Emil Kraepelin's view that psychotic disorders are naturally-occurring disease entities, and that dementia praecox and manic-depressive psychosis represent two different diseases, has been hugely influential on classificatory systems for psychosis. Corresponding to the Kraepelinian dichotomy, those systems generally differentiated schizophrenia from affective psychosis. This paper examines the debate that took place between 1980 and 2000 regarding this differentiation. During the 1980s, the scientific reliability of the diagnostic criteria was challenged. In the 1990s there were significant critiques of the validity of the Kraepelinian dichotomy. Yet the dichotomy has not been formally abandoned, and the discussion continues to the present day. This paper suggests that before psychiatry can abandon the Kraepelinian dichotomy, a new model for conceptualizing and describing psychotic symptoms may be required.