Indeterminacy and Regularization: a process-based approach to the study of sociolinguistic variation and language ideologies

!"#$ %"# &.' '(()

*!!+: ,-.(-/01) (23456) *!!+: ,-.(-/0.- ("5#457)

Sociolinguistic Studies

L O N D O N

doi : 10.1558/sols.v3i2.229 '')–'., ©'((), 78945": 29;#4$<45=

Article

Indeterminacy and regularization:a process-based approach to the study of

sociolinguistic variation and language ideologies

Alexandra M. Ja!e

Abstract"is article explores the concept of indeterminacy as a fundamental property of social life (Falk-Moore 1978) as it relates to key issues in sociolinguistic theory: the relationship between structure and agency and the way that linguistic signs are invested with social meanings. Indeterminacy emerges as a fundamental principle of sociolinguistic variables, context, speaker identity and participation structures, as a resource for speakers and as a possible objective of communicative practice. It stands in constant tension with processes of sociolinguistic regularization, as it is instantiated across texts, time and discourses and as it is embedded in particular social and political #elds. "e article argues that studying this tension is how a contemporary sociolinguistics can approach a coherent account of agency and constraint, change and continuity, reproduction and contestation of normative practices and dominant language ideologies.

K7>?"3@$: $"A4"#45=94$64A$; 45@7673B45CA>; 45@7:4AC#46>; 6<7"3>; 4@7"#"=>; $"A4"#45=94$64A %C34C;#7$

A!liation

California State University, Long Beach, USA .

Correspondence: Department of Linguistics, California State University, Long Beach, 1250 Bellflower Blvd., Long Beach CA 90840, USA.

email: [email protected]

230 !"A4"#45=94$64A !69@47$

1 Introduction

One of the central questions in sociolinguistic theory has to do with the relationship between structure and agency; determination and creativity (see Coupland, 1998, 2001; Sarangi, 2001). Dis question underlies issues of power and inequality as they are connected to language; it is also at the heart of the most basic of sociolinguistic questions: How do social meanings get attached to linguistic forms? In this essay, I take up what Coupland (2001:15) calls an ‘integrationist’ approach on the relationship between structure and agency, and focus my attention on one dimension of this relationship: indeterminacy. My argument is that a focus on indeterminacy as a principle of social life comple-ments a number of analytic orientations in contemporary sociolinguistic work, and that analyzing how indeterminacy operates in sociolinguistic data can lend depth to our understandings of social process.

De idea that indeterminacy was a fundamental component of social process was launched by legal anthropologist Sally Falk-Moore in an early book chapter entitled ‘Uncertainties in situations, indeterminacies in cultures’ (1978). In this work, Falk-Moore emphasizes the fundamental indeterminacy and ambiguity that inheres in social systems, and suggests that the core dynamic of social life is a dynamic tension between processes of regularization (in which people and groups ‘struggle against’ indeterminacy to impose order and normativity) and processes of situational adjustment in which people and groups attempt to maintain indeterminacy for any number of social purposes (p. 50–51). She points out that the outcomes of either of these processes may be either continu-ity or change, and that processes of situational adjustment, repeated suEciently oFen, may become normative. Falk-Moore’s work can be situated within a current of process and practice-based orientations to social theory that gained considerable ground in Anthropology and Sociology in the 1980s, and which included the work of Barth (1981), Giddens (1984) and Bourdieu (1977). Like many of her contemporaries, Falk-Moore was writing against narrow structural analyses of social institutions and focusing attention on reglementation (or systems of control) as practice and as process.

Since that time, practice and process-based approaches to social life have informed a signiGcant current of sociolinguistic work, in particular in critical, interactional sociolinguistic studies that focus on how sociolinguistic hier-archies are instantiated, reproduced and implicated in the maintenance of social hierarchies (see, for example, Blommaert, 2005; Heller, 2002, 2007). However, there has been less attention to both the principle of indeterminacy, and how it is implicated in both processes of regularization and situational adjustment. I will suggest that it is productive to ground sociolinguistic research and analysis on a premise of inherent indeterminacy, both as a deGning feature

45@7673B45CA> C5@ 37=9#C34HC64"5 231

of sociolinguistic data and as a baseline for our understandings of social action through language. Secondly, an indeterminacy based framework oIers a coher-ent and concise vantage point for the articulation of the importance of several contemporary concepts and currents of research in sociolinguistics/linguistic anthropology. Dese include the notion of second-order indexicality as well as work on entextualization/intertextuality/interdiscursivity and language ideolo-gies. De former allows us to understand how dominant language ideologies are reproduced even through actions that overtly contest them: those ideologies are second-order indexicals that constitute the background knowledge that both structure and are needed to interpret forms of contestation (JaIe, 1999a). At the same time, new forms of action allow new meanings to accrue to linguistic forms, and thus become part of what is indexed in action.

Dis approach thus raises the following questions: Under what conditions do speakers allow, live with or even actively seek indeterminacy in communication, and for what kinds of social purposes? At the same time, what interactional, political, social, ideological, cultural and economic factors lead people to avoid indeterminacy or to Gx meanings? What is the unique role of language (as it is used or ideologized) in processes of regularization? How do linguistic practices themselves become regularized? What factors inhibit or promote that process? Answering these questions place indeterminacy at the centre of analysis while engaging more directly with issues of power than Falk-Moore’s approach does. AFer outlining Falk-Moore’s framework, I explore its implications for socio-linguistic analysis with reference to my own work on Corsica, supplemented by examples from other scholars’ research.

2 Falk-Moore’s framework

In Law as Process (1978), Sally Falk-Moore approached social processes in terms of the interrelationship of three components: processes of regularization, processes of situational adjustment, and the factor of indeterminacy.

1) Processes of regularization

Processes of regularization include the use of rituals, rigid procedures, regular formalities, symbolic repetitions of all kinds, explicit laws, principles, rules, symbols and categories. Falk-Moore’s view was that these processes were attempts to ‘deny the passage of time, the nature of change, and the implicit extent of potential indeterminacy in social relations’ (41). Ideologies have a place in this perspective both as products of regularizing forces (51) and as rationalizations – not so much of the dominant social order, but of the

232 !"A4"#45=94$64A !69@47$

inherent orderliness of social life, as is evident in the following citation from the introduction to the book:

Ideologically, the notion that there are durable fundamental legal postu-lates may provide a sense of political continuity in the presence of many visible changes… De idea of core principles may help to preserve some of the aura of unquestionability and authority that surrounds the legal system and the legal profession without sacriGcing an iota of actual Jexibility. (11)

Reglementation, in Falk-Moore’s model, ‘emanates simultaneously from many social Gelds, including both corporate groups and less formally bounded action-arenas’ (29), and its control ‘can only be temporary, incomplete, and its consequences not fully predictable. De study of reglementation is therefore the study of the way partial orders and partial controls operate in social contexts’ (30).

