Another theoretical account of avoidance learning that has many adherents is the safety signal hypothesis. The aforementioned experiment by Rescorla and LoLordo (1965) is frequently referred to in support of this theory. Recall that as a result of the differential conditioning of CS1 (correlated with shock) and CS2 (correlated with the absence of shock), the rate of jumping over the barrier was increased in the presence of the stimulus previously associated with shock (CS1) and depressed in the presence of the tone stimulus that had been previously conditioned to predict the absence of shock (CS2). In the presence of the CS2 or safety signal, the dogs appeared to feel more relaxed or safe even though the signal had no real relevance to the actual arrangement of the avoidance contingencies involved.
The safety-relaxation theory suggests that dogs experience stimuli associated with relief from aversive stimulation as though they were positive reinforcers. These observations are relevant to traditional dog-training methodology. in addition to establishing various conditioned associations with rewards (e.g., food and ball play), praise represents a safety signal of some importance and usefulness. interestingly, within the context of behavioral training, praise appears to derive a significant portion of its associative strength and reward value from its being paired regularly with the pleasurable relief occurring immediately after the corrective event. Because praise consistently predicts the absence of aversive stimulation and is paired with emotional relief from aversive stimulation, it gradually becomes highly desirable in itself and may be treated as a kind of conditioned positive reinforcer.
A leading proponent of the safety-relaxation theory of avoidance learning is M. Ray Denny (1971). His theory owes heavily to the stimulus-response contiguity theories of Pavlov and Guthrie. According to Denny, avoidance responding is acquired through the antagonistic dynamics of fear and relief-relaxation. Within the context of aversive situations in which fearful withdrawal or escape reactions result in the termination of the fear-eliciting stimuli involved, relief or relaxation responses are subsequently elicited that mediate approach behavior. These successive relief and relaxation responses serve to reinforce avoidance behavior. Relief and relaxation are differentiated along two primary dimensions: (1) Relief occurs shortly after the offset of the aversive stimulus and decays rapidly, whereas the onset of relaxation is both delayed and longer lasting. (2) Relief involves a strong au-tonomic factor, whereas relaxation involves striatal muscles and various motoric components. Relief begins approximately 3 to 5 seconds after the withdrawal of aversive stimulation and continues for 10 to 15 seconds. Relaxation, on the other hand, is a more sluggish response, requiring approximately 2.5 minutes to produce full benefits. Ideally, avoidance training should include conditioned safety or relief signals that are presented 2 to 5 seconds after the termination of aversive stimulation and continued for several seconds thereafter (Denny, 1976). The intertrial interval between exposures should be at least 2.5 minutes for optimally efficient avoidance training. Denny noted that "the effects of safety appear to double when both relief and relaxation, rather than one of them, are associated with a particular stimulus" (Denny, 1983). Such safety signals take on conditioned positive-reinforcing properties. Experimental support for this general idea has been reported by Weisman and Litner (1969), who demonstrated that behavior maintained on a Sidman avoidance schedule could be differentially increased or decreased by presenting a CS that had been previously associated with relief from aversive stimulation.
Not only does relaxation positively support avoidance learning, it also simultaneously results in its gradual extinction. Extinction occurs as the result of backchaining and counterconditioning effects originating in the safe, relaxed situation and generalizing step by step back to the original aversive situation.
After many trials of avoidance learning, previously feared stimuli belonging to the aver-sive situation are backchained and counter-conditioned by the relaxation and comfort associated with safety.
Tortora (1983) applied the principles of safety training to the treatment of avoidance-motivated aggression in dogs. According to his assessment, aggressive behavior commonly diagnosed as dominance related is often the result of dysfunctional avoidance responding:
The dogs in this study initially behaved as if they "expected" aversive events and that the only way to prevent these events was through aggression. The consequent reaction of the victim and the family, that is, withdrawal, turmoil, and belated punishment, confirmed the dog's "expectation" and reinforced the aggression. This positive feedback loop produced progressive escalation of the aggressive response, and the avoidance nature of the aggression presumably retarded or prevented its extinction. (1983:209)
The dogs were trained under a variety of conditions to perform 15 behaviorally "balanced" exercises or, as he calls them, operands. An important aspect of Tortora's study was the systematic pairing of a 3-sec-ond safety tone with the offset of shock delivered by an electronic collar. The training trials were spaced according to a variable interval of 5 minutes (ranging from 2 to 8 minutes), well within the 2.5-minute intertrial interval recommended by the relaxation theory. Between trials, the dogs were engaged in play. As a result of safety conditioning, the tone gradually became classically associated with relief and relaxation, becoming a conditioned positive reinforcer sufficient to strengthen cooperative prosocial behavior— behavior incompatible with aggression. According to Tortora, an important aspect of intensive avoidance and safety training is that it provides dogs with an alternative nonaggressive coping pattern when exposed to provocative or aversive situations. Tortora noted that dogs appeared to become more and more confident as they progressed through the various stages of training from avoidance to safety.
Another source of theoretical support for the safety-relaxation theory of avoidance learning comes from opponent-process theory (Solomon and Corbit, 1974), which postulates that the offset of any hedonically significant stimulus results in a recoil of opposing emotional reactions (see Chapter 6). The withdrawal of aversive stimulation evokes opposing pleasurable emotional reactions. When an aversive stimulus is terminated, the opposing pleasurable recoil provides a source of covert reinforcement, either strengthening desirable alternative behavior or inadvertently reinforcing undesirable behavior.
Following the application of aversive stimulation, it is vital that some positive behavior be selected and prompted. Applying aversive stimulation without providing dogs with an opportunity to perform some alternative option risks the possibility that an undesirable competitive pattern, like running away or avoiding the owner, might be strengthened. The somewhat common practice of isolating or ignoring dogs after punishment is counterproductive from this perspective and should be assiduously avoided. The most efficient aversive events are those that simultaneously suppress an unwanted behavior while evoking a more desirable or incompatible alternative to take its place. This arrangement is commonly used during formal obedience training where unwanted behavior is suppressed by timely correction, which in turn prompts the desired response. A well-designed correction always functions in this dual manner.
Relief may by usefully employed in conjunction with aversive counterconditioning. A common behavior problem seen among puppies and dogs involves inappropriate appetitive interests, that is, attraction to some forbidden object as a chew item. By exposing a dog to a sufficiently aversive-startling stimulus at the moment the object is approached, the dog will quickly acquire a negative conditioned association with the item (determines that it is unsafe) and avoid it in the future. Interest and approach are replaced by distrust and avoidance as a result of the startling experience. Recognizing that a corresponding degree of pleasurable relief is bound up with the event, it is advisable to present the dog with an alternative, safe chew item shortly after applying the startle. Opposing the startle response are opponent approach-appetitive recoil affects associated with relief that help to make the alternative item more attractive and desirable.
Was this article helpful?