Background
Following Macaulay’s (Reference Macaulay, Chambers, Trudgill and Schilling-Estes2002, Reference Macaulay2005) repeated calls for a fuller integration of discourse-pragmatic features into the variationist research agenda, the last decade or so has seen an unprecedented upsurge in quantitative research investigating patterns of variation and change in the use of conventionalised, polyfunctional linguistic items and constructions such as innit, you know, as I say, and stuff like that, you get what I mean, at the end of the day. The growing scholarly fascination with these features is reflected in an expanding number of journal articles exploring their variability and trajectories of change (see later pages for selected references), the recent and forthcoming publication of several book-length treatments investigating their structured heterogeneity in synchronic dialect data (see, for example, Aijmer Reference Aijmer2013; Andersen Reference Andersen2001; Buchstaller Reference Buchstaller2014; D’Arcy in prep.; Macaulay Reference Macaulay2005; Pichler Reference Pichler2013; Tagliamonte Reference Tagliamonte2016) as well as a flourishing international conference series dedicated specifically to their quantitative analysis (www.dipvac.org). In her analysis of the state of the field, Pichler (Reference Pichler2013: 6–9) has attributed the international surge in variationist studies of discourse-pragmatic features to two factors: their recent theoretical reconceptualisation as integral elements of the linguistic system (see, inter alia, Brinton Reference Brinton, van Kemenade and Los2006; Diewald Reference Diewald and Fischer2006; Kaltenböck et al. Reference Kaltenböck, Heine and Kuteva2011; Traugott Reference Traugott, Joseph and Janda2003; and contributions in Degand and Simon-Vandenbergen Reference Degand and Simon-Vandenbergen2011); and recent methodological advancements in accountably quantifying their variability (see, inter alia, Buchstaller Reference Buchstaller2009, Reference Buchstaller2011; Cheshire Reference Cheshire2007; D’Arcy Reference D’Arcy2005; Pichler Reference Pichler2010, Reference Pichler2013; Pichler and Levey Reference Pichler and Levey2011; Wagner et al. Reference Wagner, Hesson, Bybel and Little2015).
Because of the myriad complexities involved in defining discourse-pragmatic features as linguistic variables (see Buchstaller [Reference Buchstaller2009]; Pichler [Reference Pichler2010, Reference Pichler2013: Ch. 2]; Tagliamonte [Reference Tagliamonte2012: Ch. 9] for details), early variationist discourse studies have tended to focus on individual items or constructions (e.g., don’t you think, eh, I mean, I think, presumably, you know, you see) and compared their frequency and functionality across social groups, without acknowledging that these features are embedded within a broader variable system (see, for example, Erman Reference Erman1992; Holmes Reference Holmes1986, Reference Holmes1990, Reference Holmes1995; Macaulay Reference Macaulay1995; Stubbe and Holmes Reference Stubbe and Holmes1995; Woods Reference Woods and Cheshire1991; but see Cheshire Reference Cheshire1981; Dines Reference Dines1980; Ferrara Reference Ferrara1997). Nonetheless, by demonstrating that the use of discourse-pragmatic features – like that of features at other levels of the linguistic system – is variable, changeable and socially indexical, these studies have laid the groundwork for subsequent work which has developed methods for quantifying discourse-pragmatic variation in maximally accountable ways. These more recent studies have generally analysed individual discourse-pragmatic features in the context of their functionally-, positionally- and/or derivationally-equivalent co-variants, and extended the focus of analysis to the intra-linguistic conditioning of discourse-pragmatic variation and change (see, inter alia, Andersen Reference Andersen2001; Buchstaller Reference Buchstaller2006a, Reference Buchstaller2008, Reference Buchstaller2014; Buchstaller and D’Arcy Reference Buchstaller and D’Arcy2009; Cheshire Reference Cheshire2007; D’Arcy Reference D’Arcy2005, Reference D’Arcy2007; Denis Reference Denis2011, Reference Denis2015; Diskin Reference Diskin2015; Drager Reference Drager2010, Reference Drager2011; Fox Reference Fox, Buchstaller and van Alphen2012; Ito and Tagliamonte Reference Ito and Tagliamonte2003; Levey Reference Levey2006a; Levey et al. Reference Levey, Groulx and Roy2013; Macaulay Reference Macaulay2001, Reference Macaulay2006; Pichler Reference Pichler2009, Reference Pichler2013; Pichler and Levey Reference Pichler and Levey2011; Rodríguez Louro Reference Rodríguez Louro2013; Tagliamonte and D’Arcy Reference Tagliamonte and D’Arcy2004, Reference Tagliamonte and D’Arcy2009; Tagliamonte and Denis Reference Tagliamonte and Denis2010; Waters Reference Waters2013). The methodological advancements have inspired an increasing number of studies that provide important insights into the sociolinguistic mechanisms underlying discourse-pragmatic variation and change, and thus make important contributions to current theories of language variation and change. Yet despite having ‘progressed substantially in recent years’ (Macaulay Reference Macaulay, Chambers and Schilling2013: 230) and having contributed important theoretical insights, quantitative variationist research of discourse-pragmatic features continues to be dwarfed by studies of phonological and morpho-syntactic variables in terms of number, scope and impact.
