Are We Smart Enough to Know How Smart Animals Are

by Frans de Waal

  Not even with regard to the Monster of Loch Ness or the Abominable Snowman will you ever hear anyone claim to have proven its nonexistence, even though this would fit the expectations of most of us. And why do governments still spend billions of dollars to search for extraterrestrial civilizations while there is no shred of evidence to encourage this quest? Isn’t it time to conclude once and for all that these civilizations simply don’t exist? But this conclusion will never be reached. That respected psychologists ignore the recommendation to tread lightly around absent evidence is most puzzling, therefore. One reason is that they test apes and children in the same manner—at least in their minds—while coming up with contrary results. Applying a battery of cognitive tasks to both apes and children and finding not a single result in the apes’ favor, they tout the differences as proof of human uniqueness. Otherwise, why didn’t the apes fare better? To understand the flaw in this logic, we need to go back to Clever Hans, the counting horse. But instead of using Hans to illustrate why animal capacities are sometimes overrated, this time we are concerned with the unfair advantage that human capacities enjoy.

  The outcome of ape-child comparisons themselves suggests the answer. When tested on physical tasks, such as memory, causality, and the use of tools, apes perform at about the same level as two-and-a-half-year-old children, but when it comes to social skills, such as learning from others or following others’ signals, they are left in the dust.34 Social problem solving requires interaction with an experimenter, however, whereas physical problem solving does not. This raises the possibility that the human interface is key. The typical format of an experiment is to let apes interact with a white-coated barely familiar human. Since experimenters are supposed to be bland and neutral, they do no not engage in schmoozing, petting, or other niceties. This doesn’t help make the ape feel at ease and identify with the experimenter. Children, however, are encouraged to do so. Moreover, only the children are interacting with a member of their own species, which helps them even more. Nevertheless, experimenters comparing apes and children insist that all their subjects are treated exactly the same. The inherent bias of this arrangement has become harder to ignore, however, now that we know more about ape attitudes. A recent eye-tracking study (which precisely measured where subjects looked) reached the unsurprising conclusion that apes consider members of their own species special: they follow the gaze of another ape more closely than they follow the human gaze.35 This may be all we need to explain why apes fare poorly on social tasks presented by members of our species.

  There are only a dozen institutes that test ape cognition, and I have visited most of them. I have noticed procedures in which humans barely interact with their subjects and ones in which they have close physical contact. The latter can safely be done only by those who raised the apes themselves or have at least known them since infancy. Since apes are much stronger than we are and have been known to kill people, the up-close-and-personal approach is not for everyone. The other extreme derives from the traditional approach in the psychology lab: carrying a rat or a pigeon into a testing room with as little contact as possible. The ideal here is a nonexistent experimenter, meaning the absence of any personal relation. In some labs apes are called into a room and given only a few minutes to perform before they are sent out again without any playful or friendly contact, almost like a military drill. Imagine if children were tested under such circumstances: how would they fare?

  At our center in Atlanta, all our chimps are reared by their own kind and so are more ape- than human-oriented. They are “chimpy,” as we say, relative to apes that have a less social background or were raised by humans. We never share the same space with them, but we do interact through the bars, and we always play or groom before testing. We talk to them to put them at ease, give them goodies, and in general try to create a relaxed atmosphere. We want them to look upon our tasks as a game rather than as work, and certainly never put them under pressure. If they are tense because of events in their group, or because another chimp is banging on the outside door or hooting his lungs out, we wait until everyone has calmed down, or we reschedule the test. There is no point testing apes who are not ready. If such procedures are not followed, apes may act as if they don’t understand the problem at hand, whereas the real issue is high anxiety and distraction. Many negative results in the literature may be explained this way.

