Presenting a peirce logic based inference engine and theorem prover for conceptual graphs

Abstract

Since the advent of Sowa's Conceptual Graph theory in 1984, various groups of workers have attempted to implement conceptual graph processors. Whilst there is no doubt that much has been learned from these, it is also true that few of them provide a deductive mechanism that is significantly more advanced than that of Prolog. Indeed, it has been the aim of some to simply reproduce Prolog in conceptual graphs. One possible reason for this is that the theory contains redundancy, which, if not reduced within the specification of a working system, increases complexity, leading to increased execution times, and data storage and knowledge base consistency difficulties. This paper suggests one approach to the design of conceptual graphs processors and describes what such a processor would do. In doing so it presents a real system developed at Loughborough University as a tool for the investigation of the properties of such processors.