This volume brings together a group of scholars widely recognised for their contributions to discourse-pragmatic variation and change research, with the purpose of stimulating the vitality and growth of this line of research. To meet this objective, the volume assembles an authoritative and original collection of articles which: (i) introduce a range of contrasting yet complementary new methods specifically tailored to the requirements of studying discourse-pragmatic variation and change; and (ii) provide new empirical and theoretical insights into the sociolinguistic dimensions of discourse-pragmatic variation and change in contemporary varieties of English. With its dual focus on presenting innovative methods as well as new results, the volume will provide an important resource for both newcomers and veterans in the field of discourse variation analysis alike, and spark discussions that will set new directions for future work in the field.
In the remainder of this introduction, I will first set out the terminology and define the scope of the volume (‘Terminology and scope’). I will then outline in more detail the overarching aims of the volume (‘Overarching aims’). Following a brief description of the general structure of the volume (‘Organisation’), I will offer a brief overview of individual contributions (‘Overview’) before sketching out the general themes to emerge from the collection of contributions (‘Implications’). In doing so, I will evaluate the strengths of the approaches advocated by the volume contributors, and indicate why and where they should be reapplied. I will also discuss how the findings presented in this volume refine current models of discourse-pragmatic variation and change, and what implications this may have for future research (design).
Terminology and scope
At the outset, it is important to clarify the contributors’ use of terminology and delimit the empirical scope of the volume. The linguistic features examined in the following chapters include a heterogeneous category of items and constructions such as so, in he, as I say, and what have you, you get what I mean which share neither a common set of formal linguistic properties nor an agreed upon macro-label (see, inter alia, Brinton [Reference Brinton1996: Ch. 2, Reference Brinton2008: Ch. 1]; Fischer [Reference Fischer and Fischer2006b]; Jucker and Ziv [Reference Jucker, Ziv, Jucker and Ziv1998]; Schourup [Reference Schourup1999] for overviews of relevant debates). Moreover, they perform vastly different micro-functions such as modifying propositions (so), seeking corroboration of propositions (in he), marking continuation (as I say), extending sets (and what have you), seeking hearer involvement (you get what I mean) etc. Many researchers indiscriminately label such features ‘discourse markers’ (DMs), ‘pragmatic markers’ (PMs) or ‘discourse particles’ (DPs). Others, by contrast, make a typological distinction among DMs which perform a structural role, PMs which signal speaker stance, and DPs which have scalar or modal meaning (Fraser Reference Fraser1990; Schourup Reference Schourup1999; but see also Blakemore Reference Blakemore1987; Schiffrin Reference Schiffrin1987; and contributions in Abraham Reference Abraham1991). Because these conceptual distinctions are not consistently applied in the literature and because individual features can function across the macro-functions ascribed to DMs, PMs and DPs (see, for example, Kärkkäinen [Reference Kärkkäinen2003] on I think), this volume has adopted the alternative, conceptually more neutral label ‘discourse-pragmatic features’ to refer to the category of linguistic items and constructions studied by its contributors. What unites this super-category of formally, functionally and syntactically heterogeneous features is that (i) they perform a range of interpersonal and/or textual functions in discourse; (ii) their use is motivated first and foremost by their functionality. (For a more detailed definition of the category of ‘discourse-pragmatic features’, see, for example, Pichler [Reference Pichler2013: 4–6].) Based on these properties, the contributors conceptualise the features studied in this volume as discourse-pragmatic variables rather than lexical or morpho-syntactic variables, two alternative classifications applied, for example, to quotatives (see Buchstaller [Reference Buchstaller2014: 251–2] for details). In addition to the macro-label ‘discourse-pragmatic features’, the contributors employ micro-labels such as ‘quotatives’, ‘utterance-final tags’ or ‘discourse like’ to refer to the specific (set of) item(s) or construction(s) that they analyse and discuss. (The micro-labels used are generally a telling indicator of how the variable (context) or scope of analysis was defined, for example, in terms of function, position or form.) When contributors do adopt the labels DM, PM or DP, they do so in full knowledge of their underlying conceptual bases.
The discourse-pragmatic variables examined in the volume include both under-explored and well-researched variables: interjections (e.g., So duh), vocatives (e.g., Trust me, bruv), text-organising features (e.g., Er as I say like), response elicitors (e.g., I will concentrate more and focus when I’m by myself. Do you get what I’m saying?) (Andersen); adverb-like features (e.g., Now, food was rationed) (Waters); question tags (e.g., It’s a sin, innit) (Pichler); utterance-final tags (e.g., I had a class with her, right) (Denis and Tagliamonte); general extenders (e.g., I had a couple of colds or whatever) (Tagliamonte; Wagner et al.); quotatives (e.g., I was like, ‘Oh my god’) (Rodríguez Louro; Levey); intensifiers (e.g., I know who really matters in my life) (Fuchs and Gut); and discourse like (e.g., She’s like skinny) (Drager). The use of these variables is examined in dialects of the following inner- and outer-circle Englishes: UK English (Andersen; Waters; Pichler; Tagliamonte; Fuchs and Gut); Canadian English (Waters; Denis and Tagliamonte; Levey); Australian English (Rodríguez Louro); Indian, Philippine and Singapore English (Fuchs and Gut); American English (Wagner et al.); and New Zealand English (Drager). The volume thus covers a wide spectrum of variables and varieties, and generates a broad knowledge base on which future work can build.
Overarching aims
The volume was developed to showcase and promote the variationist analysis of discourse-pragmatic features in synchronic and longitudinal English dialect data. Some of the key players in the field were invited to contribute chapters which explore original and variegated avenues in discourse-pragmatic variation and change research, and which address the two overarching aims of the volume. They are: (i) to offer new empirical and theoretical insights into the mechanisms underlying discourse-pragmatic variation and change; and (ii) to provide new methodological and theoretical suggestions for approaching the complexities of analysing variation and change in this component of language.
