Introduction: discourse-pragmatic variation and change

Following Macaulay’s (2002, 2005) 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 2013; Andersen 2001; Buchstaller 2014; D’Arcy in prep.; Macaulay 2005; Pichler 2013; Tagliamonte 2016) 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 (2013: 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 2006; Diewald 2006; Kaltenböck et al. 2011; Traugott 2003; and contributions in Degand and Simon-Vandenbergen 2011); and recent methodological advancements in accountably quantifying their variability (see, inter alia, Buchstaller 2009, 2011; Cheshire 2007; D’Arcy 2005; Pichler 2010, 2013; Pichler and Levey 2011; Wagner et al. 2015). Because of the myriad complexities involved in defining discoursepragmatic features as linguistic variables (see Buchstaller [2009]; Pichler [2010, 2013: Ch. 2]; Tagliamonte [2012: 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