Identification of letterlike distortions by preschool children.

In two experiments preschool children chose one of two test forms that looked like previously exposed distortions of a letterlike-form. In Exp. I a distance metric was defined to measure the physical discrepancy between forms. It was found that the distance relationship between distortions and particular test forms did not designate the children's choice behavior. In Exp. II the children were able to choose previously exposed distortions when particular test forms contained novel features. The results are related to distinctive feature and prototype hypotheses to account for children's performance in recognizing forms.