Kanzi: The Ape at the Brink of the Human Mind

by Sue Savage-Rumbaugh

  Despite Matata’s assiduous effort, and the systematic training program, she still did not progress as Sherman and Austin had. Able to communicate many things effectively through gestures, facial expressions, and vocalizations, she continued to be stumped by the keyboard. After two years of training and thirty thousand trials, she mastered only six symbols in a limited way. She could request things, but became confused when others asked her to respond to their communications. She remained tied to the anticipated consequences of symbol usage. When we attempted to alter these, by switching from a request task to a naming task or a statement task, for example, Matata no longer selected the correct symbols. She did, however, continue to try to convey things using the keyboard, albeit in an incomprehensible manner—something like the “word salad” that Terrace had described for Nim. She would become very frustrated when I did not understand her, as she pointed to a symbol that seemed to have little to do with where her attention was focused. At times she would take my hand, look me in the eyes, and vocalize urgently, as if trying to explain what she wanted and wondering why I didn’t understand.

  Matata differed from Sherman and Austin in other ways as well. For one thing, her ability to sort objects into recognizable categories was much less precise. She could sort only those objects that differed in many different dimensions, such as weight, color, shape, texture, and size. Objects that differed in only one way, such as color, were all grouped together. She was also unable to sort photographs—a favorite activity for Sherman and Austin. The inability to group photographs did not reflect an absence of picture-recognition skills, however, as looking at magazines was one of her favorite activities. She even sometimes “tasted” pictures of favorite fruits. By contrast, Sherman and Austin readily sorted objects that differed in one way, doing it first by color, then by shape, then by size, and so forth. They could sort photos into groups such as people, animals, tools, foods, and vehicles.

  Tools also baffled Matata. She often tried to put a stick into a hole, with the stick oriented perpendicular to the opening. And she failed to learn how to insert a key into a lock, despite much practice. Many of her imitation skills were poor as well. For example, Sherman and Austin learned how to operate a joystick to control video games simply by observing me do so a few times. No amount of demonstration could convey the relationship between the joystick and the movement of the cursor on the screen to Matata. She finally learned, but only after a computer program designed specifically to shape this behavior in monkeys had been devised.

  Why was it so difficult to repeat the steps into language taken by Sherman and Austin? Perhaps the problem was her age. At ten years of age when the project started, she was older than any other ape that had entered a language project. Could it be that apes, like humans, need to be exposed to language during infancy in order to acquire it?

  Whatever the reason for Matata’s difficulties, we had our own problems at this time. Our research grant was scheduled to terminate the following year, and if it were not renewed, our program of study would collapse. Matata’s progress had been minimal and we could not blame the reviewers if they elected not to invest further funds in the study of bonobos and language. Not only had Matata failed to acquire symbols, we had learned little from her failure. Sherman and Austin’s failures had always been informative, in that they had provided insights into the nature of language and how to proceed with further teaching. Matata’s failure was leaving us empty-handed. We were unable to understand why she progressed so slowly.

  While we were puzzling over Matata’s poor progress, Yerkes decided that she should conceive again. They asked us to separate her from Kanzi and transfer her to the field station where she could be with Kanzi’s father, Bosondjo. They also wanted to make the separation as easy as possible for Kanzi and they hoped to have Matata cycling at the time she joined Bosondjo. Consequently, they administered hormone tablets to inhibit lactation so that Kanzi would be weaned before she left and so that she would either be exhibiting a sexual swelling, or on the verge of doing so.

  The plan was that Matata would go to the field station for a few months where she could breed with Kanzi’s father Bosondjo, while Kanzi would stay at the Language Research Center. Kanzi would have to be weaned, and then a separation effected with the minimum of trauma. Prior to receiving hormone tablets, Matata had made no attempts to wean Kanzi. She permitted him to nurse at will. However, as the hormone tablets began to take effect, nursing became uncomfortable, and consequently Matata began attempting to stop Kanzi’s nursing from time to time. At first he was puzzled and pointed at Matata’s nipple while looking directly into her eyes, as if he were trying to tell her that he wanted to be permitted access to her nipple. After the incredulity wore off, a forlorn expression came over Kanzi’s countenance and he attempted to get closer to his mother’s nipple while pretending to be interested only in grooming. Finally, when it became clear that Matata knew what he was up to and simply was not going to let him nurse, Kanzi became furious and threw temper tantrums. Then, much to my amazement, he tried to enlist my help in this battle with his mother. He stretched out his hand toward me, gave a plaintive whimper, and with his glance directed my attention back toward his mother. He then began making threat barks at her. Having assumed, on the basis on my glances, that I agreed with him, Kanzi now threatened to bite his mother if she did not let him nurse. He was right to discern that I had felt sorry for him and was inadvertently taking his side in this psychological battle of wills. But I had no idea that he would use my support, conveyed only in glances and expressions, to escalate the conflict and threaten to bite his mother.

