Psychiatric Natural Kinds: Implications for Nosology, Practice, and Policymaking

Abstract

I seek to answer two main questions in this dissertation. First, what are mental disorders? Second, what are the implications for psychiatry and mental health policymaking? I propose that mental disorders are natural, mechanistic property cluster (MPC) kinds, underlain by one or more biological, psychological, and/or social mechanisms, with harmful effects for their bearers. Mental illnesses, on the other hand, are disease states that obtain when (1) a mental disorder obtains; (2) the bearer’s subjective experience of the disorder is a sufficiently significant impediment to their life; and (3) ascribing disease status to this individual case of mental disorder serves the legitimate strategic goals of enabling access to psychiatric care or other appropriate benefits and accommodations. I reach this conclusion by considering and rejecting a series of alternative proposals. Some alternatives fail to conform to our best scientific understanding of the etiologies of mental disorders, while others fail to accommodate all conditions of concern for psychiatry. My approach offers improvements to the accuracy of psychiatric nosology, diagnosis, and prognostication that other approaches to the metaphysics of kinds do not, while successfully accommodating the psychosocial dimension of the etiologies of mental disorders. My account achieves this by adopting a holistic naturalist perspective on mental disorders’ etiologies and a radical pluralism about disease state ascriptions, which allow it to accommodate the full range of entities of interest to psychiatry and provide a disease concept aligning with its use in psychiatry. My account also allows us to distinguish good from bad causal inferences made on the basis of membership in a psychiatric natural kind. This reveals when it is, and is not, appropriate to rely solely on general inferences about mental disorders. I argue that drawing such inferences is appropriate in contexts of psychiatric research and practice, but not in general policymaking. This is because doing so in the latter context requires reference to a broader range of epistemic resources, without which the resulting policies are epistemically and morally unjustified.