IS THE OPERANT CONTINGENCY ENOUGH FOR A SCIENCE OF PURPOSIVE BEHAVIOR?

Behavior and Philosophy, 2004 by Timberlake, William

ABSTRACT: The operant contingency remains the most powerful and flexible single technology for the production and control of purposive behavior. The immediate aim of this paper is to examine the conceptual and empirical adequacy of the operant contingency as the basis for a science of purposive behavior. My longer-term goal is to improve the effectiveness of operant contingencies and our understanding of how and why they work. I explore three aspects of the operant contingency: its development as a closed definitional system, its empirical adequacy as a technology, and the appropriateness and usefulness of related theoretical assumptions. I conclude that the efficacy of the operant contingency can be improved further by continued analysis of its implementation, mechanisms, and assumptions and by increasing its links to other approaches and concepts.

Key words: Operant contingency, discriminative stimulus, operant, reinforcer, purposive behavior, definitional system, proper function, niche-related mechanisms, tuning, causal stance

The operant contingency is unquestionably one of the most important technologies for the production and control of behavior to emerge from the twentieth century. B.F. Skinner (1938) developed the operant contingency to produce and define purposive behavior (in contrast to phenomena such as reflexes, instinctive behavior, habituation, and reflex conditioning). As Skinner pointed out in an introduction to the reprinting of The Behavior of Organisms, "operant behavior is essentially the field of purpose" (Skinner, 1966, p. xi).

Skinner and his students led the way in showing that operant contingencies could be applied in both laboratory and field contexts, to both human and nonhuman animals, and to responses ranging from lever pressing to verbal behavior. With the addition of more complex technologies such as schedules of reinforcement, chaining, and shaping, operant contingencies have been applied in a variety of additional areas, including training of motor and social skills, economic behavior, teaching, and the assessment of drug effects. Most recently, the operant contingency has been the centerpiece of attempts to combine operant learning, evolution, and culture in the common framework of selection by consequences (e.g., Hull, Langman, & Glenn, 2001; Skinner, 1966).

What is an operant contingency? It is an experimenter-imposed relation among three codefmed concepts (a discriminative stimulus, an operant response, and a reinforcer) that connects them to each other and to an accompanying orderly change in responding. This paper focuses on the conceptual and empirical adequacy of the operant contingency to serve as the basis of a science of purposive behavior. Such a question might seem superfluous given the notable successes of operant contingencies in producing and controlling behavior, but operant conditioning is frequently not as effective as we might expect of a mature technology celebrating its sixty-fifth birthday. It is certainly not as simple to use as textbook descriptions imply or as automatic in its effects as experimenters, practitioners, teachers, and parents would prefer.

Further, the achievement of a successful operant contingency is not the only important goal in the production and control of behavior. Other important aims include increased efficiency of implementing successful contingencies, predicting successful contingencies, understanding the reasons for failure and limitations of operant control, predicting asymptotic performance, and predicting and promoting better conceptual ties between operant conditioning and other approaches to the study of behavior. These goals are not equally facilitated by current procedures and practice.

The long-range purpose of this paper is to improve further our ability to use operant contingencies in the control and analysis of purposive behavior and, as part of this process, to add to our understanding of how these contingencies work at an empirical and theoretical level. My immediate purpose is to consider three aspects of the operant contingency: some ramifications of how Skinner defined it, resultant advantages and limitations on its empirical usefulness, and the importance of considering more carefully the theoretical concepts and stances that have increasingly surrounded its use.

The Operant Contingency as a Definitional System

The operant contingency is central to the field of operant psychology because it serves as the ultimate basis for defining each of the concepts it relates: namely, discriminative stimuli, operants, reinforcers, and reinforcement effects (Skinner, 1938). In other words, the elements of an operant contingency are considered to be present (and can ultimately be identified) only as part of a contingency that produces an orderly change in responding. Thus, the elements of an operant contingency form a closed system in which they are codefmed (defined in terms of each other). For example, in the case of a rat in an experimental chamber, a flashing light qualifies as a discriminative stimulus only if it is a part of a contingency in which the presence of the light results in an orderly change in the lever pressing followed by delivery of a food pellet reinforcer. In a similar vein, neither is a lever press an operant nor the food pellet a reinforcer unless they are related by an operant contingency that produces an orderly change in the rate of responding in the presence of the flashing light (Skinner, 1938).