Grice's Modified Occam's Razor

Speranza

Assuming the basic truth of Grice’s story about implicature, it seems easy enough to construct a story on which, for example, "dust" has only one meaning in the language – say, "remove dust from" – and any instance in which it is taken to mean "apply dust to" is a case of implicated utterance interpretation. Indeed, one could, thus, try to eliminate all ambiguity at the level of expression types (bank, slug, cleave, sanction, enjoin, ...). The question is whether in so doing one is respecting the facts about the semantics of natural language or distorting them to fit a particular program. Perhaps it is difficult to tell which one is doing. Why does it seem wrong to say that dust, bank, slug, etc. are univocal as types? Indeed, one might inquire after the basis of the intuitions speakers have about all of the properties (syntactic, phonological, semantic) of expressions in their language, or challenge their evidential status. If the intuitions go, however, it is hard to see what the data for linguistics would be.