Evolutionary versus prototype analyses of the concept of disorder.

The harmful dysfunction (HD) analysis of the concept of disorder (J. C. Wakefield, 1992a) holds that disorders are harmful failures of internal mechanisms to perform their naturally selected functions. S. O. Lilienfeld and L. Marino (1995) proposed instead that disorder is a Roschian prototype concept without defining properties. Against the HD analysis, they argued that many disorders are not failures of naturally selected functions because they are either designed reactions (e.g., fever) or failures of functions that are not naturally selected (e.g., reading disorder). The HD analysis is defended here against these and other objections and compared with the Roschian account. It is argued that the objections are based on conceptual confusions and can be turned around to provide strong new support for the HD analysis. A series of conceptual experiments demonstrates the superior explanatory power of the HD analysis and disconfirms the Roschian account.