The speaker's “okay” vs. the listener's “okay”: exploring lexical, phonetic, and multimodal variation of backchannels and fluencemes in conversation

This research explores the use of fluencemes and backchannels in German face-to-face conversations. Fluencemes are produced by the current speaker to facilitate speech planning and to structure the speaking turn, while backchannels are produced by the current listener to signal acknowledgment and understanding of what is being said. These two conversational devices are constituted by very short utterances, often sharing the same lexical form (e.g., “ja,” “okay,” “genau” in German); however, they display significant differences based on their function within the interaction. This study compares the distribution of backchannels and fluencemes within the conversational turn, their acoustic form, as well as their interaction with multimodal resources, across and within dyads. We find that, while the two devices share lexical candidates, their duration, F0 and pitch contour, as well as the frequency and type of co-occurring head movements vary depending on whether the item is produced by the speaker or by the listener.