An experimental outline for building and exploring multi-operant behavior repertoires.

Although theorists may be found in frequent controversy, experimenters differ in their approach to behavior, and data are sometimes ambiguous or subject to debate, the experimental organism is always right. His behavior is real, lawful, and always appropriate to the instantaneous conditions of his internal and external environments. It is basically the experimenter's job to gain control over those environmental conditions. In the laboratory, the experimenter emits a variety of behavior and then attempts to relate changes in his behavior with changes in the behavior of the organism under study. The ideal result of such interactions is the statement of definitive relationships which ultimately give rise to what is called "understanding of behavior." Unfortunately, however, we do not have definitive statements or relationships giving us an understanding of what behavior on the part of the experimenter most effectively generates relationships acceptable to the body of behavioral science. Although they are not observed under controlled conditions, variations in behavior from one experimenter to another, or within a given experimenter, suggest that he can state definitive relationships between himself and his organism only insofar as he is able to control and manipulate the relevant environmental conditions. Yet, only occasionally is the experimenter's primary effort to gain control and to manipulate. To do so, in fact, is often punished by other experimenters and theorists. The occasion for punishment would seem particularly strong when the gains in control are substantial and when the interaction between the experimenter and his organism does not immediately result in definitive relationships, but only suggests feasible ones in terms inadequate for conventional language and conceptual analysis. In spite of the occasional punishment for efforts primarily directed at bringing more of an organism's behavior under experimental control and subject to manipulation, we know that such efforts always set the occasion for the obtaining of definitive relationships; and, moreover, that somehow this behavior is maintained. The material to follow represents, in part, the results of several years of laboratory effort in which the pursuit of behavioral control progressively took precedence over the statement of problems and answers, and in which it was often pursued in their absence. The major result of this effort has been a demonstration that it is feasible to build, describe, and manipulate complex samples of behavior under controlled conditions, on a scale limited only by our individual laboratory behavior. It has been the argument of this section that to do so is in many ways basic to the building of a science of behavior. The following sections are concerned with: first, the nature of multioperant behavior and general problems of its establishment and analysis; second, the conceptual and notational description of multi-operant behavior; and, finally, the reporting of the laboratory story which largely generated the notions and points of view presented below.