Multimodal breathiness in interaction: from breathy voice quality to global breathy “body behavior quality”
Résumé
The face-to-face interaction is a complex dynamic process in which the interactants are mutually, continuously and reciprocally sharing with each other vocal and non-vocal information. This multimodal information has been defined as relevant to the speakers' interpersonal intimacy degree in the relation process ("socio-affective glue"). Otherwise, the human multimodal intimacy cues dynamics is a crucial aspect of the analysis of the human socio-affective behavior. To adress this point, our study is based on the EEE corpus involving spontaneous dialogs between the elderly and a smart home control robot whose vocalizations has been shown in previous studies. The utterances produced by the subjects are imposed commands (smart home orders). However, the elderly's vocal and non-vocal behavior changes gradually throughout the experiment by varying in breathiness, commands paraphrasing, and non-vocal cues, which we suppose could be meaningful for the nature of the human-robot relation. In this study, the communication dynamics is observed as an overall behavior. Accordingly, we suggest that the breathiness dynamics could be superposed, in the dialog time, to the dynamics of both the morpho-lexico-syntactical style and the proxemic cues (as postural proximity and gaze direction).