On the Failure of Operationism

Abstract

Operationism was proposed as the methodology for a behavioristic psychology in the 1930s, based on ideas from Bridgman's 1927 book, The Logic of Modern Physics. That the practice of operational definitions still persists in psychology has been strongly criticized. The major claim, first made in the 1940s and echoed recently, has been that psychologists misunderstood Bridgman by using operational definitions to produce or elicit new concepts or phenomena rather than to measure existing ones. However, the distinction between measuring and producing phenomena cannot be maintained in physics, and operational definitions to produce phenomena are useful in the early stages of scientific inquiry. In fact, where psychologists of the 1930s (e.g. Stevens, E.C. Tolman, Hull) failed to heed Bridgman was in their uncritical use of mechanistic explanations, and, especially, in their belief that unobservable constructs were adequately defined on the basis of a single set of operations. This latter error was corrected in the 1950s with the adoption of a `convergent operationism', which provided the methodological basis for the new cognitive psychology as well as the foundational treatment of reliability and validity in psychometrics. Thus, it is incorrect to assume that the failure of operationism, as practiced in the 1930s, is symptomatic of a deeper failure of positivism as a philosophical orientation for psychology. A convergent operationism provides the only means of making potentially valid inferences about unobservable phenomena from empirical data, and will continue to be employed by psychologists as long as it is useful to do so.