Knowledge representation and inference based on first-order modal logic

Abstract

In this paper, we present a knowledge representation system based on a first-order modal logic and discuss the deductive inference mechanism of this system. A possible-world model, which is used in discussing the semantics of a modal logic, can be regarded as structured knowledge, and modal operators can be used to describe various kinds of properties on a possible-world model. In this paper, we introduce a new concept, "viewpoints of modalities", in order to describe the knowledge structure effectively and compactly. We also show that schema formulas available in this framework are useful for the description of metaknowledge such as property inheritance. Therefore, a modal logic is suitable for representing both structured knowledge and metaknowledge. We construct a knowledge representation system based on a subset of a first-order modal logic, and give a complete deductive inference rule which is as effective as SLD resolution.