Generalization In Human Beings

When a conditioned response has been established to a particular CS, stimuli similar to that CS will also tend to elicit the response, a process that is called generalization. The more similar the new stimulus is to the original CS, the greater will be the strength of the response. The nature of generalization was illustrated clearly in a simple example by Pavlov. The CS was a mechanical scratching of the dog's skin, which had been paired with meat powder (the US). When the conditioned response (salivation) had been firmly established, the dog was tested with scratches on different parts of its body. The closer the scratch was to the point at which the original CS had been applied, the more the dog salivated.

Generalization in human beings can easily be shown. The galvanic skin response (GSR), a change in the electrical resistance of the skin related to sweating, is an unconditioned emotional reaction. When human subjects receive an electric shock (US), one of their emotional reactions (UR) is a marked GSR. The GSR is easily conditioned: If a pure tone is sounded shortly before the subjects receive a shock, after only 1 or 2 trials they show a pronounced GSR to the tone alone. When human subjects were conditioned to a particular tone, higher and lower tones also elicited a GSR. The GSR was greater when the test tone was close to the tone used as a CS during conditioning (Hovland, 1957).  With people, the similarity of one stimulus to another may depend critically on language and symbolization. Thus Diven (1936) conditioned human subjects to make a GSR in response to the word "barn'' embedded in a list of recited words. This was easily accomplished by following the word "barn'' with an electric shock. When conditioning had been established, the subjects also displayed GSRs to such rural words as "cow'' or ''hay''but not to such neutral words as ''table'' or ''chair.'' The generalization in this case has nothing to do with the physical similarity of one word to another but obviously depends on the meanings of the words.

The fact that conditioned responses generalize to stimuli similar to the CS has obvious adaptive value. The precise measurement and control over CSs that is possible in the laboratory is not much like what happens in real life, where stimuli are always changing from one occasion to the next.

If a tiger has clawed you once, you are very likely to display a GSR and other conditioned emotional reactions the next time you see a tiger. The conditioning would not be very useful, however, if as a result of your first experience you responded only to the sight of a tiger approaching from your right with a distinctive wart on its nose.