Disfluencies in gesture: Gestural correlates to filled and unfilled speech pauses

Abstract

This paper investigates the relationship between Empty and Filled pauses in speech and Simple and Augmented Holds in gesture. An Empty pause is a silent interval within the subject's contiguous speech. A Filled pause is an interval that is occupied with some non-lexical speech element (e.g. `Uhm'). Holds are detected from hand motion traces extracted as flowfield clusters from video. A Simple Hold is defined as an active configured gestural state where no intended motion is perceived. An Augmented Hold occurs when a hold is perceptually judged to have occured but with detectable motion in the `holding' hand. The extent to which the verbal and non-verbal modalities of communication overlap with respect to these phenomena is examined through the analysis of audio and video data. We show that Simple Holds and Empty Pauses are similarly distributed, and that they are not synchronized in speech. Both computational and psycholinguistic implications are discussed.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 37 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.