Calculating articulatory syllable duration and prosodic boundaries

Articulatory duration of vowels is often measured as the duration from jaw closure maximum to jaw closure maximum, flanking the production of the syllable nucleus. However, this method may not necessarily represent articulatory syllable duration, since the actual onset/offset of the syllable depends on the specific articulator (lip, tongue) that implements the articulation of the syllable onset/coda, and which does not strictly synchronize with timing of jaw movements nor with acoustically measured syllable durations (e.g., [Menezes, 2004]). We propose that by using a different approach, that suggested by the C/D model [Fujimura, 2000], it is possible to compute quantitatively the time values of prosodic boundaries from articulatory dynamics data. The algorithmic output, the “syllable pulse train”, is the phonetic realization of the utterance’s rhythmic structure (e.g., [Bonaventura and Fujimura, 2007]), which in turn reflects the phonologically derived metrical structure of the utterance (e.g., [Erickson et al., 2012]). Our small study presented here using the C/D model indeed revealed systematic articulatory patterns across speakers.