2) Processes of situational adjustment

Processes of situational adjustment are implicated by the very nature of proc-esses of regularization. First of all, abstract principles (like ideologies, or kinship systems) can only be reproduced through social action, and this ‘instance-by-instance use permits the kind of reinterpretation, redeGnition and manipulation that is associated with processes of situational adjustment’ (51). Dat is, rules, norms and ideologies are not only partial, but also enacted/reenacted by various social actors with a range of interests and agendas that are not always congruent. Dese processes include acts of selection, interpretation, emphasis and perform-ance of rules, ideologies, symbols and so forth and ‘introduce or maintain the element of plasticity and social arrangements’ (50). Maintaining this element of ‘plasticity’ is an essential social resource; Falk-Moore writes that:

People… exploit the rules and indeterminacies as it suits their immediate purposes, sometimes using one resource, sometimes the other within a single situation, emphasizing the Gxity of norms for one purpose, exploit-ing openings, adjustments, reinterpretations and redeGnitions for another. (41)

A similar view of social actors’ interests in both regularization and situational adjustment is expressed by Michael Jackson, who writes that, ‘Lived experience encompasses both the rage for order and the impulse that drives us to unsettle or confound the Gxed order of things’ (1989:2). As mentioned above, Falk-Moore also notes that ‘strategies used in situational adjustment, if repeated suEciently oFen, by suEcient numbers of people, may become part of the processes of regularization’ (51).

45@7673B45CA> C5@ 37=9#C34HC64"5 233

3) De factor of indeterminacy

De factor of indeterminacy is thus present in processes of regularization and exploited in processes of situational adjustment. It has a number of sources. First, uncertainty and indeterminacy are the outcomes of processes over time, in which norms and meanings accumulate in a piecemeal fashion, are inherently multiple (and not necessarily coherent), and are anchored in diIerent (and only partially articulated) domains of practice with diIer-ent historical depth. Secondly, all social and ideological/symbolic systems generated by processes of regularization have zones of indeterminacy, both in their internal organization and in how they articulate with all of their possible applications to lived experience. Dus, Falk-Moore’s position that ‘it is useful to conceive an underlying, theoretically absolute cultural and social indeterminacy, which is only partially done away with by culture and organized social life’ (1978:49).

3 Indeterminacy and sociolinguistics

In this section, I will explore indeterminacy as a fundamental characteristic of sociolinguistic variables, contexts and speaker identities as well as its exploita-tion in processes of situational adjustment.

3.1 The indeterminacy of the sociolinguistic variable

3.1.1 Fuzzy boundaries and ambiguous forms

I would like to start with the fundamental indeterminacy of meaning of the sociolinguistic variable, whether it is an element of language structure, the enactment of a particular ‘style’, a type of discourse or a linguistic code. Some forms of ambiguity stem, as Podesva and Chun point out, ‘from competing ideological associations between linguistic forms and meanings across speakers or communities’ (2007:1). One element of this indeterminacy has to do with the permeable and potentially fuzzy boundaries between codes. Dis means that a given utterance may be formally ambiguous – and thus bi- or polyvalent (see Woolard, 1999, 2007). Social interaction may preserve ambiguity and multiplicity of meaning, or conversely, may exhibit regularizing tendencies and deGne an utterance as having one clear code/category membership. De analytical focus thus has to be on these processes: if, how, by whom (and with what sources of authority) the essential indeterminacy of formally ambiguous language is undercut.

234 !"A4"#45=94$64A !69@47$

3.1.2 Piecemeal accrual of meaning: change across time and multiple frames of reference

Not all sociolinguistic variables or utterances are formally ambiguous, but all of them have the potential to have multiple and layered meanings (what Woolard, 1999, characterizes as ‘simultaneity’). Dis is a natural consequence of the fact that the mapping of linguistic forms onto social meanings and vice versa takes place over time, sometimes – perhaps most oFen – in a piecemeal fashion. As Blommaert points out, this is because temporal developments related to language and society do not develop synchronically (2005:128). Meanings accrue to linguistic variables from diverse domains of practice; thus a given variable may have multiple and layered meanings that may be fundamentally ambiguous and may not necessarily be coherent. Put another way, sociolinguistic variables acquire meanings in fragmented, ‘partly con-nected, partly autonomous’ social Gelds (Falk-Moore, 1978:8). To take some examples from heavily travelled ground in sociolinguistics, vernacular codes operate as signiGers in two social and political Gelds: one in which they are stigmatized in opposition to codes with high status and power; and the other in which they are valued as forms of solidarity in an alternative marketplace deGned in part (but not uniquely) in opposition to the dominant marketplace. Using a ‘vernacular’ variable thus has at least two potential meanings: stigma and solidarity. Vernacular utterances are fundamentally indeterminate in that there is no way to exclude either of their potential meanings from an analysis; they are always both potentially present. Dis is because of the partial autonomy of the social Gelds from which they derive their meaning; many situations are neither exclusively one thing or another; as Urciuoli (1996) puts it, the distinction between the ‘inner’ and ‘outer’ spheres is seldom clear-cut. Dis foregrounds the indeterminacy of context as a sociolinguistic variable, a point to which I will return. At the same time, we can see processes of regularization at work: that is, every time people make use of linguistic oppositions to produce or index social ones, they reproduce those oppositions as absolutes.

Let me take these principles up with respect to a minority language con-text – Corsica – that I have studied extensively, where Corsican historically occupied the place of the stigmatized vernacular, and was associated with both low status and high solidarity. If we examine contemporary social processes, we Gnd that the association of Corsican with social intimacy – the ‘inner sphere’ – has been greatly complicated by language shiF. Older Corsicans who experienced all of the closeness and solidarity of family and friendship in Corsican still do use that language with their elderly intimates. In these conversations, the connection between Corsican and intimacy is strong, and

45@7673B45CA> C5@ 37=9#C34HC64"5 235

stands (at least partly) in opposition to French as a language of high power and low solidarity. However, these same older Corsicans have also all experi-enced relationships of intense intimacy with younger people (including their own children and grandchildren) through French. In a Grst instance, this might lead us to propose diIerent frameworks for analysis of the meaning of Corsican use in the interactions of elder native speakers based on the age (and linguistic competencies) of their interlocutors. We would view these speakers as having varied and fragmented sociolinguistic Gelds of social experience, and as themselves activating diIerent sociolinguistic frameworks for language production and interpretation depending on the situation. As mentioned above, this approach would also make visible the indeterminate or ambiguous nature of some contexts, for example those in which participants have mixed ages or levels of Corsican competence, and to address the way that linguistic choices might draw on the solidarity-distance opposition to create, not just index, existing solidarity or intimacy.

