Alexandra D’Arcy
In historical linguistics, evolutionary linguistics, computational (phylogenetic) linguistics, variationist (socio)linguistics, and a number of related fields, we pursue the facts of language change. The fact of language change itself is a given, “too obvious to be recorded or even listed among the assumptions of our research” (Labov 1994: 9), yet despite its banality, this observation is among “the most stubborn and difficult to assimilate as we try to come to grips with the nature of language in general” (ibid.). In our attempts to ‘come to grips’, we must necessarily rely on linguistic snapshots – the state of a language at some point(s) – yet any stage of a language is a historical artefact, implicating diachrony and synchrony as complementary modes of enquiry (Labov 1989). As such, a theory of language change cannot be developed from a single perspective. Romaine (1980, 1982) has argued that to be viable, such a theory must link past with present, and Labov has argued consistently and forcibly against the Saussurian separation of diachronic and synchronic linguistics (Weinreich et al. 1968; Labov 1975, 1989, 1994). At the core of the issue then is the way in which to resolve what is known about historical change with the mounting evidence from the study of change in progress. This latter field of enquiry has proved a valuable resource in the investigation of linguistic change, and one of the ways in which variationist sociolinguistics problematises the issue is through direct examination of embedding – the route a change follows through the language and through the speech community that uses that language. In this chapter I pay particular attention to the first part of the embedding problem and consider what happens when structure changes: what does change look like, and how does it proceed? To answer such questions we must first explore what is understood by ‘the grammar’. Change is longitudinal, and this has a critical ramification: linguistic structure cannot be monolithic. It must be either composite (cf. Hale 1998) or fluid, capable of accommodating perpetual flux. Operationalisation of this fluidity entails recognition of orderly heterogeneity. Natural language is not only in a constant state of change, but it is also characterised by variation that is highly constrained, both linguistically and socially. This is particularly true of spoken language, the prime locus of change: language in social context. Thus, the primary target of variationist sociolinguistics is not simply discourse, but the vernacular.
Speech style represents a long and diverse continuum of contextual, interactional, and topical influences, but the vernacular is the baseline. It represents the model first acquired, before the standardising influence of elitism and/or education becomes an overlay on usage and systematically eradicates the regularity of the constraints on language variation, and language change (Labov 1984: 29). The greatest challenge facing the linguist is to tap this style, but once accessed, it proffers a window on language as deployed in living speech communities, the crucible of language variation and change (Weinreich et al. 1968: 177). The variationist literature is thus replete with examples of speakers using two or more forms to express a single meaning or function. This variability, exemplified in (1) (from Tagliamonte 2006: 10–11; York English Corpus), is regular and probabilistic, conditioned by language internal and language external factors. It simultaneously operates at all levels of linguistic structure, from phonetics and phonology to morphology, syntax, pragmatics, and the lexicon, as well as being visible at interfaces. Thus, one of the key contributions of variationist sociolinguistics has been to establish the empirical validity and – as a consequence – the theoretical import of heterogeneity.
(1) a I did a college course when I lef-Ø school, but I lef-[t] it because it was business studies.
b You go to Leeds and Castleford, they take it so much more serious-ly…they really are, they take it so serious-Ø.
c I think she’s gonna be pretty cheeky. I think she’ll be cheeky.
d He’s got bad breath; he has smelly feet.
e She were a good worker. She was a helluva good worker.
f I wish that forty or fifty years ago I’d as much confidence. I wish Ø I’d had it then.
Crucially, variation and change are not coterminous: variation does not ipso facto entail change. Variation is the norm; change is the exception. But change does involve a period of oscillation, as older and newer forms compete for the expression of some meaning or function in discourse. A does not become B; rather, A and B alternate for a period and the frequency of one (or more) competing variants increases, spreading in both linguistic and social space. For this reason, change is often formalised as (2), which elegantly captures the central role of variation.
(2) A > A/B > B
Considered in retrospect, looking back from one stage of a language to another, more remote stage, change may appear abrupt and replacive. In this sense the final product can appear resolutely Neogrammarian in character. But while change is regular in many respects, it is rarely so tidy and conclusive as (2) suggests. Change is not deterministic, and there is no requirement that, once initiated, a change must progress to completion (where ‘completion’ is the wholesale loss of an earlier form; see also Labov 2001; Mair 2004; Brinton and Traugott 2005; Heine and Kuteva 2005). A change may begin and simply fail to advance, receding before it has the opportunity to accelerate and spread. It may progress to a certain point and then reverse, as was the case for analytic comparatives in English. It may reach an equilibrium in which the alternants remain in a state of stable variation, such as the case of [n] ~ [η] variation in the pronunciation of the English -ing suffix. Another possibility is for an older form to become entrenched in a particular context or construction, such as has affected English shall. Once the primary marker of future temporal reference, its use is now highly restricted, virtually non-existent outside first-person interrogatives and formal speech. With respect to sound change, numerous examples of both lexical and geographic exceptions illustrate the great frequency with which innovations ‘fail’ to diffuse in full (e.g. Chambers and Trudgill 1980).
To suggest that this lack of a consistent endpoint entails chaotic, unordered trajectories in the progression of linguistic change would, however, be erroneous. Variationist research has uncovered a number of constants, a perspective that is enabled by the fundamentally longitudinal nature of change. This is because, although individual changes are instantiated through distinct pathways of change, the transition period from one language state to another can last centuries, enabling a nuanced view of the mechanisms driving and constraining change. We are thus once again confronted with linguistic structure: what happens to it during the course of change? The question becomes particularly complex when linguistic material is recruited to a particular sector from disparate linguistic sources, at different stages in the history of a language. This level of complexity is not exceptional – it is in fact quite common. Paradigmatic examples include adhortative lets in English ([you] let us go >let’s go >lets give you), pas as a generalised negator in French (ne va >ne va pas >va pas; Hock 1991), the inflectional future of Romance languages (from Latin habere + infinitive constructions; e.g. French nous chanterons; Pinkster 1987; Roberts 1993), accusative cases in West African languages (from serialised verbs; e.g. kε in Gã, ‘take’ >ACC; Lord 1993), and complementisers in a number of African and Asian languages (from locutionary verbs; e.g. Ewe bé ‘say’; Lord 1976).
In English, ongoing change in the modal auxiliary system is giving rise to ‘a new taxonomical layer’ (Krug 2000: 3) as ‘quasi’ or ‘semi-modals’ such as have (got) to, need to, dare to, ought to, and be to are recruited from lexical verbs (see also Denison 1993; Warner 1993). Indeed, the modal system is a paramount example of longitudinal instability: it has experienced constant flux since at least the Old English period. The expression of obligation or necessity (henceforth deontic modality) provides a view to just one trajectory within this complex that is characterised by robust layering of forms across time. Illustrated in Figure 18.1, these forms have developed deontic functions at distinct times throughout the history of the language, and competition between them is ongoing. A wide range of lexical verbs have been recruited for this single modal function over hundreds of years. As the examples in (3) demonstrate (from Tagliamonte and D’Arcy 2007b), the contemporary reflex of this diachronic scenario is that speakers use multiple forms to encode necessity/obligation, often in parallel linguistic contexts, within single utterances.
Figure 18.1 Layering of forms for modal obligation/necessity in English
(Tagliamonte and D’Arcy 2007b: 52; adapted in part from Jespersen 1961: 51–54, Visser 1963–1973: 1478–1479)
(3) a I said ‘You have to come up’. I said ‘You must come up’. And to the person on the phone I said ‘I’ve gotta go’.
b Do what you gotta do in class. It’s not like you enjoy doing it but you have to do it if you want to go anywhere so.
c She would be like ‘No no no, we’ve got to keep going, we have to make it to the next level’.
In the English system of future temporal reference, cross-linguistically a notoriously multifarious area of the grammar, there is also ongoing, longitudinal development. The most recent strategy for futurity is periphrastic going to, which developed from purposive directional constructions with non-finite complements (e.g. Danchev and Kytö 1994; Hopper and Traugott 2003). This change began around the onset of the Early Modern period, but layering already characterised the sector. The traditional strategy for future reference utilised the simple present tense and an optional adverb. During Middle English, shall and will entered the system and began to mark futurity. These strategies compete with a number of other established ones, including the present progressive (optionally with an adverb), be to + infinitive constructions, and collocations such as be about to and be due to (e.g. Fries 1940; Visser 1963–1973; Traugott 1992; Nesselhauf 2007, 2010). Exemplified in (4) (from Tagliamonte 2012), this wide range of morphological, lexical, and collocational options coexist in contemporary use, despite the different times at which each strategy entered the future temporal system.
(4) a Who knows what the next century’ll bring? It’s gonna be quite interesting.
b Some’ll work, and some is not gonna work.
c I shall have to put the kettle on in a minute.
d Aye she’s at Workington but she goes to Carlisle in September.
e They’re spending Christmas and New Year with her.
