Fair Testing of Apes

Comment by Frans B. M. de Waal, Ph. D., Director, Living Links, EMORY University


This study claims that apes have equal technical but lower social intelligence than human toddlers. Although it is an impressive battery of tests, and an impressive data analysis, there is a glaring lack of equivalence between the way the children and apes have been tested, one that has been criticized before by others and myself.

It may seem that testing all subjects in the same way is fair and equal, but the children are tested by their own species (a human model), whereas the apes are tested by a stranger of another species (a human model). It is entirely possible that this doesn't make much difference in physical tasks, which after all focus on physical objects, but we do know that it makes a huge difference in relation to social tasks, which by definition focus on the interaction between subject and experimenter.

It is easier to process social information of one's own species. So, it is only seemingly that all subjects were tested in the same way.

Here at the Living Links Center we have in the past five years conducted many tests, which demonstrate that if chimpanzees are tested with their own kind they show impressive social learning. We have specialized in chimp-to-chimp testing precisely to circumvent the confound introduced by the human tester.

Examples can be found on the website of Living Links:
http://www.emory.edu/LIVING_LINKS/diffus.html
http://www.emory.edu/LIVING_LINKS/conform.html

Our results don't fit the "cultural intelligence" hypothesis mentioned in the study, and in fact they contradict the central claim of this new study, because they demonstrate that apes are very good at social tasks. Testing apes with models of their own species takes more time and patience, hence is harder to integrate in a large battery of tests, but it is ecologically more relevant (that is, it relates more closely to the context in which primate cognition evolved, which is of course among members of the same species).

So, unless I see more evidence, my conclusion is that all that this new study demonstrates is how cognitive performance is sensitive to who does the testing: one's own or another species.