The Tarzan Syndrome

Only apes, it seems, alone among all the animals, can truly distinguish themselves fromt he world around them. But only the naked apes, apparently, can conceive of not just self but other.

By Karen Wright|Friday, November 01, 1996
Thus begins the syncopated lament of an orangutan named King Louie in the animated film The Jungle Book. Louie is confiding his envy of the human race to the man-cub Mowgli, whom he has recently, if forcibly, befriended. Ooh be dooh, he explains. I wanna be like you/I wanna walk like you/Talk like you, too. . . .

At the New Iberia Research Center in southwestern Louisiana, relations between humans and apes are far less flattering. Rather than serenade a visiting hominid, certain adolescent chimpanzees are likely to fill their mouths with water and then send the fluid out between their front teeth with a faucetlike force aimed at the visitor’s face, chest, or notebook. Along with the water comes a generous helping of half-chewed food and saliva. Ooh be dooh. Here’s what we think of you.

Brandy, no. No. Stop that. Stop it. Kara, you too. C’mon guys. Cut it out. The demands come from Daniel Povinelli, director of the center’s laboratory of comparative behavioral biology, who is wearing a smartly pressed white shirt and standing well within spitting range of the chimps’ chain-link compound. He and a small crew of caretakers raised these seven apes from toddlerhood, but the animals ignore him and continue their spirited greeting. Between the ages of four and five they start to figure out that they can control people’s behavior at a distance, says Povinelli, dodging another aqueous salvo.

I used to be able to get them to stop. Now I can’t even intimidate them.

It is hard to imagine Povinelli intimidating anyone. The lanky, towheaded 32-year-old seems barely removed from adolescence himself as he describes or, more often, acts out the behavior he has observed in a decade of research on ape cognition. Povinelli isn’t interested in the behavior as such, but he is always on the lookout for clues to the mental lives of his charges. He has carried out dozens of experiments with the New Iberia chimps to explore the way their minds represent the world. In doing so, he has discovered differences between human and chimpanzee mentalities that defy expectations and even common sense.

Povinelli’s work addresses the question of how--or whether--apes think about themselves and other beings. Researchers of animal behavior have long suspected that certain nonhuman primates may share with humans a trait as fundamental to our species as walking and talking: self-awareness, the quality of mind that recognizes its own existence. It is self-awareness that allows enlightened individuals like Mowgli and Louie to comprehend abstract notions such as I and wanna; in the human psyche, self- awareness is coupled with an awareness of the mental lives of others, giving rise to abstract notions such as compassion, pride, embarrassment, guilt, envy, and deceit.

Researchers have also assumed that apes, like humans, possess some awareness of the mental lives of others--that they have an inkling of what it means to be like you. This assumption has shaped prevailing models of primate intelligence, which hold that complex social interactions, informed by an awareness of self and others, drove the evolution of mental acuity in human beings and their nearest phylogenetic relatives. The sociality theory has dominated studies of primate cognition for more than ten years.

But Povinelli’s investigations have led him to challenge that model and to propose a radical new theory of the evolutionary origins of self-awareness--one that would make King Louie proud. Povinelli believes that the key to the origins of self-awareness lies not in the social behavior of the much-celebrated chimpanzee but in the locomotive behavior of the solitary and elusive orangutan. He glimpses the dawning of self- conception not in the stresses of communal living but in the perils of traversing treetops. In 1995, Povinelli and physical anthropologist John Cant of the University of Puerto Rico School of Medicine elaborated this vision in an idea they call the clambering hypothesis. Their argument is subtle and recondite, combining elements of philosophy, psychology, evolutionary biology, and physical anthropology. Its principal tenet rests, however, on the observation that the orangutan truly is, in some sense, the king of the swingers.

On a steamy saturday in April, Povinelli lugs a three-by-three- foot mirror into the chimp compound and gives his apes a chance to eyeball themselves for the first time in about a year. Reactions vary. All the chimps are excited by the new arrivals, but some seem to understand better than others just who it is that has arrived. Apollo hoots and feints in an attempt to engage his reflection in play. Brandy fixes her gaze on the mirror while repeating a series of unusual gestures, apparently mesmerized by the simian mimic who can anticipate her every move.

It is Megan, the Einstein of the cohort, who performs an eerily familiar repertoire of activities before the looking glass. She opens her mouth wide and picks food from her teeth, tugs at a lower lid to inspect a spot on her eye, tries out a series of exaggerated facial expressions. Then, assuming a not-so-familiar posture that in another primate might be considered obscene, Megan uses the mirror to draw a bead on her privates. She pokes at them with one finger and proceeds to sniff the digit with enthusiasm.

That’s classic self-exploratory behavior--getting the butt right up against the mirror, where they can see, well, parts of themselves they can’t ordinarily see, says Povinelli. They never do that--get in that bizarre posture, pick at the genitals--unless there’s a mirror there.

Povinelli and other researchers maintain that self-exploratory behavior in front of mirrors shows that the ape recognizes the self therein. And for an animal to recognize itself, they reason, it must have a sense of self--some form, however rudimentary, of self-awareness. Thus self-recognition in mirrors, they argue, can serve as an index of self- awareness in species other than our own.

The architect of this line of reasoning is psychologist Gordon Gallup of the State University of New York at Albany, who in the late 1960s devised a standard measure of self-recognition called the mark test. In the test, marks of bright red dye are applied to a chimpanzee’s eyebrow ridge and opposite ear while the animal is anesthetized. The dye is odorless and nonirritating, so the chimp can’t smell or feel it; nor can the chimp see the marks without the aid of a mirror. After the ape comes to, it is given a chance to check out its new look.

When they see themselves in the mirror, they do a double take, says Gallup. Then they touch the dyed areas, then smell and look at the fingers that have contacted the marks. That’s the basic test of self- recognition. The fact that chimpanzees touch the marks and then inspect their fingers is the clincher, says Gallup, for it demonstrates that the animals know the blood red spots they see in the mirror are not out there on some unfortunate conspecific but on their own hairy selves.

Since Gallup originated this procedure, researchers have subjected dozens of animal species--including cats, dogs, elephants, and more than 20 species of monkeys--to the mark test. So far, the only subjects that have passed are the great apes: chimpanzees, orangutans, and one gorilla (the celebrated Koko). Even for members of this elite group, self-recognition is no instant achievement. They require prolonged exposure to mirrors--from minutes to days, depending on the individual--before they begin to display self-exploratory behavior.

When they first encounter their reflections, chimps act very much as if they were confronting another chimp. Apollo’s playful outbursts are typical of these social responses. Most chimps, though, soon abandon such tactics and, like Brandy, begin to perform simple, repetitive movements, such as swaying from side to side, while watching their mirrored doubles intently. At this stage, Povinelli believes, the animals may be apprehending the connection between their actions and those of the stranger in the glass; they may understand that they are causing or controlling the other’s behavior. When they finally grasp the equivalence between their mirror images and themselves, they turn their attention on their own bodies, as Megan did.

In some sense, says Povinelli, these chimps may be recapitulating the evolutionary drama that produced self-awareness in some ape-human ancestor. In that drama, other species never get beyond the first act. Monkeys, like many animals, seem to understand how mirrors work; yet they cannot solve the riddle of their own reflections. In 1978, for example, Gallup introduced a pair of macaques to a mirror, and it’s been in their cage ever since. If the monkeys espy a human image in the mirror, they immediately turn to confront the person directly. But each monkey still threatens its mirror image as it would a macaque intruder.

It’s not that they’re incapable of responding to mirrored information--they can clearly detect the dualism as it applies to objects other than themselves, says Gallup. But when they see themselves, they’re at a complete loss.

Povinelli discovered Gallup’s work as a teenager while photocopying an article in American Scientist magazine for a high school debate. Along with the last page of that article, he copied the first page of an article by Gallup; he read the beginning of Gallup’s paper at home and then went back to the library to finish it.

I was, I don’t know, 15 or 16, and I started reading this stuff about chimps, says Povinelli. The ape language experiments were really hot and heavy then, and I got caught up in the chimps-as-hairy-human- children zeitgeist.

The attitude of the time placed the cognitive faculties of monkeys, apes, and humans on a continuum, with differences between the species portrayed as matters of degree rather than kind. Koko, the captive gorilla, had done much to reinforce this view by learning American Sign Language in the early 1970s. And in the early 1980s, when young Povinelli began devouring the literature on chimp cognition, primate researchers began to document social interactions among monkeys and apes that rivaled aspects of complex human behavior. The most compelling of these interactions involve apparent deceptions--hiding food from a compatriot, for example, crying wolf to distract an aggressor, and concealing illicit sexual encounters.

The treachery, pettiness, and politicking seems to reach an apex, as it were, in societies of chimpanzees, our closest relatives. Gallup’s self-recognition studies provided a conceptual framework for these observations. It was easy to see how a keen awareness of self--including the ability to plan your actions and anticipate their effects--might come in handy if you’re bent on making a chump of your fellow chimp. Furthermore, many primate researchers argued that the elaborate deceptions practiced in chimpanzee social groups offered clear evidence that the animals appreciate one another’s motives and intentions as well as their own. Gallup had speculated that self-recognition implied not only self- awareness but insight into the mental states of others, a capacity known as empathy.