Sociolinguistic variables are the formal representations of heterogeneity, reflecting the set of delimitable options for the realisation of some linguistic meaning or function. To paraphrase Labov (2001: 84), such variables are in the grammar (the set of generalised conventions over a set of speakers, i.e. the speech community, who understand one another), constrained by the grammar, and cannot be described apart from the grammar. Perhaps nowhere is this more evident than in the longitudinal evolution of grammatical systems, which I take here, following Langacker (1987: 3), to represent all symbolic structures of language (phonology, morphology, syntax, pragmatics, lexicon).
Grammaticalisation – the attribution of grammatical character to an erstwhile autonomous word (Meillet 1912) – provides the paramount case study of linguistic evolution, and is therefore of growing interest to variationists. The diachronic changes that are regularly instantiated as a form travels the cline from lexical status to increased grammatical status are long attested in the literature, and they are not particularly controversial. These include, among others, phonological erosion, reanalysis, metaphor, and extension. Grammaticalisation theory – the set of claims that have been made concerning these phenomena – however, has both strong proponents (e.g. Lehmann 2005) and strong opponents (e.g. Newmeyer 1998). One of the primary critiques against grammaticalisation qua theory concerns its status: does it have explanatory value in its own right, or is it epiphenomenal, explained by already well-known principles of linguistic change (Campbell 2001; for a rebuttal, see Haspelmath 2004)? Variationist theory, however, as a largely ‘pretheoretical’ method (Laks 1992; Walker 2010: 24), has little stake in this argument. The method tests predictions made by different linguistic frameworks, and while its stance on grammar as a group rather than individual construct is not compatible with all theoretical arguments, it is not itself inherently structuralist, formalist, generativist, or functionalist. Indeed, it is this characteristic that makes variationist linguistics ideally suited to provide key insights on the nature of language in general, including the embedding of language change in particular.
One such insight concerns the progression of syntactic change. Kroch (1989) demonstrated that when syntactic structure changes, all contexts are affected simultaneously and the innovation progresses at the same rate whether it is probabilistically preferred in a context or not. This is because the contexts “are merely surface manifestations of a single underlying change in grammar” (Kroch 1989: 199). As a consequence, contextual effects are constant across time (i.e. the constant rate effect).
This is not the case with grammaticalisation. Forms lose lexico-semantic and phonological content, change category membership, and spread gradually to new contexts, in some cases specialising for a given function. In this kind of scenario, the constraints governing variation must somehow be accommodated, necessarily absorbed or reorganised in the reconfigured functional complex. Outside the variationist tradition, work on grammaticalisation typically concentrates on the set of changes affecting an individual form or collocation. In variationist work, the analytical frame includes not only the grammaticalising form, but all forms involved in encoding a particular functional meaning. This focus on variable domains enables examination of the ways in which coexisting layers in a grammatical subsystem accommodate to the incursion of emergent forms. As a consequence, variationist research increasingly demonstrates that incremental shifts in the operation of the variable grammar can be used to elucidate pathways of change over the history of a language. In particular, structure can be discerned from the distribution and conditioning of variant forms, and “this structure, instantiated in the quantitative patterning of variants across elements of the context” (Poplack 2011: 212) provides a diagnostic of the transition between endpoints of change. In other words, variationist methods offer a snapshot of the structure of the system at a given period, snapshots which can then be compared. Not only does this enable us to trace the rise and fall of variant forms, it also enables us to discern their entry points into the system and the trajectories of their function.
To investigate the evolution of grammatical systems, variationists tap speech surrogates in the historical records (popular plays, journals, informal letters, and the like). This method is well-established in historical sociolinguistics (e.g. Nevalainen and Raumolin-Brunberg 2003), though it is by no means unproblematic. Most historical documents survive by accident rather than by design (Labov 1994: 11), and locating relevant diachronic evidence has been described as “making the most of bad data” (Labov 1994: 7). Nonetheless, careful and judicious use of speech-like written genres has proven insightful on a number of theoretical and empirical questions concerning the embedding of linguistic change (e.g. Poplack and Tagliamonte 1996; Nevalainen et al. 2011). Ultimately, this work demonstrates clearly that grammaticalising forms are sensitive to the variants with which they alternate. As they follow a developmental pathway, it is quite common for forms to lose, transfer, and acquire constraints, mostly as a reaction to the activity in the remainder of the sector.
