Shaping Training Through Successive Approximations

Sometimes a behavior that is unlikely to occur spontaneously will need to be trained in gradual steps. This sort of training is called shaping. Shaping is a process in which a selected behavior is obtained by differentially reinforcing successive approximations of it. Shaping involves breaking down the training objective or target behavior into more manageable and easily learned parts. Many otherwise difficult behaviors can be efficiently trained by carefully arranging these component parts of the target behavior according to a plan or program of instrumental shaping contingencies. Shaping has many applications in dog training. Almost any response or behavior pattern within a dog's behavioral capability can be shaped as long as a few basic rules are followed. An excellent introduction to shaping dog behavior through differential reinforcement of successive approximations is provided by B. F. Skinner (1951).

The first step in the process is to prepare a conditioned reinforcer (Sr+, notice the little

"r") by pairing a stimulus (e.g., "Good" or clicker/tone) with an unconditioned rein-forcer (Sr+). The Sr is often referred to as bridging stimulus. Effective conditioning of the bridging stimulus is crucial to the shaping process. Before shaping can be effective, a dog must recognize that the bridging stimulus communicates at least two messages: (1) that its presentation is contingent on the emission of a particular behavior and (2) that its occurrences are linked with a remote but forthcoming reinforcer. Murphree (1974) recommends that 50 to 100 pairings between the Sr and food be carried out before using it as a bridging stimulus for operant-shaping procedures. In the case of ordinary training activities, far fewer pairings are needed.

Once the Sr has been conditioned, it should be tested to confirm that it meets the aforementioned criteria. The test can be carried out by using the Sr+ to teach a simple behavior dependent on conditioned reinforcement for its acquisition. Normally, the behavior used to test the Sr+ is the orienting or attending response—that is, the bridge stimulus is used to reinforce the behavior of following the hand prompt or looking into the trainer's eyes. Another possible shaping objective might be to train the dog to move toward an opposite corner of the training room. While this behavior is fairly simple, it will help illustrate the most salient features of shaping.

Step 1: Define the Goal or Target Behavior it is important to define precisely the target behavior before training begins. Training objective: To train the dog to leave the handler's side on signal and walk to a predetermined corner of the room.

Step 2: Design a Plan or Program of Instrumental Contingencies

The target behavior should be broken down into as many simple parts as is practical. The plan for shaping the foregoing target behavior might include the following components:

1. The dog turns its head away.

2. The dog turns its head away and in the general direction of the corner.

3. The dog turns and moves its whole body away.

4. The dog turns and moves its whole body toward the corner.

5. The dog is required to move farther away.

6. The dog moves farther away and in the general direction of the corner.

7. The dog moves closer to the corner.

8. The dog enters the corner.

During the early stages, progress may be rapid, but as the requirements become more difficult, the acquisition curve may flatten out. Shaping is a dynamic process controlled by a feedback loop between the dog's progress and the program of instrumental contingencies. If progress is slow, renew momentum by going back a step or two. If the step still proves too difficult, break it down into even simpler elements. A program of instrumental contingencies should be flexible and opportunistic but never vague and capricious. Such adjustments to the plan, therefore, should not be made hastily or without good purpose. Each preceding step should receive enough training to make it a reliable foundation for the next step.

Once a step has been mastered, further requirements must be introduced that compel the dog to experiment with closer approximations to the target behavior. This transition is accomplished by reinforcing the previous step on an intermittent basis. This shift in reinforcement scheduling causes the dog to emit other related behaviors (response generalization) that might offer a higher rate of reinforcement. Placing previously mastered steps on an intermittent schedule is also important to prevent their extinction when selection pressures are made more demanding. Care should be taken, however, to avoid shaping transitions that cause the dog to quit, become overly anxious, or frustrated. When anxiety or excessive frustration appear, the trainer should go back to a previously successful step. Always end each training session on a positive note.

Step 3: Bring the Shaped Behavior Under Stimulus Control

Once the dog has reliably learned to go to the correct corner of the room, the new behavior can be brought under stimulus control. By overlapping the behavior with a hand gesture pointing in the same direction as the dog's movement, he will soon associate the gestural prompt with the movement of walking toward the designated corner. After many repetitions, a new contingency can be introduced requiring that the dog move toward the corner only when signaled to do so. All attempts to move toward the corner not initiated on cue are not reinforced. Such attempts can be overlapped with an Sp~ "No," indicating to the dog that reinforcement is not forthcoming for the behavior. By reinforcing only those behaviors controlled by the gestural prompt, the dog soon learns to move only when prompted to do so. Once the gestural prompt controls the behavior, it is easy to bring the behavior under the control of a verbal SD, for example, "Move." Pairing the SD "Move" with the prompt results in the former acquiring the ability to control the shaped behavior. Once a sufficient number of trials have taken place, the gestural prompt can be gradually faded out, with the verbal SD "Move" alone controlling the newly learned behavior.

The foregoing method of shaping is intended to provide the reader with a formal and structured approach. Under actual training situations, however, shaping is often carried out much more informally. The basic principles of breaking down the training goal into simple parts, organizing and teaching them in the most easily learned order, and carrying out training in a positively oriented manner are common features of all training activities. Shaping techniques provide the skilled trainer/behaviorist with powerful and efficient tools for orchestrating behavioral change through positive reinforcement. Pryor (1975, 1985) wrote at length on the use of operant-shaping techniques in the training of sea mammals and other animals, including dogs.

Ultimate Dog Care Kit

Ultimate Dog Care Kit

Many dogs today are dying early because of misinformation about proper dog care. Learning How to Take Proper Care of Your Pet Dog is Crucial to a Long and Happy Life for Your Pet. Learn How to Make Sure Your Dog Lives to His or Her Full Potential of Life.

Get My Free Ebook


Responses

Post a comment