity with speaker Z’s evaluation of aging, and at a more global level of intersubjective activity, the indexicality of the subject noun phrase co-occurring with an evaluative predicate expresses a common cultural view that invites, or perhaps marks, shared interpretation by the participants. In the excerpt in (5), on the other hand, the relationship between individual and interactive stance is illustrated by the linguistic co-construction of a general evaluation based on shared beliefs, motivated by positive politeness (e.g., high involvement, Tannen 1984). And finally, generalizations with non-personal they subjects, illustrated in (6) and (7), are used to mock institutional groups who are viewed as outsiders. In both of these extracts, speakers offer satirical generalizations that evoke laughter from the other participants, a sign of recognition as well as amusement (Coates 1996), which suggests that interactants share the speaker’s evaluation.
5.
Augmenting stance with generalizations
5.1
Propositional and interactive broadening
As discussed in the introduction, generalizations exhibit broadening properties. This broadening can apply to propositional material or to participant interaction, and both of these aspects can be used to enhance speakers’ stances in conversations. Propositionally speaking, broadening the domain of applicability of a scientific principle, for example, typically serves to make that principle theoretically more powerful. Similarly, expanding the reach of an assertion in conversation using a generalization can potentially increase the expressive power or authority of that assertion. If a speaker’s claim can be considered true for a larger class of entities or situations, then there is an assumption that the claim is stronger than it would be, for example, if the assertion were based only on one or perhaps two specific instances. In this way, generalizations can strengthen a participant’s stance by appearing to expand the reference class on which a particular claim is based. Generalizations can also be interactively broadening in the sense that participants tend to share beliefs expressed with these utterances (e.g., the whole thing is pretty ghastly). As illustrated in (9) below, a speaker can explicitly invite other participants to agree with her generalization by using a generic you pronoun. In this way, speakers enhance their positions in conversations by appealing directly to common experiences and beliefs of other participants. The two generalization types discussed in this section augment speakers’ stances in these two aforementioned ways. The analysis of the generalization in
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 contrast, the excerpt in (9) illustrates generalizations with generic you subjects, which are used by the speaker to strengthen her stance by appealing to the experiences and beliefs of the other participants. The former is a case in which the broadening properties of the generalization are applied to the propositional content of the speaker’s argument, whereas the latter generalizations are interactively expansive in the sense that they invite corroboration from conversational participants.
5.2 Use of a general class to strengthen a claim Example (8) is part of a longer conversation in which the speaker, Miles, is incredulous 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. 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. The full generalization is in lines 4–5 (I mean these guys say these women don’t care if they use rubbers or not), and the antecedent for these guys appears in lines 1–2 (guys in their twenties). (8)
(Lambada SBC0002 696.49-70594) 1 and just talking with guy=s, 2 ... in their twenties, 3 why, 4 → I mean these guy=s, 5 say these women don’t care if they use rubbers or not. 6 this is what guy=s have been telling me,
Miles’s discursive construction of the demographic, men in their twenties, potentially strengthens his stance in this conversation by broadening the reference class on which his claim is based. At the same time, 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, I mean these guys say these women don’t care if they use rubbers or not (Fox 2001; Hill and Irvine 1993). In the language of Goffman’s (1981) participation framework, Miles is animating a message that is authored by these guys about these women, 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 generalization. In fact, they provide supportive uptake, such as right, yeah, stupid, and
Generalizations in English conversations 131
participate with the speaker in elaborating his claim.11 Subsequent to the excerpt reproduced in (8), Miles provides evidence for his generalization. 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. Another participant, Jamie, adds her support by mentioning that her sisters don’t use condoms either. In this interaction, then, 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, specified, and jointly maintained by the participants.
5.3 Appealing to other participants Unlike lexical noun-phrase and they subjects, which index third-person classes and groups, generic you pronouns formally make reference to speech-act participants. On the one hand, this pronoun is used by speakers as an informal version of the generic pronoun one. On the other, it seems to retain second-person reference in that “the speaker is appealing to the hearer’s experience of life in general, or else of some specific situation” (Quirk et al. 1985: 354). Built into the speaker’s generalization using this form, then, is implicit reference to an addressee. The example 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. The episode in (9) occurs subsequent to a narrative by the primary speaker, Joanne, in which she described her brother’s down-and-out life before he kicked a disruptive drug habit. The generalizations in focus are those in Joanne’s final turn (lines 14–20), But you have to at one point let go. You can’t constantly be torn, just torn to pieces by, you know, somebody like that. In the opening line in this excerpt, I mean the guy is great, Joanne is referring to her brother as he is now since his successful recovery. After this utterance, Lenore and Ken contribute turns (lines 2–3, 7–8, 13) overlapping with Joanne’s utterances (lines 1, 4–6, 12) in which they remind her that they had tried to reassure Joanne during the time of her brother’s difficulties that he would be fine (e.g., I told you, see). (9)
(Deadly Diseases SBC0015 835.590-850.745) 1 JOANNE: I mean [the guy is] grea=t. 2 LENORE: [I told you]. 3 [[I tol=d you=]] 4 JOANNE: [[Isn’t that-5 he’s not]] hopeless.
132 Joanne Scheibman
6 [Well I had given] him up [for dead]. 7 LENORE: [I told you]. 8 KEN: [I] always tol=d her, 9 JOANNE: [Oh=, 10 KEN: [it’s gonna, 11 and she’s]-12 JOANNE: I’d given him up] [for dea-]-13 LENORE: [see]. 14 JOANNE: But you have to -15 you have to at one point let go. 16 you can’t constantly be torn, 17 just to=rn to pie=ces, 18 by, 19 you know, 20 somebody like that. 21 LENORE: (TSK) So your mother’s happy now.
Notice that in line 14 Joanne switches from using I to you. Based on spoken and written narratives, Abney (1996) finds that these shifts can be used as distancing techniques (e.g., from uncomfortable topics). 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. These reminders mark a contrast between Joanne’s friends’ expressions of confidence in her brother’s situation and her own lack of confidence, expressed in lines 6 and 12 (I’d given him up for dead). Joanne’s shift to generic you, then, may be viewed as an iconic distancing from a situation that might hold conflict for her. Another use of shifts to generic you is to generalize the speaker’s experience in order to build empathy with other participants. Joanne’s generic you utterances may justify her previous distancing from her brother by generalizing her reactions for Lenore and Ken. In generalizing her individual stance, she implicitly appeals to societal norms, and by extension, to the beliefs of other participants. This illustrates the broadening quality of the generalizations. Specifically, the speaker’s generalizations, But you have to at one point let go. you can’t constantly be torn, just torn to pieces by, you know, somebody like that, make tacit reference to discourses of drug and alcohol abuse that advise against codependency. Notice, too, that these three you utterances all contain the modal elements can’t and have to, which contribute to their having normative function. And finally, these general statements expressed with you generalizations in the conversational data often prompt speaker changes (illustrated in example (9), line 21). This suggests that generic you utterances fulfill the ritualized intersubjec-
Generalizations in English conversations 133
tive function of codas in conversational narratives – expressions conveying “general observations that are timeless in character” (Labov and Fanshel 1977: 109).
5.4 Summary The examples in this section illustrate two ways in which the broadening characteristics of generalizations can potentially strengthen a speaker’s stance in conversations. In (8), 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. Similar to the examples discussed in Section 4, the speaker’s stance in this episode is jointly maintained. In (9), on the other hand, the speaker generalizes motivations for her past reactions to her brother using generic you constructions, perhaps to invite her friends to understand, or agree, with her stance at that point in time. Additionally, as was suggested for the evaluative generalizations in Section 4, Joanne’s you generalizations are normative statements that make reference to societal discourses.
6.
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, typically evaluation. Analyses offered in Sections 4 and 5 suggest that these expressions are not referential descriptions of the world. Rather, these utterances demonstrate both subjective and intersubjective uses, and are used by participants in the conversations to express individual, interactive, and sociocultural stances. The generalizations are subjective in several ways. Selecting and indexing a general class as a starting point for evaluation or assertion is invariably an expressive act, particularly when the class is not a conventionalized one (e.g., the whole thing). The generalizations in these data frequently function as evaluations and can also be used as rhetorical devices to strengthen speakers’ stances. Additionally, because generalizations can have a broadening, or inclusive, function in these conversations, they participate in intersubjective activities at an interpersonal level (e.g., politeness, demonstrations of solidarity) and more globally in the maintenance of cultural norms through tacit sharing of societal discourses. Notably, the subjective uses of generalizations are tied to their intersubjective uses in several ways. For example, the generalizations with they subjects illustrated in Section 4 convey speakers’ evaluative stances as they function to create ingroup solidarity by othering outsiders. In (9), the speaker, Joanne, uses you generaliza-
134 Joanne Scheibman
tions to broaden her stance in an appeal to the beliefs of her friends. In the guys in their twenties example in (8) and in the extract about Vanda in (5), on the other hand, participants collaborate with speakers by specifying and elaborating their generalizations. These analyses support the idea that in English conversations, individuals’ expressions of stance using generalizations have interactive consequences. And in friendly conversations among intimates a common consequence is the construction of jointly held stances. Generalizations in these conversations, then, are expressions that link individual subjectivities to the group. Moreover, because these utterances are often statements of norms, they contribute to the construction and reproduction of cultural belief systems. Sally McConnell-Ginet observes that “[t]he reproduction of meaning refers to our dependence, in producing meanings, on previous meanings or interpretations, to our dependence in particular on one another’s experience with the linguistic forms being used” (1998: 199). And conversations are important sites in which participants engage with “one another’s experience,” both socially and linguistically. Indeed, the generalizations discussed in this paper allude to common societal discourses such as: the sexual responsibilities of young women in heterosexual relationships, appropriate ways to engage with people with drug dependencies, and attitudes toward aging, work, and scientific authorities. Participants’ uses of these generalizations, then, illustrate ways in which links between expression of individual and social attitudes are reinforced in conversational interactions.
Notes 1. I am grateful to Robert Englebretson for many valuable comments on an earlier version of this paper. 2. 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 another [target]” (1993: 5). The strong relational aspects of stancetaking suggest that the reference point notion might also be a useful way to characterize participant stances in discourse. 3. Language-acquisition researchers have also found that generalizations are sensitive to local patterns of use. For example, Tomasello (2000) presents compelling evidence from studies using nonce words suggesting that young children do not generalize verbs as a syntactic category as they do nouns. 4. In definitions of generalization used in formal systems, such as software and database design, reduction in specificity of an entity or description is called abstraction. “Abstraction aims at simplifying the description of an entity, while generalization looks for common properties among these abstractions” (Kafura 1998: 27). Because my primary objective is to examine generalizations in English conversations (for which differences between abstraction and generalization are not specifically relevant), I will not pursue this technical distinction here.
Generalizations in English conversations 135
5. Transcripts of conversational excerpts used in this paper have been edited for readability. 6. A key to transcription symbols is provided in the Appendix. 7. Cats in this conversation refers to a brand of tractor made by Caterpillar. 8. Two general subject types, which were previously addressed, were excluded from the study: (1) we subjects that designate classes (Scheibman 2004), and (2) utterances with impersonal it and that subjects (Scheibman 2002 for it and that subjects of relational clauses). 9. Excerpts from the Santa Barbara Corpus of Spoken American English are cited, while those from unpublished transcripts do not have an accompanying reference. 10. In the conversational extracts, generalizations and utterances that figure prominently in discussion in the text are bolded. 11. These contributions are not reproduced in (8).
References Abney, L. 1996. “Pronoun shift in oral folklore, personal experience and literary narratives, or ‘What’s up with you?’” SECOL Review 20: 203–226. Benveniste, E. 1971. Problems in General Linguistics. Translated by Mary Elizabeth Meek. Coral Gables, FL: University of Miami Press. Berman, R.A. 2005. “Introduction: Developing discourse stance in different text types and languages.” Journal of Pragmatics 37: 105–124. Biber, D. and Finegan, E. 1989. “Styles of stance in English: Lexical and grammatical marking of evidentiality and affect.” Text 9: 93–124. Biber, D., Johansson, S., Leech, G., Conrad, S. and Finegan, E. 1999. Longman Grammar of Spoken and Written English. London: Longman. Bouacha, M.A. 1995. “Linguistique théorique et recherche en didactique.” Le Français dans le Monde 34: 47–52. Bybee, J. and Pagliuca, W. 1985. “Cross-linguistic comparison and the development of grammatical meaning.” In Historical Semantics, Historical Word Formation, J. Fisiak (ed.), 59– 83. Berlin: Mouton de Gruyter. Bybee, J., Perkins, R. and Pagliuca, W. 1994. The Evolution of Grammar: Tense, Aspect, and Modality in the Languages Of The World. Chicago: University of Chicago Press. Chafe, W. 1994. Discourse, Consciousness, and Time. Chicago: University of Chicago Press. Clark, H.H. 1996. Using Language. Cambridge: Cambridge University Press. Coates, J. 1996. Women Talk: Conversation between Women Friends. Oxford: Blackwell. Du Bois, J.W., Chafe, W., Meyer, C. and Thompson, S.A. 2000. Santa Barbara Corpus of Spoken American English, CD, Part 1. Philadelphia: Linguistic Data Consortium, University of Pennsylvania. Du Bois, J.W., Chafe, W., Meyer, C. and Thompson, S.A. 2003. Santa Barbara Corpus of Spoken American English, CD, Part 2. Philadelphia: Linguistic Data Consortium, University of Pennsylvania. Du Bois, J.W., Schuetze-Coburn, S., Cumming, S. and Paolino, D. 1993. “Outline of discourse transcription.” In Talking Data: Transcription and Coding in Discourse Research, J.A. Edwards and M.D. Lampert (eds.), 45–89. Hillsdale, NJ: Lawrence Erlbaum. Fox, B.A. 2001. “Evidentiality: Authority, responsibility, and entitlement in English conversation.” Journal of Linguistic Anthropology 11: 167–192.
136 Joanne Scheibman
Goffman, E. 1981. “Footing.” In Forms of Talk, 124–159. Philadelphia: University of Pennsylvania Press. Goldberg, A.E., Casenhiser, D.H. and Sethuraman, N. 2004. “Learning argument structure generalizations.” Cognitive Linguistics 15: 289–316. Hill, J. and Irvine, J. (eds.). 1993. Responsibility and Evidence in Oral Discourse. Cambridge: Cambridge University Press. Hongladarom, D. 2002. “Discourse about them: Construction of ethnic identities in Thai print.” In Us and Others: Social Identities Across Languages, Discourses and Cultures, A. Duszak (ed.), 321–339. Amsterdam: John Benjamins. Hopper, P.J. 1997. “When ‘grammar’ and discourse clash: The problem of source conflicts”. In Essays on Language Function and Language Type, J. Bybee, J. Haiman and S.A. Thompson (eds.), 231–247. Amsterdam: John Benjamins. Kafura, D. 1998. Object-Oriented Software Design and Construction with C++. Upper Saddle River, NJ: Prentice Hall. Kärkkäinen, E. 2003. Epistemic Stance in English Conversation: A Description of its Interactional Functions, With a Focus On ‘I think’. Amsterdam: John Benjamins. Labov, W. and Fanshel, D. 1977. Therapeutic Discourse: Psychotherapy as Conversation. New York: Academic Press. Labov, W. and Waletzky, J. 1967. “Narrative analysis: Oral versions of personal experience.” In Essays on the Verbal and Visual Arts, J. Helm (ed.), 12–44. Seattle: University of Washington Press. Lakoff, G. 1987. Women, Fire, and Dangerous Things: What Categories Reveal about the Mind. Chicago: University of Chicago Press. Langacker, R.W. 1987. Foundations of Cognitive Grammar, Vol. I: Theoretical Prerequisites. Stanford, CA: Stanford University Press. Langacker, R.W. 1993. “Reference-point constructions.” Cognitive Linguistics 4: 1–38. Lyons, J. 1977. Semantics, Vol. 1. Cambridge: Cambridge University Press. MacWhinney, B. 1977. “Starting points.” Language 53: 152–168. McConnell-Ginet, S. 1998. “The sexual (re)production of meaning: A discourse-based approach.” In The Feminist Critique of Language: A Reader, 2nd ed., D. Cameron (ed.), 198– 210. London: Routledge. Nuyts, J. 2001. “Subjectivity as an evidential dimension in epistemic modal expressions.” Journal of Pragmatics 33: 383–400. Ono, T. and Thompson, S.A. 1995. “What can conversation tell us about syntax?” In Alternative Linguistics: Descriptive and Theoretical Modes, P.W. Davis (ed.), 213–271. Amsterdam: John Benjamins. Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. 1985. A Comprehensive Grammar of the English Language. London: Longman. Reilly, J., Zamoraa, A. and McGivern, R.F. 2005. “Acquiring perspective in English: The development of stance.” Journal of Pragmatics 37: 185–208. Scheibman, J. 2000. “I dunno ... A usage-based account of the phonological reduction of don’t in American English conversation.” Journal of Pragmatics 32: 105–124. Scheibman, J. 2002. Point of View and Grammar: Structural Patterns of Subjectivity in American English Conversation. Amsterdam: John Benjamins. Scheibman, J. 2004. “Inclusive and exclusive patterning of the English first person plural: Evidence from conversation.” In Language, Culture and Mind, M. Achard and S. Kemmer (eds.), 377–396. Stanford, CA: CSLI Publications.
Generalizations in English conversations 137
Schiffrin, D. 1990. “The principle of intersubjectivity in communication and conversation.” Semiotica 80: 121–151. Silverstein, M. 1976. “Shifters, linguistic categories, and cultural description.” In Meaning in Anthropology, K.H. Basso and H.A. Selby (eds.), 11–55. Albuquerque: University of New Mexico Press. Skarżyńska, K. 2002. “We and they in Polish political discourse.” In Us and Others: Social Identities across Languages, Discourses and Cultures, A. Duszak (ed.), 249–264. Amsterdam: John Benjamins. Strauss, C. 2004. “Cultural standing in expression of opinion.” Language in Society 33: 161– 194. Tannen, D. 1984. Conversational Style: Analyzing Talk among Friends. Norwood, NJ: Ablex. Tao, H. 2001. “Discovering the usual with corpora: The case of remember.” In Corpus Linguistics in North America: Selections from the 1999 Symposium, R. Simpson and J. Swales (eds.), 116–114. Ann Arbor: University of Michigan Press. Thompson, G. and Hunston, S. 2000. “Evaluation: An introduction.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, S. Hunston and G. Thompson (eds.), 1–27. Oxford: Oxford University Press. Thompson, S.A. and Mulac, A. 1991. “A quantitative perspective on the grammaticization of epistemic parentheticals in English.” In Approaches to Grammaticalization, Volume II: Focus on Types of Grammatical Markers, E. Traugott and B. Heine (eds.), 313–329. Amsterdam: John Benjamins. Tomasello, M. 2000. “Do young children have adult syntactic competence?” Cognition 74: 209–253. Traugott, E. and Dasher, R.B. 2002. Regularity in Semantic Change. Cambridge: Cambridge University Press. van Leeuwan, T. 1995. “Representing social action.” Discourse and Society 6: 81–106. Verhagen, A. 1995. “Subjectification, syntax, and communication.” In Subjectivity and Subjectivization: Linguistic Perspectives, D. Stein and S. Wright (eds.), 103–128. Cambridge: Cambridge University Press. Zhang, Qiao. 1998. “Fuzziness – vagueness – generality – ambiguity.” Journal of Pragmatics 29: 13–31.
Appendix: Transcription symbols (Du Bois et al. 1993) Speech overlap Truncated intonation unit Transitional Continuity Final Continuing Appeal Lengthening Medium pause Long pause
[ ] or [[ ]] -. , ? = … …(N)
138 Joanne Scheibman
Laughter Laugh quality Quotation quality Vocal noises Indecipherable syllable Researcher’s comment
@ () X (( ))
The stance triangle John W. Du Bois
University of California, Santa Barbara
1.
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, to position social actors with respect to those objects, to calibrate alignment between stancetakers, and to invoke presupposed systems of sociocultural value. Yet very little is understood at present about stance: what it is, how we do it, what role language and interaction play in the process, and what role the act of taking a stance fulfills in the broader play of social life. These are the questions which will inform this paper, 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. If we are to achieve any general understanding of stance, it will be necessary to seek the foundational principles which underlie the act of taking a stance and negotiating its meaning. Because the diversity of observable stances extends in principle without limit, 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. Realizing such a goal will require us to define a research agenda capable of bringing together multiple coordinated lines of inquiry drawn from a range of disciplines concerned with the use of language. As one step in this direction, 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. Stance can be approached as a linguistically articulated form of social action whose meaning is to be construed within the broader scope of language, interaction, and sociocultural value. Setting the problem in this way brings into play several aspects of language in interaction. As we seek the theoretical resources needed to account for the achievement of stance, we find ourselves faced with a
140 John W. Du Bois
complex web of interconnections linking stance with dialogicality, intersubjectivity, the social actors who jointly enact stance, and the mediating frameworks of linguistic structure and sociocultural value they invoke in doing so. Dialogicality makes its presence felt to the extent that a stancetaker’s words derive from, and further engage with, the words of those who have spoken before – whether immediately within the current exchange of stance utterances, or more remotely along the horizons of language and prior text as projected by the community of discourse. While the idea of dialogicality has proved theoretically evocative (Bakhtin [1934]1981; Voloshinov [1929]1973), 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). In this light, stance represents a promising testing ground to explore the potential of a more explicit dialogic method in the context of conversational interaction. Here I draw on my recent formulation of dialogic syntax as a specifically dialogic method (Du Bois 2001). Dialogic syntax looks at what happens when speakers build their utterances by selectively reproducing elements of a prior speaker’s utterance. As an analytical practice, dialogic syntax details the process of mapping resonances between juxtaposed utterances in discourse. As it happens, when the formal method of dialogic syntax is applied to conversational interaction, it turns up a remarkable number of stance pairs characterized by a recurring functional-interactional configuration, in which the stance utterance of a subsequent speaker is constituted as bearing a close analogy to the stance utterance of a prior speaker. As stances build on each other dialogically, the analogy implied by their structural parallelism triggers a series of interpretative and interactional consequences, which will be seen to carry significant implications for the interaction at hand, and, at a more general level, for the theory of stance. Injecting dialogicality into the analysis of stance leads naturally to a concern with intersubjectivity, which I understand as the relation between one actor’s subjectivity and another’s. Though its conceptual origins lie in philosophy, intersubjectivity has long been recognized as an issue for the social sciences, where it has influenced especially those approaches which locate language and cognition within a sociocultural and/or sociocognitive perspective (Bruner 1986; Clancy 1999; Hanks 1990; Lucy 1993; Ochs 1988; Schutz 1962; Silverstein 2001; Tomasello 1999; Vygotsky 1986). It should be noted that intersubjectivity presupposes subjectivity. While subjectivity has been getting the greater share of attention lately, being recognized for its role in the cognitive organization of language structure (Langacker 1991; Traugott 1989) and the discourse-functional organization of language use (Maynard 1993; Scheibman 2002), intersubjectivity is no less indispensable as a piece of the larger stance puzzle. The significance of the necessary link between the intersubjective, subjective, and objective will become evident
The stance triangle 141
once we recognize them as fundamental sociocognitive relations which organize language use. Intersubjectivity and its companions play a prominent role in the present approach to stance, insofar as they serve to ground the sociocognitive aspects of stancetaking in dialogic interaction. To complete the picture, we need to look at how the dialogic and intersubjective dimensions of stance relate to the stancetakers’ actions, and to the sociocultural frames that mediate the consequences of their actions. Stance is realized, in the usual case, by a linguistic act which is at the same time a social act. The act of taking a stance necessarily invokes an evaluation at one level or another, whether by assertion or inference. This in turn implicates those dimensions of sociocultural value which are referenced by the evaluative act. Sociocultural value is mobilized and deployed through stance processes. Via specific acts of stancetaking, value can be focused and directed at a precise target, as locally relevant values are activated to frame the significance of participant actions. Moreover, since language is reflexive (Haviland 1996; Lucy 1993), and participants routinely monitor who is responsible (Hill and Irvine 1993) for any given stance, the very act of taking a stance becomes fair game to serve as a target for the next speaker’s stance. Stance both derives from and has consequences for social actors, whose lives are impacted by the stances they and others take. Moreover, in many cases the current stance act resonates both formally and functionally with a stance taken in prior discourse. Thus the value of any stance utterance tends to be shaped by its framing through the collaborative acts of co-participants in dialogic interaction. We begin to appreciate why stance should come to wield both subtlety and power in the dynamics of social life. In this paper, I work to bring these elements together in a unified framework for stance. The model I propose is articulated in terms of a set of triangular relations among the components of stance. But before we get to the stance triangle, there is groundwork to be laid. I begin the paper with a brief presentation of several different kinds of stance phenomena. I then address the problem of contextualizing stance, 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. I go on to explore the role of intersubjectivity and other sociocognitive relations in the elaboration of stance, arguing that making sense of intersubjectivity in interaction depends on a dialogic understanding of language use. This sets the stage for the next level of theorizing stance. I introduce the stance triangle as a way of representing the components of the stance act, and more importantly, articulating their multiplex interrelations. In its guise as a representation of the fundamental structure of the stance act, the stance triangle attempts to shed light on the realization, interpretation, and consequences of stance in interaction. The result is a fundamentally dialogic perspective on stance which sees the stance act
142 John W. Du Bois
as shaped by the complex interplay of collaborative acts by dialogic co-participants. By way of conclusion, I summarize the main contributions of the stance triangle, and pose the question of what, after all, is so special about stance.
2.
Kinds of stance
One way to begin thinking about stance is to look at some likely exemplars, based on what has been recognized as stance (or a stance-related category) by one analytical tradition or another. While the range of proposals in the literature is too broad to survey here, 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 categories. For now we can think of these as different kinds or types of stance acts, but in the end we will have to consider alternative formulations of the issue of stance diversity. By the same token, although for expository purposes we begin with abbreviated examples of stance utterances viewed in isolation, it will soon become clear that the actual stance taken cannot be fully interpreted without reference to its larger dialogic and sequential context. Perhaps the most salient and widely recognized form of stancetaking is evaluation. Evaluation has received considerable attention in recent years (Conrad and Biber 2000; Hunston and Sinclair 2000; Hunston and Thompson 2000; Labov and Waletzky 1967; Lemke 1998; Linde 1997; Macken-Horarik and Martin 2003; Thompson and Hunston 2000). A closely related concept is that of assessment, as analyzed in conversation analysis (Goodwin and Goodwin 1992; Goodwin 2006; Pomerantz 1984). Work on the related notion of appraisal has been pursued from the perspective of systemic functional grammar (Martin 2000), and additional important work on stance, point of view, and related notions has been developed by a number of scholars (Berman et al. 2002; Berman 2005; Chafe 1994; Kärkkäinen 2003a, 2003b; Kockelman 2004; Shoaps 2004). Consider the following three examples of evaluation, drawn from the Santa Barbara Corpus of Spoken American English (Du Bois et al. 2000; Du Bois et al. 2003):2, 3 (1) (Conceptual Pesticides SBC003: 317.46-318.16) PETE; That’s horrible. (2) (Runway Heading SBC022: 193.440-194.341) LANCE; (TSK) That’s ideal. (3) (Appease the Monster SBC013: 1117.12-1118.12) KEVIN; @that’s @nasty.
The stance triangle 143
In these examples, taken from three different conversations, the stance predicates horrible, ideal, and nasty are used to evaluate something. The thing evaluated – the specific target toward which the evaluation is oriented – is referred to in each case by the demonstrative pronoun that. The evaluative target may be called the object of stance (for reasons that will become clear in due course). In general terms, 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. The next set of examples presents a somewhat different pattern of stance: (4) (5) (6)
(Runway Heading SBC022: 612.895-613.160) LANCE; I’m glad. (Hey Cutie-Pie SBC028: 52.330-53.355) JEFF; (TSK) I’m so glad. (Lambada SBC002: 665.79-667.35) MILES; I’m just ama:zed.
In these stance utterances, the first person pronoun I, indexing the stancetaker, is followed by an affective predicate, positioning the speaker as glad, so glad, or amazed.4 As speakers position themselves affectively, they choose a position along an affective scale – as, for example, either glad or so glad. Such utterances have often been described as indexing affective stance (Besnier 1993; Haviland 1991; Maynard 1993; Ochs 1996; Shoaps 2002). Speakers may position themselves not only along an affective scale, but also along an epistemic scale (Clift 2006; Haviland 1991; Heritage and Raymond 2005; Kärkkäinen 2003b; Thompson and Mulac 1991), for example, presenting themselves as knowledgeable or ignorant: (7) (Appease the Monster SBC013: 185.50-185.99) KENDRA; I kno:w. (8) (Risk SBC024: 588.565-588.970) DAN; .. I don’t know.
The general concept which subsumes both affective and epistemic stance acts of the sort illustrated here is positioning (Davies and Harré 1990; Du Bois 2002a). Positioning can be defined, provisionally, as the act of situating a social actor with respect to responsibility for stance and for invoking sociocultural value. In the examples presented here for both affective and epistemic positioning, the speaker who is taking the stance is indexed via a first-person pronoun in syntactic subject role (I), while the stance predicate (adjective or verb) specifies the nature of the stancetaker’s position, whether with respect to an affective (glad) or an epistemic (know) state, or both (amazed).
144 John W. Du Bois
In the next set of examples, the structural pattern is superficially similar to the one just presented: (9) (LSAC 1296-02)5 PAT; I agree. (10) (Doesn’t Work in This Household SBC019: 127.391-128.183) MELISSA; I totally agree.
Here a first-person pronoun in syntactic subject position is followed by a stance predicate (the verb agree). But what is being achieved thereby is different enough to warrant recognition as a distinct type of stance. The difference lies in the pragmatic-interactional configuration that it enacts. By uttering I agree the speaker defines her stance in relation to that of another party, typically the person she is addressing. The addressee’s role is usually left implicit, but can be made explicit on occasion: (11) (LSAC 1396-01)6 LESLIE; I agree with you.
As this example illustrates, agreement normally implies agreement with someone,7 which in conversation is usually the person being addressed. The general term for this kind of stancetaking is alignment (Du Bois 2002a; Heritage 2002; Heritage and Raymond 2005). Alignment can be defined provisionally as the act of calibrating the relationship between two stances, and by implication between two stancetakers. Although a stance verb like agree may be the most transparent way to display alignment with another speaker, it is certainly not the usual way. More commonly, speakers show alignment by stance markers like yes or no, or gestures like a nod or a headshake, or any number of other forms that index some degree of alignment. Just as often, participants allow their alignment to remain implicit, inviting the listener to infer it based on comparing the relevant stances. The role of implicit stance alignment will prove to be especially important in the management of intersubjectivity. These examples make it clear that stance utterances can vary in significant and systematic ways. But the question remains as to how to interpret this variability. Of the cases examined so far, most seem to involve one of three well-differentiated kinds of stance function, namely evaluation (that’s horrible), positioning (I’m glad), and alignment (I agree). But do these stance functions really represent separate categories of stance – or simply different facets of the speaker’s stance, more broadly conceived? What about other candidates that have been proposed as stance types – epistemic stance (I know), affective stance (I’m glad), and more? The problem is, while these are sometimes viewed as different types of stance
The stance triangle 145
acts, no one has yet resolved the question of how many such types should be recognized – whether there will turn out to be three, five, or many more subtypes of stance. Nor have the criteria for deciding the question been established. More seriously, there is the matter of whether the various forms of stance occur separately or together. In other words: Can a single stance combine aspects of more than one of these stance functions? This question raises the possibility of an alternative approach which recognizes a more complex picture of the stance act, seen as encompassing multiple facets at once. Building a more unified understanding of stance has the advantage of avoiding a limitless proliferation of stance types. Such an approach would seek to interpret the diversity of stances not as distinct types of stance, but simply as different facets of a single unified stance act. The choice between proliferation of stance types and integration of stance facets will be returned to below. The answer to some of these questions may hinge in part on the realization that, in the analysis presented so far, something seems to be missing. In stances which were characterized as positioning (4–8), for example, the question arises: What is the speaker positioning himself about? Similarly, in stances involving alignment (9–11), it seems equally essential to ask: What is the speaker agreeing about, and who are they agreeing with? These questions point to some important issues about stance. Speakers do not just perform generic stance types, they perform specific stance acts, which have specific content and are located in a particular dialogic and sequential context. To fully understand what stance is being taken on any given occasion, we need to learn more about how speakers realize stances and how hearers interpret their situated meanings. In part, this question concerns how participants contextualize the stance utterance in order to interpret it, as the stance emerges across successive utterances through processes of dialogic action. This is the issue we begin to explore in the next section.
3.
Contextualizing the emerging stance
If stance is an act, we should expect to locate it in utterances. For the sake of argument, however, let us begin this discussion where it has so often begun before in the study of language – not with the utterance, but the sentence. The sentence can, with a little imagination, be hypothesized as an abstract linguistic structure detached from any mooring in a specific context of use. What can the decontextualized sentence reveal – and what can it not? Consider the following sentences: (12) The Caribbean is incredible. (13) It was really great.
146 John W. Du Bois
(14) I would love to go.
Each of these sentences contains at least one apparently evaluative word: incredible, great, and love (and perhaps others). These words are obviously not neutral descriptions of external reality, but imply value judgments regarding some referent. The evaluative connotations of such words are evident even from sentences taken in isolation. In fact, the evaluative meaning often comes through even in an arbitrary listing of individual words (incredible, great, ideal, horrible, nasty, and so on). But stance is more than the context-free connotations of words or sentences. In the grammarian’s standard presentation of the isolated sentence, stance remains incomplete. The missing ingredients can only be found by contextualizing the utterance, defined as the situated realization of language in use. Any utterance carries cues for its own indexical contextualization (Gumperz 1992; Silverstein 1976). Contextualization cues (or indexical signs) work by pointing beyond the utterance to its presupposed conditions of use. Thus in situating the understanding of any stance, 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; Peirce [1885]1933; 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. There are at least three things we need to know about a given occasion of stancetaking, 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.
3.1
Who is the stancetaker?
In conversation, participants normally care who says what, and monitor it accordingly. Participant awareness of the attributability of utterances is routinely represented in most systems for transcribing discourse (Du Bois 1991; Du Bois et al. 1993; Jefferson 2004), typically by supplying each transcribed utterance or turn with a label indicating the identity of the speaker. In the following utterances by three different speakers, the identity of the speaker is indicated as usual at the beginning of each turn. (15) (Deadly Diseases SBC015: 210.075-212.730) JOANNE; the– the Caribbean is incre:dible.
The stance triangle 147
(16) (Conceptual Pesticides SBC003: 1326.05-1326.85) MARILYN; it was really great. (17) (Deadly Diseases SBC015: 187.025-188.070) KEN; I would love to go:.
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. A key component of the context of any utterance is the speaker who is responsible for it. In the representations of utterances given here, we are told that someone named Joanne8 is responsible for the claim that a certain place is incredible; that Marilyn is responsible for the statement that something was really great; and that Ken is responsible for a display of affect or preference about going somewhere. To be sure, just attributing speakership in this way doesn’t reveal much unless the speaker’s identity carries some significant associations for us. But in most conversational settings, participants may draw on a range of biographical associations for the current speaker, regardless of whether they happen to know the speaker’s name. Participants remember interactionally salient information about co-participants, and so may factor into their stance interpretation, in addition to what the speaker is saying right now, some or all of the following: what the speaker has said previously (whether on this occasion or some other); what sort of relationship they have displayed up to now relative to co-present others; what accent, voice quality, and intonation they are speaking with; what their displayed regional, ethnic, gender, or other identities may be; whether they appear entitled to their claimed identities; details of their life story, if known; and so on. Moreover, participants derive memorable information about each other from stances taken, which they may retain as socially salient, with the option of introducing it into future processes of stance interpretation. 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, in part because of the dialogic connections that arise between stances.
3.2 What is the object of stance? To make sense of a given stance we need to know not only who is speaking, but what they are speaking about. Among other things, we need to know the referential object or target toward which the stance is being directed – for example, what is claimed to be incredible or great, where the speaker displays a desire to go, and so on. Consider the following examples:
148 John W. Du Bois
(18) (Raging Bureaucracy SBC004: 17.67-18.82) CAROLYN; That stuff is great. (19) (Conceptual Pesticides SBC003: 1326.05-1326.85) MARILYN; it was really great.
If two people each evaluate something as great, have they taken the same stance? The cited utterances, drawn from two different conversations, differ slightly in their evaluative predicates (is great vs. was really great) and in the noun phrases which designate their respective objects of stance (that stuff vs. it). But these small differences don’t really speak to the main question. What we want to know, if we are to decide what stance is being taken, is what it and that stuff refer to. A crucial part of interpreting any stance utterance is identifying the object of stance, as part of the process of referential grounding (Hanks 1990). Often the immediate prior discourse provides sufficient contextualization to resolve the reference of a pronoun, demonstrative noun phrase, or other referring form, thus establishing the identity of the relevant object of stance: (20) (Conceptual Pesticides SBC003: 1324.18-1326.85) MARILYN; S– really relaxing weekend, you know, it was really great. (21) (Raging Bureaucracy SBC004: 15.42-18.82) CAROLYN; You know what’s good is .. hibiscus cooler. (0.2) That stuff is great.
A weekend and a drink of hibiscus cooler may both be great, but they will be great in very different ways. The respective stances of Marilyn and Carolyn, once referentially grounded, cannot be regarded as the same. Stance is a property of utterances, not of sentences, and utterances are inherently embedded in their dialogic contexts. It’s not just pronouns that need to be resolved by contextualizing the stance utterance, but also the meaning of content words and other elements as well. Compare the following two stance utterances, which both employ the word incredible in the evaluative predicate, but apply it to rather different stance objects: (22) (Deadly Diseases SBC015: 210.075-212.730) JOANNE; the– the Caribbean is incre:dible. (23) (Cuz SBC006: 168.60-170.30) ALINA; (H) The male athletes were incredible,
The stance triangle 149
The evaluative content of the message conveyed by an evaluative word like incredible may shift according to whether it applies to a vacation destination (valued for passive attributes like visual beauty, tranquility, and so on) or to athletes (valued for the active display of strength, agility, talent, and so on). In I would love to go (17), the utterance may initially appear more self-contained, as if carrying its meaning complete within itself. But this illusion is soon dispelled by examination of the prior discourse: (24) (Deadly Diseases SBC015: 164.420-188.070) 1 JOANNE; (H) I don’t know how many Americans have been to Nicaragua. ((39 LINES OMITTED)) 41 KEN; I would love to go:.
Here, the phrase love to go refers not to going in the abstract, nor to going somewhere at random, but specifically to going to Nicaragua. The stance object that Ken is evaluating is thus to go to Nicaragua, abbreviated as to go. In such cases, the contextualized stance utterance takes its interpretation in part from the prior stance of a dialogic co-participant. This contextualization remains essential to the interpretation of the current stance even if the relevant prior stance occurred quite a bit earlier, as is the case here. In sum, identifying the object of stance – what the evaluation is about – is an essential part of the process of stance interpretation, for participants and analysts alike.
