Validity and the causal structure of a disorder
There are two aspects to validation: one is establishing the very existence of a disorder, the other is determining how good our methods are of determining the presence of the disorder. Both aspects depend on the conception of the disorder as something with a causal structure implicating the symptoms of the disorder. The symptoms may figure in this structure as effects of a single latent variable, or as homeostatically related merely to one another, or in some more complex structure. But without some such conception of the causal structure of the disorder, we have no idea what we are about trying to validate criteria for the disorder. The problem in psychiatry is that we currently seem to be demanding that those causal structures should involve relations between psychological and physical variables; and we have no idea how to think of those relations in terms of causal processes and causal mechanisms.