This study debates how American English speakers employ multimodal role shift practices through bodily articulators when quoting characters in personal narratives. Four bodily articulators apparently forecast role shift impression: character intonation, character facial expressions (yet not gaze), character viewpoint gestures, and changes in body orientation. The quoted utterance type defines the degree of multimodal articulation: quotations of actual, witnessed speech events tend to collect fewer articulators than composed (‘fictive interaction’) quotations. These findings point out the significance of exploring the methodical use of the visual modality in quotation and in ordinary interaction.