Timing in conversation is dynamically adjusted turn by turn: Evidence for lag-1 negatively autocorrelated turn taking times in telephone conversation
Conversational turn taking in humans has been deemed an exquisite feat of timing. Speakers tend to anticipate another speaker’s turn ending so as to rapidly initiate the next turn as a response. This rapid response often takes only around 200ms which is considerably less time than it would take to plan and initiate a next turn in response to a turn-final silence. The timing mechanism has been heavily debated, and revolves around questions such as who is doing the timing, and how to model such a timing mechanism. Here we replicate a basic phenomenon obtained in non-communicative timing research on human rhythmic tapping abilities. We show that turn transition times in phone conversations behave similarly as rhythmic tapping to a metronome. We show that there is serial dependence between turn transition times (TTT’s), such that TTT’s are lag-1 negatively autocorrelated, suggesting that there is a joint correction mechanism operating at the level of the dyad in TTT’s during telephone conversations. This finding, if replicated, has major implications for models describing turn taking, and confirms the joint anticipatory nature of human conversational dynamics. Future research is needed to see how pervasive serial dependencies in TTT’s are, such as for example in richer communicative face-to-face contexts where visual signals affect conversational timing.