1 Conceptual Organization of Case Law Knowl - edge Bases , Carole Hafner

The legal case base contains thousands of cases, all represented as text. In the last forty years or so, the corpora of cases have been digitally represented, enabling digital searches. With new technologies come new opportunities. For the case base, this was the opportunity to search quickly and efficiently through the cases for particular information, facilitating the identification of relevant cases for academic study or legal argument. One prominent approach since the 1960s was simply Boolean searches, using “and”, “or”, and “not’, for matching strings; companies such as LexisNexis and Westlaw provided legal firms with databases and search facilities, showing it to be a commerically viable approach. To the current day, string search continues to be the primary search approach used by web-based search engines, e.g. Google, albeit vastly improved with the use of pre-indexation of webpages on the Internet. However, string searches are limited in a range of respects. In [2], several of these limitations are highlighted, and an alternative approach is proposed concept indexation and concept search. Among the limitations, it requires some expertise to know what search terms to use, it is not feasible to know all the relevant combinations of terms, and there is an over-estimation of the quality of the results. A better approach would be to annotate the cases with conceptual cover terms, which are semantically meaningful generalisations over the textual particulars and variant expressions textual differences would be homogenised, and relationships between textual elements could more easily be idenified. Then searches would be with respect to fewer terms or related terms, providing a better, more accurate rate of returned cases. However, to annotate the cases requires that the textual content be annotated. Prior to the annotation task, we must provide some “map” of the relevant legal concepts and issues and their relationships to the legal case corpora. The paper proposes an organization of a case law knowledge base in terms of three interacting components. First, there is domain knowledge model that defines the basic concepts of a case law domain; these may be the actors, events and relationships from a particular domain, e.g. intellectual property or personal injury liability. Second, a legal case frame that represents the properties, roles and relationships of elements of the case and which the case instantiates. And finally, Issue/Case Discrimination Trees that represent the significance of each case relative to a model of the normative relationships of the legal domain, e.g. shephardised relationships and case-based reasoning. In current terminology, the first two would be ontologies, an ontology of the domain and of legal cases, which the last reflects reasoning. Case notes from Annotated Law Reports are used to derive the various components, among which we find: