Evaluating the logic of perspective-taking experiments.

In their recent study, Udell, Dorey, and Wynne (in press)[COMP: Ref. is LBE0034. Will have to be updated to (2011) here and in LBE0038, once page nos. are available.] showed that in a begging task, at least in some conditions, dogs as well as wolves preferentially approached a human partner who could see them in contrast to one whose eyes were occluded, and Udell et al. concluded that this success was dependent on the subjects' experiences with the specific occluder used. Here we argue, however, that since both partners expressed similar attentiveness towards the subjects by calling their names, Udell and colleagues' conclusion does not refer to the sensitivity of canines to others' attentiveness, but instead reflects the fact that the animals obeyed a familiar command better in a familiar context than in an unfamiliar one. Moreover, in contrast to Udell et al.'s conclusion, we believe that their data demonstrate that pet dogs can generalize the use of the visibility of human eyes to novel situations, showing a preference towards an attentive partner even if the eyes of the other partner are occluded in a novel way (e.g., having a bucket on his or her head). Finally, after presenting alternative interpretations of the results of the wolves tested by Udell and colleagues, we conclude that there is no evidence that wolves are sensitive to the attentional states of humans.