Abstract:
Rules and cases arc essential elements in legal reasoning, but computational models have barely begun to reflect the complexities of their roles. Based on experience with a real case, this paper idcntifics four areas that deserve attention from anyone concerned with understanding the processes of a general legal reasoner. These are (1) combining rules that were adopted for differing purposes but that all have application to the problem at hand; (2) allowing for argument over the logical structure of rules, and managing to reason with them even when unsure what the logical structure is; (3) allowing cases to be used mainly lbr their facts and outcome, mainly for their reasoning, or mainly tbr the rules they lay down, and employing each technique when appropriate; and (4) extending the legal sources that are treated cases. The paper does not propose solutions but merely attempts, by way of cxamplcs, to suggest significant research areas.