Blackwell Reference Online

Extract

The work of phonologists is, in part, to discover empirical generalizations about the sound shape of utterances in each of the world's languages, and to formulate theories which predict these generalizations. The “in part” comes from the mentalist stance of generative linguistics: its ultimate goal is not merely a theory which accounts for generalizations in surface data, but a theory that accounts for these data in the same way that native speakers' internalized grammars do. In a sense, then, generative analyses are not really theories of (corpus-internal) linguistic data per se; they are theories of native speakers' tacit theories about those data – the grammars which speakers arrive at during language acquisition. In Chomsky's (1965) terms, generative linguistics aims at theories that possess explanatory adequacy – the ability to account for how language acquisition occurs (and thus also for the knowledge resulting from acquisition) – and not merely observational adequacy – the ability to correctly predict the grammatical or ungrammatical status, for the language in question, of any given form. This mentalist stance creates two complications for how phonologists deal with generalizations in linguistic data. First, a language's lexicon might contain generalizations which speakers have not incorporated into their mental grammars, perhaps because Universal Grammar does not ... log in or subscribe to read full text