A knowledge-based approach to language processing: a progress report

Abstract

We present a model of natural language use meant to encompass the language-specific aspects of understanding and production. The model is motivated by the pervasiveness of nongenerative language, by the desirability of a language analyzer ana a language production mechanism to share their knowledge, and the advantages of knowledge engineering features such as ease of extention and modification. This model has been used as the basis for PHRAN, a language analyzer, and PHRED, a language production mechanism. We have implemented both these systems using a common knowledge base; we have produced versions of PHRAN that understand Spanish and Chinese with only changing the knowledge base and not modifying the program; and we have implemented PHRAN using the query language of a conventional relational data base system, and compared the performance of this system to a conventional LISP implementation.