Philosophical prolegomena to a cognitive theory of metaphor processing.

摘要

The dissertation seeks answers to several foundational questions whose resolution is a necessary prerequisite to the development of a computational theory of metaphor processing. Working within a naturalistic framework, I address three main issues. Does metaphor fall within the domain of semantic theory or pragmatic theory? Is the concept of metaphor embedded in a 'folk' understanding of language and thought, and, if so, will the notion of metaphor-processing figure in any mature scientific psychology? Does the distinction between the metaphorical and the literal inherently involve claims about truth and reference, and, if so, how can this distinction be addressed by a psychology that presupposes methodological solipsism?;I then conclude that a semantic theory of metaphor must quantify over extra-psychological facts about the world. This raises the problems mentioned above concerning the psychological status of metaphor in a very pressing way. I reject both of the standard responses to these problems, eliminativism and sentential realism, and argue that explanations involving an extra-psychologically individuated type--such as metaphor--can constrain explanations involving a solipsistically individuated type. This offers some hope that an extensional theory of metaphor may be psychologically relevant.;In Chapter Five I develop such an extensional theory. Through philosophical interpretation of work by some psychologists--principally A. Tversky and A. Ortony--I arrive at a set of necessary and sufficient conditions for the metaphoricity of an utterance. This definition constitutes the principal positive accomplishment of the dissertation.;Chapter Six reviews some unsolved foundational problems for future attention, sets my project in the context of other philosophical work on metaphor, and then looks forward to a computational theory of metaphor processing.;I defend the traditional, Aristotelean view which seeks to analyze both metaphors and similes in terms of underlying statements of comparison. This view implies that metaphors have semantic content, that they can, like literal claims, but true or false. I argue that Donald Davidson's analysis of meaning can, contrary to Davidson's own position, support this Aristotelean comparativist thesis.