A number of personality theorists stress the importance of childhood experience over later experience in the development of personality, and most child psychologists agree with them. Freud, who saw the first few years of life as determinative, was among the most extreme exponents of this view. Those working with learning theory tend, however, either to ignore developmental effects or to dismiss them flatly as relatively unimportant. Watson was the extreme here. He, however, recognized a change in rate of learning with age, which he attributed to some tendency to get settled into a humdrum way of liv. ing after 30 ( 2). If what can be learned can also be unlearned in about the same length of time, what is important about childhood? It is just this point I wish to seize upon. I wish to suggest that what is learned in childhood quite probably cannot ever be completely unlearned if the unlearning process is long delayed because with time the child will develop an increased ability to differentiate. [One of the Stanford-Binet items (1) for example, is a test of whether the child can discriminate squares, circles, and triangles.] Hence the stimul~~s to which he was conditioned will be broken into a set of stimuli, and unless the child's response is extinguished to each element in the set, some effect of the original will remain. Thus, if a child learns to drop a black circular token into a machine for a reward before he can discriminate circle from square or triangle and then if later after the discriminative ability develops this behavior with black circular tokens is extinguished, the effect will still remain for black square or triangular tokens. Or if terrorized by his father during [he period when all men appear alike, and then lacer after developing discrimination, is exposed only co kind men, the child will not react to all men positively but rather is likely to react with fear except to those who have blond hair, mustaches, or other features similar ro those of the kind men experienced later. Thus, once the pillow of feachers is opened, the feathers can never all be gathered. If this suggestion proves out experimentally, it not only offers a point of reconciliation between the two views but a more exact way of predicting the effect of early experience.