ARGUMENTATION‐BASED INDEXING FOR INFORMATION RETRIEVAL FROM LEARNED ARTICLES Available to Purchase

Current indexing methods used in automated bibliographic and full text information retrieval assume that knowledge can be adequately represented as a semantic network which is manipulable by means of Boolean operators. However, this semantic approach requires the user to state formally what it is that he wants to find. This paper presents an alternative argumentation‐based method. It involves representing a learned article by means of rhetorical structure rather than by a semantic representation of content.