Language acquisition: Studies in first language development Edited by Paul Fletcher and Michael Garman (review)

Abstract

470LANGUAGE, VOLUME 58, NUMBER 2 (1982) guistics.) Wanner & Maratsos also provide processing evidence that fits the notion of WH-movement . Their evidence needs to be reckoned with in whatever theory emerges as best. The notion of a hold mechanism (where one puts who in Who did you see? before knowing what role it plays) is obviously a useful concept. Halle's masterful paper emphasizes the notion of unconscious knowledge revealed by the analysis of phonology. He equates physical mechanisms and phonological principles in his discussion . Things might, of course, be radically different; the knowledge of sounds might be represented in ways that allow more than one physical representation. However, if an isomorphism exists between an abstract system and its putative mechanical realization in phonology, this is an auspicious sign for syntactic research in psycholinguistics. Altogether, the lessons which psycholinguistics can learn from phonology are not given as much emphasis as they deserve in a volume of this kind. A final word: although each of these papers is enmeshed in a particular research program, they nonetheless give us a rich and stimulating sense of the 'state of the art'. REFERENCES Chomsky, Noam. 1981. Lectures on binding and government. Dordrecht: Foris. Fodor, Jerry A. 1975. The language of thought. New York: Crowell. Weinberg, A., and R. Berwick. 1981. Parsing efficiency, computational complexity, and the evaluation of grammatical theories. MIT, ms. [Received 4 August 1981.] Language acquisition: Studies in first language development. Edited by Paul Fletcher and Michael Garman. Cambridge: University Press, 1979. Pp. xi, 508. Cloth $49.50, paper $14.95. Reviewed by N. V. Smith, University College London* It is useful to make explicit the distinction between 'language acquisition'— the transition, often idealized to instantaneity, from infantile ignorance to the adult's knowledge of his language—and 'language development', referring to the stages through which the individual passes in his progress from the 'initial state' to the final 'steady state' (in the sense of Chomsky 1980b, e.g. p. 9). The phenomenon of language acquisition has been of fundamental importance in the development of generative grammar because of the complexity of the final steady state: the set (of sets) of rules characterizing a speaker's competence is sufficiently great to warrant interesting innatist speculations on the nature ofthe initial state. Specifically, the subtlety and sophistication of thejudgments of which the adult is capable are too great to be accounted for by any currently available learning theory on the basis of the putative input to the child; one therefore seems to be justified in ascribing to the neonate a rich theory of grammar which determines the growth of his language ability as he matures. Viewed in this way, language acquisition is simply a fact which requires explanation—a fact which lends support to an innatist account to the extent that plausible alternative explanations are lacking. Language development is * I am grateful to Maria Black for her comments on a draft of this review. REVIEWS471 then of interest on any of a number of counts: (a)By showing how children progress to the final steady state, it can provide evidence for the correct characterization ofthat state—and, by extension, the initial state before children's exposure to language. For instance, the co-occurrence of both agentive and agentless passives can be described in a number of different ways; however, the widely-noted development of agentless passives before agentive ones is described more economically as a simple progression from a PS rule of the form A —» XY to one of the form A —» XY(Z) than as, say, the transition to optionality of a previously obligatory transformation. It must, however, be emphasized that both analyses are compatible with the data (as are indefinitely many others): there is no necessary reason to assume a direct relationship between the rules of one's internalized grammar and the utterances which one produces by reference to those rules. In the artificially simplified juxtaposition ofjust these two possibilities, it is clear that one is supported by the developmental facts in a way that the other is not, so the relevance of developmental considerations to the final grammar can be established in principle. (b)It is an open question whether...