The effects of syllable boundary, stop consonant closure duration, and VOT on VCV coarticulation

This study investigated whether the vocalic gestures in VCV sequences are produced with a single diphthongal movement as predicted by the superimposition model of coarticulation or as separate events as predicted by a phoneme-by-phoneme view. ‘Troughs’ or discontinuities in anticipatory vowel-to-vowel coarticulation during the closure period of bilabial stops in symmetrical VCV sequences have provided evidence in favor of a phoneme-by-phoneme view. This investigation sought to uncover the acoustic correlates of troughs in VCV utterances produced by five English and two Persian speakers. Acoustic evidence for troughs was found in frequency changes of F2 transitions in symmetrical [V.bV] and [V.phV] sequences. F2 transitions indicative of tongue-related trough effects were influenced by the syllabic affiliation of the intervocalic stop and to a lesser degree by its voicing properties. Changes in consonantal closure duration did not elicit troughs. Locus equations (LE) were employed as a second methodology to uncover the acoustic correlates of troughs in a variety of consonantal and vocalic contexts. LE slopes capture CV coarticulation. Troughs would be expected to lower slope values. Consistently lower LE slopes at the CV2 interface were observed with closed versus open and voiceless versus voiced stops. LE slopes remained stable across changes in consonantal closure duration. Bidirectional V-to-V coarticulatory effects were also explored. The superimposition model predicts no changes in V-to-V coarticulation as a function of the syllabic affiliation, closure duration, and the voicing properties of an intervocalic stop. Results showed reduced anticipatory V-to-V effects in closed versus open, geminate versus singleton, and voiceless versus voiced conditions in English and with lesser degrees of consistency in Persian. Increased carry-over V-to-V effects were observed in closed versus open syllables in English and in voiceless versus voiced conditions in Persian. Carry-over effects were generally smaller in geminate versus singleton utterances in English and Persian and across voiceless versus voiced stops in English. The results of this study support a phoneme-by-phoneme view of segmental organization in which vocalic and consonantal gestures are distinct, but temporally and spatially overlapping events and affect one another in varying degrees based on the degree of their overlap.Linguistic