Variation theory and second language acquisition (review)

REVIEWS163 Variation theory and second language acquisition. By Hugh Douglas Adamson. Washington, DC: Georgetown University Press, 1988. Pp. viii, 92. $10.95. Reviewed by René Coppieters, Pomona College The rationale for VT&SLA springs from what A perceives as an ongoing shift within SLA. The psycholinguistic tradition, understood here primarily as the influence of early First Language Acquisition studies on SLA, is, according to A, rapidly adopting 'theoretical constructs and analytic tools from sociolinguists ' (vi). Furthermore, 'the field is in danger of being split into two camps— psycholinguists and sociolinguists, who both study language acquisition, but who are unable to integrate their findings. This book is, I hope, a first step towards ... a unifying theory' (ibidem). VT&SLA is centered around an outline of relevant studies in the sociolinguistics and SLA literature. Unfortunately, and unwittingly proving that the split he fears is indeed taking place, A completely ignores all studies based on the more recent developments in theoretical linguistics. ' There is no recognition here of the work of Stephen Pinker (e.g. 1984) in first-language acquisition (FLA), or that of Lydia White and Suzanne Flynn (cf. White 1982 and, most recently, Flynn 1987 on the possible role of Universal Grammar in SLA), to mention just a few relevant examples.2 As a result, rather than confronting traditions in need of dialogue, if not reconciliation, A restricts himself to discussing works—more exactly, parts of works—which are, in fact, already compatible . This exclusion is particularly problematic for SLA, which has, in recent years, taken considerable advantage of the opportunities for productive research offered by ongoing developments in linguistic theory (for a recent overview , see Lightbown & White 1987). The basic issue tackled by A is how to analyze variation in the speech of second language learners. The unifying theme is methodology. Within the constraints mentioned above, A's discussion is informative and well organized. It includes a review of the early first- and second-language morpheme-acquisition studies; an overview ofthe techniques developed by sociolinguists for the analysis of variation; an outline of the possible similarities between the speech of second-language learners and pidgins and créoles; an essay on the psychological reality of linguistic rules, in particular variable rules; a discussion of the variable rule notation to express the results of 'prototype' studies; and a short discussion of Krashen's monitor model (see, for instance, Krashen 1981) in relation to variation theory.3 In each case, A presents his analyses in the context of specific and detailed examples. 1 The last theoretical work discussed in VT&SLA is Chomsky 1965. A also refers to Bresnan 1978, although her views in that article are not discussed or reflected in his book. 2 White 1982 is mentioned in VT&SLA, but A does not review her work or take her concerns into consideration. 3 A also includes an experimental study, conducted by himself and his co-worker Mary Ciske. Using adult second language acquirers as subjects, they tested (with some success) Bickerton's 1981 claim that children use the contrast between the presence or absence of an article to mark the [± specific] distinction for NPs. Cf. Cziko 1986 for a thorough evaluation of Bickerton's thesis in first-language acquisition. 164LANGUAGE, VOLUME 66, NUMBER 1 (1990) If VT&SLA's strong point is a clear review of relevant studies on variation, its weak point is the lack of a consistent and explicit framework in which the important questions alluded to would find their place. Indeed, on many critical issues in linguistics VT&SLA offers sometimes dogmatic and often nebulous opinions which are neither properly justified nor sufficiently integrated. In Ch. 1, A points out that 'constructing an average order of acquisition requires abstracting away from the speech data of individuals, and in this sense it is like describing langue, or the competence of the ideal speaker-hearer' (6). What is the rationale, then, for this abstraction, which he otherwise strongly condemns, and what does it represent? A rather meekly answers: 'However, such abstracting is commonly done in psychological studies involving groups, and it results in a description that is closer to the actual speech data than are the very abstract descriptions of Saussure and Chomsky' (6). A ignores here the issue of the interpretation...