[Access article in PDF]
Reliability and Validity in Psychiatric Classification:
Values and Neo-Humeanism
Keywords: Validity, reliability, values, taxonomy, classification, McDowell.
"On those remote pages [of the Celestial Emporium of Benevolent Knowledge] it is written that animals are divided into (a) those that belong to the Emperor, (b) embalmed ones, (c) those that are trained, (d) suckling pigs, (e) mermaids, (f) fabulous ones, (g) stray dogs, (h) those that are included in this classification, (i) those that tremble as if they were mad, (j) innumerable ones, (k) those drawn with a very fine camel hair brush, (l) others, (m) those that have just broken a flower vase, (n) those that resemble flies from a distance." (Borges 1966, 108)
There is very valuable philosophical work to be done in carefully articulating differences. Haslam's discussion of classification provides a useful account of some of the differences between taxonomic kinds that may be found in psychiatric theory. There is, however, an underlying concern about psychiatric classification that is not addressed in the graded distinctions highlighted and that, in this short note, I wish to address. (This is not a criticism of Haslam's paper—which clearly states its project and carries it out—but is an addition to it.)
I will sketch three kinds of classification that appear less objective than Haslam's examples of non-kinds before flagging some general issues that may underpin some of the worries about psychiatric classification. My purpose is thus twofold: first, to sketch an addition to the possible kinds of kinds as a contribution to understanding disagreements in the philosophy of psychiatry; second, to suggest that concerns that these kinds of kinds do not describe an objective world may themselves be mistaken. Psychiatry could contain such kinds without threatening its claim to describe reality.
Consider the two ends of the spectrum Haslam discusses. At one end are what Haslam calls "non-kinds." He identifies these with continua: "Structures that cannot be accommodated under even the most expansive account of categories can be called continua." Classifications (e.g., binary [End Page 229] distinctions) can be imposed on continua. But such classifications lack objectivity in this way. "Although a binary distinction can be imposed on such a continuum, its placement is purely arbitrary: there is no correct location where the line should be drawn, and any such line create a discontinuity that is merely artifactual." The placement does not answer to anything.
At the other end are natural kinds, which are construed essentially. "Category boundaries . . . could also arise, in principle, out of essences. It is at least conceivable that a categorically distinct psychiatric syndrome might occur only when a specific, causally efficacious pathologic process, mechanism, or structure is present." The best known essentialist accounts of natural kinds are those of Putnam (1975) and Kripke (1980), who argue that it is the real nature of samples of, say, iron or tigers that fixes the extension of the kind rather than our conceptions of them. The world, as it were, individuates itself.
Stretching between these extremes, Haslam describes kinds that enjoy intermediate degrees of objectivity or judgment independence: practical, fuzzy, and discrete kinds. What this spread does not include, however, are classifications that fail to be objective in a more radical way. But it is surely this way—described below—that is at the heart of most of the worries about psychiatric classification.
Consider first, for example, a distinction imposed on a continuum. Adult men might be classed as short if less than 5 feet 7 seven inches, of medium height from 5 feet 7 inches to 5 feet 11 inches, and tall over that height. The boundaries are arbitrary. (I am assuming that there is no practical significance to these boundaries, for example, no sudden increase in mortality rates.) Nevertheless, the classification meets at least the spirit of one very influential mark of objectivity from the C20. It can be reduced to basic physical descriptions. The reduction is not neat because the higher-level categories are disjunctive. Nevertheless...