We are now at an interesting crossroads in the history of psychiatric nosology. There have been many oscillations over time between etiologic and descriptive models, between lumping and splitting, and between categorical and dimensional systems. It is unfortunate, but inevitable, that most of the etiologic models of the past have been based on unproven and unprovable theories. Like its predecessors, DSM-IV is a descriptive system, and it too gradually will be replaced by an etiologic model--one, it is hoped, that is more scientifically valid than previous attempts at etiologic explanation. This will be an important step for the profession, for scientific understanding, and for the patient.