Psychiatric Disorders Are Soft Natural Kinds

Tilmes concludes his interesting and informative piece with the sentence that “analysis of psychiatric vagueness merits further consideration.” I agree with this point, as well as with his earlier assertion that how one understands psychiatric vagueness may implicate the diagnostic model that one adopts, and the research that one pursues. Fortunately, there has been recent attention to vagueness in psychiatry, addressing both degree-vagueness (e.g., how much depression is required for a diagnosis of depression) and combinatorial vagueness (e.g., what sorts of symptoms are needed for this diagnosis) (Geert, Lara and Rico, 2017). Vagueness in psychiatry is related to a range of nosological debates, including about the value of categorical versus dimensional constructs. Notably, the editors of DSM-5 initially aimed to shift to a more dimensional approach, in keeping with the continuous nature of biological domains of function (Regier, Narrow, Kuhl, & Kupfer, 2009). Nevertheless, dimensional constructs can be reformulated as categories by using cut-points, symptoms of categories can be tallied up to form dimensions, and both categorical and dimensional measures are useful (Kessler, 2002). While many psychiatric traits are continuous, diagnostic categories have considerable clinical utility and were largely retained in the Diagnostic and Statistical Manual of Mental Disorders (DSM)-5 (First, 2005). Tilmes divides theories of vagueness into semantic, epistemic, and ontic ones, and defends the claim that if there is a such a thing as psychiatric vagueness then some cases of such vagueness are least in part semantic. That is, that at least some of this indeterminacy is due to our descriptions of the world, rather than due to the state of the world or what we know about it. While it is entirely reasonable to study the language of diagnosis, I would want to emphasize the complex and intertwined relationships between our concepts and the world, and to point out that many other considerations are at stake here other than the use of language by different communities. It is notable, for example, that the International Classification of Disease (ICD)-11 made the decision to use clinical guidelines that avoid the “pseudo-precision” of the DSM-5 (Reed et al., 2019). Thus, for example, whereas for generalized anxiety disorder DSM-5 specifies that symptoms have lasted for 6 months, ICD-11 refers instead, rather more vaguely, to “several months.” It is not, however, the case that this vagueness indicates means that ICD-11 takes an anti-realist position on mental disorders (Tilmes equates semantic