This article is a review of the research literature with regard to the definition and operationalization of the concept of psychotic relapse, and the use of the BPRS in it. On the basis of the research literature it seems meaningful to make a distinction between `target symptom', and `specific clinical state' approaches. The former seems particularly fit in the management of individual patients, while the latter is more suitable for scientific research. The use of the BPRS, as a general psychopathological symptom scale, is recommendable for the latter method, e.g. longitudinal research.