Stevens (1989; 2002) proposed that phonemic categories play a fundamental role in speech perception. According to this view, there is a process for specialized speech perception mechanisms to extract prototypical categories from variable acoustic inputs, and discard the variability to unmask the underlying phonetic or featural prototypes. One important support for this model is said to come from studies on categorical perception. For example, Cienfuegos et al. (1999) presented evidence that the deficit in categorizing acoustic properties of speech sounds may hinder the analysis of phonemes, and lead to the difficulty in language processing in schizophrenia.
展开▼