Auditory and visual perception of silent-center syllables.

Abstract

Silent-center syllables are defined as CVC syllables in which the vowel nucleus is attenuated to silence, leaving three or four pitch periods of consonant transition at either edge of the syllable. Previous studies [W. Strange, J. Acoust. Soc. Am. 85, 2135–2153 (1989)] have shown that subjects are able to identify vowels in silent-center (SC) syllables with nearly the same accuracy as for full syllables. These results have been used to support the theory of dynamic specification, where vowels are conceived of as gestures having intrinsic timing parameters. The current studies suggest alternate interpretations of the SC results. In the first study, the perception of a set of SC syllables was compared with that of full syllables at several signal-to-noise ratios. Contrary to previous experiments, it was found that identification performance was significantly better for full syllables than for SCs across signal-to-noise ratios. This suggests that, given a more demanding perceptual task, identification is hindered when the vowel nucleus is not physically present in the signal. The second study was a replication in the visual mode of earlier studies on SC syllables [W. Strange, J. Acoust. Soc. Am. 85, 2135–2153 (1989)]. Here, spectrographic analogs of full, SC, and other altered syllables were created. Subjects performed a matching task. Results were similar to those from the auditory domain, suggesting that more general perceptual processes may underlie the pattern of responses in both sensory modalities.