Can tests be devised to measure empathy in primates in the same way the mark test plumbs self-awareness? That question has long preoccupied Povinelli. It became the topic of his dissertation at Yale and the principal focus of his subsequent work at the New Iberia center. The University of Southwestern Louisiana, which administers the primate center, hired the fledgling Ph.D. to set up a research program in 1991; Povinelli also established the university’s Center for Child Studies, where he runs experiments that parallel his primate research--matching the wits, in effect, of apes and children. By comparing the performances of the two species on cognitive tasks, Povinelli hopes to clarify the features of mind that distinguish people from pongids.

In human beings, self-awareness and other-awareness are inextricably linked in a cognitive feature that psychologists call theory of mind. That lofty term describes the tendency to assume that other people--and also pets and even, sometimes, inanimate objects--experience desires, intentions, and beliefs just as they do. We use our assumptions about these subjective experiences to interpret behavior (as in, the dog is barking at the door because it wants to go out), to predict behavior (as in, he won’t call because he’s angry with me), and to judge behavior (as in, the killing was self-defense, not murder). And yes, human beings also use their theories about the minds of others to manipulate and deceive.

In toddlers, these conceptions of self and other as conscious, mental agents seem to develop in tandem. We think that theory-of-mind skills are emerging in kids right around 18 to 24 months of age, says Povinelli. That’s where you see their first understanding of desire, reference, and attention. And that’s also the age at which kids first recognize themselves in mirrors.

Children who can pass the mark test, for example, clearly understand conventions of nonverbal communication that require a concept of other. They understand pointing as a referential gesture--a gesture meant to connect, intangibly, two or more subjects with an object in space. And they recognize that the direction of a person’s gaze indicates where that person’s attention is directed as well.

Povinelli decided that such hallmarks of human cognitive development could serve as models for tests of empathy in primates. Could chimps understand, say, the intentions that underlie pointing and gazing in humans? He designed a series of experiments that yielded intriguing results. In one such test, a chimp has to choose between two overturned cups to find a treat underneath. An experimenter offers a hint by pointing at one cup. At first, it looked as though the apes could learn how to interpret the gesture; after several dozen trials, they picked the right cup almost every time. But additional experiments showed that the chimps were not taking their cue from the direction of the pointing finger. Instead they were choosing the cup closest to the experimenter’s hand. If the experimenter held her pointing hand equidistant from the two cups, the chimps chose randomly. They seemed unable to learn the significance of pointing alone.

In another experiment, Povinelli tried to ascertain whether chimpanzees’ ability to track another’s gaze reflects a conscious understanding of another’s point of view. This time the chimps had to choose which of two boxes contained a hidden treat. An experimenter gazed at a spot midway between the receptacles. A wooden partition blocked one box from the experimenter’s view, and the chimp’s task was to figure out which box he could be gazing at. Children know to pick the box in front of the partition. But chimps, while they clearly register the direction of the experimenter’s gaze, tend to pick the box behind the barrier almost as often as the one in front of it.

They’ll follow your gaze, but there’s no evidence that they understand your vision as a mental state of attention, says Povinelli. Another experiment confirmed this: given a choice between two experimenters, chimpanzees will beg for food from someone wearing a bucket over his head--someone who not only looks foolish but clearly cannot see their entreaties--as often as they will solicit a person carrying a bucket on his shoulder.

Why would an animal so adept at learning in the lab fail to respond to the cues in these experiments? Povinelli acknowledges the difficulty of probing the mind of another species. With such unorthodox experimental designs, it is not always clear who is testing whom. So far, though, the results of his experiments suggest that chimpanzees don’t comprehend the intentions or points of view of others--though an anthropomorphic reading of their social behavior may suggest that they do.

Contrary to what Gallup believed about empathy among apes, chimpanzees may inhabit a cognitive realm that includes a subjective notion of me but not you. Anecdotal accounts of chimpanzee deception, says Povinelli, can be explained without invoking the capacity for empathy--and should be, in light of his research. Chimpanzees are hard-wired to be ultrasensitive to social contexts and cues, he adds; they are expert at manipulating behavior--just like spitting at you in the compound.

But while deception and manipulation indicate a powerful, specialized intelligence, they do not necessarily implicate a theory of mind. A chimpanzee can get a cheap thrill from watching a human being evade a projectile of water without knowing (or caring) why the human responds that way--without appreciating the embarrassment, annoyance, and discomfort of conducting an interview in a spit-spattered blouse with a handful of soggy pulp for a notepad. As Povinelli sees it, chimps may be self-centered in the purest sense of the word.

