Kanzi: The Ape at the Brink of the Human Mind

by Sue Savage-Rumbaugh

  As researchers have become more interested in self-awareness they have begun to recognize that it is closely linked to other-awareness and that language skills are predicated on awareness of both self and others. Why would anyone bother to tell anyone else something unless they assumed that the other person did not already know what it was they were about to say?

  This idea—that the knowledge states of the speaker and the listener can in fact be different—has been given the term “theory of mind.” How can we know if animals believe that other animals have minds? Richard Byrne and Andrew Whiten of the University of Saint Andrews have suggested that the key lies in the ability to deceive. If one animal sets about to deceive another with intent, it must recognize that the other has a mind that is composed of knowledge states different from its own.

  Lies are notorious in the animal kingdom. Nonpoisonous butterflies imitate the wing patterns of the monarch butterfly, so as to avoid being eaten by birds who have learned to leave the poisonous monarch alone. Plovers pretend to have a broken wing in order to lead potential predators away from their young. While it is clear that the butterfly that imitates the monarch is not doing so intentionally, it is less clear with the plover.

  Scientists have assumed that the broken wing display is an innate pattern, carried out with little understanding in time of danger. However, this view has recently been challenged by Carolyn Ristau of Rockefeller University, who points out that like all supposedly innate patterns, the bird appears to be making decisions that vary from instance to instance and that do take into account characteristics of the predator. There are also great differences among plovers in how, where, and when they elect to lead predators, and some are far better deceivers than others. Yet none of these facts tells us whether the plover knows that it is deceiving the predator. The only time that plovers appear to have broken wings, however, is when there is a predator present.

  Other animals engage in deceptive acts that are no less situation-specific. Byrne and Whiten have coined the term “tactical deception” to separate acts that are deceptive but apparently without reasoned intent, from those where the intent of the perpetrator is clearly to mislead. They define tactical deception as “an individual’s capacity to use an ‘honest act’ from his normal repertoire in a different context, such that even familiar individuals are misled.”3

  Byrne and Whiten became interested in deception after seeing several instances of it among a troop of baboons they were observing in the Drakensberg Mountains of southern Africa. For instance, one day they saw Paul, a juvenile male, approach Mel, a mature female, who was engaged in unearthing a succulent tuber. Paul looked around: No other baboons were in sight, but they were not far away. Suddenly, Paul let out a piercing scream, as if he were in danger. Predictably, Paul’s mother, who was dominant to Mel, rushed to the scene and drove Mel away. Paul then calmly ate the abandoned tuber. “Watching the incident, it was difficult to suppress an intentional interpretation,” Byrne and Whiten commented later.4 Had Paul concluded, “If I scream, my mother will assume Mel is attacking me; she’ll run to defend me; and I will be left with the juicy tuber to eat”? or had he merely been upset that he was not getting any of the tuber for himself and screamed out of frustration?

  The work of Dorothy Cheney and Robert Seyfarth with vervet monkeys in the Amboseli National Park in Kenya represents a long-term attempt to understand the minds of monkeys in their natural habitat. They too, have seen behaviors that appear to be intentional deception, though they have concluded that the monkeys only partially understand what they are doing. As an example, they cite the behavior of Kitui, a low-ranking male at Amboseli. One day a new male seemed poised to join the group from elsewhere, an event that would surely have been a threat to Kitui’s already low social status. When he saw the male, Kitui gave a leopard-alarm call even though there was no leopard in sight. The call had the effect of keeping the new male in the trees and delaying his entry into the group. “So far, so good,” observed Cheney and Seyfarth. “The alarm calls appeared deceitful because they signal danger that Kitui, but not the interloper, knew to be false, and they kept the interloper temporarily at bay.”5 But Kitui’s behavior was not consistent. Instead of remaining in his own tree throughout, which is what vervets do when a genuine leopard alarm is made, Kitui descended, crossed open ground, and climbed a tree near to the new male, alarm-calling all the time. He had got only part of the story right, not unlike the child who vigorously denies raiding the cookie jar, with crumbs evident all over her face.

