Abstract

This article reviews the theoretical and methodological status of the traditional distinction between denotational ambiguity and vagueness, concluding that it is valid and necessary to use it in modeling natural lexical representation. The definitional criterion is argued to be unreliable, but the equivocality that linguistic and logical criteria appear to exhibit does not challenge the basis of the distinction. Rather, it is argued, their use must be qualified, and in particular confined to denotational rather than referential phenomena. While it may restrict their practical utility, the weakness of these metacognitive criteria does not establish the absence of a cognitive distinction between ambiguity and vagueness. An account of polysemy is then sketched which is consistent in many respects with Tuggy's cognitive grammar model of polysemy.