de la Cruz-Pavía, I. 1, 2 , Werker, J. F. 1 , Vatikiotis-Bateson, E. 1 & Gervain, J. 2
1 University of British Columbia, Canada
2 Université Paris Descartes
Speech is audiovisual from infancy, but the role of visual information in speech perception remains largely unexplored. This study investigates the role of co-verbal visual information in parsing speech into structural units. Specifically, we examine whether the presence of head-nods in the input helps adult speakers break speech into phrases. We familiarize English monolinguals and English/other-language bilinguals with ambiguous artificial languages that contain: (a) phrase-level auditory-only (AO) prosodic information, or (b) phrase-level audiovisual (AV) prosodic information. The auditory prosody consists of changes in duration (short vs. long syllables), a phrasal prosodic patterning characteristic of English. The audiovisual prosody consists of the addition of head nods displayed via an animated line drawing of the face, where the apex of the nods occurs in the middle of the long, acoustically prominent syllables. Participants are subsequently tested on their segmentation preferences. Preliminary results show that both monolinguals and bilinguals segment the ambiguous language according to the phrasal patterning characteristic of English, and this preference is reliably stronger in the AV than the AO condition. These results suggest that the available visual information influences the segmentation preferences of both monolinguals and bilinguals and thus provide further evidence for the intermodal nature of speech.