New-Dialect Formation in Canada: Evidence from the English Modal Auxiliaries by Stefan Dollinger

New-Dialect Formation in Canada: Evidence from the English Modal Auxiliaries . Stefan Dollinger . Amsterdam/Philadelphia : John Benjamins , 2008 , xxii + 355 pp. There is a dearth of scholarly documentation of Canadian English (‘CanE’) as a major variety of English, and particularly of the historical development of CanE. Studies of earlier stages have tended to be based more on settlement history than linguistic data. Also slow to gather momentum has been public acceptance of CanE as a variety in its own right, many Canadians even today believing that there is nothing that distinguishes Canadian from American English. The present book examines the development of a set of modal expressions in early CanE (more specifically Ontario English, the socially and historically dominant variety), used to test the validity of Trudgill's (2004) new-dialect formation theory and such notions as ‘colonial lag’ and the ‘founder principle’, and to seek answers to general questions of the type: ‘How conservative is CanE usage?’ and ‘When did CanE become a distinct variety?’ These aims are ambitious indeed, given the restriction of the analysis to a single grammatical category. The origins of the book in a thesis (University of Vienna PhD, New-Dialect formation in early Canada: the modal auxiliaries in Ontario, 1776–1850) are evident, particularly in the extensiveness of the research review in the early chapters. Dollinger is aware that some readers may find this distracting: “from some perspectives one might consider the research review as in disproportion to the immediate needs of the empirical part of the study” (p. 6), but his justification is that the thoroughness of the review enables it to “serve as a starting point for newcomers to the study of CanE” (p. 6). Chapters 3 and 4 respectively provide an overview of Ontario's external language history from 1776–1850 with particular reference to the contributions of demographically distinguishable immigrant groups, and an account of the corpus used as the empirical base of the study, the Corpus of Early Ontario English (pre-Confederation Section, or ‘CONTE-pC’). The first machine-readable corpus of historical Canadian English, CONTE-pC is of modest proportions (125,000 words of running text from diary entries, letters, and local newspapers). It is organized into three internal periods (1776–1799, 1800–1824 and 1825–1849) to comply with Labov's desideratum of periods no less than 25 years for the detection of real, non-lexical, language change. To facilitate cross-dialectal comparisons Dollinger also uses ARCHER-1 (A Representative Corpus of Historical English Registers, Version 1) and CL18P (A Corpus of Late 18th-Century Prose) as a further source of data. He concedes that his study is “at the mercy of the available data” (p. 166), and that his capacity to make effective comparisons is hampered by the absence of a corpus of early AmE (p. 283). As a result of the small size of the corpora at his disposal, many results do not reach statistical significance, forcing Dollinger to compensate by considering percentile changes in distributions. In ch. 5 Dollinger reviews two ‘scenarios’ for the origins of CanE, both of which figure prominently in accounts of the development of the variety, despite having been published well before advances in the study of new-dialect formation. That propounded by Morton Bloomfield, an early adherent to the ‘Loyalist base theory’ who focuses on the influence of politically and linguistically conservative Loyalist immigrants from the US, contrasts markedly with that of Matthew Scargill, who emphasizes the importance of the British influence on CanE. These positions are considered in relation to more recently propounded theories associated with polygenetic approaches to colonial variety evolution. These include ‘swamping’, the ‘founder principle’, and ‘colonial lag’. It is however Trudgill's model of new-dialect formation that is the most important to the present study. The attractiveness of Trudgill's (2004) approach, which incorporates six key processes (mixing, levelling, unmarking, interdialectal development, reallocation and focusing) is that it is conducive to the type of empirical testing that Dollinger reports in chs. 6–9. Chapter 6 presents the methodological background and explains the selection of variables. In referring to these variables there is some inconsistency: on p. xvii Dollinger claims to be studying “eleven modal auxiliaries” but later we discover that one of these items is actually the semi-modal have to. On p. 7 he refers to the “ten core modals and one semi-modal” (ought to, included as one of the ‘core modals’ is surely more of a ‘marginal’ modal!), and then again on the same page to “the data on eleven modals”. On p. 206 and p. 224 Dollinger refers to have to as a ‘modal auxiliary’. Chapter 7 focuses on can, could, may and might. The data corroborates Kytö's (1991: 209) finding that a stronger preference by can/could than by may/might for occurrence in negative contexts was already established in EModE. Dollinger's claim (p. 195) that AmE leads the way in