  I was incredulous. Did he really think I would support him in attacking his mother and that I could make her let him nurse? Yes, I was sort of “in charge” of Matata from his perspective, but I certainly could not make her suckle Kanzi. It was a very curious position to find oneself in; being asked to intercede in a weaning battle between mother and son of another species. I decided quickly to withdraw my nonverbal and unintentional expressions of support. Kanzi then decided not to threaten his mother. He sat in her lap looking forlorn. I tried to help by making a bottle of milk and offering it to Kanzi but he refused it, preferring to sulk.

  For several months before Matata’s departure, Kanzi had been willing to travel with me to other parts of the center and around the forest, leaving Matata behind. At first, Matata protested loudly when she saw Kanzi go out of her sight, but eventually she was content for us to act as babysitters. Kanzi became intrigued by the “childside” of the Language Research Center, where children with mental retardation came for special language training. He seemed especially interested in the fact that the children used keyboards, just like his mother’s.

  By the time Kanzi was eighteen months old he began inventing simple iconic gestures, the first of which indicated the direction of travel in which he wished to be carried. He did this not with a finger point, but with an outstretched arm. Often, when he rode on my shoulders he added emphasis to his gesture by forcefully turning my head in the direction he wished to go—as though to guide my eyes. At other times, as he sat on my shoulders, he would lean his whole body in the desired direction of travel so that there was no mistaking his intent.

  He employed another gesture to solicit help in opening a jar. This he did by making a twisting motion with one hand while pointing to the jar to be opened with the other one. When he wanted nuts cracked, he made hitting motions in the direction of the desired nuts. And if he wanted an object given to him, he gestured first at the person and then at the object, as human infants do. Often Kanzi vocalized while gesturing, which served to catch our attention and to convey the emotional affect that accompanied each request.

  When the day of Matata’s departure arrived, we indicated to Kanzi that we wanted to take him for a walk in the woods; of course, he was eager to go. While we were gone, Matata was sedated and placed in a van without Kanzi’s knowledge. When Kanzi returned from his forest travels, he was no
t initially distressed at being unable to see Matata. However, after about thirty minutes of looking for her off and on, it became clear to him that she was not going to reappear just any moment. He began to appear anxious as his hair puffed out, and he started walking rather stiffly as his eyes assumed a worried expression. He seemed to think that his mother must be hiding somewhere, and he felt a need to look in every nook and cranny, indoors and out. He asked to visit every room in the lab, several times, as if he thought she might be moving about also looking for him. Kanzi then wanted to go back outside, and he searched many places there, too, including the trash cans.

  That night he elected to sleep with Austin, but Austin didn’t make the kind of big comfortable bed that Matata had, from intertwined blankets. After about an hour Kanzi joined me in a bed I had prepared nearby, in case he became frightened. For the next two and a half days, Kanzi glued himself to me and would not let me out of his sight. Finally, exhausted, I insisted that he stay with someone else. Kanzi objected. This was the first time I had heard Kanzi scream or express separation anxiety. I felt remorse over being the cause of any worry to him. Fortunately, he calmed down and began playing happily a few minutes later, so I used the opportunity to slip out when he wasn’t looking, leaving him in my sister Liz’s care. Liz had also helped raise Kanzi since his arrival at six months of age, and Kanzi knew her well. After searching briefly for me, Kanzi settled down happily in Liz’s lap.

  The day after Matata’s departure, we set up the keyboard in the expectation that Kanzi would begin his language instruction—if he could learn to sit in one place long enough. Kanzi, however, had his own opinion about the keyboard and he began at once to make it evident by using it on more than 120 occasions that first day. I was hesitant to believe what I was seeing. Not only was Kanzi using the keyboard as a means of communicating, but he also knew what the symbols meant—in spite of the fact that his mother had never learned them. For example, one of the first things he did that morning was to activate “apple,” then “chase.” He then picked up an apple, looked at me, and ran away with a play grin on his face. Several times he hit food keys, and when I took him to the refrigerator, he selected those foods he’d indicated on the keyboard. Kanzi was using specific lexigrams to request and name items, and to announce his intention—all important symbol skills that we had not recognized Kanzi possessed.

  How could this be? We had spent two years systematically trying to teach Matata a small number of symbols, with meager success. Kanzi appeared to know all the things we had attempted to teach Matata, yet we had not even been attending to him—other than to keep him entertained. Could he simply have picked up his understanding through social exposure, as children do? It seemed impossible.

  For several weeks we monitored Kanzi’s behavior and his use of the keyboard, checking to see if perhaps he was somehow fooling us into thinking that he knew how to use the symbols. We wanted to test his usage in a controlled situation, but because Kanzi had not been taught to do formal tests, he simply ran away every time we asked him to sit down and take a test. If we insisted, he cried and stayed there, but refused to participate. I consequently looked for ways to gather the data in a more casual way. We made tests into games that Kanzi enjoyed and that involved a great deal of play in exchange for a modicum of cooperation and data.

