Response latency differences in speech shadowing as an indicator of distinctive features of consonants.

Abstract

The purpose of this study was to determine which articulatory features serve as important distinctive features of perception by measuring latency in speech shadowing. Fifteen subjects shadowed a recording of consonant-vowel syllables orally and by means of a hand-held switch. Vocal shadowing was tape recorded and the onsets of the stimuli and the manual responses were recorded on a polygraph. The oral responses were used to indicate perceptual confusions and the manual responses were used as the basis for measuring the shadowing latency for each consonant. After eliminating the dental fricatives because of perceptual confusions, the remaining consonants (p, b, t, d, k, g, f, v, s, z, ∫, 3, m, n, 1, r, j, w,) were classified according to a feature system and the latency scores were analysed. An analysis of variance indicated significant differences among sounds arranged according to features. A post hoc comparison between all possible features (interrupted, grave, acute, diffuse, strident, voiced, tense, nasal) showed that three features were primarily responsible for the significant variance: stridency (or, minus interrupted) voicing, and nasalization. The results do not seem to support the concept of a hierarchy of features which order the decisions made in consonant identification.