Compositionality, innocence and the interpretation of NPs

Neale (1999) outlines a theory of meaning which he claims can account for a range of problematic semantic data while respecting two intuitively plausible constraints on semantic theory: compositionality and semantic innocence. In this paper I argue that Neale’s theory, hindered as it is by its underlying conception of the relation between language and thought, fails to satisfy the criteria it sets for itself. Once this conception is abandoned, a new analysis becomes available which accounts for the same semantic phenomena within a more parsimonious framework, as well as respecting both compositionality and semantic innocence, properly understood.