Are We Smart Enough to Know How Smart Animals Are

by Frans de Waal

  To this end, the age-old definition of imitating as “doing an act from seeing it done” had to be changed to something narrower, something more advanced. The category true imitation was born, which requires one individual to intentionally copy another’s specific technique to achieve a specific goal.45 Merely duplicating behavior, such as one songbird learning another’s song, was not enough anymore: it had to be done with insight and comprehension. While imitation is common in lots of animals according to the old definition, true imitation is rare. We learned this fact from experiments in which apes and children were prompted to imitate an experimenter. They’d watch a human model open a puzzle box or rake in food with a tool. While the children copied the demonstrated action, the apes failed, hence the conclusion that other species lack imitative capacities and cannot possibly have culture. The comfort this finding brought to some circles greatly puzzled me, because it did not answer any fundamental questions either about animal culture or about human culture. All it did was draw a flimsy line in the sand.

  One can see here the interplay between the redefinition of a phenomenon and the quest to know what sets us apart, but also a deeper methodological problem, because whether apes imitate us or not is wholly beside the point. For culture to arise in a species, all that matters is that its members pick up habits from one another. There are only two ways to make a fair comparison in this regard (if we disregard the third option of having white-coated apes administer tests to both apes and children). One is to follow the wolf example: raise apes in a human home so that they are as comfortable as children around a human experimenter. The second is the so-called conspecific approach, which is to test a species with models of its own kind.

  The first solution yielded results right away, because several human-raised apes turned out to be as good at imitating members of our species as were young children.46 In other words, apes, like children, are born imitators and prefer to copy the species that raised them. Under most circumstances, this will be their own kind, but if reared by another species, they are prepared to imitate that one as well. Using us as models, these apes spontaneously learn to brush their teeth, ride bicycles, light fires, drive golf carts, eat with a knife and fork, peel potatoes, and mop the floor. It reminds me of suggestive stories on the Internet about dogs raised by cats, which show feline behavior such as sitting in boxes, crawling under tight spaces, licking their paws to clean their face, or sitting with their front legs tucked in.

  Another critical study was conducted by Victoria Horner, a Scottish primatologist, who later became my team’s lead expert on cultural learning. Together with Andrew Whiten of St. Andrews University, Vicky worked with a dozen orphan chimps at Ngamba Island, a sanctuary in Uganda. She acted like a mix between a mother and caretaker for the juvenile apes. Sitting next to her during tests, the juvenile apes were attached to Vicky and eager to follow her example. Her experiment created waves because as in Ayumu’s case, the apes proved to be smarter than the children. Vicky would poke with a stick into holes in a large plastic box, going through a series of holes until a candy would roll out. Only one hole mattered. If the box was made of black plastic, it was impossible to tell that some of the holes were just for show. A transparent box, on the other hand, made it obvious where the candies came from. Handed the stick and the box, young chimps mimicked only the necessary moves, at least with the transparent box. The children, on the other hand, mimicked everything that Vicky had demonstrated, including useless moves. They did so even with the transparent box, approaching the problem more like a magic ritual than as a goal-directed task.47

  With this outcome, the whole strategy of redefining imitation backfired! After all, it was the apes who best fit the new definition of true imitation. The apes were showing selective imitation, the sort that pays close attention to goals and methods. If imitation requires understanding, we have to give it to the apes, not to the children, who for lack of a better term, showed only dumb copying.

  What to do now? Premack complained that it was way too easy to make children look “foolish”—as if that were the goal of the experiment!—whereas in reality, he felt, there must be something wrong with the interpretation.48 His distress was genuine, showing to what degree the human ego gets in the way of dispassionate science. Promptly, psychologists settled on a narrative in which overimitation—a new term for children’s indiscriminate copying—is actually a brilliant achievement. It fits our species’ purported reliance on culture, because it makes us imitate behavior regardless of what it is good for; we transmit habits in full, without every individual making his or her own ill-informed decisions. Given the superior knowledge of adults, the best strategy for a child is to copy them without question. Blind faith is the only truly rational strategy, it was concluded with some relief.

  Even more striking were Vicky’s studies at our field station in Atlanta, where we started a decade-long research program in collaboration with Whiten, focusing entirely on the conspecific approach. When chimps were given a chance to watch one another, incredible talents for imitation manifested themselves. Apes truly do ape, allowing behavior to be faithfully transmitted within the group.49 A video of Katie imitating her mother, Georgia, offers a nice example. Georgia had learned to flip open a little door in a box, then stick a rod deep into the opening to retrieve a reward. Katie had watched her mother do this five times, following her every move and smelling Georgia’s mouth every time she got a reward. After her mother was moved to another room, Katie could finally access the box herself. Even before we had added any rewards, she flipped open the door with one hand and inserted the rod with the other. Sitting like this, she looked up at us on the other side of the window and impatiently rapped it, while grunting, as if telling us to hurry up. As soon as we pushed the reward into the box, she retrieved it. Before ever being rewarded for these actions, Katie perfectly duplicated the sequence she had watch Georgia perform.