  The methodology sections of scientific papers rarely offer a look in the “kitchen,” but I think it is crucial. My own approach has always been to be firm and friendly. Firm, meaning that we are consistent and don’t make capricious demands but also don’t let the animals walk all over us, such as when they only want to play around and get free sweets. But we are also friendly, without punishment, anger, or attempts to dominate. The latter still happens all too often in experiments and is counterproductive with such headstrong animals. Why would an ape follow the points and prompts of a human experimenter whom he sees as a rival? This is another potential source of negative outcomes.

  My own team typically cajoles, bribes, and sweet-talks its primate partners. Sometimes I feel like a motivational speaker, such as when Peony, one of our oldest females, ignored a task that we had set up for her. For twenty minutes, she lay in the corner. I sat down right next to her and told her, in a calm voice, that I didn’t have all day and it would be great if she would get going. She slowly got up, glancing at me, and strolled to the next room, where she sat down for the task. Of course—as discussed in the previous chapter in relation to Robert Yerkes—it is unlikely that Peony followed the details of what I had said. She was sensitive to my tone of voice and knew all along what we wanted.

  However good our relations with apes, the idea that we can test them in exactly the same way we test children is an illusion of the same order as someone throwing both fish and cats into a swimming pool and believing he is treating them the same way. Think of the children as the fish. While testing them, psychologists smile and talk all the time, giving instructions where to look or what to do. “Look at the little froggy!” tells a child so much more than an ape will ever know about the green plastic blob in your hand. Moreover, children are usually tested with a parent in the room, often sitting on their lap. Having permission to run around and an experimenter of their own species, they have an enormous leg up over the ape sitting behind bars without verbal hints or parental support.

  True, developmental psychologists try to reduce the influence of parents by telling them not to talk or point, and they may give them sunglasses or a baseball cap to cover their eyes. These measures, however, reveal their woeful underestimation of the power of a parent’s motivation to see their child succeed. When it comes to their precious offspring, few people care about the objective truth. We can be glad that Oskar Pfungst designed far more rigorous controls while examining Clever Hans. In fact, Pfungst found that the wide-rimmed hat of the horse owner greatly benefited Hans, since hats amplify head movements. In the same way that the owner vociferously denied his effect on the horse even after it had been proven, parents of children may be completely honest about suppressing cues. But adults have far too many ways to unintentionally guide the choices of a child on their lap, through slight body movements, gaze direction, halted breathing, sighs, squeezes, strokes, and whispered encouragements. Letting parents attend the testing of a child is asking for trouble—the sort of trouble we avoid in animal testing.

  The American primatologist Allan Gardner—who was first to teach American Sign Language to an ape—discussed human biases under the heading “Pygmalion leading.” Pygmalion, in ancient mythology, was a Cypriot sculptor who fell in love with his own statue of a woman. The story has been used as a metaphor of how teachers raise the performance of certain children by expecting the world of them. They fall in love with their own prediction, which serves as a self-fulfilling prophecy. Remember how Charlie Menzel felt that only people who hold apes in high esteem will fully appreciate what they are trying to communicate? His was a
plea for raised expectations, which unfortunately is not the situation apes typically face. Children, in contrast, are treated in such a nurturing manner that they inevitably confirm the mental superiority ascribed to them.36 Experimenters admire and stimulate them from the outset, making them feel like fish in the water, whereas they often treat apes more like albino rats: keeping them at a distance, and in the dark, while depriving them of the verbal encouragement we offer members of our own species.

  The cognition of children and apes is tested in superficially similar ways. Yet children are not kept behind a barrier; they are talked to and often sit on their parents’ laps, all of which helps them connect with the experimenter and receive unintentional hints. The greatest difference, however, is that only apes face a member of another species. Given how much these comparisons disadvantage one class of subjects, they remain inconclusive.