The contributors readily and enthusiastically heeded the invitation to design studies that would allow them to demonstrate the richness and diversity of discourse variation research, and advance current understanding of the social and system-internal dimensions of discourse-pragmatic variation and change. Thus, the analyses of the well-researched quotative, general extender, intensifier and discourse like variables were designed to examine their variable use in new data sources, varieties and populations (Tagliamonte; Rodríguez Louro; Levey); probe new dimensions of their variability (Fuchs and Gut; Drager); and test long-standing, conflicting claims in the literature about their sociolinguistic conditioning (Wagner et al.). The analyses and discussions of the less widely studied interjection, vocative, tag and adverb variables were configured to identify and document innovations in their repertoires and usage patterns (Andersen; Pichler); probe the mechanisms underlying their innovation (Denis and Tagliamonte); and evaluate different approaches to systematically studying patterns of variation and change in their use (Waters). To maximise the impact of their studies, the contributors were asked to situate their findings in the broader context of the variationist literature and point out potential implications of their findings for future work in the field. The contributions in this volume thus broaden and deepen our understanding of the details of discourse-pragmatic variation and change, and encourage new ways of thinking about variability at this level of linguistic structure.
In order to explore the research avenues outlined, the contributors chose to refine and reassess established ways of approaching the analysis of discourse-pragmatic features within a variationist framework (Labov Reference Labov1972b). As a result, the volume introduces alternative ways of identifying discourse-pragmatic innovations and analysing them in mega-corpora (Andersen; Pichler; Fuchs and Gut); coding discourse-pragmatic features for their semantic-pragmatic variability and mutability (Denis and Tagliamonte; Wagner et al.); elucidating the nature, trajectory and acquisition of discourse-pragmatic change (Denis and Tagliamonte; Tagliamonte; Rodríguez Louro; Levey); and exploring how discourse-pragmatic features are manipulated in the construction of stances and styles (Drager). Moreover, several contributors address an issue that remains contentious for some: whether (all) discourse-pragmatic features should be conceptualised as linguistic variables, and how much flexibility is required and/or acceptable to define them as variables (Waters; Pichler; Denis and Tagliamonte; Cheshire). The great value of the methods and approaches advocated in the volume is demonstrated in the results they reveal (see ‘Overview’ for details). To facilitate future replication (see Macaulay [Reference Macaulay2003] on the value of replication), contributors were asked to describe their methods in some detail, outline their rationale, and discuss how they could be applied to the analysis of other discourse-pragmatic variables and/or the pursuit of different research objectives. By presenting contrasting but complementary methods and approaches to analysing the complexities of discourse-pragmatic variation and change, the volume constitutes an important resource for scholars new to the field as well as those keen to expand their approach to studying variation and change in the use of discourse-pragmatic features.
Organisation
The chapters in this volume deal with multiple, overlapping issues pertinent to discourse variation research. As a result, any attempt to partition the volume into narrowly thematic sections risks distracting from the combined strength of the contributions. Notwithstanding the fact that all chapters address key methodological issues in the study of discourse-pragmatic variation and change, I have structured the volume into four parts that reflect the range of issues covered. What serves to unify the four parts and all contributions is their fundamental concern with providing new empirical results and advancing methodological approaches.
I: Methods. The two chapters in this part by Andersen and Waters set out methods for identifying innovative discourse-pragmatic forms and for conceptualising both new and old forms as discourse-pragmatic variables.
II: Innovations. This part of the volume features two chapters, one by Pichler and the other by Denis and Tagliamonte, which detail methods for uncovering innovations in the use of well-established discourse-pragmatic forms and for elucidating the nature of discourse-pragmatic innovations.
III: Change: In this part of the volume, Tagliamonte, Rodríguez Louro and Levey provide longitudinal perspectives on discourse-pragmatic change and explore children’s participation in discourse-pragmatic changes in progress.
IV: Variation: Fuchs and Gut’s, Wagner et al.’s and Drager’s chapters in the final part of the volume examine the effect of register (formality, familiarity) on discourse-pragmatic variation and demonstrate the role of discourse-pragmatic variables in signalling speaker stance.
The volume concludes with an epilogue by Jenny Cheshire, one of the first scholars to investigate discourse-pragmatic variation (see Cheshire Reference Cheshire1981) and the scholar whose foundational work in the field has inspired this volume.
Overview
To help readers locate the chapters that are most relevant to their interests and needs, I will provide a brief overview of the volume. My descriptions will focus on detailing which variables and varieties were studied in each chapter, which goals were pursued and what results were produced. Because research themes cut across chapters, I will evaluate their collective implications and impact in the next section.
