A Failure Of Intelligence Testing, With Chimpanzees : 13.7: Cosmos And Culture Apes' intelligence tests vary so much from kids' that comparative results may be invalid. Anthropologist Barbara J. King explores new insights from psychology.

A Failure Of Intelligence Testing, This Time With Chimpanzees

Guenterguni/Getty Images
A family of chimpanzees sits on a tree in Gombe Stream National Park in Western Tanzania.
Guenterguni/Getty Images

In 1981, the evolutionary biologist Stephen Jay Gould's book The Mismeasure of Man hit the presses.

A take-down of studies purporting to demonstrate that the intelligence of humans is genetically determined — and that some human groups (read "white Western Europeans") are innately superior — the book exposed interpretive bias and scientific racism in the measurement of human intelligence. Different environmental histories across human groups, in fact, affect testing outcomes in significant ways: There is no innate superiority due to genes.

The Mismeasure of Man ignited ferocious discussion (and the occasional subsequent correction) that has continued even in recent years across biology, anthropology, psychology and philosophy: Its argument mattered not only for how we do science, but how science entangles with issues of social justice.

Now, psychologists David A. Leavens of the University of Sussex, Kim A. Bard of the University of Portsmouth, and William D. Hopkins of Georgia State University have framed their new Animal Cognition article, "The mismeasure of ape social cognition," around Gould's book. Ape (especially chimpanzee) social intelligence, the authors say, has been routinely mismeasured because apes are tested in comprehensively different circumstances from the children with whom they are compared — and against whose performance theirs is found to be lacking.

Leavens et al. write:

"All direct ape-human comparisons that have reported human superiority in cognitive function have universally failed to match the groups on testing environment, test preparation, sampling protocols, and test procedures."

Confounding factors in these experiments, in other words, are essentially fatal: They render the conclusions unreliable. The testing procedures are so different between apes and children that it becomes impossible to isolate evolutionary history as the explanatory factor when differences in social cognition are uncovered.

Gould's The Mismeasure of Man is evoked as a key antecedent by Leavens et al., in that the "obvious bias and special pleading toward nativist explanations for systematic group differences in test preparedness [that Gould exposed] seems antiquated to contemporary scientists, but it is entirely characteristic of cross-species comparisons between humans and apes."

Let's get specific. What are some of these testing differences to which Leavens et al. refer?

Last week, I asked Leavens to choose a particularly powerful example from the paper. In an email message, he described a 2014 study that compared chimpanzees and children:

"Human 1-year-old infants were compared with a group of apes that were, on average, 19 years old (a sampling confound). In this study there was a test phase in which the participants could ask a distant experimenter to replenish a cache of toys (humans) or food (apes — another confound).

[The researchers'] claim was that if a participant stayed where they were, without moving, then this indicated that the participant possessed an appreciation of the psychological state of common ground. Astonishingly, the humans were tested at distances of .95 meters and 1.8 meters between themselves and the experimenter, but the great apes were tested at distances of about 6 meters.

It turned out that about half of the human babies communicated about the distant toys without moving from their original places, whereas none of the apes did so; all of the apes locomoted the full distance to the experimenter and communicated about the food from that position."

Does this result mean that the apes did not understand common ground between themselves and the experimenters? Leavens et al., of course, say no such conclusion can be reached, because the testing factors were wildly divergent.

Working with a team of researchers, Leavens subsequently carried out similar research with 166 chimpanzees, but using 1.5 meters as the distance for the food. That is, he equalized the testing conditions in one central way.

"We found," Leavens told me, "that, like the human babies, about half of the chimpanzees who communicated, communicated from a distance. So, a simple tweak of distance rendered similar response profiles between humans and apes, notwithstanding a number of other key differences" between the two studies.

The catalog of testing differences in the Leavens, Bard, and Hopkins review is striking: Apes are tested through cage bars, and humans are not. Most apes had been isolated, as the authors put it, "from early intensive exposure to human nonverbal conventions of give and take and of daily exposure to nonverbal references to entities" whereas the human kids, of course, had not been so isolated. This last difference means that the children had a good deal of task-relevant preparation that the apes lacked. But, remember: Developmental histories were not taken into account in interpreting cognitive differences found.

When I asked Leavens what he saw as the "value added" to the article by framing it around The Mismeasure of Man, he referred to:

"Gould's straightforward, easy-to-understand methodological critique of racist psychological science in the early 20th century, which presaged the contemporary speciesist psychological science in the early 21st century."

As someone who writes about animal cognition, I found the Leavens, Bard, and Hopkins article to be incredibly important. It, like Gould's book, has implications for both science and also for the understanding and treatment of living populations or individuals — in this case, highly sentient apes.

One aspect of the article gave me pause, and I wanted to ask Leavens about it. The authors make suggestions for how to avoid the problems that currently plague this sort of comparative research. Researchers could train apes in task-relevant experiences in order to level the playing field. It takes kids about 18 months, for instance, to follow another person's pointing gesture to a location behind themselves. An ape who is tested on such a pointing gesture would, then, be given 18 months of comparable cultural exposure to such gestures.

Cross-fostering of apes by humans is also suggested, though, and that's where I feel concern. Leavens et al. do note "ethical implications" of raising apes by bringing them into the human world and thus isolating them from other apes. I asked if he worried that highlighting cross-fostering will encourage studies that many (including me) consider unethical in this day and age?

Leavens' reply to this question was lengthy, and I can only offer a fraction of it here. Basically, he weighs the cost to apes — too high when apes are taken from their mothers and peers to be raised by humans, only to be housed again later with their own species with whom they no longer have a connection — against the benefits to science, because "cross-fostering has the ability to inform our understanding of developmental plasticity." The best ethical outcome, Leavens believes, may be a two-species, two-cultures scenario in which apes are raised with other apes as well as with their human social partners.

I mention the cross-fostering conundrum because it highlights, for me, the complexities of asking apes to inform our science in ways we simply never (thankfully) ask children: We don't lock kids up in captivity in order to carry out experiments on them, whether alone or with their mothers or peers.

The point isn't that I'm against all noninvasive experiments on apes to assess their cognition. I'm not.

But I think we would do well to keep ethics as much front and center in our discussions of measuring ape intelligence as Leavens et al. — quite brilliantly — do with the specifics of the testing procedures.

Barbara J. King is an anthropology professor emerita at the College of William and Mary. She often writes about the cognition, emotion and welfare of animals, and about biological anthropology, human evolution and gender issues. Barbara's new book is Personalities on the Plate: The Lives and Minds of Animals We Eat. You can keep up with what she is thinking on Twitter: @bjkingape