Abstract

Classificatory realism is the view that nature divides herself up into classes, or “natural kinds”, and claims that it is the goal of scientific classification systems to correctly identify, name, and describe these classes. On this view, the legitimacy of a classification is independent of us and our needs, and instead depends entirely on how well the structure of the classification “matches” the natural kind structure of reality. Progress with respect to classification consists in finding classifications that better match reality. Ultimately, classifications are seen as representations of reality, and it is the world itself that dictates what is to count as a good or correct classification of a given domain. In this thesis I argue against classificatory realism. In the course of this argument, I develop and defend an alternative view of our diagnostic and classificatory practices in psychiatry. This alternative view takes seriously the fact that psychiatric classifications and diagnostic concepts are technical extensions of natural language — tools that are introduced by particular humans, at a particular point in time, for particular human purposes. Viewed from this alternative perspective, some of the most vexing problems in the philosophy of psychopathology suddenly become tractable.