DESCRIPTORS: private events, emotion, hypothetical constructs, verifiability
------------------------------------------
Friman, Hayes, and Wilson (1998) provide a cogent critique of emotion, but then go on to propose that an analysis of derived relational responding provides a means for fruitfully incorporating this admittedly vague concept into behavior analysis. This proposal raises a number of questions that have long been associated with discussions of private events.
An approach focused on such putative private events as thoughts and emotions is undermined by the problem of verifiability (Zuriff, 1985, chap. 2), the same problem that led to the demise of introspection. The problem is that reports of private events cannot possibly be verified by others. And calling the private events behaviors ("Thoughts are private verbal behavior"; Friman et al., 1998, p. 148) does not solve the problem. "Thoughts" is a hypothetical construct because the term "refer[s] to unobserved events within the organism" (Zuriff, 1985, p. 72). If it is countered that thoughts are observed by the person experiencing them, we are still faced with the problem of verifiability.
This problem is particularly salient in such applied domains as clinical work. What independent measure allows the clinician to assert that "feeling anxious is a behavioral event" (Friman et al., 1998, p. 140) that can be a predictor? If a child (per their example) is crying, complains of feeling anxious, and then runs away, these are, indeed, behavioral events, but to label the set of events as "feeling anxious" does not mean that we are dealing with a feeling, a concept that the authors have earlier so effectively undermined. How does the clinician know whether the client is being truthful when he or she asserts that he or she is too anxious to go to work, has obsessive thoughts, feels depressed? The clinician must rely on the client's relevant public behavior, much of which will be verbal self-report. But "first-person reports are not incorrigible guides to covert events, and inferences are drawn by the experimenter ;obor clinician;cb, not the subject" (Zuriff, 1985, p. 234).
Furthermore, "Skinner's theory implies that the connection between a private event and a verbal response will not usually be a reliable one. He therefore does not trust first-person reports of private events and does not use them as observation reports" (Zuriff, 1985, p. 28).
This criticism does not deny the existence and importance of private events. Rather, it questions the value of referring to undefined internal processes that can be observed by others indirectly, if at all. Even if the day were to come when various categories of self-reports were found to be highly correlated with various patterns of neurophysiological and endocrine activity, it is difficult to imagine clinicians routinely obtaining such information. Furthermore, the private would then be public; the clinician would no longer be dealing with private events.
Friman et al. (1998) are concerned with the ordinary-language term anxiety, which they readily admit is vague and for which they offer no technical definition. Serious problems arise, however, when ordinary-language terms, such as those used to describe various emotions, are incorporated into behavioral science (Chiesa, 1994, pp. 24, 25).
A focus on such hypothetical constructs as thoughts entails threats to the empirical nature and objectivity of behavior analysis. This is because "hypothetical constructs are associated with 'existence propositions' which assign `surplus meaning' over and above observed correlations" (Zuriff, 1985, p. 73). The anchoring of such hypothetical constructs to the behavioral data language can be seriously questioned.
REFERENCES
Chiesa, M. (1994). Radical behaviorism: The philosophy and the science. Boston: Authors Cooperative.
Friman, P. C., Hayes, S. C., & Wilson, K. G. (1998). Why behavior analysts should study emotion: The example of anxiety. Journal of Applied Behavior Analysis, 31, 137--156.
Zuriff, G. E. (1985). Behaviorism: A conceptual reconstruction. New York: Columbia University Press.
Received April 13, 1998 Final acceptance June 15, 1998 Action Editor, David P. Wacker
Correspondence concerning this article should be addressed to P. A. Lamal, Department of Psychology, University of North Carolina at Charlotte, Charlotte, North Carolina 28223-0001
(E-mail: palamal@email.uncc.edu).