In their analysis of future temporal reference in Brazilian Portuguese, for example, Poplack and Malvar (2007) reveal how, over the course of five centuries, change consisted of a series of small adjustments to the operation of the variable grammar. The prolonged and finely nuanced nature of this change highlights the sensitive role of structure in flux, but it is also noteworthy given that frequencies tell a much more dramatic story. Distributions alone, as in Figure 18.2, suggest cataclysmic shifts in variant use, particularly during the nineteenth and twentieth centuries, but analysis of the constraints affecting variant selection reveal ongoing, longitudinal, systemic shifts across time.
During the sixteenth, seventeenth and eighteenth centuries, future temporal reference was encoded one of two ways: via the synthetic future (cantarei ‘I will sing’) or via periphrasis with haver (hei de cantar). Synthesis was the majority and default variant, favoured in frequent and unmarked contexts such as contingent constructions and affirmative declaratives. It also occurred with the majority of lexical verbs. By the nineteenth century, the effects of contingency and sentence type had been transferred and lost: the futurate present (cantaria) subsumed the effect of contingents, while affirmative declaratives became the favoured domain of the incoming periphrastic form (vou cantar). The synthetic future found itself relegated to residual contexts, such as non-specific adverbials and distal futures, created when the incoming forms staked out their niches. For the futurate present, this was motion verbs; for the periphrastic form, it was proximal futures. By the twentieth century, only one effect operated on synthesis, which maintained a firm grip on futures with non-specific adverbial marking, a context for which it had no historically significant association. Periphrasis, which had emerged as the majority form, patterned in complementarity with the futurate present, in a complex grammar of variable constraints. These pathways are summarised in (5), in which each line represents a diachronic period (sixteenth—eighteenth century; nineteenth century; twentieth century).
Figure 18.2 Distribution of Brazilian Portuguese future temporal reference variants by century
(Poplack 2011: 216, fig.17.1; adapted from Poplack and Malvar 2007)
The key observation overarching this diachronic scenario concerns the longitudinal reorganisation of the way in which future temporal reference operated in Brazilian Portuguese. Through the sixteenth to eighteenth centuries, five constraints significantly affected the choice mechanism: sentence type, verb type, contingency, temporal distance, and grammatical person of the subject. These effects, the forms associated with each, and even the direction of effect itself continuously reorganised over five centuries. By the twentieth century, sentence type and temporal distance no longer exerted constraint effects, though adverb specification had come to the fore (with binary effects on the two primary forms (+/-) and a ternary effect on the obsolescent synthetic future, which maintained a niche in the system, specialised for use with non-specific adverbs). On its route to obsolescence, the synthetic future was jostled about in the system, losing, transferring, and acquiring constraints in reaction to activity throughout the future temporal system. Thus, although Figure 18.2 suggests a straightforward pattern of replacement, in which the periphrastic future simply generalised across the sector, this is not what happened. Individual forms negotiated new functions for the expression of futurity, redistributing the workload across time.
Where grammaticalisation has firmly caught the attention of variationists, fewer have explicitly paid heed to lexicalisation.1 However, lexicalisation – institutionalised adoption of forms into the lexicon (Brinton and Traugott 2005) – offers a number of intriguing avenues for investigations of evolutional pathways, particularly as lexicalisation subsumes a number of distinct phenomena (e.g. ‘ordinary’ word formation processes, such as compounding and conversion, as well as univerbation, demorphologisation, decliticisation, etc.). Like grammaticalisation, the emergence of a new contentful form – with syntactic and semantic properties distinct from its source content – involves syntagmatic fixation, coalescence, and loss of semantic compositionality (though the outcomes may be quite different; see Brinton and Traugott 2005: 105–106). Lexicalisation is not characterised, however, by decategorialisation or semantic bleaching. Thus, where grammaticalisation necessarily entails reorganisation of the functional complex in which forms evolve, lexicalisation does not. In this light, speech act verbs offer an insightful case study for variationist sociolinguistics to explore lexicalisation pathways.
Over the last two decades, variationists have become increasingly focused on direct quotation – the reproduction of speech, thought, attitude, writing, gesture, and sounds. This work, spurred by the recent and dramatic innovation of be + like as a quotative verb, is traditionally presented with the framework of grammaticalisation (e.g. Romaine and Lange 1991; Ferrara and Bell 1995; Tagliamonte and Hudson 1999; Tagliamonte and D’Arcy, 2004). Romaine and Lange (1991: 261), for example, proposed a possible grammaticalisation pathway in which the function of like was extended from that of preposition to complementiser, a shift that entailed recategorisation (subcategorisation for nominal elements to subcategorisation for sentential complements) and extended the domain of like via analogy.2 From this textual function it then continued to grammaticalise as a discourse marker, operating on the interpersonal plane.