But we could also go further, and question our assignment of these elderly Corsican’s experiences (past and present; with older and younger interlocu-tors) to autonomous Gelds of experience and meaning, proposing instead that the gradual accrual of new meanings to the use of French across these people’s lifetimes has also subtly reconGgured the ‘old’ meanings of speaking Corsican. Speaking Corsican today, even with other elders, does not have the same meaning as speaking Corsican in an era where it was more or less unmarked, and spoken by everyone. Today it is not just opposed to French as a language of status, power and distance, but to French as a language of intimacy.

By the same token, the introduction of Corsican as a language of instruc-tion in school has not just raised the status of Corsican and leF the rest of the system of oppositions intact. De use of Corsican in education has also reconGgured the sociolinguistic landscape and people’s experiences in and through the language. A new generation of Corsican children is now accruing experiences of institutional life in Corsican, where it is both part of an adult regime to which they have no choice but to submit and where it is explicitly promoted as a language of personal identity and unique cultural heritage. Both of these experiences of Corsican are far removed from the experiences of their parents’ and grandparents’ generations. Adapting Corsican to a new domain of practice, the school, has also changed the code itself, and created new kinds of awareness of variation within Corsican, including the variation associated with ‘new’ linguistic forms whose authority is anchored in the institution vs. ‘old/traditional’ forms whose authority derives from ‘authentic’ elderly speakers.

236 !"A4"#45=94$64A !69@47$

De meaning of the sociolinguistic variable in this context – whether it is using Corsican or using a particular kind of Corsican – is thus fundamentally indeterminate in the sense that it simultaneously brings into play multiple social and historical Gelds of reference. Arguably, due to the social and politi-cal history that has shaped the Corsican sociolinguistic context, the tension between, in this case, power and solidarity is an enduring, and irresolvable feature of what it means to speak Corsican 1: in short, indeterminacy is built in to this kind of sociolinguistic context. One of the features of this context is, as I have indicated, a relatively rapid rate of change related both to processes of language shiF and to processes of language revitalization.

Such processes of change are widespread, although they vary in their scope and rate, and produce layered (and sometimes contested) meanings. For exam-ple, Bucholtz (2009) analyses the use of ‘güey’ in a Coors beer commercial. De advertisers deployed güey as a stance marker, ‘translating’ it humorously with a screen subtitle as ‘dude’ and showing it being used by young men projecting the ‘cool masculine solidarity’ that sells well in their target demographic. Responses to this advertising were mixed. Older, Mexican Spanish-speaking viewers associated güey (ostensibly derived from ‘ox’) only with its pejorative use as an insult. Bucholtz’s analysis of Chicano adolescents’ use of güey dem-onstrates that in this generation, while still used as an insult, it has accrued a wide range of additional social and interactional uses: this category of viewer seemed to both recognize themselves in and appreciate the advertising. Here, we can track multiple processes of change at work: Grst, we have the shiFing status of Spanish between generations (shared with the Corsican context): Spanish is the primary language for the older Mexican speakers but one among many linguistic (and stylistic) resources for Chicano youth. Secondly, the younger speakers live in social Gelds that at least loosely overlap with youthful ‘dude’ users: their use of güey as a discourse marker and interactional resource is likely conditioned by knowledge of ‘dude’ use that older Spanish speakers are very unlikely to share. Deir appreciation of the Coors commercials sug-gests that they must also, at some level, recognize masculine youth culture as a superordinate identity category that encompasses ethnic diIerentiation, which may contrast with the older Spanish-speaking viewers’ experiences of ethnicity as a form of primary diIerentiation that trumps other possible forms of social solidarity.

3.1.3 Multiple variables, mixed discourses

Sociolinguistic variables have an indeterminate quality with respect to the discourses in which they are embedded. To return to the ‘standard’-‘vernacular’ contrast, we know that many utterances and longer stretches of discourse

45@7673B45CA> C5@ 37=9#C34HC64"5 237

contain a mixture of forms subject to identi#cation as ‘standard’ or ‘vernacular,’ some of which will have long histories as sociolinguistic stereotypes. We can document the statistical frequencies of these two types of variables, both in the language practices of individual speakers, and across diIerent categories of speakers. However, the impact of the presence of vernacular vs. standard forms on how the speaker is perceived remains an empirical question. For example, Campbell-Kibler’s perceptual study of listener evaluations of the ING variable in matched guise tests show that despite its strong presence in the speech of Americans from diIerent regions, –in is taken as a core linguistic feature of a Southern accent which ‘dampens’ the perception of a ‘gay-sounding’ accent characterized (in part) by a fully articulated –ing (2007:55–56).

3.1.4 Indirect indexicality

Finally, a signiGcant current of sociolinguistic and linguistic anthropological work has emphasized the indirect way in which many linguistic signs in the speech event index social identities and categories (see for example Eckert, 2000; Eckert and Rickford, 2001; Hill, 1993, 2005; Ochs, 1990, 1991; Silverstein, 2003). De use of particular variables indexes stances or attitudes that are in turn mapped onto to social categories (for example, gender, class or ethnic-ity). Processes of indexicalization create multiple and overlapping mappings (Johnstone, Andrus and Danielson, 2006; Podesva and Chun, 2007; Moore and Podesva, 2009). As Silverstein points out, indexical associations at one level, or ‘order’, exist in interaction and even competition with higher order indexical levels (2003:194). Dese ‘indexical Gelds’ (Eckert, 2008) are by no means random; their elements are tied together with what Podesva calls a core association, or ‘kernel of similarity’ (2008). Dey are thus present – incipient – in speciGc moments of interaction, but those moments of interaction evoke and draw on diIerent elements from them in ways that relate to the inten-tions, positionality and purposes of interactants. De elements in that Geld (such as, to take Bucholtz’s example above, the use of güey) can index both cool male solidarity and aggression, both its historical pejorative meanings in Spanish-dominant communities and its new status as a resource in Chicano conversations; both being male and being Mexican and/or Chicano. Crucially, indirect indexicals do not make overt reference to, but rather presuppose con-nections between stances/attitudes and social or moral categories. Both the indirectness of the link between linguistic forms and identity categories and the complexity of processes of indexicalization contribute to the indeterminacy of the sociolinguistic variable.

