The successive approximations reinforced are increasingly accurate approximations of a response desired by a trainer. As training progresses the trainer stops reinforcing the less accurate approximations. For example, in training a rat to press a lever, the following successive approximations might be reinforced:
- simply turning toward the lever will be reinforced
- only stepping toward the lever will be reinforced
- only moving to within a specified distance from the lever will be reinforced
- only touching the lever with any part of the body, such as the nose, will be reinforced
- only touching the lever with a specified paw will be reinforced
- only depressing the lever partially with the specified paw will be reinforced
- only depressing the lever completely with the specified paw will be reinforced
The trainer would start by reinforcing all behaviors in the first category, then restrict reinforcement to responses in the second category, and then progressively restrict reinforcement to each successive, more accurate approximation. As training progresses, the response reinforced becomes progressively more like the desired behavior.
The culmination of the process is that the strength of the response (measured here as the frequency of lever-pressing) increases. In the beginning, there is little probability that the rat would depress the lever, the only possibility being that it would depress the lever by accident. Through training the rat can be brought to depress the lever frequently.
Successive approximation should not be confused with feedback processes as feedback generally refers to numerous types of consequences. Notably, consequences can also include punishment, while shaping instead relies on the use of positive reinforcement. Feedback also often denotes a consequence for a specific response out of a range of responses, such as the production of a desired note on a musical instrument versus the production of incorrect notes. Shaping, on the other hand, involves the reinforcement of each intermediate response that further resembles the desired response.
Not all approximations are successful. Marian and Keller Breland (students of B.F. Skinner) used their knowledge of autoshaping to try to make a pig and a raccoon deposit a coin in a bank. However the sign-tracking failed. The coin, which was being reinforced with food, began to be perceived as the food reward itself by the animals. They acted towards the coin in the same way that they may have acted towards a snack. Animals that act this way are more prone to addictive behaviors than others. Sometimes these animals may even be called "sign-trackers". If the animal did not behave in this manner and actually placed the coin in the bank, it may have been labeled a "goal-tracker".
Read more about this topic: Shaping (psychology)
Famous quotes containing the word successive:
“He represents the privilege of the intellect, the power, namely, of carrying up every fact to successive platforms, and so disclosing, in every fact, a germ of explanation.”
—Ralph Waldo Emerson (18031882)