ISCA Archive SpeechProsody 2010
ISCA Archive SpeechProsody 2010

Production–perception entrainment in speech rhythm

Pablo Arantes, Plinio A. Barbosa

The article reports favorable initial results to the hypothesis that rhythm perception can be seen as a listener–speaker entrainment process. The data comes from an experiment in which subjects had to detect a click in test sentences. Each sentence contained one click that was associated to one of the syllables of two consecutive stress groups defined by duration criteria. Reaction time (RT) to click detection is assumed to reflect the degree of listener-speaker entrainment: faster detection meaning stronger entrainment. Results show that the closer to the phrasally stressed syllable the click is the faster the RT is. The crucial result concerning our working hypothesis, though, is that RT slows down after the stress group boundary, resuming its decrease trend afterwards. Multiple linear regression analysis performed on different acoustical parameters of the test sentences shows that duration and F0 explain around 50% of RT variance. We interpreted these results as a positive preliminary corroboration of the entrainment hypothesis by showing that boundaries in the spoken utterances seem to trigger a reset in entrainment activity and that duration seems to be the main acoustical feature driving listeners' behavior.

Index Terms: speech rhythm, rhythm modeling, rhythm processing

Cite as: Arantes, P., Barbosa, P.A. (2010) Production–perception entrainment in speech rhythm. Proc. Speech Prosody 2010, paper 221

  author={Pablo Arantes and Plinio A. Barbosa},
  title={{Production–perception entrainment in speech rhythm}},
  booktitle={Proc. Speech Prosody 2010},
  pages={paper 221}