Povinelli’s portrait of the self-centered chimp recasts the question of how primate intelligence evolved. If his data accurately represent simian sensibilities--and he is not excluding the possibility that they don’t--there is a deep cognitive chasm separating apes from humans. It’s possible that there’s a disjunction, evolutionarily speaking, between self-conception on the one hand and a general theory of mind on the other, he says. In other words, there was an understanding of self before there was an understanding of other.

Maybe chimps have a pretty good theory of their own minds, in the sense that they can contemplate what their attention is focused on, what they want, that kind of thing. But maybe they simply don’t have any understanding of that quality in others. And maybe humans, for some reason, have fused an understanding of self and other.

Povinelli’s findings don’t exactly refute the sociality theory; instead they render it somewhat less relevant. It is easy to imagine that the pressures of navigating primate social hierarchies--dodging the wrath of the dominant male, for example--may have advanced some aspects of intelligence in certain primates. Yet there is nothing about social pressures that would have driven the dawning of self-awareness per se, notes Povinelli. After all, monkeys have fairly complex social lives, and they fail the mark test. Orangutans, on the other hand, are among the most solitary of primates, yet they pass with flying colors.

No one has ever explained why on earth sociality would have anything to do with this phylogenetic break in the self-concept, says Povinelli. In fact, there were no explanations at all for how a primitive sense of self may have evolved in the common ancestor of great apes and humans--until Povinelli went into the Indonesian jungle.

In 1989 and again in 1991, Povinelli spent a field season with John Cant documenting the movements of arboreal primates in the rain forests of northern Sumatra. Cant was studying the locomotion of monkeys, gibbons, and orangutans for his research on the evolution of the primate musculoskeletal system. Though such studies are outside his own area of interest, Povinelli was eager for field experience; in particular, he looked forward to watching orangutans, which are scarce in captivity.

Primatology lore holds that these large, solitary, and slow- moving apes are as smart as, if not smarter than, their phylogenetic cousins, the chummy chimpanzees. Yet if the orangutan’s social life isn’t responsible for its perspicacity, Povinelli began to wonder, what forces are responsible? Braving scorpions, leeches, and warm Bintang beer, he and Cant struck upon a way to explain not only the intelligence of orangutans but also the self-awareness of chimps and human beings. The clambering hypothesis was born.

The idea’s ungainly name derives from an equally ungainly activity unique to orangutan locomotion. As Cant defines it, clambering is the slow, deliberate navigation by which an orangutan manages to move from tree to tree. In no way, Cant contends, does clambering resemble the more automatic and repetitive movements, such as running, leaping, and swinging, that are typical of other primates. And according to his observations, clambering is the method orangutans prefer for traveling through the treetops.

When an orangutan is moving around up there, says Cant, it sounds like a small tornado is going through the canopy--branches swaying back and forth, brushing against each other, some breaking. And if you look, quite often you see what you think is the animal stopping and making up its mind. It starts doing something, stops, pauses, and--whether or not it looks around in some befuddled human way--it then does something different.

There is much in navigating treetops to give an orangutan pause. Adult males of the species can weigh upwards of 180 pounds; tree trunks and branches bow mightily under their weight, and falls can be fatal. In spite of these risks, Sumatran orangutans rarely, if ever, travel on the ground. They climb from tree to tree like sluggish acrobats, using the exceptional mobility of their hip and shoulder joints to distribute their mass among multiple supports. It is not unusual to see an orangutan grasping a woody vine with one hand, holding a branch with the other, and bracing one foot against a tree trunk while the other reaches for a nearby limb. By shifting their weight back and forth, orangutans can bend a tree to their will, making it sway closer to its neighbors and thus aid passage.

None of these maneuvers were lost on Povinelli. While becoming acquainted with orangutan locomotion, he was also boning up on the work of Jean Piaget. The Swiss psychologist had described the dawning of self- conception in children as arising from the inadequacy, or failure, as he put it, of the sensorimotor system. In Piaget’s theory, this system governs the repetitive and seemingly instinctual movements of infants younger than 18 months or so. Before that age, Piaget argued, children are not conscious of causing their own actions. But as a child’s mental life becomes more complex, those actions become more ambitious, and some will inevitably fail to provide the intended outcome. Confronted with such failures, children become conscious of both their actions and their intentions--they become, in a word, self-aware. Somewhere around the age of two they also enter a new stage of development, in which they learn to control and plan the outcome of their actions.

When we got to the field and started talking about clambering, says Povinelli, it suddenly struck me that that, in a way, may be the same damn thing. Clambering is the failure of the sensorimotor system, in an evolutionary sense.

In Povinelli and Cant’s hypothesis, clambering represents the self-aware locomotive style of a common ancestor of humans, chimps, orangutans, and gorillas. Like orangutans, this ancestor probably lived in the trees and weighed at least three times as much as the most massive tree-dwelling monkey. Climbing procedures scripted by the sensorimotor system--exemplified by the limited repertoire of repetitive movements that characterize monkey locomotion--would most likely have failed the ancestor, much as they would fail present-day orangutans. And in this context, failure meant an express trip of 30 feet or more to the forest floor. Fall flat on your face from a height of a few dozen feet for a few million years, say Povinelli and Cant, and sooner or later you will evolve the capacity to figure out what went wrong. Figuring that out means conceiving of the self as a causal agent: understanding that the breaking of boughs and subsequent plummeting action is caused by one’s own heft, inexpertly deployed.

Once this sense of personal identity and agency emerges, the coauthors have written, an understanding of that object (the self) can be elaborated and expanded upon almost indefinitely.

It is this budding awareness of the self as a causal agent that Povinelli sees in his chimpanzees’ antics in front of mirrors. Reflections give the apes an opportunity to observe the direct consequences of their actions: I caused that. Self-recognition occurs when an ape understands that it causes everything about its mirror double: I am that.

For monkeys, it seems, there is no I. Povinelli and Cant assert that tree-to-tree travel was never hazardous enough for monkey ancestors to warrant the evolution of a specialized cognitive coping mechanism. Because of these ancestors’ low body weight, falls would have been infrequent and not particularly harmful.

Monkeys jump onto the end of the branch, and when it bends on them they just hold on, says Povinelli. It’s the difference between assimilating the reaction of the environment into your behavior and actively using your behavior to plan how to change the environment in order to solve a particular problem. You don’t need to have a sense of self to do what you have to do to be a monkey.

Having elaborated this distinction between monkeys and apes, however, Povinelli emphasizes that his claims for ape self-awareness are still quite modest.

It’s nothing like, ‘My God, I’m an orangutan. I’m an orangutan, and gosh, I was born 17 years ago, and here I am, still up in the trees, climbing. I wonder what my fate is?’ says Povinelli. We’re just arguing that a combination of factors drove the evolution of an ability to objectify the self--the first step, he says, along the road to self- discovery.

Qualifiers aside, Povinelli and Cant are well aware that they are out on a rather fragile limb themselves. The clambering hypothesis is by far Povinelli’s most speculative piece of work to date, and it has garnered more than a few hoots from other naked apes.

We hardly know what self-awareness is, let alone how it came about, says ethologist Frans de Waal, research scientist at the Yerkes Primate Center in Atlanta. I am personally not convinced by the argument. De Waal believes that the climbing behaviors of several species of South American spider monkeys may be as complex and premeditated as the clambering of orangutans. I don’t think orangutans are doing anything that these monkeys don’t do. De Waal also objects to defining self-awareness so narrowly. I look at self-awareness as a kind of continuum that probably runs from fish to humans, he says. The mirror test somehow taps into a higher level of it. But I cannot imagine that this is an all-or-nothing phenomenon.

This is what I say to people who are extremely skeptical about the clambering hypothesis, says Povinelli. I say, well, okay, fine. But there’s a real problem here. Self-recognition in mirrors is restricted to the great ape-human clade. There’s no other proposal on the table that explains why.

That doesn’t mean, he adds, that the clambering hypothesis is right.

Indeed, even claims of mirror self-recognition in apes have come under fire of late. Using a modified version of the mark test, cognitive neuroscientist Marc Hauser of Harvard has prompted unusual behavior in tamarins that he says could be taken as a sign of self-recognition. I want to remain kind of agnostic about what’s actually going on, says Hauser. But he says his observations cast doubt on the long-standing notion that mirror self-recognition is a reliable marker for self-awareness.

Povinelli says he and Gallup have tried to replicate Hauser’s work in marmosets, so far with no success. But he is the first to admit that he doesn’t have the final word on either self-recognition studies or primates’ concept of self.

The problem seems so simple, you know? A mirror, a monkey . . . a mirror, a chimp. . . . But there’s three decades’ worth of work to be done in figuring out what the heck’s going on.

Anybody who thinks that they’ve got the final word on this-- Povinelli pauses to engage his own theory of mind--I think they’re stark raving mad.
Comment on this article

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar

Log in to your account

Email address:
Remember me
Forgot your password?
No problem. Click here to have it emailed to you.

Not registered yet?

Register now for FREE. It takes only a few seconds to complete. Register now »