Psychiatry still lives in a Kraepelinian world and scholarship on the work of Kraepelin has consequently been unhistorical. Selective exegesis of the various editions of his textbook has led to a rigid view of his contribution. However, Kraepelin lived and wrote during an important period of European intellectual history and his work can only be understood against this background. This paper analyses the development of his views in terms of the 'Research Programme' he planned early in his life and whose objective was the creation of a stable description and classification of the psychoses. This Kraepelin eventually achieved by longitudinally studying patient cohorts in terms of methodological criteria such as illness course and incurability. In the event, this methodology allowed him to identify by correlation 'clinical pictures' that both represented the 'essence' of the disease and provided a taxonomic criterion. Although avowedly atheoretical, Kraepelin thus managed to construct (influenced by Kahlbaum and Wundt) an empirical support for his Kantian categorization of the psychoses. A discussion of the cultural variables that moulded these ideas is included.