Since Romaine and Lange’s proposal, there has been a wealth of research establishing that like functioned as a discourse marker long before it functioned as a verb of quotation (e.g. D’Arcy 2007, 2008), which raises questions about a pathway from preposition to complementiser to discourse marker (i.e. textual to interpersonal). Buchstaller (2014) has argued that it is the discourse marker, already fully established before the advent of be like as a verb of quotation, that fills the slot adjacent to be to form the specialised verbal collocate. Although it has not (yet) been argued that this syntagm has undergone univerbation, most analyses treat be like as a single, compound verb form and not as be plus a discourse marker (though see Buchstaller [2014: chapter 1] for a different perspective). It is possible that the trajectory is not one of grammaticalisation, but rather of lexicalisation: an optional grammatical form has come to function in tandem with a lexical form, with meaning, patterns of use, and constraints on use that are distinct from those of its source composites. It is thus ‘lexical’ in that it must be learned by speakers, along with the variable grammar that conditions it use (alongside other forms in the quotative repertoire).
In this light, the nature of the developmental pathway that be like has followed is of critical import. Ultimately, grammaticalisation affects an entire system, not simply a form that is acquiring a (new) grammatical function. As discussed above, forms lose, transfer, and possibly acquire constraints as they are redistributed through a sector, perhaps obsolescing, perhaps specialising, perhaps rising to prominence. Recent work targeting the development of be like, however, has revealed a remarkably stable variable grammar across its evolutionary history (Tagliamonte and D’Arcy 2007a; D’Arcy 2012; Durham et al. 2012), which, though short (c. 30 years), has been meteoric. In particular, three constraints have repeatedly been shown to have conditioned its use, both in apparent (synchronic) time and in real (diachronic) time: grammatical person, mimesis, and content of the quote. Thus, be like is favoured for first person mimetic thought, and appears to have developed to fill the emergent niche as narrators began to tell stories based on a series of internalised events, attitudes, and thought sequences. Despite fluctuations in the strength (i.e. magnitude) of the subject, voice, and pragmatic effects on be like, these probabilistic trends have been evident throughout its history. In so far as being a significant predictor on the use of be like, tense has also been a constant in its development. The details of its effect vary regionally, but there is overarching systematicity within the operation of tense across time. Thus, although in most varieties the trajectory of this constraint entails gradual differentiation of the simple present and the historical present (HP; past tense reference with non-present morphology), the effect itself is continuously present and does not wholly reorganise (i.e. where this configuration obtains, the past tense consistently disfavours be like ). This historical pattern is evident in Figure 18.3 (see also Tagliamonte and D’Arcy 2007a).
Figure 18.3 The tense effect on be like across time in New Zealand English (adapted from D’Arcy 2012: 364, fig. 6d)
Further evidence for stability in the development of be like concerns its putative entrypoint to the quotative system. It has been suggested that be like was first used to introduce thought and sound, and only later generalised to introduce direct speech (Ferrara and Bell 1995: 279). This would suggest distinct pathways, or at least, distinct junctures in the quotative system. Haddican et al. (to appear) argue against this view, providing evidence for a single trajectory in which both stative and eventive interpretations of be like emerge simultaneously. In particular, the reported thought and direct speech readings have diffused at similar rates, which is consistent with the view of a single change in which different contexts are affected in tandem.
The diachronic picture that emerges across the history of be like is thus quite distinct from that which emerged across the history of the periphrastic future in Brazilian Portuguese (indeed, for grammaticalising forms in general). In this latter case, different constraints affect use across the transition period, until the endpoint of the change when the reconfigured system stabilises. In the case of be like, the same constraints have affected its use, in the same way, across its full history, regardless of its status as incipient, mid-range, or nearly completed (on stages of change, see Labov 1994: 67, 79–83). Thus, where for example the Brazilian Portuguese periphrastic future was jostled about in response to changes in its sector (future temporal reference), be like was not. Its workload remained constant, irrespective of other ongoing, widespread, and diverse changes that were affecting its own sector (direct quotation). In fact, the system of direct quotation has undergone whole-scale restructuring over the past century and a half (Buchstaller 2011; D’Arcy 2012). In the mid-nineteenth century, the quotidian function of quotation was direct speech; quotative say was used 9 out of 10 times and the recounting of inner experience was exceptional. The ascendancy of say and speech reporting lasted well into the twentieth century, but underlying the overarching stability of the repertoire, mimetic encoding was emerging as an active constraint within the system, and the pragmatic restriction to speech was loosening as a range of content types began to spread and be encoded by an expanding repertoire of forms (D’Arcy 2012). As part of this latter development, the reporting of inner states and monologues was developing into a productive narrative genre, creating the niche that ultimately emerged as the nucleus of be like . By the end of the twentieth century, the system of direct quotation had developed an active and richly articulated variable grammar, in which a host of primary forms had a distinct role to play. Nonetheless, there was little fluctuation in the roles of individual forms across time. Thus, not only has be like been consistently constrained; the same holds for the traditional and archetypal verb of direct quotation, say. This form acquired new constraints over time, but these exerted a constant effect from their inception and older constraints remained stable as well.