3.3 What stance is the stancetaker responding to? Knowing the identity of the stancetaker and the object of stance is a good start, but we remain on uncertain ground until we know what prior stance the current stance is being formulated in response to – its counterstance, if you will. Why this stance is being taken, why just now, 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. Consider the examples of alignment given previously (I agree, I agree with you). Despite the fact that as sentences these are grammatically complete, as stances they are pragmatically incomplete. People don’t agree in the abstract, they agree with someone about something. While a sentence like I agree with you (11) foregrounds – or profiles, in Langacker’s (1987) terms – the dimension of alignment almost exclusively, for its interpretation it must still indexically incorporate a prior stance content, including the relevant object of stance. Normally the relevant stance content will be locatable in the prior discourse, specifying what specific stance is being agreed with:
150 John W. Du Bois
(25) (LSAC 1396-01) 1 CORA; 4 LESLIE;
we are considered white collar because we’re social workers. ((2 LINES OMITTED)) I agree with you.
Here, 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. The result can be informally paraphrased9 as I agree with you that we are considered white collar (because we’re social workers). Similarly in the following example: (26) (LSAC 1296-02) 1 KIM; 2 PAT; 3 4
No more or no less than any other school is the way I see it. Yeah, I agree. I agree.
When Pat says I agree in line 3 (and again in line 4), she is doing more than just “being agreeable” (or even “doing agreement”). Her stance is quite specific, and can be paraphrased as, I agree (with you that it’s) no more or no less than any other school....10 Finally, consider Melissa’s utterance of I totally agree (10), whose relevant context is as follows: (27) (Doesn’t Work in This Household SBC019: 98.652-129.138) 1 JAN; Take it downstairs. ((28 LINES OMITTED)) 30 MELISSA; I totally agree. 31 I should go downstairs.
In saying I totally agree, Melissa is not agreeing in the abstract, but specifically agreeing with her mother Jan’s directive that she should take her homework downstairs. 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, if need be, to find the proposition spelled out overtly. Lest there be any doubt as to what proposition she is agreeing to in line 30, Melissa makes it explicit in her next utterance (line 31), yielding a composite stance that is roughly paraphrasable as I totally agree (that) I should go downstairs.11 To generalize, given a decontextualized sentence which apparently expresses simple agreement (I agree, I totally agree, I agree with you), it is not possible to tell from the sentence alone what is being agreed with.12 Only by referencing the relevant prior stance, locatable ana-
The stance triangle 151
phorically in the dialogic context, can the meaning of the present agreeing stance be understood. We have considered three questions about the context of stance which are likely to be relevant, in some formulation or other, to the interpretation of any act of stancetaking. In asking, Who is the stancetaker, what is the stance about, 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 taken actually is. In more general terms, these questions about stance can be linked to notions of stance subject, stance object, and alignment, connections which will be further developed below. But we are far from claiming that these are the only questions, or necessarily the main ones, or the best formulations of them. Certainly 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, and some of these will come up below. But it is important to point out here that even the questions informally posed so far frequently require participants (and analysts) to go beyond what is explicit in the words of the stance utterance itself. Still the questions must be answered. Their relevance to stance does not depend solely on the presence of explicit words, gestures, prosody, or other communicative elements, however important these may be, but is grounded ultimately in the systematic knowledge which participants control regarding what can be expected to be present in any stance. The constant relevance of the general components of stance influences what we expect to know about any act of stancetaking, and thereby shapes its specific interpretation. This holds true whether the information is directly expressed in the stance utterance, or is only to be found distributed across multiple utterances by different speakers within extended sequences of dialogic exchange. The claim is that in each case, certain well-defined items of information are actively sought out by participants in response to the projectable structure of stance. We have been considering how the contextualization of actual stance utterances (e.g., 15–27) contributes to the interpretation of stance. In contrast, little or nothing of the relevant contextualizing information can be gleaned from inspecting idealized sentences (e.g., 12–14) in isolation, even if the sentences are fully grammatical and fully meaningful – at least as meaningful as they can be, if limited to the sentential level of referring-and-predicating functions (Silverstein 1976, 2001). Yet once we identify the relevant contextual features, the necessary ingredients start to fall into place for a full pragmatic and interactional interpretation. In constructing a logic of stance interpretation, what we want to describe is the participant’s interpretive process, which we track by close observation of their own interpretive actions in stance-rich environments. This interpretive inquiry is akin to that which has motivated scholars as diverse as Geertz (1973) and Sacks (1992), whose overt theoretical positioning may otherwise appear so disparate as
152 John W. Du Bois
to be incommensurable. We need contextual grounding (Gumperz 1992; Hanks 1990; Silverstein 1976) in dialogic and sequential context (Linell 1998; Sacks 1992; Schegloff 1996; Voloshinov [1929]1973) to arrive at a sufficiently enriched interpretation of the utterance (Recanati 1989; Sperber and Wilson 1995). We cannot overlook the influence of even so basic interpretive processes as these if we want to understand how stances come to be understood. Ultimately, the specificity of stance is grounded in its dialogicality. Only when the dialogic context is taken into account does stance become complete.
4.
Subjectivity and positioning
In light of this understanding of the processes of stance contextualization, we are now ready to re-examine the issue of positioning, which we consider in light of speaker subjectivity. Subjectivity and positioning go hand in hand, inasmuch as the act of positioning regularly invokes a dimension of speaker subjectivity. Consider the following: (28) (Lambada SBC002: 865.91) JAMIE; I like this song. (29) (This Retirement Bit SBC011: 444.12-445.36) SAM; I don’t like those.
In these stance utterances, the subjective dimension is registered overtly through several discrete linguistic elements, including personal pronouns (I) and affective verbs (like, don’t like). The personal pronoun I points directly to the speaking subject (Benveniste 1971; Culioli 1990; Ducrot 1972, 1984)13 who, in these cases at least, is the one taking the stance. The affective stance predicate indexes specific aspects of the subject’s feelings, positioning the speaker subjectively along some scale of affective value. Yet it cannot be overlooked that subjective predicates which are transitive (like, don’t like, love, hate, and so on) regularly specify also the object to which the subject is orienting affectively.14 The transitive clause I like this song (28), uttered in the auditory presence of a song playing on the stereo, directly specifies the object of stance – the entity being oriented to. This example compactly illustrates the co-existence of subjective (I) and objective (this song) elements within a unified stance. Similarly, in I don’t like those (29), the coherent expression of speaker subjectivity requires both a subject and an object of stance (specified as I and those, respectively). In sum, to articulate this kind of subjectivity, what is required is an orientation to a specific object of the speaking subject’s
The stance triangle 153
stance, combined with specification of a particular intentional relation (Searle 1983), such as desiring, loving, hating, liking, not liking, and so on. The following stance diagrams represent the relations of evaluation and positioning between the stance subject and stance object: (30) Speaker JAMIE;
Stance Subject I
Positions/Evaluates like
Stance Object this song
(31) Speaker SAM;
Stance Subject I
Positions/Evaluates don’t like
Stance Object those
This kind of stance diagram is useful in making explicit several aspects of our analysis. First, we find it essential to represent the identity of the speaker, because this tells us who is the speaking subject (the stancetaker). Second, in analyzing the various discrete components of the utterance, we label the words which overtly express or index the stance subject and the stance object. Third, we label the verb or other stance predicate according to the kind of stance action it performs. In these examples, the predicate (like, don’t like) obviously serves to position the subject, but it also commits the stancetaker to a certain evaluation of the object. In recognition of this apparent dual stance function, the representations of evaluation and positioning are combined in a single column. While some stance utterances evidently perform combinations of functions (for example, evaluation of a stance object combined with positioning of a stance subject), we might still ask whether it is possible for a stance utterance to express only pure subjectivity. It is true that there are stance utterances which overtly position the speaking subject without explicitly including any reference to a stance object. In what appear to be simple one-place predicates (Thompson and Hopper 2001), speaking subjects position themselves subjectively – and that seems to be the end of it. Consider examples (4) and (5), repeated here for convenience: (32) (Runway Heading SBC022: 612.895-613.160) LANCE; I’m glad. (33) (Hey Cutie-Pie SBC028: 52.330-53.355) JEFF; (TSK) I’m so glad.
Up to now we have analyzed these cases as simply reflecting the speaker’s subjective self-positioning, with respect to a scale of affective value (glad). In contrasting Lance’s I’m glad with Jeff ’s I’m so glad, what is immediately obvious as a difference is the presence of the intensifying adverb so in the latter case. 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. While this may be true as far as it goes, to focus exclusively on the subjective side of the equation is to
154 John W. Du Bois
leave out a key variable: the object of stance. Here the stance object is, once again, what the speaker is affectively orienting to. 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. 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. The claim is that it is a regular feature of subjectivity to orient to an object. If the stance object is not overtly specified within the immediate stance utterance, participants will feel that something is missing. If subjectivity requires orientation to an object, the full meaning of any subjective stance must remain mysterious until we locate the object, even if this requires us to search the discourse context to find it. 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. Such appears to be the case in examples (32)–(33) above. 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. But that’s not to say that the participants are not orienting to a stance object. Once we take into account the sequential context in which the stances developed in the first place, their actual significance becomes clear. In the first I’m glad example, Lance, an apprentice air traffic controller, is being debriefed after a work session by Randy, his trainer: (34)
(Runway Heading SBC022: 607.505-613.160) 1 RANDY; N:o significant problems. 2 (0.5) 3 #to #dea– — 4 .. to talk about. 5 (2.0) 6 LANCE; I’m glad.
Here the stance content of the utterance I’m glad emerges from successive contributions by Randy and then Lance, with the cumulative result being paraphrasable as something like I’m glad (that there are) no significant problems to talk about.15 In the second example, Jeff is talking on the phone to Jill about her friend who is visiting her, and asks: (35)
(Hey Cutie-Pie SBC028: 49.985-53.355) 1 JEFF; Are you guys having fun? 2 JILL; Y:es:. 3 (0.6) 4 JEFF; (TSK) I’m so glad.
The stance triangle 155
Quite a different stance emerges from this sequential contextualization, paraphrasable as I’m so glad you guys are having fun.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, but is instead based on differences in the dialogic sequence through which the stance emerged. 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. People are not glad in general, but glad that (something). 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. Even self-positioning presupposes an object, namely, the specific entity or state of affairs toward which the speaker expresses their subjective stance. The following diagrams represent the relevant relationship between stance subject and stance object: (36) # Speaker 6 LANCE;
Stance Positions/ Stance Subject Evaluates Object I ’m glad {there are no significant problems}
(37) # Speaker 4 JEFF;
Stance Positions/ Stance Subject Evaluates Object I ’m so glad {you guys are having fun}
Along the same lines, the next stance utterance appears at first to be a simple display of amazement: (38) (Lambada SBC002: 665.79-667.35) 69817 MILES; I’m just ama:zed.
But the sequential context makes it clear that Miles is not just amazed, he is amazed about something: (39) (Lambada SBC002: 660.75-667.35) 696 MILES; Cause there’re a lot of women out there who apparently don’t believe in using condoms. 697a (1.0) 697 PETE; Hm. 698a (0.6) 698 MILES; I’m just ama:zed.
Summing across the full discourse context, 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. Clearly, the stance act of affective self-positioning (as
156 John W. Du Bois
glad or amazed) is incomplete until we include the object of stance – what the speaker is glad or amazed about. Although it may be merely implicit in the current stance utterance, the stance object is an indispensable component of even a subjective stance. Even in its absence, the stance object remains relevant and hence may trigger a search for it in the prior discourse. The conclusion: Subjectivity takes an object. If subjectivity must have its object, this should be as relevant to epistemic subjectivity as to affective subjectivity. Consider examples (7)–(8), repeated here for convenience: (40) (Appease the Monster SBC013: 920.13-921.13) KENDRA; I kno:w. (41) (Risk SBC024: 588.565-588.970) DAN; .. I don’t know.
On the face of it, Kendra positions herself as knowing, while Dan positions himself as not knowing. At least this is what the words taken literally and in isolation seem to mean. But such interpretations in isolation are meaningless. How far off the mark they are becomes evident once we take into account the dialogic context: (42) (Appease the Monster SBC013: 919.23 -921.13) WENDY; Those are good spatula[s]. KENDRA; [I] kno:w. (43) (Risk SBC024: 585.630-588.970) JENNIFER; .. Are we not attacking each other until we get rid of the striped guy? DAN; .. I don’t know.
Kendra and Dan each speak no more than three words, yet manage to produce apparently complete stance utterances thereby. But the stances they achieve through their slender utterances are more complex than what is immediately evident in the stance utterances themselves. Their respective stances emerge only from the larger dialogic sequence, as shown in the following representations: (44) Stance Positions/ Stance # Speaker Subject Evaluates Object 2 KENDRA; I know {those are good spatulas}
The stance triangle 157
(45) Stance Positions/ # Speaker Subject Evaluates 2 DAN; I don’t know
Stance Object {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. Rather, they know (or don’t know) particular things. Generally the precise specification of what they know, if not present in the sentence itself, is already there in the immediate prior discourse – often in the utterance of a dialogic partner. It takes Wendy’s and Kendra’s utterances together for Kendra’s stance to emerge. Likewise, it takes Jennifer and Dan working together to articulate Dan’s emerging stance. This kind of co-action, as realized in the context of conversational interaction, is a big part of what it means for linguistic action to be dialogic. Once our analysis systematically incorporates the dialogic co-participants’ contributions to the emerging stance, several things start to become clear. First, a stance verb like know often points to a dialogically prior stance, articulating a precise indexical relation to it. Second, the prior stance must be incorporated anaphorically into the interpretation of the overall emergent stance which culminates in the current stance utterance. Third, the stance act is not necessarily complete within a single intonation unit, clause, sentence, or even turn. Stance utterances like I know and I don’t know are designed to incorporate their dialogic antecedents, through which they gain the interpretative specificity they need to be complete. This has important consequences for our understanding of subjectivity. Despite initial appearances, 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, whether affective or epistemic. Rather, it necessarily combines a subjective and an objective component. In constituting subjectivity, 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, informed about, and so on). The link that is constituted between subject and object creates a vector of subjectivity. The subject-object link is often achieved dialogically, through separate but coordinated contributions by several co-actors, as successive stance utterances – stance and counterstance – are deployed in response to each other. Despite popular conceptions of subjectivity as a purely internal, solipsistic state of the individual psyche, 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. In the end, subjectivity proves meaningful only when subject and object are defined in relation to each other.
158 John W. Du Bois
Though we have not yet arrived at a full understanding of what is needed to contextualize and interpret stance, we have come far enough to draw some preliminary conclusions. The three questions about stance contextualization that were introduced at the outset – regarding the identity of the stancetaker, the object of stance, 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 utterance. 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. We need to ask what else may be needed to complete the picture of the stance in question. So far, 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. This predictable pattern remains relevant even if one of the elements is not explicitly verbalized within the current stance utterance, but is to be located anaphorically in the discourse of a prior speaker. It is in fact routine to find complex stance culminations, precipitated in the end by a seemingly simplex stance utterance, but dependent on the prior context. In many such cases, it is possible to distinguish the object-centered element from the subject-centered element, and to attribute each element to a distinct portion of the discourse, whether as a specific word or phrase in the current or prior stance utterance. The concrete localization of the overt words or other meaningful elements 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. In the end, however, no stance stands alone. It is the stance utterance with its dialogic context that is the relevant unit for stance interpretation. In order to arrive at a situated interpretation for any particular stance utterance, it is necessary to monitor the developmental history of the emergence of the stance, as its content is enriched via the structure of expectations regarding the necessary recurrent components in the general model of stance. 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. To get to the next level of understanding, a key issue that remains to be addressed concerns the role of intersubjectivity. As with the subjective and objective elements of stance, the place of the intersubjective 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 dialogically realized.
The stance triangle 159
5.
Intersubjectivity and alignment
We have seen how subjectivity figures in stance, most evidently in the case of positioning, where what is positioned is typically the speaking subject – the stancetaker. A positioning utterance like I’m glad foregrounds its subjectivity via overt cues such as first-person pronouns, affective predicates, and other elements that index salient aspects of the speaking subject. Moreover, the subjective stance presumes an orientation to an object, whether overtly present in the stance utterance (I don’t like those) or not (I’m amazed). But even after identifying the missing stance object, we still only have part of the story. Some hint of what remains to be incorporated is suggested by a simple observation. In the dialogic realization of stance, the subjective orientation to a stance object may be shared among more than one participant. Indeed this object-orientation may extend across multiple stance acts by different speakers. This gives rise to what I call the shared stance object. (For the related notion of joint attention, see Hobson 1993; Kidwell and Zimmerman 2006; Moore and Dunham 1995; Tomasello 1999; Tomasello et al. 2005.) As we will see, the shared stance object becomes the cornerstone of the dialogic construction of intersubjectivity. What are we to make of intersubjectivity? How does it enter into the dialogic realization and interpretation of stance? As the word itself suggests, intersubjectivity presupposes subjectivity, at least etymologically. But to observe intersubjectivity in action, 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, when the subjective stances of two participants collide within a dialogic exchange. When we learn to see how one speaker’s subjectivity reacts to another’s subjectivity, we will have a real opportunity to witness the dialogic emergence of intersubjectivity. Consider the following exchange, which provides a larger window of context for example (29): (46)
(This Retirement Bit SBC011: 444.12-446.30) 1 SAM; I don’t like those. 2 (0.2) 3 ANGELA; I don’t either.
Through this kind of sequential juxtaposition of evaluative stances, intersubjectivity rises to focal prominence. Analogical relations are established between the juxtaposed stances (I don’t like those : I don’t either). The foregrounding of this dialogic relation potentially invites inferences based on the comparison. A similar
160 John W. Du Bois
dialogic relation arises between the two participants’ stances in the next example as well: (47)
(A Tree’s Life SBC007: 581.32-585.99) 1 ALICE; I don’t know if she’d do it. 2 (0.6) 3 MARY; I don’t know if she would either.
It is useful to have a way to represent the implicit structure of the stance parallels in such cases, in a way that perspicuously captures the potential for participant inferencing about stance. One tool that has proven useful for representing dialogic relations between stances is what I have called the diagraph (Du Bois 2001). While the diagraph was developed for independent purposes – primarily for analyzing the structural relations that characterize dialogic syntax – it can be equally effective in bringing out the similarities and contrasts between stances in dialogic sequence. Foregrounding the relevant structural relations, diagraphs display the parallelism of elements using a representation like the following (abstracted from example 47): (48) 1 ALICE: 3 MARY:
I do n’t know if she ’d do it . I do n’t know if she would either.
A word is in order regarding the organization and significance of diagraph analysis. The diagraph is designed to represent the mapping of structured resonances across utterances. (Diagraph, from dia- ‘across’ + graph ‘mapping’, means literally ‘mapping across’.) The two most immediately relevant dimensions of the diagraph are, in simplistic terms: rows correspond to utterances,19 and columns correspond to the specific elements which resonate between different utterances, which are specified as parallel within the diagraph. To display the resonance iconically, the diagraph is formatted so that elements that resonate with each other are aligned vertically in columns. A further element that is always included in the diagraph is speaker labels, which serve to index the identity of the participants who enact the dialogic resonance in question. The recognition of a central role for participant identity in resonance relations is part of what makes dialogic syntax dialogic. Note that the diagraph is not intended to capture everything about the utterances represented, but is simply one tool to be used in conjunction with others. The diagraph focuses primarily on those elements which display significant resonance, foregrounding their relationships to each other. As such, the diagraph representation is not supposed to replace the original transcription with all its detail, but rather should remain implicitly linked to it for purposes of interpretation. The connection between the two representations of the same discourse excerpt is made explicit via such notational devices as line numbers which cross-reference
The stance triangle 161
each line of the diagraph with the corresponding line of the transcription. (For more detailed discussion and exemplification of diagraph analysis, see Du Bois 2001. For related work on structural parallelism in discourse, see Blanche-Benveniste et al. 1991; Harris 1952; Jakobson 1966, 1981; Johnstone 1994; Linell 1998; Schegloff 1996; Silverstein 1984; Tannen 1987.) Now, what does the diagraph in (48) tell us? One speaker takes a stance, then a second speaker takes a seemingly equivalent stance. But a closer examination of the diagraph makes it clear that, along with their parallels, the two stances are also fundamentally different. Though subtle, the differences turn out to be important because of what they tell us about the dialogic relations that are established through the sequential realization of stances. 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, there is a limit to their convergence. The second stance utterance ends with the word either, and this is no mere adornment. 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, in part because of the absence of the word either.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, in more or less the same way. Why is the presence of the word either so crucial here? The same issue arises in I don’t either in (46), suggesting that there is a general principle involved. In both cases, the word either in this construction serves to index a specific intersubjective relation between two speakers engaged in dialogic interaction. While space precludes full exploration of the detailed workings of this pattern here, 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). There is no other explanation for the virtually obligatory presence of either in such sentences than its role in indexing the intersubjective relationship between two stances in dialogic juxtaposition. In general terms, whenever an interactionally salient dialogic resonance arises between two stances, 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. 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. In this light, the diagraph in (48) can be interpreted in terms of who leads and who follows. The first stance (Alice’s in line 1) can be characterized as a stance lead, while the second (Mary’s in line 3) is a stance follow.21 The importance of this contrast becomes clear when we note that, despite their similar stance content, the participants mark their stances differently. Mary’s stance follow is virtually required to include
162 John W. Du Bois
the word either, which serves an intersubjective indexical function here. In contrast, Alice’s stance lead is virtually precluded, pragmatically if not grammatically, from including the intersubjective indexical either. Example (46) is like (47) in this respect: the intersubjective use of either is pragmatically required. The subjective positioning in (46) is affective (don’t like), while the subjective positioning in (47) is epistemic (don’t know), as well as modal (if) in effect. But the need to index the intersubjective relationship between a stance follow and a prior stance lead remains the same. There is an important connection between intersubjectivity and the stance act of alignment that is visible in these examples. Note that alignment, as I use the term, represents a point along a continuous scale or range of values. In contrast to common usage which forces a binary choice between a positive pole (referred to as aligned) vs. a negative pole (disaligned), the approach I favor treats alignment as continuously variable in principle. By recognizing the variability of scalar alignment we can take into account the fact that stances are aligned by subtle degrees, so that stance alignment can be relatively positive or negative – or, more precisely speaking, convergent or divergent to some degree. Alignment is in play whether the direction is convergent, divergent, or as often happens, ambiguous between the two. Thus two participants in dialogic interaction should be understood as engaging in the alignment process when they converge to varying degrees, and, by the same token, when they diverge to varying degrees. In this light, 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 engaged co-participants. In such cases, words like either (or too, as discussed below) can be said to function as intersubjective alignment markers.
6.
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. Key components include the concepts of evaluation, positioning, and alignment, as well as the sociocognitive relations of objective, subjective, and intersubjective intentionality. In developing new ways of analyzing these and other elements, we have been working to lay the groundwork for the next stage of theorizing stance. Now it is time to bring all the elements together to forge a unified framework for stance. 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 processes. On the level of action, stance is to be understood as three acts in one – a triune act, or tri-act. To the question posed at the outset (in Section 2) as to whether
The stance triangle 163
evaluation, positioning, and alignment represent three different types of stance, the view from the stance triangle suggests that they are simply different aspects of a single stance act. Rather than three separate types of stance, we interpret them as subsidiary acts of a single overarching, unified stance act. Each subsidiary act is distinguishable from the others by virtue of its own distinctive consequences, yet the three are yoked together through their integration in the dialogic stance act. The stance act thus creates three kinds of stance consequences at once. In taking a stance, the stancetaker (1) evaluates an object, (2) positions a subject (usually the self), and (3) aligns with other subjects. The following definition sums it up: (49) Stance is a public act by a social actor, achieved dialogically through overt communicative means, of simultaneously evaluating objects, positioning subjects (self and others), and aligning with other subjects, with respect to any salient dimension of the sociocultural field.
Alternatively, adopting the first-person point of view of the stancetaker as speaking subject, we can informally gloss this definition as follows: (50) I evaluate something, and thereby position myself, and thereby align with you.
Figure 1. The stance triangle
164 John W. Du Bois
The beginnings of a framework for analyzing stance are implicit in these definitions. In the rest of this section, I will further sketch out a preliminary version of this framework, along with some initial suggestions as to how it can expand the analytic reach of the stance concept. The clearest way to represent the stance model I am proposing is in the form of a triangle (Figure 1). The three nodes of the stance triangle represent the three key entities in the stance act, namely the first subject, the second subject, and the (shared) stance object. The three sides of the triangle represent vectors of directed action that organize the stance relations among these entities. While the stance triangle comprises the three subsidiary acts of evaluating, positioning, and aligning, these are not distributed evenly among the three sides, as in the expected one-to-one correspondence found in conventional triangular models. Rather, two of the three sides represent evaluative vectors directed from one of the two stance subjects toward the single shared stance object. The first evaluative vector originates from the first subject, the second from the second subject. The third side of the triangle (the vertical line on the left) represents alignment between the two subjects. Significantly, each of the three stance act vectors is relational and directed, linking two nodes of the triangle. Vectors of alignment may originate in either the first or second subject and be directed toward the other subject. For each vector of directed action in the diagram, an arrowhead points in the direction of action’s object or target. Because there are two social actors represented in the stance triangle – the first and second stance subjects – there are two tokens of each action vector type. This makes for a total of six arrowheads, corresponding to the three acts of evaluation, positioning, and alignment, doubled by the co-presence of two subjects. The stance triangle is unusual in that it depicts three stance acts for the first subject, and again the very same three stance acts for the second subject.22 And yet the acts are different, the second time around. The stance triangle provides the basis for understanding the causal and inferential linkage that may arise between the various subsidiary acts. Concomitant to evaluating a shared stance object, stancetakers position themselves. Concomitant to positioning themselves, stancetakers define alignment with each other, whether the alignment is convergent or divergent. Depending on the circumstances, it is possible to draw inferences regarding any unspecified portion of the stance triangle, as long as the rest of the triangle is known. 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 stance triangle shows how a stance utterance that specifies only one of the three vectors can allow participants to draw inferences about the others. For example, if Melissa agrees with Jan, she positions herself as taking
The stance triangle 165
the same stance (roughly speaking) as Jan, including the evaluation Jan has performed in her own prior stance. Conversely, if Melissa expresses an evaluation that is effectively the same as Jan’s previous evaluation, this allows us to infer her convergent alignment with the previous speaker. In sum, the stance triangle posits a model of the components of stance and of the organization of the relations between them, with strong implications for inferencing regarding participants’ positioning, alignment, and evaluation. Attending to the structured interrelations among the acts and entities which comprise stance allows participants, and analysts, to draw inferences by triangulating from the explicit components of stance to the implicit. What I am proposing is that the structure of dialogic action represented in the stance triangle offers a framework for analyzing the realization and interpretation of stance. Our understanding of stance is enhanced, I suggest, by taking seriously the interrelations among components of the stance act as specified in the stance triangle model. To assess these claims for the theoretical significance of the stance triangle, we need to see it in action – to test its utility in the analysis of actual instances of stance in interaction. We begin by taking a second look at an example introduced earlier (46), considering it now in light of the stance triangle: (51)
(This Retirement Bit SBC011: 444.12-446.30) 1 SAM; I don’t like those. 2 (0.2) 3 ANGELA; I don’t either.
The three entities at the nodes of the stance triangle are more or less transparently represented in this example – the first stance subject (Sam’s I), the second stance subject (Angela’s I), and the shared stance object (in Sam’s utterance, those; in Angela’s, what some would call a zero, or a deletion, representing the understanding that Angela is referring implicitly to the same referent as Sam’s those). 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). While the stance acts of evaluation and positioning are more or less evident from direct inspection of the conversational example, to see the alignment clearly it will be useful to create a diagraph: (52) 1 SAM; I do n’t like those . 3 ANGELA; I do n’t either.
Angela marks her contribution as a stance follow to Sam’s stance lead, deploying the word either in its intersubjective alignment function.
166 John W. Du Bois
To display the analysis in terms of the stance triangle more precisely, we can incorporate labels specifying which entities and actions are present in the stance utterance, and how they are expressed (or implied) in it, as in the following diagram, termed a stance diagraph.23 (53) # Speaker 1 SAM; 3 ANGELA;
Stance Positions/ Subject Evaluates I1 don’t like I2 don’t {like}
Stance Object Aligns those {those} either
As for the three stance actions, in these data, the verb specifies both the evaluation of the object and the positioning of the subject, so the two labels are combined in a single column. Angela’s use of the word either indexes alignment, taking account of the fact that Angela’s stance utterance is a stance follow which builds dialogically off of Sam’s prior stance lead. The column heading thus marks the alignment function of the word either accordingly. Note that this stance diagraph representation 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. For example, if we really want to specify how alignment is achieved in this kind of utterance, we would have to acknowledge that it’s not just using the word either that does it, but also the resonance generated through the act of reproducing words and structures of the prior speaker. There is always more to stance in dialogic interaction than can be captured in any labeled diagram as simple as this one. Nevertheless, the stance diagraph helps to make visible key aspects of the mapping between forms which resonate across utterances. This brings out the similarities, but just as important, the differences that constitute what I have called the stance differential. The stance diagraph serves as a useful intermediate stage in the analysis leading to the stance triangle, especially in foregrounding the stance act of alignment. Having parsed out the various component acts and entities of a stance exchange via the stance diagraph, one can in principle then map this analysis onto the stance triangle. The next example provides a further test case involving the three subsidiary acts of evaluation, positioning, and alignment, as well as the stance differential. (The first line of this example was analyzed above in (17) and (24).) As is clear from the prior discourse, Ken is talking specifically about going to Nicaragua. Now look at what develops next, in the subsequent exchange of stances: (54)
(Deadly Diseases SBC015: 186.540-198.625) 1 KEN; .. I would love to go:. 2 (0.3) 3 LENORE; Yeah.
The stance triangle 167
4 5 JOANNE; 6 7
(1.0) Yeah? (0.9) I want to go too.
There are interesting aspects of resonance to be found in various of the possible pairings of stance utterances that could be made here, but two resonance pairings in particular are of immediate interest: the two yeah’s (lines 3 and 5), and I would love to go : I want to go too (lines 1 and 7). First, consider Lenore’s yeah in line 3. Uttered with a final falling intonation, this amounts to an intersubjective alignment marker, displaying her convergence with Ken’s stance in line 1, as the relevant prior stance. After a one-second pause – a saliently long pause for a lively conversation like this – Joanne utters the same word as Lenore, but with a different intonation. Even though the words are identical, a stance differential is created nonetheless, primarily by intonation but also by other factors such as voice quality and sequential placement. Where Lenore had a terminative intonation contour (realized as a final fall) on Yeah, Joanne uses a rising appeal contour, suggesting a questioning of, or at least an ambivalence toward, Lenore’s more definite yeah.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, to which Ken and Lenore are both by now on-record adherents. After another substantial pause, Joanne goes on in line 7 to commit to a sort of stance follow. But she withholds the word love, choosing want instead, a presumable downgrade of the level of commitment to the proposition at issue.25 This sheds light on her lack of commitment as implied by her use of a questioning intonation for what could have been, with different intonation, a strong marker of convergent alignment. The crucial resonance relations in this exchange are highlighted in the following diagraph of Ken and Joanne’s utterances: (55) 1 KEN; I would love to go . 7 JOANNE; I want to go too.
While not realized in immediately successive utterances, the analogy between the two stance utterances (I would love to go : I want to go too) is unmistakable – sufficiently transparent to be recognizable across a few intervening utterances. Although it might appear that Joanne is saying more or less “the same thing” as Ken, the differences which constitute the stance differential are actually quite crucial, as the diagraph above suggests. To create a more explicit representation labeling the relevant stance entities and actions, we recast (55) as a stance diagraph:
168 John W. Du Bois
(56) # Speaker 1 KEN; 7 JOANNE;
Stance Positions/ Subject Evaluates I1 would love I2 want
Stance Object Aligns to go {to Nicaragua} to go {to Nicaragua} too
This diagram shows clearly how Ken performs a stance lead, and Joanne a stance follow. Joanne explicitly registers the difference between her follow and his lead via the word too, which is not omissible in this context without pragmatic anomaly. The evidence from many such cases makes it clear that when a stance follow is juxtaposed to a dialogically resonant prior stance lead, then, given certain welldefined conditions,26 the intersubjectivity involved must be registered obligatorily, using the appropriate linguistic index: too for positive utterances, either for negative utterances.27 Although the role of intersubjectivity in language usually remains implicit or is subtly expressed (e.g., via prosody or sequential placement), sometimes it is realized overtly in the tangible form of a specific word, as in this case. As this and earlier examples attest, the indexical function of too and either becomes a valuable diagnostic for the individual speaker’s obligatory engagement with dialogically constructed intersubjectivity. One further point about the stance triangle calls for comment here. In the present analysis of stance, the shared stance object obviously plays a critical role, binding the subjectivities of dialogic co-actors together, thereby articulating an intersubjective relation between them. 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. Yet the argument can be made that the stance triangle applies even in such less-than-transparent cases. While this question deserves a more extended response than can be presented here, a word of commentary may give some idea of what kind of answer will be required. It has been claimed that all meaningful use of human language, from the age of about one year, presupposes shared orientations, for example toward a word’s referent (Tomasello 1999; Tomasello et al. 2005; see also Hobson 1993). It will be an important task for future research to show how the stance triangle extends naturally to incorporate such observations. In the meantime, in cases where it may not be obvious that the full stance triangle is in play, it is usually possible to break the triangle down into its component vectors (e.g., an individual stance vector constituting the subject-object evaluative relation), and thereby to achieve an insightful, if partial, analysis of stance. At the outset (Section 2) we considered whether we could set up a distinction between different types of stance. Among the more promising candidates were evaluation, positioning, and alignment. But should we consider these as distinct types, or merely different facets, of stance? The answer may hinge on whether
The stance triangle 169
they regularly occur separately, in complementary distribution, or whether they rather co-occur as different aspects of a single act of stancetaking. The evidence we have seen suggests that they are not so much alternatives as different facets of a more general phenomenon of stance. This is what the stance triangle tries to represent. Since its first introduction (Du Bois 2002a, 2002b) the stance triangle has come to be used in a diverse and informative body of work on stance in interaction (e.g., Haddington 2004, 2005, 2006; Kärkkäinen 2006; Takanashi 2004; see also the papers by Kärkkäinen, Haddington, Keisanen, and Rauniomaa in this volume). The stance triangle has proved well suited to framing effective research questions, without imposing predetermined limitations on what the researcher can undertake to investigate about stance.
7.
Discussion
The conception of stance we need is one capable of being situated within a larger mediating framework for linguistic action and interpretation, which is itself grounded in the dialogic dimensions of sociocognitive relations, interactional collaboration, and shared responsibility between conversational co-participants. To come to terms with the complexity of the stance act and its interpretive frame, it is necessary to articulate a systematic approach to understanding what stance is. To do this, we need a way to represent how stance works. We need this if we want to be able to frame questions about how discourse participants achieve stance, and about how the multiplex meanings and consequences of stance play out in the public space of interaction. The framework I have proposed is encapsulated in a triangle, but it must not be forgotten that there is a larger interpretive apparatus that subtends this geometric metaphor. The stance triangle is built on certain basic assumptions about what is needed to constitute a stance, which I have elaborated in some detail in this paper. I have argued that stance can be analyzed, in its fundamental structure, as a single unified act encompassing three subsidiary acts – in effect, a triune act, or tri-act. I define stance as a public act by a social actor, achieved dialogically through overt communicative means (language, gesture, and other symbolic forms), through which social actors simultaneously evaluate objects, position subjects (themselves and others), and align with other subjects, with respect to any salient dimension of value in the sociocultural field. I have argued for a particular configuration of actors and actions as the defining feature of stance. Key to this configuration is a set of three entities (first subject, second subject, stance object) and a set of three actions (evaluation, positioning, alignment). The analysis of stance in terms of these elements lays the basic foundations on which the
170 John W. Du Bois
stance triangle is built. But there is more to the stance triangle than this. Crucially, the stance triangle specifies the effective relationships between entities and acts through vectors of stance action. On another level, sociocognitive relations link the subject and object of stance relations, and map one subject-object vector to another to constitute the intersubjective relation. 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. What consequences flow from the triangular model of stance? Here I will point to just a few of the implications that arise. The parallelism of two of the three stance vectors allows us to analyze the phenomenon of alignment, based on an analogy which is partly structural and partly functional. This parallelism may in turn invite analogical inferences (Itkonen 2005; Jakobson 1966; Silverstein 1984). The stance triangle provides a framework for understanding the sociocognitive relations (objective, subjective, intersubjective) that are present in all dialogic interaction, and tries to clarify how these relations are constituted through the stance acts of evaluating objects (objective), positioning subjects (subjective), and aligning with other subjects (intersubjective). Especially in cases involving what I have called the stance differential, the triangle provides leverage to analyze the fine calibration of convergent and divergent alignment in the stancetakers’ positioning of themselves relative to others.28 In the world of theory-making, there are many triangles. Why one more? But this is not your average triangle. 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. Here, on the other hand, two of the three points of the triangle are used to represent what amounts to the same thing twice – two stance subjects. This is not done to flout the principle of economy, but because it displays the interchangeability of perspectives (Mead 1934; Schutz 1962) of dialogic stance partners who alternately fill the roles of speaker and hearer, of first subject and second subject, as the polarity of co-action cycles rapidly via the dynamics of dialogic exchange. Unlike many triangles which remain in the realm of the theoretical, this triangle is meant to be used. The idea is that its general architecture underlies the actual practices of realizing stances and negotiating their significance in particular events of language use. To the extent that it articulates a predictable framework for action, the stance triangle is available to participants as a resource for organizing their evaluative actions on any specific occasion. Looked at from the analyst’s perspective, this makes it relevant to describing what is happening in actual instances of stance in interaction. Appropriately deployed, the stance triangle can clarify the array of entities and sociocognitive relations that are activated, constituted, and brought into relation to each other by a particular stance action. In addition, by
The stance triangle 171
allowing us to triangulate between one participant’s observable (if partial) stance actions and those of their dialogic partner, the stance triangle provides a way to apprehend the regularity of stance inferencing that brings systematicity to the consequences of stance.
8.
