Model generation for discourse representation theory
Michael Kohlhase · 2000
. We present a tableaux-based model generation calculus for DRT, which gives us an incremental approach to discourse processing in the presence of world knowledge. We show the usefulness of such as system for several discourse phenomena. 1 Introduction Natural Language Processing (NLP) is one of the core areas of Artificial Intelligence. While some of the subtasks of NLP, such as speech recognition and syntactical analysis have made considerable progress towards practical applicability, the area of natural language semantics has so far mainly concentrated on representational issues and on the semantics construction process proper. The field of "dynamic semantics " studies logical systems that are well-suited for representing the meaning of discourses and in particular anaphoric binding. It has developed various "discourse logics", e.g. Discourse Representation Theory (DRT) [8] or Dynamic Predicate Logic (DPL) [6]. Semantic analysis, -- inference on the basis of semantic information ...