Computational models of ontology evolution in legal reasoning

Abstract

This thesis analyses the problem of creating computational models of ontology evolution in legal reasoning. Ontology evolution is the process of change that happens to a theory as it is used by agents within a domain. In the legal domain these theories are the laws that define acceptable behaviours and the meta-legal theories that govern the application of the laws. We survey the background subjects required to understand the problem and the relevant literature within AI and Law. We argue that context and commonsense are necessary features of a model of ontology evolution in legal reasoning; and propose a model of legal reasoning based upon creating a discourse context. We conclude by arguing that there is a distinction between prescriptive and descriptive models of ontology evolution; with a prescriptive model being a social and philosophical problem, rather than a technical one, and a descriptive model being an AI-complete problem.