Maintaining multiple viewpoints with gaze

Co-speech bodily gesture has remarkable flexibility in displaying or enacting viewpoint, since — unlike speech but like signed languages — it deploys multiple relatively orthogonal articulators, including head and gaze, two arms and hands, and torso posture. Combined with the viewpoints expressed in the linguistic track, this allows oral narrators to embody viewpoints of two characters at once, or to embody both narratorial viewpoint and an embedded character viewpoint simultaneously. This paper examines video data of semi-spontaneous personal narratives told by speakers of American English. We observe some of the ways in which gaze specifically is used to mark and maintain either the narrator's or some character's viewpoint (including the narrator's Past Self as a story character) even while other articulators may be marking a different viewpoint. These include discourse uses of gaze marking memory access, or 'checking' for approval from an interlocutor, as well as content uses such as alternation between enacted characters' gazes. It is always the storyteller's own eyes and face doing the gaze-enaction, but the understood meaning attributes a particular gaze to one of a complex of narrative viewpoints. This is transparent to listeners/viewers because they have access to the complex set of mental spaces evoked, not just to the physical space.