REVERSALS PRIOR TO SOLUTION IN CONCEPT IDENTIFICATION.

These studies investigated the effects of reversal and nonreversal shifts before solution upon performance in a later concept identification task. In 2 experiments, a reversal or a nonreversal shift after an error on a critical trial had no interfering effect upon subsequent learning. The reversal and nonreversal groups made about the same number of errors and required as many trials to learn as did controls who were not shifted. In a 3rd experiment, 1 group of Ss received reversals on every alternate error, but still made the same number of informed errors as did controls who learned with no shifts. These results support the hypothesis that learning is insightful or an all-or-nothing event in simple concept identification. In the typical two-category concept identification experiment, 5 is shown a series of complex patterns which vary in several, binary attributes. As each pattern is presented, 5 attempts to anticipate the correct classification; following his response, he is informed of the correct response. The patterns