Inference in a social context: A comparative study of capuchin monkeys (Cebus apella), tree shrews (Tupaia belangeri), hamsters (Mesocricetus auratus), and rats (Rattus norvegicus).

Four species (capuchin monkeys, tree shrews, rats, and hamsters) performed an inference task situated in a social context. In Experiment 1, capuchin monkeys first explored food sites under 1 of 2 conditions: In 1 condition, food was refilled after it was eaten (replenished condition), whereas it was not refilled (depleted condition) in the other condition. Two food sites were presented for each condition. In the test phase, a subject watched a conspecific demonstrator visit 1 of the food sites in either the replenished or depleted condition. A screen placed in front of the sites prevented the subject from seeing the demonstrator actually eat the food. When the demonstrator was removed, the subject explored the cage. Three of 4 monkeys tended to go to the unvisited sites in the depleted condition, but tended to go to the visited site in the replenished condition. This suggests that they inferred that there was no food because the demonstrator had eaten it. In Experiment 2, using the same procedure, 2 nongroup-living species (tree shrews and hamsters) were indifferent to demonstrator behavior and visited sites only randomly, and group-living rats showed a strong tendency to follow demonstrators, irrespective of the type of food site. These tendencies were unchanged when olfactory information was added in Experiment 3 and when motivation to compete increased in Experiment 4. These results suggest that only capuchin monkeys have the ability to solve an inference task when cued by social information.