Seligman and Johnston (1973) articulated a cognitive theory of avoidance learning. According to this viewpoint, avoidance signaling results from both emotional conditioning and cognitive information processing of a form roughly corresponding to that outlined by Tolman. Although the acquisition phase of avoidance learning undoubtedly involves the conditioning of fear-eliciting avoidance cues, according to Seligman and Johnston (1973) this emotive phase is slowly subsumed under a more cognitive one. Avoidance training depends on dogs acquiring an expectation that their behavior controls the occurrence of such aversive events. This expectancy is confirmed (negatively reinforced) whenever a dog performs the assigned task within the time frame allotted for its emission. Essentially, the dogs learn to control the incidence of aversive stimulation by responding appropriately to available avoidance cues, thereby confirming the operative expectancy underlying the avoidance behavior.
According to Mowrer's two-factor theory of avoidance learning, fear reduction is viewed as the active reinforcing substrate maintaining avoidance behavior. However, as has been noted, this scenario is inconsistent with what actually appears to occur during avoidance training. In particular, this view conflicts with the relatively anxiety-free character of behaviors acquired through such learning and their unique resistance to extinction. Both factors suggest that the dynamics maintaining avoidance acquisitions do not depend exclusively on fear reduction. According to Seligman and Johnston's cognitive theory, instead of reducing fear of impending aversive stimulation, the learned avoidance behavior is maintained because it consistently confirms an expectancy that such behavior will successfully avoid the aversive event. As additional successful avoidance trials take place, this expectancy and its reinforcing confirmation produce increasing lev els of confidence in the presence of fear-eliciting stimuli, and, as long as this expectancy is not disconfirmed by punishment (i.e., the presentation of the negative rein-forcer), the behavior will be maintained at a high operant level on the basis of confirmation alone.
in effect, the avoidance signal functions in an identical manner to that of the discriminative stimulus (Sd) during positive instrumental learning. The Sd announces a moment where a reward is forthcoming, given that the dog emits the selected behavior in a timely manner. Dogs learn over several trials to expect a reward when they respond appropriately. When this expectation is confirmed by reinforcement, the linkage between the Sd and the behavior is strengthened or stamped in and extinguished or stamped out when the expectation is disconfirmed by the omission of reinforcement. For example, if a dog was trained to sit under two different signals and then exposed to a situation in which one of the signals is followed by the omission of reinforcement while the other continues to be associated with its presentation, the dog will subsequently learn to sit under the signal confirmed by reinforcement but not sit under the signal predicting the omission of reinforcement. The instrumental response of sitting per se is not affected by this training arrangement. What is affected is the stimulus control exercised by the two signals over the emission of the sit response. in the case of positive reinforcement, learning is based on the acquisition of a promised or hoped-for outcome in the form of a reward. In avoidance training, learning is based on behavior that successfully avoids the presentation of an aversive stimulus together with the concurrent production of emotional relief or relaxation as the result of having removed (postponed or avoided) the impending threat. Both positive and negative reinforcement paradigms depend on learned expectancies based on a history of confirmatory outcomes. These paradigms of learning are usually considered as two separate ways in which learning takes place. Viewing them as two sides of a single process within a broader context of expectancy and confirmation helps to clarify the nature of learning itself, and the respective role each reinforcement paradigm plays in the learning process.
Was this article helpful?