Conclusions
Stance is not something you have, not a property of interior psyche, but something you do – something you take. Taking a stance cannot be reduced to a matter of private opinion or attitude. Using the language of Wittgenstein (1953) we might say: There are no private stances. We deploy overt communicative means – speech, gesture, and other forms of symbolic action – to arrive at a dialogic achievement of stance in the public arena. Stance can be imagined as a kind of language game in Wittgenstein’s sense, which is to say, it unfolds within a recognized framework for interpreting action. To realize stance dialogically means to invoke a shared framework for co-action with others. Stance must be intelligible within that framework, as it comes into existence in its natural environment of dialogic interaction. 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. Stance is an activity built for two (or more). As we maneuver within the constantly shifting field of stances, we find that even our own stance must be enacted collaboratively. Through joint and several acts we engage in the activity of stance, invoking relevant components of the stance frame as we both shape and respond to the multiplex consequences which flow from our actions. 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. Participants use their knowledge of the elements, actions, and vectors of stance – as described here by the stance triangle – to project the multiplex consequences of their own and their partners’ unfolding stances. Stance is best understood in terms of the general structure of the evaluative, positioning, and aligning processes that organize the enactment of stance, rather than as a catalog of the contents of stance, or even – as important as these may be – of the sociocultural value categories that are referenced by 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.29 Stance always combines elements of generality and specificity; but while the stance triangle as a theoretical object is general by design, it is also intended to frame the concrete analysis of specific stances. The general principles governing the stance framework, including principles of dialogic
172 John W. Du Bois
co-action and the intersubjective organization of sociocognitive relations, impinge on the realization and interpretation of individual stances. 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. Stance utterances gain added levels of significance through their juxtaposition with other stance utterances. Much of the dialogic quality of stance comes from the way a present stance may resonate with a prior stance, as resonance across stances shapes the sociocognitive alignment between speakers, and thus helps frame intersubjectivity. The method used here to analyze resonance across successive stance utterances is dialogic syntax (Du Bois 2001), and specifically the diagraph. The diagraph representation depicts the precise resonance relations between two stance utterances, allowing us to probe in finer detail the specific reciprocal relations between sub-portions of the utterances mapped in the diagraph. Given an explicit set of mapping relations between counterpart elements across paired stance utterances, we can go on to draw detailed inferences regarding aspects of the stance differential between the two participants, including convergences and divergences of epistemology, affect, and other regular concerns of stancetakers. The potential is extremely rich for analytical and theoretical cross-fertilization between the methods of the stance triangle and the diagraph. 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. From a dialogic perspective, no stance stands alone. Each stance is already specific with respect to, for example, the participants it indexes, the objects it evaluates, and the dimensions of sociocultural value it invokes. But beyond this, any stance realization is capable of extending its particularity further by entering into unique relationships with other stance realizations, within particular configurations of dialogic exchange. This sets up dialogic juxtapositions of stances which may foreground the structural analogy between pairs of successive stances, yielding a combinatorial explosion of resonance relations, as these are activated across the paired stance utterances. Yet the resources available to participants for interpreting such a diversity of stances are sufficient to the task, while remaining constant and general. The stance triangle – including its component vectors – provides a general account of the framing processes which adapt to both the particularities of the individual stance act and its unique configuration of dialogic-sequential development. The tension between generality and particularity, characteristic of all culturally framed social practice, is necessarily reflected in the interpretation and realization of stance. To resolve this tension, or simply to embrace it, constitutes an ongoing challenge for any theory of stance, including the present one.
The stance triangle 173
Stance is undeniably complex. There is no avoiding the rapid ramification of meaning and consequence as stances emerge from the flux of dialogic interaction. Still, 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. What is it that sets stance apart as a research agenda, 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. This characterization, while by no means complete, unites three key aspects of social life: act, responsibility, and value. The act of stance is performed in the public space of dialogic interaction, where it inherits the peculiar contingencies of all dialogic action, both influencing and being influenced by the co-actions of others. Despite its apparent dependency as a local symbolic action delimited by the uptake of other individuals, stance may be among the most broadly consequential of social actions in its cumulative effects – an act whose force can be measured by its effective penetration into virtually all domains of sociocultural life. Responsibility, the second element in this capsule cameo of stance, comes with ownership. In the dialogic shop of stances, there’s a rule: If you take it, you own it. Responsibility for the stance act is serious business, with potentially profound consequences for the relationships of stance actors with their dialogically co-responsible partners and with their expanding networks of social relations along wider horizons of time and space. Because stance responsibility is of interest to all within the stancetaking community of discourse, the question of who took which stance is perennially salient, is remembered over time, and counts as negotiable coin in the currency of reported discourse. Through its open-ended social circulation, stance (and meta-stance narratives) can only expand its role in the broader calculus of social meaning. Third, value is what stance is all about – literally. Stance always invokes, explicitly or implicitly, presupposed systems of sociocultural value, while at the same time contributing to the enactment and reproduction of those systems. Social actors are accountable for how they manage and indeed reshape the systems of social value on which we all depend. In sum, ownership of stance is the glue that binds the stance act together with actor responsibility and sociocultural value, so that all is linked to a social actor with a name, a history, an identity. As players in the language-game of stance, we’ve all got some skin in the game. We make it our business to know where the other players stand, who they stand with, and where they’re headed. And we care about the state of the game, too: how it is played, who plays it well and fairly, 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. We are left with a view of the stance act as, perhaps, the smallest unit of social action. Such a fundamental status will be justified if stance can be shown to
174 John W. Du Bois
bind together the minimum structures necessary to attain the force of social action. This is what the stance triangle aims to represent: the minimum structure of stance as dialogic action. In depicting the co-participants’ joint evaluative orientation to a shared stance object, 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. To be sure, it may seem counterintuitive to locate contestation within intersubjectivity, rather than to simply demand mute agreement to the normative assumptions of social reality as the conventional prerequisite to communication. But the evidence from stance in interaction is clear: convergence and divergence of evaluative alignment are equally at home in the dialogic engagement of co-participants. If the stance triangle is to have analytic value in the end, it must come from striving to represent at once the unity, and the ambivalence, of stance as it emerges in dialogic interaction.
Notes 1. My understanding of stance as presented here has developed during the course of a series of presentations at conferences and symposia, including at the meetings of the American Anthropological Association, New Orleans (2002); Volkswagen Foundation conference on Rhetoric Culture, Johannes Gutenberg University, Mainz (2002); Kroeber Anthropological Society, Berkeley (2003); a lecture series on stance and dialogic syntax, Oulu, Finland (2003); LangNet Symposium, Oulu (2003); Fulbright lecture, Södertörn University, Södertörn, Sweden (2004); and, finally, culminating in a presentation on “The Intersubjectivity of Interaction” at the 10th Rice Symposium, on Stancetaking in Discourse, at Rice University, Houston (2004). My research 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), University of Verona (Du Bois 2004). In addition, 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, Interaction, and Social Organization (LISO) at the University of California, Santa Barbara. I have benefited greatly from discussion with all these audiences, for which I am grateful. I am especially thankful for comments on earlier versions of this work from Mira Ariel, Mary Bucholtz, Patricia Clancy, Robert Englebretson, Rachel Giora, Pentti Haddington, William Hanks, John Haviland, Susan Hunston, Cornelia Ilie, Adam Jaworski, Elise Kärkkäinen, Tiina Keisanen, Amy Kyratzis, Gene Lerner, John Lucy, Mirka Rauniomaa, Geoffrey Raymond, Joanne Scheibman, Robin Shoaps, Michael Silverstein, Hiroko Takanashi, Sandra Thompson, and an anonymous reviewer for the present volume. 2. See the Appendix for transcription conventions. Note that some of the transcriptions have been simplified for the sake of clarity. For example, square brackets for overlapping speech have been omitted when the cited example doesn’t include the other half of the pair of overlapping utterances. For full transcription details, the original sources may be consulted, as described in the following note. 3. Most of the examples in this paper are taken from conversations in the Santa Barbara Corpus of Spoken American English (Parts 1 and 2). The source citation gives the title of the dis-
The stance triangle 175
course in italics (e.g., Lambada), followed by the identifying number of the discourse in the Santa Barbara Corpus of Spoken American English (e.g., SBC002). The final portion of the citation for each excerpt consists of two numbers, representing respectively the start time and end time in seconds. With this information it is possible for interested readers to listen to the appropriate portion of the relevant audio file (e.g., SBC002.WAV) by accessing the Santa Barbara Corpus as published on CD and DVD (Du Bois et al. 2000; Du Bois et al. 2003), or on the internet. Further information on contents of and access to the Santa Barbara Corpus of Spoken American English is available at http://www.linguistics.ucsb.edu/research/sbcorpus.html. 4. The predicate amazed can be understood as incorporating, in addition to its affective dimension, a salient epistemic dimension as well. 5. The source of this example, cited as LSAC, is the Longman Spoken American Corpus. The LSAC is a five million word corpus of spoken American English conversation, recorded under my direction by researchers from UC Santa Barbara. 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 LSAC was developed under contract to Longman publishers. 6. The names of speakers are arbitrarily assigned pseudonyms. The following is a key to the original labels in the LSAC corpus for the speakers cited here: in LSAC 1396-01, CORA= and LESLIE=; in LSAC 1296-02, KIM= and PAT=. 7. Alternatively, agreement can be articulated in relation to the proposition expressed by a coparticipant – the stance content of a prior utterance, as denoted, for example, by the word that in I agree with that. 8. Names used for speakers are pseudonyms. 9. In explicating the interpretation of a stance in its dialogic context, I have sometimes ventured a paraphrase of the “content” of the stance as it has emerged from the successive contributions of several participants. The problem is that the paraphrase usually comes across as too explicit or too literal in character. A summary paraphrase cannot do justice to the specific quality of the actual stance as it emerges dialogically in discourse. 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. The challenge of precisely and perspicuously describing stances that emerge in dialogic interaction is a serious one, which deserves more attention than we can give it here. 10. The content of Kim’s stance is constituted as no more or no less than any other school. (This phrasing is admittedly rather opaque in itself, but unfortunately this cannot be clarified without expending a lot more time and space than is warranted for the present example; nevertheless, the general point about alignment should be sufficiently clear.) This is to be distinguished from the final portion of the utterance (is the way I see it), which pertains narrowly to Kim’s framing of the proposition. This is a parenthetical framing of self-positioning, not part of the stance proposition as such, so the framing phrase is interpreted as outside the scope of Pat’s agreeing move. Alignment markers like the verb agree typically pick out just the stance content or stance proposition, leaving the previous speaker’s metalinguistic framing or self-positioning out of the calculus of stance alignment. How this works in detail on a structural level is a rich topic for further analysis. 11. To be sure, Melissa’s use of the modal should introduces a subtle shift which slyly mitigates the directive force of Jan’s original unmitigated imperative.
176 John W. Du Bois
12. For a rich treatment of agreement from the perspective of conversation analysis, see Heritage (2002) and Heritage and Raymond (2005). 13. In most of the examples presented in this paper, the stance subject can be taken to be more or less equivalent to the speaking subject. But things are not always so simple, as is clear from the sophisticated analyses of the sujet de l’énonciation by the cited authors, as well as from characterization of the multivocality of certain kinds of utterances (Bakhtin [1934] 1981; Goffman 1981; cf. Agha 2005). While the complex case of multivocality is very interesting to consider from the perspective of the stance triangle, this must await a separate treatment. 14. In other words, they are intentional predicates, in the philosophical sense (Searle 1983). 15. There are additional interesting issues here having to do with the incremental realization of Randy’s turn, the timing of pauses, and what could be considered a lengthy delay on Lance’s part in coming in with I’m glad. But these issues are largely orthogonal to the present discussion. 16. How this stance develops over time and across speakers is of some interest. Jeff ’s affective stance utterance in line 4 builds off of Jill’s prior yes in line 2, which in turn endorses Jeff ’s own setting of the question in line 1. This can be considered an instance of other-positioning (line 1). 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), with varying degrees of impositive force. Given that this often seems to provide the best analysis for, e.g. ordinary questions, the phenomenon of other-questioning is surprisingly commonplace. (For application of this concept to questioning by interviewers in television news formats, see Haddington 2005, and Haddington this volume.) 17. This number represents the intonation unit number (or line number), counting from the beginning of the published transcription. 18. This is not to suggest that an epistemic paraphrase tells the whole story of Dan’s I don’t know, 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, see note 16) that Jennifer presents him with. But that’s another story. 19. More precisely, it is intonation units (Chafe 1993; Du Bois et al. 1993) that generally define the rows (or “strands”) of a diagraph. Based on my research (Du Bois 2001), intonation units represent the most salient and productive unit for dialogic mapping in conversation. Where information about the intonation unit is unavailable, the most viable alternative for identifying the rows of a diagraph would generally be the clause. 20. One can try to imagine special circumstances in which the pronoun I is given heavy contrastive stress (suggesting that it is only I, but not you, who don’t know if she’d do it), but this would be quite unusual. 21. My terms here are modeled on the notion of gaze follow(ing) (Tomasello 1999: 62–67; Tomasello et al. 2005), which is parallel to the phenomenon of stance follow in interesting ways. For a related notion, see discussion in conversation analysis of the interactional negotiation of turn status as epistemically “first” (=lead) or “second” (=follow) (Heritage and Raymond 2005). 22. Note that the self-positioning act defines a vector which emanates from a subject and reflects back on that same subject. To explicitly represent both source and target of the reflexive vector of self-positioning would involve a circular arrow originating in the subject and reflecting back on itself. Rather than display such an arrow in this simple diagram of the stance triangle, we show only the head of the arrow for the self-positioning vector, as it reflects back on the stance subject. This reflection can be seen as being triggered by a sort of “blowback” from the
The stance triangle 177
subject’s act of evaluating the stance object. (Other-positioning requires a significantly more elaborate notation, and thus is not represented in Figure 1.) 23. The stance diagraph incorporates aspects of the stance diagram (see discussion of examples (30)–(31)), such as the labeling of stance roles (stance subject, stance object) and functions (evaluation, positioning, alignment), combined with aspects of the diagraph, such as the inclusion of multiple lines representing dialogically resonating utterances, with vertical alignment of elements to iconically display which are resonating with which. 24. 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. : Yeah?), it will not be diagrammed here for reasons of space. 25. The passage continues with several further developments regarding this stance negotiation, but a fuller analysis of these events must be reserved for another occasion. 26. While it is beyond the scope of this article to lay out all the conditions governing the use of too and either, one key factor can be mentioned here: the use of subjective intentional stance predicates such as like, love, know, think, and want (Du Bois 2004). 27. This is not to say that too and either always function as diagnostics for intersubjectivity. Aside from their high frequency use in marking intersubjective pragmatic relations, 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)). Conversely, there are many intersubjective contexts in which these particular indexical forms do not appear. While too and either should be recognized as powerful diagnostics for intersubjectivity when their conditions of use are applicable, the analysis of intersubjectivity in language must draw on a wide variety of additional tools, including diagraph analysis. 28. The stance triangle bears an important relation to the systems of sociocultural value that stances invoke and reproduce, as social actors position themselves and evaluate entities with respect to specific values along any socially salient dimension of the sociocultural field. 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, but one which represents a prime topic for further research. 29. While the inclusion of the shared stance object – seemingly a specialized property of certain kinds of stance exchanges – might seem to limit the general applicability of the stance triangle, from a dialogic perspective it can be argued that a shared orientation to a stance object is a general property, not only of stance acts but of the use of language in general (see the discussion at the end of Section 6). But that is a long story, and must be reserved for another occasion.
References Agha, A. 2005. “Voice, footing, enregisterment.” Journal of Linguistic Anthropology 15: 38–59. Bakhtin, M. [1934]1981. The Dialogic Imagination: Four Essays by M.M. Bakhtin. Austin, TX: University of Texas Press. Benveniste, E. 1971. “Subjectivity in language.” In Problems in General Linguistics, E.M. Meek (trans), 223–230. Coral Gables, FL: University of Miami Press.
178 John W. Du Bois
Berman, R.A., Ragnarsdóttir, H. and Strömqvist, S. 2002. “Discourse stance.” Written Language and Literacy 5: 255–290. Berman, R.A. 2005. “Introduction: Developing discourse stance in different text types and languages.” Journal of Pragmatics 37: 105–124. Besnier, N. 1993. “Reported speech and affect on Nukulaelae atoll.” In Responsibility and Evidence in Oral Discourse, J.H. Hill and J.T. Irvine (eds.), 161–181. Cambridge: Cambridge University Press. Blanche-Benveniste, C., Bilger, M., Rouget, C. and van den Eynde, K. 1991. Le français parlé: Études grammaticales. Paris: Éditions du Centre National de la Recherche Scientifique. Bruner, J. 1986. Actual Minds, Possible Worlds. Cambridge, MA: Harvard University Press. Chafe, W.L. 1993. “Prosodic and functional units of language.” In Talking Data: Transcription and Coding in Discourse Research, J.A. Edwards and M.D. Lampert (eds.), 33–43. Hillsdale, NJ: Lawrence Erlbaum. Chafe, W.L. 1994. Discourse, Consciousness, and Time: The Flow and Displacement of Conscious Experience in Speaking and Writing. Chicago: University of Chicago Press. Clancy, P.M. 1999. “The socialization of affect in Japanese mother-child conversation.” Journal of Pragmatics 31: 1397–1421. Clift, R. 2006. “Indexing stance: Reported speech as an interactional evidential.” Journal of Sociolinguistics 10: 569–595. Conrad, S. and Biber, D. 2000. “Adverbial marking of stance in speech and writing.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, S. Hunston and G. Thompson (eds.), 56–73. Oxford: Oxford University Press. Culioli, A. 1990. Pour une linguistique de l’énonciation. Paris: Ophrys. Davies, B. and Harré, R. 1990. “Positioning: Conversation and the production of selves.” Journal for the Theory of Social Behavior 20: 43–63. Du Bois, J.W. 1991. “Transcription design principles for spoken discourse research.” Pragmatics 1: 71–106. Du Bois, J.W. 2001. Towards a Dialogic Syntax. Ms., Department of Linguistics, University of California, Santa Barbara. Du Bois, J.W. 2002a. “Stance and consequence.” Paper presented at Annual Meetings of the American Anthropological Association, New Orleans, LA, November 20–24. Du Bois, J.W. 2002b. “Dialogic rhetoric: Stance and intersubjectivity in the syntax of engagement.” Paper presented at Volkswagen Foundation conference on “Rhetoric Culture,” Johannes Gutenberg University, Mainz, July 4–7. Du Bois, J.W. 2004. “Searching for intersubjectivity: ‘Too’ and ‘either’ in stance alignment.” Paper presented at 25th Conference of the International Computer Archive of Modern and Medieval English (ICAME), University of Verona. Du Bois, J.W., Schuetze-Coburn, S., Cumming, S. and Paolino, D. 1993. “Outline of discourse transcription.” In Talking Data: Transcription and Coding in Discourse Research, J.A. Edwards and M.D. Lampert (eds.), 45–89. Hillsdale, NJ: Erlbaum. Du Bois, J.W., Chafe, W.L., Meyer, C. and Thompson, S.A. 2000. Santa Barbara Corpus of Spoken American English, CD, Part 1. Philadelphia: Linguistic Data Consortium. Du Bois, J.W., Chafe, W.L., Meyer, C., Thompson, S.A. and Martey, N. 2003. Santa Barbara Corpus of Spoken American English, CD, Part 2. Philadelphia: Linguistic Data Consortium. Ducrot, O. 1972. Dire et ne pas dire. Paris: Hermann. Ducrot, O. 1984. Le dire et le dit. Paris: Minuit. Geertz, C. 1973. The Interpretation of Cultures. New York: Basic Books.
The stance triangle 179
Goffman, E. 1981. “Footing.” In Forms of Talk. Philadelphia: University of Pennsylvania Press. Goodwin, C. and Goodwin, M.H. 1992. “Assessments and the construction of context.” In Rethinking Context: Language as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 147–189. Cambridge: Cambridge University Press. Goodwin, M.H. 2006. The hidden life of girls: Games of stance, status, and exclusion. Oxford: Blackwell. Gumperz, J.J. 1992. “Contextualization and understanding.” In Rethinking Context: Language as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 229–252. Cambridge: Cambridge University Press. Haddington, P. 2004. “Stance taking in news interviews.” SKY Journal of Linguistics 17: 101– 142. Haddington, P. 2005. The Intersubjectivity of Stance Taking in Talk-in-Interaction. Ph.D. dissertation. English Department, University of Oulu: Oulu University Press. Haddington, P. 2006. “The organization of gaze and assessments as resources for stance taking.” Text and Talk 26: 281–328. Hanks, W.F. 1990. Referential Practice: Language and Lived Space Among the Maya. Chicago: University of Chicago Press. Harris, Z.S. 1952. “Discourse analysis.” Language 28: 1–30. Haviland, J.B. 1991. “‘Sure, sure’: Evidence and affect.” Text 9: 27–68. Haviland, J.B. 1996. “Projections, transpositions, and relativity.” In Rethinking Linguistic Relativity, J.J. Gumperz and S.C. Levinson (eds.), 271–323. Cambridge: Cambridge University Press. Heritage, J. 2002. “Oh-prefaced responses to assessments: A method of modifying agreement/disagreement.” In The Language of Turn and Sequence, C.E. Ford, B.A. Fox and S.A. Thompson (eds.), 196–224. Oxford: Oxford University Press. Heritage, J. and Raymond, G. 2005. “The terms of agreement: Indexing epistemic authority and subordination in talk-in-interaction.” Social Psychology Quarterly 68: 15–38. Hill, J.H. and Irvine, J.T. (eds.). 1993. Evidence and Responsibility in Oral Discourse. Cambridge: Cambridge University Press. Hobson, R.P. 1993. Autism and the development of mind. Hillsdale, NJ: Erlbaum. Hunston, S. and Sinclair, J.M. 2000. “A local grammar of evaluation.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, S. Hunston and G. Thompson (eds.), 74–101. Oxford: Oxford University Press. Hunston, S. and Thompson, G. 2000. Evaluation in Text: Authorial Stance and the Construction of Discourse. Oxford: Oxford University Press. Itkonen, E. 2005. Analogy as Structure and Process. Amsterdam: John Benjamins. Jakobson, R. [1957]1990. “Shifters and verbal categories.” In On Language: Roman Jakobson, L.R. Waugh and M. Monville-Burston (eds.), 386–392. Cambridge, MA: Harvard University Press. Jakobson, R. 1966. “Grammatical parallelism and its Russian facet.” Language 42: 398–429. Jakobson, R. 1981. “Poetry of grammar and grammar of poetry.” In Selected Writings, Vol. 3: Poetry of Grammar and Grammar of Poetry, 87–97. The Hague: Mouton de Gruyter. Jefferson, G. 2004. “Glossary of transcript symbols with an introduction.” In Conversation Analysis: Studies from the First Generation, G.H. Lerner (ed.), 13–31. Amsterdam: John Benjamins. Johnstone, B. (ed.). 1994. Repetition in Discourse: Interdisciplinary Perspectives. Norwood, NJ: Ablex.
180 John W. Du Bois
Kärkkäinen, E. 2003a. “‘Is she vicious or dense?’: Dialogic practices of stance taking in conversation.” In Santa Barbara Papers in Linguistics 12: Recent Studies in Empirical Approaches to Language, T. Nakayama, T. Ono and H. Tao (eds.), 47–65. Santa Barbara: University of California, Santa Barbara. Kärkkäinen, E. 2003b. Epistemic Stance in English Conversation: A Description of its Interactional Functions, with a Focus on ‘I think’. Amsterdam: John Benjamins. Kärkkäinen, E. 2006. “Stance taking in conversation: From subjectivity to intersubjectivity.” Text & Talk 26: 699–731. Kidwell, M. and Zimmerman, D.H. 2006. “‘Observability’ in the interactions of very young children.” Communication Monographs 73: 1–28. Kockelman, P. 2004. “Stance and subjectivity.” Journal of Linguistic Anthropology 14: 127–150. Labov, W. and Waletzky, J. 1967. “Narrative analysis: Oral versions of personal experience.” In Essays on the Verbal and Visual Arts, J. Helm (ed.), 12–44. Seattle: University of Washington Press. Langacker, R.W. 1987. Foundations of Cognitive Grammar, Vol. 1: Theoretical Prerequisites. Stanford, CA: Stanford University Press. Langacker, R.W. 1991. Concept, Image, and Symbol: The Cognitive Basis of Grammar. Berlin: Mouton de Gruyter. Lemke, J.L. 1998. “Resources for attitudinal meaning: Evaluative orientations in text semantics.” Functions of Language 5: 33–56. Linde, C. 1997. “Evaluation as linguistic structure and social practice.” In The Construction of Professional Discourse, B.-L. Gunnarsson, P. Linell and B. Nordberg (eds.), 151–172. New York: Addison Wesley Longman. Linell, P. 1998. Approaching Dialogue: Talk, Interaction, and Contexts in Dialogical Perspectives. Amsterdam: John Benjamins. Lucy, J.A. (ed.). 1993. Reflexive Language: Reported Speech and Metapragmatics. Cambridge: Cambridge University Press. Macken-Horarik, M. and Martin, J.R. 2003. Negotiating heteroglossia: Social perspectives on evaluation (Special Issue). Text 23. Martin, J.R. 2000. “Beyond exchange: APPRAISAL systems in English.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, S. Hunston and G. Thompson (eds.), 142–175. Oxford: Oxford University Press. Maynard, S.K. 1993. Discourse Modality: Subjectivity, Emotion, and Voice in The Japanese Language. Amsterdam: John Benjamins. Mead, G.H. 1934. Mind, Self, and Society. Chicago: University of Chicago Press. Moore, C., and Dunham, P.J. (eds.). 1995. Joint attention: Its origins and role in development. Hillsdale, NJ: Erlbaum. Ochs, E. 1988. Culture and Language Development: Language Acquisition and Language Socialization in a Samoan Village. Cambridge: Cambridge University Press. Ochs, E. 1996. “Linguistic resources for socializing humanity.” In Rethinking Linguistic Relativity, J.J. Gumperz and S.C. Levinson (eds.), 407–437. Cambridge: Cambridge University Press. Peirce, C.S. [1885]1933. “Three kinds of signs.” In Collected Papers of Charles Sanders Peirce, Vol. 3, C. Hartshorne and P. Weiss (eds.), Cambridge, MA: Harvard University Press. Pomerantz, A. 1984. “Agreeing and disagreeing with assessments: Some features found in preferred/dispreferred turn shapes.” In Structures of Social Action, J.M. Atkinson and J. Heritage (eds.), 57–101. Cambridge: Cambridge University Press.
The stance triangle 181
Recanati, F. 1989. “The pragmatics of what is said.” Mind and Language 4: 295–328. Sacks, H. 1992. Lectures on Conversation. Oxford: Blackwell. Schegloff, E.A. 1996. “Confirming allusions: Toward an empirical account of action.” American Journal of Sociology 102: 161–216. Scheibman, J. 2002. Point of View and Grammar: Structural Patterns of Subjectivity in American English Conversation. Amsterdam: John Benjamins. Schutz, A. 1962. Collected Papers I. The Problem of Social Reality. The Hague: Nijhoff. Searle, J.R. 1983. Intentionality: An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press. Shoaps, R.A. 2002. “‘Pray earnestly’: The textual construction of personal involvement in Pentecostal prayer and song.” Journal of Linguistic Anthropology 12: 34–71. Shoaps, R.A. 2004. Morality in Grammar and Discourse: Stance-taking and the Negotiation of Moral Personhood in Sakapultek (Mayan) Wedding Counsels. Ph.D. dissertation. Linguistics Department, University of California, Santa Barbara. Silverstein, M. 1976. “Shifters, verbal categories, and cultural description.” In Meaning in Anthropology, K.H. Basso and H. Selby (eds.), 11–55. Albuquerque: School of American Research. Silverstein, M. 1984. “On the pragmatic poetry of prose: Parallelism, repetition, and cohesive structure in the time course of dyadic conversation.” In Meaning, Form, and Use in Context: Linguistic Applications, D. Schiffrin (ed.), Washington, DC: Georgetown University Press. Silverstein, M. 2001. “The limits of awareness.” In Linguistic Anthropology: A Reader, A. Duranti (ed.), 382–401. Oxford: Blackwell. Sperber, D. and Wilson, D. 1995. Relevance: Communication and Cognition. Oxford: Blackwell. Takanashi, H. 2004. The Interactional Co-Construction of Play in Japanese Conversation. Ph.D. dissertation. Linguistics Department, University of California, Santa Barbara. Tannen, D. 1987. “Repetition in conversation: Towards a poetics of talk.” Language 63: 574– 605. Thompson, G. and Hunston, S. 2000. “Evaluation: An introduction.” In Evaluation in Text: Authorial Stance and the Construction of Discourse, S. Hunston and G. Thompson (eds.), 1–27. Oxford: Oxford University Press. Thompson, S.A. and Mulac, A. 1991. “A quantitative perspective on the grammaticization of epistemic parentheticals in English.” In Approaches to Grammaticalization, Vol. 2: Focus on Types of Grammatical Markers, E.C. Traugott and B. Heine (eds.), 313–329. Amsterdam: John Benjamins. Thompson, S.A. and Hopper, P.J. 2001. “Transitivity, clause structure, and argument structure: Evidence from conversation.” In Frequency and the Emergence of Linguistic Structure, J.L. Bybee and P.J. Hopper (eds.), 27–60. Amsterdam: John Benjamins. Tomasello, M. 1999. The Cultural Origins of Human Cognition. Cambridge, MA: Harvard University Press. Tomasello, M., Carpenter, M., Call, J., Behne, T., and Moll, H. 2005. “Understanding and sharing intentions: The origins of cultural cognition.” Behavioral and Brain Sciences 28: 675–735. Traugott, E.C. 1989. “On the rise of epistemic meanings in English: An example of subjectification in semantic change.” Language 65: 31–55. Voloshinov, V.N. [1929]1973. Marxism and the Philosophy of Language. Cambridge, MA: Harvard University Press.
182 John W. Du Bois
Vygotsky, L.S. 1986. Thought and Language. Cambridge, MA: MIT Press. Wittgenstein, L. 1953. Philosophical Investigations. New York: MacMillan.
Appendix: Transcription conventions The transcription symbols and conventions used in this paper are largely as in Du Bois et al. (1992), although there have been a number of significant updates in my more recent transcription practice. The most relevant symbols are given below. (For further details see http://www. linguistics.ucsb.edu/projects/transcription/representing). Meaning intonation unit speaker/turn attribution simultaneous speech pause, timed hold (micro-pause) lag (prosodic lengthening) breath laugh unintelligible uncertain hearing pseudograph
Symbol line JILL; [ ] (1.2) .. : (H) @ ### #you’re #kidding ~Jill
Comments one new line for each intonation unit semicolon follows name in CAPS brackets show overlap start and end pause duration in seconds less than 150 milliseconds colon marks slowing of local tempo audible inhalation one symbol for each pulse of laughter one symbol per unintelligible syllable transcribed words are uncertain name change to preserve anonymity
The role of I guess in conversational stancetaking Elise Kärkkäinen University of Oulu
1.
Introduction1
My overall starting-point in this paper is an epistemic stance marker that has been strangely neglected in linguistic research so far, namely I guess. Here, 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 markers. In Kärkkäinen (2003), I established that American English speakers use a limited set of high-frequency markers in everyday speech for their expression of epistemic/evidential stance.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. This is the case with another frequent expression, I think, which in everyday American English has been shown in first-position turns to routinely frame an upcoming stanced turn or longer opinion sequence, or in second-position turns to project that a different speaker perspective will follow, while it may also do less routine-like recipient design and face-work at points of trouble in interaction (Kärkkäinen 2003; see also Rauniomaa this volume for essentially similar findings concerning the Finnish minun mielestä ‘in my opinion’ and minusta ‘I think’). In what follows, I will view I guess, an epistemic/evidential fragment that is very frequent in American English (but not in British English), in an overall theoretical framework proposed by Du Bois (2001, 2003, this volume) of stance as a social act that is achieved dialogically and publicly, through our engagement with our co-participants. This view is not unlike the view held in discursive social psychology, of evaluations as not carried around ready-made by participants, but worked up in a way that is suitable for what is being done in interaction (see e.g.,
184 Elise Kärkkäinen
Potter 1998). In this paper, I indeed view stance essentially as emerging from dialogic interaction between interlocutors (Holquist 1990; Voloshinov [1930]1973; see also Kärkkäinen 2006), and I am primarily interested in the practices and processes of (the activity of) stancetaking between co-participants. I draw on conversation 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. In this paper, I argue that I guess, a subjective marker par excellence (but see Section 3.2), functions as an intersubjective stance frame that organizes the stancetaking activity between conversational co-participants in a surprisingly consistent fashion. By frame, I here mean a fragment of speech that provides a perspective or stance toward the action that is produced in the associated utterance (see Section 3.2 for a more detailed definition of frame). It is significant that when we look beyond the initial frame, I guess, the rest of the utterance typically also contains explicit stance-indicating or stance-evoking material, such as evidential markers, markers of epistemic modality and evaluative lexis. The speaker thereby frequently produces an action such as an assessment, an opinion, or a (strong) claim, that inherently involves taking a stance or a position. Indeed, Ochs and Schieffelin (1989: 22) have alerted us to the possibility that affect (or stance) may permeate the entire linguistic system. The authors argue that the linguistic resources for expressing affective and epistemic stance include, not only the lexicon, but the following: grammatical and syntactic structures such as choice of pronouns, determiners, verb voice, tense/aspect, sentential adverbs, hedges, cleft constructions, diminutives, augmentatives, quantifiers, and word order; phonological features such as intonation, voice quality, sound repetition, and sound symbolism; and discourse structures such as code-switching as instantiated by taboo words, dialect, couplets, and repetition of own/other’s utterances (Ochs and Schieffelin 1989: 12-14; Ochs 1992: 412). Students of language have further acknowledged the necessity and difficulty of identifying evoked or implied evaluation, or cases where stance is not explicitly inscribed but seems nevertheless to be evoked (Martin 2003; see also Clift 2006 for interactional evidentials which are dependent on sequential position to index stance, as opposed to stand-alone evidentials which inherently index a stance). What, then, 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,’ ‘coming to a realization,’ or ‘drawing a conclusion,’ i.e., this marker displays and makes public a (reasoning) process of the speaker at a particular point in interaction. 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, it
The role of I guess in conversational stancetaking 185
becomes pertinent to view I guess as not only indexing inference or conclusion, but in effect organizing the stancetaking activity between the co-participants. As we will see below, I guess can be viewed as a discursive practice used for performing a practical task in discourse (cf. Potter and Hepburn 2003): in sequence-initiating actions for displaying the discovery of a speaker’s stance, in responsive actions for displaying and projecting that the speaker will give up on a stance and/or adopt a new one, and in extended turns to insert a stance relevant for the understanding and appreciation of a telling. If some of the most frequent “traditional” stanced items in a language, such as epistemic I think and I guess, turn out to essentially only act as stance frames, i.e., to display and project an upcoming stanced action and organize the stancetaking activity between participants, rather than themselves indexing an interactionally strong and salient stance on many occasions, it is essential that we acquire some breadth in our descriptions of the manifestations of stance in naturally-occurring discourse and language use. We clearly need to also look beyond these frequent and routinized stance markers (such as markers of epistemic modality and evidentiality, evaluative lexis, and affect markers) that linguists have largely occupied themselves with. Indeed, my aim in this paper is to go deeper into the more interactional resources of stancetaking. I wish to shed some light on the practices and processes of stancetaking beyond individual speakers’ contributions, across speakers and speaking turns (cf. repetition of own/other’s utterances mentioned above). I will therefore expand the scope of examination from the actions and turns containing I guess, to the turns-at-talk coming before and after the current turn, and ultimately to the larger activity that the participants may be seen to orient to in a given situation. In other words, attention will be paid to the multiple ways that stances are taken in turns-at-talk containing this initial stance frame, but also in prior and subsequent turns. A frequent resource used thereby for taking stances is the syntactic, semantic, and prosodic resonance between contributions by different speakers, as proposed by Du Bois (2001; see also Anward 2000).
2.
Data and transcription
My database consists of 54 occurrences of I guess gleaned from 17 different conversations of about 15–25 minutes each, from Parts I and II of the Santa Barbara Corpus of Spoken American English (Du Bois et al. 2000; Du Bois et al. 2003). Thus, some 50 sequential environments of I guess were analyzed in detail. One example is taken from a corpus of British and American television news interviews (Haddington 2003).
186 Elise Kärkkäinen
The data are transcribed according to the discourse transcription conventions presented by Du Bois et al. (1993; see the Appendix for a list of symbols used), with the intonation unit (IU) as the basic unit of transcription. The way we as analysts transcribe our data, of course, already affects our perception of how speech and interactional actions are produced and emerge through time. It is advantageous to transcribe data into IUs, because the very visual representation, with each IU lined up on a separate line, immediately alerts us to the emergent, incremental, and directional nature of turns-at-talk and the turn-construction units (TCUs) that turns are made of. Thus, the construction of a turn is frequently incremental, or produced piece-by-piece rather than as one long chunk (Schegloff 1996: 55). All TCUs are also directional, available to participants as they unfold, and not only once they have been produced and completed (Lerner 1996: 307). However, I have modified the Du Bois et al. (1993) convention in one respect. The marking of pauses is generally done in accordance with the conversation analytical methodology: pauses appear on a separate line rather than at the beginning of a certain speaker’s IU, to alert the analyst to the fact that it is frequently possible to assign the pause to some other speaker instead.
3.
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, in order to better account for their use in actual everyday and institutional interactions (see e.g., Ford et al. 2002; Keisanen this volume). The very frequency of I guess makes it a worthy object of such study, and I will here give a brief description of this collocation as a linguistic phenomenon, even though most studies to date only deal with it in passing.
3.1
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, even though not quite as frequent as I think or I don’t know. In Kärkkäinen (2003), personalized stance markers that make reference to the speaker, such as I think, I guess, I know, I feel (like), and I found, were by far the most frequent epistemic markers. This finding gains strong support from some other recent studies based on spoken language corpora, namely those of Biber et al. (1999), Thompson (2002), and Scheibman (2001, 2002).
