Linking linguistic and non-linguistic information

作者:

Highlights:

摘要

Natural language semantics in systems for text understanding depends on linguistic surface structure and on knowledge about the domain of application. Whereas linguistic theory tends to prefer the former aspect, developments in artificial intelligence seem to favor the latter. Work in the knowledge engineering group of the LILOG project on a text understanding system, which integrates both aspects, gave rise to the following considerations. The main goal of this article is to point at some difficulties in linking domain knowledge to knowledge about language suggesting some directions one may take to overcome these problems.

论文关键词:Knowledge types,linguistic side,non-linguistic side,subsumption,word meaning,contextual factors,interpretation,sign relation,task dependence

论文评审过程:Available online 12 February 2003.

论文官网地址:https://doi.org/10.1016/0169-023X(92)90023-5