CHALLENGING THE GVS
Of the various writers who in the 1990s carried on Stockwell and Minkova’s work of deconstructing the traditional conceptualisation of the GVE, Paul Johnston presents the most detailed argument and, to my mind, the most plausible one. I shall focus on it almost exclusively in this section.
Before we examine Johnston’s argument, however, we need to consider a few hard facts about standard languages. In effect standard languages are nothing more or less than “synecdochic dialects” (Joseph 1987). That is, they are dialects that are “chosen”—generally, but not always, by chance—to “represent” all the others in the construction of the “language” itself. During the process of standardisation a synecdochic dialect is also prey to the introduction into its lexical, grammatical and phonological systems of elements from other dialects and other “languages”. In this sense, a standard language is never free from innovations introduced either naturally or by force of codification processes that are often controlled institutionally by language academies, education systems and political engineering.
In addition, many standard languages begin life as written variants syn- ecdochically representing the dialect varieties, both social and geographical, which exist within the territory in which they are destined to become a sociopolitical “standard”. The final stage in the standardisation process is often, but not always, the construction of an oral standard, and this final stage can lead to divisive dispute over what is or is not part of that oral standard (see, e.g., Crowley 2003; Milroy & Milroy 1985; Honey 1997; Carter 1999). This has been the case within the last 30 years in Britain, as we shall see in chapter 10 , and the dispute has by no means reached a settlement (Bex & Watts 1999). In the case of English, the beginnings of a concern with the oral standard can be traced back to the second half of the eighteenth century (Sheridan 1762; Walker 1791), although much of the groundwork had already been laid by the orthoepists in the late sixteenth and seventeenth centuries.
If we accept this very brief sketch of the standardisation of English, it follows that even if we were to restrict the GVS to the sixteenth and seventeenth centuries, we would still need to look closely at the dialects to find evidence of vowel shifts, since there was no standard oral English at all till, at the very earliest, the latter half of the eighteenth century—that is, of course, if we are prepared to accept the existence of an oral standard at all. There is also evidence in the present-day dialects that vowel shifts are still continuing, and if we consider some of the currently more prestigious varieties of English in the United Kingdom (the “infamous” Estuary English) and the United States (parts of the Midwest), it is noticeable that they may be exerting a significant influence on the shifting of vowels in oral standard English English (to use a term coined by Trudgill ) and oral standard American English. If we carry this argument back into the past beyond the sixteenth century, which is what we need to do to identify the beginnings of the GVS, the only place we will ever be able to locate those vowel changes will be in localised forms of writing: in the dialects. This, in turn, means accepting the validity of occasional spellings as evidence for vowel shifting, which is what Wolfe (1972: chap. 4) rejects in favour of a close study of the orthoepists.
Johnston, on the other hand, argues that, if we seriously intend to look for the beginnings of the GVS, “orthoepic evidence is, in fact, too late” (1992: 205). He supports his argument by locating the likely areas of vowel shifting in the dialects through a close study of important sources of evidence such as the Survey of English Dialects, Anderson’s A Structural Atlas of the English Dialects (1987), Kristensson’s Survey of Middle English Dialects ( 1988), and McIntosh, Samuels and Benskin’s A Linguistic Atlas of Late Mediaeval English (1986).
Johnston is well aware of the vagaries of relying on occasional spellings. The first difficulty in interpreting possible phonological realisations from graphological evidence is the limited representational power of the Roman alphabet, complicated after the Conquest by the adoption of Norman-based spelling conventions. “Middle English” length interchange processes followed by short vowel raisings make it almost impossible to differentiate between long vowels that have gone through this process and those that may have been raised without shortening during the GVS. If we add to these difficulties in using occasional spellings the simple fact that what we are looking for is spellings that are “necessarily ‘occasional’ and sporadic” (Johnston 1992: 206), it is not difficult to appreciate the unwillingness of such researchers as Patricia Wolfe to consider this sort of evidence.
