Date: Wed, 15 Dec 1993 19:37:00 EST From: "Dennis.Preston" <22709MGR[AT SYMBOL GOES HERE]MSU.BITNET> Subject: modren metathesis Don Lance's question about his local hypercorrector's performance is a very interesting one - perhpas the most interesting in the context of variation studies and the so-called dominant paradigm. Historical metathesis (we wake up some morning and 'bridd' has become 'bird' ignores 'individual metathesis.' - that horrible moment in history when there must have been bridd-bird speakers. According to Chomsky et al. those speakers simply (sic!) had two grammars, one being preferred over the other for 'uninteresting' reasons (status, style, gender). I suppose I continue to find it odd that those who are interested in (and know a lot, about) variation engage the essential psycholinguistic questions which lie behind a genrative grammar so infrequently. Chomsky's assumption (which is explicit in any number of places) assumes, for example, that the compound-ccordinate questions of bilingualism is completely resolved in monolinguals, and this assumption goes unchallenged by variationists who seem to be happy with more and more data (and less and less impact on the modern study of language). Dennis Preston