Model generation for natural language interpretation and analysis

Abstract

Model generation refers to the automatic generation of mathematical structures that prove the satisfiability of logical theories. The research documented in this thesis investigates the use of model generation in the analysis and interpretation of formal semantic representations of natural language. Based on standard techniques for first-order model generation, we develop a model generation technique for a restricted higher-order logic and show how this method can be used to investigate the criteria that distinguish valid natural-language interpretations from interpretations that do not correspond to the intended meaning of the represented sentences. In particular, we investigate the analysis of singular definite descriptions and reciprocal sentences and show that model generation gives a computational method for describing theories of prefernece for natural-language interpretations.

nicht vorhanden