Friday, September 29, 2006

Paul M. Pietroski, "The Character of Natural Language Semantics"

Natural language semantics should give up the assumption that it is to provide truth conditions for sentences, because the truth of a sentence is a massive interaction effect of sentence meaning and a host of worldly factors. Travis is right that predicates aren't functions to extensions, and there are other, Chomskyish reasons to doubt that the truth conditions of sentences are compositional. NLS should be a purely internal, non-truth-involving theory.

No comments: