Variables Influencing the Nature of Learned Categorical Perception Effects

Abstract

Considerable research has demonstrated so-called learned categorical perception (CP) effects, where learning to classify a set of stimuli leads to either compression (within-category stimuli judged to be more similar and/or confusable than before learning) or expansion (between-category stimuli judged to be less similar and/or confusable than before learning) or both. The issue of why category learning causes one type of effect or the other has not been systematically investigated, but previous research suggests that highly discriminable stimuli may tend to produce compression while stimuli that are difficult to discriminate may tend to produce expansion. We report a series of studies testing the effect of stimulus discriminability on the type of learned CP effect produced using both similarity and XAB measures of the effects. Preliminary results suggest that different measures of categorical perception reveal different effects, and that category structure may also be relevant.


Back to Table of Contents