ABSTRACT
Within any model theoretic framework for natural language semantics, we associate with each category C in the language a set of logically possible denotations defined in terms of the semantic primitives of the model. We shall refer to this set as the type for C (relative to the primitives) and denote it TC . The problem we are concerned with here is how to specify the interpretations of syntactically simple expressions in a category C where, on the one hand, the expressions are not logical constants, and on the other hand, are not interpretable freely in TC . Two examples below illustrate the problem.