The role of I guess in conversational stancetaking 187
Biber et al. (1999: 667–669) found that in a large database of some 6.4 million words of conversational American and British English, 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, the statistics provided in the Biber et al. (1999) grammar show a dramatic drop in frequency after the epistemic verb guess: the next verbs, see, find, believe, feel, suggest, and show only occur less than a hundred times per million words of talk. This can be taken as evidence for my claim of the relatively small size of the set of frequent stance markers. Thompson (2002: 138) also observes that the most common complement-taking predicates (CTPs) in her data are the following, 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, 2002) data, I guess was also highly frequent, along with the two other epistemic phrases I think and I don’t know. For comparison, 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, or 100 face-to-face and telephone conversations of 5–20 minutes in length (totaling 205,608 words). I think 729 tokens or 0.35 % of 205,608 words I don’t know 309 tokens or 0.15 % 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. As this search yielded 20 occurrences of I guess, the number of I guess per million words in these data would very roughly amount to 100, which is considerably less than the 500 in the Biber et al. (1999) database. But we may conclude that even though this item is much more frequent in American English, it is not entirely unheard of in everyday British English. Here, however, I confine myself to American English usage.5
188 Elise Kärkkäinen
3.2 The fragment I guess as a stance frame As becomes obvious from many recent linguistic studies, I guess is commonly regarded as a conventionalized, sedimented pragmatic expression, i.e., a stance marker, which frequently precedes complement clauses in English (Scheibman 2002: 32). Complement clauses have commonly been regarded as subordinate to the main clause, i.e., to an epistemic phrase like I think, I (don’t) know, and I guess (see Thompson 2002 for a summary of studies that hold this view).6 Thompson and Mulac (1991a, 1991b) claim I think and I guess to have grammaticized into epistemic phrases that act in the same way as adverbs, i.e., as adverbials, and behave much like epistemic morphemes in other languages. They note, however, that it is not clear what the grammatical status of an epistemic phrase is. In a later study, Thompson (2002) goes on to argue that we might term the frequent CTPs – I think, I thought, I guess, I remember, and I know/knew – epistemic/evidential/evaluative fragments, rather than main clauses obtaining a complement clause. Thompson further claims that the most frequent of these phrases become relatively fixed epistemic formulas (e.g., I think/I don’t think/I thought/I didn’t think), notably with first-person subjects, while the less frequent ones show more diversity of form (e.g., make sure, tell, be interesting). Scheibman (2001: 70– 71, 76) also found that there are highly frequent formulaic collocations of firstperson-singular subjects and especially verbs of cognition (I guess, I don’t know, I think) in her data, and as was mentioned above in Section 3.1, this finding is corroborated in Biber et al. (1999: 667–669).7 It is worth pointing out, however, that the actual degree of subjectivity and referentiality of frequent epistemic fragments like the ones above can be quite low. Scheibman (2001) claims that there is a weakness or generality of referentiality in the subject of expressions like you know, I mean, and I think: the speaker is not specifically referring to himself in these usages, but the entire collocation takes on pragmatic function with a concomitant reduction in specificity of the pronominal referent. Traugott (1995: 38) further claims of the English I think that it is becoming more subjective both in function (toward a fixed phrase indicating speaker’s epistemic attitude) and also in the overwhelming selection of the first-person-subject form, which may eventually become eroded and leave only a discourse particle (presumably think). We may also ask how much referential meaning is left in the predicate and ultimately in the whole collocation. Schiffrin (1987) argues that some cases of discourse markers like y’know are less referentially meaningful than others (1987: 319), while Stenström (as cited in Helt 1997) observes of I think, I mean, you know, and you see that they fill various slots along a semantic continuum that represents a strong relationship to their literal meanings at one end, and a minimal relationship to literal meaning at the other.
The role of I guess in conversational stancetaking 189
Lenk further points out that the (short) lexical items that may figure as discourse markers, i.e., that are used to “signal the sequential and ideational relationship of the two utterances between which they occur, or to other segments within the discourse” (1998: 52), often also have a (separate) propositional meaning (cf. to guess the answer). The collocation I guess, however, mostly appears with a pragmatic meaning in my data, even though it still carries remnants of its referential or propositional meaning in many contexts of use.8 Indeed, in Kärkkäinen (2003: 175–179), 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, it is syntactically detachable from sentences, it commonly appears in initial position of an utterance, it may have a range of prosodic contours, and either has no meaning or only a vague meaning (see Schiffrin 1987 for criteria for discourse markers, to be discussed in more depth in Section 5). Let us then look at the following example, presented in Thompson (2002: 132): (1) (at a birthday party, after Kevin was discovered to have lettuce on his tooth, everyone has jokingly commented on it, and Kendra has asked for a toothpick) WENDY: ... everybody’s getting uh, tooth obsessed. KEN: I guess we a=re.
Thompson draws our attention to the fact that Ken’s aligning agreement to Wendy’s summary of the previous turns is expressed in the complement clause we are, and that it is not the complement-taking phrase, here I guess, that constitutes the actual action of the turn. She ends up suggesting (2002: 142, 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, i.e., the assessments, claims, counterclaims, and proposals, being done in the associated utterance. In effect, Thompson (2002: 141) proposes a view of grammar as reusable fragments, to be used as turns or parts of turns, i.e., as practices of turn construction. This is very much the view that I also adopt here, 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. I guess will be examined in connection with different kinds of conversational actions done in different sequential environments: in initiating and responsive actions, and in extended or multi-unit turns.
190 Elise Kärkkäinen
3.3 A note on the semantic meaning of I guess Few studies actually deal with the semantics of this stance marker. Chafe (1986), however, notes that I guess is a marker of “belief.” He further claims (1986: 266) that belief is a mode of “knowing” in which concern for evidence is downgraded, and speakers believe things because other people do or simply because they want to believe them. Thompson and Mulac (1991a: 325) observe in passing that, as epistemic phrases, I think and I guess have still retained vestiges of their earlier 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. This, the authors argue, is “traceable to the difference between think and guess as verbs: guess implies an assertion based on little or no evidence, and hence less commitment to a proposition than think does” (1991a: 325). But on a closer analysis, I guess in my data frequently appears in contexts where the speakers do show an orientation toward evidence, which is presented in, or inferable from, the immediately prior turns, or is even present in the actual physical or wider social environment where talk takes place. Here, it differs from I think, 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, see Kärkkäinen 2003). 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.
Practices of stancetaking: Beyond the stance frame I guess
4.1
I guess in sequence-initiating actions: Discovering a stance
I will here examine cases in which I guess frames sequence-initiating actions, such as (first) assessments, opinions, assertions, and questions, which are followed by second actions or second-pair parts, such as second assessments, opinions, assertions, and answers. The following examples show that I guess marks actions in which some kind of change in the speaker’s current state of knowledge or awareness or orientation has (just) taken place (cf. Heritage 1984 for the change-of-state token oh). The stance displayed frequently arises from an inference made by the speaker, and is based on evidence gleaned from a just prior turn or turns, from the actual environment where talk takes place, or from the wider social context. Such a stance, as any act of stancetaking, necessarily invokes an evaluation at some level, whether this is actually asserted or otherwise inferable (Du Bois this volume; see also Martin 2003 for evoked evaluation). ‘Evaluation’ in turn means charac-
The role of I guess in conversational stancetaking 191
terizing the stance object as having some specific quality or value (Du Bois this volume; see also Goodwin and Goodwin 2002: 154 for assessments as evaluations of persons and events being discussed within talk). And, finally, actions involving I guess propose something as a possible stance object that the co-participants can subsequently position themselves towards, i.e. such actions frequently initiate stancetaking activity between the conversational co-participants (see also Goodwin and Goodwin 2002 for assessment activity as something that the participants orient to, and participate in, together). Displaying a sudden change in the speaker’s state of knowledge or awareness 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 utterance was somehow wrong or mistaken and no longer relevant, and in effect, cancels out the need for the action performed in it, typically a question: (2) (Wonderful Abstract Notions SBCSAE 0017 ) 1 MICHAEL: 2 3 4 JIM: 5
... So they’re `color-cycling right ^now. ... Does it ^zoom `later? [I `guess it ^must]. [(H) Yeah, `I’ve got] ^zooms on here `too.
While producing the question in line 2, Michael seems to reason and infer, on the basis of how videos are usually designed, that sooner or later the picture will zoom in on something. The semantic meaning of must ‘induction’ (Chafe 1986) of course strengthens this interpretation, 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 Chinese? .. I guess with a name like Yoyo=; SBCSAE 0019). Notice how interactive this turn design actually is: the speaker engages in interaction with him- or herself, and simultaneously makes public his or her inferential process and discovery of some state of affairs.10 Such turns are also regularly responded to by the recipient, who may legitimately speak at the end of the first turn-construction unit, the actual question, even though the current speaker also continues to speak. A further example of a change in the speaker’s awareness, or the speaker making a discovery based on inference, is provided by the following example (3). The topic of prior discourse has been Harold’s nephew Thomas, a three-year-old who is now learning to tap dance. The boy’s parents had been inspired by a famous young tap dancer, who is “awesome” and incredibly fast. Prior to this extract the participants have been discussing and assessing him from their respective angles.
192 Elise Kärkkäinen
(3) (Lambada SBCSAE 0002 ) 1 HAROLD: I’m sure `Thomas is all ^over it. 2 (3.7) 3 JAMIE: Prob[ably XX][2XXX2] -4 HAROLD: [ I mean `he][2has a bro-2] -5 MILES: [2XXXX could have2] ^see=n `him. 6 HAROLD: I `guess that `means his broken `leg is [3@doing @^okay3]. 7 PETE: [3I was ^wonder3]ing about `that, 8 I was `imagining [4he had `broke an ^arm4] or something. 9 JAMIE: [4 Oh yeah= 4]. 10 PETE: But it was his ^leg?
Harold brings the prior topic to a close on line 1 by offering a concluding evaluative claim or assessment concerning the little nephew’s enthusiastic attitude toward tap dancing. There is clearly a topic closure here (cf. Goodwin and Goodwin 1992: 169), which is marked prosodically by Harold’s lowered volume and a long pause after this turn. Then several participants self-select to speak almost simultaneously, but Harold’s new beginning is clearly signaled by a rather loud and high-pitched I mean `he has a bro- -- on line 4; he refers to the nephew, as will become obvious a little later. This is overlapped by Jamie (line 3, largely inaudible) and Miles (line 5, very likely talking about the nephew as well, wishing that he could have seen him tap dance). 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. The broken leg had not been topicalized in the conversation so far, even though we can now be fairly certain that Harold intended to announce it as news on line 4, I mean he has a broken leg. On line 6, he produces the action of simultaneously (in passing) announcing a piece of news and evaluating it (doing okay), a turn design that highlights the unexpected nature of the state of affairs. Harold also produces the last two words with some laughter, thereby inviting laughter from the other participants (Jefferson 1979). Pete and Jamie decline this invitation and take up the topical import of Harold’s talk instead. 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, even though they clearly had some prior knowledge of the boy’s broken leg. 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, which he is about to put on the
The role of I guess in conversational stancetaking 193
table. I guess signals that this particular stance is such that it arises from an almost on-line reasoning process of the speaker. When producing such an inference, the speaker digresses from his original trajectory: rather than simply self-repairing and re-producing his original news announcement, he proceeds to evaluate the newsworthy item, the broken leg. We can therefore say that stance and evaluation take the center stage here, and thereby affect the original action trajectory and sequential organization of talk. What is more, by introducing such a stance, 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, namely first assessments, which make second assessments conditionally relevant (Pomerantz 1984). In example (3) the co-participants join in to further evaluate the nephew’s recovery, 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”. In the next example, the evidence is clearly seen in the environment, the Indonesian masks that Harold and Jamie have hanging on their wall. (4) (Lambada SBCSAE 0002 ) 1 MILES: ... `Those two top ^ma=sks , 2 .. `there. 3 (1.7) 4 HAROLD: ^Yep. 5 (0.8) 6 JAMIE: @@@ 7 HAROLD: [We `need a ^verb]. 8 MILES: [I ^guess the] [2`ones below are `women2], 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: .. ^Uh-oh. 14 MILES: I mean, 15 .. i%- -16 (1.1) 17 Well you see their eyeballs, 18 [I guess]. 19 PETE: [Right]. 20 .. Yeah [2the other2] -21 MILES: [2In addition2] to a mustache and a beard. 22 (1.0)
194 Elise Kärkkäinen
23 But the thing is, 24 [that second one looks like the guy] who was in [2one of the2] [3Oba Oba3] skits.
Miles introduces the two top masks as the actual referential focus of talk on line 1, 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. But before he initiates talk on that exact referent, he displays some trouble in continuing his talk, as can be heard from the long pauses and the fact that Harold humorously prompts him to continue (lines 4 and 7). In line 8 Miles then produces a noticing, and engages in a brief comparison of the top masks and the ones hanging below them; 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. 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. The two utterances are clearly designed to be produced together as one chunk: the noticing in line 8 (which ends in a continuing intonation) acts as a necessary background stance for the more explicit stancetaking action, the actual question, in line 11, which is said in a kind of latching prosody, in an overall high pitch, continuing at the height where the previous intonation unit left off. In what follows, both Jamie and Harold display that there is some trouble involved in Miles’s proposed stance: Jamie laughs on line 12, and Harold displays, by a mock alarm cry uh-oh, that there was something quite problematic in the prior turn. Miles himself realizes soon enough that the women in the masks have “a mustache and a beard”, that they do not portray women at all but male figures. After this side sequence he resumes the original topic, the figure portrayed in one of the top masks. In sum, what we can say about the role of I guess here is that it again appears in a context, in a side-sequence, where the speaker makes public an inference (a noticing) that was based on some evidence provided by the immediate environment. At the same time, the insertion of such an inference takes place in a sequential position where the speaker rather abruptly changes the trajectory of his talk. The inference is also part of a turn design that generally invites others to take a stance on the matter, even biases them toward a certain stance. We can see essentially a similar pattern in the following excerpt from institutional data, namely from an American television news interview recorded in the aftermath of the 9/11 terrorist attacks. The same example is also discussed by Haddington (this volume) with respect to the turn-design features of the interviewer’s and the interviewee’s turn.
The role of I guess in conversational stancetaking 195
(5) CNN, Larry King Live, Sep 12, 2001: Bomb him out IR: Larry King, IE: Brian Jenkins (001 / 5 / 3:49) 1 IR: 2 3 4 5 6 IE: 7 8
(0.6)(TSK) ^Brian, could `you=, ... well I `guess the `public, would `look at this `simply , `Could you ^bomb him out. .. (TSK)(H)U=h, I d- -I don’t know that you could ^bo=mb him out,
The interviewer (IR) designs his turn on-line, as he re-directs it after having already initiated the question Brian, could you=. He subsequently inserts an utterance which incorporates the stance of an ‘other,’ namely of the general public, who are represented as favoring the rather extreme action of simply bombing the terrorist leader out from wherever he is hiding (a rather widespread public opinion established on some rapid opinion polls). Well here signals an orientation shift within the speaker’s extended turn (Schiffrin 1987: 102–127), and displays to the interviewee that the action that follows is not fully coherent with the main trajectory of the question. It is further framed by I guess, which projects an upcoming stance, 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 “positioning” activity). Iconically, this inserted stance is also produced in a parenthetical prosody, namely in slightly lower pitch and faster tempo. Goodwin (1981: 127) suggests that such modifications of the turn construction are regularly done by participants to “coordinate their production with the actions of a recipient”. For example, a phrasal break and subsequent insertion of a new unit (a word, a phrase, 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, and coordinate his or her actions with those of, the recipient. 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 :, I guess he’s, [ X__________
In example (5) above, however, something else is going on. The IR engages in looking at the IE only until the end of line 2. Simultaneously with the onset of
196 Elise Kärkkäinen
the parenthetical digression starting with well I guess, he withdraws his gaze from the IE and looks down. He also turns his palm upward, possibly in acknowledgement that he had already brought up the results of the opinion poll earlier in the interview, or as an ‘offer’ or a ‘presentation’ of the discourse (or stance) object to the recipient (Kendon 2004: 123). Only after the insertion is completed does he turn his gaze toward the IE again, while the palm remains in its upward position. The embodied production of the insertion, especially the downward gaze and the (parenthetical) prosodic production, together contribute to its interpretation as an interactionally engendered ‘remembering’ (Heritage 2005), a realization of the IR inserted in the larger question sequence. On the other hand, 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. Lerner 1991). The above insertion again causes talk to diverge from its original trajectory. 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, rather than the actual stance of the current speaker himself. 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=, on line 2, very likely resulting in could you bomb him out, which would have constituted a valid question presented in the IR’s “own name.” By inserting the I guess-framed digression, the IR switches from this to proposing another stance that he may or may not himself share, and to now asking the question “on behalf of ” the general public rather than himself. Finally, this turn design invites the IE to take a stance on the matter and, while doing so, take into account the ostensibly simple solution favored by the general public (but not necessarily shared by the IR). In sum, I guess in initiating actions like the ones above indexes a just discovered stance, a sudden change in the speaker’s epistemological state of knowledge or awareness or orientation. This stance arises from a reasoning process of the speaker, which is often based on evidence or some other stimulus gleaned from just prior turn(s) or even the ongoing one, from the actual environment where talk takes place, or from the wider social context. I guess here comes quite close to another marker of evidentiality and stance in British English data, surely, 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. 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, or via inference (Downing 2001: 277). 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. A special flavor of I guess, not unlike
The role of I guess in conversational stancetaking 197
that of surely, is that the stance that “surfaces” in the associated utterance often has a humorous and surprised quality, and sometimes comes close to sarcasm. What is essential is that even though we see evidence of some private processes of inference in the turns framed with I guess, these nevertheless derive from and surface in the intersubjective to and fro of dialogue. The I guess-framed utterance often constitutes a clear side sequence within the current speaker’s ongoing sequence, initiated and closed by the speaker him- or herself (Lerner 1991; Local 1992; Duvallon and Routarinne 2005; cf. also Jefferson 1972 for subsidiary side sequences initiated by another speaker). I guess can further be seen to organize the stancetaking activity of the co-participants, as the actions containing this frame often invite them to also take a stance in subsequent discourse, and in doing so to take into account the stance already implicit or explicit in the design of the ongoing turn. We may also argue that at such points in interaction, 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.
4.2 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.e., in responsive actions to some other actions) in conversational sequences that often involve some degree of disagreement and disaffiliation between the participants. Here, I guess similarly acts to display and project – on the basis of just prior discourse – that the current speaker wishes to modify, withdraw, and redefine his or her original stance at this point, i.e., to align with another speaker and, frequently, to display affiliation and a convergent stance. Note that even though many researchers use the terms ‘alignment’ and ‘affiliation’ almost interchangeably, in what follows, by aligning action is simply meant one that is produced within the trajectory of the sequence-so-far (e.g. as one in an adjacency pair sequence), while an affiliating action is used to refer to one that establishes and increases social solidarity between the social actors involved. In the next example, Ken is telling a story about his childhood, when he had witnessed the feeding of a fish, an Oscar, in a pet store: a goldfish was put in the tank and the Oscar started chasing it. Joanne, as one of the two recipients, is taking an active role at points in the telling (Joanne and Ken are a couple), even though she has no prior knowledge of the incident.
198 Elise Kärkkäinen
(6) (Deadly Diseases SBCSAE 0015 ) 1 KEN: and `then= uh, 2 (H) But the `goldfish got ^s=tuck. 3 .. <MRC> ^half-way ^into his ^mou[=th ]. 4 JOANNE: [It `went the] wrong ^way. 5 KEN: ^No, 6 it [was `going], 7 JOANNE: [^Fin `first]? 8 KEN: it was `going, 9 JOANNE: .. ^Mouth `first? 10 KEN: .. Well it was go -11 went ^tail `first. 12 JOANNE: `That’s the [^problem. 13 KEN: [I `guess that’s the ^wrong way, 14 JOANNE: (H) `That’s the ^problem ]. 15 KEN: `I don’t `know . 16 (H)] ^Anyway, 17 JOANNE: [2(H)2] 18 KEN: [2(H)2] .. (TSK)`So, 19 (H) you got this ^Oscar there, 20 `swimming around [in the ^tank,
There is a danger from line 4 onwards that Joanne may halt or derail the storytelling (cf. Mandelbaum 1989: 118). She hastens to propose a reason why the goldfish might have been stuck in the Oscar’s mouth, It `went the wrong ^way. Ken starts to disagree on lines 5–6 (^No, it was `going,) and again in line 8 (it was `going,), but he is interrupted by Joanne, who proposes further candidate alternatives for how the fish was going (^Fin `first?, line 7 and ^Mouth` first?, line 8). For Joanne, 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 goldfish, her point having been that the snake would always eat the goldfish by the head. 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, as she has been talking about the two interchangeably) rather than by the head, and actively pursues that in the discourse that follows. On lines 10–11, Ken displays recognition that the fish actually went in 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, went ^tail `first., which is produced with a drop in pitch on tail that indicates some resignation at the realization of this fact. For Joanne, this of course proves
The role of I guess in conversational stancetaking 199
her earlier point about the snake, and she rather enthusiastically displays this by producing `That’s the ^problem, with the pitch going up and the pitch range widening during the second production of ^problem (partly also because she is overlapped by Ken). Ken finally concedes on lines 13 and 15 that Joanne may be right, I `guess that’s the ^wrong way, `I don’t `know . This turn indicates that Ken is giving up on what for him is possibly a relatively small detail in the bigger scheme of things, in order to retain his role as the main storyteller and to continue with his story. There is some evidence for this in how Ken designs his turn. He hastens to agree already upon hearing Joanne’s that’s the (line 12); this particular syntactic chunk or collocation strongly projects that Joanne is about to produce an evaluative turn, 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. line 4). Ken indeed displays an understanding of Joanne’s turn-so-far as an evaluative one, even though he misprojects what type of evaluative item will follow (the problem), and produces the wrong way instead. In fact, 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), while Ken still orients to the more local ‘which way is the wrong way.’ Ken finally tags on I don’t know in a very accelerated form, thereby marking that this side sequence in the story has come to a completion (cf. Ford and Thompson 1996: 169; Scheibman 2000). After an audible inbreath he continues the story with ^Anyway, (cf. Jefferson 1972: 316–320 for resumption of the ongoing sequence; cf. Lenk 1998: 71–78 on anyway closing general conversational digressions and resuming the earlier conversational topic). The role of I guess in this sequence is to display that, in light of what was established in just prior discourse, the speaker is giving up his original stance and adopting a new one. In this sense, its use here is not far removed from that witnessed in initiating actions (see Section 4.1), because here as well, 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). In the following, I will present the turns above in which stances are being negotiated and a (more) convergent stance is achieved, in the form of two diagraphs. Such a diagrammatic form of presentation has been proposed by Du Bois (2001), to make visible the frequent modification of stances that takes place between discourse participants: a stance is often a “product of an immediately prior act of stancetaking toward a shared stance object” (Du Bois 2003). Diagraphs map the relevant intonation units produced by different speakers onto each other (roughly) according to syntactic structure.
200 Elise Kärkkäinen
In Diagraph 1, the participants first establish which way the fish was going in the Oscar’s mouth. In doing so, the speakers resonate both with their own and with each others’ talk at several levels of linguistic structure. Diagraph 1 (from example 6): 4 JOANNE: It `went 5 KEN: ^No, 6 KEN: it was `going, 7 JOANNE: 8 KEN: it was `going, 9 JOANNE: 10 KEN: Well it was go -11 went
the wrong ^way.
^Fin `first? ^Mouth `first? ^tail `first.
Joanne completes the syntactic structure initiated by Ken (lines 6–7 and 8–9). This format is broken by Ken himself on line 11, 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 past tense went signals finality compared to the past progressive was going and resonates with Joanne’s original wording on line 4, also in the past tense. He also produces the final item, ^tail first, in resonance with both the syntactic form and prosodic realization (in terms of accent) of Joanne’s candidate items in lines 7 and 9. Their versions of reality now match factually (cf. Mandelbaum 1993: 263 for negotiations about versions of “reality”), but what the implications of this are for their initial disagreement need to be made more explicit. The next diagraph presents how this is done. Diagraph 2 (from example 6): 4 JOANNE: It `went the wrong ^way. 12 JOANNE: `That ’s the ^problem. 13 KEN: I `guess that ’s the ^wrong way,
Joanne’s initial anticipatory judgment of the fish going in ‘the wrong way’ was expressed on line 4. After the participants have established together that the fish had in fact entered the Oscar’s mouth ‘tail first,’ Joanne produces an explicit ‘Itold-you-so’ concluding evaluation on line 12. Ken produces an affiliating second evaluation on line 13, even though, as we have seen, slightly shifting the stance object here. Here, 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,
The role of I guess in conversational stancetaking 201
and also recycles her earlier evaluative element, the wrong way. He further signals his convergent stance by framing the turn with I guess. The following is another instance of storytelling. Kevin and Wendy (a couple) are jointly telling about an incident that they had witnessed together recently in the parking lot of a mall; however, their interpretations of what exactly the protagonist was doing or trying to do differ. (7) (Appease the Monster SBCSAE 0013 ) 1 KEVIN: ... Some `guy came `out and he- he was, 2 @[=] 3 WENDY: [ Oh ]. 4 KEVIN: he was trying to `sell us ^cologne [2f-2]– 5 KENDRA: [2(COUGH)2] [3(COUGH)3] 6 WENDY: [2No2], 7 [3he `wasn’t3] trying to ^sell [4us `co4]lo[5gne5], 8 KEVIN: [4`Well it-4]– 9 [5No5]=, 10 I `guess he was `trying to like, 11 ^lure us to a .. `place where they ^would `sell, 12 like, 13 .. ^imitation `cologne, 14 but he `said, 15 it’s `not ^imitation, 16 `be[cause] ,
Even though in the immediately preceding story preface sequence Wendy had aligned herself as a story consociate, here it is apparent that for her there is now some trouble involved (cf. Lerner 1992). Kevin produces an assertion that frames the story and the incident as having been mainly about selling, or about trying to sell, Some guy came out and he was trying to sell us cologne (lines 1 and 4). This turn design contains an implied evaluation, an evoked stance, namely that trying to sell cologne to them is something rather strange, funny, or doomed to fail. We can find evidence for this in Kevin’s production of this turn. He produces a laughing particle in the middle, restarts several times as if to further emphasize his upcoming point, and can be heard to put on a voice quality of “slight desperation.” He is thereby orienting the other participants toward the upcoming story in a certain way.11 But Wendy, the story consociate, disagrees on such a framing and puts in a correction in lines 6–7. Wendy explicitly disaffiliates with Kevin, No he wasn’t try-
202 Elise Kärkkäinen
ing to ^sell us cologne, with an emphasis on the contentious item, sell. It is not the case that Wendy is disagreeing on the facts of the story (cf. Lerner 1992 on story consociates correcting facts), it is rather that their fundamental perceptions differ of what actually happened in the parking lot, and their versions of “reality” are being negotiated here (Mandelbaum 1993: 263). Upon hearing the emphasized word sell, Kevin immediately starts to redesign his turn; `Well it- -- displays that he may be about to start with a further disagreement, but he then retracts and produces an explicit No=, in effect agreeing with Wendy’s previous disaffiliating turn. In what follows, he produces a new claim that takes back some but not all of what he had said initially, I guess he was trying to like, ^lure us to a .. place where they ^would sell, like, imitation cologne. He offers a new and more precise understanding of the past events, that the guy was going to sell something later, in other words Kevin does not altogether withdraw the idea of selling. 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. 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.’ It seems enough of a concession to Kevin, who then assumes the role of the main storyteller and continues the story. Diagraph 3 (from example 7): 4 K: he was trying to `sell us 7 W: he was n’t trying to ^sell us 10 K: I guess he was trying to like, 11 ^lure us where they ^would sell, 12 like, 13
^cologne cologne, to a .. place
imitation cologne.
We can see from the diagraph that on line 7 Wendy is building her divergent stance on Kevin’s just expressed stance. She does so with the negation and with just slightly shifting the stress from cologne to sell. Kevin in turn builds his now more converging stance on his own and Wendy’s immediately preceding turn, by resonating with their syntactic form (he was trying to X us), and in addition, bringing in some new information, a place and imitation cologne, that further specify how he had perceived the original incident. The form ^would sell even escalates the content in trying to sell, thus displaying that Kevin still maintains his original position on this point.
The role of I guess in conversational stancetaking 203
In sum, I guess is brought in when the current speaker signals that, in response to some evidence or other material presented in just prior turns, a (sudden) 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. At the risk of simplifying somewhat, the difference between I think and I guess in second position is that I think tends to project or frame disaligning or disaffiliative second actions, namely a different slant or take on a given issue, or a new speaker perspective (Kärkkäinen 2003), whereas I guess basically affiliates with the prior speaker. Even though very similar, the two are not interchangeable but have each specialized in a certain type of interactional function. Lastly, I will briefly mention two other types of second-position actions framed by I guess, in which the function of this frame comes close to the semantic meaning of (weak) “belief ” as proposed by Chafe (1986) and Thompson and Mulac (1991a). First, there is a typical reactive response to evaluative first actions, Yeah, I guess so., where I guess marks the participant’s striving toward a convergent stance (e.g., compliance). On occasion, 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. B: Yeah I guess you do .; SBCSAE 0005). Second, in answers to requests for information, the speaker may simply answer I ^guess or preface his or her answer with I ^guess, and the function again comes close to the semantic meaning of “belief.” (8) (Car Sales SBCSAE, unpublished) 1 G: 2 3 D: 4 5 6 7
.. (H) They have `many uh, .. ^Spanish `speaking `people go ^in there? U=h, u=h, I ^guess‑ , .. ^Small ^percentage, .. ^Basically they’re .. ^Caucasians.
Here, I guess appears simply to express that the actual answer to be offered is probably true.
4.3 I guess in extended turns: Inserting a stance In many cases, I guess occurs in conversational sequences in which one main speaker produces a multi-unit turn or tells a story, while the other participants align as recipients. Here, I guess projects an inserted stance that is often based on an infer-
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. Such an aside often works toward increasing understanding or appreciation of the telling or story. Here, I guess sometimes comes very close to an actual hearsay evidential used to mark that the current speaker positions him- or herself as not possessing or relating first-hand knowledge. Here is an example from the same story as in example (6), about feeding the Oscar with goldfish. 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). (9) (Deadly Diseases SBCSAE 0015 ) 1 KEN: [Anyway, 2 I remember they had a, 3 (H) they had this fish in there called] an Oscar. 4 (H)= And they are these like, 5 .. [they are these really, 6 JOANNE: [<X> I remember- , 7 yeah th- they’re big]. 8 KEN: .. They’re these] big gnarly suckers. 9 They’re like [2um, 10 JOANNE: [2Yeah, 11 I remember what they look2] like. 12 KEN: .. (TSK) (H) They’re like m-2] -13 f- -14 I don’t know, 15 [six or eight inch]es, 16 JOANNE: [big eyes], 17 [2(H)2] 18 KEN: [2(H)2] and they’ve got, 19 ... they `only eat like other ^fish . 20 They eat like small [^fish. 21 JOANNE: [(SNIFF)=] 22 KEN: And they eat like], 23 ... you know like ^goldfish. 24 And I guess #the goldfish or ^g=uppies, 25 get the `brunt of ^everything. 26 .. (H) Poor ^guys. 27 (H)(Hx) They’re `like the, 28 you know, 29 they’re like the ^cattle of the @`animal @^world.
30 31 JOANNE: 32 33 34 35 KEN: 36 37 JOANNE: 38 39 KEN: 40 41
The role of I guess in conversational stancetaking 205
[They’re br=ed only to f=eed other things, [Yeah, f=eeds. Yeah, they’re feeders. (H) Anyway, (H) anyway so, They’re nothing but tanks of feeders, in every pet shop (Hx)]. (H) I remember w-] watching them feed the Oscar once, when I was a kid. ... (TSK) And they dropped in the goldfish,
Throughout the orientation sequence Ken casts a very negative stance toward the Oscars by using negatively evaluative lexemes (big gnarly suckers). In line 19, he yet again evokes a negative stance toward them through only; not only are the Oscars big and ugly, but they are also predatory fish. From here onwards, he engages in a kind of incremental turn design and escalating stancetaking activity. He first asserts that Oscars eat small fish, 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”). On lines 24–25, Ken makes a generalizing evaluative claim about the hard destiny of goldfish, And I guess #the goldfish or ^g=uppies, get the `brunt of ^everything. (cf. Scheibman this volume). He is struck by the likeness of what he is about to tell to what Joanne had just told before, and this is reflected in the way this generalization is produced, 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, it is generally very low in pitch, and the extreme case formulation ^everything receives a rather strong primary stress (cf. Pomerantz 1986). The utterance can now be heard as really making an inference on the basis of the previous story and the present story-so-far, very likely to intensify interest in the upcoming story about the dramatic feeding of the Oscar. Ken then stops to evaluate the goldfish some more: he offers a token of sympathy toward them, Poor ^guys, and further evaluates their lot in the animal world (lines 27–30). It is possible and even likely that by engaging in such escalating stancetaking activity that gradually also becomes more specific, Ken wishes to involve the story recipients in taking a stance on the same object, the goldfish, at least to express their validation or ratification but possibly also a convergent
206 Elise Kärkkäinen
stance. But the recipients may not have been able to do so at line 26 (notice the short pause), as ‘getting the brunt of everything’ is still rather vague in reference and, what is more, the recipients have not yet heard Ken’s actual story. Once Ken has elaborated his point some more, Joanne expresses a strong convergent stance at the earliest possible point on lines 31–34, and Ken continues his story (Anyway, anyway so). 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. The speaker opens this digression or side sequence with I guess (or in this case And I guess) and closes it with anyway (cf. again Jefferson 1972 for side sequences initiated by a recipient and Lenk 1998: 71–78 for anyway after general conversational digressions).12 Such initial and final marking or bracketing of discourse sequences, of course, strengthens the interpretation of I guess as a prospective discourse marker for which more support will be given in Section 5. The inserted stance, the generalizing inference, is of the same “recently discovered” quality as we have seen in earlier sections of this paper. 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). The next example briefly illustrates a case in which no visible uptake follows from the recipient, but in which the turn design is nevertheless a very interactive one. 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) (Actual Blacksmithing SBCSAE 0001 ) 1 LYNNE: 2 3 4 5 6 7 LENORE: 8 LYNNE: 9 10 11 12 13 14 15
(H) Anyway, I was gonna tell you about that %- -... the girl, you know, there isn’t mary- .. very many, .. ferrier girl=s you know? Uhhuh. That do this? (H) This summer I met one. (H) Jorgensen’s, they have ... one .. shoer, that comes to their house, all the time, they’ve had him for years and years and years. (H) And I guess like in `Minnesota it’s real ^we=t ,
16 17 18 19 20 21
The role of I guess in conversational stancetaking 207
(0.7) and ` stuff you know? So like, .. they really have to watch their shoes, for not coming off, because the hoof wall, is so much softer.
The story is told in Montana. The reference to Minnesota is explained by the fact that Lynne has spent most of the summer there with her boyfriend and his family, the Jorgensens. The protagonist of Lynne’s story is a girl ferrier, who, as it turns out much later in the story, is incredibly strong and competent after only nine months of ferrier college. Yet the inserted assessment in lines 15-16 acts as a preface to a longer assessment sequence that is consequential for the understanding 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). 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. The assessment, produced somewhat faster than prior speech, 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 turnconstruction unit, and lines 15–16 are clearly pragmatically incomplete, requiring elaboration (Ford and Thompson 1996). I guess is in large part doing organizational work in the telling, as it comes at a transition point from the main story line, i.e., at the beginning of a side sequence (see also example 11 below). Again, this digression is closed with anyway (see Lenk 1998: 65–68, 93–94 for anyway after digressions supplying relevant background information and but anyway used for continuing the narration), even though this time the resumption of the main story line about the girl ferrier comes considerably later after 46 lines: 67 LYNNE: 68 69 70 71 72 .
(H) But anyway, ... (TSK) they always have this one horseshoer, and, (H) this girl’s been .. in this f=errier, ... uh, she’s going to an actual .. ferrier college
Again by way of summary, while they are producing multi-unit turns such as conversational stories, speakers mark through I guess that they are diverging from the trajectory of the story or telling, to insert a stanced utterance or even a longer sequence of such utterances. These involve taking an explicit stance toward some-
208 Elise Kärkkäinen
thing that is consequential for the story, arises from the story as a conclusion, or serves as an explanation for the recipients to better appreciate the story, but is often something that the narrator him- or herself just inferred from the story-sofar. Similarly to first and second position I guesses, a shift in the epistemological stance of the speaker is thus again evident.13 I guess may simultaneously mark that the teller does not possess first-hand knowledge or was not the actual experiencer of an event, but that the telling at this point is based on hearsay or on some indirect evidence, and is also potentially disputable. In all, we can say that I guess again organizes interaction in that it is used to draw the recipients’ attention to the current speaker’s stance, possibly also to invite them to appreciate and ratify this stance. However, it is often also enough for recipients to align as passive recipients of an extended turn such as a conversational story, or display only minimal ratification (or ratification through embodied action only). And finally, it is worth noting that the I guess-framed stanced utterances or sequences deviate from the original trajectory of talk, that is, stancetaking assumes a central role in interaction. 5. I guess as a discourse marker In this paper the view has been adopted (see Section 3.2) 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). Such a view of grammar is a novel one and emphasizes grammar as consisting of a “collection of crystallizations of linguistic routines” used in the service of social interaction (Ford et al. 2003: 120). However, 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.’ As was established in Section 3.1 above, I guess is one of the very frequent stanced items in spoken American English, even though not quite as frequent as I think and I don’t know. While the latter have been shown in recent research to have specialized into some rather routine discourse-organizing functions in conversation (Beach and Metzger 1997; Kärkkäinen 2003; Scheibman 2000), we have also seen evidence in the examples above of a similarly discourseorganizing function of I guess, as it may signal that an upcoming sequence is not fully coherent with the main flow of the telling or talk (cf. Lenk 1998 for a view of discourse markers as primarily marking global coherence relations). In this section, 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. an essentially similar list by Brinton represented in Jucker and Ziv 1998: 3, and a summary of
The role of I guess in conversational stancetaking 209
criteria by different scholars in Helt 1997: 16–17). However, because the focus in the present article has been on the role of I guess in the stancetaking activity between conversational co-participants, and not on all possible environments of use of this marker, some of the observations that I will make below of its discourse marker status are not directly observable in the example segments above. First, I guess is syntactically detachable from sentences, as it has largely lost all overt marking of a syntactic connection to them, namely the complementizer that. It also has some discourse and sentence mobility, because even though it tends strongly to occur intonation-unit-initially and at the same time clause- and sentence-initially (as in all the examples of this paper), it may also at times be encoded as an independent intonational entity, a separate IU, and may appear (even though rarely) in the middle of an emergent syntactic sentence (for lack of space, I do not present such occurrences here). Schiffrin’s (1987) second criterion for discourse markers is that they have to be commonly used in initial position of an utterance, which was established immediately above. Third, I guess has a range of prosodic contours, as it is produced with secondary, primary, or no stress on guess (see examples (3), (5)–(7), and (11) for secondary stress; (4) and (8) for primary stress; and (9) for an unstressed case), while at times it is produced in a very accelerated and phonologically reduced form (examples (9) and (10)) or with parenthetical prosody (example (5)). Givón indeed observes of “epistemic quantifiers” that “the conventionalized subject pronoun is so specific to particular verbs, that it is often dropped in rapid speech,” as in (I) think she’s there, or (I) guess you were right (1993: 38). However, only one such instance was found in the present database, Doris: … Guess I’ll put it o=n. it’s not on, . Fourth, I guess operates at both local and global levels of discourse, because it can project an inserted stanced digression not only one utterance long (e.g., examples (5), (6), and (9)), but also a side sequence of several units (as in example (10), where the side sequence extends over some 50 IUs). Fifth, according to Schiffrin (1987) a discourse marker either has to have no meaning, a vague meaning, or to be reflexive (of the language, of the speaker). This criterion is also clearly filled by I guess. In many contexts of use, it seems to make only a very general or vague reference to the actual speaker: it primarily frames and projects an upcoming stancetaking action and simultaneously the onset of a digression, a side sequence. But on other occasions, I guess appears to retain some of its referential, lexical meaning and brings in a particular kind of evidential stance, e.g., in example (4), the speaker bases his inference on visual evidence of the masks on the wall (and stresses ^guess); or in example (10), the teller does not have first-hand knowledge or experience about Minnesota. In the
210 Elise Kärkkäinen
latter cases, I guess acts much like a hearsay evidential such as people say or I’m told, 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, e.g., it seems was originally a marker of induction but may sometimes be used as a hearsay evidential). Of course we can argue that there is some semantic meaning to I guess, as speakers do not simply choose I think or some other discourse marker in its place – in other words, we cannot say that I guess has no meaning at all but a latent (vague) meaning. The following two examples further display the difference of I guess and I think as discourse markers. In example (11), Rebecca, an attorney, is engaged in the delicate issue of preparing Rickie, the victim of an exhibitionist, to appear as a witness in court. Here, 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 ) 1 RICKIE: He went ^through <X> th[ere , 2 REBECCA: [He `went through ^these `doors]? 3 RICKIE: into `it unhunh], 4 into `another ^car. 5 REBECCA: ^Oh[=]. 6 RICKIE: [(H)] And `then he came `back ^again, 7 .. and `then, 8 (H) I `guess that was `like another ^stop, 9 and more `people were getting ^off , 10 .. and ^the=n, 11 ... that’s when he `came .. and ^s=at b=y me, 12 ^sat .. in this `seat right ^here.
