The Analysis of Verbal Behavior

209 When one is in a familiar context, and in the presence of a familiar audience, many verbal operants may be strong concurrently, for each feature of such an environment, by itself, is likely to have pervasive evocative effects on verbal behavior. The combined effect of many such features is no doubt incalculably complex. In relation to this welter of interacting variables, verbal behavior, as emitted, is remarkably coherent. Sometimes one’s utterances display slips or blends, but more commonly, only one verbal operant is emitted at a time. Despite the unity implied by the dominance of one verbal operant over all others at any moment, that operant is likely to have been potentiated by more than one variable. No matter how much we would like a glass of water, we are not likely to ask for one in the absence of an audience or in the absence of a source of water. One can plausibly assume that all verbal behavior, to varying degrees, is evoked by concurrent variables. Sometimes the form of a verbal response betrays its multiple sources of control. When we read that Representative Ball dismissed the president’s recent bounce in the polls, we suspect that bounce was emitted in preference to jump or rise because of the intraverbal control from Ball. Similarly, the statement that a track star jumped at the chance to compete suggests at least two sources of control of the metaphor. When writing, I often find myself repeating distinctive response forms when many adequate synonyms are available, indicating that the form is under both thematic and echoic control. For example, in the paragraph above, I originally wrote the phrase at any time on two successive lines. Since such repetition of a conspicuous phrase “clangs in the ear,” I changed the second to at any moment. The “clang” is a distraction to the reader for the very reason that one source of control, the echoic, is irrelevant to the theme, and consequently, writers usually try to avoid iterations of distinctive phrases. Skinner (1957) devoted most of Chapter 9 of Verbal Behavior to similar examples culled from his own experience, and doubtless the reader can supply many more. The concept of joint control—the confluence of two controlling variables on a single response form—is therefore not new. Rather, it is a special case of the nearly ubiquitous phenomenon of multiple control, which has been discussed at length by Skinner. What is new is an appreciation of the role that joint control plays in complex behavior. Specifically, Lowenkron has proposed that the onset of joint control can be a discriminable event; moreover, that event can serve as a controlling variable in novel delayed matching-to-sample performance and other complex behavior. (See Lowenkron, 1998, for a comprehensive account.) In the light of my comments above that multiple control is the nearly universal rule rather than the exception, it might be hard to see how joint control could be of interest. However, important cases arise when we confine our consideration to those examples of joint control in which each of two stimuli can plausibly be assumed to exert control over a single dominant response rather than over broad classes of responses of roughly equal strength. Address correspondence to: David C. Palmer, Department of Psychology, Smith College Northampton, MA 01063; e-mail: [email protected]. Joint Control: A Discussion of Recent Research