  Needless to say, I view most ape-child comparisons as fatally flawed.37

  Recall that apes have been tested for ToM by having them guess what humans know or don’t know. The problem here is that captive apes have every reason to believe that we are omniscient! Suppose my assistant calls to tell me that Socko, the alpha male, has been wounded in a fight. I head over to the field station, walk up to him, and ask him to turn around, which he does—having known me since he was a baby—to show me his behind with the gash. Now try to look at this from Socko’s perspective. Chimps are smart animals, always trying to figure out what’s going on. Of course, he wonders how I know about his injury—I must be an all-knowing god. As such, human experimenters are about the last to be used to find out if apes understand the connection between seeing and knowing. All we are testing is the ape’s theory of the human mind. It is no accident that we made substantial progress only after egg-hunt scenarios pitted apes against other apes.

  One area of cognitive research that has been lucky to escape the species barrier is the study of ToM in animals that are so different from us that everyone understands that humans are unsuitable partners. This has been the case with corvids. Since a true animal watcher never takes a break, the British ethologist Nicky Clayton made a major discovery over lunch at the University of California at Davis. While sitting at an outdoors terrace, she saw Western scrub jays fly off with scraps stolen from the tables. They not only cached them but also guarded them against thieves. If another bird saw where they hid their food, it was bound to disappear. Clayton noticed that after their rivals left the scene, many of the jays returned to rebury their treasures. In follow-up research with Nathan Emery in their lab at Cambridge, she let jays cache mealworms either in private or while being watched by another jay. Given a chance, the jays quickly re-cached their worms at a new location—but only if they had been watched. They seemed to understand that the food was safe if no other birds had any information. Moreover, only birds who themselves had pilfered others’ food re-cached their own. Following the dictum “It takes a thief to know a thief,” the jays seemed to extrapolate from their own criminality to that of others.38

  A Western scrub jay caches a mealworm while being watched from behind glass by another. As soon as he is alone, the jay will quickly rehide his treasures, as if realizing that the other knows too much.

  Again, we recognize the Menzel-like design of this experiment, which is even more obvious in a study of perspective-taking ravens. The Austrian zoologist Thomas Bugnyar had a low-ranking male who was expert at opening canisters that contained goodies, but this male often lost his prize to a bullying and stealing dominant male. The low-ranking male, however, learned to distract his competitor by enthusiastically opening empty containers and making as if to eat from them. When the dominant bird found out, “he got very angry, and started throwing things around.” Bugnyar further found that when ravens approach hidden food, they take into account what other ravens know. If their competitors have the same knowledge, they hurry to get there first. But if the others are ignorant, they take their time.39

  All in all, animals do plenty of perspective taking, from being aware of what others want to knowing what others know. A few frontiers are left, of course, such as whether they recognize when others have the wrong knowledge. In humans, researchers test this issue with the so-called false-belief task. But since these refinements are hard to evaluate without language, we face a dearth of animal data. Still, even if the remaining differences hold up, there is little doubt that the blanket assertion that ToM is uniquely human must be downgraded to a more nuanced, gradualist view.40 Humans probably possess a fuller understanding of one another, but the contrast with other animals is not stark enough that extraterrestrials would automatically pick ToM as the chief marker that sets us apart.

  While this conclusion is based on solid data from repeated experiments, let me add one anecdote that captures the phenomenon in an entirely different way. At the Yerkes Field Station—where apes live in grassy open-air enclosures in the warm Georgia weather—I developed a special bond with an exceptionally bright female chimp named Lolita. One day Lolita had a new baby, and I wanted to get a good look at it. This is hard to do since a newborn ape is really no more than a little dark blob against its mother’s dark tummy. I called Lolita out of her grooming huddle, high up in the climbing frame, and pointed at her belly as soon as she sat down in front of me. Looking at me, she took the infant’s right hand in her right hand and its left hand in her left hand. It sounds simple, but given that the baby was ventrally clinging to her, she had to cross her arms to do so. The movement resembled that of people crossing their arms when grabbing a T-shirt by its hems in order to take it off. She then slowly lifted the baby into the air while turning it around its axis, unfolding it in front of me. Suspended from its mother’s hands, the baby now faced me instead of her. After it made a few grimaces and whimpers—infants hate to lose touch with a warm belly—Lolita quickly tucked it back into her lap.