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 defendant. 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, possibly, because there were now fewer people). Again the aside contains the speaker’s speculation about what may have caused the man to come back to her car, and the role of I guess is to mark it as such. Let us compare this example to an almost similar one that came a little earlier in the same conversation, where we have I think instead of I guess, and the special nature of each will become more apparent.
The role of I guess in conversational stancetaking 211
(12) (Tell the Jury That SBCSAE 0008 )14 1 REBECCA: 2 3 4 RICKIE: 5 6 7 8 9 10 11
A=nd, .. when he came in, .. when .. where was he when you first saw him .. come in. `He got on the ^trai=n `o=n like, (THROAT) I ^think ^Twelfth Street or something? in .. (H) ^Oakland? (H) And like in the, .. let me see . (3.9) Okay. .. He got on another car.
This is similarly a case of subsidiary information inserted in a lengthy answering turn, as Rebecca’s question on lines 1–3 was really meant to establish where on the train, rather than where in Oakland, the defendant had been when Rickie first laid eyes on him. Rickie goes on to provide more information about whether the defendant got in the same car or not on line 7. By inserting I think (and or something) she simply marks part of the subsidiary information as uncertain. The difference between the two markers is in the potentiality inherent in I guess to display an inference, that is, to display that the speaker is realizing or inferring something at the moment of speaking, often on the basis of available evidence. The speaker therefore orients toward making a conclusion rather than simply offering uncertain information. By contrast, I think marks speaker uncertainty and orients more toward the speaker’s memory as the basis of uncertainty.
6.
Conclusion
In this paper, I have argued that I guess, a reusable evidential (and not really epistemic) fragment, is frequently employed in American English conversation as a stance frame, which may appear in rather diverse sequential positions in interaction. My aim was to examine what kind of stance is expressed by this marker, and what its role is in the stancetaking activity between conversational co-participants. Despite the different sequential positions established, namely sequenceinitiating actions, responsive actions, and extended tellings or multi-unit turns, the observed commonalities can be summarized as follows.
212 Elise Kärkkäinen
From a subjective, single-speaker vantage point: – The form and decontextualized meaning of I guess is that of a subjective evidential marker. The evidence in question is provided by prior discourse or by visual or sensory evidence, and also via inference (i.e., with no overt evidence in sight). Yet the actual referential meaning of this evidential is not very strong in many instances of use, nor is it easy to determine without recourse to the actual context of use. – I guess projects a certain type of stance to be expressed in the utterance to follow. – Such a stance can variously be said to involve “making an inference,” “coming to a realization,” or “drawing a conclusion,” i.e., I guess indexes a reasoning or inferential process of the speaker, which involves a shift in the speaker’s epistemological stance. From an intersubjective perspective: – The stance projected by I guess is not carried around ready-made but “surfaces” in the sequentially constructed intersubjective weave of dialogue. It arises from, or is based on, evidence or some stimulus gleaned by the speaker (storyteller) from the ongoing or just prior turn(s), from the actual physical environment, or from the wider social context. I guess takes into account these interactional contingencies and displays the recency of the speaker’s inference and displayed stance. – Depending on the sequential position, I guess-framed turns and actions may function to invite others to take a stance, to display that the current speaker gives up or modifies his or her stance (and not only aligns but also often affiliates with the recipient), and to draw the recipients’ attention to the current speaker’s inserted stance (to increase and facilitate their understanding and appreciation of a telling). – Within these actions and turns-at-talk, I guess can be viewed as an organizer of the stancetaking activity between co-participants. It is used to signal that by making public a “just discovered” subjective stance, the speaker invites others to also take a stance, gives up on or adopts a new stance, or alerts recipients to an inserted stance at crucial points in a telling and invites them to join in the stancetaking activity. 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 upcoming stanced digression or side sequence to follow. – Finally, I guess appears at points in the interaction at which stance and stancetaking is highlighted in favor of proceeding with business as usual; in that sense, we can say that stance impinges on the projected trajectory of talk and
The role of I guess in conversational stancetaking 213
thus organizes interaction on a par with other interactional organizations such as coherence, turn-taking, sequential organization, preference organization, and repair organization. As we saw in Section 3.2, Thompson (2002) argues that fragments like I guess provide a certain type of perspective or stance toward the actions (i.e., assessments, claims, counterclaims, and proposals) being done in the associated utterance. In light of my findings, however, it seems that I guess not so much provides a perspective toward the actions, as it signals and displays a just discovered stance to the co-participants. The frame I guess is an essential part of the action being accomplished in the turn. An action, such as an assessment or a claim, can be well accomplished without I guess, but it would then constitute an essentially different type of assessment or claim (cf. example (2) above: I guess that means his broken leg is @doing @okay versus That means his broken leg is doing okay). In other words, I guess signals, projects, and makes more explicit the kind of stancetaking action that is being done in the current turn. We have seen, on one hand, that stance marking can be of a very routinized nature linguistically, in that speakers only use a rather small set of inherently stanced words with some frequency. If, however, we turn our focus to stancetaking as a process and an activity oriented to by participants, 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, unambiguous “stance” anymore. Finally, if we approach stance as something jointly oriented to by the co-participants, we need even more spread in our linguistic description. As I hope to have shown in the above analysis, some linguistic practices of stancetaking go beyond specific, discrete grammatical or lexical devices analyzable in single speakers’ contributions. I have demonstrated some syntactic, semantic, and prosodic resonances between contributions by different speakers (especially when I guess occurs in second position), which are also to be seen as resources for stancetaking and therefore deserve our full attention.
Notes 1. This paper, as also those by Pentti Haddington, Tiina Keisanen, and Mirka Rauniomaa, is based on work done in my research project entitled Interactional Practices and Linguistic Resources of Stance Taking in Spoken English (2002–2006), and has been financed by the Academy of Finland (grants 00381 and 53671). The project is greatly indebted to our collaborator John Du Bois, who has provided us with a view of stance as achieved out in the social world and in interactions with other social agents. My heartfelt thanks are also due to Robert Englebretson for making the 10th Biennial Rice University Linguistics Symposium happen, as well as for
214 Elise Kärkkäinen
many invaluable comments on earlier versions of this paper. Thanks also to all the participants of the Rice symposium for many helpful comments on my paper, and to the two anonymous referees for making many important points and helpful suggestions for revisions. I myself am of course responsible for what use I have made of the comments and suggestions I have received, as also for all the remaining inadequacies. 2. This finding is nicely corroborated by Precht’s work within the “appraisal” framework (cf. Martin 2000, 2003): The resources of language enable a virtually unlimited number of ways in which we could express ourselves, however, my results suggest that we are culturally ‘programmed’ to use a very limited, very specific subset of these options. My previous analysis found more than 1,400 different stanced words in English, and yet we use only about 150 words for ninety percent of our stance expression [..]. Although we have a myriad of options for expressing our emotions, attitudes, and commitments, we tend to use the same small set of stance markers repeatedly. Our expression of stance, I would argue, is shaped by culture and custom – we are socialized to use particular stance markers in particular ways. (Precht 2003: 240, emphases mine) 3. This extremely high frequency is said to be largely due to the use of the clause I think. 4. The frequency of the verb (or noun) guess turned out to be only very little higher than that for I guess (23). By contrast, the total frequencies for the verbs think and know (1,115 and 1,713 respectively) were considerably higher than the ones for I think and I don’t know (729 and 309). 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), namely I don’t think (118), you know (954), and I know (224). 5. The ICE-GB data are essentially collected for corpus linguistics purposes, i.e., to form a computer-searchable syntactically tagged and parsed database, with less attention to transcription detail or quality of sound, and it is in many parts not ideally suited for close interactional analysis. 6. But see for example Quirk et al. (1985: 1112–1113) who hold the opposite view, i.e., considering “comment clauses” like I think to be subordinate to the rest of the sentence. 7. Similar observations about the frequency and formulaic nature of such explicitly subjective collocations have further been made of many other languages. Thus, Weber and Bentivoglio (1991) discuss essentially similar discourse patterns of the Spanish verbs of cognition, creer ‘believe’ and pensar ‘think,’ in spoken data. For Swedish, 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,’ tycka ‘think,’ tänka ‘think,’ minnas ‘remember,’ etc., 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. First-person subjects typically co-occur with mental verbs also in colloquial Finnish (mä luulen ‘I believe/think’) and Estonian (ma arvan ‘I think,’ mai tea ‘I don’t know’; Keevallik 2003), even though one of the frequent Finnish epistemic stance markers, minun mielestä ‘in my opinion’ and minusta ‘I think,’ only displays a first-person subject but no verb (Rauniomaa this volume). 8. I guess is often translatable with just a particle to other languages, e.g., with the Finnish kai or the Swedish väl, both meaning ‘probably, maybe.’ This could be taken as further evidence for its discourse marker status.
The role of I guess in conversational stancetaking 215
9. Where evidentiality fits in with epistemicity and which one is considered the superordinate category varies from one researcher to the next. Chafe (1986), in fact, subsumes epistemic modality under evidentiality, 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, whereas Biber et al. (1999) include under epistemic stance not just markers of certainty, actuality, precision, and limitation but also of source of knowledge and of the perspective from which the information is given. 10. 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.” Subjectivity in this case comes out of and presupposes intersubjectivity, or the speakers engaging with other subjectivities in conversational dialogue (see also Du Bois this volume). 11. This way of framing the upcoming story is very similar to a more explicitly evaluative device for prefacing stories, namely by what Goodwin (1996) has called “prospective indexicals.” Items like problem or a wonderful/terrible thing in a story preface offer a framework for interpretation to the recipients, who do not yet have access to the story but who are expected to respond to it in an appropriate way upon its completion. 12. I thank Robert Englebretson for drawing my attention to this, in fact for three of my examples (6), (9), and (10). 13. I thank the anonymous reviewer for pointing out this obvious similarity to me. 14. This example is analyzed in more detail in Kärkkäinen (2003).
References Anward, J. 2000. “How types of lexical resources emerge from turn-construction.” Paper presented at the Euroconference on Interactional Linguistics, Spa, Belgium, September 16– 21. Beach, W.A. and Metzger, T.R. 1997. “Claiming insufficient knowledge.” Human Communication Research 23: 560–585. Biber, D., Johansson, S., Leech, G., Finegan, E. and Conrad, S. 1999. The Longman Grammar of Spoken and Written English. London: Longman. Clift, R. 2006. “Indexing stance: Reported speech as an interactional evidential.” Journal of Sociolinguistics 10 (5): 569–595. Chafe, W. 1986. “Evidentiality in English conversation and academic writing.” In Evidentiality: The Linguistic Coding of Epistemology, W. Chafe and J. Nichols (eds.), 261–272. Norwood, NJ: Ablex Publishing Corporation. Dahl, Ö. 2000. “Egophoricity in discourse and syntax.” Functions of Language 7 (1): 37–77. Downing, A. 2001. “‘Surely you knew!’ Surely as a marker of evidentiality and stance.” Functions of Language 8 (2): 251–282. Du Bois, J.W. 2001. Towards a Dialogic Syntax. Ms., Department of Linguistics, University of California, Santa Barbara. Du Bois, J.W. 2003. “Stance and consequence in interaction.” Paper given at the Langnet Symposium, University of Oulu, September 11. Du Bois, J.W., Schuetze-Coburn, S., Cumming, S. and Paolino, D. 1993. “An outline of discourse transcription.” In Talking data: Transcription and Coding in Discourse Research, J. Edwards and M. Lampert (eds.), 45–87. Hillsdale, NJ: Lawrence Erlbaum.
216 Elise Kärkkäinen
Du Bois, J.W., Chafe, W.L., Meyer, C. and Thompson, S.A. 2000. Santa Barbara Corpus of Spoken American English, CD, Part 1. Philadelphia: Linguistic Data Consortium. Du Bois, J.W., Chafe, W.L., Meyer, C., Thompson, S.A. and Martey, N. 2003. Santa Barbara Corpus of Spoken American English, CD, Part 2. Philadelphia: Linguistic Data Consortium. Duvallon, O. and Routarinne, S. 2005. ”Parenthesis as a resource in the grammar of conversation.” In Syntax and Lexis in Conversation. Studies on the use of linguistic resources in talkin-interaction, A. Hakulinen and M. Selting (eds.), 45–74. Amsterdam: John Benjamins. Ford, C. and Thompson, S.A. 1996. “Interactional units in conversation: syntactic, intonational, and pragmatic resources for the management of turns.” In Interaction and Grammar, E. Ochs, E. Schegloff and S.A. Thompson (eds.), 134–184. Cambridge: Cambridge University Press. Ford, C., Fox, B. and Thompson, S.A. 2002. “Social interaction and grammar.” In The New Psychology of Language: Cognitive and Functional Approaches to Language Structure, Vol. 2, M. Tomasello (ed), 119–143. Mahwah, NJ: Lawrence Erlbaum. Givón, T. 1993. English Grammar. A Function-Based Introduction. Vol. 1-2. Amsterdam: John Benjamins. Goodwin, C. 1981. Conversational Organization. Interaction between Speakers and Hearers. New York: Academic Press. Goodwin, C. 1996. “Transparent vision.” In Interaction and grammar, E. Ochs, E. Schegloff and S.A. Thompson (eds.), 370–404. Cambridge: Cambridge University Press. Goodwin, C. and Goodwin, M.H. 1992. “Assessment and the construction of context.” In Rethinking Context: Language as an Interactive Phenomenon, A. Duranti and C. Goodwin (eds.), 147–190. Cambridge: Cambridge University Press. Haddington, P. 2003. Oulu Corpus of US / British Television Interviews from 1999–2004. Ms., Department of English, University of Oulu. Helt, M.E. 1997. Discourse Marker and Stance Adverbial Variation in Spoken American English: A Corpus-Based Analysis. Ph.D. dissertation. Department of English, Northern Arizona University. Heritage, J. 1984. “A change-of-state token and aspects of its sequential placement.” In Structures of Social Action: Studies in Conversation Analysis, J.M. Atkinson and J. Heritage (eds.), 299–345. Cambridge: Cambridge University Press. Heritage, J. 2005. “Cognition in discourse.” In Conversation and Cognition, H. te Molder and J. Potter (eds.), 184–202. Cambridge: Cambridge University Press. Holquist, M. 1990. Dialogism. Bakhtin and His World. London: Routledge. International Corpus of English–Great Britain (ICE-GB). Survey of English Usage, University College London. Jefferson, G. 1972. “Side sequences.” In Studies in Social Interaction, D. Sudnow (ed.), 294–338. New York: Free Press. Jefferson, G. 1979. “A technique for inviting laughter and its subsequent acceptance/declination.” In Everyday language: Studies in ethnomethodology, G. Psathas (ed.), 79–96. New York, NY: Irvington Publishers. Jucker, A. and Ziv, Y. 1998. “Discourse markers: Introduction.” In Discourse Markers: Descriptions and Theory, A. Jucker and Y. Ziv (eds.), 1–12. Amsterdam: John Benjamins. Kärkkäinen, E. 2003. Epistemic Stance in English Conversation: A Description of Its Interactional Functions, with a Focus on ‘I think’. Amsterdam: John Benjamins. Kärkkäinen, E. 2006. “Stance taking in conversation: From subjectivity to intersubjectivity.” Text & Talk 26 (6): 699–731.
The role of I guess in conversational stancetaking 217
Karlsson, S. 2003. “Interactional uses for an epistemic marker: The case of ‘jag tycker’/‘tycker jag’ in Swedish.” Melbourne Papers in Linguistics and Applied Linguistics 1: 5–23. Keevallik, L. 2003. From Interaction to Grammar. Estonian Finite Verb Forms in Conversation. Acta Universitatis Upsaliensis. Studia Uralica Upsaliensia 34. Kendon, A. 2004. Gesture: Visible Action as Utterance. Cambridge: Cambridge University Press. Lenk, U. 1998. Marking Discourse Coherence. Functions of Discourse Markers in Spoken English. Tübingen: Gunter Narr. Lerner, G. 1991. “On the syntax of sentences-in-progress.” Language in Society 20 (3): 441– 458. Lerner, G. 1992. “Assisted storytelling: Deploying shared knowledge as a practical matter.” Qualitative Sociology 15 (3): 247–271. Lerner, G. 1996. “Finding ‘face’ in the preference structures of talk-in-interaction.” Social Psychological Quarterly 59 (4): 303–321. Local, J. 1992. “Continuing and restarting.” In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 273–296. Amsterdam: John Benjamins. Mandelbaum, J. 1989. “Interpersonal activities in conversational storytelling.” Western Journal of Speech Communication 53: 114–126. Mandelbaum, J. 1993. “Assigning responsibility in conversational storytelling.” Text 13 (2): 247–266. Martin, J.R. 2000. “Beyond exchange: APPRAISAL systems in English.” In Evaluation in Text. Authorial Stance and the Construction of Discourse, S. Hunston and G. Thompson (eds.), 142–175. New York: Oxford University Press. Martin, J.R. 2003. “Introduction.” Text. Special issue: Negotiating Heteroglossia: Social Perspectives on Evaluation, 23 (2): 171–181. Ochs, E. 1992. “Linguistic resources for socializing humanity.” In Rethinking Linguistic Relativity, J. Gumperz and S. Levinson (eds.), 407–437. Cambridge: Cambridge University Press. Ochs, E. and Schieffelin, B. 1989. “Language has a heart.” Text 9 (1): 7–25. Pomerantz, A. 1984. “Agreeing and disagreeing with assessments: Some features of preferred/ dispreferred turn shapes.” In Structures of Social Action. Studies in Conversation Analysis, J.M. Atkinson and J. Heritage (eds.), 57–101. Cambridge: Cambridge University Press. Pomerantz, A. 1986. “Extreme case formulations.” Human Studies 9: 219–230. Potter, J. 1998. “Discursive social psychology: From attitudes to evaluations.” European Review of Social Psychology 9: 233–266. Potter, J. and Hepburn, A. 2003. “‘I’m a bit concerned’ – early actions and psychological constructions in a child protection helpline.” Research on Language and Social Interaction 36 (3): 197–240. Precht, K. 2003. “Stance moods in spoken English: Evidentiality and affect in British and American conversation.” Text. Special issue: Negotiating Heteroglossia: Social Perspectives on Evaluation 23 (2): 239–257. Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. 1985. A Comprehensive Grammar of the English Language. London: Longman. Schegloff, E. 1996. “Turn organization: On the intersection of grammar and interaction.” In Interaction and Grammar, E. Ochs, E. Schegloff and S.A. Thompson (eds.), 52–133. Cambridge: Cambridge University Press. Scheibman, J. 2000. “I dunno ... A usage-based account of the phonological reduction of don’t in American English conversation.” Journal of Pragmatics 32: 105–124.
218 Elise Kärkkäinen
Scheibman, J. 2001. “Local patterns of subjectivity in person and verb type in American English conversation.” In Frequency and the Emergence of Linguistic Structure, J. Bybee and P. Hopper (eds.), 61–89. Amsterdam: John Benjamins. Scheibman, J. 2002. Point of View and Grammar: Structural Patterns of Subjectivity in American English Conversation. Amsterdam: John Benjamins. Schiffrin, D. 1987. Discourse Markers. Cambridge: Cambridge University Press. Thompson, S.A. 2002. “‘Object complements’ and conversation: Towards a realistic account.” Studies in Language 26 (1): 125–163. Thompson, S.A. and Mulac, A. 1991a. “A quantitative perspective on the grammaticization of epistemic parentheticals in English.” In Approaches to Grammaticalization. Vol. 1–2, E.C. Traugott and B. Heine (eds.), 313–329. Amsterdam: John Benjamins. Thompson, S.A. and Mulac, A. 1991b. “The discourse conditions for the use of the complementizer that in conversational English.” Journal of Pragmatics 15: 237–251. Traugott, E.C. 1995. “Subjectification in grammaticalisation.” In Subjectivity and Subjectivisation: Linguistic Perspectives, D. Stein and S. Wright (eds.), 31–54. Cambridge: Cambridge University Press. Voloshinov, V.N. [1930] 1973. “Verbal interaction.” In Marxism and the Philosophy of Language, Chap. 3, 83–98. Cambridge, MA: Harvard University Press. Weber, E. and Bentivoglio, P. 1991. “Verbs of cognition in spoken Spanish: A discourse profile.” In Discourse–Pragmatics and Verb. The Evidence from Romance, S. Fleischman and L. Waugh (eds.), 194–213. London: Routledge.
Appendix: Symbols used in transcription (From Du Bois et al. 1993) Units Intonation unit Truncated intonation unit Truncated word Transitional continuity Final Continuing Appeal (seeking a validating response from listener)
{carriage return} --
. , ?
Speakers Speech overlap (numbers inside brackets index overlaps)
[ ]
Accent and lengthening Primary accent (prominent pitch movement carrying intonational meaning)
^
Secondary accent Unaccented Lengthening Pause Long and medium (length indicated in seconds) Short (brief break in speech rhythm) Vocal noises e.g., (TSK), (SNIFF), (YAWN), (DRINK) Glottal stop Exhalation Inhalation Laughter (one pulse) Laughter during speech (1–5 words) (e.g. @two @words) Laughter during speech (+6 words) (e.g. words )
The role of I guess in conversational stancetaking 219
` =
…(1.0) ..
% (Hx) (H) @ @ @
Quality Special voice quality Forte: loud Higher pitch level Lowered pitch level Parenthetical prosody Allegro: rapid speech Marcato: each word distinct and emphasized Yawning
words words words words words words <MRC>words words
Transcriber’s perspective Uncertain hearing Uncertain Indecipherable syllable
<X> words #word X
Stance markers in spoken Finnish Minun mielestä and minusta in assessments Mirka Rauniomaa
University of Oulu, Finland
1.
Introduction1
The expressions minun mielestä and minusta, which frequently occur in almost any Finnish interaction, have drawn analysts’ attention before but mainly ended up in side comments and footnotes. The purpose of this paper is to contribute to the understanding of the two expressions by focusing solely on their locally situated use, particularly in the context of assessments. 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. Indeed, the analysis is guided by the idea that social interaction organizes linguistic phenomena, as posited in interactional linguistics (see, e.g., Ford et al. 2002; Ochs et al. 1996; Selting and Couper-Kuhlen 2001). The central theme pursued in this paper is the notion of stancetaking, i.e., participants’ joint construction of evaluations, attitudes, affective stances, etc. in dialogic interaction (Du Bois this volume; Kärkkäinen 2006, this volume). Before proceeding to a discussion of stancetaking, however, a closer look at the expressions is in order. In what follows, I will discuss minun mielestä and minusta in terms of their morphosyntactic form and with relation to previous analyses. I will also briefly introduce the data used in this study (Section 2). The expression minun mielestä consists of the first-person-singular pronoun minä in its genitive case and the noun mieli ‘mind’ in its elative, i.e., inner locative, case. The literal translation of minun mielestä would thus be ‘from my mind’ (or ‘out of my mind’); other translations into English include ‘in my opinion,’ ‘to my mind,’ and ‘it seems to me.’ Similar meanings can be attributed to the expression minusta, 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, the ending takes either the back vowel a or
222 Mirka Rauniomaa
the front vowel ä). Literally minusta would translate as ‘from me’ (or ‘out of me’); the most appropriate translation is ‘I think.’ Both expressions are often reduced phonetically, to the extent that minun mielestä may come close to sounding like minusta, and minusta may take an even more reduced form. A comprehensive analysis would certainly be useful, but I will not explore the phonetic realizations of the expressions in any detail here. It is worth pointing out, however, 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. For this reason, and because of data limitations, I do not attempt to distinguish between them at this point. Nonetheless, whenever the analysis permits, I keep the two expressions separate to encourage further investigations into their possibly differentiated use (see Luukka 1992a: 118–119, 139–140, 1992b: 373, for observations on the diverse distribution of the expressions in spoken and written academic discourse). In previous literature, minun mielestä and minusta have commonly been described as hedges that are used to obscure the import of an utterance in order to address issues of politeness, i.e., saving the speaker’s or a co-participant’s face (Aalto 1997: 58–61; Hakulinen et al. 1989: 118–123; Lampinen 1990: 84–85; Luukka 1992a: 91–92, 108, 1992b: 363–373; Nikula 1996: 115–121, 135–136). More specifically, they have been considered to represent a particular type of hedge that mitigates speakers’ commitment to the truth of a proposition (Hakulinen et al. 1989: 118–119; drawing on Prince et al. 1982: 85). In this way, the expressions delimit the generalizability of an utterance by framing it not as a fact but as an opinion that the co-participants may disagree with (Luukka 1992a: 108, 1992b: 372– 373). 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. As an example, 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. Overall, previous studies provide extensive overviews of linguistic modality in general and, consequently, only touch upon such individual expressions as minun mielestä and minusta. What is more, they consider the two expressions to mark speakers’ individual, subjective stances. The current paper builds on this foundation to bring in the intersubjective aspects of stancetaking. I refrain from the predetermined categories of politeness and carefully examine minun mielestä and minusta in the specific sequential contexts in which they occur. In this way, the present paper parallels Kärkkäinen’s (this volume) exploration of I guess and, more importantly, her (Kärkkäinen 2003) study on I think, an expression that is more or less equivalent to the Finnish minun mielestä and minusta (see also
Stance markers in spoken Finnish 223
Karlsson 2003 for a discussion of the Swedish jag tycker/tycker jag ‘I think’). Kärkkäinen (2003) identifies several local purposes for the use of I think: depending on its sequential positioning, I think marks topical and other boundaries, brings in a personalized speaker perspective, displays on-line planning, signals turn completion or pursues a response from the recipients. As Section 5 will show, minun mielestä and minusta can be seen to serve similar interactional functions. The remarks that have been made on minun mielestä and minusta in previous literature suggest that the two items offer a routinized way for speakers to put their stamp on an utterance. This paper proposes that the routinization is also evident at a sequential level. 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. As access to knowledge plays a major part in some cases in the present data, minun mielestä and minusta could be further categorized as epistemic stance markers. Nonetheless, 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. After a presentation of the data in Section 2, I will discuss stancetaking and assessments in some detail in Section 3. 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). The bulk of the paper, Section 5, presents the three interactional functions of the stance markers that can be found in the data: projecting disagreement in a second assessment, marking transition to a first assessment across turns, and marking transition to a first assessment within an extended turn. I will provide more evidence of the contingent nature of the stance markers by presenting a special case in Section 6, before a discussion of the findings in the conclusion (Section 7).
2.
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). They amount to approximately two and a half hours, or around 34,000 words, of casual conversation: there are 2 face-to-face conversations, each lasting for about an hour, and 12 telephone conversations, ranging in length between 1 and 14 minutes. The data were originally transcribed using the transcription conventions developed by Gail Jefferson (as laid out in Seppänen 1997), but for the present study, I have re-transcribed the relevant parts according to the intonation-unit-based discourse transcription system (Du Bois et al. 1992; Du Bois et al. 1993). In addition to a free English translation, I have
224 Mirka Rauniomaa
added a morpheme-by-morpheme gloss for each line of transcript. Keys to relevant transcription and glossing symbols are provided in Appendices A and B. It should be pointed out that the speakers in these data come mainly from southern and eastern Finland, which represent two distinct dialect areas. This explains the varied spellings of the stance markers under examination (e.g., minun mielestä ~ mun/miun mielest(ä); minusta ~ must(a)/miust(a)) and possibly of some other words in the transcripts. In referring to the stance markers in general, I use the full forms minun mielestä2 and minusta, but I deem it important that the evident dialect differences be reflected in the transcripts. All in all, 67 occurrences of the stance markers can be found in the data. Minun mielestä is somewhat more frequent with a total of 42 cases, minusta being represented by a total of 25 cases. Examples (1) and (2) typify the expressions minun mielestä and minusta, respectively. (1) sg067_a2 Lawyer Issues R: Se on mu-n miele-stä tosi hyvä kirja. it be:3sg I-gen mind-ela really good book ‘In my opinion it is a really good book.’ (2) sg151_a Chat about Something I2: Mut ol-i miu-sta ne hyvä-n-näköse-t. but be-pst:3sg I-ela they good-gen-looking-pl ‘But I think they were good-looking.’
Although it is possible to use mielestä or simply the elative case ending -stA with other person references, e.g. with another personal pronoun or with a proper name, such cases are rare in comparison with the first-person-singular forms. In the present data, there is only one case of a second-person reference (example 3) and one case of a third-person reference, in which the speaker refers to a cat called Perttu (example 4). (3) sg151_a Chat about Something I2: Mut on-k-s nää nyt siu-n miele-st sit hyvä-t. but be:3sg-q-cli these now you-gen mind-ela then good-pl ‘But are these ones good in your opinion.’ (4) sg067_a2 Lawyer Issues R: Se-n mie- Pertu-n miele-stä se on aina just ede-ssä, 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,’
Stance markers in spoken Finnish 225
Examples (3) and (4) represent individual cases; the stance-marker elements mielestä and -stA are overwhelmingly used with first-person reference in the current data. This observation corresponds with Helasvuo’s (2001a: 32) finding that first-person subjects typically co-occur with mental verbs, or verbs of cognition, in spoken Finnish. Similar collocations between first-person referents and stanceinferring expressions in subject-verb combinations have also been observed in other languages (see Dahl 2000 for Swedish; Keevallik 2003: 74–100 for Estonian; Kärkkäinen 2003; Scheibman 2001 for English; Weber and Bentivoglio 1991 for Spanish). The Finnish minun mielestä and minusta, as adverbial noun phrases including an explicit reference to cognition and/or the speaker, can also be seen as indices of the strong tendency in interactive discourse for speakers to personalize what they are saying, i.e., to talk from their particular point of view (see Bybee and Hopper 2001). It has to be added, however, that even this personalization, or subjectivity, can only be constructed in interaction with other participants. The interactive nature of stancetaking will be discussed next.
3.
Stancetaking through assessments
It is a well-established observation that when producing an utterance, speakers leave in it traces of their attitudes, emotions, and so on, toward whatever they are talking about (see Thompson and Hunston 2000 for a review). What is new is the shift in focus from individual stances to stancetaking, i.e., from single-speaker contributions to the joint construction of stance at both linguistic and interactional levels (see Wu 2004: 3–19 for a review). Following this line of study, I examine the linguistic items that are the focus here, minun mielestä and minusta, within their local sequential contexts and consider them to arise contingently from those contexts. In other words, minun mielestä and minusta are here regarded to acquire their status as stance markers only in and through talk-in-interaction. 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, positioning the self, and aligning with other subjects. Almost any utterance helps to locate the speaker’s perspective or point of view among an infinite set of possible ones. Within this view of stance, Du Bois emphasizes the significance of the stance differential, a subtle difference in the participants’ alignment with each other, which gradually rises out of their contributions. To put it differently, even if participants fully agree on a matter, their specific utterances are affected by the fact that one precedes, follows, or overlaps the other. Stancetaking is a social activity that is not restricted to any specific action type. Nonetheless, certain action types can be considered to be particularly rele-
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. In Finnish, assessments typically follow more or less the same syntactic form as in English (Tainio 1996: 85), 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, assessments include an evaluation (assessment 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.g., by choice of tense in the copula (Goodwin and Goodwin 1992: 165). What is more, assessments make relevant a type-connected uptake (Pomerantz 1984) and thus facilitate participants’ alignment with each other. Assessments clearly represent the speaker’s perspective, and there is no need for an explicit reference to the speaker. Nevertheless, as the following sections will show, minun mielestä and minusta frequently occur within assessments, accomplishing orientational work.
4.
Positions of minun mielestä and minusta
The way in which a single utterance is constructed is consequential to its interpretation. It is therefore important to consider the immediate context of the stance markers minun mielestä and minusta. I will here discuss two levels of positioning, with a special focus on stance markers within assessments. Firstly, I will examine where the stance markers are placed within the intonation unit in which they occur. Secondly, I will make a few remarks on the stance markers’ possible positions in the sequence of actions, i.e., whether the stance markers occur in initiating or responding turns. The four possible positions that the stance markers can take in the intonation unit (IU) are presented in Table 1. Table shows that there is a tendency for both stance markers, especially minusta, to be placed at the start of an intonation unit. Helasvuo (2001b: 145–149) has found that the clause core (i.e., the predicate and its core arguments) is typically produced within one intonation unit in Finnish. 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, 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. 1989: 136). In the case of the English I think, Kärkkäinen argues that its IU-initial positioning “helps recipients to align themselves to what is coming” (2003: 115). In other words, stance markers orient participants to treat
Stance markers in spoken Finnish 227
Table 1. Position of stance markers minun mielestä, and minusta in the intonation unit (assessments in parentheses) minun mielestä minusta Total
Initial*
Medial
Final
Separate
Total
24 (13) 20 (5) 44 (18)
15 (6) 5 (4) 20 (10)
1 (0)
2 (1)
1 (0)
2 (1)
42 (20) 25 (9) 67 (29)
* Cases in which a stance marker is preceded by a discourse particle such as no ‘well’, niin ‘so’ and mutta ‘but’ are included within this group (see Hakulinen et al. 1989 for a discussion of discourse particles in Finnish).
what follows as possibly requiring a response and in any case relevant to their stancetaking. Additionally, if an IU-initial position coincides with a turn-initial one, 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. 1974: 722). When the stance markers are positioned somewhere in the middle of an intonation unit, only a limited set of items can typically be placed in front.3 These include the negation element ei,4 the verb on ‘is,’ pronouns, or a combination of these. 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, and the favoured word order is SVO, with given information preceding new information in a clause and especially with pronominal subjects preceding the verb (Helasvuo 2001b: 76–78). It is striking that in all the cases in which a stance marker is placed IU-medially in an assessment, the evaluative element comes only after the stance marker, either in the same intonation unit or in the subsequent one. This suggests that IU-medial stance markers orient participants to what follows in much the same way as IU-initial ones. As Table 1 reveals, there is only one case of an IU-final stance marker in the data. It occurs in the utterance no ei mun mielestä ‘well not in my opinion,’ which is produced as an answer to a previous speaker’s question, and which is syntactically tied to that question. 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. Motivation for such use of the stance markers is also inextricably tied to the ongoing interaction, as example (6) will later demonstrate. As for their possible positions in a somewhat larger context, both stance markers can be in turns that initiate new sequences or respond to a previous action. Two types of action that can take place in either of these positions figure prominently in the data: statements, or information delivery in a wider sense (Heritage 1984), and assessments, which typically come in pairs of first and second assessments (Pomerantz 1984). In the present data, it is common for both stance markers to occur in assessments: 20 cases of minun mielestä (48%) and 9 of minusta
228 Mirka Rauniomaa
(36%) are produced within an utterance accomplishing such an action. That is to say, the stance markers are frequently used in a context that can also otherwise be identified as being about stancetaking. These observations quite clearly indicate that the stance markers minun mielestä and minusta are made relevant by the ongoing interaction. They do not in themselves reveal speakers’ stances but point out how an utterance is to be understood there and then. This is the case especially when the stance markers occur within assessments, i.e., when there are clearly evaluative elements either within the intonation unit itself or in an adjacent intonation unit within the same turn. I will examine such cases in detail in the following sections.
5.
Minun mielestä and minusta orienting participants to stancetaking
Section 4 has shown that in the present data minun mielestä and minusta are often either embedded in or followed by an assessment, and an assessment clearly contributes to stancetaking already on its own. An interesting question can then be raised: What is it that makes stance markers relevant in assessments? Generally put, they orient participants to the stancetaking. More specifically, the stance markers can be seen to perform three different functions: projecting disagreement in a second assessment, marking transition to a first assessment across turns, and marking transition to a first assessment within an extended turn.
5.1
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; Hakulinen et al. 1989: 136). Among the assessments in the present data, however, this function is relatively rare. I present it here first because of its clear trajectory: a speaker makes a first assessment, and a recipient makes a disagreeing second assessment that includes a stance marker. In example (5), Kaarina and Reija are discussing the characteristics of people coming from different parts of Finland. Kaarina reports on what someone has said about people who come from Savo, the area in eastern Finland that Reija is from.
