Consonants and vowels behave differently in silent center syllables

Abstract

Compared to consonants, steady-state vowels show no right ear advantage, activate different brain regions, and only trigger categorical perception when extremely short. Auditory explanations for this distinction assume different perceptual mechanisms for brief, dynamic consonants and long, slow-changing vowels. Gestural views suggest steady-state vowels are perceived as only partially speechlike. The present study compares consonant and vowel identification and discrimination when based on the same brief, dynamic information in the form of silent center (SC) syllables. The CVCs simultaneously varying in equal-sized F2 steps along both a /b-d/ and an /ε-Λ/ continuum were synthesized (full syllables). The SC syllables were produced by replacing the middle 60% with silence. Relative to full syllables, overall SC consonant discrimination improved from chance to 70%; vowel discrimination dropped from 85% to 70%. The SC vowel identification was affected by consonantal context but the reverse effect was weaker. Improved SC consonant discrimination may result from decreased masking or greater prominence of the consonant gesture. Decreased SC vowel discrimination could result from reduced acoustic information or weaker prominence of the inferred vowel gesture. While these results do not distinguish auditory and gestural theories, they do place the consonant/vowel distinction in a new perspective. [Work supported by NIH Grant No. HD-01994.]