Seeing the experimenter influences the response to pointing cues in long-tailed macaques

Methodological variations in experimental conditions can strongly influence animals' performances in cognitive tests. Specifically, the procedure of the so-called object-choice task has been controversially discussed; here, a human experimenter indicates the location of hidden food by pointing or gazing at one of two or more containers. Whereas dogs usually succeed, results for nonhuman primates are ambiguous. In the standard version of the task the majority of subjects do not respond appropriately to human pointing. However, modifying the task setup, such as placing the containers further apart, seems to improve subjects' performances, suggesting that cue salience may be an important variable. Here we investigated whether the visibility of the experimenter inhibits long-tailed macaques' (Macaca fascicularis) usage of the pointing cue. In our baseline condition, with the experimenter fully visible, the monkeys chose the correct container in 61% of the trials. The performance increased significantly, however, when the experimenter was hidden behind a curtain and only the arm of the experimenter, a doll's arm, or a stick was visible. Furthermore, the monkeys performed significantly better when the tip of the pointing finger or device was close to the target compared to the more distant condition. Intriguingly, after these experiments the monkeys' performance was also significantly improved in the baseline condition (70%). Apparently, the monkeys were first distracted by the presence of the experimenter, but then learned to use the cue. These findings highlight the importance of the test conditions, and question some of the assumptions about species-specific differences in the object-choice task.