the development of epistemic could is difficult to reconcile with the frequencies adduced by Collins (2009: 109) showing BrE to have a far higher overall frequency than AmE. The chapter has disappointingly few examples, making it difficult in many places to determine precisely what type of meaning is being talked about. Consider for example Table 7.4 “COULD and MIGHT – semantic categories and examples”, which has four cells resulting from the intersection of the categories ‘non-epistemic/epistemic possibility’ and ‘non-past/past’. Instead of providing eight examples to demonstrate the parallels between could and might, Dollinger provides only four. Furthermore, the interpretation of a number of examples is disputable; for instance in the following example cannot is claimed to be epistemic, but is more appropriately paraphraseable by ‘possible for’ rather than ‘possible that’: (7.6.b) and yet, if the people have settled themselves according to these offsets, in front, they cannot near be disturbed, unless it can be clearly ascertained that with a view of encroaching upon (CanE, letters-2) Chapter 8 deals with must and have to. The weak obligation markers should and ought to are relegated to a later chapter where, curiously, they are discussed along with would. The preference for spoken over written genres that have to displays in ModE (cf. Collins 2009: 67) appears to have been in place at least from the late 18th century, with informal written genres of CanE outstripping formal, leading Dollinger to surmise, quite plausibly, that: “this distribution is indirect evidence that HAVE TO entered the language not only via informal written styles, but ultimately, via informal spoken styles”. The data shows that the path undertaken by have to in its contest with must was independent, simultaneously more conservative than AmE and less progressive than BrE, and this is noted by Dollinger to corroborate generalized statements about CanE characterizing it as “more conservative linguistically than the United States and Australia [and other ex-colonies]” (Chambers 1998: 253). Chapter 9 deals with shall and will. Given the influence of the prescriptive rules enshrined in such widely used references as Lindley Murray's English Grammar, it is not surprising that Dollinger should find these two modals to have been subject to high levels of awareness. CanE is noted to start conservatively with respect to ‘1st person shall’, but then become progressive in Stages 2 and 3, casting some doubt on the applicability of the Loyalist base theory. Dollinger stresses, however, that this finding does not necessitate rejection of the theory as a whole; rather, it “suggests the introduction of a different process for variables that operate above the level of consciousness” (p. 248). Chapter 10 deals with the left-over items should, would and ought to. Dollinger's semantic classification, as for the other modals, is somewhat unadventurous – based primarily on Coates (1983)– and not explored in detail. With would, Dollinger's excuse for failing to undertake any “detailed semantic analysis” is that, like will, it is often difficult to discriminate senses (e.g. root willingness and intention). This problem could surely have been solved simply by the inclusion of an ‘indeterminate’ category. Dollinger discusses the development of the epistemic use of ought to, noting a dearth of attention to it in the literature. If, as he claims, the moribund ought to is dependent for its viability on its expansion into the epistemic realm, the miniscule size of the expansion confirms that the future of this modal is bleak indeed. Among the several significant findings reported in this book one may number Dollinger's identification of an increase in the epistemic meanings of the modals, along with a concomitant loss of root meaning (with may as the significant exception). Another is the demonstration that colonial lag is not a strong influence on the behaviour of the modals. Table 11.2 on p. 276 documents the non-conservatism of LMod CanE: of the nineteen variables listed (e.g. root possibility can and may, 3rd person shall and will), four are classified as ‘conservative’, eight as ‘neutral’, and seven as ‘progressive’. Another is the data-driven ranking of possible scenarios for the origins of CanE: drift (parallel development) → loyalist base (AmE influence) → independent Canadian development → British influence (a finding which, if accurate for early CanE in general, lends slightly more credibility to Bloomfield's account than that of Scargill). It is unfortunate that this book is beset by a large number of distracting typographical errors and infelicities (to mention but a few instances: on p. 22, “Geikie expressed his despise for all things Canadian”; on p. 166, “despite of unfortunate gaps in the AmE data”; on p. 169, “In the present study, it will be attempted to meet several aims”; on p. 226, “thesis of root loos”). The book is nevertheless a timely and welcome addition to the literature on CanE, particularly for its real-time diachronic basis and for its use of morphosyntactic, rather than the more familiar phonological data.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.