What is the relation between models, as used in model-theoretic semantics, and the “world” which models represent? More specifically, let us consider the question of whether a single individual, event, time or other element in a model might be used to represent more than one individual, event, time or other object in the real world. It should be understood that the question here is not whether groups of objects may be represented with some sort of “plural individual” in the style of Link (1983); rather, the idea is to use a model in which pragmatically irrelevant distinctions between individuals are ignored, so that the model “conflates” more than one realworld entity into a single model-theoretic entity, representing them as though they were the same object. This idea was argued for most explicitly and systematically, to my knowledge, in Nunberg (1984); though related ideas have been hinted at or suggested briefly or casually in a number of other places. The present paper provides arguments against this technique, as developed by Nunberg, and suggests an alternative approach to dealing with cases where the distinction between real-world individuals is pragmatically ignorable, based on the device of “Pragmatic Halos” presented in Lasersohn (1999). In considering this issue, we will also have to consider exactly what the point is of using models in model-theoretic semantics; and although this seems like a very elementary issue, and one which lies at the foundation of our whole technique, I think it is still an issue about which there is a great deal of obscurity and differences in thinking. To a surprising degree, semanticists often use essentially identical formalism, yet understand that formalism very differently.
原文链接
分享
引用