The volume begins with two chapters focusing on methods in discourse variation research. In the first chapter (‘Using the corpus-driven method to chart discourse-pragmatic change’), Andersen advocates combining corpus-based and corpus-driven methods for identifying discourse-pragmatic innovations. Andersen argues that a major drawback of the corpus-based approach (the standard approach in discourse variation studies) is that data extraction is determined by researchers’ prior knowledge of variant forms; this risks inclusion in the analysis of only a sub-set of relevant variants and not accounting for previously undocumented ones. The corpus-driven method overcomes these limitations by calculating the frequency of individual words and word sequences in corpora; subsequent comparison of these frequency and co-occurrence patterns across corpora recorded at different time points facilitates the identification of potentially innovative variants. In addition to outlining the key steps of this approach, Andersen illustrates its value by applying it to an analysis of discourse-pragmatic innovations in contemporary London English. His analysis reveals a number of forms and constructions that are strong candidates for innovation, such as duh, blood, at the end of the day and you get what I’m saying. As such, it provides a useful addition to recent studies exploring discourse-pragmatic change in this variety (see, inter alia, Andersen Reference Andersen2001; Cheshire et al. Reference Cheshire, Fox, Kerswill and Torgersen2008, Reference Cheshire, Kerswill, Fox and Torgersen2011; Fox Reference Fox, Buchstaller and van Alphen2012; Pichler Chapter 3; Torgersen et al. Reference Torgersen, Gabrielatos, Hoffmann and Fox2011).
Chapter 2 by Waters (‘Practical strategies for elucidating discourse-pragmatic variation’) addresses the two fundamental questions underpinning any variationist analysis (of discourse-pragmatic features): how to define the variable, and how to circumscribe the variable context? Waters’s review of previous discourse variation studies demonstrates that scholars have variously defined discourse-pragmatic variables on the basis of semantic, functional or derivational equivalence between variants, and that they have variously appealed to form, function and/or position to circumscribe the envelope of variation. Waters does not challenge this lack of conceptual uniformity but argues that it is a necessary reflection of the heterogeneous nature of the category of discourse-pragmatic features which have been examined. She supports this view in her detailed discussion of how to quantify the variable use of adverbs with discourse-pragmatic functions (e.g., actually, really, now). The use of these features is notoriously difficult to quantify because of their positional mobility, inherent multifunctionality and lack of semantic bleaching. Waters concludes that there is no one-size-fits-all approach to studying discourse-pragmatic variables; accountable variationist analyses of discourse-pragmatic features must be designed to accommodate the specific characteristics of the feature studied as well as the specific goals of the study conducted.
The second part of the volume on innovations opens with a chapter by Pichler (Chapter 3) (‘Uncovering discourse-pragmatic innovations: innit in Multicultural London English’) which investigates the use of innit and other negative-polarity interrogative tags in a socially stratified corpus of contemporary London English. By closely investigating variants’ positional, scopal, functional and social properties, Pichler uncovers that innit and a small number of its derivationally-equivalent co-variants are rapidly innovating in this variety. Their use is no longer restricted to right-periphery, clause-final positions but extends to the clausal left periphery and positions adjacent to left-dislocated and lone noun phrases. In these positions, innit and other variants broadly function to secure hearer involvement (e.g., Innit, I don’t like trains.) and to facilitate referent activation (e.g., The sister, innit, she’s about five times bigger than me.). These innovations had gone unnoticed in two previous analyses of innit in the same dataset which had paid insufficient attention to the form’s variable position and scope (Palacios Martínez Reference Palacios Martínez2015; Torgersen et al. Reference Torgersen, Gabrielatos, Hoffmann and Fox2011). Pichler argues that identification of these innovations and exploration of how they become embedded in any pre-existing system is made possible by adopting an empirically- and theoretically-grounded but flexible approach to defining the variable (context), thus supporting Waters’s proposal in Chapter 2.
In the next chapter (Chapter 4) (‘Innovation, right? Change, you know? Utterance-final tags in Canadian English’), Denis and Tagliamonte draw on a socially stratified corpus of Toronto English to explore innovations in the system of utterance-final tags (UFTs), i.e., ‘any utterance-final discourse feature [such as you know, right, yeah, eh] that primarily communicates to a hearer that the preceding proposition contains shared knowledge’. Their analysis reveals that the UFT system is dominated by right and you know, with the former increasing and the latter decreasing in frequency in apparent time. To assess whether this pattern reflects ongoing grammaticalisation, whereby right is gradually expanding into the discourse contexts of you know, or lexical replacement, whereby right simply replaces you know, Denis and Tagliamonte code UFTs for the discourse contexts in which they occur. Distributional results suggest that as the status of right changes from innovative to majority variant, it is expanding in use across discourse contexts. However, a grammaticalisation hypothesis is rejected by the results of sophisticated statistical tests which show a lack of differences in the number of discourse contexts in which different generations of Torontonians use right, suggesting that right was available across contexts from the outset. Denis and Tagliamonte thus conclude that the rise of right in Toronto English is a case of lexical replacement.
Part III on change opens with two chapters offering a diachronic perspective on contemporary discourse-pragmatic variation patterns. Tagliamonte’s chapter (Chapter 5) (‘Antecedents of innovation: exploring general extenders in conservative dialects’) draws on a database of four relic dialects spoken in peripheral communities in the north of United Kingdom in order to explore the use of general extenders (GEs), i.e., constructions such as and stuff, or something like that which prototypically serve a set-extending function but also frequently perform interpersonal and textual functions. Because relic dialects tend to preserve earlier stages of language development longer than mainstream dialects, they can serve as a proxy for diachronic data from which an earlier stage of GE development can be deduced. Analysis of these data thus allows Tagliamonte to test conflicting claims in the literature that synchronic patterns of GE variability are a product of grammaticalisation (Aijmer Reference Aijmer2002; Cheshire Reference Cheshire2007 vs. Pichler and Levey Reference Pichler and Levey2011; Tagliamonte and Denis Reference Tagliamonte and Denis2010). Tagliamonte’s analysis of two measures of grammaticalisation, syntagmatic length and co-occurrence with other discourse-pragmatic features, demonstrates that short GE variants such as and (all) (that) dominate the GE system in the relic data and that co-occurrence rates are higher with shorter than with longer GEs. Thus, Tagliamonte argues, the predominance of short GEs in synchronic dialect data is not a reflex of grammaticalisation but the retention of a conservative pattern of GE use in northern UK Englishes.