  Overall, Cheney and Seyfarth concluded that “vervets’ cognitive abilities are limited compared with our own and that there is no evidence that the monkeys have a ‘theory of mind’ that allows them to recognize their own knowledge and attribute mental states to others.”6 Another illustration of this conclusion is the vervet’s failure to take account of the listeners’ state of knowledge or ignorance when making alarm calls: The caller is uninfluenced by how much others know about what danger is threatening.

  I have observed similar incidents involving Kanzi’s mother, Matata. Once, when I was introducing a new person to Matata, she became jealous of the new person and refused to let her touch any item that she was fond of, including her blankets, her bowls, her food, and her mirror. One day we were sitting together on the floor when Matata decided to ask me to go get some food by holding her empty bowl out to me and making a food sound. I told Matata I would get her some food and left the room, leaving the bowl with her. I had been gone less than five minutes when suddenly Matata began screaming loudly. When I rushed back into the room the new person was holding Matata’s bowl and Matata was screaming at her and threatening to bite. Matata looked back and forth from me to this person and then to the bowl, screaming—intimating that the bowl had been grabbed from her in my absence and that I should support her in attacking this mean individual who had taken her bowl. Of course, given the gift of language, this person was able to explain that she had indeed done nothing. Matata had placed the bowl in her hands and then starting screaming for me as if she had been wronged.

  When Matata saw us talking about what had happened, she began to look very crestfallen, concluding her ruse had not worked. She stopped screaming and moved to the corner where she suddenly became very preoccupied with grooming herself.

  When Byrne and Whiten talked informally with fellow psychologists and primatologists, they heard similar reports. Few such reports make it into the pages of the scientific literature, however, because they are single observations that the researchers have no means of replicating. In behavioral science, one is supposed to be able to predict that a certain behavior is going to occur. Clearly, when one is dealing with deception, this is virtually impossible, since one almost never knows when deception is going to occur. In fact, if the deception is truly successful, the researcher as well as the other animals are so satisfactorily deceived that they do not know that it even occurred.

  Byrne and Whiten conducted a survey of more than a hundred of their colleagues in 1985 and again in 1989, asking for anecdotal accounts of incidents that might be judged as tactical deception. The response was enthusiastic, and several hundred putative cases of deception were assembled. The question was: Did these cases reflect instances of deception with reasoned intent, or could there be other explanations? When Byrne and Whiten applied strict criteria to the supposed examples of deception, ruling out as carefully as they could possibilities of learning, they concluded that of the 253 cases assembled in the 1989 survey, only 16 could be said to reflect mind reading in the sense of tactical deception. All of these were in apes, and most with chimpanzees. Two are noted here.

  During her lifelong monumental studies at Gombe National Park, in Tanzania, Jane Goodall observed many examples of putative deception. One of the most intriguing involved the chimp named Figan, who noted that a banana had been left in the crook of a tree. Unfortunately for Figan, Goliath, the Alpha male, was resting under that very tree. After glancing briefly at the
fruit and then at Goliath, Figan moved some distance away, perhaps fearing that if the fruit were still in his line of sight, he wouldn’t be able to resist looking at it and would thus alert Goliath to its presence. After fifteen minutes Goliath left, and without hesitating Figan retrieved the banana and ate it quickly.

  The Dutch primatologist Frans Plooij observed a similar incident, also at Gombe. An adult male was alone in a feeding area when a box was opened electronically, revealing the presence of bananas. Just then a second chimp arrived, and the first one quickly closed the box and ambled off nonchalantly, looking as if nothing unusual were afoot. Like Figan before him, he waited until the intruder departed and then quickly opened the box to retrieve the bananas. Unlike Figan, however, he had been tricked. The other chimp had not left but had hidden, waiting to see what was going on. The deceiver had been deceived.

