Abstract

There is a growing body of research demonstrating that during comprehension, language users
predict upcoming information. Prediction has been argued to facilitate dialog in that listeners try to
predict what the speaker will say next to be able to plan their own utterance early. Such behavior may
enable smooth transitions between turns in conversation. In face-to-face dialog, speakers produce a
multitude of visual signals, such as manual gestures, in addition to speech. Previous studies have shown
that comprehenders integrate semantic information from speech and corresponding iconic gestures when
these are presented simultaneously. However, in natural conversation, iconic gestures often temporally
precede their corresponding speech units with substantial lags. Given the temporal lags in gesture-
speech timing and the predictive nature of language comprehension, a recent theoretical framework
proposed that listeners exploit iconic gestures in the service of predicting upcoming information. The
proposed study aims to test this proposal. We will record electroencephalogram from 80 Dutch adults
while they are watching videos of an actress producing discourses. The stimuli consist of an
introductory and a target sentence; the latter contains a target noun. Depending on the preceding
discourse, the target noun is either predictable or not. Each target noun is paired with an iconic gesture
whose presentation in the video is timed such that the gesture stroke precedes the onset of the spoken
target either by 520 ms (earlier condition) or by 130 ms (later condition). Analyses of event-related
potentials preceding and following target onset will reveal whether and to what extent targets were pre-
activated by iconic gestures. If the findings reveal support for the notion that iconic co-speech gestures
contribute to predictive language comprehension, they lend support for the recent theoretical framework
of face-to-face conversation and offer one possible explanation for the smooth transitions between turns
in natural dialog.

Abstract

Repair is a core building block of human communication, allowing us to address problems of understanding in conversation. Past research has uncovered the basic mechanisms by which interactants signal and solve such problems. However, the focus has been on verbal interaction, neglecting the fact that human communication is inherently multimodal. Here, we focus on a visual signal particularly prevalent in signaling problems of understanding: eyebrow frowns and raises. We present a corpus study showing that verbal repair initiations with eyebrow furrows are more likely to be responded to with clarifications as repair solutions, repair initiations that were preceded by eyebrow actions as preliminaries get repaired faster (around 230 ms), and eyebrow furrows alone can be sufficient to occasion clarification. We also present an experiment based on virtual reality technology, revealing that addressees’ eyebrow frowns have a striking effect on speakers’ speech, leading them to produce answers to questions several seconds longer than when not perceiving addressee eyebrow furrows. Together, the findings demonstrate that eyebrow movements play a communicative role in initiating repair in spoken language rather than being merely epiphenomenal. Thus, they should be considered as core coordination devices in human conversational interaction.