PUB - Publikationen an der Universität Bielefeld

In face-to-face interactions, the conversational feedback produced by the listener to signal attention and participation to the current speaker is multimodal: in the vocal channel, it consists of verbal expressions (e.g., ``yes'' or ``exactly'') and vocalizations without lexical content, such as non-lexical backchannels (e.g., ``mhm'') and laughter; in the visual channel, listener feedback includes movements of the head, such as nods or tilts. In the current research, we investigate the frequency and the distribution (i.e., the location and the transition type with respect to the other interlocutor's turn) of lexical and non-lexical items, laughter and head movements, as well as the phonetic variation of vocal feedback, in face-to-face dialogues in German. We find that the feedback type influences the distribution and the variation of intensity values and voice quality, and, for multimodal items, it also influences the temporal alignment of the head movement with the vocal component.