238 !"A4"#45=94$64A !69@47$

3.1.5 The polyvocality and dialogicity of texts

As Scollon points out, Bahktin’s characterization of the inherent polyvocal-ity and dialogicity of texts ‘means that all texts are always intertextual and interdiscursive’ (1998:15). Dis has two major implications that relate to inde-terminacy. First, every text, whether spoken or written, has multiple potentials as what Scollon calls a ‘mediational means’ in social action. De analyst’s job, in Scollon’s framework, is to attend to ‘just those aspects of texts which are of relevance to actions taken by participants in any particular situation (1998:15). We could argue, however, that even when interactants orient towards only a subset of the possible meanings or uses of a text in a particular situation, its other potential meanings have a kind of virtual presence. To paraphrase the title of Judith Irvine’s (1996) publication, there are ‘shadow conversations’ between the aspects of those texts that are made relevant at a given moment and those that lie dormant. Secondly, as Scollon writes, the fact there ‘are no “pure” or “original” texts or genres… produces a level of indeterminacy into any analysis of the texts of mediated action. While it is theoretically given that the utterer is not the original producer of a text, at the same time… it is always ultimately undeterminable who or what discourse is the original voice we are hearing’ (1998:15).

3.2 Indeterminacy and the speaker

3.2.1 Stance, authorship and the inherent multiplicity and indeterminacy of participant roles

In the previous section, the focus was primarily on the sociolinguistic variable and secondarily, on context of speech. In this section, I focus on the speaker. Here, a fundamental source of indeterminacy has to do with the complex and potentially ambiguous relationship between the speaker and his or her words. De groundwork for this perspective was laid in GoIman’s seminal decomposition of the speaker role into discrete functions: author, animator and principle (1974:517, 1981:226). Since that time, scholars like Levinson (1988) have elaborated on this typology of participant roles and functions, showing how they can be fragmented and distributed across diIerent persons. Irvine has extended the theoretical reach of these concepts by focusing our attention on the processes of fragmentation themselves, and on the relationship between a given communicative act and other acts – past, future, hypothetical and so on – that make up the multiple and intersecting frames within which a particular utterance takes place (1996:134–5).

45@7673B45CA> C5@ 37=9#C34HC64"5 239

What this perspective postulates is that we cannot base our analyses on the notion of a uniGed speaker, who is at once the author (person responsible for the creation of a message), the principle (person responsible for its content/position) and the animator (person responsible for its transmission). Dese roles may sometimes be uniGed in a single speaker, but they are not necessar-ily so. As a consequence, our analytical focus must be drawn to the way that ‘participation structures are constructed, imagined and socially distributed’ (Irvine, 1996:136). In other words, we must posit participation structures – and the speaking self – as indeterminate, and investigate when (and if) they are represented or treated as uniGed, or relatively more Gxed/stable/unambiguous.

Dis lends support to the position taken in many lines of contemporary sociolinguistic analyses (particularly of style, stylization and stance) that speakers and their social identities cannot be treated as constants, subject to straightforward correlations with their uses of sociolinguistic variables, since it is possible for speakers to take up multiple positions, or stances, with respect to the form or content of their utterances. To take one well documented example, the robust literature on reported speech shows how the potential fragmentation of speaker functions can be strategically exploited by speakers who may align or disalign with the speech that they revoice, or even do both simultaneously (see Besnier, 1990, 1993; Irvine, 1989; Holt and CliF, 2007; Tannen, 2007). Dat is, by imposing a frame in which participant roles are destabilized, speakers can allude to multiple possible stances while fully committing to none. We might call this strategic indeterminacy, and I will return to this issue below.

3.2.2 The speaker as emergent across time

A second theoretical framework that supports a view of the speaker as ana-lytically indeterminate can be anchored in processual models that locate the speaking individual in the cumulative and oFen fragmented and partially articulated patterns of choice made in life projects and across life trajectories. Johnson-Hanks (2002) articulates this perspective on the person in her critique of the concept of ‘life stages’. Following Goody (1982) and Falk-Moore (1986), she argues that coherence in identities (like parenthood) and life stages (like adulthood) ‘should be an object, rather than an assumption, of ethnographic inquiry’ (865). Describing her own ethnographic study of the Beti in Southern Cameroon, she writes,

Rather than a clear trajectory toward adulthood, there are multiple, variable, and oFen hidden paths… the pace of trajectories varies not only between people but also within an individual. Dese piecemeal trajectories imply that people consist of multiple, partially independent parts – that the person is not uniGed but, rather, internally complex. (Johnson-Hanks, 2002:868)

240 !"A4"#45=94$64A !69@47$

Dis perspective has several sociolinguistic implications. First, sociolinguistic data gathered at a single point in time with a single speaker may or may not be consistent with the same individuals’ choices or practices at diIerent points of time and across diIerent situations. Dat is, there may be speaker-internal variability that does not correlate with the typical contextual triggers that may be evoked as explanations. In short, coherence of choices/practices for the speaking individual is possible, but is not necessary, which means that it cannot be assumed for either methodological or analytical purposes (including, for example, the practice of using ‘apparent time’ in some variationist sociolin-guistic work). Following Johnson-Hanks’ perspective on the ‘coordinated life transitions’ that she studied, we could say that it is the ‘degree and forms of accomplishment’ (Johnson-Hanks, 2002:869) of coherence of sociolinguistic choice that should be the object of inquiry and subsequently, grounds for the interpretation of discrete instances of individual linguistic practice.

It also follows that speakers’ choices in discrete speech events should ideally be interpreted with reference to their demonstrated repertoires and histories of usage. De exceptional use of a sociolinguistic variable for one person in a given instance has a diIerent meaning than the habitual use of the same variable for another person. For example, in a discussion of speaker stance (JaIe, 2009) I describe a speaker of my acquaintance who uses the informal ‘tu’ dramatically less oFen than any other speaker in her social circles. Dis exceptional pattern of use of a sociolinguistic variable diIerentiates the meaning of any given utterance of ‘tu’ by this woman from its use by other speakers. Dus her use of ‘tu’ may carry a more intense aIective stance of intimacy than the ‘tu’s of normative speakers. Alternatively, while normative speakers’ use of ‘tu’ with acquaintances would simply be read as cordial, when she does the same thing it constitutes a departure from her preferred usage, and thus may be seen as a more signiGcant act of social alignment with (or consideration for) interlocutors who desire a reciprocal ‘tu’ usage with her. In other cases, speakers demonstrate incoherent patterns of practice and low levels of control over their linguistic choices. In another analysis, I contrast the codeswitching practices of a Corsican bilingual teacher with another teacher educator. De former exhibited a high level of control over language choice, engaging in ‘pure’ monolingual practice in both Corsican and French across a range of contexts. De teacher educator, however, always codeswitched, even in contexts in which he claimed not to have alternated languages and in which he felt it was preferable not to do so. My argument is that because of these patterns of usage, there are far weaker analytical grounds for ascribing any intentionality to any given instance of codeswitch-ing by the teacher educator than there are in the case of the bilingual teacher (JaIe, 2007).

45@7673B45CA> C5@ 37=9#C34HC64"5 241

Tracking the histories of speaker choices and styles across the life trajectory is the hallmark of Barbara Johnstone’s work (1996 and 2009) on Barbara Jordan, a well-known African-American politician. Johnstone characterizes Jordan’s ‘linguistic identity’ as the cumulative patterning of Jordan’s socio-linguistic, stylistic and stance choices over time. Dis patterning reveals a commitment to consistency of style and register across widely divergent contexts which might be reasonably expected to trigger diIerent kinds of speech. Jordan, in other words, deliberately regiments her own practice so as to read out the inherent indeterminacy of everyday practice in order to project an ‘ethos of self ’ in which language is to be seen as the reJection of a durable, stable elements of personal character rather than as a mere response to social contingencies.

Wortham also focuses on the indeterminacy of particular speech events with respect to speaker identity, which he treats as emergent in social interaction. He argues both that ‘meaning is oFen indeterminate until subsequent utterances contextualize focal signs as having meant something robustly presupposable’ and that there is an inherent ‘indeterminacy of contextualization that inhibits the clear interpretation of signs at the moment of utterance’ (2005). His analy-ses focus both on the cumulative patterns of speaker choices across multiple discursive interactions, and on the cumulative results of how speaker talk is received, evaluated and framed by other social actors for how those speakers are socially and morally positioned. By paying attention to both the agency of the speaker and the agency of his/her interlocutors over time (and in speciGc institutional contexts), Wortham demonstrates the inherent indeterminacy of particular utterances or moments of discourse as well as how social processes come to solidify their meanings. Sharing Johnson-Hanks’ emphasis on the multiplicity of possible life trajectories and on the creation of coherence as a project, not a feature of those histories, he writes that:

Accounts of academic socialization that focus only on recurrent events cannot explain how the potential indeterminacy of socialization – the fact that an individual could move along various trajectories, toward being diIerent types of person – gets overcome as coherent trajectories of socialization emerge across events. (2005:99)

Not only is the meaning of particular utterances dependent on their social uptake in the moment, but they are also subject to retroactive revision and reinterpretation: their meanings are not inherently stable, but rather, subject to stabilization.

242 !"A4"#45=94$64A !69@47$

3.3 Language practice and strategic indeterminacy

Here we return to the ‘impulse that drives us to unsettle or confound the Gxed order of things’ (Jackson, 1989:2) to explore the social uses of indeterminacy of language.

3.3.1 The social power of indirection

De brief discussion of indirect indexicality introduced the notion of presupposi-tion. When interactional and social meanings are embedded in presuppositions of talk, interlocutors are implicated through the very process of interpretation. Dis implication of the interlocutor can be one of the social goals that makes indeterminacy strategic. In a classic article on metaphor in Apache culture, Basso (1976) shows how elders with the authority to criticize younger people oFen choose instead to comment on the actions of others indirectly, through metaphors. In doing so, by obliging the recipients of these metaphors to decode them (and apply them to their own lives) they make the targets of their critique complicit in the negative framing of their behaviour (see also Shoaps, 2007).

3.3.2 As resistance

In other cases, speakers may exploit indeterminacy in language use as a way of resisting processes of regularization, reglementation and categorization, using their agency to suspend deGnition when being clearly deGned creates dissonance, personal or interactional conJict, discomfort or disadvantage. Because multiple social and linguistic positions, identities and stances are relevant or useful for social actors, they can have an interest in exploiting the fundamental indeterminacy or multivalency of language use to maintain Jex-ibility of self-presentation in potentially unpredictable or volatile social Gelds of reception and interpretation.

Stroud (1992) takes this perspective in his analysis of codeswitching in a public appeal by one man (Kem) for social and Gnancial support for an impor-tant project. Stroud takes the position that the use of Tok Pisin and Taiap in this discourse can be linked in a general way with contrasting sources of social and moral authority, and that in using both languages, Kem invokes both social obligations associated with kinship (through Taiap) and obligations associated with Christianity (through Tok Pisin). However, he argues that below this very general level, most of Kem’s switches elude correlational analysis, and can be viewed as strategies aimed at maintaining multiple and indeterminate alliances. Indeterminacy works in Kem’s favour because it allows his various interlocutors to ‘hear’ (through presupposition) whichever moral framework they Gnd most persuasive and to ignore the other. In a very diIerent context,

45@7673B45CA> C5@ 37=9#C34HC64"5 243

Lo (1999) explores how, in a multiethnic peer group, speakers use ‘outgroup talk’ (in the heritage languages of their interlocutors) to be deliberately opaque in interactions in which there is some tension about who controls the language of the interaction as well as who has legitimate claims to various identities.

Both Woolard (1999) and JaIe (2000) have explored bilingual media comedy performances in minority language contexts where language choice has been heavily and overtly implicated in identity politics. In both cases, the comedians exploit bivalent linguistic forms and codeswitching practices to undermine the ‘either-or’ imperatives of sociolinguistic identity politics in Catalonia and Corsica, and to present personas who elude clear classiGcation. De Corsican comedian also exploits the indeterminacy of reported (or performed) speech in his performance of an elderly bilingual speaker with some limitations in her command of French to mediate the conJict many Corsicans experience between models of ‘good French’ and associated fears of language mixing and their more recent desires to revitalize Corsican and celebrate older speakers. In these cases, identity conJict and tension at the societal level are expressed through deliberately indeterminate linguistic forms and practices.

In other cases, speakers themselves are fundamentally conJicted, and stance multiplicity and indeterminacy expressively mediates that conJict. Dis is apparent in McIntosh’s (2009) analysis of the narratives of White Kenyans whose identities as whites are constructed in contrast with ‘irrational’ Black African belief systems which nevertheless permeate their experiences and social practices as Kenyans. As a consequence, taking a position with respect to witchcraF and the occult causes existential conJict. In response to this, McIntosh’s interviewees introduce multiple ‘I’s into their accounts, privileging the ‘I’s that can’t believe while simultaneously speaking from the ‘I’s who have had persuasive encounters with the occult.

