On the perception of coarticulation effects in English VCV syllables.

Abstract

Ohman’s 1966 investigation of the acoustic correlates of coarticulation in VCV sequences indicates that terminal formant frequency transition values are strongly influenced by the nature of the transconsonantal vowel. This experiment was designed to explore the perceptual correlates of Ohman’s spectrographic findings. We discovered that when a VCV sequence (where C is a voiceless plosive) is cut in two during the period of consonantal closure, there are not enough remaining cues in either the resulting VC or CV sequences to allow for identification of the deleted segment or of its articulatory features. However, it appears that coarticulation effects may hinder recognition of nonfinal allophones placed artificially in final position. Consonants in VC sequences spliced from original VCV utterances are more difficult to identify than unreleased final consonants of the same quality.