This article consists of two important parts. The first is a specific defense of some of the central claims made by stanley cavell in "must we mean what we say" against the criticisms of fodor and katz in "the availability of what we say." the major issue concerns the question of whether evidence of some sort is needed to support a claim by a native speaker about what we mean when we say something. Further speculations on this topic occupy the (...) other part of the paper. (shrink)