Pavlovian contingencies and anticipatory contrast

Abstract

Pigeons were trained on a four-component multiple schedule in which two target components with identical reinforcement schedules were followed by other components with either higher or lower reinforcement rates. The Pavlovian signal properties of the target-component stimuli were varied by changes in their duration relative to the following components, and by whether the two following components were cued by the same or different stimuli. When different stimuli occurred in the following components, response rates were higher in the target component preceding the following component with the lower reinforcement rate, and these contrast effects were larger with shorter relative durations. But with nondifferential stimuli in the following components, contrast consistently occurred only with the longer durations of the target components. Moreover, several subjects with the shorter duration target stimuli had higher response rates in the target followed by the richer schedule-that is, Pavlovian conditioning occurred to the target stimuli. This interaction suggests that the processes underlying anticipatory contrast and Pavlovian conditioning are in opposition, and that the Pavlovian effect can dominate ifthe signaling properties of the target components are sufficiently enhanced.Numerous experiments have demonstrated that a major portion of behavioral contrast is due to the following schedule (Williams, 1976(Williams, , 1979(Williams, , 1981Williams & Wixted, 1986;Wilton & Gay, 1969). The typical procedure used in such experiments has been for two target components, associated with different stimuli but similar reinforcement schedules, to be followed by components with differential reinforcement schedules. Thus, one target component leads to a richer reinforcement schedule; the second leads to a leaner reinforcement schedule. "Anticipatory contrast" has been the result, as response rates in the target components are inversely related to the following rate of reinforcement.The occurrence of anticipatory contrast is paradoxical when viewed from the perspective of Pavlovian conditioning. In order for differential behavior to occur during the target components, their keylight signals must "predict" the following conditions of reinforcement. But such a predictive relation has been used in other procedures to generate autoshaping (Brown & Jenkins, 1968). That is, the keylight that predicts the following higher rate of reinforcement should elicit higher response rates, which should result in the opposite of the contrast effects actually obtained.A major difference between the Pavlovian contingencies embedded in the anticipatory-contrast procedure and the usual autoshaping procedure is that the target stimuli in the former procedure do not lead directly to food, but lead Hassin, and Goldhammer (1982), using a procedure quite similar to that for anticipatory contrast. In their study, different colored keylights were presented for 8-sec durations, and were then followed by 30 sec of a white keylight. No reinforcement was avai...

Behavioral contrast redux

“…For the fourth condition in Figure 5 and in similar experiments (Williams, 1979, Experiment 2, and1990b), the stimulus present in both following components was high in value, as is shown by high response rates whenever it was present. When the schedules in the following components were signaled VI versus extinction, the stimulus present most of the time was that correlated with extinction,and therefore, very little responding occurred.…”

“…As a result, the reversal of response rates in the target components seen in the second segment of Figure 4 eventually disappears. Most typically, results for individualsubjects are highly variable (see Williams, 1990b, 1992b.…”

Competing sources of stimulus value in anticipatory contrast

“…The strongest effects are seen when the schedules in the target components are extinction and the different following schedules are associated with the same discriminative stimulus (Brown, Hemmes, Coleman, Hassin, & Goldhammer, 1982;Williams, 1992, Experiment 1): The target stimulus preceding the food reinforcement schedule engenders a great deal of responding, whereas the target that precedes extinction does not. Pavlovian effects are also seen when the target components have food schedules of their own, but with their different following schedules associated with the same stimulus (Williams, 1979, Experiments 2 and 3; Williams, 1990), although these effects are often variable across individual subjects and disappear with continued training. In such procedures, the Pavlovian effect (a higher rate to the target component followed by the richer schedule) is replaced by the anticipatory contrast pattern of behavior (Williams, 1992, Experiment 1) when the different reinforcement schedules in the two following components are discriminated conditional on which target stimulus preceded their presentation.…”

Behavioral Contrast: Pavlovian Effects and Anticipatory Contrast

“…This effect is termed anticipatory contrast. Williams (1990) addressed the apparent conflict between anticipatory contrast and the data reported by Brown et al by focusing on procedural differences between the relevant studies. Studies showing anticipatory contrast have usually employed multiple schedules in which all components were equal in duration, and all were differentially cued.…”

“…These conditions establish a strong S_SR contingency between a target-component cue and reinforcement in the following component; that is, color of a short duration targetcomponent cue is the only reliable predictor of reinforcement in the succeeding component. Williams (1990) systematically examined the effects of these procedural differences in a 2 x 2 factorial design. One factor was targetcomponent duration (10 or 30 s) relative to following-component duration (always 30 s).…”