Leigland, S. (1996). The functional analysis of psychological terms: In defense of a research program. The Analysis of Verbal Behavior, 13(1), 105-122. doi:10.1007/BF03392909
[Abstract]In 1945, B. F. Skinner outlined a proposal that psychological or mentalistic terms found in natural language might be analyzed empirically in terms of the variables, conditions, and contingencies of which they may be observed to be a function. Such an analysis would enable discriminations to be made between different classes of variables that enter into the control of the term. In this way, the analysis would clarify what is traditionally called the "meanings" of such terms as they occur as properties of verbal behavior. Despite his expressed confidence in the success of such a program, Skinner largely abandoned the functional analysis of psychological terms in favor of the development of a promising new field; the experimental analysis of behavior. The present paper argues that the original program is of great importance as well, and for the following reasons: (a) to make full, immediate, and (most importantly) effective contact with the range of issues and terms of central importance to the traditionally and culturally important concepts of "mind" and "mental life" (and thereby demonstrating the relevance of radical behaviorism to the full range of human and verbal behavior); and (b) to extend the methodology of the functional analysis of verbal behavior more generally. Such a research program would demonstrate, through an empirically-based scientific analysis, that the philosophical problems concerning "mental life" may be productively analyzed as problems of verbal behavior. Issues of methodology are discussed, and possible methodological strategies are proposed regarding the confirmation of behavior analytic interpretations of mentalistic terms.
[Citing Place (1993c)]  
Citing Place (1993c) in context (citations start with an asterisk *):
Subsection Complicating Factors
* To be sure, the functional analysis of psychological terms is a research program which faces a variety of complicating factors, Skinner's (1945) confidence notwithstanding. Three such factors will be noted briefly. First, there is the role of private events in the control of verbal behavior (e.g., Skinner, 1953, 1957). To the extent that the terms of interest could be characterized as "mentalistic" or "subjective," such events could be expected to be a factor in a functional analysis. Although the role of such events would provide a methodological challenge to a functional analysis of such terms, the problem should not be insurmountable since the relations between private events and verbal behavior is presumed to be established and maintained through contingencies as well (e.g., Day, 1976a/ 1992, 1983/1992; Moore, 1980, 1990, 1995; Place, 1993; Skinner, 1945).