Seeing first person changes gesture but saying first person does not

Events with a motor action component (e.g., handling an object) tend to evoke gestures from the point of view of a character (character viewpoint, or CVPT) while events with a path component (moving through space) tend to evoke gestures from the point of view of an observer (observer viewpoint, or OVPT). Events that combine both components (e.g., rowing a boat across a lake) seem to evoke both types of gesture, but it is unclear why narrators use one or the other. We carry out two manipulations to explore whether gestural viewpoint can be manipulated. Participants read a series of stories and retold them in two conditions. In the image condition, story sentences were presented with images from either the actor’s perspective (actor version) or the observer’s perspective (observer version). In the linguistic condition, the same sentences were presented in either the second person (you…) or the third person point of view (h/she…). The second person led participants to use the first person (I) in retelling. Gestures produced during retelling were coded as CVPT or OVPT. Participants produced significantly more CVPT gestures after seeing images from the point of view of an actor, but the linguistic manipulation did not affect viewpoint in gesture. Neither manipulation affected overall gesture rate, or co-occurring speech. We relate these findings to frameworks in which motor action and mental imagery are linked to viewpoint in gesture.