The present study is motivated by the question of “boundaries” of the reported segment in direct reported speech and its multimodal materialization. Reported speech can be delimited with internal and external syntactic boundaries. Dehé & Wichmann have described three types of segmental and supra-segmental boundaries for parentheticals, which they call “discontinuities” (2010). Hanote has further developed this model for the specific description of reported speech (2012, 2015) using pitch, temporal and segmental discontinuities as markers of boundaries for the reported segment.

We would like to further implement this model with another type of discontinuity, i.e. visual-gestural discontinuities, following our previous work on reenactment (Kosmala & Vigneron-Bosbach, 2022). Reporting talk is a common feature of naturally occurring spoken interaction whereby interactants often animate and re-voice (Goffman, 1981) themselves or someone else, or even a generic speaker engaged in a prior or fictive talk. Direct reported speech can be conveyed by gestures and facial expressions (Li, 1986) which can further be considered as a form of demonstration (Clark & Gerrig, 1990) and reenactment (Sidnell, 2006). As Clark & Gerrig (1990) pointed out, reporting talk is not only concerned with the retelling of an event, but also a display or a depiction, which involves demonstrating things (noises, gestures, facial expressions). More recently, Sidnell (2006) highlighted the multimodal dimension of such demonstrations, and focused on the production of reenactments. Reenactments are said to provide “virtual re-presentations of the relevant events” (Sidnell, 2006: 406) providing opportunities for interactants to convey perspective shifts and visual-spatial representations of a given event, within the same turn-at-talk.

Analyses are based on the DisReg Corpus (Kosmala, 2020) which includes videotaped recordings of French university students engaged in various conversations. These sequences include retellings and reenactments of amusing anecdotes and stories, which are often associated with reported speech (see Holt, 2000). Our analyses offer a multi-level approach, focusing on the left boundary of reported speech, with boundaries at different levels: (1) the syntactic level, with syntactic internal and external boundaries and the presence of quotatives, (2) the prosodic level, with the presence of pauses and inbreath, pitch reset, and change in voice quality, (3) the visual-gestural level, with shifts in body posture and gaze, and production of manual gestures. Based on quantitative annotations on a selection of the data (four pairs), We attempt to show how syntactic, prosodic and visuo-gestual boundaries coincide or not and may reveal recurrent patterns.