Stop place coding: an acoustic study of CV, VC#, and C#V sequences

This study investigated stop + vowel coarticulation as a coding mechanism for differentiation of stop place categories in an F2-defined stimulus space. Locus equations (LEs) were used to index the extent of coarticulation in three contexts: (1) onset stop + vowel utterances [.CV]; (2) within-syllable vowel + coda stop utterances [VC#], and (3) across-syllable/word coda stop + vowel utterances [C#V]. Five speakers of American English and 2 speakers of Persian produced [CV1.CV2] (English and Persian), [tV1C#V2t] (English) and [dV1C#V2t] (Persian) tokens with voiced labial, alveolar/dental, and velar stops surrounded by a variety of vowels. In the across-syllable/word boundary condition [C#V2] the extent of anticipatory coarticulation was much reduced relative to traditionally measured onset [CV2] syllables. LE slopes derived across the syllable/word boundary, however, still differed between stop places in the same order as onset CVs. LE slopes derived from within-syllable [V1C#] sequences, however, did not differ between stop places.