A New Look at the Well-Formedness Constraint

Abstract

◇ Partee: What I want to talk about today concerns not just semantics so much as the relation between semantics and syntax. Although the work that I have been doing and that I will talk about is mainly within the theory of Montague grammar, I think that some of these ideas are equally applicable to transformational grammar. Emmon and I both have been talking a lot about Montague grammar while we have been here. I don't want to make it seem as though we think transformational grammar is, in the past, all of no value. It is certainly true that much of the recent work in Montague grammar by linguists has drawn very heavily on past work in transformational grammar and many linguists are now formulating theories which try to use best parts of Montague's work together with best parts of transformational work. So the theory of transformationel grammar, I think, should still be considered a very important part of the enterprise which now I call Montague grammar. The general topic within which this paper is a part is the topic of constraints on grammars. The goal of linguistic theory is to try to characterize what it is that makes human languages distinctive, how human languages are the actual human languages and what we think to be all of the possible human languages differ from other imaginable kinds of languages including the languages of computers or the languages that are invented by logicians. The class of human languages seems to have a lot of very particular properties and to characterize these is one of the important goals of linguistic theory. In general, the formulation of these properties comes in two parts: one being the very general form of a linguistic theory and the other part being formulation of certain universal constraints within such a theory.