Interrelations Among The Processes

of Aversive Control The processes involved in control by aversive stimuli are all interrelated. When behavior is influenced by the occurrence of an aversive event by one process, such as punishment, other processes of aversive control, the effect of a preaversive stimulus, for example, will also operate. An aversive stimulus may have a potential range of effects via all of the behavioral processes which have been discussed so far. The control of behavior by aversive stimuli is also closely related to the processes by which behavior is maintained by positive reinforcement. This is especially true in the case of punishment which is, by definition, the application of an aversive event to a form of behavior maintained by some other process. Whatever the role of the aversive stimulus, however, it still will have effects on the positively maintained repertoire of the organism. Conversely, the avoidance and escape behavior generated by aversive stimuli must be considered in the context of its prepotency over concurrently maintained reinforced behavior.

Escape And Avoidance

Whenever an aversive stimulus occurs repeatedly, elements of the environment acquire properties of a preaversive stimulus, potentially having the broad emotional effects on behaviors which were discussed above. Such a repeated occurrence of an aversive stimulus may be seen where escape and avoidance are used to control behavior. A rat whose bar presses postpone an electric shock will be repeatedly shocked whenever a bar press fails to occur. The apparatus will be correlated specifically with the delivery of the shock as will the class of behaviors going on when the rat is not pressing the bar. The parent who punishes the child for not picking up its toys is attempting to strengthen the behavior of putting away toys by arranging an avoidance schedule. Putting away toys postpones or terminates an aversive stimulus. In the process of establishing the avoidance behavior, however, the toys become a preaversive stimulus because of their close correlation with an aversive event. Even though the behavioral control procedures were designed to strengthen specific responses through avoidance there was inevitably a concomitant effect, in the broad emotional influence of conditioned aversive stimuli. Even when an aversive stimulus occurs without any explicit attempt to strengthen behavior by escape or avoidance, the general circumstances in which the aversive stimuli occur serve as a preaversive stimulus which will reinforce escape and avoidance behavior, that is, moving out of the situation in which the conditioned aversive stimuli are present, or interfere with some of the ongoing behavior of the organism.

The Role Of The Baseline Performancea Preaversive Stimulus

Experiments which have studied the factors that influence the suppressive effect of a preaversive stimulus in animals give information as to what determines how disruptive a preaversive stimulus may be. A preaversive stimulus will have a less prolonged effect when the ongoing behavior of the animal is very strongly maintained than when it is maintained under less optimal conditions of positive reinforcement (Brady, 1955). When the buzzer is no longer followed by a shock, the amount of time required before the animal behaves normally in the presence of the buzzer is much longer where positive reinforcement is infrequent than when the animal is producing food on a very optimal schedule of reinforcement such as a small fixed ratio. When the baseline performance is strongly maintained, the duration of the preaversive stimulus is a crucial factor in how much disruption the pre-shock stimulus produces. The preaversive stimulus will suppress the performance only if the temporary disturbances in the baseline performance do not reduce the animal's net take. As the duration of the preaversive stimulus increases, however, so that the suppression leads to a significant loss of the total possible reinforcements available to the animal, then the performance occurs despite the previous disruptions of the preaversive stimulus (Stein et al., 1958). These observations show that the suppressive effect of an aversive stimulus and the general conditions of positive reinforcement maintaining the organism's repertoire interact in a complex way. It is necessary, therefore, to take into account many over-all conditions affecting the strength of the individual's behavior in addition to the presence or absence of aversive control in order to evaluate the disrupting effects of preaversive stimuli. The differences among individuals who exhibit "sexually inhibited" behavior, for example, are probably not as much due to a differential history of punishment and aversive control as they are to differences in how strongly maintained the behavior is by positive reinforcement. Given the wide variety of operant performances involved in dealing with the opposite sex, and, as a result, the differences from individual to individual in the general strength of their sexual behaviors, there would be a differential sensitivity to punishment and other aspects of aversive control. A special history in which the required successive approximation of the complex repertoire did not occur would produce the indicated weakness in the behavior. Similar factors probably govern the wide range of effects in human behavior resulting from apparently similar magnitudes of aversive control.

A preaversive stimulus does not always suppress the ongoing behavior of the organism. Animal experiments have allowed us to analyze situations in which a preaversive stimulus sometimes leads to an increased rate of responding, rather than the usual depressive effect as in many cases of clinical "anxiety." These increased rates of responding in the presence of a preaversive stimulus occurred when the baseline schedule of reinforcement was maintained on an avoidance schedule. The discovery of this phenomenon is aptly described by the psychologist who originally made the discovery (Sidman, 1960).

We first conditioned the monkeys to press a lever by the simple expedient of giving them a brief shock whenever 20 seconds elapsed without a lever depression. Each time they pressed the lever they postponed the shock for 20 seconds. After the animals had settled down to a relatively stable rate of avoidance responding we introduced the clicker and unavoidable shock sequence [the warning stimulus], using the earlier schedule of 5-minute periods of silence.

The immediate result was that the animals pressed the lever at approximately three times their normal rate, both when the clicker was on and when it was silent. In fact, they responded sufficiently often to avoid all avoidable shocks; the only shocks they received were the unavoidable ones. The monkeys then gradually slowed down to their normal rate of lever pressing. But they returned to their normal rates more rapidly when the clicker was silent than when it was sounding. There was, therefore, an intermediate phase in which they pressed the lever at a higher rate during the clicking periods than during the periods of silence. This reversal of the Estes-Skinner observation [the usual effect of a preaversive stimulus] caught our interest.

We eliminated the avoidable shocks [the base line avoidable schedule] but continued to administer the unavoidable ones. The monkeys ceased lever pressing, as was to have been expected, during the periods of silence [extinction of the avoidance behavior]. But for a long time they persisted in lever pressing during the clicking periods . . . the animal practically never responds during periods of silence or during the initial minutes of the clicking periods. But as the time approaches for shock, the monkey begins to press the lever rapidly and continues until it receives the shock. Immediately after the shock, it again ceases pressing, and another cycle begins. This phenomenon is called conditioned facilitation.

Does conditioned facilitation during the clicking period represent a breakdown of the lawfulness to which we have become accustomed in our experience with the Estes-Skinner technique? From an adaptive point of view, the facilitation of lever pressing makes no more sense than does suppression. The shock is inevitable, and the animal's high response rate during the stimulus represents only so much wasted energy. It would take very little stretching of the imagination to class this behavior as pathological. Yet, as we shall see, it results from normal processes at work in a slightly unusual setting.

When an animal that is pressing a lever for food is first exposed to the clicker-shock sequence it may initially cease pressing both when the clicker is on and when it is silent, even though it receives shock only while the clicker is on. This may be thought of as a generalized effect of unavoidable shock. A corresponding generalized effect, an over-all increase in response rate, is initially observed when the lever pressing has served to postpone shock. In their first stages, then, the two effects are opposite in direction but similar, perhaps, in origin.

With this second effect of a preaversive stimulus, we begin to see some of the ways in which a functional analysis of behavior is beginning to unravel some of the complexities of human behavioral phenomena. "Anxiety" very clearly has diverse effects from individual to individual and in various situations, but a common denominator is emerging in describing the major effects as the emotional effects of a preaversive stimulus. The apparently paradoxical observation of a fearful situation, provoking in one individual depression and in another, excitement and general increases in performance, now becomes understandable when we can specify what performances are in strength in the individual at the time he is exposed to the preaversive stimulus. A pre-aversive stimulus will produce increased activity in an individual when the baseline performances that are observed are avoidance responses. A "compulsive talker," who avoids difficulty in social situations by a steady stream of conversation which prevents the conversation from turning to anxiety-provoking topics, will react to a preaversive stimulus by increasing his rate of verbal behavior. There is, in fact, a suppression of verbal behavior which is not immediately apparent but the suppression is in the area of verbal forms maintained by a positive effect on the listener. The high rates of activity from behaviors previously acquired by the individual as avoidance responses can occur in performances of almost any form, for example, tics, telling long anecdotes, sustained professional activity, busy work, occupation with routine tasks, or compulsive gestures such as body manipulation and impulsive head washing.