Suegami, T. 1 & Michimata, C. 2
1 University of Oslo
2 Sophia University
Categorical perception (better discrimination for the cross-category stimuli relative to the within-category) is observed in face recognition (e.g., Beale & Keil, 1995). There are two possible mechanisms: The one is an expansion effect, in which the discrimination for the cross-category stimuli is intensified. The other is a compression effect, in which the discrimination for the within-category stimuli is attenuated. We examined which mechanism underlies the categorical perception of face by successive 3 experiments. Eleven participants learned to classify 17 successive morphed faces into two categories, and then they rated how each stimulus represented its category. Finally, they performed the face discrimination task. Another 11 participants performed only the discrimination task without learning and rating task, as a control. The results for the rating task revealed the boundary and the prototypes of the two categories. The results for the discrimination task showed that the accuracy for the cross-category discrimination was higher than the within-category discrimination in the category learner group. Such a categorical perception was not obtained in the control group. The accuracy for within-category discrimination, moreover, was lower in the category learner group than the control group, suggesting that the categorical perception of face is mediated the compression effect.