  With this elegant motion, Lolita demonstrated that she realized I would find the front of her newborn more interesting than its back. To take someone else’s perspective represents a huge leap in social evolution.

  Spreading Habits

  Decades ago friends of mine were outraged by a newspaper article that ranked the smartest canine breeds. They happened to own the breed that was dead last on the list: the Afghan hound. Naturally, the top breed was the border collie. My insulted friends argued that the only reason Afghans were considered dim-witted is that they are independent-minded, stubborn, and unwilling to follow orders. The newspaper’s list was about obedience, they said, not intelligence. Afghans are perhaps more like cats, which are not beholden to anyone. This is no doubt why some people rate cats as less intelligent than dogs. We know, however, that a cat’s lack of response to humans is not due to ignorance. A recent study showed that felines have no trouble recognizing their owner’s voice. The deeper problem is that they don’t care, prompting the study’s authors to add: “the behavioral aspects of cats that cause their owners to become attached to them are still undetermined.”41

  I had to think of this story when dog cognition emerged as a hot topic. Dogs were depicted as smarter than wolves, perhaps even apes, because they paid better attention to human pointing gestures. A human would point at one out of two buckets, and the dog would check that particular bucket out for a reward. Scientists concluded that domestication had given dogs extra intelligence compared to their ancestors. But what does it mean that wolves fail to follow human pointing? With a brain about one-third larger than a dog’s, I bet a wolf could outsmart its domesticated counterpart anytime—yet all we go by is how they react to us. And who says that the difference in reaction is inborn, a consequence of domestication, and not based on familiarity with the species doing the pointing? It is the old nature-nurture dilemma. The only way to determine how much of a trait is produced by genes and how much by the environment is to hold one of these two constant to see what difference the other one makes. It is a complex problem that is never fully resolved. In the dog-wolf comparison, this
would mean raising wolves like dogs in a human household. If they still differ, genetics might be at play.

  Raising wolf puppies in the home is a hellish job, though, since they are exceptionally energetic and less rule-bound than dog puppies, chewing up everything in sight. When dedicated scientists raised wolves this way, the nurture hypothesis came out the winner. Human-raised wolves followed hand points as well as dogs. A few differences persisted, though, such as that wolves looked less at human faces than dogs and were more self-reliant. When dogs tackle a problem they cannot solve, they look back at their human companion to get encouragement or assistance—something that wolves never do. Wolves keep trying and trying on their own. Domestication may be responsible for this particular difference. Instead of intelligence, though, it seems more a question of temperament and relations with us—those weird bipedal apes that the wolf evolved to fear and the dog was bred to please.42 Dogs, for example, engage in lots of eye contact with us. They have hijacked the human parental pathways in the brain, making us care about them in almost the same way that we care about our children. Dog owners who stare into their pet’s eyes experience a rapid increase in oxytocin—a neuropeptide involved in attachment and bonding. Exchanging gazes full of empathy and trust, we enjoy a special relationship with the dog.43

  Cognition requires attention and motivation, yet it cannot be reduced to either. As we have seen, the same problem troubles the comparison between apes and children, an issue that popped up again in the controversy around animal culture. Whereas in the nineteenth century, anthropologists were still open to the possibility of culture outside our own species, in the twentieth they began to write culture with a capital C while claiming that the trait is what makes us human. Sigmund Freud considered culture and civilization a victory over nature, while the American anthropologist Leslie White, in a book ironically entitled The Evolution of Culture, declared: “Man and culture originated simultaneously—this by definition.”44 Naturally, when the first reports of animal culture came along, defined as habits learned from others—from potato-washing macaques and nut-cracking chimpanzees to bubble-net-hunting humpback whales—they faced a wall of hostility. One line of defense against this offensive notion was to focus on the learning mechanism. If it could be shown that human culture relies on distinct mechanisms, so the thinking went, we might be able to claim culture for ourselves. Imitation became the holy grail of this battle.