Are We Smart Enough to Know How Smart Animals Are?(6)

One Asian elephant, named Happy, recognized her reflection. Marked with a white cross on her forehead above her left eye, she repeatedly rubbed the mark while standing in front of the mirror. She connected her reflection with her own body.12 By now, years later, Josh has tested many more animals at Think Elephants International, in Thailand, and our conclusion holds: some Asian elephants recognize themselves in the mirror. Whether the same can be said of African elephants is hard to tell, because up to now our experiments have resulted in a lot of destroyed mirrors due to this species’ tendency to examine new items with vigorous tusk action. This makes it hard to decide between poor performance and poor equipment. Obviously, the destruction of mirrors is no reason to conclude that African elephants lack mirror self-recognition. We are just dealing with species-typical treatment of novel items.

The challenge is to find tests that fit an animal’s temperament, interests, anatomy, and sensory capacities. Faced with negative outcomes, we need to pay close attention to differences in motivation and attention. One cannot expect a great performance on a task that fails to arouse interest. We ran into this problem while studying face recognition in chimpanzees. At the time, science had declared humans unique, since we were so much better at identifying faces than any other primate. No one seemed bothered by the fact that other primates had been tested mostly on human faces rather than those of their own kind. When I asked one of the pioneers in this field why the methodology had never moved beyond the human face, he answered that since humans differ so strikingly from one another, a primate that fails to tell members of our species apart will surely also fail at its own kind.

But when Lisa Parr, one of my coworkers at the Yerkes National Primate Research Center in Atlanta, tested chimpanzees on photographs of their own species, she found that they excelled at it. Selecting images on a computer screen, they would see one chimpanzee portrait immediately followed by a pair of others. One portrait of the pair would be a different picture of the same individual as presented before, while the other would show a different individual. Having been trained to detect similarity (a procedure known as matching to sample), the chimpanzees had no trouble recognizing which portrait most resembled the first. The apes even detected family ties. After having seen a female portrait, they were given a choice between two juvenile faces, one of which was the offspring of the female shown before. They picked the latter based purely on physical similarity, since they did not know any of the depicted apes in real life.13 In much the same way, we can leaf through someone else’s family album and quickly notice who are blood relatives and who are in-laws. As it turns out, chimpanzee face recognition is as keen as ours. It is now widely accepted as a shared capacity, especially since it engages the same brain areas in humans and other primates.14

In other words, what is salient to us—such as our own facial features—may not be salient to other species. Animals often know only what they need to know. The maestro of observation, Konrad Lorenz, believed that one could not investigate animals effectively without an intuitive understanding grounded in love and respect. He saw such intuitive insight as quite separate from the methodology of the natural sciences. To marry it productively with systematic research is both the challenge and the joy of studying animals. Promoting what he called the Ganzheitsbetrachtung (holistic contemplation), Lorenz urged us to grasp the whole animal before zooming in on its various parts.

One cannot master set research tasks if one makes a single part the focus of interest. One must, rather, continuously dart from one part to another—in a way that appears extremely flighty and unscientific to some thinkers who place value on strictly logical sequences—and one’s knowledge of each of the parts must advance at the same pace.15

The danger of ignoring this advice was amusingly illustrated when a famous study was replicated. In the study, domestic cats were placed in a small cage; they would wander about impatiently meowing—and in the process rub against the cage interior. In so doing, they accidentally moved a latch that opened a door, which allowed them to get out of the cage and eat a scrap of fish nearby. The more trials a cat performed, the quicker she’d escape. The investigators were impressed that all the tested cats showed the same stereotyped rubbing pattern, which they thought they had taught them with food rewards. First developed by Edward Thorndike in 1898, this experiment was considered proof that even seemingly intelligent behavior (such as escaping from a cage) can be fully explained by trial-and-error learning. It was a triumph of the “law of effect,” according to which behavior with pleasant consequences is likely to be repeated.16



Edward Thorndike’s cats were considered to have proven the “law of effect.” By rubbing against a latch inside a cage, a cat could open a door and escape, which would gain her a fish. Decades later, however, it was shown that the cats’ behavior had nothing to do with the prospect of reward. The animals escaped just as well without the fish. The presence of friendly people was all that was needed to elicit the flank rubbing that marks all feline greeting behavior. After Thorndike (1898).

When the American psychologists Bruce Moore and Susan Stuttard replicated this study decades later, however, they found that the cats’ behavior was nothing special. The cats performed the usual K?pfchengeben (German for “head giving”) that all felines—from house cats to tigers—use in greeting and courting. They rub their head or flank against the object of affection or, if the object of affection is inaccessible, redirect the rubbing to inanimate objects, such as the legs of a kitchen table. The investigators showed that the food reward was not needed: the only meaningful factor was the presence of friendly people. Without training, every caged cat that saw a human observer rubbed its head, flank, and tail against the latch and got out of the cage. Left alone, however, the cats were unable to escape, since they never performed any rubbing.17 Instead of a learning experiment, the classical study had been a greeting experiment! The replication was published under the telling subtitle “Tripping over the Cat.”