Stance markers in spoken Finnish 229
(5) sg067_a1 Lawyer Issues 1 K: Sit se v- sano vaan et, then it say:pst:3sg just that ‘Then he j- just said that,’ 2 ...(0.3) 3 et niinku, that prt ‘that like,’ 4 (H) et ne on helv[eti-n ka]teellis-i-i ihmis-i-ä. that they be:3sg hell-gen jealous-pl-ptv person-pl-ptv ‘(H) that they are damn jealous people.’ 5 R: [(SNIFF)] 6 ...(0.4) 7 Mu-n miele-st se on paska-puhet-[2ta2]. I-gen mind-ela it be:3sg shit-talk-ptv ‘In my opinion it is bullshit.’ 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], apparently it prt ‘E- apparently it like,’ 10 R: [3(SNIFF)3] 11 K: .. jotenki tule-e esiin tollase-s pikku-kaupungi-s. somehow come-3sg into.view like.that-ine small-town-ine ‘.. somehow comes up in a small town like that.’ 12 ...(0.8) 13 R: Nii joo. prt prt ‘Oh yeah.’5
Embedded in Kaarina’s report is an assessment made by a third party: it is introduced with a reporting clause on line 1, sit se v- sano vaan et ‘then he j- just said that,’ and realized as a that-clause on line 4, et ne on helvetin kateellisii ihmisiä ‘that they are damn jealous people.’ The report does not convey the speaker’s own stance toward the matter at hand. Nonetheless, the report is stance-inducing in
230 Mirka Rauniomaa
that it makes relevant an uptake, i.e., a display of the telling’s impact, by the recipient (see Heritage 1984). The first syllable of kateellisii ‘jealous’ is produced with relatively loud voice, which highlights the point of the utterance and the relevance of an uptake. A response from Reija is particularly relevant here as she represents the group of people that has been assessed. Kaarina’s turn accomplishes two actions that Reija can respond to, i.e., the report as such and the assessment within it. On line 7, Reija deals with both by producing an assessment of the reported assessment, mun mielest se on paskapuhetta ‘in my opinion it is bullshit.’ The predicate noun paskapuhetta ‘shit talk’ or ‘bullshit’ marks negative evaluation of and strong disagreement with the third-party assessment, which is here referred to with the pronoun se ‘it.’ There are several features in Reija’s turn that project disagreement. Firstly, Reija does not respond immediately after Kaarina’s turn nor in overlap with it; instead there is a 0.4-second pause on line 6, which may be interpreted as a means of stalling disagreement (Pomerantz 1984: 65). Secondly, in Finnish, first assessments are usually responded to by particles, or particle and finite verb combinations, rather than fully-fledged second assessments (Tainio 1993: 153–154; as reported in Tainio 1996: 108). This suggests that as soon as the utterance is recognized to consist of something other than such a particle, i.e., in Reija’s producing mun mielest, it can be understood to be an atypical and possibly dispreferred response (see Pomerantz 1984 for a discussion of preference in assessments). Thirdly, the stance marker mun mielest projects contrast by explicitly marking the start of the speaker’s viewpoint, which in this case would not be necessary were the assessment displaying agreement (see Aalto 1997: 65; Hakulinen et al. 1989: 136). The stance marker does special work in projecting disagreement: it shows the speaker’s sensitivity toward the fact that, on the one hand, a stance has already been reported on, and, on the other hand, another stance is yet to be taken. In other words, the stance marker acknowledges the existence of several possible stances and locates the speaker’s stance among them. The fact that Kaarina from line 8 onward defends the reported speaker by giving an account for the third-party assessment suggests that she may sympathize with both Reija’s and the reported speaker’s stances. Had Reija not used the stance marker on line 7, 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, a plain assessment by her would have borne a sense of authority and finality, making it hard for Kaarina to differ on the issue. Starting the disagreeing second assessment with a stance marker, Reija leaves Kaarina more room to manoeuvre between the reported stance and the one that Reija takes. Here, minun mielestä indeed has a softening or modifying effect, as reported in earlier remarks on the stance markers (e.g., Aalto 1997: 65; Hakulinen et al. 1989: 136).
Stance markers in spoken Finnish 231
The amount of disagreements that are framed with the stance markers is limited in these data. 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 reported speaker. However, there are also cases in which the assessment sequence is more straightforward: in example (6), the stance marker mun mielestä projects disagreement with a co-participant. Here, Kaarina and Reija talk about Joni, with whom they are both acquainted. Kaarina has introduced Joni into their talk a little earlier by asking Reija to confirm a piece of information about his past whereabouts. Kaarina has then asked several questions about him, and Reija has provided answers as well as some extra information about his social life, which Kaarina has marked as news to her. Reija has thus proved to be the more informed participant on this topic by the time the participants get to the segment presented here. The main focus is on the stance marker on line 5, but I will also briefly comment on the one on line 2.6 (N.B. Reija’s first intonation unit is divided into two misaligned lines due to limitations of space.) (6) sg067_a1 Lawyer Issues 1 K: Se on jotenki semmonen vähän ällö, it be:3sg somehow such a.little disgusting ‘He is somehow a bit disgusting,’ 2
mu-n mie[le-stä
]. I-gen mind-ela ‘in my opinion.’ 3 [((KNOCK))] ((KNOCK)) 4 R: (H) Mut se-ki on %m= - but it-cli be:3sg ‘(H) But it/he is also m= --’ 5 .. %Se on mu-n miele-stä ol-lu pitkään it be:3sg I-gen mind-ela be-pcp long ‘.. 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, actually when it-ill acquaint-3sg better ‘actually when Ø gets to know him better,’Joo
3]. prt ‘Yeah.’ettei ihan pala
(Hx)]. that-neg quite burn ‘so that Ø doesn’t burn completely.’<X> X (( ))
]. [(GROAN)] (1.2) Well it’s no ^worse than her ^screaming at em, `is it? (1.0) `Yeah but `now you’ll have ^both. (0.6) `Yeah ^right. (0.4) Probably be like, , you `know, XX? `Oh= ^Go=d.
On lines 1–4, Harold topicalizes the neighbor’s pregnancy that was first mentioned some time ago. He asks Jamie for the basis or the source of information for her announcement that their annoying neighbor is pregnant again. Harold thereby expresses doubt toward the accuracy of the news. Jamie responds not by giving such an account, but by stating simply that She’s ^pregnant. She’s ^totally `pregnant (lines 5–6). Harold acknowledges this briefly with Oh (line 7). Jamie, in turn, continues to attend to the inadequacy of her response by excluding one of the most obvious counterarguments against pregnancy with .. It’s ^not .. ^eating too `much, she’s ^pregnant (lines 8–9). 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=, I `guess, (1.2) I `mean thi- this- thi- this just ^happened? (lines 10–13). However, Jamie does not respond to Harold’s inquiry. Instead, she moves on to complain about the situation by mentioning one negative outcome that the pregnancy will have on their lives, We’re gonna have `babies ^crying (line
Challenging the prior speaker 275
15). After a pause, during which no one responds (line 16), Jamie adds,
(line 17), thereby upgrading the complaint. During the latter part of Jamie’s complaint Harold groans (line 18), which shows recipiency and acknowledges that the complaint requires uptake. However, the long pause (line 19) pre-monitors that the complaint will not get the aligning response it invites. Harold produces a tag-question challenge that targets the basis of the complaint, Well it’s no ^worse than her ^screaming at em, `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, 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. In addition to being a disaligning second-position turn, the tag question is marked with turn-initial well to show disalignment or some discontinuity with the prior turn (Pomerantz 1984b). Moreover, no worse sets up an opposition with the prior turn, and the evaluative verb screaming explicates Harold’s negative stance toward the noise the mother makes when communicating with her children. As was the case in example (7), it can be noted that in the current example Harold’s challenge is directed at the projected activity (complaining), and the actual content of the turn may be seen to include elements that show support or the like. 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. In other words, Harold uses the tag-question turn to challenge the appropriateness of Jamie’s complaint by indicating and/or reminding Jamie that the situation is already bad, and that it cannot get much worse, no matter how much the neighbor’s new baby will cry. However, instead of Jamie responding to Harold’s challenge, Pete takes a turn to counter the challenge with `Yeah but `now you’ll have ^both (line 23). The turn indicates that the two scenarios presented by Jamie and Harold are equally bad, and therefore equally correct. In other words, Pete counters Harold’s challenge by challenging its grounds. Jamie, in turn, seconds Pete with `Yeah ^right (line 25). To sum up, similarly to Section 3.1, in the present section the challenging tag questions and negative yes/no interrogatives likewise point toward some discrepancy in information between the participants. In response to initiating actions such as complaints, the discrepancy is used to challenge the relevance or the appropriateness of doing the action in the prior turn. As such, the negative yes/no interrogatives and tag questions offer a disaligning response to the initiating action. 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, the different constructions are not interchangeable. In excerpt (2), for example, the challenging turn consists of a negative declarative followed
276 Tiina Keisanen
by a positive tag, and is used to challenge a claim (explicit or implied) or position of positive polarity. On the other hand, the negative yes/no interrogatives in excerpts (5) and (7), for example, challenge a claim (explicit or implied) or stance of negative polarity. However, the differences between the constructions are not well studied, and other differences are also likely to exist. Future work on this topic is necessary to elucidate the exact differences in more detail. The examples in the sections above provide evidence that sometimes discourse participants can prioritize holding prior speakers accountable for their claims and stances or positions the have taken, even implicitly, rather than align with the projected course of action. So, even though the preference for agreement (Sacks [1973]1987) limits the occurrence of disagreement in social interaction, prior stances can be contested and challenged, as the interrogative constructions discussed here demonstrate. Moreover, on the sequential level such challenges can function as a vehicle with which to exert influence on the course of the unfolding action. Sequential positioning is the central factor in the interpretation of the negative yes/no interrogatives and tag questions discussed here as challenges. However, in addition to sequential positioning, the placement of the interrogative turns in the stream of talk also contributes to their interpretation. More specifically, the smooth transition from a turn to a next with no gap and no overlap (Sacks et al. 1974) is regularly disturbed by the challenging negative yes/no interrogatives and tag questions. In about half of the cases in the database, the challenges are placed in overlap with prior talk (e.g., excerpts 3 and 4). Recipients are also observed to rather frequently withdraw or delay their response to the challenging turns (e.g., excerpts 5 and 9). These are some further indications of the disaligning and dispreferred nature of the interrogatives discussed here. The linguistic and prosodic turn-design features also contribute to the construction of disalignment. First, the challenges are regularly designed to negate or contradict a claim or position embedded in the prior turn, while the interrogative design is used for requesting the recipient to account for this claimed discrepancy. The turns may also include lexical markers, such as different (example 3), only (example 4), really (example 2), and no worse (example 9), which help to identify the problematic portion of the prior turn. In terms of linguistic practices of stancetaking, these can be seen to function as restrictive stance markers in their local context. On the other hand, the use of marked prosody, especially of high pitch, can be used to add an affective stance to the challenging turns.
4.
Challenging the prior speaker 277
Conclusion
The speakers of negative yes/no interrogatives and tag questions discussed here initiate a sequence that ostensibly concerns some piece of information. However, due to the action of challenging carried out with the interrogatives, on the level of social relations these sequences involve the issue of alignment between the discourse participants. What is notable is that the negotiation of alignment does not surface in the interaction explicitly, but as I hope to have demonstrated, it is still available to analysis through the examination of sequence organization, the distribution of silence and talk, as well as turn-design features of the turns in question. In other words, the patterns of stancetaking discussed here expand on two different, but interlocking levels. On the one hand, stancetaking can be conceptualized as alignment or disalignment between discourse participants with respect to the projected course of action or of the sequence. The interrogatives discussed here provide the discourse participants with a resource for doing disalignment, specifically to challenge the projected or ongoing course of action. On the other hand, when the features of the aligning or disaligning actions are examined in more detail, it is observed that actions may embody epistemic as well as affective stance displays toward not only one’s own turn, but also toward others’ turns. In the data here, the challenging actions are used to display doubt toward prior claims or stances taken, in addition to which the prosodic turn-design features can be used to add an affective component to the turn. As demonstrated by Pomerantz (1984b) in her discussion on agreement in assessment sequences, delay devices such as pauses and repair initiators are commonly found in disagreeing turns that are produced in response to turns that invite agreement. Delays, repair initiators, and reformulations are found also in the sequences discussed here. They indicate troubles in turn transition, and point toward an interpretation that the recipients regard the interrogative turns as indicative of disalignment. However, 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, such as, for example, the information conveyed in statements of opinion often is. This may cause trouble for the recipient in interpreting the relevance of an interrogative. That is, interrogatives 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. Such sequences draw attention to the fact that discourse participants can be held accountable for anything that they produce in interaction, even implicitly. The data discussed here provide evidence for stancetaking as an intersubjective achievement between discourse participants. By singling out some (explicitly or implicitly expressed) claim or position in the prior turn, discourse participants
278 Tiina Keisanen
can call the interactional positioning implicated in the voicing of such claims or positions into question. On occasion, then, taking a stance can be treated as an accountable action.
Notes 1. This chapter is based on a section in my dissertation research (Keisanen 2006). Acta Universitatis Ouluensis is gratefully acknowledged for allowing the reprinting of the relevant materials. I would also like to thank the participants of the 10th Biennial Rice Linguistics Symposium held March 31 – April 3, 2004 at Rice University where the ideas for this paper were first presented, and the Academy of Finland (project number 53671) for the funding of the research. Moreover, John Du Bois, Robert Englebretson, Pentti Haddington, Elise Kärkkäinen, Maarit Niemelä, Mirka Rauniomaa, Sandra Thompson and two anonymous reviewers have provided constructive critique and insightful comments on the paper for which I am very grateful. All the remaining inaccuracies are mine alone. 2. While outside factors to the ongoing conversation are not favored within conversation analysis, in this case the fact that the discourse participants are siblings seems to contribute to the challenging nature of the tag question turn. In other words, as sisters the three women might be expected to more or less know and remember each other’s prior teachers. Based on this knowledge base that seems to be evoked and claimed as common (cf. Heritage 2002), the tag question turn is produced from a strong epistemic position and thereby carries out a challenge (cf. Koshik 2003), rather than being used, for example, as an epistemically less strong (skeptical) newsmarker, which it might well be between less intimate participants. However, despite the label chosen for the action (cf. Pomerantz 1990), the main argument here, as well as in the other cases, is that the interrogative turn displays doubt towards the prior speaker’s claim or position. 3. Similar to example (2), 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. However, the high rising final pitch might be used here to mitigate the degree of the hostility component that has been associated with challenging actions (Heritage 2002). 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, for example, high rising final pitch can be used to make the turn appear less hostile and/or challenging.
References Beach, W. and Metzger, T. 1997. “Claiming insufficient knowledge.” Human Communication Research 23 (4): 560–585. Clift, R. 2000. “Stance-taking in reported speech.” Essex Research Reports in Linguistics 32. Department of Language and Linguistics, University of Essex. Drew, P. 1997. “‘Open’ class repair initiators in response to sequential sources of troubles in conversation.” Journal of Pragmatics 28: 69–101.
Challenging the prior speaker 279
Drew, P. 1998. “Complaints about transgressions and misconduct.” Research on Language and Social Interaction 31 (3, 4): 295–325. Du Bois, J.W., Chafe, W., Meyer, C. and Thompson, S.A. 2000. Santa Barbara Corpus of Spoken American English, CD, Part 1. Philadelphia: Linguistic Data Consortium, University of Pennsylvania. Du Bois, J.W., Chafe, W., Meyer, C., Thompson, S.A. and Martey, N. 2003. Santa Barbara Corpus of Spoken American English, CD, Part 2. Philadelphia: Linguistic Data Consortium, University of Pennsylvania. Du Bois, J.W. and Englebretson, R. 2004. Santa Barbara Corpus of Spoken American English, DVD, Part 3. Philadelphia: Linguistic Data Consortium, University of Pennsylvania. Du Bois, J.W., Schuetze-Coburn, S., Cumming, S. and Paolino, D. 1993. “An outline of discourse transcription.” In Talking Data: Transcription and Coding in Discourse Research, J.A. Edwards and M.D. Lampert (eds.), 45–87. Hillsdale, NJ: Lawrence Erlbaum. Ford, C. 2004. “Contingency and units in interaction.” Discourse Studies 6 (1): 27–52. Ford, C., Fox, B. and Thompson, S.A. 2003. “Social interaction and grammar.” In The New Psychology of Language. Cognitive and Functional Approaches to Language Structure, M. Tomasello (ed.), 119–144. Mahwah, NJ: Lawrence Erlbaum. Freese, J. and Maynard, D. 1998. “Prosodic features of bad news and good news in conversation.” Language in Society 27: 195–219. Givón, T. 2001. Syntax: An Introduction, Vol. 2. Amsterdam: John Benjamins. Goodwin, C. 2002. “Time in action.” Current Anthropology 43 (S4): S19–S35. Heritage, J. 1984. “A change-of-state token and aspects of its sequential placement.” In Structures of Social Action. Studies in Conversation Analysis, M.J. Atkinson and J. Heritage (eds.), 299–345. Cambridge: Cambridge University Press. Heritage, J. 1998. “Oh-prefaced responses of inquiry.” Language in Society 27: 291–334. Heritage, J. 2002. “The limits of questioning: Negative interrogatives and hostile question content.” Journal of Pragmatics 34: 1427–1446. Hopper, P.J. 2000. “Grammatical constructions and their discourse origin: Prototype or family resemblance?” In Applied Cognitive Linguistics: Theory, Acquisition and Language Pedagogy, M. Pütz, S. Niemeier and R. Dirven (eds.), 109–129. Berlin: Mouton de Gryter. Kärkkäinen, E. 2003. Epistemic Stance in English Conversation: A Description of Its Interactional Functions, with a Focus on ‘I think’. Amsterdam: John Benjamins. Keevallik, L. 2003. From Interaction to Grammar. Estonian Finite Verb Forms in Conversation. Ph.D. dissertation. Department of Finno-Ugric Languages, Uppsala University. Keisanen, T. 2006. Patterns of Stance Taking: Negative Yes/No Interrogatives and Tag Questions in American English Conversation. Acta Universitatis Ouluensis B 71. Oulu, Finland: Oulu University Press. Kim, K.-h. 1995. “Wh-clefts and left-dislocation in English conversation: Cases of topicalization.” In Word Order in Discourse, P. Downing and M. Noonan (eds.), 247–296. Amsterdam: John Benjamins. Koshik, I. 2003. “Wh-questions used as challenges.” Discourse Studies 5 (1): 51–77. Koshik, I. 2005. Beyond rhetorical questions: Assertive questions in everyday interaction. Amsterdam: John Benjamins. Pomerantz, A. 1984a. “Giving a source or basis: The practice in conversation of telling ‘How I know’.” Journal of Pragmatics 8 (5, 6): 607–625.
280 Tiina Keisanen
Pomerantz, A. 1984b. “Agreeing and disagreeing with assessments: Some features of preferred/ dispreferred turn shapes.” In Structures of Social Action. Studies in Conversation Analysis, J.M. Atkinson and J. Heritage (eds.), 57–101. Cambridge: Cambridge University Press. Pomerantz, A. 1990. “Conversation analytic claims.” Communication Monographs 57 (3): 231– 235. Quirk, R., Greenbaum, S., Leech, G. and Svartvik, J. 1985. A Comprehensive Grammar of the English Language. London: Longman. Raymond, G. 2003. “Grammar and social organization: Yes/no interrogatives and the structure of responding.” American Sociological Review 68 (6): 939–967. Sacks, H. [1973]1987. “On the preferences for agreement and contiguity in sequences in conversation.” In Talk and Social Organization, G. Button and J.R.E. Lee (eds.), 54–69. Clevedon: Multilingual Matters. Sacks, H., Schegloff, E.A. and Jefferson, G. 1974. “A simplest systematics for the organization of turn-taking for conversation.” Language 50 (4): 696–735. Sadock, J.M. and Zwicky, A.M. 1985. “Speech act disctinctions in syntax.” In Language Typology and Syntactic Description. Volume 1. Clause structure, T. Shopen (ed.), 155–196. Cambridge: Cambridge University Press. Schegloff, E.A. 1968. “Sequencing in conversational openings.” American Anthropologist 70 (6): 1075–1095. Schegloff, E.A. 1988. “Goffman and the analysis of conversation.” In Erving Goffman. Exploring the interaction order, P. Drew and A. Wootton (eds.), 89–135. Cambridge: Polity Press. Schegloff, E.A. 1995. “Discourse as an interactional achievement III: The omnirelevance of action.” Research on Language and Social Interaction 28 (3): 185–211. Schegloff, E.A. [1995] 2007. Sequence Organization in Interaction: A primer in conversation analysis I. Cambridge: Cambridge University Press. Schegloff, E.A., Jefferson, G. and Sacks, H. 1977. “The preference for self-correction in the organization of repair in conversation.” Language 53 (2): 361–382. Schegloff, E.A. and Sacks, H. 1973. “Opening up closings.” Semiotica 7 (4): 289–327. Selting, M. 1994. “Emphatic speech style–with special focus on the prosodic signaling of heightened emotive involvement in conversation.” Journal of Pragmatics 22 (4): 375–408. Thompson, S.A. 1998. “A discourse explanation for the cross-linguistic differences in the grammar of interrogation and negation.” In Case, Typology, and Grammar, A. Siewierska and J.J. Song (eds.), 309–341. Amsterdam: John Benjamins.
Appendix: Transcription conventions (from Du Bois et al. 1993) Units Intonation unit Truncated intonation unit Truncated word
{carriage return} -%
Challenging the prior speaker 281
Transitional continuity Final Continuing Appeal
. , ?
Speakers Speech overlap (numbers inside brackets index overlaps [2two words2])
[ ]
Accent and lengthening Primary accent Secondary accent Unaccented Lengthening
^ ` =
Pause Long and medium Short (brief break in speech rhythm)
(N) ..
Vocal noises e.g., (TSK), (SNIFF), (YAWN), (DRINK) Glottal stop Exhalation Inhalation Laughter (one pulse) Laughter during speech (1–5 words) Laughter during speech (+6 words)
% (Hx) (H) @ @ (e.g. @two @words) @ (e.g. )
Quality Special voice quality Forte: loud Higher pitch level Parenthetical prosody
Transcriber’s perspective Uncertain hearing Researcher’s comment Indecipherable syllable
<X (( )) X
X>
Positioning and alignment as activities of stancetaking in news interviews Pentti Haddington
University of Oulu, Finland
1.
Introduction1
This paper investigates two interactional stancetaking activities called positioning and alignment in American and British news interviews. It represents a functional-interactional approach to studying stancetaking (Haddington 2004, 2005a), which combines Du Bois’ (this volume) theory of stance with conversation analytic methods of investigating the sequential and interactional aspects of talk-ininteraction and the grammar and grammatical practices for doing actions in interaction (Schegloff 1996). Stancetaking is here understood as an intersubjective activity in which interlocutors collaboratively construct and take stances about a stance object, by engaging with and building upon stances taken by their co-participants in the immediately prior talk (Du Bois this volume; Haddington 2005a). When interlocutors take stances, 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). I concentrate on describing the linguistic, sequential, and turn organizational features of positioning and alignment. 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. Consequently, it is also a forward-type intersubjective activity, because it establishes and creates an interactional context for the interviewee. A positioning question encodes or evokes a preferred stance, presumption, or a presupposition that puts the interviewee in a quandary relative to her institutional background, a publicly taken prior stance, audience expectations, and broader sociocultural beliefs and so on. Although the propositional context of the co-participants’ stance is not of interest per se, it is worth considering the contextual background and the indexical properties of language, and
284 Pentti Haddington
how they together with other practices invoke particular situated interpretations (Haddington 2004). The starting point for looking at positioning is the understanding 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).2 Clayman and Heritage (2002: 209) refer to the possibility that combinations of practices in questions may strongly favor particular answers over others, but do not discuss this possibility in their examples. Nevertheless, 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. This body of work has described how interviewers on the one hand design their questions as neutralistic (Clayman 1988; Clayman and Heritage 2002: 126–131; Heritage and Greatbatch 1991), and on the other hand, how they also adopt adversarial stances and exert pressure on their respondents by incorporating third-party statements, particular topical agendas, presuppositions, and accusations in their questions (Berg 2003, 2001; Clayman and Heritage 2002; Heritage 2003; Heritage and Roth 1995). However, many of these studies have only described the ways in which these practices and actions are accomplished within a TCU, and have thus given little attention to the design of entire questioning turns in news interviews (cf. question delivery structures, Heritage and Roth 1995) and the recurrent linguistic practices in them. In fact, Schegloff (1996) suggests that a linguistic vantage point can be used for looking at how speaker turns are composed of individual TCUs, how TCUs are sequentially positioned in their turns, how they are related to each other, and how all this affects the understanding of the turns (Schegloff 1996: 64). Similarly, Ford (2002, 2004) shows that it is fruitful to consider how recurrent action combinations and linguistic formats in turns are utilized for constructing co-participation in interaction. Schegloff ’s (1996) and Ford’s (2002, 2004) work provide a useful starting point for considering how the interlocutors’ stancetaking is organized across TCUs, transition relevance places (TRPs) and speaker turns.3 It also enables us to consider how the combinations of TCUs, and the actions within those TCUs, contribute to the interlocutors’ stancetaking. This approach, which goes beyond the examination of individual TCUs, is particularly relevant for the investigation of news-interview interaction, because talk in news interviews is largely composed of multi-unit turns (cf. Heritage and Roth 1995; ten Have 2001: 6). By recognizing the practices by which interviewers set up positions for their guests, it is also possible to understand the underlying motivations behind some instances of interviewee resistance. Although the examples of the interviewees’ answers in this paper could be seen to be evasive, it is also fruitful to consider the ways in which they engage and align with the question. The notion of alignment is
Stancetaking in news interviews 285
here understood somewhat differently from its use in CA. In CA, an aligning action is understood to be an appropriate or preferred next action, which fulfills the expectations raised by the previous action. For example, an aligning action to a request is an acceptance, and a disaligning action is a refusal (Heritage 1984: 269). The notion of alignment here is also not synonymous with agreement. Rather, it aims to explicate the range of possible types of convergent and divergent positions that interactants can take relative to each other, i.e., “how I put my stance in relation to your stance” (Du Bois, p.c.). Interlocutors use language for aligning with each other and therefore, alignment, as it is outlined by Du Bois (this volume) and adopted here, is very much a linguistic process in which interactants use morphosyntax, lexis, and prosody to construct their stances. One central element of linguistic alignment, as Du Bois (2001) further claims, is that speakers frequently use and recycle the linguistic elements that their coparticipants have used. By recycling linguistic forms, speakers bring two utterances into close relation to one another. When this happens, these forms engage. This paradigmatic connection between two utterances can generate new local meanings, even ones that at first sight seem disparate. Thus the connections between the “formal” structures of language can naturally affect the way in which individual stances are understood in the interactional context (Du Bois 2001). And as Du Bois (2001) shows, interlocutors use linguistic forms dialogically in order to negotiate their stances and create new meanings. It is noteworthy that by recycling elements, speakers do not necessarily aim to display agreement (Du Bois 2001). Rather, by recycling linguistic forms, interactants can frequently display subtle differences between their stances. In news interviews, the interviewee’s alignment activity can be seen as a backward-type intersubjective activity, and a contingent achievement by which she responds to the position in the question. Although conversation analysis has described such “evasive” interviewee conduct, interviewer responses to such evasive maneuvers (Clayman 2001: 238–298; Clayman and Heritage 2002), change of topical agendas, and question reformulations (Clayman 1993; Greatbatch 1986), it has largely looked at them separate from the question. In comparison, this paper focuses on some aspects of the interviewer’s turn that act as an impetus for an indirect answer, and some of the interviewees’ ways of responding to and answering difficult questions.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. As we will see, interviewees use various types of turn-design features, practices, and actions to respond to difficult questions and to intersubjectively engage with them. In sum, this paper supplements previous work by examining positioning and alignment from the following vantage points:
286 Pentti Haddington
1. It looks at how the combinations of practices, actions, action combinations, and particular turn-organizational features are used as resources and contribute to stancetaking in news interviews. 2. It describes a patterned and reusable linguistic practice that interviewees use to manage their stancetaking in news interviews. More specifically, interviewees frequently use a stance marker (with cognitive or communication verbs) and a recycled element of the interviewer’s question, to align with a proposition or presupposition incorporated in the question. Moreover, the interviewee’s turn in these answers is organized into a negative-positive structure, which helps the interviewee not only align with the question but also construct a stance of her own. 3. It shows how the co-participants intersubjectively engage in stancetaking. This paper supports prior findings (Du Bois this volume; Kärkkäinen 2003b) that by taking a stance an interlocutor (in this case the interviewee) displays an understanding of a prior stance (in the interviewer’s question). Thus, the findings in this paper support the claim that stancetaking is a contingent achievement occasioned by prior stances. Finally, 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, but rather that the interviewers come to the interview situation with certain predetermined questions and agendas in mind and the interviewees – although they naturally construct their stances vis-à-vis the interviewer and the question – design their answers based on their own objectives and attitudes. The database used in this paper was collected between October 1999 and March 2004. The corpus contains approximately 20 hours of news interviews from both the United States and Britain. The data fragments are transcribed in the Discourse Transcription (DT) style in which one line represents one intonation unit (Du Bois et al. 1993). The transcription conventions are given in the Appendix.
2.
Practices of positioning and alignment in news interviews
In the following sections, I discuss how interviewers set up difficult positions for the interviewees by asking hostile questions, evoking and encoding preferred stances, and by incorporating presuppositions in the questioning turns. I also identify some recurrent practices by which interviewees align with these positions.
Stancetaking in news interviews 287
2.1 Positioning by hostile questions and interviewees’ practices of aligning with them In example (1) below, which is from BBC2’s Newsnight, Jeremy Paxman interviews Alistair Darling, the Secretary of State for the Department for Transport. It presents an interesting example of how the interviewer positions the interviewee by designing a somewhat hostile question that the interviewee cannot answer directly. The question is about whether the British government should intervene in the tendency of falling flight prices.5 Nevertheless, the interviewee does not bluntly evade the question but manages to align and engage with it. (1) BBC2, Newsnight, Dec 17, 2003: Cheap air travel IR: Jeremy Paxman, IE: Alistair Darling (033 / 1 / 0:34) 1 →2 IR: Well earlier I spoke to the transport secretary Alistair Darling, 2 (TSK)(H) I kicked off by asking him, 3 →1 .. if he thought air travel (H) was too cheap. 4 ((CUT TO ALISTAIR DARLING)) 5 ...(1.4) 6 IE: I think that aviation, 7 ought to=, 8 meet the= `costs of the environmental damage it ^%causes. 9 Although .. (GLOTTAL) the ^difficulty there is that, 10 taxation of aviation fuel for example, 11 (H) is ^dealt with by `international ^treaty, 12 and we ^can’t do sen- something, 13 (H) %uh [u]nilaterally. 14 IR: [S-] -15 IE: (H) But in ^relation to=, 16 aviation costs in the last few years, 17 the reason it’s come ^down, 18 (H) is because, 19 the industry has ^dramatically cut its ^costs, 20 and passed that on to the passengers. 21 →3 IR: (0) Do you think it’s too cheap. 22 IE: .. (H)(TSK) No I do- I- -23 Now I would -24 →4 I wouldn’t argue that it’s too cheap, 25 (H) I ^do think though, 26 .. <MRC>it ought to meet the costs of the environmental 27 damage, 28 that it causes.
288 Pentti Haddington
The interviewer’s first question in lines 2–3 is a reported yes/no question. It is not possible to know whether the wording of the actual question was the same. If it was not, 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.6 Nevertheless, 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. The reported question is designed so that both of the projected alternative answers (an agreement or disagreement) are potentially harmful for the interviewee. As a transport secretary, the interviewee has to consider the interests of the airline companies, which of course makes giving a negative answer problematic. However, as an elected politician, answering yes would obviously be counter to the general public, which prefers cheap flight prices. Thus, in terms of positioning, 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 problematic situation. At first, 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. Clayman 2001; Clayman and Heritage 2002: 238–298). After this, the interviewee moves (lines 15–20) toward the original topical agenda in the question. However, he only reports some reasons why flight prices have come down and, in effect, does not answer the question. Then (in line 21) the interviewer orients to the interviewee’s dodging action and repeats the 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. Then in lines 22–23, the interviewee begins to align with the question. Judging from the truncations and the hesitation in the beginning of the interviewee’s answer, he at first has problems in formulating his response to the interviewer’s hostile repeated question. He does not finish the already started utterance No I do- I- -- that projects the utterance No I don’t think and goes on to modify the way he frames his stance. He does this by producing a self-repair and begins to produce an alternative modal auxiliary wouldn’t. Consequently (arrow 4), instead of recycling the verb think, the interviewee uses the communication verb argue (Biber et al. 2003: 316) in line 24. It could be that the change of the modal auxiliary is closely connected to the change of the verb as well. It is likely that the interviewee cut off a denial that was going to be produced as the stance marker I don’t think, but then changed it to a different stance marker I wouldn’t argue. By framing his stance in this way, the interviewee refrains from producing a direct denial
Stancetaking in news interviews 289
to the question (which I don’t think it’s too cheap would have done), and rather claims that he would not engage in an argument about the issue. In other words, 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. Therefore, 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; he must word his stance carefully in order to take into account the airline companies’ and the public’s possibly differing opinions. Stance markers with a first-person subject pronoun combined with a cognitive verb (e.g., think, know, guess, and accept) or a communication verb (e.g., argue, claim) are not only frequent in news interviews, but also play a significant role in interviewee alignment. As many studies on different languages have shown, the collocation between a first-person subject and a cognitive verb is strong in everyday talk, and these expressions tend to precede complement clauses (Kärkkäinen 2003a, this volume; Karlsson 2003; Rauniomaa this volume; Scheibman 2001). However, it would seem that speakers are not specifically referring to themselves when using these markers. Rather, these markers are used as epistemic/evidential/evaluative fragments in contexts where a stance is going to be produced by the speaker (Thompson 2002). This view is also supported by Simon-Vandenbergen (2000), who shows that in news interviews, the stance marker I don’t think is used for conveying and framing the speaker’s subsequent opinion, and by Schegloff (1996: 61–62) who points out that in everyday talk, I don’t know seems to project more talk to come. As the above example and the examples below show, these types of stance markers occur in a specific context in news interviews: they precede a complement clause that is composed of language recycled from the interviewer’s question turn. In the above example, the interviewee recycles the phrase it’s too cheap (arrow 4) from the interviewer’s turn. In Diagraph 1 below, 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.7 Diagraph 1 (from example 1): Cheap air travel 21 IR: you 22 IE: No I 23 Now I 24 I
{do} think it ’s too cheap do- Iwould -wouldn’t argue that it ’s too cheap
The above diagraph clearly demonstrates Greatbatch’s (1986) and Clayman’s (1993; 2001) findings that interviewees frequently recycle a unit from the question in
290 Pentti Haddington
order to shift the topical agenda or to evade the question altogether. Thus, this practice has a clear function. 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; Heritage and Roth 1995). However, an alternative reading that considers the impact of the question on the answer is perhaps more accurate in example (1). In other words, 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. Therefore, it is the way in which the stance is formulated in the question that affects the design of the answer. The engagement and alignment 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). This pattern is frequently used by interviewees for aligning with a difficult position posed by the interviewer’s question. By using this pattern, the interviewee in the above example locates the trouble source, 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. And even though the interviewee’s turn, in example (1) above, is constructed so that it displays a more explicit engagement with the question than his previous answer (in lines 6–20), he basically gives the same answer as he gave before the interviewer’s repeated question. In other words, by carefully designing the answer so that it engages with the difficult question, the interviewee rather than just evading the question, displays careful alignment with the question and, in fact, answers it.
2.2 Positioning by incorporating preferred stances in questions and practices of aligning with them In the following, I show how the interviewer sets up a position for the interviewee by designing the turn so that it evokes a preferred stance. The notion of preference is one of the central analytic concepts in CA (cf. Hutchby and Wooffitt 1998; Sacks [1973] 1987; Sacks and Schegloff 1979). It refers to the idea that in interaction particular actions invite (i.e., prefer) particular types of responses (Hutchby and Wooffitt 1998: 44). Preference does not refer to the co-participants’ internal or subjective preferences, but rather to the structural features of how a turn or an action is produced in its sequential context. The clearest evidence for preference organization can be seen in adjacency pairs. For example, an invitation prefers an acceptance; assessments prefer agreeing second assessments, and so on. Second pair parts in adjacency pairs that are structured as preferred tend to be unmarked, whereas dispreferred turns often incorporate different types of dispreference mark-
Stancetaking in news interviews 291
ers (cf. Hutchby and Wooffitt 1998: 44; Sacks [1973] 1987). Preference is also often signaled with particular use of syntax. 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 question of whether a stance incorporated in the first turn can evoke a preferred next stance has not yet received much attention. Since stancetaking is not an action but a larger activity (Haddington 2004) the whole idea of preference as it relates to stancetaking requires reconsideration. In other words, a stance preferring a particular subsequent stance is a different issue from one action preferring a certain type of next action. Sacks ([1973] 1987: 58) rightly claimed that features of preference organization can be examined without knowledge of the participants, the topic, and where a sequence takes place. However, preference as it relates to stancetaking benefits from considering at least some of the extra-textual and participant-specific aspects in the analyzed segment, as is shown below. In the following example, the interviewer asks a question that is framed as a third-party stance. 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. This example comes from CNN’s Larry King Live. The third-party stance in the question preface is based upon recent polls on whether the United States should attack bin Laden in Afghanistan. After this, the interviewer, Larry King, formulates a yes/no question to the interviewee Brian Jenkins. (2) CNN, Larry King Live, Sep 12, 2001: Bomb him out IR: Larry King, IE: Brian Jenkins (001 / 5 / 3:49) 1 IR: 2 3 →1 4 →1 5 6 IE: 7 8 →2 9 10 11 12 13 14 15 16
...(0.6)(TSK) ^Brian, could you=, ... well I guess the `public, would look at this `simply, Could you ^bomb him out. .. (TSK)(H)U=h, I d- -I don’t know that you could ^bo=mb him out, I mean, I I ^think that, uh, .. the ^magnitude of this operation, is going to `call for, a ^qualitatively different `response, than we have seen in the ^past `two=, two acts of ^terrorism.