Speakers may also have vested interests in keeping issues of agency and responsibility opaque, and thus, in not committing to full authorship of the words they utter, as I have suggested in my mention of reported speech above. Indeterminacy created by the layering of participant roles allows speakers the power to suggest without the risks of assertion; and with the beneGts of aFer-the-fact denials or reconstruals of their meanings and stances. Rampton (1995) shows this process at work in British adolescents’ ‘crossing’ into Panjabi and SAE (Stylized Asian English) in interactions with their teachers. Crossing itself represents an indeterminate relationship between speaker and code, since it is by deGnition the use of a code to which a speaker who cannot claim full ownership rights. Dus crossing makes it clear that the speaker is appropriating another voice, and brings that other voice into the interaction. When that voice is a socially stigmatized one (as is the case for Panjabi and SAE), its introduc-tion into an adolescent-adult interaction indirectly brings issues of racism

244 !"A4"#45=94$64A !69@47$

and discrimination into play: adults who name and sanction the language varieties being used can be accused of bias. Students, meanwhile, cannot be held responsible for any discrete propositional content. Crossing thus emerges as a resource students sometimes use to resist adult authority, or to engage in rekeying of situations to suit their own ends.

Conversely, role fragmentation in the utterance can be a tool through which speakers or writers tap in to a range of collective socially sanctioned attitudes, positions and meanings associated with diIerent forms and genres of speech. Dat is, individuals can beneGt from animating, consuming or commissioning words (and thus being connected to them in socially sig-niGcant ways) while simultaneously being released from the exigencies and consequences of individual authorship (see for example Irvine’s 1989 analysis of Wolof nobles’ relationship with griots and JaIe’s 1999b study of the social uses of greeting cards).

De slippage between the personal and the collective with respect to author-ship can also be exploited for other purposes, and may be oriented towards integration (rather than maintenance of fragmentation) of participant roles. In my analyses of Corsican classroom interactions, for example, I have docu-mented how teachers create the conditions in which less competent speakers of Corsican are discursively integrated into an expert community of practice through practices like repetition, teacher revoicing and chorusing. In doing so, teachers symbolically transfer both some of their own linguistic expertise to the (novice) group, and the collective competencies of the group to individual speakers (JaIe, 2003b). Dese practices respond to the potential tension, at the level of identity, of low competence in a language of cultural heritage.

Finally, an emphasis on indeterminacy can be a form of resistance to domi-nant language ideologies and to domination itself. Dis is the crux of Samuels’ (2001) analysis of the San Carlos Apaches, who embrace indeterminacy as a cornerstone of an aesthetic ideology ‘highlights the ambiguities of translation and code switching’. He argues that in doing so, ‘San Carlos Apaches engage with domination by explicitly appropriating and transforming its naturalized signs into expressions and claims of a shared San Carlos Apache history and connection to the past. San Carlos Apaches make sense of their world by playing parodic tricks with the symbols of domination’ (278). In doing so, they reveal concepts of closure as ideological positions rather than as essential qualities of language use. Dis ‘refusal of resolution is also a refusal of domination, of assimilation’ (294).

45@7673B45CA> C5@ 37=9#C34HC64"5 245

4 Processes of regularization: fixing language and its meanings

De emphasis, in this essay, has been on indeterminacy, but let me make slightly more abbreviated reference to the issue of regularization – both nor-mative and prescriptive – with which a great deal of sociolinguistic analysis has been concerned. De advantage of approaching regularization through the lens of a primary indeterminacy is that it prevents us from taking that regularization for granted, and prompts us to examine in some detail the inter-actional, institutional, social, cultural and political processes through which language (and its relationships to identity categories and social hierarchies) is regularized and regimented. Dis is in keeping with a number of analytic approaches in contemporary sociolinguistics and linguistic anthropology. Central to these approaches is a concern with two simultaneous processes: the creation of sameness/equivalence (‘adequation’ in Bucholtz and Hall’s 2004 terms) and the construction of linguistic and social diIerence. As Irvine and Gal (2000) and Irvine (2005) underscore, both similarity and diIerence are culturally and historically inJected, and the ‘construal of likeness’ is always built upon ‘a construal of relevant oppositions and their scope (2005:76). De connections between linguistic forms/codes and social groups are also subject to iconization (in Irvine and Gal’s 2000 framework): that is, they are treated as essential rather than historical and contingent.

A focus on how sameness and diIerence are created takes us in a number of directions. Returning to Ochs’ insights about the mediating role played by stance in creating indexical relationships between linguistic forms and social identity categories, we can investigate how particular stances become associated with some identities over others (since many stances, like ‘deference’ are used by many people in multiple contexts), and seek answers both in statistical patterns of practice (who takes what stances and how oFen?) as well as in processes of stereotypiGcation in which social ideologies shape which linguistic traits are taken to ‘typify’ certain groups (see also Moore and Podesva, 2009).

Relationships of similarity and diIerence are also constructed across texts, speech events, discourses and semiotic modes. A signiGcant body of contempo-rary research examines these linkages (and the strength of their associations). Dis includes work on entextualization and reentextualization (Bauman and Briggs, 1990; Silverstein and Urban, 1996); intertextuality (Bauman, 2004; Fairclough, 1992; Hill, 2005) interdiscursivity/discursive re-contextualization (Silverstein, 2005; Van Leeuwen, 2008; Wortham, 2005); resemiotization (Iedema, 2003), mediation (Scollon, 1998) and enregisterment (Agha, 2005). In addition to emphasizing the way that sociolinguistic data is (potentially)

246 !"A4"#45=94$64A !69@47$

transformed each time it is reinscribed, revoiced or otherwise reproduced, these approaches identify entextualization as an ideological process and an exercise of power, as it is deGned by speciGc institutional and social frameworks. Much of this work also incorporates an attention to agency in both the act of production (of text or discourse) and in acts of uptake, interpretation, appropriation and so forth, thus emphasizing the social, dialogic and dynamic nature of these processes.

