How do you give an animal an IQ test?

Illustration for article titled How do you give an animal an IQ test?

We humans are notorious for underestimating the intelligence of animals. As Frans de Waal writes in the Wall Street Journal, scientists have consistently botched up intelligence tests administered to animals — tests that are often human-centric and stacked unfavorably against them. Is there a better way to estimate animals' mental capacities?


Top image: An 11-year-old chimpanzee named Ayumu performs a memory test with randomly placed consecutive numerals. Chimps consistently outperform humans in this task — and by a wide margin.

Frans de Waal is a primatologist and ethologist who works at the Yerkes National Primate Research Center in Atlantic. In his latest book, The Bonobo and the Atheist: In Search of Humanism among the Primates, de Waal explores the idea that moral behavior does not begin and end with religion, but is instead a product of evolution.

Illustration for article titled How do you give an animal an IQ test?

And as he points out in a recent Wall Street Journal OpEd, it’s not just the moral life of animals that we’re underestimating, but the scope and scale of their intelligence as well.

Testing for intelligence, admits de Waal, is one of the “thorniest questions facing science today.” But the time has come for us to start getting it right.

He writes:

Can an octopus use tools? Do chimpanzees have a sense of fairness? Can birds guess what others know? Do rats feel empathy for their friends? Just a few decades ago we would have answered "no" to all such questions. Now we're not so sure.

Experiments with animals have long been handicapped by our anthropocentric attitude: We often test them in ways that work fine with humans but not so well with other species. Scientists are now finally meeting animals on their own terms instead of treating them like furry (or feathery) humans, and this shift is fundamentally reshaping our understanding.


De Waal points to several examples of poorly designed tests, including a mirror test for elephants in which a small mirror was placed on the ground such that they could only see their legs; the elephants promptly failed (when done right, elephants pass this test with flying colors). Or a facial recognition test in which chimps were asked to identify human faces rather than chimp faces (which, needless to say, resulted in some very poor outcomes; the chimps performed significantly better on subsequent tests involving chimp faces).

These flawed tests are especially problematic when trying to establish how smart apes are relative to human children:

To see how their cognitive skills compare, scientists present both species with identical problems, treating them exactly the same. At least this is the idea. But the children are held by their parents and talked to ("Watch this!" "Where is the bunny?"), and they are dealing with members of their own kind. The apes, by contrast, sit behind bars, don't benefit from language or a nearby parent who knows the answers, and are facing members of a different species. The odds are massively stacked against the apes, but if they fail to perform like the children, the invariable conclusion is that they lack the mental capacities under investigation.

A recent study, tracking the pupil movements of chimpanzees, found that they followed the gaze of members of their own species far better than that of humans. This simple finding has huge implications for tests in which chimpanzees need to pay attention to human experimenters. The species barrier they face may fully explain the difference in performance compared with children.


The signs of nonhuman animal intelligence are right under our noses, argues de Waal, we just need to learn where and how to look.

There’s plenty more to this article at the WSJ.

Image: University of Kyoto.




It seems to me like the best we can do is to compare very specific areas of cognition across species. For example, we already know scrub jays (and many mammals) have spatial reasoning and spatial memory that are insanely better than humans. The desire to test for "higher" cognition is shown in doing theory of mind experiments and mirror recognition tests. As others have noted, the idea of ascribing a general "intelligence" measurement is impossible - for now I think we should try to describe the intelligence of different species by what cognitive tasks they're relatively good at and what they're relatively bad at.

Of course, all of this is moot if the extreme anthropocentric bias mentioned above isn't removed first, but from my experience with modern research and prominent ethologists, this is well underway in the respected scientific community.