Abstract
The present report reviewed a short history of studies on variability in response and response sequence. Early research focused on the variability of a single response and found that variability in response topography may be a natural outcome of innate adaptive behavioral mechanisms. Subsequent studies shifted the research aim to produce variability in the response sequence by differential reinforcement. Early research failed to find enhanced variability in the response sequence. Rather, some studies showed a stereotyped response sequence and suspected that reinforcement created the stereotyped sequence. However, later studies proved that modification of training procedures and introduction of “Lag n” schedules improved variability in the response sequence. Two explanations were proposed. One was based on random selection of manipulanda and the other, on response selection probabilities. However, recent studies revealed shortcomings in both of them. The discussion addresses the need to form a new explanation to account for the findings. It also suggests some implications for future research including the use of variability measures to test higher-order cognitive processes.