Sociolinguistics and linguistic anthropology have also been concerned with processes through which dominant language ideologies that posit languages and codes as bounded entities that have Gxed and essential relationships with social categories are produced, reproduced and potentially challenged (JaIe, 1999a; Irvine and Gal, 2000). Along with some of the critical dis-course perspective mentioned above, this approach shares an attention to the power of indeterminacy and second order indexicality as it is coupled with the exercise of power. In this respect, language (at least within a Western political and philosophical tradition) appears to have a unique status as a tool of diIerentiation and adequation. On the one hand, it subject to being represented as closed, bounded, autonomous and rule-governed; on the other it is signiGcantly manipulable in practice. Dis is what Falk-Moore terms ‘the denial of indeterminacy’; this works through processes of the ‘erasure’ (Irvine and Gal, 2000) of practice in the social representations of language. De result is that language can be used by those who have (or wish to claim) power to prescribe while purporting to describe; to Gx and deGne others while retaining all the advantages of indeterminacy and all the manoeuvre room of processes of situational adjustment. At the same time, work on language ideologies shows us that we cannot just assume that dominant language ideologies exercise a seamless hegemony, or that their ‘recognition’ or ‘mis-recognition’ (Bourdieu, 1990) are ever complete. For example, my analysis of the instantiation of a pluralist ideology of linguistic and cultural identity in Corsican schools demonstrates both the cracks in the dominant ideological armour (the exposure of ‘misrecognition’) and the complexities and limits associated with recognizing diversity in institutional contexts (JaIe, 2003a). Ideologies of language, as Woolard (1985) and Gal (1998) pointed out early on, are also oFen multiple and contested. Following Falk-Moore, they are instantiated in Gelds of social action that are partially, not fully articulated or coherent. De same indeterminacy and multiplicity that operates at the level of participant roles and structures can also be found in social actors’ engagement with and participation in dominant language ideologies and sociolinguistic hierarchies. People’s states of belief in or adherence cannot be assumed to be consistent across diIerent domains, social Gelds and experi-ences, and so they may simultaneously adhere to and reproduce dominant

45@7673B45CA> C5@ 37=9#C34HC64"5 247

language ideologies in one domain and challenge, undermine and resist them (to varying degrees) in another. As a consequence, the same individual can participate in both the regularizing process and its contestation; at any given moment in time, even relatively uniGed sociolinguistic Gelds have these patches of indeterminacy.

5 Conclusions

In this essay, I have argued that a view of social life as inherently indeter-minate, subject to regularization (and reglementation) as well as processes of situational adjustment, is a productive framework for sociolinguistics. Indeterminacy emerges as a fundamental principle of sociolinguistic vari-ables, context, speaker identity and participation structures, as a resource for speakers and as a possible objective of communicative practice. It stands in constant tension with sociolinguistic regularization, as it is instantiated across texts, time and discourses and as it is embedded in particular social and politi-cal Gelds. Studying this tension is how a contemporary sociolinguistics can approach a coherent account of agency and constraint, change and continuity, reproduction and contestation of normative practices and dominant language ideologies. Sociolinguistic analyses, furthermore, examine language practices that are crucial resources for both processes of regularization and situational adjustment, and highlight the inherent multivocality of all semiotic systems implicated in the conduct of social life.

About the author

A. M. JaIe is Professor in the Department of Linguistics and Anthropology at California State University Long Beach.

Notes

1 See Falk-Moore (1978:26–27) for a similar point about tensions between hierarchy and equality/equivalence in organizations.

References

Agha, A. (2005) Voice, footing, enregisterment. Journal of Linguistic Anthropology 15(1): 38–59.

Barth, F. (1981) Process and Form in Social Life. London: Routledge and Kegan Paul.

248 !"A4"#45=94$64A !69@47$

Basso, K. (1976) ‘Wise words’ of the Western Apache: Metaphor and semantic theory. In K. Basso and H. Selby (eds) Meaning in Anthropology 93–121. Albuquerque: University of New Mexico Press.

Bauman, R. (2004) A World of Other’s Words. Malden MA: Blackwell.

Bauman, R. and Briggs, C. (1990) Poetics and performance as critical perspectives on language and social life. Annual Review of Anthropology 19: 59–88.

Besnier, N. (1990) Language and aIect. Annual Review of Anthropology 19: 419–451.

Besnier, N. (1993) Reported speech and aIect on Nukulaelae Atoll. In J. Hill and J. Irvine (eds) Responsibility and Evidence in Oral Discourse 161–181. Cambridge: Cambridge University Press.

Blommaert, J. (2005) Discourse: A critical introduction. Cambridge: Cambridge University Press.

Bourdieu, P. (1977) Outline of a "eory of Practice. Cambridge: Cambridge University Press.

Bourdieu, P. (1990) "e Logic of Practice. (Translated by R Nice.) Cambridge: Polity Press.

Bucholtz. M. (2009) From stance to style: Gender, interaction, and indexicality in Mexican immigrant youth slang. In A. JaIe (ed.) Sociolinguistic Perspectives on Stance 146–170. Oxford: Oxford University Press.

Bucholtz, M. and Hall, K. (2004) Language and identity. In A. Duranti (ed.) A Companion to Linguistic Anthropology 268–294. Malden, MA: Blackwell.

Campbell-Kibler, K. (2007) Accent, (ING), and the social logic of listener perceptions. American Speech 82: 32–64.

Coupland, N. (1998) What is sociolinguistic theory? Journal of Sociolinguistics 2(1): 110–117.

Coupland, N. (2001) Introduction: Sociolinguistic theory and social theory. In N. Coupland, S. Sarangi and C. Candlin (eds) Sociolinguistics and Social "eory 1–29. London: Longman.

Eckert, P. (2000) Linguistic Variation as Social Practice. Oxford: Blackwell.

Eckert, P. (2008) Variation and the indexical Geld. Journal of Sociolinguistics 12(4): 453–476.

Eckert, P. and Rickford, J. (eds) (2001) Style and Sociolinguistic Variation. New York and Cambridge: Cambridge University Press.

Fairclough, N. (1992) Discourse and Social Change. Cambridge: Polity Press.

Falk-Moore, S. (1978) Law as Process. Boston: Routledge and Kegan Paul.

Falk-Moore, S. (1986) Social Facts and Fabrications: ‘Customary’ law in Kilimanjaro, 1880–1980. Cambridge: Cambridge University Press.

Gal, S. (1998) Multiplicity and contestation among linguistic ideologies. In K. Woolard and B. SchieIelin (eds) Language Ideologies: Practice and theory 317–331. Oxford: Oxford University Press.

45@7673B45CA> C5@ 37=9#C34HC64"5 249

Giddens, A. (1984) "e Constitution of Society: Outline of the theory of structuration. Berkeley, University of California Press.

GoIman, E. (1974) Frame Analysis: An essay on the organization of experience. New York: Harper and Row.

GoIman, E. (1981) Forms of Talk. Philadelphia, PA: University of Pennsylvania Press.

Goody, E. (1982) Parenthood and Social Reproduction: Fostering and occupational roles in West Africa. Cambridge: Cambridge University Press.

Heller, M. (2002) Eléments d’une sociolinguistique critique. Paris: Didier.

Heller, M. (2007) Bilingualism as ideology and practice. In M. Heller (ed.) Bilingualism: A social approach 1–17. London: Palgrave.

Hill, J. (1993) ‘Hasta la vista, baby’: Anglo Spanish in the U.S. Southwest. Critique of Anthropology 13: 145–176.

