Peter Trudgill's essay raises important issues, some of which are uncontroversial while others are less convincing, partly because of a narrow concept of identity and partly because of an infelicitous choice of case studies. For him, the denial of identity seems axiomatic, while I suggest that what his criticism boils down to is the relationship between accommodation and identity, and the question of when and how identity as a social concept is effective, specifically in processes of colonial and postcolonial dialect formation.
WHICH FORCES WHEN? A CRITICAL LOOK AT TRUDGILL'S ARGUMENTS
Trudgill begins by insisting on the overall importance of dialect mixture, caused by dialect contact and leading to new-dialect formation. I couldn't agree more. I also agree that monogenetic explanations of colonial dialect evolution are unconvincing; in fact, I suspect this should be largely uncontested in present-day scholarship. Most of the examples of such positions that Trudgill quotes date from the early 20th century, and the two recent ones qualify their statements: Bauer and Lass argue for an “essentially” southeastern basis of some Southern Hemisphere Englishes, but I cannot see them denying any impact of dialect mixture or any influence, albeit weaker, from other regions. I also wholeheartedly subscribe to his point that accommodation is the fundamental mechanism in dialect mixture – though I think that the relationship between accommodation and identity deserves to be looked at more closely.
While Trudgill concedes that identity is a viable sociolinguistic concept, it is not clear what kind of an effect he is willing to assign to it and how he sees its role restricted. He agrees that “there clearly are sociolinguistic situations where identity plays a role.” What kind of role could that be other than strengthening select forms as sociolinguistic markers of a specific group identity (and thereby contributing to dialect evolution, dialect change, and also, depending upon the nature of input components, dialect mixture)? Actually, his denial is more restricted, applying to “colonial dialect formation” and emerging “new national” identities only, and to a goal-directed interpretation of identity directly causing certain forms to be used more prominently, as implied by Tuten's phrase “in order to.”
To make his points, Trudgill occasionally constructs oppositions and interpretations that are slightly idiosyncratic. For instance, Mufwene's belief that creoles are not “created” but emerge “by accident” is presented as arguing against identity, which it is not, with Trudgill implying an understanding of the effects of identity as deliberate and conscious, an idea that is worth noting but open to discussion. The same applies to Labov's qualification of customary interpretations of his findings from his well-known Martha's Vineyard study as proof of the role of identity. It is true that Labov rejects too generic an interpretation along these lines, but he does not say (as Trudgill implies) that the Martha's Vineyard study fails to show such a correlation; it does, pace Labov, though the implication is that it may be exceptional in that respect (we “do not often find” such a correlation, he says). Further down, to document the relevance of his historical case studies for the evolution of new colonial varieties, Trudgill adduces but turns around Labov's “uniformitarian principle.” This is putting the cart before the horse, however: The uniformitarian principle was postulated to project deductions based on familiar, well-studied present-day situations into a dark past about which much less is known; here, however, we are expected to assess the well-documented and brightly lit present day by using the dim torch of medieval studies.
Trudgill's case studies convincingly document the effect of dialect mixture in newly emerging dialect contact situations involving European languages, but they all relate to the Middle Ages. Trudgill himself concedes that they do “not produce any positive evidence concerning the absence of any role for identity factors,” and I agree this would hardly be possible given the huge differences in historical settings and sociolinguistic circumstances. Much of this is reminiscent of the well-known documentation of new-town dialect emergence (e.g., in Milton Keynes in England), but not indicative of, by Trudgill's own restriction, the formation of new national dialects, given that in those days the concept of nation in a modern sense simply wasn't applicable. Hence, arguing on that basis that identity cannot have been the driving force of but is rather “parasitic upon” accommodation is logically not possible.
ON THE RELATIONSHIP BETWEEN ACCOMMODATION AND IDENTITY
However, this juxtaposition raises the issue of what the relationship is between accommodation and identity, an issue that is clearly at stake also in Trudgill's discussion of Tuten. In a process of accommodation, individuals approach each other's speech behavior by adopting select forms heard in their environment, thus increasing the set of shared features. It is a process of linguistic approximation with the social goal of signaling solidarity by diminishing symbolic distance; it contributes to group formation and group cohesiveness. Identity, in contrast, represents an individual stance with respect to the social structures of one's environment, an attitude that also contributes to group formation and group delimitation through establishing an “us vs. them” construct of human alignments and through establishing relationships of similarity or difference – that is, social classification and individual affiliation (Jenkins 1996). Accommodation is one of the mechanisms of expressing one's identity choices. Both are closely related, but not quite the same; the two notions emphasize different aspects of similar constellations and processes, different sides of the same coin. If Trudgill accepts accommodation (as he does) it is hard to see how he could deny any impact of identity altogether. Accommodation only works within groups, through group definitions and delimitations, and these are defined on the basis of an individual's identity choice as a prime constituent of group cohesiveness. Of course, identities can be multiple and dynamic; they need to be negotiated, and they may change in the course of time and vary from one social context to another. Linguistic accommodation, on the other hand, is commonly viewed as a rather long-term, goal-directed process, fueled by one constant identity projection.
