Epistemic Value Commitments in the Debate over Categorical vs. Dimensional Personality Diagnosis John Z. Sadler (bio) Abstract Contemporary philosophy of science tells us that scientific theories are “underdetermined” by their accompanying data in a variety of ways. Briefly put, theories are not constructed on data alone. Psychiatric classification is subject to this same kind of underdetermination. Theories may be determined by a combination of data, historical factors, practical constraints, value commitments, and other factors. While practical constraints (like user-friendliness or compatibility across diagnostic systems) are commonly admitted to be influential in shaping psychiatric classification, the idea that values shape psychiatric classification is still a controversial one. This article addresses this controversy by (1) defining a particular domain or type of value commitment (epistemic values), (2) showing how they might be assumed in psychiatric classification, (3) developing a set of conceptual tools to analyze epistemic value content, and (4) applying this set of tools to concretely illustrate how this type of value is “at work” in shaping classification. A classic paper on the dispute between categorical and dimensional models for personality disorders is systematically examined to illustrate this “epistemic value” rubric. Keywords psychiatry, mental disorders, categories, dimensions, classification, nosology, personality disorders, values, Thomas Kuhn, Hilary Putnam Scientific theories, Quine (1993) tells us, are “underdetermined” by the data they use in their claims. Scientific observation, interpretation, and generalizing are shaped not only by the data themselves, but a host of other determinants, both intrinsic to, and extrinsic to, the process of doing science. What, though, are these other determinants, and what role do they play in the attainment of scientific knowledge? Such otherwise diverging thinkers as van Fraassen (1980) and Habermas (McCarthy 1978) tell us that certain practical or pragmatic interests shape scientific observation and theorizing. In the case of van Fraassen, these pragmatic interests are relevant to developing and appropriately constraining the credible scientific explanation. In the case of Habermas, the practical interests may refer to the broader political-economic arena in which science is embedded. On the other hand, Kuhn (1962, 1970, 1989) and Fleck (1979), among others, demonstrate that prior historical experiences shape the scientist’s apprehension and interpretation of data. Historical experiences, in their sense, may include the influence of the culture [End Page 203] at large, the microculture of the scientist’s professional-ideological community, or occasionally even the personal quirks of the researcher! There is at least one other generative source of scientific theory. This source relates to values and commitments to them. Traditionally, values stood in opposition to facts as judgments or beliefs about the desirability, worth, beauty, or importance of the particular subject. So we talk of “facts” on the one hand, and value “judgments,” on the other. In medicine and psychiatry this thinking manifests in an ideally strict distinction between such terms as “symptoms,” “signs,” and “clinical findings” on the fact side, and things like “ethical” or “evaluative” issues on the value side. We will see below, however, that the ways values appear in psychiatric discourse are not limited to just ethical or moral values, but include a broad range of complex evaluative entities. Moreover, for many contemporary philosophers the distinction between fact and value is not so clear-cut. For instance, Williams (1985) has distinguished “thin” value concepts or descriptors from “thick” ones. Thin value concepts are terms that are abstract and transparently evaluative, like good or sinful. Thick value concepts are more concrete, yet complex descriptors where the evaluative meaning-connotation is embedded in a rich panoply of plainly descriptive meanings, as in chaste, elegant, or fatuous. In psychiatry, Fulford (1989, 1994) has described this kind of mixture of descriptive and evaluative elements within psychopathologic terminology. This mix of descriptive and evaluative meanings of terms sets up much of the controversy surrounding diagnosis-relevant concepts like impairment or trauma because of the greater amount of interpersonal disagreement in recognizing “impairment” (an evaluative descriptor in psychiatry) compared to more consensually agreed-upon (but perhaps equally evaluative) descriptors in physical medicine (like wheezing or congestive heart failure). As Fulford’s work suggests, thick value terms often appear in psychiatric nosology within category language and diagnostic criteria. The thick terms in this context...