  Evidence of self-awareness and of deception therefore suggests that apes think of themselves and others as having knowledge states that differ. Is there another window into the animal consciousness, one that perhaps requires less inference regarding intent on the part of the observer? One area that has begun to be studied is that of animal pretense or imagination.

  All animals, like all children, love to play, and in play all of the elements of true aggression are acted out in pretense. It is as though all of what might happen, or could happen, is experienced many times on many planes prior to the actual occurrence, so that the animal or person is ready for the event when it actually does occur. Animal behaviorists have typically assumed that animals do not know they are pretending at aggression, as children do when they play at being superheroes. This is because children can say they are pretending while animals cannot. Many researchers even claim that it is not until four years of age that children can discuss the difference between pretense and reality. Prior to this time, what appears to be pretense to us is said to be reality to the child.

  Games of pretend among apes are not as elaborate as those seen in children, but they are engaged in with enthusiasm nonetheless. Vickie, the first chimpanzee who participated in a language project, was reported to have a great time pulling an imaginary toy with an imaginary string. On occasion, it even appeared that the imaginary toy became stuck, such as between the toilet and the wall, as Vickie ran around the room with it.

  When he was young, Austin often pretended to eat imaginary food, occasionally even using an imaginary dish and an imaginary spoon. He would carefully place the nonexistent food in his mouth and then roll it around on his lips, watching it just as though it were real food. Sherman was not interested in imaginary food—he always wanted the real thing—but he loved to pretend that dolls, particularly King Kong dolls, were biting his fingers and toys as well as having fights with each other.

  Most frequently, these pretend games were played alone, but not always. On one occasion, as Sherman and Austin were watching a King Kong movie on television, they began to pretend that King Kong was actually in a cage located in their room. This cage looked just like the one that Kong was in in the movie. At this point, Sherman and Austin stopped watching the television and began to make threat barks at the empty cage and to throw things at it, as though it housed Kong himself. Sherman even got out the hose and began to spray the cage.

  As a youngster, Kanzi liked to pretend that he was hiding imaginary food in his blankets or under some toys. Occasionally he would pretend to give Panbanisha or others bites of the imaginary food. If he elected to eat it, he did not do so the way Austin did, who consumed it very slowly while watching himself in the mirror. Instead, Kanzi pretended to gulp it down hurriedly, as if he had stolen it.

  Panbanisha’s favorite pretend game was to act as though she had heard a monster in the next room. Going toward the door with her hair out, she would comment “monster” and invite others to search with her. Sometimes she would then put on a monster mask and pretend to chase her sister Tamuli. She also liked to pretend that she was taking bites out of pictures of food she saw in magazines, or even out of the peaches that are depicted on all the Georgia license plates at the laboratory.

  But whether it is self-awareness, awareness of the minds of others, pretense, or deception—all of these cognitive activities are manifest in language, for it is with language that Kanzi and Panbanisha and Sherman and Austin can tell us things that we would otherwise not know. Kanzi has told us where he left a ball the day before and reminded us of yesterday’s promise that we forgot. Panbanisha has told us that she wanted to watch an “ice TV” when it started snowing outside. Kanzi told us that he was looking for his mother, Matata, when asked why he was trying to crawl under the railroad ties. Panbanisha tore the “good” lexigram off the keyboard and gave it to us as a way of sealing her promise to be good. Sherman told me there was a “scare” outdoors when he saw a chimp being carried away in a transport cage, and Austin always tells me he wants Coke instead of the juice I am offering. Panbanisha tells me when it is raining outdoors. She also told me that the lady who visited has hair that looks like a mushroom—and she was really right. These are small things perhaps, but they offer a constant glimpse of other minds that I would not have without language.

  Kanzi and Panbanisha tell me and others these things because they assume that I do not know them. Things we both know they never bother to state. Not only do they use language to present their thoughts, sometimes they use language to pretend and sometimes to lie. Kanzi, for example, knowing that he cannot have any more M&M’s, will ask to go play in the T-room and get some toys when he (but not I) knows that M&M’s have been placed in a T-room cabinet. Once in the T-room, he will quickly steal the M&M’s and run out. He does this so deftly and quickly, it is clear that it was his intent all along.