  Rewards are often secondary. Imitation without reward is of course common in human culture, such as when we mimic hairstyles, accents, dance steps, and hand gestures, but it is also common in the rest of the primate order. The macaques on the Arashiyama mountaintop in Japan customarily rub pebbles together. The young learn to do it without any reward other than perhaps the noise associated with it. If one case refutes the common notion that imitation requires reward, it is this weird behavior, about which Michael Huffman, an American primatologist who has studied it for decades, notes, “It is likely that the infant is first exposed in utero to the click-clacking sounds of stones as its mother plays, and then exposed visually as one of the first activities it sees after birth, when its eyes begin to focus on objects around it.”50

  The word fashion was first used in relation to animals by Köhler, whose apes invented new games all the time. They’d march single file around and around a post, trotting in the same rhythm with emphasis on one stamping foot, while the other foot stepped lightly, wagging their heads in the same rhythm, all acting in synchrony as if in a trance. For months our own chimps had a game we called cooking. They’d dig a hole in the dirt, collect water by holding a bucket under a faucet, and dump water into the hole. They’d sit around the hole poking in the mud with a stick as if stirring soup. Sometimes there were three or four such holes in operation at the same time, keeping half the group busy. At a chimpanzee sanctuary in Zambia, scientists followed the spread of yet another meme. One female was the first to stick a straw of grass into her ear, letting it hang out while walking around and grooming others. Over the years, other chimps followed her example, with several of them adopting the same new “look.”51

  Fashions come and go in chimps as in humans, but some habits we find in only one group and not in another. Typical is the hand-clasp grooming of some wild chimpanzee communities, in which two individuals hold hands above their heads while grooming each other’s armpits with their other hands.52 Since habits and fashions often spread without any associated rewards, social learning is truly social. It is about co
nformity instead of payoffs. Thus an infant male chimp may mimic the charging display of the alpha male who always bangs a specific metal door to accentuate his performance. Ten minutes after the male has finished his performance—a dangerous activity, during which mothers keep their children near—the little son is let go. With all his hair on end, he goes to bang on the same door as his role model.

  Having documented numerous such examples, I developed the idea of Bonding- and Identification-based Observational Learning (BIOL). Accordingly, primate social learning stems from an urge to belong. BIOL refers to conformism born from the desire to act like others and to fit in.53 It explains why apes imitate their own kind far better than the average human, and why, among humans, they imitate only those whom they feel close to. It also explains why young chimps, especially females,54 learn so much from their mothers, and why high-status individuals are favorite models. This preference is also known in our own societies, in which advertisements feature celebrities showing off watches, perfumes, and cars. We love to emulate the Beckhams, Kardashians, Biebers, and Jolies. Might the same apply to apes? In one experiment, Vicky spread brightly colored plastic chips around in an enclosure, which the chimps could collect and carry to a container in exchange for rewards. Exposed to the sight of a top-ranking group member trained to drop tokens into one container and a bottom-ranking one trained to use a different container, the colony massively followed in the footsteps of the more prestigious member.55

  As evidence mounted regarding imitation in apes, other species inevitably joined the ranks, showing similar capacities.56 There are now compelling studies on imitation in monkeys, dogs, corvids, parrots, and dolphins. And if we take a broader view, we have even more species to consider because cultural transmission is widespread. To return to dogs and wolves, a recent experiment applied the conspecific approach to canine imitation. Instead of following human instructions, both dogs and wolves saw a member of their own species manipulate a lever to open the lid of a box with hidden food. Next, they were allowed to try the same box themselves. This time the wolves greatly outsmarted the dogs.57 Wolves may be poor at following human pointing, but when it comes to picking up hints from their own kind, they beat dogs. The investigators ascribe this contrast to attention rather than cognition. They point out that wolves watch one another more closely as they rely on the pack for survival, whereas dogs rely on us.

  Clearly, it is time for us to start testing animals in accordance with their biology and move away from human-centric approaches. Instead of making the experimenter the chief model or partner, we better keep him or her in the background. Only by testing apes with apes, wolves with wolves, and children with human adults can we evaluate social cognition in its original evolutionary context. The one exception may be the dog, which we domesticated (or which domesticated itself, as some believe) to bond with us. Humans testing dog cognition may actually be a natural thing to do.

  Moratorium

  Having escaped the Dark Ages in which animals were mere stimulus-response machines, we are free to contemplate their mental lives. It is a great leap forward, the one that Griffin fought for. But now that animal cognition is an increasingly popular topic, we are still facing the mindset that animal cognition can be only a poor substitute of what we humans have. It can’t be truly deep and amazing. Toward the end of a long career, many a scholar cannot resist shining a light on human talents by listing all the things we are capable of and animals not.58 From the human perspective, these conjectures may make a satisfactory read, but for anyone interested, as I am, in the full spectrum of cognitions on our planet, they come across as a colossal waste of time. What a bizarre animal we are that the only question we can ask in relation to our place in nature is “Mirror, mirror on the wall, who is the smartest of them all?”