Indeed, even Johnston suggests that occasional spellings are hard to distinguish from “simple errors”. This comment by Johnston, however, is a distinctly odd one to make. Surely an “error” in spelling presupposes a well established set of conventions for spelling. Even Wolfe’s refusal to take spellings into account on the grounds that “anyone who has corrected many high school essays will be reluctant to ascribe much . . . phonetic accuracy to occasional spellings” (1972: 115) becomes vacuous if we consider that most of those “incorrect” spellings are an attempt to represent graphologically the way the speller pronounces the word. I would argue that precisely because they are “errors”—a point that in any case has little substance in the absence of a generally accepted set of spelling conventions—they are the kind of evidence we should be looking for. Attempts by local scribes to find a way of representing graphologically what they say provide invaluable insights into the earlier phonological structure of the dialects.
Johnston’s careful analysis of occasional spellings from all parts of the country (see his table 2) reveals the need to reinterpret the GVS as two (possibly more) smaller chain shifts that must have taken place in different parts of the country at different periods of time. In general, he is able to differentiate a shift in the high front and back vowels in two areas that would hardly warrant the assumption of frequent sociolinguistic contact among their speakers, the Midlands and East Anglia, on the one hand, and the Southwest, on the other.
For these areas the focus of attention in terms of trade, migration, prestige, and so on must have been, primarily, local centres such as Norwich, Exeter, Chester, Stafford and, secondarily, London. The second chain shift affects the bottom half of the vowel trapezoid and overlaps the shift in the high vowels in the northern Midlands. The focus of attention here becomes Yorkshire and in particular the Plain of York, which lay outside the area of overlap. Evidence from occasional spellings from this area allows Johnston to surmise that “at least on the Plain of York, low-vowel raising occurred so early that it not only precedes the high-vowel raising chain but also Open Syllable Lengthening” (1992: 214). That is, the shifting can be traced to the early thirteenth century or earlier.
This is strong evidence to support Stockwell and Minkova’s rejection of the GVS as a coherent, unitary phenomenon. It is also strong evidence on which to base a rejection of interpretations of the GVS that see it as an element in the process of the standardisation of English, even if, as in the case of Jespersen, Wolfe and also Chomsky and Halle in The Sound Pattern of English (1968), this is not their explicit averred purpose. Johnston, however, goes further than this. In support of the statement that “Vowel shifting . . . can be described as a continuous process rather than as something unique to a specific, long-past period, and the consequent reduction in the number of explanations needed to account for all the data a gain in simplicity rather than a loss” (1992: 219), he suggests that studying the phonetic processes involved in any vowel shifting in the present can help us to gain a fuller picture of vowel shifting in the past. Some of this work has already been done by Labov, Yaeger and Steiner (1972) in their discussion of the connections between peripherality (or tensing) of vowels and raising, on the one hand, and deperipheralisation (laxing) of long vowels resulting in double-mora nuclei and consequent dipht hongisation, on the other. Johnston suggests that more attention should be paid to phonetic environments such as final or pre-pausal position, added stress (often for emphasis) and slower speech rate, all of which tend to produce an increase in vowel length rather than periph- erality. Lengthening “pure” vowels is likely to produce an increase in sonority. In the case of the high vowels this opens up one mora of the vowel to produce nuclei such as [H] and [uu].
All of these processes are involved in the apparent, and in all likelihood necessary, instability of vowel systems. They might lead us to suggest that, no matter how we try to explain the GVS, we are ultimately forced to admit that it really was not quite as “great” as has been assumed in the literature on the history of the language following Ellis, Sweet, Jespersen and Luick. Johnston’s work provides strong evidence of at least two partially independent, but at some stage interlocking, small vowel shifts, and it is quite possible that further research of this kind will necessitate the introduction of more small vowel shifts. At this point, however, the option that this leaves us of denying that the GVS ever took place at all makes me a little uneasy, and this is the final step in my argument.
-  Two examples of a standard’s being “artificially” constructed from the dialects making up a languageare Nynorsk (created by Ivar Asen in the nineteenth century) and Rhaeto-Rumantsch (created by HeinrichSchmid in the early 1980s).
-  Voeste (2008), who has looked in detail at variant spellings in German texts in the sixteenth century,comes to the conclusion that writers may actually have gained prestige in spelling a word in as many differentways as possible, which is further evidence of the lack of conventionalised standards of orthography.