This chapter examines in detail the writings of Emil Kraepelin, Eugen Bleuler and Kurt Schneider to assess the extent to which they provide evidence that the necessary conditions for inferring dementia praecox or schizophrenia – i.e. the observation of a set of regularities – have been fulfilled. The term dementia praecox, which Kraepelin was to bring to such prominence, had been introduced by Morel in 1852 but with referents different from those suggested by Kraepelin in 1896. Kraepelin himself had used the term in 1893 with referents somewhat similar to Morel’s but different from those he was later to employ. In 18%, Kraepelin adopted the criterion of similarities in onset, course and outcome to assign meaningfulness to any postulated cluster of bizarre behaviour whose antecedents were unknown. Kraepelin appears to have paid scant attention to two fundamental problems of discussing ‘onset’ in this context.