292 Pentti Haddington
The TCU, which the interviewer starts in line 2, projects a yes/no question. However, 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. It has been shown that interviewers use third-party statements in order to raise controversial topics, but since these stances are attributed to third parties, the interviewers still maintain a neutralistic stance toward the topical agenda and the guest (Clayman 1988; Heritage 2003; Heritage and Greatbatch 1991).8 In the above example, the third-party stance is placed sequentially before the question (at arrow 1), and therefore, presents the question not so much as the interviewer’s question, but as a stance that the majority of the general public share. The third-party is mentioned in the complement clause in lines 3–4. The complement clause is preceded by the epistemic stance marker I guess in line 3. Kärkkäinen (this volume) claims that I guess often frames actions in first position (e.g., questions) that invite second-pair parts, 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. Indeed, not only does the marker here occur in first position (a questioning turn), 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, of which according to the polls at the time 94 percent favored a military response. The sudden transformation of the question’s design and the invoked third-party stance change the way in which the question can be understood, and thus have important interactional consequences for stancetaking. In other words, in addition to the fact that the interviewee has to answer the question, he now also has to orient to the third-party stance given in the question preface. This is because although the question TCU Could you ^bomb him out. in line 5 is a yes/no question and makes relevant either an affirmative or a negative answer, in this case, a subsequent negative answer would mean that the interviewee disagrees with a large portion of the public and the TV-viewers, a position one does not seek willingly. In other words, not only does the question favor an agreeing answer (Sacks [1973] 1987), but the question TCU together with this particular third-party stance (i.e., the combination of the two actions) also strengthens the preference conditions for an answer that agrees with the question. 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 interviewee. The evidence for this can be found in the beginning of the interviewee’s turn. He goes through some trouble in formulating an answer to the question, which is indicated by the pause before the interviewee’s answer, the hesitation marker U=h (line 6), and the self-repair (line 7). These dispreference markers and faltering talk are indicative of a stance that is divergent from the public stance
Stancetaking in news interviews 293
reported in the question. After this, the interviewee answers the question by uttering I don’t know that you could ^bomb him out (arrow 2). As Diagraph 2 below shows, after the stance marker, the interviewee again recycles an element from the question, which establishes a connection between the question and the answer. Diagraph 2 (from example 2): Bomb him out 5 IR: {you} could bomb him out 7 IE: I do- -8 I do n’t know that you could bomb him out
The interviewees’ formulation of the TCU in line 8 is interesting, because the recycled part of the TCU is a finite dependent that-clause in which the verb is in the subjunctive mood.9 The use of subjunctives is very rare in spoken English, but they are sometimes used to express unreal or hypothetical meanings (Biber et al. 2003: 261). In comparison to a conditional clause (I don’t know if you could bomb him out), which in this context would display epistemic uncertainty or insufficient knowledge about the issue (Haddington 2005b), the interviewee now orients to the preferred stance by a subtle expression of doubt about the public’s view. In light of this, the use of the that-clause and the subjunctive contributes to the aligning activity in an interesting way. It treats the third-party stance as doubtful or even hypothetical, even though the reported third-party stance was actually based on topical polls. After this, the interviewee produces an explanation by introducing alternative actions to “bombing.” In sum, the interviewee orients to the preferred stance and shapes the answer so that it disagrees with it as little as possible. In front of the audience (the public referred to in the question), this is much more convenient for the interviewee than an explicit disagreement or an evasive answer would be. Example (3) below comes from CNN’s Crossfire.10 It was recorded in December 2002, a couple of months after the terrorist attacks on New York and Washington. The interviewer is Paul Begala, who served in the Clinton administration, and the interviewee is Frank Gaffney, who is the former Assistant Secretary of Defense and an expert in foreign and defense policy. The question concerns the so-called “bin Laden tapes” and whether these tapes include covert messages for inactive terrorists to take action and, therefore, should not be shown on television. In this example, the interviewer positions the interviewee by invoking a third-party stance, subtly incorporating his own stance, and designing the question to contain a preferred answer. The question also shows how the interviewee carefully aligns with this rather complex positioning activity.
294 Pentti Haddington
(3) CNN, Crossfire, Dec 27, 2001: The new bin Laden tape IR: Paul Begala, IE: Frank Gaffney (003 / 1 / 1:13) 1 IR: 2 3 4 5 6 →1 7 →1 8 →1 9 →1 10 →1 11 →1 12 →1 13 →1 14 →1 15 →1 16 →1 17 →1 18 →2 19 →2 20 →3 21 →3 22 →4 IE: 23 →4 24 →4 25 →4 26 →4 27 28 →5 29 30 31 32 33
.. Uh the ^new bin Laden tape. (H) ... that’s, aired by Al-Jazeera ^today? ...(H)(TSK) When, .. he ^began sending these tapes out, .. the President’s national security adviser, ^told the `networks. ...(0.7) They ^shouldn’t run these, because `she ^feared, ... ^first that he would whip up -.. uhv, uhm, .. anti-American `views, but then ^second, and probably more .. `ominously, (H) that there were <MRC>secret coded messages, ^potentially in these tapes. ... Now they’ve had some of these tapes for eight weeks. We have the best cryptographers in the world. (0) Is there <MRC>slightest shred of evidence, that she was right? ...(1.0) (H) (TSK) Well ^nothing, .. has been blown up, ... so 'fa=r. .. So, ... you might deduce from that, `no. ... I don’t think this is a matter of ^cryptography, I think this is a ^question of whether, <MRC>people have gotten instructions, ...(0.7) (H) who are ^here in this country, or perhaps ^elsewhere in the world, ... (H) and will be prepared to ^operate on them.
The above interviewer’s turn is again composed of two main parts: the question preface (in lines 1–19), which sets up the topical agenda and gives background information to the question, and the actual question (lines 20–21). At the beginning of his turn, the interviewer introduces the stance object “the new bin Laden tape” in lines 1–3, after which, in lines 6–17, he goes on to report a third-party stance,
Stancetaking in news interviews 295
i.e., what Condoleezza Rice, the President’s National Security Adviser, has said about the stance object (that the networks should not run these tapes, and that there are potentially secret coded messages in them). After this, the interviewer concludes the turn by formulating an interrogative in lines 20–21. Even though the question is not hostile or adversarial, the combination of the several practices in the question set up a position for the interviewee. First, similarly to example (2), in example (3), 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. In December 2001, when this interview was recorded, politicians across the political field (as well as the majority of the American public and the TV networks) were almost unanimous in their support of the President and his administration. 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). Second, the interviewer makes a footing shift (Goffman 1981) and produces an assessment, which presupposes that the issue is about cryptography (We have the best cryptographers in the world.) (arrows 2). These two TCUs are produced in distinctly rapid speech, which suggests that he slips them in between the question preface and the question proper. Nevertheless, the fact that the footing shift is sequentially positioned before the actual question frames the question in a particular light. Even though the interviewer in these TCUs does not (and cannot) express explicit disagreement with the third party, he frames the question by providing additional background information that undermines the reliability and the correctness of the third-party stance. Furthermore, as we will see later, when the answer is considered, these two TCUs indeed are contextually relevant and affect the way in which the following yes/no interrogative is understood. Finally, the question that ends the interviewer’s turn in lines 20–21 contains a preference. It is designed as a yes/no-type interrogative which narrows down the possible relevant answers to an affirmative yes or a negative no. The question also contains the adjective slight in the superlative form and the noun shred, which together invoke the idea of smallness and insignificance. These words could be perceived as negative polarity items (Horn 1989), which usually embody a preference for a no answer (Clayman and Heritage 2002: 211–212). The preference for the negative answer is also strengthened with the superlative construction. The negative polarity items and the superlative construction together suggest that in spite of all the intelligence resources, Condoleezza Rice does not have any evidence to support her argument that the interviewer had reported in lines 6–17. 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
296 Pentti Haddington
is considerably higher than in the other parts of his turn. In sum, the third-party statement, the sequential proximity of the footing shift and question TCU, and the linguistic design of the question together contribute to the positioning activity that the turn is doing. The interviewee is basically asked to display his disagreement with the President’s National Security Adviser, which at that historical moment could have displayed the interviewee in an unfavorable light. In spite of the difficult position, the interviewee does not evade the question, but rather addresses and responds to both parts of it: on the one hand, the issue regarding Condoleezza Rice (lines 22–27) and, on the other hand, the cryptography/secret coded messages-issue (lines 28–33). In addition to this, the interviewee orients to the question’s problematic preference structure and tacks through it by designing the answering turn in a very careful manner. As we saw above, the question prefers a negative answer. The interviewee complies with this by eventually answering no in line 27. However, the interviewee resorts 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. In addition to this, before producing the no-answer, the interviewee carefully designs the first two TCUs (at arrows 4) so that he avoids encoding them as personal stances and rather relies on evidential information (nothing, .. has been blown up, ... so ‘fa=r.) and induction (you might deduce from that,) (cf. Chafe 1986). He further distances himself from the stance he takes by using the generic pronoun you and the modal verb might (cf. can). Thus, even though the interviewee provides the preferred answer no, the two preceding TCUs frame the answer as a depersonalized stance. Only after this does the interviewee continue to respond to other parts of the question (cf. fn 4 in Clayman and Heritage 2002: 106–107). He does this by denying the presupposition about the relevance of cryptography (arrow 5), and then providing his view of what the issue is really about. He does this again by first producing the stance marker I don’t think, which is followed by a recycled lexical item in a slightly modified form (cryptography) from the question in line 28. 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 question preface and by repeating it, the noun cryptography, in his turn. By doing so, he denies the presupposition in the question and its relevance in relation to the issue at hand. This precedes and strongly projects a resolution or an account to the
Stancetaking in news interviews 297
denied issue (Ford 2002; Haddington 2005b), which is then provided immediately after the denial. In sum, the interviewee not only orients to the preferred stance evoked in the question, but also uses the interactional space he has been allocated to design and present the stance in his own terms. In other words, the above practices, which the interviewee relies on, imply that he orients to the preferred stance in the question and agrees with it (although minimally). However, at the same time, these practices turn out to be handy for expressing an alternative stance. In sum, the two examples in this section have shown different ways in which interviewers incorporate preferred stances in the questions, and thereby put their guests in difficult positions to answer them. In spite of this, the interviewees do not just avoid answering the questions, but in fact intersubjectively align with the position set up in the question. In the following section, I concentrate in greater detail on how interviewers use presuppositions for setting up positions for the interviewees and how interviewees deny them.
2.3 Using presuppositions for positioning and linguistic practices for aligning with them As Clayman and Heritage (2002: 127, 203–208) note, and as we saw in example (3), interviewers can and often do embody presuppositions in the questions. According to Levinson (1983: 167), presuppositions are pragmatic inferences that are based on linguistic structure and are sensitive to contextual factors. Here a presupposition is understood to be a background assumption, which, on the one hand, is used in the interviewer’s question for rendering the question sensible and understandable for the interviewee and the overhearing audience, and on the other hand, for setting up a position for the interviewee to take a stance. Clayman and Heritage (2002) claim that in news interviews presuppositions differ in terms of their “embeddedness” in the question. Some presuppositions are closer to the surface than others, and some are easier to deny than others. Although presuppositions are not “explicit” stances, but rather compositions of “given” information, “built-in” elements, and “self-evident” assumptions on which speaker stances are built, the following examples show that the subsequent speaker can orient to a presupposition. In other words, a presupposition, which is incorporated in an utterance, can affect the next utterance and the stance in that utterance. In the following, I briefly go through examples in which presuppositions are used for setting up a position for the interviewee. It is worth noting that in the following questions, presuppositions are not the only practice used for setting up a position for the interviewee, rather the interviewers actually use a combination of various practices to construct a problematic position for the guest. After that,
298 Pentti Haddington
I consider the resources with which the interviewees deal with the presuppositions. As in the examples above, the interviewees use the same linguistic practice (stance marker + linguistic recycling) for aligning with the question. All in all, my aim is to show that even though the answers can be claimed to be somewhat evasive, the answers intersubjectively orient to the position set up in the question both linguistically and in terms of turn design. Example (4) below comes from Larry King Live. Here, the interviewer asks a question from Christopher Whitcomb, a special agent for the FBI. The question concerns Osama bin Laden, the alleged leader of the Al-Qaida terrorist organization, and whether bin Laden should be killed or taken captive and then brought to justice in the United States. (4) CNN, Larry King Live, Sep 12, 2001: Rid them from this planet IR: Larry King, IE: Christopher Whitcomb (001 / 5 / 1:24) 1 IR: 2 3 →1 4 5 →1 6 →1 7 8 9 10 →1 11 →2 12 →2 13 →2 14 15 16 17 →3 IE: 18 19 20 21 22 23 24 25 26
Christopher, if it ^gets to the `point, as we have ^promised that, wl- -`we will ^take them out, we will=, u=h, <MRC>rid them from this planet, in a sense, We are at ^war with them, ^How `do you `do that Christopher, ^How do you ... ^take out a, a ^bin Laden or `someone like a bin Laden, who’s hidden, who `moves arou=nd, and who ^may be under ^cover of the `country he lives in. ...(0.8) (TSK)(H) Well `I don’t know about ^taking him out Larry, But I think we have a ^lot of `options, And one of them `obviously is ^military. (H) The ^other=, because this is a ^law enforcement `operation, and because `we ^have to=, (H) uh `go with the ^justice system, and we want to bring `ultimately these people to ^justice, (H) Uh we’ve ^done, in the ^pa=st,
27 28 29 30
Stancetaking in news interviews 299
(H) `operations which are referred to as ^renditions, where we go into `other ^countries, (H) with the ^cooperation of those countries of course, and bring people ^back to the United States.
The question contains two distinct units: a question preface in lines 2–10, which provides a contextual background for the question, and the actual question in lines 11–16. The question contains two presuppositions, which are not hostile (Clayman and Heritage 2002; Heritage 2003, 2002) but nevertheless set up a position for the interviewee. First of all, the question preface contains stances that are attributed to we (arrows 1). This pronoun, which can also be called a shifter (cf. Jakobson et al. 1995; Sacks 1992; Silverstein 1995), indexes a category that can loosely be equated with “Americans.” Therefore, it implies that the interviewee belongs to a group that is assumed to share the stance, which is made relevant through the action of promising, reported in the question preface, namely as we have ^promised that, in line 3, we will=, u=h¸<MRC>rid them from this planet, in lines 6–8.11 Moreover, the third-party stance, which by implication includes the interviewee, is also used as a resource for constructing the actual question (^How do you ... ^take out a, a ^bin Laden or ‘someone like a bin Laden, arrows 2). The fact that this stance is here reproduced as an interrogative renders it presupposed information, i.e., 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. Levinson 1983: 184). 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, which indeed marks the question to be inquiring about ways in which bin Laden can be taken out, and not whether it is reasonable or possible to take him out. By building on the assumption that the interviewee shares both the stances in the question preface and the presupposition in the question, the interviewer invites the interviewee to think about measures for ‘taking out bin Laden,’ and thus raises potential problems for the interviewee. As we can see in the example, the interviewee does not share the question’s stances, nor its presuppositions, but designs his answer so that it contains a divergent stance. First, at the very beginning of his answer (arrow 3) the interviewee displays his divergence from the presupposition by denying it (cf. Haddington 2005b). The divergence is also signaled by the long pause and the dispreference marker well (cf. Sacks [1973] 1987). After this, the interviewee provides an account (in line 18) by proposing other alternatives for catching and dealing with Osama bin Laden. The recycled elements can be seen in the following diagraph.
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, the recycled forms are again preceded by a stance marker that contains a first-person subject (I), an auxiliary verb with a negative marker (don’t), and a (cognitive) verb (know). Moreover, since the index of the first-person subject I (in line 17) is by implication included in the firstperson-plural we (lines 3, 6, and 10) used earlier, it can be claimed that these two pronouns resonate with each other, and thus indicate the stance differential between the interviewer’s and the interviewee’s turns, and the problem that the interviewee has with the indexical we. As in example (3), 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. And in line 18, 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, but by new terms the interviewee himself chooses. In example (5) below, the interviewee attempts to align with the interviewer’s positioning question by producing a denial. However, in contrast to example (4), the interviewer holds the interviewee accountable for not answering the question by asking the same question again. This example comes from BBC2’s Newsnight. Here, Jeremy Paxman is interviewing the conservative MP, Tim Yeo, about the introduction of higher tuition fees in British universities. This issue caused a lot of dispute and controversy in Britain in spring 2004. The focus is on the question in lines 24–25, which is a repeated question occasioned by the interviewee’s evasive answer (data not shown). In other words, the interviewer holds the interviewee accountable for not answering the question by asking the same question again. (5) BBC2, Newsnight, Jan 21, 2004: speaking for the universities IR: Jeremy Paxman, IE: Tim Yeo (037 / 1 / 7:18) 1 IR: 2 3 4 5
Can I ask you just one small factual point, of the eighty-nine Vice Cha=ncellors, (H)(GLOTTAL) Uh, how many of them, .. actually support your position?
((18 LINES OMITTED)) 24 →1 IR: You keep on claiming to speak for the universities, 25 →1 How many of the Vice Chancellor [s s]upport [2you2].
Stancetaking in news interviews 301
26 IE: [(GLOTTAL)] 27 →2 [2I2] don’t claim to 28 →2 speak ^for the universities, 29 but [I do claim]-30 IR: [You’ve been doing no]thing [2^but2] the entire [3ev3]en 31 [4ing4]. 32 IE: [2Uh2], 33 [3%3] 34 [4I4] do claim to be concerned about their future, 35 (H) An[d],
The sequential progression of the questioning, which amounts to coercing the interviewee to provide a relevant answer to the question, in itself puts the interviewee in a difficult position (cf. example 1, above). However, there are also other aspects that set up a position for the interviewee. First of all, the interviewer’s question, both the original one in lines 1–5 and the repeated one (arrows 1), presupposes that at least some Vice Chancellors support the interviewee’s views, which then projects an answer that should give their exact number. This question is in itself probably difficult to answer, because not only are the Vice Chancellors numerous, but it also seems improbable that the interviewee knows exactly which support him and which do not, if any indeed do. Moreover, the interviewer’s questions are hostile, 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. Heritage 2003: 81). This is fortified by the design of the question, 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. All the linguistic and interactional elements in the question intersubjectively set up a position for the interviewee that he needs to take into account in his response. 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. 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. Again, 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, and at the same time to undermine the claim in the interviewer’s question preface. Note also that the primary stress on ^for in line 28 further emphasizes the interviewee’s divergence from the interviewer’s claim.
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, this time (contrary to example 4, above) the interviewer responds with a forceful counter-argument. Despite the interviewer’s intervention, the interviewee continues to produce an account of what his actions really are about (line 29). In this case, the interviewee uses the denial + account action combination as a strategy for evading the hostile question. About four seconds later, the interviewer explicitly voices the fact that interviewee is not answering the question: (6) BBC2, Newsnight, Jan 21, 2004: speaking for the universities IR: Jeremy Paxman, IE: Tim Yeo (037 / 1 / 7:51) 46 IR: You’re not going to answer that question [n, 47 IE: [I am going to] answer it, 48 IR: are you].
As we have seen in this and the previous section, there is a strong and recurrent interactional-linguistic pattern by which interviewees respond to difficult questions. By using a stance marker combined with a recycled linguistic unit from the interviewer’s question turn, the interviewee locates a trouble source in the question (an assertion or a presupposition) and denies it. The trouble source is located and foregrounded by recycling it in the denial: in examples (1) and (2), a proposition in the question, in (3), (4), and (5), a presupposition in the question. The denial, which is composed of the stance marker in association with a recycled unit from the question, not only sequentially precedes, but also strongly projects a next action (more evidence for this is given in Section 5). The recycled element can engage with any element in the question, either in the question preface (Diagraphs 1 and 2), the question TCU (Diagraphs 4 and 5) or both (Diagraph 3). These phenomena are closely related to the interlocutors’ stancetaking and especially to interviewee alignment with the questions. Indeed, the interviewees do not just evade the question for the sake of it. Rather, their evasive answers are contingent products and occasioned by the asserted or presupposed stances in the question. This practice, therefore, stands as firm evidence of the backwardtype intersubjectivity in news interviews (Haddington 2004). Interviewees recycle language from the question in order to orient to and undermine a stance encoded or evoked 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, an assertion, a presupposition or another problematic position in the question. It is
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 answer. In the following section, I turn to a more detailed linguistic analysis of what happens in the second parts of the action combination. I show that interviewees use a particular turn-constructional format for aligning with preferred, evoked, and presuppositional stances.
2.4 The neg + pos pattern as a resource for alignment Interviewees frequently use the stance marker + recycled language, discussed in the previous section, as a resource for aligning with the interviewer’s difficult question. This structure recurs in turn-initial position (Haddington 2005b). This is not an accident, because in that sequential position it has a clear function in the organization of the answering turn and thereby also in the organization of the interviewee’s stance. As was noted, this pattern is used for displaying engagement with the question, but also for denying a presupposition or a position set up in the question. However, denials do not occur alone. Ford (2002) claims that in everyday talk a denial is actually only one part of a combination of two actions. The second action, which follows the denial, is a correction or an account by the same speaker that gives an alternative interpretation to what has been denied (Ford 2002: 62). According to Ford (2002), this action combination is a robust and coherent discourse structure in everyday talk. This happens in news interviews as well. In my data, however, this action combination is used exclusively by the interviewees and, contrary to everyday talk, is frequently produced with the help of a linguistic pattern called the neg + pos pattern. A search of this pattern in the Santa Barbara Corpus of Spoken American English (Du Bois et al. 2000; Du Bois et al. 2003; Du Bois and Englebretson 2004) yielded only one example, which suggests that it is rare in everyday talk and indicates that it has a special function in news interviews. By using it, 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, they actually align with them. 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. The bolded parts picture the linguistic pattern.12 Diagraph 6 (from example 1): Bomb him out 7 IE: I d8 I do n’t know that you could bomb him out
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- I23 Now I would -24 I wouldn’t argue that it’s too cheap, 25 I do think though, 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, but also organizes the interviewee’s intersubjective stancetaking. It is composed of two parts, which always occur in different intonation units. Quite often, but not always, these two parts also follow each other in consecutive intonation units.13 The first part doing the denial contains a first-person-singular pronoun (or sometimes some other pronoun) and a negated cognitive or communication verb, which is followed by a recycled linguistic element that identifies the issue that is being denied. The second part doing the account, then contains a first-person-singular pronoun with a cognitive or communication verb, which is followed by the account. The pattern can be schematized in diagraph form as: “I” + negative particle + cognitive / communication verb + recycled language “I” + verb + predication
Stancetaking in news interviews 305
It is worth noting that the original linguistic unit that the interviewee recycles is rarely negatively formulated. This suggests that this practice in this interactional context can only be used if the unit that is recycled does not contain negative markers. One reason for this might be that the denial that projects the account always contains a negative marker and that a positively formulated utterance in this context is easier to recycle. However, in cases in which the recycled linguistic unit comes from a negatively formulated TCU, the recycled element can be slightly modified. Consider the following deviant case: (7) BBC2, Newsnight, Jan 19, 2004: Keep a promise IR: Jeremy Paxman, IE: Tony Blair (036 / 1 / 3:34) 1 IR: 2 3 4 5 →1 6 IE: 7 →2 8 9 10
At `what ^point, after ^writing that `manifesto, in which <MRC>you requested our votes, did you `realise, (H) that you couldn’t keep the promise? (TSK)(H) ... Well I, I ^don’t `accept that we have broken the promise, (0) as I said, because the `new system doesn’t come into ^effect, .. until ^after the next general `%election.
In the above example, 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), which, of course, communicates an opposite meaning compared to the original phrase. 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, and that the stance marker in the denial is a “compulsory” part of the combination and can, in fact, influence the design of the recycled linguistic elements in the complement.14 The neg + pos pattern does not, of course, only occur in this interactional context, but it is a recurrent and routinized pattern in political news interviews and reproducible in this interactional context. By using this particular pattern, the interviewees deny an assertion or a presupposition in the interviewer’s question and then provide the next relevant action, an account for the denial. It is
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, the interviewees do not just bluntly avoid answering the question, but in fact engage with it. Consequently, this pattern shows that the interviewee orients to the position (or some part of it) set up in the question, undermines it and goes on to propose another way of looking at the issue, i.e., she aligns with the question. Following Ford (2002), this action combination and the linguistic pattern in it could be perceived as a turn-constructional format, a pattern in which two actions are closely connected. As Ford (2002) claims, a denial alone strongly projects a resolution or an explanation component. This is particularly true in news interviews in which denial alone would be perceived as inadequate, because interviewees are expected to support and give grounds to their statements. Therefore, the denial projects a move toward the second part, a resolution or a counterstance. Consequently, these two actions are not just individual actions, but they organize a larger intersubjective stancetaking activity in which the interviewee responds to an evoked, asserted, or presupposed stance in the question. Further evidence for the connectedness of the two actions can be heard in how this pattern is designed prosodically. In all examples except (1) (see Figures 2–5), the second part of the action combination is latched onto the first part. In other words, in order to keep the turn, the interviewee produces a rush-through (Local 1992; Schegloff 1982), which helps him get to the next TCU. In example (1), the interviewee produces a clearly audible inbreath, which also projects more talk to come (see Figure 1). In addition to this, as the acoustic measurements suggest, in some cases (see the pitch curve in Figures 2 and 5), the end of the first part is produced with distinctly rising-continuing intonation, which also projects further talk from the speaker. These phonetic features further suggest that the first part projects a move toward the second part. The interviewer can use these features as a resource for understanding what the interviewee is doing in terms of turn projection, i.e., that the interviewee is going to produce more talk beyond the next TRP. Although these features display the close relationship between the two TCUs within the interviewee’s turn, they also serve an interactional function. As was noted above, interviewers have the right to interrupt or challenge the interviewee if they feel that she is not answering the question. Thus, it is possible that when an interviewee produces a denial, the interviewer can intervene after the denial and challenge or disagree with the interviewee’s denial. This actually happens in example (5) in line 30, when the interviewer says [You’ve been doing no]thing [2^but2] the entire [3ev3]en[4ing4]. By producing the action combination prosodically as they do (no pause, rising intonation, and latching onto previous unit), the interviewees minimize the interactional space at the TRP in which the inter-
Stancetaking in news interviews 307
viewer has an opportunity to intervene. Moreover, as we can see in example (5), the interviewee continues and finishes the second part of the action combination, which shows that despite the interviewer’s intervention, this turn-internal organization of the two actions is very strong. Another interesting point that arises in these examples is that there is a tendency in how this pattern is produced prosodically. In all of the examples above, the maximum pitch in the first part is produced on the negative modal or auxiliary verb (either don’t or wouldn’t). In addition to this, the unit that is recycled from the interviewer’s question also receives strong emphasis. Furthermore, in the second part, the primary emphasis tends to be on the first (cognitive or communication) verb or an accompanying modal. These similarities in the prosodic design of this action combination act as evidence of intersubjective stancetaking, i.e., the interviewee is taking a stance based on an assertion or presupposition in the question turn, and thereby aligning with it. In other words, the practices in the action combination and the action combination itself are contingent achievements made relevant because of the position set up in the question. As was claimed above, this linguistic pattern almost never occurs in everyday talk. Therefore, an interesting question arises: What motivates the use of this pattern in news interviews? Sacks ([1973] 1987) noted that questions tend to prefer agreement. He also noted that if answerers treat questions as problematic and show this by using various dispreference markers, questioners tend to modify the question so that it contains the opposite preference. In other words, the questioners end up using a form that answerers can more readily agree with. However, in news interviews the interviewers rarely modify their questions for the interviewees. Rather, the whole issue of preference is turned upside-down. As we have seen, the
Figure 1. Prosody in example (1): Cheap air travel
308 Pentti Haddington
Figure 2. Prosody in example (2): Bomb him out
Figure 3. Prosody in example (3): The new bin Laden tape
Figure 4. Prosody in example (4): Rid them from this planet
Stancetaking in news interviews 309
Figure 5. 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. Since the interviewees are forced to answer the questions, they have to align with the problematic stance in that question immediately. Because of the pressure, the above linguistic pattern is a useful resource for the interviewees to align with the question. In other words, they can produce a stance that diverges from but still does not totally disagree with an assertion or a presupposition in the question. By designing the answer in this way, they can reduce the severe consequences that a strict disagreement might bring with it.
3.
Discussion and conclusions
News interviews are publicly broadcast interactions in which the journalist’s primary task is to ask questions of a public figure. The public figure is then expected to provide appropriate answers to the questions (e.g., Scannell 1991: 4). One purpose of this question-answer activity is to give information, express opinions, and discuss and debate topical and often controversial issues. Against this background, news interviews are the venues for politicians and other experts to publicly convey, formulate, defend, and negotiate their stances. This paper makes the following claims about stancetaking in news interviews. First of all, it shows that the methodological approach to study stancetaking with data from real interactional situations is able to provide new and interesting findings about the ways in which speakers take stances. It also shows that speakers do not (just) express their subjective stances, but indeed orient to and engage with each other’s stances, and thus engage in intersubjective stancetaking. It has also shown that a linguistic analysis combined with the interactional analysis of
310 Pentti Haddington
the data can provide new findings about stancetaking. The linguistic analysis especially shows that participants in news interviews rely on recurrent linguistic formats and patterns in specific interactional situations, for example in producing the denial + account action combination. Second, this paper has described some practices by which interviewers set up positions for the interviewees and how the interviewees align with these positions. The positioning activity is essentially a one-directional and forward-type intersubjective activity. By setting up a position the interviewers attempt to constrain and delimit the possibilities for the interviewee to construct a responsive stance. It is worth noting, however, that not all questions position the interviewees. Often (but not always) simple information-seeking wh-questions position the interviewees only minimally, if at all, because they establish a fairly loose agenda and project a broader answer than, for example, negative interrogatives or tag questions, which have a fairly clear preference structure (cf. Heritage 2002, 2003: 69). It is also important to notice that positioning becomes clearly visible only if the interviewee treats the questions as doing positioning. In some cases, the interviewer may be seen to be positioning the interviewee, but the interviewee does not treat the question as problematic. The interviewees indeed can decode preferred stances and presuppositions in the questions and orient to and deny them. One practice for denying an element in the question is the above-mentioned denial + account action combination in which interviewees use the stance marker (usually personal pronoun + negative particle + cognitive/communication verb) together with recycled language. As was shown, this linguistic pattern is a routinized and reusable linguistic pattern in this interactional context. Third, these activities are made possible by three special features of news interview interaction. First, the special (institutional) turn-taking organization in which sequences of questions and answers follow each other. Second, the turntype pre-allocation, i.e., the fact that interviewers produce questions and interviewees produce answers means that only interviewers do the positioning and the interviewees the aligning. Third, 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. There are also some general observations that can be drawn from the analysis above. First of all, recent research on stancetaking, attitudes and evaluation suggests that participants in everyday conversation do not converse with particular pre-determined stances in their heads, but rather work together in interaction in order to reach some sort of a joint and negotiated stance (Du Bois this volume; Kärkkäinen 2003b; Potter 1998). However, the findings in this paper suggest that stancetaking is different in news interviews. As was shown above, interviewers
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. The way in which these practices, actions, and the combinations of them function in the question turns suggests that interviewers come to the interview situation with particular questions and agendas in mind. They use their knowledge of who the interviewees are, what their background is, who or what they represent, and what kind of answers can be expected from them, 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 reconsider the idea of interviewee evasion. Not only are evasive or agenda-shifting actions often occasioned by the interviewers’ questions and therefore intersubjective, but they are also contingent achievements occasioned by the evoked and preferred stances and presuppositions in the questions. Consequently, it is better not just to assume that interviewees evade questions, 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 stances, but also reflect the speaker’s own identity, background, aims, and previously stated stances. The linguistic pattern discussed above stands as good evidence of this. The interviewees use this pattern for vitiating preferred stances and presuppositions, and for expressing the stances that better represent their views of the topical matter at hand, i.e., views that they have very likely had before they came to the program. We could therefore call into question whether the stances in news interviews are as emergent as they are in everyday talk, although the design of the interviewee’s stance is obviously contingent upon the design of the question. On the contrary, it seems that the interviewees do not assume and accept the positions and stances put on the table by their host but firmly state and support their own stances. It seems that stancetaking in news interviews relies on some types of prior beliefs, or contextual, political, and cultural values that are bound to the chosen topical agendas and the viewpoints assumed by the participants. In other words, stancetaking in news interviews is perhaps even more about ‘who you are,’ ‘what you say,’ and ‘how you say it,’ than it is in everyday talk. The stancetaking activities described above also have interesting implications for the audience. As Heritage (1985) has shown, for example, the fact that interviewers sometimes reformulate interviewee answers in news interviews shows that news interviews are produced for an overhearing audience. It would seem that the stancetaking activities of positioning and alignment are also targeted for the audience in that they produce and distribute information, create entertainment, report opinions and political viewpoints, tell news for the audience, and
312 Pentti Haddington
thereby also (attempt to) influence and shape public opinions. In fact, these activities are characteristic of news interviews, but are not likely to occur in other forms of question-driven institutional interaction (e.g., doctor-patient interaction). It seems, therefore, that it is not the institutionality per se that occasions these activities, but the fact that news interviews are publicly broadcast programs and aired for an audience. Although it is quite possible that one rationale behind producing these activities is the entertainment value of news interviews, 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. However, 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.
Notes 1. I am greatly indebted to the following persons for help, comments and encouragement: John Du Bois, Robert Englebretson, Tiina Keisanen, Elise Kärkkäinen, Maarit Niemelä, Arja Piirainen-Marsh, and Mirka Rauniomaa. I also want to thank the anonymous reviewer for helpful comments. Naturally, I alone am responsible for any errors and inadequacies that remain. This research has been partly funded by a grant from the Academy of Finland (research project 53671). 2. Turn constructional unit (TCU) is the basic unit out of which speakers set out to construct talk (Sacks et al. 1974). One TCU can constitute a recognizably complete turn and can characteristically be lexical, phrasal, sentential, or clausal. These linguistic characteristics of turn construction provide for “projectability of a turn,” i.e., the possibility for co-participants to realize and understand what is under way and to project the possible completion point of a TCU (Sacks et al. 1974). In everyday conversation, each speaker gets the right to construct a single TCU to a possible completion. In institutionalized interaction, however, the allocation of turns is pre-determined. 3. TRPs refer to the ends of TCUs, i.e., that there is a possibility for transition between speakers at the end of a TCU (Sacks et al. 1974). However, in news interviews, TRPs are more “neutral” than in everyday talk (cf. Schegloff 1996: fn. 14; 2001). This is a consequence of the ‘multi-unitness’ of turns and the pre-allocated turn types in news interviews. For example, interviewees are entitled to and even assumed to give answers that contain more than one TCU. Therefore, the turn-taking rules that apply to everyday talk (Sacks et al. 1974), in which one speaker is basically allotted one TCU, and in which different rules of “current speaker selects next” or “self-selection” apply at each TRP, do not apply to news interview interaction in the same way (see for example Clayman and Heritage 2002; Greatbatch 1988; Heritage 1985). 4. Although there are only scant mentions of how the interviewer’s question acts as an impetus for the interviewee’s evasiveness, Nuolijärvi (1994) discusses how the interviewer’s turn in fact creates the foundation and thereby acts as the sequential impetus for the interviewee’s nonanswers, and Clayman and Heritage (2002: 188–237) address the issue occasionally, but only in passing, together with the analysis of examples.
Stancetaking in news interviews 313
5. Note that the talk and the question in lines 1–3 is not part of the actual interview, but added in the program after the actual interview. Nevertheless, I treat the reported speech as faithful to the actual question in the interview. 6. Editing is used frequently in documentaries and news reports. 7. Diagraph is a technical term used in dialogic syntax (Du Bois 2001, this volume). Diagraphs are used to depict and illustrate linguistic relations between the talk of two speakers or within a speaker’s turn. One line in a diagraph represents one intonation unit. The purpose of diagraphs here is to show how speakers use each other’s syntax and other linguistic material to construct their stances. Curly brackets in the diagraphs indicate that some unit has been moved for representational purposes. 8. See more about the notion of neutralism and how it relates to stancetaking in Haddington (2004). 9. I thank Marja-Leena Sorjonen for this observation. 10. Example (3) is analyzed in detail in Haddington (2004). 11. In addition to seeing this as a promise that is made to the allies of United States, it is also possible to see this action as involving an (indirect) threat to al-Qaida. 12. A more quantitative analysis of this pattern in this action combination is given in Haddington (2005b). It also discusses the different types of denials and the ‘claim for insufficient knowledge + explanation’ action combination that occur in news interviews. 13. It seems that this pattern is used in this order particularly for denying something in the interviewer’s question. Other orders also occur and their analysis may reveal other functions. This, however, requires further research. 14. For more detailed discussion, see Haddington (2005b).
References Berg, M. 2001. “Vallan vahtikoira haukkuu mutta ei pure: Poliitikkoa syyllistävät toimittajan kysymykset televisiokeskustelussa [The watchdog of power barks but does not bite: Interviewer’s accusatory questions toward politicians in television conversation].” In Institutionaalinen vuorovaikutus: Keskustelunanalyyttisia tutkimuksia [Institutional interaction: Conversation analytic studies], J. Ruusuvuori, M. Haakana and L. Raevaara (eds.), 121–137. Helsinki: Suomalaisen Kirjallisuuden Seura. Berg, M. 2003. Syytöksiä ja epäilyjä: Toimittajan ja poliitikon vuorovaikutuksesta televisiokeskustelussa [Accusations and doubts: Interaction between journalists and politicians in television conversation]. Helsinki: Suomen Kirjallisuuden Seura. Biber, D., Conrad, S. and Leech, G. 2003. Longman Student Grammar of Spoken and Written English. Essex: Longman. Chafe, W. 1986. “Evidentiality in English Conversation and Academic Writing.” In Evidentiality: The Linguistic Coding of Epistemology, W. Chafe and J. Nichols (eds.), 261–72. Norwood, NJ: Ablex Publishing Corporation. Clayman, S.E. 1988. “Displaying neutrality in television news interviews.” Social Problems 35 (4): 474–492. Clayman, S.E. 1993. “Reformulating the question: A device for answering/not answering questions in news interviews and press conferences.” Text 13: 159–188. Clayman, S.E. 2001. “Answers and evasions.” Language in Society 30: 403–442.
314 Pentti Haddington
Clayman, S.E. and Heritage, J. 2002. The News Interview: Journalists and Public Figures on the Air. Cambridge: Cambridge University Press. Du Bois, J.W. 2001. Towards a Dialogic Syntax. Ms., Department of Linguistics, University of California, Santa Barbara. Du Bois, J.W., Chafe, W., Meyer, C. and Thompson, S.A. 2000. Santa Barbara Corpus of Spoken American English, CD, Part 1. Philadelphia: Linguistic Data Consortium, University of Pennsylvania. Du Bois, J.W., Chafe, W., Meyer, C., Thompson, S.A., and Martey, N. 2003. Santa Barbara Corpus of Spoken American English, CD, Part 1. Philadelphia: Linguistic Data Consortium, University of Pennsylvania. Du Bois, J.W. and Englebretson, R. 2004. Santa Barbara Corpus of Spoken American English, DVD, Part 3. Philadelphia: Linguistic Data Consortium, University of Pennsylvania. Du Bois, J.W., Schuetze-Coburn, S., Cumming, S. and Paolino, D. 1993. “Outline of Discourse Transcription.” In Talking Data: Transcription and Coding in Discourse Research, J.A. Edwards and M.A. Lampert (eds.), 45–89. Hillsdale, NJ: Lawrence Erlbaum Associates. Ford, C.E. 2002. “Denial and the construction of conversational turns.” In Complex Sentences in Grammar and Discourse. Essays in honor of Sandra A. Thompson, J.L. Bybee (ed.) 61–78. Philadelphia: John Benjamins. Ford, C.E. 2004. “Contingency and units in interaction.” Discourse Studies 6 (1): 27–52. Goffman, E. 1981. “Footing.” In Forms of Talk, E. Goffman (ed.) 124–159. Philadelphia: University of Pennsylvania Press. Greatbatch, D. 1986. “Aspects of topical organization in news interviews: the use of agenda– shifting procedures by interviewees.” Media, Culture & Society 8: 441–455. Haddington, P. 2004. “Stance taking in news interviews.” SKY Journal of Linguistics 17: 101– 142. Haddington, P. 2005a. The Intersubjectivity of Stance Taking in Talk-in-interaction. Ph.D. dissertation. Oulu: Oulu University Press. Haddington, P. 2005b. “The linguistic NEG + POS pattern and two action combinations as resources for interviewee stance taking in news interviews.” In Kieli ja sosiaalinen toiminta [Language and Social Action], L. Kuure, E. Kärkkäinen and M. Saarenkunnas (eds.), 85–107. AFinLA Yearbook. Publications de l’association finlandaise de linguistique appliquée. Heritage, J. 1984. Garfinkel and Ethnomethodology. Cambridge: Polity Press. Heritage, J. 1985. “Analyzing news interviews: Aspects of the production of talk for an ‘overhearing’ audience.” In Handbook of Discourse Analysis, Vol. III, T.A.V. Dijk (ed.) 95–119. London: Academic Press. Heritage, J. 2002. “The limits of questioning: Negative interrogatives and hostile question content.” Journal of Pragmatics 34 (10–11): 1427–1446. Heritage, J. 2003. “Designing questions and setting agendas in the news interview.” In Studies in Language and Social Interaction: In Honor of Robert Hopper, P. Glenn, C.D. LeBaron and J. Mandelbaum (eds.), 57–90. London: Lawrence Erlbaum. Heritage, J. and Greatbatch, D. 1991. “On the institutional character of institutional talk: The case of news interviews.” In Talk and Social Structure, D. Boden and D. Zimmerman (eds.), 93–137. Oxford: Polity Press. Heritage, J.C. and Roth, A.L. 1995. “Grammar and institution: Questions and questioning in the broadcast news interview.” Research on Language and Social Interaction 28 (1): 1–60. Horn, L.R. 1989. A Natural History of Negation. Chicago: The University of Chicago Press.
