| Summary: | The aim of this thesis is to discuss the semantics of FQs (fuzzy quantifiers), formal semantics in particular. The approach used is fuzzy semantic based on fuzzy set theory (Zadeh 1965, 1975), i.e. we explore primarily the denotational meaning of FQs represented by membership functions. Some empirical data from both Chinese and English is used for illustration. A distinguishing characteristic of the semantics of FQs like about 200 students and many students as opposed to other sorts of quantifiers like every student and no students, is that they have fuzzy meaning boundaries. There is considerable evidence to suggest that the doctrine that a proposition is either true or false has a limited application in natural languages, which raises a serious question towards any linguistic theories that are based on a binary assumption. In other words, the number of elements in a domain that must satisfy a predicate is not precisety given by an FQ and so a proposition con¬ taining one may be more or less true depending on how closely numbers of elements approximate to a given norm. The most significant conclusion drawn here is that FQs are compositional in that FQs of the same type function in the same way to generate a constant semantic pattern. It is argued that although basic membership functions are subject to modification depending on context, they vary only with certain limits (i.e. FQs are motivated—neither completely predicated nor completely arbitrary), which does not deny compositionality in any way. A distinctive combination of compositionality and motivation of FQs makes my formal semantic framework of FQs unique in the way that although some specific values, such as a norm, have to be determined pragmatically, semantic and inferential patterns are systematic and predictable. A number of interdisciplinary implications, such as semantic, general linguistic, logic and psychological, are discussed. The study here seems to be a somewhat troublesome but potentially important area for developing theories (and machines) capable of dealing with, and accounting for, natural languages.
|
|---|
