Visual influences on interactive speech alignment.

Abstract

Speech alignment describes the tendency to produce speech that shares characteristics with a perceived speech signal. Visual speech information has also been found to modulate speech alignment [R. M. Miller et al., Attention, Perception, & Performance (in press)]. The present study evaluated whether visual speech information enhances alignment over auditory‐alone information in a live context. Pairs of subjects performed an interactive search task for which they were required to utter nine key words multiple times. Half of the subjects performed the task while interacting face‐to‐face, while the other half performed the task separated by a cloth preventing them from seeing each other without inhibiting the heard speech. Subjects also uttered the key words before and after the interaction. Alignment was evaluated using an AXB matching task in which raters judged whether post‐interaction or pre‐interaction utterances of a word spoken by a subject were more similar to the post‐interaction word spoken by that subject’s partner. Raters also judged similarity between partner utterances from early and late in the interaction. Evidence showed that visual information did enhance interactive speech alignment. These findings add to the evidence that amodal, gestural properties of perceived speech can induce alignment.