Stancetaking in news interviews 315
Hutchby, I. and Wooffitt, R. 1998. Conversation Analysis. Cambridge: Polity Press. Jakobson, R., Waugh, L.R. and Monville-Burston, M. 1995. On Language. Harvard: Harvard University Press. Karlsson, S. 2003. “Interactional uses for an epistemic marker: The case of ‘jag tycker’/‘tycker jag’ in Swedish.” Melbourne Papers in Linguistics and Applied Linguistics 1: 5–23. Kärkkäinen, E. 2003a. Epistemic Stance in English Conversation: A Description of Its Interactional Functions, with a Focus on ‘I think’. Amsterdam: John Benjamins. Kärkkäinen, E. 2003b. “‘Is she vicious or dense?’: Dialogic practices of stance taking in conversation.” In Functional Approaches to Grammar: Theoretical and Methodological Considerations., H. Tao (ed.) 48–66. Santa Barbara Working Papers: University of California, Santa Barbara. Levinson, S.C. 1983. Pragmatics. London: Cambridge University Press. Local, J. 1992. “Continuing and restarting.” In The Contextualization of Language, P. Auer and A. di Luzio (eds.), 273–296. Amsterdam: John Benjamins. Nuolijärvi, P. 1994. “Talousrikollinen televisiohaastattelussa: Vastauksen muotoilemisen problematiikkaa [White-collar criminal in a television interview: Problems in designing an answer].” In Talous ja kieli II [Economics and Language II], M. Majapuro and T. Nikko (eds.), 228–244. Helsinki: Helsinki School of Economics and Business Administration. Potter, J. 1998. “Discursive social psychology: From attitudes to evaluative processes.” European Review of Social Psychology 9: 233–266. Sacks, H. [1973] 1987. “On the preferences for agreement and contiguity in sequences in conversation.” In Talk and Social Organization, G. Button and J.R.E. Lee (eds.), 54–69. Clevedon: Multilingual Matters Ltd. Sacks, H. 1992. Lectures on Conversation [1964–1968]. Cambridge: Blackwell Publishers. Sacks, H. and Schegloff, E.A. 1979. “Two preferences in the organization of reference to persons and their interaction.” In Everyday Language: Studies in Ethnomethodology, G. Psathas (ed.). New York: Irvington Publishers. Sacks, H., Schegloff, E.A. and Jefferson, G. 1974. “A simplest systematics for the organization of turn-taking for conversation.” Language 50: 696–735. Scannell, P. 1991. “Introduction: The relevance of talk.” In Broadcast Talk, P. Scannell (ed.), 1–13. London: Sage. Schegloff, E.A. 1982. “Discourse as an interactional achievement: Some uses of ‘uh huh’ and other things that come between sentences.” In Georgetown University Roundtable on Languages and Linguistics 1981; Analyzing Discourse: Text and Talk, D. Tannen (ed.), 71–93. Washington D.C.: Georgetown University Press. Schegloff, E.A. 1996. “Turn organization: One intersection of grammar and interaction.” In Interaction and Grammar, E. Ochs, E.A. Schegloff and S.A. Thompson (eds.), 52–133. Cambridge: Cambridge University Press. Scheibman, J. 2001. “Local patterns of subjectivity in person and verb type in American English conversation.” In Frequency Effects and Emergent Grammar, J. Bybee and P.J. Hopper (eds.), 61–89. Amsterdam: John Benjamins. Silverstein, M. 1995. “Shifters, linguistic categories and cultural description.” In Language, Culture and Society, B.G. Blount (ed.), 187–222. Prospect Heights, IL: Waveland Press, Inc. Simon-Vandenbergen, A.-M. 2000. “The functions of I think in political discourse.” International Journal of Applied Linguistics 10 (1): 41–62.
316 Pentti Haddington
ten Have, P. 2001. “Applied conversation analysis.” In How to Analyse Talk in Institutional Settings: A Casebook of Methods, A. McHoul and M. Rapley (eds.), 3–11. London: Continuum. Thompson, S.A. 2002. “‘Object complements’ and conversation: Towards a realistic account.” Studies in Language 26 (1): 125–164.
Appendix: Transcription conventions Based on Du Bois et al. (1993). Speech Units Intonation unit Truncated intonation unit Truncated word
{line break} –– – (en dash)
Transitional continuity Final Continuing 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) Secondary accent Unaccented Lengthening Pause Long pause (0.7 seconds or longer) Medium pause (0.3 – 0.6 s) Short (brief break in speech rhythm)(0.2 or less) Latching Vocal noises Alveolar click Glottal stop Inhalation Laughter (one pulse) Laughter during speech (1–5 words) (e.g., @two @words)
[ ] [2 two words 2] ^ ` = ...(N) ... .. (0) (TSK) (GLOTTAL) (H) @ @
Quality Tempo and rhythm Allegro: rapid speech Lento: slow speech Marcato: each word distinct and emphasized Rhythmic: stresses in a beatable rhythm Voice quality Creak Creak during speech (e.g., %two %words) Transcriber’s perspective Uncertain hearing (e.g., #two #words) Researcher’s comment Indecipherable syllable Specialized notations Restart
Stancetaking in news interviews 317
<MRC> % %
# (( )) # {Capital initial}
Name index
A Antaki, Charles 52 Athanasiadou, Angeliki 16 B Bailey, Guy 54 Becher, Tony 35–36 Bednarek, Monika A. 16 Benveniste, Émil 16, 122, 125 Benwell, Bethan 83 Berman, Ruth 18, 111–113, 118 Biber, Douglas 16–17, 28–30 Bouacha, M. A. 118 Brown, Roger 74 Bucholtz, Mary 52, 84 C Chafe, Wallace L. 120, 190 Channell, Joanna 36–37 Charles, M. 28, 30–34 Clayman, Steven 284, 297 Clift, Rebecca 184, 260 Conrad, Susan 28–30
F Field, Margaret 17 Finegan, Edward 17 Fitzmaurice, Susan 19 Ford, Cecelia E. 284, 303, 306 G Gilman, Albert 74 Groom, Nick 35–36 H Haddington, Pentti 22, 194 Hall, Kira 52, 84 Hazen, Kirk 54 Helasvuo, Marja-Liisa 225–226 Heritage, John 255–256, 284, 297, 311 Hunston, Susan 16, 20, 39–44, 123 Hyland, Ken 17, 20 J Johnstone, Barbara 18, 20, 55
D Dasher, Richard 125 Du Bois, John W. 20, 51–52, 140, 199, 225, 231, 283, 285– 286 Duranti, Alessandro 97, 103–104
K Kafura, Dennis 116–117, 134 Kärkkäinen, Elise 17–18, 21, 183, 186, 189, 226, 292 Keisanen, Tiina 21 Kiesling, Scott F. 55 Koshik, Irene 254–256
E Echols, John M. 76 Eckert, Penelope 55 Edwards, Derek 19 Englebretson Robert 18, 20–21, 72–73 Errington, J. Joseph 74
L Labov, William 53 Langacker, Ronald W. 16, 115, 120 Lenk, Uta 189 LePage, Robert B. 53–54 Louw, Bill 37–39, 41–42 Lyons, John 15–16, 118
M MacWhinney, Brian 120 Martin, J. R. 16 Matoesian, Gregory 18 McConnell-Ginet, Sally 134 Milroy, James 53 Milroy, Lesley 53 Mulac, Anthony 190 N Nuyts, Jan 125 O Ochs, Elinor 18, 51 P Pomerantz, Annita 277 Precht, Kristen 18 R Rampton, Ben 54 Rauniomaa, Mirka 21 Rose, Mary 54 S Sacks, Harvey 256, 290–291, 307 Schegloff, Emanuel A. 256, 284, 289–291 Scheibman, Joanne 17–18, 21, 187–188 Schiffrin, Deborah 125, 188– 189, 207–209 Schilling-Estes, Natalie 54 Shadily, Hassan 76 Shoaps, Robin 18 Simon-Vandenbergen, AnneMarie 289 Sinclair, John 37, 39 Sneddon, James N. 76–77 Stokoe, Elizabeth 83
320 Name index
Strauss, Claudia 125 Stubbs, Michael 17, 70
Trudgill, Peter 53 Tse, Polly 17, 20
T Tabouret-Keller, Andrée 53–54 Thompson, Geoff 16, 123 Thompson, Sandra A. 187, 189, 213 Traugott, Elizabeth Closs 125, 188
V van Leeuwen, Theo 111, 117–118 Verhagen, Arie 19, 122 W White, Peter R. R. 16, 19 Widdicombe, Susan M. 52
Wittgenstein, Ludwig 171 Wouk, Fay 70 Wu, Ruey-Jiuan Regina 18, 88–89, 95, 225 Z Zhang, Qiao 111
Subject index
A adjacency pairs 256, 290–291 adjectives 11–14, 33–36 adverbs 17, 29, 91, 188 stance adverbials 29–30, 71 affect see affective stance affective stance 17, 94–95, 105, 143–144, 152, 162, 260–261, 272 affiliation 54, 70, 74–75, 197 see also solidarity agency 97–99, 103–105 alignment 139, 144–145, 150– 153, 159, 162–167, 169–170, 197, 226, 253, 277, 283, 285, 310 the neg + pos pattern 303– 309, 311 to hostile interview questions 287–290, 296–297 to preferred stances 293– 297 to presuppositions 299–303 intersubjectivity appraisal 142 assessment of interactional relevance 93–95, 105 assessments 142, 225–226 attitudinal stance 29, 36 B British National Corpus 4–6, 11–13, 22–23 and VIEW (Variation in English Words and Phrases) 4, 22–23 C challenges 254–255 collocations 11–15, 188–189 see also phraseologies complement clauses 17, 30, 188
complement-taking predicates 17, 30, 186–189 constructed dialogue 85–86 Conversation Analysis 19–20, 184, 253–254, 283–285 corpus linguistics 20, 27–28 corpus of conversational Finnish (Helsinki) 223–224 D diagraph 160–161, 166, 172, 199–200, 231 dialect identity see local identity dialogicality 19, 139–141, 147, 152, 157–158, 160–161 see also intersubjectivity dialogic syntax 140, 160, 172 diathesis (in Colloquial Indonesian) 95–105 discourse markers 188–189, 206, 208–211 discourse stance 18–19, 111–113 E epistemicity see epistemic stance epistemic phrases 186–190, 212 epistemic stance 17–18, 20–21, 29, 50–51, 61–66, 71, 88–89, 91–95, 105, 143–144, 156, 162, 183, 222–223, 253, 257 evaluation 6, 9–10, 12–14, 16–17, 35–36, 40–41, 112–113, 122–124, 127–129, 141–146, 153, 155, 158, 163–166, 169–170, 191–194, 226, 253, 270 evaluative stance see evaluation evidentiality 91–93, 95, 105 evidentials 190, 211–212
F finite clauses 29 first–person reference (in Colloquial Indonesian) 73– 88, 104–105 G gaze 195–196 generality of reference 111–117 general subjects 118–122, 126 generic you 130–133, 296 grammar pattern 30–34 H hedges 222 I identity 50–55, 73–74, 83, 87–88 performance of 51–55, 62–66, 83–84 see also local identity I guess (English) in second–position actions 197–203 in sequence–initiating actions 190–197 in side sequences 204–208, 210 inference 191–194 institutional stance see moral stance interactional linguistics 253– 254 interrogatives negative yes/no questions 255–257 as challenge to prior action 271–273 as challenge to telling sequence 266–269
322 Subject index
tag questions 255–257, 310 as challenge to prior action 269–271, 273–275 as challenge to telling sequence 258–266 wh-questions 254–255, 310 intersubjectivity 19–20, 111, 113–114, 124–125, 133–134, 140–141, 158–162, 168, 170, 184, 212, 283, 286, 309–311 see also alignment L local identity 20, 49–50, 53–56, 62–66 M meaning group 30–34 minun mielestä and minusta (Finnish) marking transition to a first assessment across turns 233–237 marking transition to first assessments within extended turns 237–241 opening up assessments for mutual attention 241– 246 position in IU 226–227 position in turn sequence 227–228 projecting disagreement in second assessments 228–233 modals 17, 71 moral stance 6, 8, 10, 11–14, 18, 96–99, 103–104 N news interviews 255–256, 284, 309–312 non-personal they 127–129 -nya clitic (Colloquial Indonesian) 88–95, 105 P particles 70 personal stance see evaluation
phraseologies 27, 36–37, 39–40, 42–45 see also collocations physical stance 6–8, 12, 14, 18 positioning 95–99, 103–105, 113–114, 139, 143, 145, 152–153, 155, 158–159, 162–166, 169–170, 283, 286, 310 by hostile questions 287– 290, 295–296 by preferred stances in questions 290–292, 294–297 by presupposition in questions 297–301 preference structure 290–291 prepositional phrases 29 presupposition 298–300 prosody 10, 260–261, 263, 270–272, 307–309 Q questions see interrogatives R register 6, 14, 29–30 reported speech see constructed dialogue S Santa Barbara Corpus of Spoken American English 4–10, 22, 255–257 semantic prosody 39 sequence organization 277–278 sequentiality 149, 154–156, 159–161, 168, 212 social network theory 53 sociolinguistics (variationist) 52–54, 83 solidarity 114–115, 124, 126, 128–129 see also affiliation stance across genre 19 and first-person subjects 225 as collaborative activity see stance: interactional nature of
consequentiality of 6–8, 11, 14–15, 173 definitions and characterizations of 1–4, 6–15, 21, 51–52, 142–145 diversity of 1–4, 15, 144–145, 169 historical development of 18–19 indexicality of 6, 8, 11, 14–15, 18–19, 51, 55, 111–112, 114, 124, 146 interactional nature of 6, 8, 10–11, 13–15, 19–22, 112–114, 125–126, 128–129, 134, 157, 171, 185, 254, 277–278, 283 interdisciplinary nature of 1–4 in writing 19–20 joint construction of see stance: interactional nature of metalinguistics of 3–15 projectability of 151 public nature of 6–8, 10, 13–15, 163, 169, 171, 173 relational nature of see stance: interactional nature of responsibility for 147, 171, 173, 277–278 sequential nature of 283 sociocultural dimensions of 6, 8, 11, 15, 18, 43–44, 112–114, 117–118, 124–126, 128–129, 134, 139–141, 173 see also affective stance, appraisal, attitudinal stance, discourse stance, epistemic stance, evaluation, moral stance, physical stance, style stance stance differential 166–167, 170, 225, 300 stance frame 184–185, 188 stance object 143, 147–148, 151–157, 159, 163–166, 169–170, 283
stance subject 147, 151–153, 155, 157, 159, 163–166, 169–170 style stance 29 subjectivity 15–18, 111, 113, 122, 125, 133–134, 140–141, 152–154, 156–157, 159, 170, 184, 225
Subject index 323
T transcription 4, 20 turn design 194–197, 277–278 U usage-based linguistics 4, 69–71
V voice (grammatical) see diathesis
Pragmatics & Beyond New Series A complete list of titles in this series can be found on the publishers’ website, www.benjamins.com 171 Félix-Brasdefer, J. César: Politeness in Mexico and the United States. A contrastive study of the realization and perception of refusals. xiv, 190 pp. + index. Expected January 2008 170 Hougaard, Anders and Todd Oakley (eds.): Mental Spaces in Discourse and Interaction. Expected January 2008 169 Connor, Ulla, Edwin R. Nagelhout and William V. Rozycki (eds.): Contrastive Rhetoric. Reaching to intercultural rhetoric. viii, 319 pp. + index. Expected December 2007 168 Proost, Kristel: Conceptual Structure in Lexical Items. The lexicalisation of communication concepts in English, German and Dutch. xii, 296 pp. + index. Expected December 2007 167 Bousfield, Derek: Impoliteness in Interaction. xiii, 278 pp. + index. Expected November 2007 166 Nakane, Ikuko: Silence in Intercultural Communication. Perceptions and performance. xii, 233 pp. + index. Expected November 2007 165 Bublitz, Wolfram and Axel Hübler (eds.): Metapragmatics in Use. vi, 290 pp. + index. Expected December 2007 164 Englebretson, Robert (ed.): Stancetaking in Discourse. Subjectivity, evaluation, interaction. 2007. vii, 323 pp. 163 Lytra, Vally: Play Frames and Social Identities. Contact encounters in a Greek primary school. xi, 291 pp. + index. Expected November 2007 162 Fetzer, Anita (ed.): Context and Appropriateness. Micro meets macro. 2007. vi, 265 pp. 161 Celle, Agnès and Ruth Huart (eds.): Connectives as Discourse Landmarks. 2007. viii, 212 pp. 160 Fetzer, Anita and Gerda Eva Lauerbach (eds.): Political Discourse in the Media. Cross-cultural perspectives. 2007. viii, 379 pp. 159 Maynard, Senko K.: Linguistic Creativity in Japanese Discourse. Exploring the multiplicity of self, perspective, and voice. 2007. xvi, 356 pp. 158 Walker, Terry: Thou and You in Early Modern English Dialogues. Trials, Depositions, and Drama Comedy. 2007. xx, 339 pp. 157 Crawford Camiciottoli, Belinda: The Language of Business Studies Lectures. A corpus-assisted analysis. 2007. xvi, 236 pp. 156 Vega Moreno, Rosa E.: Creativity and Convention. The pragmatics of everyday figurative speech. 2007. xii, 249 pp. 155 Hedberg, Nancy and Ron Zacharski (eds.): The Grammar–Pragmatics Interface. Essays in honor of Jeanette K. Gundel. 2007. viii, 345 pp. 154 Hübler, Axel: The Nonverbal Shift in Early Modern English Conversation. 2007. x, 281 pp. 153 Arnovick, Leslie K.: Written Reliquaries. The resonance of orality in medieval English texts. 2006. xii, 292 pp. 152 Warren, Martin: Features of Naturalness in Conversation. 2006. x, 272 pp. 151 Suzuki, Satoko (ed.): Emotive Communication in Japanese. 2006. x, 234 pp. 150 Busse, Beatrix: Vocative Constructions in the Language of Shakespeare. 2006. xviii, 525 pp. 149 Locher, Miriam A.: Advice Online. Advice-giving in an American Internet health column. 2006. xvi, 277 pp. 148 Fløttum, Kjersti, Trine Dahl and Torodd Kinn: Academic Voices. Across languages and disciplines. 2006. x, 309 pp. 147 Hinrichs, Lars: Codeswitching on the Web. English and Jamaican Creole in e-mail communication. 2006. x, 302 pp. 146 Tanskanen, Sanna-Kaisa: Collaborating towards Coherence. Lexical cohesion in English discourse. 2006. ix, 192 pp. 145 Kurhila, Salla: Second Language Interaction. 2006. vii, 257 pp. 144 Bührig, Kristin and Jan D. ten Thije (eds.): Beyond Misunderstanding. Linguistic analyses of intercultural communication. 2006. vi, 339 pp. 143 Baker, Carolyn, Michael Emmison and Alan Firth (eds.): Calling for Help. Language and social interaction in telephone helplines. 2005. xviii, 352 pp. 142 Sidnell, Jack: Talk and Practical Epistemology. The social life of knowledge in a Caribbean community. 2005. xvi, 255 pp.
141 Zhu, Yunxia: Written Communication across Cultures. A sociocognitive perspective on business genres. 2005. xviii, 216 pp. 140 Butler, Christopher S., María de los Ángeles Gómez-González and Susana M. Doval-Suárez (eds.): The Dynamics of Language Use. Functional and contrastive perspectives. 2005. xvi, 413 pp. 139 Lakoff, Robin T. and Sachiko Ide (eds.): Broadening the Horizon of Linguistic Politeness. 2005. xii, 342 pp. 138 Müller, Simone: Discourse Markers in Native and Non-native English Discourse. 2005. xviii, 290 pp. 137 Morita, Emi: Negotiation of Contingent Talk. The Japanese interactional particles ne and sa. 2005. xvi, 240 pp. 136 Sassen, Claudia: Linguistic Dimensions of Crisis Talk. Formalising structures in a controlled language. 2005. ix, 230 pp. 135 Archer, Dawn: Questions and Answers in the English Courtroom (1640–1760). A sociopragmatic analysis. 2005. xiv, 374 pp. 134 Skaffari, Janne, Matti Peikola, Ruth Carroll, Risto Hiltunen and Brita Wårvik (eds.): Opening Windows on Texts and Discourses of the Past. 2005. x, 418 pp. 133 Marnette, Sophie: Speech and Thought Presentation in French. Concepts and strategies. 2005. xiv, 379 pp. 132 Onodera, Noriko O.: Japanese Discourse Markers. Synchronic and diachronic discourse analysis. 2004. xiv, 253 pp. 131 Janoschka, Anja: Web Advertising. New forms of communication on the Internet. 2004. xiv, 230 pp. 130 Halmari, Helena and Tuija Virtanen (eds.): Persuasion Across Genres. A linguistic approach. 2005. x, 257 pp. 129 Taboada, María Teresa: Building Coherence and Cohesion. Task-oriented dialogue in English and Spanish. 2004. xvii, 264 pp. 128 Cordella, Marisa: The Dynamic Consultation. A discourse analytical study of doctor–patient communication. 2004. xvi, 254 pp. 127 Brisard, Frank, Michael Meeuwis and Bart Vandenabeele (eds.): Seduction, Community, Speech. A Festschrift for Herman Parret. 2004. vi, 202 pp. 126 Wu, Yi’an: Spatial Demonstratives in English and Chinese. Text and Cognition. 2004. xviii, 236 pp. 125 Lerner, Gene H. (ed.): Conversation Analysis. Studies from the first generation. 2004. x, 302 pp. 124 Vine, Bernadette: Getting Things Done at Work. The discourse of power in workplace interaction. 2004. x, 278 pp. 123 Márquez Reiter, Rosina and María Elena Placencia (eds.): Current Trends in the Pragmatics of Spanish. 2004. xvi, 383 pp. 122 González, Montserrat: Pragmatic Markers in Oral Narrative. The case of English and Catalan. 2004. xvi, 410 pp. 121 Fetzer, Anita: Recontextualizing Context. Grammaticality meets appropriateness. 2004. x, 272 pp. 120 Aijmer, Karin and Anna-Brita Stenström (eds.): Discourse Patterns in Spoken and Written Corpora. 2004. viii, 279 pp. 119 Hiltunen, Risto and Janne Skaffari (eds.): Discourse Perspectives on English. Medieval to modern. 2003. viii, 243 pp. 118 Cheng, Winnie: Intercultural Conversation. 2003. xii, 279 pp. 117 Wu, Ruey-Jiuan Regina: Stance in Talk. A conversation analysis of Mandarin final particles. 2004. xvi, 260 pp. 116 Grant, Colin B. (ed.): Rethinking Communicative Interaction. New interdisciplinary horizons. 2003. viii, 330 pp. 115 Kärkkäinen, Elise: Epistemic Stance in English Conversation. A description of its interactional functions, with a focus on I think. 2003. xii, 213 pp. 114 Kühnlein, Peter, Hannes Rieser and Henk Zeevat (eds.): Perspectives on Dialogue in the New Millennium. 2003. xii, 400 pp. 113 Panther, Klaus-Uwe and Linda L. Thornburg (eds.): Metonymy and Pragmatic Inferencing. 2003. xii, 285 pp. 112 Lenz, Friedrich (ed.): Deictic Conceptualisation of Space, Time and Person. 2003. xiv, 279 pp. 111 Ensink, Titus and Christoph Sauer (eds.): Framing and Perspectivising in Discourse. 2003. viii, 227 pp.
110 Androutsopoulos, Jannis K. and Alexandra Georgakopoulou (eds.): Discourse Constructions of Youth Identities. 2003. viii, 343 pp. 109 Mayes, Patricia: Language, Social Structure, and Culture. A genre analysis of cooking classes in Japan and America. 2003. xiv, 228 pp. 108 Barron, Anne: Acquisition in Interlanguage Pragmatics. Learning how to do things with words in a study abroad context. 2003. xviii, 403 pp. 107 Taavitsainen, Irma and Andreas H. Jucker (eds.): Diachronic Perspectives on Address Term Systems. 2003. viii, 446 pp. 106 Busse, Ulrich: Linguistic Variation in the Shakespeare Corpus. Morpho-syntactic variability of second person pronouns. 2002. xiv, 344 pp. 105 Blackwell, Sarah E.: Implicatures in Discourse. The case of Spanish NP anaphora. 2003. xvi, 303 pp. 104 Beeching, Kate: Gender, Politeness and Pragmatic Particles in French. 2002. x, 251 pp. 103 Fetzer, Anita and Christiane Meierkord (eds.): Rethinking Sequentiality. Linguistics meets conversational interaction. 2002. vi, 300 pp. 102 Leafgren, John: Degrees of Explicitness. Information structure and the packaging of Bulgarian subjects and objects. 2002. xii, 252 pp. 101 Luke, K. K. and Theodossia-Soula Pavlidou (eds.): Telephone Calls. Unity and diversity in conversational structure across languages and cultures. 2002. x, 295 pp. 100 Jaszczolt, Katarzyna M. and Ken Turner (eds.): Meaning Through Language Contrast. Volume 2. 2003. viii, 496 pp. 99 Jaszczolt, Katarzyna M. and Ken Turner (eds.): Meaning Through Language Contrast. Volume 1. 2003. xii, 388 pp. 98 Duszak, Anna (ed.): Us and Others. Social identities across languages, discourses and cultures. 2002. viii, 522 pp. 97 Maynard, Senko K.: Linguistic Emotivity. Centrality of place, the topic-comment dynamic, and an ideology of pathos in Japanese discourse. 2002. xiv, 481 pp. 96 Haverkate, Henk: The Syntax, Semantics and Pragmatics of Spanish Mood. 2002. vi, 241 pp. 95 Fitzmaurice, Susan M.: The Familiar Letter in Early Modern English. A pragmatic approach. 2002. viii, 263 pp. 94 McIlvenny, Paul (ed.): Talking Gender and Sexuality. 2002. x, 332 pp. 93 Baron, Bettina and Helga Kotthoff (eds.): Gender in Interaction. Perspectives on femininity and masculinity in ethnography and discourse. 2002. xxiv, 357 pp. 92 Gardner, Rod: When Listeners Talk. Response tokens and listener stance. 2001. xxii, 281 pp. 91 Gross, Joan: Speaking in Other Voices. An ethnography of Walloon puppet theaters. 2001. xxviii, 341 pp. 90 Kenesei, István and Robert M. Harnish (eds.): Perspectives on Semantics, Pragmatics, and Discourse. A Festschrift for Ferenc Kiefer. 2001. xxii, 352 pp. 89 Itakura, Hiroko: Conversational Dominance and Gender. A study of Japanese speakers in first and second language contexts. 2001. xviii, 231 pp. 88 Bayraktaroğlu, Arın and Maria Sifianou (eds.): Linguistic Politeness Across Boundaries. The case of Greek and Turkish. 2001. xiv, 439 pp. 87 Mushin, Ilana: Evidentiality and Epistemological Stance. Narrative Retelling. 2001. xviii, 244 pp. 86 Ifantidou, Elly: Evidentials and Relevance. 2001. xii, 225 pp. 85 Collins, Daniel E.: Reanimated Voices. Speech reporting in a historical-pragmatic perspective. 2001. xx, 384 pp. 84 Andersen, Gisle: Pragmatic Markers and Sociolinguistic Variation. A relevance-theoretic approach to the language of adolescents. 2001. ix, 352 pp. 83 Márquez Reiter, Rosina: Linguistic Politeness in Britain and Uruguay. A contrastive study of requests and apologies. 2000. xviii, 225 pp. 82 Khalil, Esam N.: Grounding in English and Arabic News Discourse. 2000. x, 274 pp. 81 Di Luzio, Aldo, Susanne Günthner and Franca Orletti (eds.): Culture in Communication. Analyses of intercultural situations. 2001. xvi, 341 pp. 80 Ungerer, Friedrich (ed.): English Media Texts – Past and Present. Language and textual structure. 2000. xiv, 286 pp. 79 Andersen, Gisle and Thorstein Fretheim (eds.): Pragmatic Markers and Propositional Attitude. 2000. viii, 273 pp.
78 77 76 75 74 73 72 71 70 69 68 67 66 65 64 63
62 61 60 59 58 57 56 55 54 53 52 51 50 49 48 47 46 45 44 43 42
Sell, Roger D.: Literature as Communication. The foundations of mediating criticism. 2000. xiv, 348 pp. Vanderveken, Daniel and Susumu Kubo (eds.): Essays in Speech Act Theory. 2002. vi, 328 pp. Matsui, Tomoko: Bridging and Relevance. 2000. xii, 251 pp. Pilkington, Adrian: Poetic Effects. A relevance theory perspective. 2000. xiv, 214 pp. Trosborg, Anna (ed.): Analysing Professional Genres. 2000. xvi, 256 pp. Hester, Stephen K. and David Francis (eds.): Local Educational Order. Ethnomethodological studies of knowledge in action. 2000. viii, 326 pp. Marmaridou, Sophia S.A.: Pragmatic Meaning and Cognition. 2000. xii, 322 pp. Gómez-González, María de los Ángeles: The Theme–Topic Interface. Evidence from English. 2001. xxiv, 438 pp. Sorjonen, Marja-Leena: Responding in Conversation. A study of response particles in Finnish. 2001. x, 330 pp. Noh, Eun-Ju: Metarepresentation. A relevance-theory approach. 2000. xii, 242 pp. Arnovick, Leslie K.: Diachronic Pragmatics. Seven case studies in English illocutionary development. 2000. xii, 196 pp. Taavitsainen, Irma, Gunnel Melchers and Päivi Pahta (eds.): Writing in Nonstandard English. 2000. viii, 404 pp. Jucker, Andreas H., Gerd Fritz and Franz Lebsanft (eds.): Historical Dialogue Analysis. 1999. viii, 478 pp. Cooren, François: The Organizing Property of Communication. 2000. xvi, 272 pp. Svennevig, Jan: Getting Acquainted in Conversation. A study of initial interactions. 2000. x, 384 pp. Bublitz, Wolfram, Uta Lenk and Eija Ventola (eds.): Coherence in Spoken and Written Discourse. How to create it and how to describe it. Selected papers from the International Workshop on Coherence, Augsburg, 24-27 April 1997. 1999. xiv, 300 pp. Tzanne, Angeliki: Talking at Cross-Purposes. The dynamics of miscommunication. 2000. xiv, 263 pp. Mills, Margaret H. (ed.): Slavic Gender Linguistics. 1999. xviii, 251 pp. Jacobs, Geert: Preformulating the News. An analysis of the metapragmatics of press releases. 1999. xviii, 428 pp. Kamio, Akio and Ken-ichi Takami (eds.): Function and Structure. In honor of Susumu Kuno. 1999. x, 398 pp. Rouchota, Villy and Andreas H. Jucker (eds.): Current Issues in Relevance Theory. 1998. xii, 368 pp. Jucker, Andreas H. and Yael Ziv (eds.): Discourse Markers. Descriptions and theory. 1998. x, 363 pp. Tanaka, Hiroko: Turn-Taking in Japanese Conversation. A Study in Grammar and Interaction. 2000. xiv, 242 pp. Allwood, Jens and Peter Gärdenfors (eds.): Cognitive Semantics. Meaning and cognition. 1999. x, 201 pp. Hyland, Ken: Hedging in Scientific Research Articles. 1998. x, 308 pp. Mosegaard Hansen, Maj-Britt: The Function of Discourse Particles. A study with special reference to spoken standard French. 1998. xii, 418 pp. Gillis, Steven and Annick De Houwer (eds.): The Acquisition of Dutch. With a Preface by Catherine E. Snow. 1998. xvi, 444 pp. Boulima, Jamila: Negotiated Interaction in Target Language Classroom Discourse. 1999. xiv, 338 pp. Grenoble, Lenore A.: Deixis and Information Packaging in Russian Discourse. 1998. xviii, 338 pp. Kurzon, Dennis: Discourse of Silence. 1998. vi, 162 pp. Kamio, Akio: Territory of Information. 1997. xiv, 227 pp. Chesterman, Andrew: Contrastive Functional Analysis. 1998. viii, 230 pp. Georgakopoulou, Alexandra: Narrative Performances. A study of Modern Greek storytelling. 1997. xvii, 282 pp. Paltridge, Brian: Genre, Frames and Writing in Research Settings. 1997. x, 192 pp. Bargiela-Chiappini, Francesca and Sandra J. Harris: Managing Language. The discourse of corporate meetings. 1997. ix, 295 pp. Janssen, Theo and Wim van der Wurff (eds.): Reported Speech. Forms and functions of the verb. 1996. x, 312 pp. Kotthoff, Helga and Ruth Wodak (eds.): Communicating Gender in Context. 1997. xxvi, 424 pp.
41 Ventola, Eija and Anna Mauranen (eds.): Academic Writing. Intercultural and textual issues. 1996. xiv, 258 pp. 40 Diamond, Julie: Status and Power in Verbal Interaction. A study of discourse in a close-knit social network. 1996. viii, 184 pp. 39 Herring, Susan C. (ed.): Computer-Mediated Communication. Linguistic, social, and cross-cultural perspectives. 1996. viii, 326 pp. 38 Fretheim, Thorstein and Jeanette K. Gundel (eds.): Reference and Referent Accessibility. 1996. xii, 312 pp. 37 Carston, Robyn and Seiji Uchida (eds.): Relevance Theory. Applications and implications. 1998. x, 300 pp. 36 Chilton, Paul, Mikhail V. Ilyin and Jacob L. Mey (eds.): Political Discourse in Transition in Europe 1989–1991. 1998. xi, 272 pp. 35 Jucker, Andreas H. (ed.): Historical Pragmatics. Pragmatic developments in the history of English. 1995. xvi, 624 pp. 34 Barbe, Katharina: Irony in Context. 1995. x, 208 pp. 33 Goossens, Louis, Paul Pauwels, Brygida Rudzka-Ostyn, Anne-Marie SimonVandenbergen and Johan Vanparys: By Word of Mouth. Metaphor, metonymy and linguistic action in a cognitive perspective. 1995. xii, 254 pp. 32 Shibatani, Masayoshi and Sandra A. Thompson (eds.): Essays in Semantics and Pragmatics. In honor of Charles J. Fillmore. 1996. x, 322 pp. 31 Wildgen, Wolfgang: Process, Image, and Meaning. A realistic model of the meaning of sentences and narrative texts. 1994. xii, 281 pp. 30 Wortham, Stanton E.F.: Acting Out Participant Examples in the Classroom. 1994. xiv, 178 pp. 29 Barsky, Robert F.: Constructing a Productive Other. Discourse theory and the Convention refugee hearing. 1994. x, 272 pp. 28 Van de Walle, Lieve: Pragmatics and Classical Sanskrit. A pilot study in linguistic politeness. 1993. xii, 454 pp. 27 Suter, Hans-Jürg: The Wedding Report. A prototypical approach to the study of traditional text types. 1993. xii, 314 pp. 26 Stygall, Gail: Trial Language. Differential discourse processing and discursive formation. 1994. xii, 226 pp. 25 Couper-Kuhlen, Elizabeth: English Speech Rhythm. Form and function in everyday verbal interaction. 1993. x, 346 pp. 24 Maynard, Senko K.: Discourse Modality. Subjectivity, Emotion and Voice in the Japanese Language. 1993. x, 315 pp. 23 Fortescue, Michael, Peter Harder and Lars Kristoffersen (eds.): Layered Structure and Reference in a Functional Perspective. Papers from the Functional Grammar Conference, Copenhagen, 1990. 1992. xiii, 444 pp. 22 Auer, Peter and Aldo Di Luzio (eds.): The Contextualization of Language. 1992. xvi, 402 pp. 21 Searle, John R., Herman Parret and Jef Verschueren: (On) Searle on Conversation. Compiled and introduced by Herman Parret and Jef Verschueren. 1992. vi, 154 pp. 20 Nuyts, Jan: Aspects of a Cognitive-Pragmatic Theory of Language. On cognition, functionalism, and grammar. 1991. xii, 399 pp. 19 Baker, Carolyn and Allan Luke (eds.): Towards a Critical Sociology of Reading Pedagogy. Papers of the XII World Congress on Reading. 1991. xxi, 287 pp. 18 Johnstone, Barbara: Repetition in Arabic Discourse. Paradigms, syntagms and the ecology of language. 1991. viii, 130 pp. 17 Piéraut-Le Bonniec, Gilberte and Marlene Dolitsky (eds.): Language Bases ... Discourse Bases. Some aspects of contemporary French-language psycholinguistics research. 1991. vi, 342 pp. 16 Mann, William C. and Sandra A. Thompson (eds.): Discourse Description. Diverse linguistic analyses of a fund-raising text. 1992. xiii, 409 pp. 15 Komter, Martha L.: Conflict and Cooperation in Job Interviews. A study of talks, tasks and ideas. 1991. viii, 252 pp. 14 Schwartz, Ursula V.: Young Children's Dyadic Pretend Play. A communication analysis of plot structure and plot generative strategies. 1991. vi, 151 pp.
