Quantitative studies of the interaction of simple habits. I. Recovery from specific and generalized effects of extinction.

Abstract

200 white rats were given 30 food-rewarded training trials on a horizontal-bar-pressing habit and 30 on a vertical-bar habit. Half of the animals were later allowed to perform the H-bar habit unrewarded until they reached a criterion of no response in a period of 5 minutes. Recovery of the H-bar habit was measured by a second extinction test made 5.5, 25, 65 or 185 minutes after the last response, a separate group of 25 animals being used for each interval. With the remaining 100 animals, following extinction of the V-bar habit, the recovery of the H-bar habit was measured (recovery from the generalized effects of extinction). Measured by number of responses and time to reach the 5-minute criterion, recovery from the specific effects of extinction was negatively accelerated, showing 50% and 75% recovery respectively in 185 minutes. Curves of recovery from generalized effects were approximately parallel to the curves of recovery from specific effects of extinction, but showed 85% and 100% recovery in terms of number of responses and time measures respectively. Measured by mean rate of response, there was no recovery in the ordinary meaning of the term.

Spontaneous recovery varies inversely with the training-extinction interval

“…However, introducing a period of time following that extinction routinely results from the partial restoration of responding, so-called spontaneous recovery (e.g., Pavlov, 1927). A similar pattern results from instrumental learning when a response is first paired with a reinforcer and then the reinforcer is omitted (e.g., Ellson, 1938).Results such as these are commonly interpreted in an associative framework, according to which pairings of a CS and a US result in the development of an association between the internal representations of the two events. The observation of spontaneous recovery after the extinction is usually interpreted as meaning that extinction did not fully remove the original CS-US association (e.g., Bouton, 1991;Mackintosh, 1974).…”

“…However, introducing a period of time following that extinction routinely results from the partial restoration of responding, so-called spontaneous recovery (e.g., Pavlov, 1927). A similar pattern results from instrumental learning when a response is first paired with a reinforcer and then the reinforcer is omitted (e.g., Ellson, 1938).…”

Lack of reacquisition in learned taste aversions

“…It has been dernonstrated in a variety of conditioning situations that the reinstatement of a conditional response following extinction is relatively rapid. That is, a response which has been classically conditioned through repeated conditioned stimulus (CS)-unconditioned stimulus (UCS) pairings, then extinguished by CS-alone presentations, is usually reestablished by CS-UCS pairings much more quickly than if there were no prior history of CS-UCS pairings (e.g., Ellson, 1938; Pavlov, 1927, p. 59). In contrast with these findings, the results of the present experiments, using the tasteaversion situation, indicated that prior acquisitionextinction experience did not facilitate reacquisition -indeed, it profoundly retarded reacquisition.…”

MYSTERIES OF THE ORGANISM: CLARK L. HULL'S PRINCIPLES OF BEHAVIOR AND SOME PROBLEMS IN CONTEMPORARY SCHEDULE THEORY1

“…Herrnstein's matching-consistent equation (1970) and optimization models (Baum, 1981) both correctly predict a zero steady-state response rate under extinction because the rate of reinforcement delivered by the schedule is zero. Neither, however, constitutes an adequate theory of extinction apart from this, because neither is able to predict that different types of schedule will generate different amounts of responding in extinction, nor are phenomena like spontaneous recovery of responding after extinction (e.g., Boakes & Halliday, 1975;Ellson, 1938) obviously within their grasp. Computer models simulating interresponse-time reinforcement (Peele et al 1984;Wearden & Clark, 1988) fare even more poorly, because in these models all systematic changes in behavior are driven by the occurrence of reinforcers, and thus in extinction responding continues indefinitely.…”