A COMPUTATIONAL MODEL OF VERBAL UNDERSTANDING

Abstract
This paper presents a theory of verbal understanding based on a formal model of conceptual structures that represent verbal meanings expressed in English sentences. Verbal understanding is defined as the capability for disambiguation, paraphrase, question answering, translation, etc., with regard to natural language sentences. The model has been implemented as Protosynthex 3 in LISP on the Q-32 time-shared system. Experimental results from the system include examples of the analysis of complex sentences, disambiguation of multisensed words via sentence context, question answering via logical inference, and meaning-preserving paraphrase generation. The authors conclude that sophisticated natural language processing by computers is a realistic goal that has been partly achieved.

This publication has 0 references indexed in Scilit: