Support for Constructing Theories in Case Law Domains

Abstract

Reasoning with cases has been a central focus of work in Artificial Intelligence and Law since the field began in the late eighties. Reasoning with cases is a distinctive feature of legal reasoning and is of interest because such reasoning is both inherently defeasible, and because it is an example of practical reasoning in that it aims to provide a rational basis for a choice rather than to deduce some conclusion from premises. As reasoning with cases has developed, it has moved beyond techniq ues for matching past cases to the current situation to consider how arguments for a position are constructed on the basis of past cases. Recently it has been argued that this should be seen as a process involving the construction, evaluation and applicati on of theories grounded in the phenomena presented by the past cases. Our aim is to develop and refine this idea, with the ultimate goal of building a system which is able to reason with cases in this manner. This paper describes the implementation of a th eory con-struction tool (CATE) to aid in the construction and evaluation of theories to explain the decisions obtained in legal cases, so as to give an understanding of a body of case law. CATE gives a rapid way of creating and testing different theories. Use of CATE is illustrated by showing the construction of alternative theories in a small case study. CATE is useful in itself for anyone wishing to explore their understanding of a set of cases, such as lawyers practising in the domain and knowledge engine ers tasked with constructing a rule based system in the domain. We also believe that it offers good prospects for automating the process of theory construction.