Although variationist research is not aligned with any particular linguistic theory or framework, it is predicated on a view of language that aligns quite naturally with perspectives that incorporate the role of meaning construal in interaction. In particular, when Weinreich et al. (1968: 187) explicitly shifted the definition of language change away from individual lects (innovation) to the level of the community (change), they also focused the analytical lens squarely on speech in action (see also Labov, 1994: 45, fn.2). Discourse is the crucial site of language change, since it is here that speakers must negotiate meaning. It is also where they create new meaning. In this sense, ‘language change’ is an interactive process, incremented as communities of speakers develop different representations for encoding meanings. This differs from the structuralist model, which emphasises the role of endogenous properties of language in explaining change, as well as from the generative model, which foregrounds the cognitive states of individuals. In shifting from analysis of structure and lect, the window also opened on the role of social constructs within communities, and it was this view that predominated the first two decades of variationist research. Central to embedding is the question of social structure. But the question of linguistic structure is also vital, and it is here that this chapter has focused, not to relegate or minimise the critical role of either social factors or acquisition in shaping and constraining change (on social factors, see Michael, this volume; on acquisition, see Stanford, this volume), but rather to emphasise the link between established findings in diachronic linguistics and the variationist enterprise. What emerges from consideration of the discursive variable grammar underlying language variation and language change is that the established hallmarks of diachronic change (e.g. weakening, levelling, generalisation, bleaching, recategorisation, reanalysis, analogy, metaphor, etc.) leave distinct footprints across the history of a language, and these footprints have key implications for pathways of change. It is the emphasis on form/function asymmetries over the course of change that enables variationist linguistics to trace these pathways, underscoring the value of a holistic approach to diachrony: “The key to understanding language change is not to look at elements atomistically, but to see them in connection with other elements in actual use” (Joseph 2004: 62). Grammatical change typically does not proceed in isolation, but affects a system as a whole. Thus, to summarise this chapter, through observation not of forms but of systems we can discern “not only how they function today, but also how the ever changing and conflicting needs of their users are permanently at work silently shaping, out of the language of today, the language of tomorrow” (Martinet 1952: 125).
1 Indeed, lexicalisation has received far less attention in the linguistic literature in general than has grammaticalisation.
2 Once like could occupy the slot for verbs of quotation, a dummy be was inserted in order to license it within the frame.
Kroch, Anthony. 1989. Reflexes of grammar in patterns of language change. Language Variation and Change 1: 199–244.
Labov, William. 2007. Transmission and diffusion. Language 83(2): 344–387.
Tagliamonte, Sali A. and Alexandra D’Arcy. 2009. Peaks beyond phonology: adolescence, incrementation, and language change. Language 85(1): 58–108.
Torres Cacoullos Rena, and James A. Walker. 2009. On the persistence of grammar in discourse: a variationist study of that. Linguistics 47(1): 1–43.
Weinreich, Uriel, William Labov and Marvin I. Herzog. 1968. Empirical foundations for a theory of language change. In Winfred P. Lehmann and Yakov Malkiel (eds) Directions for historical linguistics: a symposium. Austin: University of Texas Press, 95–189.
Brinton, Laurel J. and Elizabeth Closs Traugott. 2005. Lexicalization and language change. Cambridge: Cambridge University Press.
Buchstaller, Isabelle. 2011. Quotations across the generations: a multivariate analysis of speech and thought introducers across 5 decades of Tyneside speech. Corpus Linguistics and Linguistic Theory 7(1): 59–92.
——2014. Quotatives: new trends and sociolinguistic implications. Malden: Wiley-Blackwell.
Campbell, Lyle (ed.). 2001. Grammaticalization: a critical assessment. Special issue Language Sciences 23.
Chambers, J. K. and Peter Trudgill. 1980. Dialectology. Cambridge: Cambridge University Press.
