Pavlov’s pioneering work on how dogs come to anticipate the presentation of food has proliferated into an extensive body of research exploring the formation of associations between stimuli (Rescorla, 1988). A major focus has been the condi-tioning of emotional reactions such as fear. Watson and Raynor (1920) demon-strated the ease with which a child can learn to fear an object. ‘Albert’, a 9-month-old boy, initially showed no fear of a white rat, in fact he wanted to touch it, but he did show distinct displeasure upon hearing the loud sound of a hammer hitting a steel bar. The researchers then went about presenting the rat, followed by the sound. After just two repetitions, Albert was reluctant to touch the rat. With just a few more repetitions, Albert cried and tried to move away when-ever he saw the rat. This is a classic example of a ‘conditioned emotional response’.
Emotional responses, such as excitement, anxiety, fear and aggression, are often triggered by specific stimuli. Envision the dog that gets wildly excited when its owner picks up the leash, anticipating a walk in the park. Or the dog that begins to pace and whine when its owner prepares to leave for the workday, anxious at the prospect of being left alone. Or the dog that hides under the bed when the owner picks up the nail clippers, fearing a painful grooming episode.
Conditioned emotional responses not only underlie many serious behavioural
problems, they can also hinder the learning and performance of desirable behav-iour. A pet dog may be too anxious to benefit from training classes or a working dog may be so aggressive that it can’t be trusted to perform in the required cir-cumstances.
Classical conditioning procedures geared toward reducing interfering emo-tional responses are typically more effective at remedying problems such as fear and aggression than operant conditioning procedures geared toward addressing the behaviour directly. For instance, suppose you have a dog that barks and lunges aggressively at people entering your home. One option is to train the dog to maintain a sit-stay and to suppress barking, presumably through negative or positive punishment, when people enter the home. This is an operant condition-ing approach and it would be effective in altercondition-ing the dog’s behaviour, but not necessarily the dog’s underlying emotional state (Wright et al., 2005). A second option is to teach the dog a pleasant association: each time a person enters the home, wonderful things happen to the dog. Maybe the visitor always brings a large meaty bone for the dog. With sufficient repetitions, the dog comes to antic-ipate bones when people come to the door. The dog now experiences a new kind of conditioned emotional response, one of pleasure and excitement, rather than anxiety and aggression. One would predict the dog’s behaviour to change from lunging and barking to opening the door wide and welcoming guests to your home!
Desensitization and counterconditioning
The most popular approach for eliminating problematic behaviours, such as fearful and aggressive responses, is desensitization and classical counterconditioning (DSCC). Wolpe (1958) was the first to describe the use of DSCC to eradicate con-ditioned fear in cats. He taught hungry cats to fear the presence of food by associ-ating the act of feeding with the delivery of electric shock. After the initial conditioning, the cats refused to eat in the experimental room, despite significant food deprivation. The cats also generalized to rooms that were similar to the experimental room. Wolpe identified a room that was sufficiently unlike the exper-imental room that the cats, though still anxious, were able to eat. Successive feed-ings in the new room eliminated all signs of distress. The cats were then moved to a room slightly more similar in appearance to the experimental room and offered food. This was repeated in a series of rooms of increasing similarity to the experi-mental room, each time remaining in the same room until distress was no longer visible. Eventually the cats were able to eat comfortably in the original room.
This same technique can be applied to a fearful dog, such as that shown in Fig. 8.5. This dog became tense when people reached for it. If escape were possi-ble, it would retreat. If not, it would snap and bite. I started by pairing a hand lifted in the air, but not directed toward the dog, with highly palatable treats.
Once the dog was comfortable with this arrangement, I progressed to orienting the hand toward the dog. Gradually the hand was moved progressively closer to
the dog, remaining at each step until the dog appeared to relax. Eventually, the dog appeared eager to have a hand reaching out to touch it, even when treats weren’t offered immediately (see Fig. 8.6). Because the dog no longer feared being touched, defensive behaviours were replaced with approach and tail wagging.
Fig. 8.5. Photograph of chihuahua recoiling from a hand.
Fig. 8.6. Photograph of chihuahua tolerating hand touch.
Classical counterconditioning, in particular, is an extremely powerful agent for behaviour change. In most laboratory studies of counterconditioning (also known as cross-motivational transfer experiments), an initially neutral stimulus is first paired with one unconditioned stimulus, such as shock, and then, at a later time, the same previously neutral stimulus is paired with a motivationally disparate stimulus, such as food (Dickinson and Dearing, 1979). In even more convincing demonstrations, though, two incompatible unconditioned stimuli are paired together. Erofeeva (1921, cited in Dickinson and Dearing, 1979) used a strong electric shock to signal the delivery of food to hungry dogs. The dogs initially responded with defensive behaviours, such as struggling and yelping. Yet, as conditioning progressed, the dogs began to show typical appetitive responses in response to the shock, including lip licking and salivation. Even more surprising, after counterconditioning the aversive stimulus to signal an appetitive stimulus, the aversive stimulus was incapable of func-tioning as a punishing stimulus in an operant condifunc-tioning procedure. Classical counterconditioning is so potent that it appears to actually produce a change in the motivational and reinforcing properties of an unconditioned stimulus.