Practice effects for auditory localization: a test of a differentiation theory of perceptual learning and development.

Practice and transfer of practice in an auditory localization task were observed for three different information conditions (normal, transformed and reduced) to examine four aspects of Gibson's (1969) differentiation theory of perceptual learning and development. Support was found for the propositions that: (i) dimensions of difference, rather than prototypes, are learned; (ii) specificity is a more appropriate response measure than number of correct responses; (iii) perceptual learning will occur regardless of whether feedback is given or not; (iv) transfer of practice will be greater if dimensions of difference (as opposed to prototypes) are learned. It was also found that feedback during practice did not affect transfer. In an apparent contradiction though, specificity increased in the absence of dimension of difference. This increase was interpreted as being associated with the detection of non-task-related information and was seen as lending support to Gibson's hypothesis that, perceptual learning is motivated by intrinsic cognitive drive and is terminated by a reduction in subjective uncertainty.