Set Bound Variables and Its Application to Natural Language Processing

Abstract

We extend class bound variables in the inheritance hierarchy prolog PAL by introducing set bound variables (SBVs), which are variables with set constraints. Using SBVs we can get a useful method of semantic representation for objects and sentences. SBVs also contribute to dealing with multiple meanings of words without backtracking and to preserving the consistency of interpretation of words. SBVs have the ability to play an important role in the programming for natural language processing systems because they can represent interpretations which are updated successively in the process of understanding sentences.