Ben A. Williams, Ph.D. - US grants

This research will investigate the mechanisms by which learning is affected by the context in which it occurs. The focus of the investigation will be operant conditioning procedures where previous analysis of context effects have been ascribed to the molar variable of relative rate of reinforcement. An alternative account, taken from the analysis of parallel context phenomena in Pavlovian conditioning, is that such effects can be understood in terms of the notion of associative competition. This research will investigate whether such an associative approach is applicable to free-operant behavior by determining whether the critical empirical phenomena supporting an associative-competition account in Pavlovian procedures can be demonstrated in free- operant procedures as well. The proposed research will also include an analysis of several major operant phenomena, including the matching law, anticipatory behavioral contrast, and conditioned reinforcement, to determine whether they also are consistent with the notion of associative competition. The results of the research should determine whether a unified explanatory framework is possible for Pavlovian and free-operant procedures, and should provide a deeper understanding of the concept of reinforcement and how its application to applied behavior settings can be utilized most efficiently.

DESCRIPTION (adapted from applicant's abstract): Any theory of behavior requires a measure of behavioral strength. It also requires some specification of variables that determine the degree of strength. Recent research has revealed major problems for each of these issues. With respect to the measure of strength, it is now clear that the two commonly most used measures, response rate and preference, are inversely related under several circumstances. Experiments are thus proposed to determine the meaning of the different measures by a variety of convergent operations. The proposed experiments will have major implications for the theory of behavioral contrast, and the role of the context of reinforcement in determining conditioned value. With respect to the critical independent variable determining response strength, research in the past five years has shown that neither probability of reinforcement, nor rate of reinforcement, directly controls behavior. This evidence challenges the melioration theory of the matching law, and has major implications for choice theory more generally. The issue at hand is the identification of what really is the controlling variable, and the circumstances that occasion control by one variable versus another. Experiments described here provide a reexamination of the role of changeover behavior as an alternative dimension of strength, and an investigation of different theories of the determinants of the strength of changeover behavior. Such behavior is analogous to the role of search strategies in behavioral ecology. More generally, these experiments have relevance for all types of behavioral research, which must necessarily be concerned with the functionally critical independent and dependent variables.