Back and Forth: Prediction and Interactive Alignment in Conversation
On a daily basis, we engage in multiple conversations with others that seem to occur effortlessly. Despite this apparent ease, the cognitive processes involved in conversation are complex. The challenges associated with conversation include comprehending our partner while simultaneously planning our own contributions and ensuring that what we say is appropriate to the conversational context. The paradoxical ease of this cognitively demanding activity suggests that we may employ a specialized mechanism when engaging in conversation. This dissertation investigates two potential processes that may be involved during social interaction: conversational prediction and alignment. Chapters 2 and 3 of this thesis investigate the perceptual conditions that are necessary for making predictions during conversation. In particular, we investigate how the availability of information influences how we are able to predict a turn exchange during conversation. We present findings that demonstrate that perceivers can predict upcoming turn exchange behavior between two talkers. Further, we suggest that auditory and visual cues serve distinct roles in the predictive processes involved in conversation. Chapters 4, 5 and 6 investigate behavioral alignment in social interaction. In general, it has been shown that conversing talkers align several aspects of their behavior. Talkers have independently been shown to become similar in linguistic attributes such as the words they use and nonlinguistic attributes such as their movement patterns during an interaction. Here, our findings demonstrate that talkers, in fact, align multiple attributes during a single conversation. In addition, there are relationships between alignment behavior within these attributes; when talkers become similar within one attribute, they are more likely to become similar within another attribute. Further, we show that this alignment behavior is influenced by the relationship between talkers suggesting that social factors play a significant role in our conversational behavior. Taken together, this dissertation demonstrates that prediction and alignment behavior are key features of conversational interaction. In particular, we demonstrate that the informational and social context of a conversation determine how these important processes operate. These findings have important implications for the identification of the specific underlying mechanism that is responsible for ensuring the success of our daily communication.