  Keeping humans in their preferred spot on that absurd scale of the ancient Greeks has led to an obsession with semantics, definitions, and redefinitions, and—let’s face it—the moving of goalposts. Every time we translate low expectations about animals into an experiment, the mirror’s favorite answer sounds. Biased comparisons are one ground for suspicion, but the other is the touting of absent evidence. I have lots of negative findings in my own drawers that have never seen the light since I have no idea what they mean. They may indicate the absence of a given capacity in my animals, but most of the time, especially if spontaneous behavior suggests otherwise, I am unsure that I have tested them in the best possible way. I may have created a situation that threw them off, or presented the problem in such an incomprehensible fashion that they didn’t even bother to solve it. Recall the low opinion scientists held of gibbon intelligence before their hand anatomy was taken into account, or the premature denials of mirror self-recognition in elephants based on their reaction to an undersize mirror. There are so many ways to account for negative outcomes that it is safer to doubt one’s methods before doubting one’s subjects.

  Books and articles commonly state that one of the central issues of evolutionary cognition is to find out what sets us apart. Entire conferences have been organized around the human essence, asking “What makes us human?” But is this truly the most fundamental question of our field? I beg to differ. In and of itself, it seems an intellectual dead end. Why would it be any more critical than knowing what sets cockatoos or beluga whales apart? I am reminded of one of Darwin’s random musings: “He who understands baboon would do more towards metaphysics than Locke.”59 Every single species has profound insights to offer, given that its cognition is the product of the same forces that shaped ours. Imagine a medical textbook that declared that its discipline’s central issue is to find out what is unique about the human body. We would roll our eyes, because even though this question is mildly intriguing, medicine faces far more basic issues related to the functioning of hearts, livers, cells, neural synapses, hormones, and genes.

  Science seeks to understand not the rat liver or the human liver but the liver, period. All organs and processes are a great deal older than our species, having evolved over millions of years with a few modifications specific to each organism. Evolution always works like this. Why would cognition be any different? Our first task is to find out how cognition in general operates, which elements it requires to function, and how these elements are attuned to a species’s sensory systems and ecology. We want a unitary theory that covers all the various cognitions found in nature. To create space for this project, I recommend placing a moratorium on human uniqueness claims. Given their miserable track record, it is time to rein them in for a few decades. This will allow us to develop a more comprehensive framework. One day years from now, we may then return to our species’s particular case armed with new concepts that allow a better picture of what is special—and what not—about the human mind.

  One aspect we might focus on during this moratorium is an alternative to overly cerebral approaches. I have already mentioned that perspective taking is likely tied to bodies, and the same applies to imitation. After all, imitation requires that another individual’s body movements are perceived and translated into one’s own body movements. Mirror neurons (special neurons in the motor cortex that map another’s actions onto one’s own bodily representations in the brain) are often thought to mediate this process, and it is good to realize that those neurons were discovered not in humans but in macaques. Even though the precise connection remains a point of debate, imitation likely is a bodily process facilitated by social closeness.

  This view is quite different from the cerebral one according to which it all depends on the understanding of cause-effect relations and goals. Thanks to an ingenious experiment by the British primatologist Lydia Hopper, we know which view is correct. Hopper presented chimps with a so-called ghost box controlled by fishing lines. The box magically opened and closed by itself, producing rewards. If technical insight were all that mattered, watching such a box should suffice, as it shows all the necessary actions and consequences. But in fact, letting chimps watch the ghost bo
x ad nauseam taught them nothing. Only after seeing an actual chimp operate the same box, did they learn how to get the rewards.60 Thus for imitation to occur, apes need to connect to a moving body, preferably one of their own species. Technical understanding is not the key.61

  To find out how bodies interact with cognition, we have incredibly rich material to work with. Adding animals to the mix is bound to stimulate the up-and-coming field of “embodied cognition,” which postulates that cognition reflects the body’s interactions with the world. Until now, this field has been rather human-focused while failing to take advantage of the fact that the human body is only one of many.

  Consider the elephant. It combines a very different body with the brainpower to achieve high cognition. What is the largest land mammal doing with three times as many neurons as our own species? One may downplay this number, arguing that it has to be corrected for body mass, but such corrections are more suited to brain weight than to number of neurons. In fact, it has been proposed that absolute neuron count, regardless of brain or body size, best predicts a species’ mental powers.62 If so, we’d better pay close attention to a species that has vastly more neurons than we do. Since most of these neurons reside in the elephant’s cerebellum, some feel they carry less weight, the assumption being that only the prefrontal cortex matters. But why take the way our brain is organized as the measure of all things and look down on subcortical areas?63 For one thing, we know that during hominoid evolution, our cerebellum expanded even more than our neocortex. This suggests that for our species, too, the cerebellum is critically important.64 It is now up to us to find out how the remarkable neuron count of the elephant brain serves its intelligence.