University of Pennsylvania Working Papers in Linguistics


In a preliminary investigation of isochrony, the rhythmic integration of talk, we evaluated rhythmic phenomena previously theorized to coordinate turn-transitions for correlates in the acoustic signal. Rhythmic sequencing is one of many elaborate contextualization cues regarded as facilitating a successful turn-transition. Previous studies of rhythm in conversation have attended only to its perceptual and interactional facets. In addressing this gap, our study finds quantitative justification for such claims of rhythmic turn-taking. We selected for acoustic analysis the twelve non-task-based, dyadic conversations of the Santa Barbara Corpus of Spoken American English (SBCSAE). Following Marcus’s (1981) assertion that the onset of the vowel is the closest acoustically-measurable location to the perceptual center of the syllable where the rhythmic downbeat occurs, duration was measured between vowel onsets to create prosodic syllables. Not all prosodic syllables can contain a rhythmic beat, and those that can are characterized as “prominent” in nature (Couper-Kuhlen 1993). Out of 42,807 prosodic syllables measured, our methods yielded 15,972 prominent prosodic syllables. The units of duration between prominent syllables, hereafter intervals, were judged to form an isochronous sequence when the durations between at least three consecutive intervals varied by less than the conservative measure of the perceptual threshold for tolerance of isochrony, up to a 30% variance (Couper-Kuhlen 1993). This measure revealed 564 rhythmic sequences across the twelve SBCSAE conversations, which ranged in duration between one and ten seconds and consisted of up to eleven intervals. Of these, 208 or 37% appeared within turn-transitions, and results from our preliminary analysis indicated that rhythmic sequencing was significantly more likely to appear within a turn-transition than outside of one. Our analysis shows that isochrony is not simply perceptual in nature, but that it has a quantifiable correlate in the acoustic signal. Our findings of significant isochrony in the turn-transitions of the SBCSAE, a corpus often used in discourse analysis, confirms what many interactional sociolinguists have long argued: that rhythmic cues aid the coordination of talk between speakers in turn-transitions. We can confirm that these rhythmic cues are a component of turn-transitions not only perceptually, but acoustically as well.