Danchev, Andrei and Merja Kytö. 1994. The construction be going to + infinitive in Early Modern English. In Dieter Kastovsky (ed.) Studies in Early Modern English. Berlin: Mouton de Gruyter, 59–77.
D’Arcy, Alexandra. 2007. Like and language ideology: disentangling fact from fiction. American Speech 82(4): 386–419.
——2008. Canadian English as a window to the rise of ‘like’ in discourse. Anglistik: International Journal of English Studies 19(2): 125–140.
——2012. The diachrony of quotation: evidence from New Zealand English. Language Variation and Change 24(3): 343–369.
Denison, David. 1993. English historical syntax: verbal constructions. Harlow, Essex: Longman.
Durham, Mercedes, Bill Haddican, Eytan Zweig, Daniel Ezra Johnson, Zipporah Baker, David Cockeram, Esther Danks and Louise Tyler. 2012. Constant linguistic effects in the diffusion of be like . Journal of English Linguistics 40(4): 316–337.
Ferrara, Kathleen and Barbara Bell. 1995. Sociolinguistic variation and discourse function of constructed dialogue introducers: the case of be + like. American Speech 70(3): 265–290.
Fischer, Olga, Muriel Norde and Harry Perridon (eds). 2004. Up and down the cline – the nature of grammaticalization. Amsterdam/Philadelphia: John Benjamins
Fries, Charles C. 1940. American English grammar. New York: Appleton, Century, Crofts.
Haddican, William, Eytan Zweig and Daniel Ezra Johnson. to appear. Change in the syntax and semantics of be like quotatives. In Theresa Biberauer and George Walkden (eds) to appear. Syntax over time: lexical, morphological and information-structural interactions. Proceedings of the 12th meeting of Diachronic Generative Syntax (DiGSXII). Oxford: Oxford University Press.
Hale, Mark. 1998. Syntactic change. Syntax 1(1): 1–18.
Haspelmath, Martin. 2004. On directionality in language change with particular reference to grammaticalization. In Fischer et al. (eds), 17–44.
Heine, B. and T. Kuteva. 2005. Language contact and grammatical change. Cambridge: Cambridge University Press.
Hock, Hans Henrich. 1991. Principles of historical linguistics. Berlin: Mouton de Gruyter.
Hopper, Paul J. and Elizabeth Closs Traugott. 2003. Grammaticalization. 2nd Edition. Cambridge: Cambridge University Press.
Jespersen, Otto H. 1961. A modern English grammar on historical principles. Part VI: Morphology. London: Bradford and Dickens.
Joseph, Brian D. 2004. Rescuing traditional (historical) linguistics from grammaticalization theory. In Fischer et al. (eds), 45–69.
Kroch, Anthony. 1989. Reflexes of grammar in patterns of language change. Language Variation and Change 1: 199–244.
Krug, Manfred G. 2000. Emerging English modals: a corpus-based study of grammaticalization. Berlin/New York: Mouton de Gruyter.
Labov, William. 1975. On the use ofthe present to explain the past. In Luigi Heilmann (ed.). Proceedings of the 11th International Congress of Linguists. Bologna: Il Mulino, 825–851.
——1984. Field methods of the project on linguistic change and variation. In John Baugh and Joel Sherzer (eds) Language in use: readings in sociolinguistics. Englewood Cliffs: Prentice-Hall, 28–54.
——1989. The child as linguistic historian. Language Variation and Change 1: 85–97.
——1994. Principles of linguistic change. Volume 1: Internal factors. Oxford: Blackwell.
——2001. Principles of linguistic change. Volume 2: Social factors. Oxford: Blackwell.
——2007. Transmission and diffusion. Language 83(2): 344–387.
Laks, Bernard. 1992. La linguistique variationniste comme méthode. Languages 108(1): 34–50.
Langacker, Ronald W. 1987. Foundations of cognitive grammar. Volume I: Theoretical prerequisites. Stanford: Stanford University Press.
Lehmann, Christian. 2005. Theory and method in grammaticalization. Zeitschrift für germanistische Linguistik 32(2): 152–187.
Lord, Carol. 1976. Evidence for syntactic reanalysis: from verb to complementizer in Kwa. In Sanford B. Steever, Carol A. Walker and Salikoko S. Mufwene (eds) Chicago Linguistics Society 12: Papers from theparasession on diachronic syntax. Chicago: Chicago Linguistic Society, 179–191.
——1993. Historical change in serial verb constructions. Amsterdam/Philadelphia: John Benjamins.
