Binding without Pronouns (and Pronouns without Binding)

This paper is rooted in the hypothesis of “direct compositionality” (cf. Montague, 1974). In its broadest formulation, this hypothesis is that the syntax and semantics work in tandem in such a way that the syntactic combinatory operations specify a set of well-formed expressions and each such operation is coupled with a semantic operation which provides a model-theoretic interpretation for the expression. I also assume a general Categorial Grammar (CG) implementation of this, whereby the syntactic category of an expression encodes its combinatory possibilities and its semantic type and where the semantic operation associated with any syntactic operation is predictable. (Moreover, as in most modern versions of CG, I assume that there is a small and very general set of syntactic/semantic operations—although the question at issue here is just how small.) Of crucial importance is the claim that each syntactic expression is directly assigned a model-theoretic interpretation, and hence no expression needs to be mapped into some other level such as LF in order to be interpreted.