I Don't See It Your Way: The Dot Perspective Task Does Not Gauge Spontaneous Perspective Taking

Data from studies employing the dot-perspective task have been used to support the theory that humans are capable of automatically computing the visual perspective of other individuals. Recent work has challenged this interpretation, claiming instead that the results may arise through the automatic reorienting of attention triggered by observed head and gaze cues. The two experiments reported here offer a stronger test of the perspective taking account by replacing the computer-generated avatars used in previous research with, respectively, photo-realistic stimuli and socially co-present individuals in a "live", face-to-face version of the task. In each study observers were faster to judge the number of dots in a display when either a digitized image depicting a human "gazer" (Experiment 1), or a socially co-present gazer (Experiment 2) could see the same number of dots as the observer, than when the number of dots visible to each was different. However, in both experiments this effect was also obtained in conditions where barriers clearly occluded the gazers' view of the target dots so that the perspectives of participants and gazers were always different. These results offer no support for the idea that participants are engaged in spontaneous perspective taking in the dot perspective task. It is argued that, instead, the results are likely caused by a spontaneous redirection of a viewer's attention by the observed gazes, which is unlikely to involve representations of the gazer's mental state.