Abstract

Four papers that first appeared elsewhere, though within the last few years. Probably the best source for Chomsky’s recent technical work and a good source for his recent views about linguistic theory and about language and cognition generally. The introduction and the first two essays, "Questions of Form and Interpretation" and "On the Nature of Language," speak to the more general questions and the linguistics will not overburden the philosophical reader. From the middle sixties to the early seventies, many linguists proposed ever more abstract and idiosyncratic syntactical analyses, often invoking vague and overly powerful "rules" without much concern for how they could fit within a learnable system of rules. What motivated these linguists was a desire to give some account within a rich grammar for nearly any sort of semantic, logical, or cognitive feature that comes up in our use of language. Much of Chomsky’s recent work has been concerned with a principled reduction in the power of grammatical rules; since this reduces the sorts of rules and grammars that may be proposed, it makes grammatical analyses more testable. The reduction is "principled" in that what one is after are the universal constraints of human language that derive from the child’s language acquisition capacities. In "Conditions on Transformations" and "Conditions on Rules of Grammar" Chomsky proposes various limitations on the power of transformational rules within the context of a similarly-restricted base. The logico-semantic role in the strict explication of linguistic competence is restricted to "logical form," which is read off a surface level of syntax into something like the familiar structures of predicate logic. The trace theory of movement rules is a powerful and relatively simple way of accounting for features of pronominalization and anaphor, reference and logical form, without sacrificing testability or system; it also permits the expression of some elegant generalizations and lends itself to the exploitation of the universal—particular grammar distinction. Trace theory also permits logical form to be read off from surface structure alone. And the logical form so read off appears to be the familiar variable-employing one and not the sort suggested by Montague.—J.L.