Conditioned Aversive Stimuli

Just as a stimulus which sets the occasion for positive reinforcement (conditioned reinforcement) acquires some of the properties of positive reinforcement, stimuli which precede an aversive stimulus similarly come to acquire the properties of an aversive stimulus. An animal experiment illustrates the basic process of the development of a conditioned aversive stimulus. A rat is trained to press a lever because this behavior terminated an electric shock and leaves the rat free of electric shock for a period of time. The buzzer can now precede the electric shock, and any lever press in the presence of the buzzer now terminates the buzzer and leaves the rat free of aversive control for a period of time. Under appropriate conditions, the rat soon will terminate the buzzer whenever it appears and will rarely be exposed to the electric shock. Pressing the lever is now reinforced by the termination of the buzzer which is reinforcing because it has preceded the electric shock in the past history of the rat. Since the termination of the buzzer is reinforcing because it is followed occasionally with shock, the process is potentially unstable. The buzzer will gradually lose its conditioned aversive properties when the shock no longer follows. As the buzzer loses its conditioned aversive properties, however, and the rat fails to press the lever, the shock again occurs following the buzzer and reinstates its conditioned aversive properties. The rate at which the buzzer loses its conditioned aversive properties is a function of many conditions, including the intensity of the aversive stimulus, but in general this is a relatively slow process so that the oscillations have a long period.

Most control by aversive stimuli in human behavior tends to be by conditioned aversive stimuli the stimuli preceding the aversive events - rather than the aversive event itself. This shift to conditioned aversive stimuli tends to occur because most aversive control is fairly predictable and occurs under special circumstances which are fairly reliable in any one individual's life. Most demands or commands tend to be conditioned aversive stimuli, because they specify an aversive event which will occur if some particular behavior is not emitted. The parent makes a demand, "Pick up your toys." If the toys are not picked up, they are taken away, the child is sent to his room or he is spanked. Subsequently the child terminates such a stimulus by picking up its toys because the threat is a conditioned aversive stimulus which derives its function because it is followed by some more aversive consequence. The effectiveness of such conditioned aversive stimuli as threats depends, in part, on their proximity to the actual aversive event. With many parents the conditioned aversive stimulus is a progressive event, the form of which changes continuously until the aversive event is finally delivered. The parent, for example, will first say, "Pick up your toys," in a mild voice. If the avoidance response is not made, the demand is repeated a second or a third time. Then the intensity increases, and, finally, at a high intensity with some change in the form and quality of the voice, the aversive event is delivered if the child has not yet picked up its toys. The effectiveness of such a threat, of course, depends upon the stage at which the threat is actually followed by the aversive event. Eventually a discrimination forms, and the early threats lose their conditioned aversive properties. The early threats are never followed by the aversive event and, hence, the avoidance or escape response is postponed until later forms which are more consistently related to an aversive stimulus. The effectiveness of the threat, as with the buzzer and the rat, depends upon following it periodically with the aversive stimulus. When we observe someone who is frequently making threats with little effect in the direction of the intended control, it is likely to be because the threat is not followed by the aversive stimulus often enough. The threat, never followed by the aversive consequences, ceases to serve as an aversive stimulus. As with conditioned positive reinforcers the topography of the threat is arbitrary. "Speak softly and carry a big stick" illustrates the point. Most often threats are emitted in anger or under other strong emotional states, but this correlation is not a necessary condition for the effective function of threat. The emotional tone of most threats is probably a secondary factor related to a heightened disposition to punish under strong emotional states. A parent who is indisposed to punish a child may eventually be reinforced by producing an emotional state in himself in order to bring about conditions where he might control the child by generating avoidance or escape responses.