Hill, J. (2005) Intertextuality as source and evidence for indirect indexical meanings. Journal of Linguistic Anthropology 15(1): 113–124.

Holt, R. and CliF, R. (eds) (2007) Reporting Talk: Reported speech in interaction. Cambridge: Cambridge University Press.

Iedema, R. (2003) Multimodality, resemioticization: Extending the analysis of discourse as a multisemiotic practice. Visual Communication 2: 29–57.

Irvine, J. T. (1989) When talk isn’t cheap: Language and political economy. American Ethnologist 16(2): 248–267.

Irvine, J. T. (1996) Shadow conversations: De indeterminacy of participant roles. In M. Silverstein and G. Urban (eds) Natural Histories of Discourse 131–159. Chicago: Chicago University Press.

Irvine, J. T. (2005) Commentary: Knots and tears in the interdiscursive fabric. Journal of Linguistic Anthropology 15(1): 72–80.

Irvine, J. T. and Gal, S. (2000) Language ideology and linguistic diIerentiation. In P. Kroskrity (ed.) Regimes of Language 35–83. Santa Fe, NM: School of American Research Press.

Jackson, M. (1989) Paths Toward a Clearing. Bloomington: Indiana University Press.

JaIe, A. (1999a) Ideologies in Action: Language politics on Corsica. Berlin: Mouton de Gruyter.

JaIe, A. (1999b) Packaged sentiments: De social meaning of greeting cards. Journal of Material Culture 4(2): 115–141.

JaIe, A. (2000) Comic performance and the articulation of hybrid identity. Pragmatics 10(1): 39–60.

JaIe, A. (2003a) Misrecognition unmasked? ‘Polynomic’ language, expert statuses and orthographic practices in Corsican schools. Pragmatics 13(3/4): 515–538.

JaIe, A. (2003b) ‘Imagined competence’: Classroom evaluation, collective identity and linguistic authenticity in a Corsican bilingual classroom. In S. Wortham and B. Rymes (eds) Linguistic Anthropology of Education 151–184. Westport CT: Praeger.

250 !"A4"#45=94$64A !69@47$

JaIe, A. (2007) Codeswitching and stance: Issues in interpretation. Journal of Language, Identity and Education 6(1): 1–25.

JaIe, A. (2009) Introduction: A sociolinguistics of stance. In A. JaIe (ed.) Sociolinguistic Perspectives on Stance 3–28. Oxford: Oxford University Press.

Johnson-Hanks, J. (2002) On the limits of life stages in ethnography: Toward a theory of vital conjunctures. American Anthropologist 104(3): 865–880.

Johnstone, B. (1996) "e Linguistic Individual: Self-expression in language and linguistics. Oxford: Oxford University Press.

Johnstone, B. (2009) Stance, style and the linguistic individual. In A. JaIe (ed.) Sociolinguistic Perspectives on Stance 29–52. Oxford: Oxford University Press.

Johnstone, B., Andrus, J. and Danielson, A. E. (2006) Mobility, indexicality, and the enreg-isterment of ‘Pittsburghese.’ Journal of English Linguistics 34: 77–104.

Levinson, S. (1988) Putting linguistics on a proper footing: Explorations in GoIman’s concepts of participation. In P. Drew and A. Wooton (eds) Erving Go!man: An interdis-ciplinary appreciation 161–227. Oxford: Polity Press.

Lo, A. (1999) Codeswitching, speech community membership, and the construction of ethnic identity. Journal of Sociolinguistics 3(4): 461–479.

McIntosh, J. (2009) Stance and distance: Social boundaries, self-lamination and metalin-guistic anxiety in White Kenyan narratives about the African Occult. In A. JaIe (ed.) Sociolinguistic Perspectives on Stance 72–91. Oxford: Oxford University Press.

Moore, E. and Podesva, R. (2009) ‘We do say, “in’ he”, don’t we?’ Style, indexicality, and the social meaning of tag questions. Unpublished ms.

Ochs, E. (1990) Indexicality and socialization. In J. Stigler, R. Schweder and G. Herdt (eds) Cultural Psychology 287–308. Cambridge: Cambridge University Press.

Ochs, E. (1991) Indexing gender. In A. Duranti and C. Goodwin (eds) Rethinking Context 335–358. Cambridge: Cambridge University Press.

Podesva, R. (2008) Dree sources of stylistic meaning. Texas Linguistic Forum 51: 1–14.

Podesva, R. and Chun, E. (2007) On Indeterminacy in the social meaning of variation. Paper presented at NWAV 2007.

Rampton, B. (1995) Crossing: Language and ethnicity among adolescents. London and New York: Longman.

Samuels, D. (2001) Indeterminacy and history in Britton Goode’s Western Apache placenames: Ambiguous identity on the San Carlos Apache reservation. American Ethnologist 28(2): 277–302.

Sarangi, S. (2001) A comparative perspective on social theoretical accounts. In N. Coupland, S. Sarangi and C. Candlin (eds) Sociolinguistics and Social "eory 30–59. London: Longman.

Scollon, R. (1998) Mediated Discourse as Social Interaction. New York: Longman.

Shoaps, R. (2007) Moral irony: Modal particles, moral persons and indirect stance-taking in Sakapultek discourse. Pragmatics 17(2): 297–335.

45@7673B45CA> C5@ 37=9#C34HC64"5 251

Silverstein, M. (2003) Indexical order and the dialectics of sociolinguistic life. Language & Communication 23: 193–229.

Silverstein, M. (2005) Axes of evals: Token vs. type interdiscursivity. Journal of Linguistic Anthropology 15(1): 6–22.

Silverstein, M. and Urban, G. (1996) De natural history of discourse. In M. Silverstein and G. Urban (eds) Natural Histories of Discourse 21–44. Chicago: Chicago University Press.

Stroud, C. (1992) De problem of intention and meaning in codeswitching. Text 12(1): 127–155.

Tannen, D. (2007) Talking Voices. Cambridge: Cambridge University Press.

Urciuoli, B. (1996) Exposing Prejudice. Boulder CO: Westview Press.

Van Leeuwen, T. (2008) Discourse and Practice: New tools for critical discourse analysis. Oxford: Oxford University Press.

Woolard, K. (1985) Language variation and cultural hegemony. American Ethnologist 12: 738–748.

Woolard, K. (1999) Simultaneity and bivalency as strategies in bilingualism. Journal of Linguistic Anthropology 8(1): 3–29.

Woolard, K. (2007) Strategic bivalency in Latin and Spanish in early modern Spain Language in Society 36: 487–509.

Wortham, S. (2005) Socialization beyond the speech event. Journal of Linguistic Anthropology 15(1): 95–112.