Abstract

This thesis has two interconnected themes. First and foremost, it is a study of the philosophical and cognitive aspects of word-meaning. An examination of the differing grammatical behavior of words of the same general type shows that much structure in our language comes from aspects of word-meaning. Using some tools from theoretical linguistics, we can uncover numerous semantic concepts that interact with grammatical properties of language. In linguistics, these concepts are often only acknowledged for their purely grammatical uses, and in philosophy they have not yet been a focus of detailed study. By attending to their features, we can decisively address a number of well-known philosophical claims, and we can supply evidence for many foundational theses about the nature of language and mind. ;The second theme concerns the appropriate methodology of linguistics. I defend a linguistic methodology centered around four features. First, a theory of this sort is a modified version of the Davidsonian project. The Davidsonian position is justified on very general grounds. However, the project is also Chomskian in that it is a part of a theory of the mind. Secondly, there are strict conditions on importing semantic structure into the meaning of an expression. In general, there are not many reasons for treating the meaning of a word as having structure. In particular, explaining the inferential behavior of an expression does not by itself support adding structure to the expression's meaning. Third, the logical apparatus used is not assumed. A common strategy is to represent the meanings of sentences within a given logic, and to reason about the meaning-structures that the logic generates. In contrast, no assumptions are made about the underlying logic of the semantic theory. Instead, which logic is appropriate to the theory is a straightforward empirical question. These three components of the theory serve to lessen the importance of many of the considerations that often shape linguistic theories. Thus, the fourth component is that the day by day actual practice of linguistics should be very heavily focused on the empirical data of natural languages