The Kraepelinian syndromal approach to diagnosis taken by DSM-III and its successors, which defines disorders by their clinical phenomenon, has come under rising criticism with increasing calls for an etiologically based nosology. The relative virtues of a syndromal versus etiologic psychiatric nosology have actually been debated within our field for a long time. To deepen and historically contextualize our current discussion, I review in detail the proposal for etiologic diagnostic systems for insanity by David Skae (1814–1873). While his proposal was intuitive and appealing to some, others questioned its viability and utility pointing out a number of potential problems in its implementation that remain relevant today. Something critical might be lost for psychiatric disorders, they argued, if mental symptoms were removed from diagnostic criteria. Etiologically based diagnoses work best for mono-causal disorders and those where the causes all operate on the same scientific level. However, psychiatric disorders are highly multifactorial with a wide diversity of risk factors spread across biological, psychological and social-environmental domains so identification of a particular cause on which to base diagnoses would be difficult. Not only do individual risk factors contribute to many different disorders but most disorders are influenced by many etiologic factors. With respect to causes and disorders, psychiatry is characterized by a ‘many to many’ relationship which would make an etiologic nosologic system inherently problematic. Finally, causes and effects can be devilishly difficult to distinguish for psychiatric illness and, while clinically based nosologies aid in differential diagnosis, etiologically based system might not operate similarly.