  Kanzi’s naturalistic acquisition of words and his emergent comprehension of complex spoken sentences indicate that the chimpanzee has all the basic neurological machinery for a primitive language. Kanzi has not learned to speak, but this limitation appears to be one of the physical structure of vocal/respiratory circuitry and anatomy. His expression of a language facility through being steeped in language, as human children are, illuminates the nature of the chimpanzee mind—and it puts the mind of Homo sapiens in proper biological perspective.

  Kanzi’s linguistic capacities give us a clearer view of human language. First, they demonstrate the narrowness of the Chomskian assertion that spoken language is an evolutionary novelty that arose uniquely in humans. Second, they emphasize the interactional aspects of language acquisition, for both apes and humans. Steeped in a language-rich environment, human children first come to comprehend language, and then to produce it, a process requiring elaborate control over the vocal/respiratory tract and complex planning of muscle movement.

  As language is learned, many general cognitive processes become shaped as a consequence. Language in humans, rather than being innate, is seen as the product of a plastic cognitive substrate interacting with environmental exposure to speech. In their natural state, chimpanzees do not develop the kind of language we see in captivity, but given the appropriate environmental exposure, an ability for symbolic language use becomes evident. At the same time other cognitive skills become honed. For instance, language-competent chimpanzees develop an ability to learn to use a joystick (connected to a computer monitor) through simple observation rather than through active teaching. Language-naive apes must be trained to do this bit by bit.

  Clearly, exposure to speech in infancy shapes the developmental processes in the ape brain just as it does in the human brain. And it is more than likely that the neural networks influenced by a language-steeped environment in apes are evolutionary parallels to those involved in language acquisition in human children. Because chimpanzees are so closely related to humans in an evolutionary sense, it should not be surprising—except to a Chomskian—that they are sensitive to the same environments that foster cognitive competence in human children.

  The ease with which Kanzi acquired a facility for symbolic
communication not only tells us something about humans, and the assumed uniqueness of the human mind, but also something about apes and their cognitive competence in their natural state. We have yet to appreciate the intellectual challenge of the natural lives of intensely social primate communities. One aspect of this is that the natural communicative skills of chimpanzees in the wild are almost certainly greatly underestimated.

  The boundary wall between humans and apes has finally been breached. As a result, more and more biologists are bowing to the logic of the evidence of genetics, by suggesting a reclassification of humans and the great apes. At the very least, humans, chimpanzees, and gorillas should share the same family status, with orangutans in a separate family. If there is a line to be drawn in this evolutionary scheme of things, it will have to be one that places the African apes in our company, not one that excludes them. Humans are African apes, of an unusual kind. Similarly, if a mental Rubicon was crossed in our evolutionary history, humans are not the sole occupants of one bank; they are accompanied there by apes.

  The distinctiveness that we have so assiduously ascribed to ourselves as humans is, in reality, an accident of history. Imagine, for instance, how much more distinct we could have claimed our species to be had all the great apes become extinct before we began pondering our position in the world of nature. If vervet monkeys were our closest relatives, humans would indeed appear to stand separate. Equally, if the species of hominid that links us to our common ancestor with the African apes had not become extinct, the gap between us and chimpanzees would be closed all the way. Gradations between human and ape would be present at every step, and our revered distinctiveness would vanish.

  It is simply a contingent fact of history that certain species did become extinct during the past five million years, leaving us to compare ourselves with the African apes as our closest living relatives. And it is a sobering fact of current history that the comparison between humans and apes may soon become virtually artificial, as each species of ape faces extinction in its natural populations. If this happens, it means we will lose the opportunity to learn about ourselves from our nearest living relatives, just at the time that we have indeed recognized them as our relatives. It also means that we will have frittered away our one remaining chance to allow our sibling species to live the way of life for which they, and we, co-evolved across the millennia.