Donald Bren School of Information and Computer Sciences

This paper presents a formalism for representing knowledge in the legal domain with an explicit account of time. We do not propose an original contribution on temporal representation, but we present a temporal representation method for Legal Reasoning Systems based on some techniques proposed in the area of temporal logics for AI which happen to adequately fulfill some other representational requirements from the legal domain as well.

A central notion of our approach is that of temporal token. It allows to make direct reference to specific temporal occurrences, supports talking about temporal occurrence without the technical problems of reification, avoids the ontological problems of temporal types, allows for a neat separation of the temporal component - which eventually will be processed by a specialized temporal resoner -, and turns out to be very appropriate for expressing references to legal propositions, which is an important issue in representing legal knowledge, while staying within standard many-sorted first-order logics.