| Abstract | A speeded classification paradigm was used to examine the hypothesis that segmental & tonal dimensions of speech are more strongly integrated in perception by speakers of a tone language (Mandarin Chinese) than by speakers of a nontone language (English). Four binary classification tasks required attention to either a segmental (consonant [C] or vowel [V]) or tonal dimension of CV syllables; the other dimension was to be ignored. Both S groups showed strong perceptual integrality (ie, interference from orthogonal variation in the unattended dimension). The Chinese Ss (N = 8) showed significantly more integrality than the English Ss (N = 8) in only one of the four tasks. However, it was necessary to take into account that all Ss (especially the English-speaking group) took longer to respond to tonal distinctions than to segmental distinctions. With this adjustment, the results suggest that Chinese & English listeners both show an underlying processing asymmetry between consonants & tones in CV syllables
(consonants being more integral with tones than vice versa, as shown in previous research), whereas only Chinese listeners show such an asymmetry between vowels & tones (vowels being more integral with tones than vice versa). This difference may be a reflection of specific language experience. |