Mair, Christian. 2004. Corpus linguistics and grammaticalisation theory: statistics, frequencies, and beyond. In Hans Lindquist and Christian Mair (eds) Corpus approaches to grammaticalization in English. Amsterdam/Philadelphia: John Benjamins, 121–150.
Martinet, Andre. 1952. Function, structure and sound change. Word 8(1): 1–32.
Meillet, Antoine. 1912. Linguistique historique et linguistique générale. Paris: Champion.
Nesselhauf, Nadja. 2007. The spread of the progressive and its ‘future’ uses. English Language and Linguistics 11(1): 191–207.
——2010. The development of future time expressions in Late Modern English: redistribution of forms or change in discourse? English Language and Linguistics 14(2): 163–186.
Nevalainen, Terttu and Helena Raumolin-Brunberg. 2003. Historical sociolinguistics: language change in Tudor and Stuart England. London: Pearson Education.
Nevalainen, Terttu, Helena Raumolin-Brunberg and Heikki Mannila. 2011. The diffusion of language change in real time: progressive and conservative individuals and the time-depth of change. Language Variation and Change 23(1): 1–43.
Newmeyer, Frederick J. 1998. Language form and language function. Cambridge/London: MIT Press.
Pinkster, Harm. 1987. The strategy and chronology of the development of future and perfect tense auxiliaries in Latin. In Martin Harris and Paolo Ramat (eds) The historical development of auxiliaries. Berlin: Mouton de Gruyter, 193–223.
Poplack, Shana. 2011. A variationist perspective on grammaticalization. In Heiko Narrog and Bernd Heine (eds) The Oxford handbook of grammaticalization. Oxford: Oxford University Press.
Poplack, Shana and Elisabete Malvar. 2007. Elucidating the transition period in linguistic change: the expression of the future in Brazilian Portuguese. Probus 19: 121–169.
Poplack, Shana and Sali A. Tagliamonte. 1996. Nothing in context: variation, grammaticisation and past time marking in Nigerian Pidgin English. In Philip Baker and Anand Syea (eds) Changing meanings, changing functions: papers relating to grammaticalisation in Creole languages. Westminster: University of Westminster Press, 71–94.
Roberts, Ian. 1993. A formal account of grammaticalisation in the history of Romance futures. Folia Linguistica Historica 13(1/2): 219–258.
Romaine, Suzanne. 1980. The relative clause marker in Scots English: diffusion, complexity and style as dimensions of syntactic change. Language in Society 9(2): 221–247.
——1982. Socio-historical linguistics: its status and methodology. Cambridge: Cambridge University Press.
Romaine, Suzanne and Deborah Lange. 1991. The use of like as a marker of reported speech and thought. American Speech 66(3): 227–279.
Tagliamonte, Sali A. 2006. Analysing sociolinguistic variation. Cambridge: Cambridge University Press.
——2012. Sociolinguistics: variation, change and interpretation. Malden: Wiley-Blackwell.
Tagliamonte, Sali A. and Alexandra D’Arcy. 2004. He’s like, she’s like: the quotative system in Canadian youth. Journal of Sociolinguistics 8(4): 493–514.
——2007a. Frequency and variation in the community grammar. Tracking a new change through the generations. Language Variation and Change 19(2): 199–217.
——2007b. The modals of obligation/necessity in Canadian perspective. English World-Wide 28(1): 47–87.
——2009. Peaks beyond phonology: adolescence, incrementation, and language change. Language 85(1): 58–108.
Tagliamonte, Sali A. and Rachel Hudson. 1999. be like et al. beyond America: the quotative system in British and Canadian youth. Journal of Sociolinguistics 3(2): 147–172.
Torres Cacoullos, Rena and James A. Walker. 2009. On the persistence of grammar in discourse: a variationist study of that. Linguistics 47(1): 1–43.
Traugott, Elizabeth. 1992. Syntax. In Richard M. Hogg (ed.) The Cambridge history of the English language. Cambridge: Cambridge University Press, 168–289.
Visser, F. Th. 1963–1973. An historical syntax of the English language. Leiden: E. J. Brill.
Walker, James A. 2010. Variation in linguistic systems. New York/London: Routledge.
Warner, Anthony. 1993. English auxiliaries: structure and history. Cambridge: Cambridge University Press.
Weinreich, Uriel, William Labov and Marvin I. Herzog. 1968. Empirical foundations for a theory of language change. In Winfred P. Lehmann and Yakov Malkiel (eds) Directions for historical linguistics: a symposium. Austin: University of Texas Press, 95–189.