Trudgill's underlying explanation of these processes, via Keller's maxim “Talk like the others talk,” which in turn is motivated by a more general biologically motivated principle of “behavioral congruence,” is perfectly to the point: Humans are social beings, and in a potentially hostile world group cohesiveness contributes to an individual's chance of survival. But given the above definitions, this principle in fact underlines the importance of identity as a driving force in linguistic evolution. Language choice is both an expression and a manipulation of social bonds, of relationships of power and solidarity, and identity decisions operate in both directions – either strengthening accommodation as an expression of solidarity or increasing differences caused by socially dissociating decisions. Under these circumstances it is difficult to subscribe to Trudgill's argument that acknowledges the existence of new identities and their linguistic correlates and the connection between identity and accommodation, but views identity as not the driving force but as “a consequence of accommodation” (Trudgill's emphasis), chronologically subsequent to it. This strikes me as unconvincing: It suggests that speakers who for no obvious social reason have been mutually adjusting their speech behavior become aware of this fact and then, based on their recognition of linguistic similarities, decide to increase their social bonds. Trudgill may be right in arguing that identity may not be “the” driving force of new-dialect formation – that is, the only or the main such force. Other such forces play a role (demography, for instance, as well as power relationships), and a strong one as well – no question. But identity is one such force, and depending on individual contexts it clearly may be a strong one. And one of the contexts in which this applies, even if it is not the only and perhaps not the primary one, is sociolinguistic cohesiveness on a national scale.
THE CHANGING ROLE OF IDENTITY IN DIFFERENT EVOLUTIONARY PHASES OF POSTCOLONIAL ENGLISHES
The assumption that Trudgill rejects is primarily that of speakers expressing a new national identity by specific linguistic choices. However, identity effects operate in all directions and cannot be restricted like this. In the emergence of postcolonial Englishes as described in my “Dynamic Model” (Schneider 2003, 2007), the stage of nation building, involving identity redirections with linguistic effects, comes fairly late, while the impact of conservative or other identity decisions in earlier stages can be observed as well – but with different consequences. In phases 1 and 2, the foundation and stabilization of a colony, settlers' mental alignments with the former mother country cause exonormative attitudes and behavior.1
%1In New Zealand, for example, settlers desired “to make a new Britain in the South Pacific” (cf. Schneider 2007:128–29).
Another issue is the question of the “development of new linguistic features,” rejected by Trudgill. It is certainly true that Australian wide vowels were not caused by Australians suddenly feeling like Australians rather than like expatriate Brits. But which socially diagnostic linguistic variable (certainly, phonetic variable) was ever created afresh? The low back diphthong onsets of Wolfram's hoi toider O'cockers (Wolfram & Schilling-Estes 1997) and the centralized diphthong onsets produced in Canadian Raising and by Labov's Martha's Vineyarders definitely existed as phonetic variants in the community before; but they were functionalized to become socially diagnostic, and possibly redistributed to specific phonological environments in the new context. And they certainly were strengthened, in terms of frequency and regularity of use and symbolic value attached, by their identity-marking role – just as some diagnostic features of new national varieties are.
Trudgill's case studies represent situations comparable to phases 1 or 2 in the Dynamic Model at best, and so they cannot test a process that is strongest at phase 4. Of course, there are linguistic forms considered diagnostic of individual postcolonial Englishes, and it is difficult to see how precisely these forms rather than any others should have been selected on a purely deterministic basis, excluding national identity as a factor. The strongest argument for the impact of identity in these processes is the observation that the origin and/or recognition and spread and/or scholarly documentation of these forms typically fall into periods of heightened national or social awareness. Cases in point (just a small selection) include2
For evidence for the first three examples, including more detailed references, see Schneider (2007:244–47, 131–33, and 275–81, respectively).
- Canadian Raising or the recently discovered Canadian sound shift, both of which occurred in the later 20th century, at a time when Canadian nationhood was at stake;
- New Zealanders' /e/ raising, /ı/ centralization or diphthong mergers of the late 20th century, when they had set themselves off from Britain but were determined not to be mistaken for Australians;
- the emphasis on a distinctive American English after the War of Independence, promoted by Noah Webster and others;
- and, currently, the emergence of a pan-ethnic Pasifika dialect in Auckland (Donna Starks and Allan Bell, p.c.).
Designing a study that will test a straightforward connection between sociopsychological attitudes (including national identity) and the use of specific linguistic forms in these contexts will clearly be a worthwhile task.
