| In this dissertation, the thesis is defended that lexical semantics is not a subdomain of semantics, but rather a subdomain of pragmatics. That is, knowledge about the meaning of words is not part of one's knowledge of the language, but rather part of one's knowledge of the world. This distinction marks the traditional difference between dictionary and encyclopedia. One consults a dictionary to find out what the word lemon means (on the assumption that one knows what lemons are), while one goes to the encyclopedia to find out about lemons (on the assumption that you know how English speakers name them). The view suggested here is that this is a distinction that is not possible to maintain in a formal description of a language. Such a formal description, it is claimed, should contain only minimal semantic information about any lexical item. Almost all semantic information would be regarded as encyclopedic world-knowledge.; Part of establishing this thesis is to show that the interpretation of lexical items in utterances can be contextually determined, given a suitably broad definition of context, and is thus pragmatic in nature, understanding the domain of pragmatics to be language in context. It is shown that is is possible both for interpretation of an utterance to depend on the interpretation of its parts (to accept some form of the Principle of Compositionality) and for the interpretation of the smallest parts to be determined pragmatically from context.; Thus much of this dissertation is devoted to the task of providing a truth-conditional model-theoretic account of utterance meaning that allows for the interpretation of lexical items in context without recourse to any object that might be called its meaning. This account uses a Montague-type semantics and incorporates a modification of the methods suggested in Nunberg (1978). |
