O. Hobart Mowrer (1960) found that many phenomena observed during avoidance acquisition and extinction could not be adequately explained by previous theories of learning. For instance, Thorndike's law of effect postulates that behavior followed by reward is stamped in, whereas behavior followed by punishment is stamped out. The effects of punishment, however, are often more complicated than Thorndike's assessment indicates. Mowrer argued that two direct features are added to the training situation as a consequence of punishment: (1) punishment does not simply suppress an ongoing behavior, it also strengthens behavior directly associated with its termination, and (2) antecedent stimuli and cues occurring prior to the onset of punishment become emotionally conditioned with fear.
Startle devices like the shaker can are commonly used in dog training. Their purpose is to generate a startle effect immediately and directly associated with an unwanted behavior. For example, a common behavior complaint presented for modification is the habit of jumping on countertops in search of food. Since it is intermittently reinforced (sometimes with very large rewards), it can be a very persistent habit and resist suppression. One means of suppressing the tendency is to booby trap the countertop with a suspended shaker can or something else that causes a significant startle. Sometimes, forbidden items themselves are attached directly to the shaker can by a length of fishing line or dental floss. Dogs that attempt to steal a snack are very much startled by the resounding crash caused by their efforts.
Several things happen during such training. Both external cues (the tabletop and tempting food), internal cues (the desire to jump up for food), and the behavior of jumping itself are all associatively linked with the startling event. Also, dogs learn to escape the startling event by excitedly leaping away from the table (if the startle is sufficiently intense, and they are sufficiently sensitive to it). After a few trials, dogs learn to stay clear of tabletops (unless the potential reward of jumping up offsets the threat of punishment). In the foregoing case, a dog's tendency to jump up is conflicted by competing conditioned emotional responses (CERs) and two learned instrumental components: the tendency to jump off has been strengthened by successful escape and not jumping up rewarded with continued safety from the feared event.
In the laboratory, animals are often trained to jump over a barrier dividing the experimental chamber into two identical compartments. The grid floor is attached to a shock generator. This arrangement is called a shuttle box and is commonly used in the study of punishment and escape-avoidance learning. During the escape phase of training, an electric charge is passed into the grid work of the floor. Animals learn by trial and error to escape the shock by jumping over the barrier into the safety of the other side. After several trials, they learn to escape more and more efficiently by jumping over the barrier as soon as the shock occurs. The avoidance phase of training involves pairing a neutral stimulus (e.g., a light or tone) with the delivery of shock. If a tone is presented just prior to the onset of shock, animals quickly learn to anticipate the occurrence of shock and avoid it by jumping over the barrier as soon as the auditory cue is heard. In the beginning this association may need periodic reinforcement, but as training progresses the animals respond almost without error. Once established, avoidance training is extremely resistant to extinction. Solomon and Wynne (1953) found that dogs trained to avoid traumatic shock under such conditions persisted in the habit long after the threat of shock was eliminated. Resistance to extinction is a peculiar feature of avoidance learning—a feature it shares with learned fears and phobias.
Although the foregoing scenario sounds straightforward enough, several perplexing aspects about avoidance learning prompted Mowrer's attention. One theoretical issue is how avoidance learning is maintained, since the avoidance response is rarely reinforced with shock. Mower proposed a two-factor theory of avoidance learning to account for it. His theory is composed of two distinct parts: a Pavlovian component involving conditioned emotional reactions and a Thorndikian component involving habit formation. The tone in the foregoing arrangement possesses no aversive or fear-eliciting properties until it is classically associated with shock. The tone signal gradually acquires motivational properties originally belonging only to shock itself. Consequently, the tone becomes a CS eliciting various fearful emotional responses and concomitant physiologi cal changes like accelerated heart rate and respiration. Mower theorized that animals find such emotional reactions aversive and learn to escape them in precisely the same way they learn to escape direct aversive stimulation— negative reinforcement. Since jumping the barrier reduces an aversive tension generated by the CS, the response is negatively reinforced. An experiment that is often cited in support of this view of avoidance learning was carried out by Kamin (1956), who found that if the CS was continued beyond the emission of the avoidance response, avoidance learning would be disrupted—that is, the extended CS punished the avoidance response. If the termination of the CS were delayed for as long as 10 seconds, avoidance learning was seriously impeded.
Subsequently, Rescorla and LoLordo (1965) performed a series of experiments that provided additional support for the two-factor theory of avoidance learning. In their study, dogs were trained to jump over a barrier without the aid of external avoidance cues (Sidman avoidance task). During this initial avoidance training, the dogs were exposed to regularly spaced shocks that they could avoid with well-timed responding. Once a strong pattern of avoidance responding was established, they were exposed to a classical conditioning procedure. Some of the dogs were presented a tone stimulus (CS1) that was regularly followed by shock after a variable delay. Another group of dogs was exposed to the same CS1, but instead of receiving shock, they were exposed to another tone stimulus (CS2)—a stimulus that was never followed by shock. The dogs' differential rate of avoidance responding in the presence of each CS arrangement was then measured. Dogs exposed to CS1 and shock were significantly more active avoidance responders. Their rate of jumping over the barrier was significantly increased whenever the tone stimulus was turned on. The other group in which CS1 was followed by another tone (but never shock) made fewer avoidance responses. The first preparation (CS1-shock) augmented avoidance responding while the latter (CS1-CS2) depressed such responding. In a sense, the dogs were less worried about the occurrence of shock in the presence of the CS1-CS2 arrangement. CS1 followed by CS2 predicts the absence of shock—that is, it is a "safety signal." These experiments demonstrate that some variable emotional factor alleviates or potentiates avoidance responding. In the presence of inhibitory CS1-CS2 (predicting the absence of shock) avoidance responding decreases, whereas in the presence of the excitatory CS1-shock (predicting the presence of shock) avoidance responding increases.
Mowrer progressively refined his analysis of avoidance learning and gradually modified his theoretical interpretation of two-factor learning. The bulk of these changes leaned in the direction of the cognitive learning theory of Tolman (1934). The two parts of his avoidance paradigm, corresponding to classical and instrument conditioning, were referred to, respectively, as sign learning (or the what to escape) and solution learning (or the how to escape). He viewed two-factor learning theory as a creative synthesis bridging traditional views of learning with Tolman's cognitive viewpoint:
Reflexology (used here to include Thorndikian habit theory as well as Pavlovian conditioning) and cognition are, in some ways, poles apart— one being behavioristic and the other mentalis-tic—but two-factor theory represents an effort to bring about a creative synthesis thereof. We discard the notion that behavior itself is learned, whether as habit or as conditioned reflex; but we retain the concept of conditioning and, with Tolman, use it to explain how certain internal events get attached to new (extrinsic or intrinsic) stimuli. But whereas Tolman identified these internal events as "pure cognition," we see them, simply but more dynamically, as hopes and fears. And these then guide, select, or control behavior along lines which are, generally speaking, adaptive—a phenomenon which both Thorndike and Pavlov, in their different but equally oversimplified ways were also attempting to account for. (Mowrer,
The expectancy theory of avoidance has received a great deal of scientific interest, with many experiments having been carried out to determine the relative contribution and importance of emotional conditioning versus cognitive information in the formation of avoidance signals.
Was this article helpful?
Discover The Secrets To Successfully Adopting A Dog! Never Before Revealed Information! You love dogs-- and you want a dog! Many people feel exactly the way you do! This is certainly no surprise, because dogs are wonderful animals and they make excellent pets. A dog can bring lots of fun, friendship, and joy to your life!