Stimulus Discriminability, Contingency Discriminability, and Complex Stimulus Control
Michael C. Davison University of Auckland, New Zealand
In the development of the Experimental Analysis of Behavior, a great deal of effort has been expended on two historically separate questions. One question has been how the consequences of behavior affect subsequent behavior -- the study of reinforcement control. The other has been how antecedent stimuli that signal differing contingencies affect behavior -- the study of stimulus control. Studies of reinforcer control have generally provided either no explicit antecedent stimuli, or have provided antecedent stimuli that are assumed to be highly distinctive, and have studied parametric variations in reinforcer frequencies. Studies of stimulus control have generally arranged maximally differing reinforcer conditions (e.g., multiple variable-interval extinction) and tested the control by small differences in exteroceptive stimuli. While this characterization by no means embraces all the research literature in these areas, it is sufficiently general to highlight the absence of research specifically designed to quantify the joint effects of stimulus and reinforcer control on behavior.
Within the reinforcer-control literature, one area of endeavor has been the search for an empirical description of the ways in which changing reinforcer parameters (such as rates, delays, and amounts) affect the relative allocation of behavior between two response alternatives. The empirically based generalized matching law ( Baum, 1974), was found to be a good description of relative behavior allocation in many situations, and it provided a reasonably wide domain of parameter invariance. The relation is written: