Abstract
Specificity of face recognition was discussed in terms of prosopagnosia and Williams syndrome researches. First, we examined prosopagnosic patients on a new face recognition test with superimposed computed images by morphing. The results showed that prosopagnosic patients had wide differential threshold suggesting the low accuracy of judgment compared with normal subjects ; but more important findings were that patients tended to choice the more similar comparative stimuli on the similarity judging task as well as normal subjects. These findings reflect categorical perception and they suggest that the disturbed categorical level of recognition for prosopagnosia may be defined by expertise level. Recently some fMRI studies gave evidence in support of our findings and cast some doubt on specificity of face recognition. Second, I took up some recent reports concerning face recognition for Williams syndrome patients. Williams syndrome is known as a genetically based disorder with cognitive characteristics expressed as “peaks and valleys” ; they show relative strengths in language and facial processing and profound impairment in spatial cognition, which supports modularity of face recognition. But some reports investigated local/global cognitive ability respectively on both face and non-face recognition, which concluded that Williams syndrome patients recognized faces in terms of local features instead of global features unlike normal subjects. They also constrain the specificity of face recognition. In conclusion, we have many problems to be solved to declare that a face is special.