Chapter 6 by Rodríguez Louro (‘Quotatives across time: West Australian English then and now’) investigates quotative variation (e.g., And he says, ‘Well, that’s her thing.’ She’s like, ‘Oh.’) in spontaneous narratives of personal experience collected from Western Australian adults born between 1870 and 1980. The diachronic perspective afforded by these data enables Rodríguez Louro to establish dramatic changes in the choice of quotative variants as well as the performance of narratives between the late nineteenth and early twenty-first centuries. Although say constitutes the majority variant throughout the time-frame covered by the data, it is gradually being replaced by other variants, including zero, think, go and, negligibly, be like. The ingress of non-say quotative variants is accompanied by an increase in self-revelations through reports of inner thoughts, feelings and attitudes. This observation leads Rodríguez Louro to posit that the changes affecting the constitution of the quotative variant pool and the linguistic conditioning of the say variant are inseparably intertwined with changes to how quoted content is used in narratives (see D’Arcy [Reference D’Arcy2012] for similar developments in New Zealand English). According to Rodríguez Louro, it is the increase in internal thought-encoding that ultimately gives rise to the incursion of be like into late twentieth-century Australian English.
The third and final chapter on change by Levey (Chapter 7) (‘The role of children in the propagation of discourse-pragmatic change: insights from the acquisition of quotative variation’) explores how preadolescents acquire the innovating quotative variant be like (e.g., She was like, ‘Don’t do this to me.’). Levey examines the formal constitution of the quotative system and its underlying variable grammar in three complementary datasets of Ottawa English: recent recordings of children aged eight to nine and eleven to twelve (to probe fine-grained age differences in the acquisition process); recent recordings of adults (to incorporate an apparent-time component and situate children’s usage patterns in relation to community norms); and recordings of children and adults made in the early 1980s (to introduce a diachronic control and verify the existence of change). Levey’s analysis of these datasets reveals that preadolescents participate in and advance changes affecting the quotative system: the children have acquired the form be like and its linguistic constraints. However, subtle differences between younger and older children’s variable grammars for be like show that younger children’s variable grammar is less closely aligned with that of adults in the community than older children’s. Based on these results, Levey proposes that the acquisition of adult-like patterns of discourse-pragmatic variation is a more prolonged process than that of phonological variation.
The final part of the volume on variation begins with two chapters on register variation. Fuchs and Gut’s chapter (Chapter 8) (‘Register variation in intensifier usage across Asian Englishes’) introduces two complementary methods for the semi-automatic analysis of discourse-pragmatic variation: cluster analysis, used to uncover variation patterns across independent variables; and phenograms, used to visualise these patterns and facilitate their interpretation. Following a very detailed outline of these methods, Fuchs and Gut apply them to their investigation of the combined effect of register and variety on variation in the use of intensifiers, i.e., items such as very and hardly which are used to modify the following word. Their analysis of intensifier use in twelve registers of three Asian Englishes recorded in the International Corpus of English (Indian, Philippine and Singapore English) confirms Aijmer’s (Reference Aijmer2013) recent findings that register and variety dramatically affect discourse-pragmatic variation patterns. It reveals: (i) variety-consistent effects of formality on intensifier frequency (higher frequency in spoken/less formal than written/more formal registers); and (ii) more marked cross-variety intensifier frequency differentials in informal than formal registers. Moreover, the authors find that the three Asian varieties do not share a uniform pattern of intensifier use; they differ in terms of intensifier frequency and variant distribution. Fuchs and Gut explain these differences with reference to the varieties’ differential socio-political development.
While Fuchs and Gut’s study focuses on comparing variable and variant frequencies across registers, Wagner et al.’s study in Chapter 9 (‘The use of referential general extenders across registers’) is concerned with comparing variable frequencies and functions across registers. The comparison is of GE use in two corpora of US English which are comparable in terms of speaker demographics but differentiated by register: talk between familiars vs. talk between non-familiars. The analysis challenges the reliability of previous work which has relied on GE frequency comparisons to propose that register exerts an important effect on GE variability. Wagner et al. demonstrate how frequency results can be affected by cross-corpora differences in transcription methods as well as failure to remove outliers from frequency counts. They try to overcome these limitations by developing innovative coding methods to test whether register affects how speakers use GEs. Although Wagner et al.’s analysis is, by their own admission, limited to contrasting the frequency of set-extending vs. non-set-extending GEs and ignores any potential variation in the distribution of interpersonal/textual GE functions, the results demonstrate that the functionality of GEs (broadly conceived as set-extending vs. non-set-extending) is fairly consistent across registers as well as styles (personal vs. non-personal).
