The role of default logic in knowledge representation

Various researchers in Artificial Intelligence have advocated formal logic as an analytical tool and as a formalism for the representation of knowledge. Our thesis in this paper is that commonsense reasoning frequently has a nonmonotonic aspect, either explicit or implicit, and that to this end Default Logic (DL) provides an appropriate elaboration of classical logic for the modeling of such phenomena. That is, DL is a very general, flexible, and powerful approach to nonmonotonic reasoning, and its very generality and power makes it suitable as a tool for modeling a wide variety of applications. We propose a general methodology for using Default Logic, involving the naming of default rules and the introduction of special-purpose predicates, for detecting conditions for default rule applicability and controlling a rule's application. This allows the encoding of specific strategies and policies governing the set of default rules. Here we show that DL can be used to formalize preferences among properties and the inheritance of default properties, and so we essentially use DL to axiomatize such phenomena.