Abstract
This paper attempts to offer an explanation for the relatively disappointing progress of information retrieval research to develop as a coherent and firmly based empirical discipline. The explanation is based on the argument that there is an inherent categorical duality in the subject domain of information retrieval research which is stultifying for the development of a single category science. Attempts to develop a single category research tradition flounder in the face of this elementary categorical duality while attempts to recognise the duality result in rather complex research designs and perhaps intractable research problems revealed in distinct methodological dilemmas, particularly in relation to the problems of measurement. The different ways in which this categorical duality has presented problems for information retrieval research are explored through an analysis of the two main research traditions which have dominated information retrieval research – the archetypal and cognitive approaches – and in relation to the ideas of Kuhn and Laudan on what constitutes, and what conditions favour, progress in a scientific research tradition.