The penultimate chapter (Chapter 10) of the volume by Drager (‘Constructing style: phonetic variation in quotative and discourse particle like’) combines variationist methods and insights with acoustic phonetic analysis to investigate New Zealand adolescent girls’ use of discourse particle like (e.g., I was just like singing) and quotative like (e.g., She was like, ‘Whoa!’). Drager demonstrates how the girls exploit both the frequency and variable phonetic realisation of discourse like depending on a combination of their social group membership and their shifting interactional stances. Thus, degrees of monophthongisation and /k/-reduction in discourse like were found to vary not just in accordance with speakers’ membership in a Common Room vs. non-Common Room social group but also in accordance with the stance they take towards the content of their narratives or the people they discuss and quote. For example, girls used quotative like realisations associated with their own social group when reporting their own past speech; and they used quotative like realisations associated with other social groups when introducing the speech of others from whom they wanted to distance themselves. Based on these findings, Drager argues in favour of combining different methods and perspectives (variationist techniques and acoustic analyses; style and stance criteria) to advance current understanding of how speakers manipulate the phonetic realisation of discourse-pragmatic features to construct their styles and stances.
In her Epilogue to the volume (‘The future of discourse-pragmatic variation and change research’), Cheshire identifies four new directions for discourse variation research that she derived from findings reported in the volume: (1) investigation of whether different discourse-pragmatic variables and variants are involved in different types of linguistic change (e.g., lexical replacement vs. grammaticalisation), and to what extent such changes may be localised; (2) exploration of how the migration of discourse-pragmatic features to new positions affects their functionality, and what the social triggers are that motivate such positional and functional changes; (3) increased analysis of the acquisition of discourse-pragmatic variation patterns, specifically the order of children’s and second language learners’ acquisition of pragmatic vs. syntactic constraints on variation, and the effect of frequency and salience on the acquisition of individual features; (4) closer attention to the phonetic realisation of socially salient discourse-pragmatic variables and variants, and how this relates to the expression of stance. Beyond detailing what insights can be gained from pursuing these directions for future research, Cheshire also explores to what extent the analysis of discourse-pragmatic features can be accommodated within the variationist framework. While she acknowledges the benefits of using variationist methods in the analysis of discourse-pragmatic variation and change, Cheshire – unlike other contributors to the volume – remains unconvinced that such analyses must, by definition, use the linguistic variable as an analytical concept.
Implications
To conclude this introduction, I will sketch out what I see as the major themes that emerge from the volume, focusing on methods. (The implications of the findings reported in the volume for future work in the field are discussed in Cheshire’s Epilogue.) I will review the methodological and analytical approaches advocated by the contributors and highlight what new insights into the mechanisms underlying discourse-pragmatic variation and change their rigorous application affords us. Where applicable, I will also comment on how the field will benefit from future replication of the approaches advocated in the volume as well as from further testing of the hypotheses and theories developed by its contributors.
Andersen’s, Fuchs and Gut’s, and Denis and Tagliamonte’s contributions showcase the great value of employing sophisticated computational and statistical tools in discourse variation research. Andersen (Chapter 1) demonstrates how his inductive and exploratory corpus-driven analysis of two London English corpora uncovered previously undocumented discourse-pragmatic features, including response elicitor variants which were not reported in Torgersen et al.’s (Reference Torgersen, Gabrielatos, Hoffmann and Fox2011) top-down corpus-based analysis of this variable in the same datasets. Andersen’s findings suggest that the former approach is better equipped than the latter to uncover innovative discourse-pragmatic forms as well as the full set of available co-variants of a targeted variable. Consequently, the application of corpus-driven methods in the analysis of discourse-pragmatic variation and change promises to yield more accountable and reliable results. With its potential to uncover previously undocumented forms and constructions, the corpus-driven approach will also support endeavours to broaden the scope of discourse variation research to a wider range of variables than have been studied to date (see Pichler Reference Pichler2013: 234–5).
Like Andersen’s corpus-driven method, the mathematically-based computational tools adopted by Fuchs and Gut (Chapter 8) will facilitate the analysis of discourse-pragmatic variation and change in mega-corpora of several million words (see Poplack Reference Poplack, Fasold and Schiffrin1989), which – due to their size – do not lend themselves to manual data analysis. Not only can these tools handle multi-million word corpora but they can also compute the operation of multiple factor groups with multiple factors on discourse-pragmatic variation. Moreover, they neatly visualise the results obtained. The potential of cluster analysis and phenograms to be extended to the analysis of diverse variables and contextual constraints makes them a welcome addition to the variationist toolkit, especially at a time when the number of available mega-corpora is steadily increasing.
Equally welcome is the introduction of new statistical techniques to the analysis of discourse-pragmatic variables. To establish whether the apparent-time frequency increase in the use of right is accompanied by its expansion across discourse contexts, Denis and Tagliamonte (Chapter 4) employ a zero-inflated Poisson regression model. Application of this model makes it possible to disentangle the interaction between variant form, its frequency differential across generations and its associated contexts of use. It reveals that the distributional patterns suggestive of right’s apparent-time semantic-pragmatic context expansion are in fact an artefact of how right is used relative to its co-variants. Denis and Tagliamonte’s study thus cautions scholars against uncritically interpreting distributional patterns suggestive of apparent-time context expansion as conclusive evidence of ongoing grammaticalisation. It also challenges dominant views in the literature regarding the gradualness of semantic-pragmatic change and the role of grammaticalisation (or pragmaticalisation) in these developments (see also D’Arcy Reference D’Arcy2015; Denis Reference Denis2015; Margerie Reference Margerie2014; Tagliamonte and Denis Reference Tagliamonte and Denis2010).