  This was one of the most intriguing times of my professional life. I recognized that if what we thought we had accomplished proved indeed to be accurate, it could revolutionize our understanding of the nature of language acquisition, indeed perhaps of all learning processes. Equally significant, it could seriously undermine a main tenet in the body of knowledge around which both the social sciences and the physical sciences are constructed—that of the uniqueness of human mind. We would have to reformulate our view of apes as organisms. If apes could acquire language in the manner that humans do, without instruction, this meant that man did not possess a unique sort of intelligence, dramatically different from that of all animals. Perhaps Homo sapiens were given a gift for making speechlike sounds and the making of tools, but this did not mean that they understood things on a different plane from all other creatures. Kanzi’s language acquisition seemed to announce dramatically that language acquisition was first and foremost a feat of understanding. The actual production of the sound was a matter of possessing the right peripheral apparatus to do so. The understanding of language, however, was a matter of comprehending the intended meaning behind the sounds, and Kanzi clearly was doing this.

  For so revolutionary a scientific claim as this one, a persuasive body of data would be required, and as yet, I had only my notes of what Kanzi had done. Would anyone believe those? Would anyone believe anything without a number of detailed blind tests? I doubted it. I knew that convincing others would be a difficult task, but I also knew that if I were to focus too intently on proving everything Kanzi said or did, I would lose his natural engagement in the language process—a process that is comprised of communication and negotiation, not proof.

  Given what Kanzi could already do, the only logical research strategy seemed to be to abandon any and all plans of teaching Kanzi and simply to offer him an environment that maximized the opportunity for him to learn as much as possible.

  On a practical level, this meant rapidly increasing the number of symbols available to Kanzi, and figuring out how to make them interesting and important from his perspective. Since apes spend the better part of their day traveling from one food resource to another, I reasoned that the names of foods and locations should be among the most intrinsically interesting items to the ape mind. The Language Research Center’s fifty-five-acre forest provided an ideal context for such learning, and the nature of the vocabulary would become pertinent if foods could reliably be found at specific locations as they are in the wild.

  We set up seventeen locations in the forest where Kanzi could reliably find specific foods. He could travel between them as he wished. Different games and activities evolved at each location, thus helping make each unique. For example, at LookOut Point, which was high off the ground, we often played a game of hiding objects. We hid pine cones in our shirts, pine needles under blankets, balls in the leaves, and so on. Kanzi loved these games of object hiding and would initiate them sometimes by hiding an object himself, and sometimes by saying things like pine-needle hide or shirt hide. At the location we called Treehouse, we rarely played hiding games with objects, but rather hid ourselves. Kanzi would sit high in the treehouse and wait while all of us hid ourselves in the dense foliage below. Then he would find us one by one. We always told him not to peek, but he hated that and would always look. Near the location known as Midway, Kanzi was always the one who hid. Here the thicket and undergrowth were dense and we hated to go into it. Kanzi would therefore use this opportunity to vanish quietly, usually no more than four to ten feet away, but nonetheless out of sight. We would call and call and pretend that we could not see him. He thought it great fun that he had eluded us and would sometimes sit quietly under a bush for ten or fifteen minutes before showing himself.

  Travel was always accompanied by tree climbing, looking at small animals and insects, and learning about the naturally edible plants that were in the forest. We assigned words to these activities and objects and used them whenever the occasion warranted. No day was planned; it simply evolved, and each day was different. Sometimes we had to contend with dangers like thunderstorms, snakes, or floods; at other times there were visitors and surprises. Initially, either I or one of the other caretakers decided where to go, as Kanzi did not know how to use the symbols to initiate travel to a particular location.

  On a philosophical level, the switch from structured training to laissez-faire learning meant that Kanzi, not an experimenter, would decide which words were acquired and what they meant. By believing I had to teach language to Sherman and Austin, I also had implicitly assumed that their abilities were limited and that we should decide which words they were ready to learn given our experim
ental questions. It now seems odd that anyone had ever decided which words an ape was to learn—certainly no one does that with children, who elect to learn very different sorts of words, apparently focusing on the aspects of language that fascinate them.

  Could it be that our assumptions of limited abilities had inextricably led to circumscribed learning, even though we had been attempting to press the skills of Sherman and Austin to the limit? Could the same assumptions have limited the progress of Washoe or Sarah? If I had simply assumed that Kanzi was able to learn as humans do, might not those very assumptions have produced a very different sort of ape?

  With Matata gone, I, and the other humans who helped care for Kanzi, became the focus of his life. He was now free to be fully attentive to the things we were interested in, without the constant maternal monitoring that had previously occupied him. Even though we wanted Kanzi to learn to communicate, we did not aspire to rear him as a human child. Instead, we wanted him to be a happy, well-adjusted bonobo who liked people. Nonetheless, we did encourage some humanlike activities, such as toilet training, simply because they made daily life more pleasant for everyone. Unnecessary human cultural predispositions, such as the need for privacy with regard to elimination, bathing, and so on, were not requested of Kanzi. Similarly, we did not attempt to clothe him for the sake of appearances or decency. Sometimes, when it was chilly out, he wore a sweat shirt or sweater, but he ruled out pants and shoes early on.