Putting back the clock in variation studies

PUTTING BACK THE CLOCK IN VARIATION STUDIES Derek Bickerton University of Hawaii The presentation of Washabaugh 1977 represents a retrograde step in the analysis of creóle systems. Though he may be correct in his belief that the constraints on complementizer variation were inadequately stated in Bickerton 1971, his decision to replace that analysis by one based on lexical diffusion and the perception of surface structures is vitiated by a number of errors, both analytic and statistical, and ignores what is most interesting about linguistic change, i.e. the capacity of the human mind to make quite abstract generalizations about broad linguistic categories. Washabaugh 1977 constitutes an interesting attempt to re-analyse the variation between fi/fu and tu as complementizers in Caribbean Anglo-Creoles, and in so doing to revise the original findings on this topic in Bickerton 1971. Those findings, as is inevitable in any first approximation, were almost certainly inadequate in a number of respects, and I would no longer be prepared to defend them in detail. However, there should be some 'via media' between keeping the baby in his dirty bathwater and throwing both out together. Unfortunately, W opts for the second course; and, in so doing, he threatens to regress the study of creóle systems to the level of aimless botanizing on which it existed until fairly recently. This being so, his arguments demand something more than a superficial scrutiny. W (333) lists three assumptions which he attributes to me.1 Of these, only the first and third have any substance. As for the second—that 'the ^-replacement process applies to just the pre-infinitival complementizer'—W claims this is incorrect because replacement applies also to the ^-genitive and what he calls the 7?-dative' (i.e. 'benefactive'?) This may be true of Providence Island Creole (PIC), but my assumption was perfectly justifiable for Guyanese Creole (GC), since the ^-genitive is unknown there, and the preposition almost invariably takes the form fu; since the reflex of Eng. for is also/i/, it makes no sense to talk about 'replacement ' in this environment. But differences between PIC and GC are not really at issue. It may well be that, in PIC, replacement of/?-complementizer and replacement of/? in other environments proceed pari passu; but from this it does not follow that the two phenomena are necessarily interdependent. To show any such relationship, W would have to show that both replacements resulted from the same kind of syntactic motivation; this he does not attempt to do, and, I would imagine, cannot do.2 Indeed, his own practice contradicts this criticism, since elsewhere (18) he explicitly states that his own analysis 'will deal with just the replacement of the 1 AU page references are to W's paper in this issue of Language. 2 W seems to be under the impression that if both ^-complementizer and preposition-/? are removed during decreolization, this is sufficient proof that it is the lexical item fi that is being removed, rather than distinct representations of discrete grammatical categories. But it is trivially true that all features characteristic of creóle speech will be removed if decreolization is carried far enough. What matters, from the point of view of the study of creóle systems, is the features with which any given variable is in conflict, and the various factors that govern that conflict. It cannot be too often repeated that decreolization is not a process of replacing words by other words, but one of replacing systems by other systems. 353 354LANGUAGE, VOLUME 53, NUMBER 2 (1977) ^-complementizer'. His objection to my second assumption is, therefore, without merit. With regard to my first assumption, that 'variation in decreolization is bidirectional ' (W, 333)—which I take to mean that the continuum is linear, and that all style-shifting takes place along it, rather than at an angle to it—I find nothing in W's paper to make me alter it. The assumption referred specifically to syntactic rather than phonological phenomena. However, even if it were extended to all phenomena, W's counter-argument is inconclusive. He claims (336) that PIC has two axes of variation: one—from 'broad talk' to 'speakin'—represents...