Although corpus-driven methods, cluster analyses and Poisson regressions have been used in other areas of linguistic enquiry (see the authors’ contributions for details), these tools do not at present have a wide currency in discourse variation research. Andersen, Fuchs and Gut, and Denis and Tagliamonte provide compelling evidence of their respective merits for discourse variation research. The full articulation and illustration of the methods in Chapters 1, 4 and 8 provide models for replication in future studies.
The volume also convincingly illustrates the great hermeneutic and explanatory value of qualitative data analysis, i.e., of closely examining discourse-pragmatic variables in their interactional contexts of use and/or of quantifying their functionality across independent variables. In her analysis of innit and other negative-polarity interrogative tags in Multicultural London English, Pichler (Chapter 3) appeals to function in order to establish whether the previously undocumented non-clause-final occurrence of some interrogative tag variants is indicative of an interactionally-motivated linguistic innovation. By closely investigating every token of innit and its co-variants in non-canonical positions, she establishes that, far from being random or idiosyncratic performance errors, these tokens perform clearly identifiable discourse functions related to those performed by interrogative tags in their canonical, clause-final position. Pichler’s analysis thus demonstrates that qualitative data analysis is a compelling means of determining the discourse status of unconventional and unexpected uses of otherwise well-established discourse-pragmatic variables and variants.
In her attempt to elucidate the role of stance in discourse-pragmatic variation, Drager (Chapter 10) closely examines the interactional contexts in which quotative and particle like occur and what stances speakers take in these contexts (e.g., an adversarial vs. non-adversarial stance towards the person whose reported speech is introduced by be like). By correlating speakers’ shifting stances with their variable frequency of particle like and their variable phonetic realisations of quotative like, Drager demonstrates how the use of discourse like is subtly affected by factors that would go undetected in more abstract analyses. Drager’s contribution is an exemplary demonstration of how a more nuanced understanding of discourse-pragmatic variation can be gained by studying variables in their interactional contexts of use.
Finally, Wagner et al. (Chapter 9) also include function in their study of GE register variation. Broadly differentiating between GEs that perform at least a set-extending function vs. those that perform interpersonal/textual functions without simultaneously implicating a larger set, they demonstrate that how GEs are used is not significantly different across registers; GEs are used with set-extending functions at roughly similar rates across talk between familiars vs. talk between non-familiars. These findings challenge previous reports that have drawn attention to the register-sensitivity of GEs based on register-specific frequency differentials in GE use (see Wagner et al.’s chapter for references). Wagner et al.’s study thus suggests that endeavours to improve our understanding of register effects on discourse-pragmatic variation may need to be based on more in-depth analyses and results than frequency comparisons alone can afford (see also Fuchs and Gut Chapter 8). Collectively, then, Chapters 3, 9 and 10 support Pichler’s (Reference Pichler2010: 597) call for a fuller integration of qualitative data analysis in variationist discourse studies.
Yet the contributions in this volume go beyond highlighting the general importance of functional analyses; two contributions introduce new reliable ways of coding variables for function and pragmatic shift. Wagner et al. (Chapter 9) present a function coding scheme which sets out a series of objective criteria (i.e., number of referents in the syntactic/discourse context; syntactic ambiguity) for differentiating GEs with set-extending functions from those GEs with only interpersonal/textual functions. In their study of UFTs, Denis and Tagliamonte (Chapter 4) operationalise the number of discourse contexts in which right appears (e.g., statement of opinion, statement of fact, command, narrative etc.) as a proxy of pragmatic shift. These coding methods address concerns in the literature about the subjectivity of coding discourse-pragmatic variables for pragmatic function/shift and of confidently replicating coding criteria across studies (see, inter alia, Cameron et al. Reference Cameron, McAlinden, O’Leary, Coates and Cameron1988; Cheshire Reference Cheshire2007: 182–4; Labov Reference Labov1994: 549–50; Macaulay Reference Macaulay, Chambers and Schilling2013: 227). Wagner et al.’s and Denis and Tagliamonte’s coding schemes are based on objective criteria which can be consistently applied by trained coders without recourse to subjective decisions. As a result, they promise to improve coder reliability. Moreover, they can be faithfully replicated across studies as well as, potentially, extended to the analysis of other discourse-pragmatic variables. The studies reported in Chapters 4 and 9, then, provide important models for future work investigating the variable functionality and potential grammaticalisation of discourse-pragmatic features.
Several of the chapters in this volume demonstrate the empirical and theoretical value of extending the analysis of discourse-pragmatic variation and change patterns to relic, longitudinal and real-time dialect data. Recent studies of GE variation have attracted criticism for inferring evidence of the variable’s grammaticalisation from synchronic variation patterns without accessing an appropriate real-time benchmark (see Pichler and Levey [Reference Pichler and Levey2011] for details). In the absence of comparable diachronic data sources, Tagliamonte (Chapter 5) draws on contemporary recordings of elderly speakers of relic dialects to increase the time-depth of data analysed and test a grammaticalisation scenario for GEs in data assumed to retain earlier stages of language use. Tagliamonte’s exploitation of relic dialect data proves to be an ingenious way of expanding the synchronic database into the past and testing hypotheses that contemporary discourse-pragmatic variation patterns are a result of ongoing grammaticalisation.
Rodríguez Louro’s contribution (Chapter 6) is another compelling illustration of how unconventional datasets can be exploited to elucidate the development of synchronic discourse variation patterns. Rodríguez Louro draws on oral history recordings of four generations of Australians born as early as 1870 in order to establish changes in the formal composition, variable grammar and type of content introduced by quotatives, before the influx of be like in the late twentieth century. The results allow for a meaningful, longitudinal perspective on the nature of quotative change in Australian English, and help elucidate the sociolinguistic dynamics of quotative innovation.
Levey’s (Chapter 7) rigorous analysis of children’s acquisition of ongoing discourse-pragmatic changes provides a useful model for any future study of the acquisition of variation as well as the investigation of discourse-pragmatic change more generally. Levey exploits complementary datasets, i.e., synchronic corpora of child and adult speech as well as comparable real-time data, to probe the existence and directionality of ongoing changes in the quotative system as well as the acquisition of these changes by children. His analysis of multiple datasets overcomes the limitations of studies that infer evidence of change from synchronic data without verifying this scenario with reference or access to a diachronic benchmark. With the increasing availability of real-time vernacular datasets, Levey’s approach can be replicated in future studies of discourse-pragmatic change.
The analyses of discourse-pragmatic variation and change patterns in unconventional datasets and under-researched populations reported in Chapters 5–7 will inspire scholars to increase the time-depth of their data, and explore patterns of discourse-pragmatic variation and change in complementary datasets (see also Denis Reference Denis2015). Discourse-pragmatic changes do not (of necessity) unfold quickly and, as a result, they are not always easily observable in the relatively shallow time-span covered by synchronic corpora (see, for example, Pichler and Levey Reference Pichler and Levey2011; Raumolin-Brunberg and Nurmi Reference Raumolin-Brunberg, Nurmi, Narrog and Heine2011). Analysing them in relic dialect data, oral history archives and real-time corpora will therefore help explore unanswered questions surrounding the nature and mechanisms of discourse-pragmatic change.
The chapters gathered together here share a view of discourse variation research that is characterised by going beyond frequency comparisons of variables and/or variants across social groups or functional categories (see ‘Background’). For example, Pichler (Chapter 3) argues that she would not have discovered dramatic innovations in the use of innit if she had relied on comparing the form’s frequency across the age groups represented in the apparent-time data she analysed. It was her close attention to the positional distribution and scopal properties of innit that made possible the discovery of these innovations. Denis and Tagliamonte’s (Chapter 4) comparison of UFT variants’ frequency across three generations of Torontonians revealed a generational change from you know to right. Consideration of the discourse contexts in which you know and right occurred enabled them to clarify what type of change is underway: abrupt lexical replacement. The analysis of be like by Levey (Chapter 7) illustrates that while a comparison of different age cohorts’ variant frequencies can uncover children’s participation in ongoing linguistic changes, the nature and details of the acquisition process can only be elucidated through careful analysis of the sociolinguistic conditioning of competing variants. Wagner et al. (Chapter 9) show that while registers might be differentiated by overall GE frequencies, they are not significantly differentiated by GE functionality (broadly defined). Drager (Chapter 10) demonstrates that it is not only speakers’ fluctuating frequencies of discourse like but also their phonetic realisations of the variable that tie in with speakers’ shifting stance during an interaction. Thus, the contributions in this volume consistently show that what is key to elucidating the nature of discourse-pragmatic variation and change is not necessarily how often a discourse-pragmatic feature is used but how it is used.
The central claim made by this collection is that a diversity of conceptual and methodological approaches are required to offer a maximally comprehensive view of how and why the use of discourse-pragmatic features varies and changes. This key observation is most clearly articulated in Waters’s contribution (Chapter 2), but is integral to most chapters collected here. Discourse-pragmatic variables, including those studied in this volume, constitute a heterogeneous set of linguistic items and constructions which differ in their functional repertoires, formal structures and degrees of syntactic integratedness. Therefore, accountable results cannot be produced if scholars apply identical methods of data analysis across the range of variables studied or across the range of research questions pursued (see further Waters’s Chapter 2). Moreover, some flexibility in defining discourse-pragmatic variables and their variable contexts is required to accommodate the emergence of innovative uses of established variables in our accountable analyses, and to examine how these innovations fit in the larger linguistic sub-systems in which they become embedded (see further Pichler’s Chapter 3). (See, however, Cheshire [Epilogue] who expresses strong concerns about conceptualising (all) discourse-pragmatic features as linguistic variables and about changing the definition of the discourse-pragmatic variable across analyses.) While some uniformity in research design and execution is, of course, required to ensure reliability, intersubjectivity and comparability (see Pichler Reference Pichler2010), the methods employed must accommodate the specific characteristics of the variables studied as well as any potential changes in their use, and they must be tailored to the research objectives set.
Recognition of the requirement for ‘bespoke analyses’ (Waters Chapter 2) of discourse-pragmatic variation and change will challenge the continued preoccupation with a rather limited set of discourse-pragmatic features that can be defined as variables on the basis of Dines’s (Reference Dines1980) foundational criterion of functional equivalence between variants (see also Lavandera Reference Lavandera1978; Sankoff and Thibault Reference Sankoff, Thibault, Johns and Strong1981), and whose variable contexts can be circumscribed by positional criteria (e.g., quotatives, intensifiers, GEs). Thus, by demonstrating that different approaches to quantifying discourse-pragmatic variation and change are possible, desirable and in fact necessary in order to advance the current state of the field, this volume encourages scholars to investigate a wider and more diverse range of variables, and to examine new and unexplored dimensions of discourse-pragmatic variability. This volume provides readers with complementary methodological and theoretical insights for approaching the complexities of discourse-pragmatic variation and change in order to ensure continued progress in the exciting arena of discourse variation research.
