Kanzi: The Ape at the Brink of the Human Mind

by Sue Savage-Rumbaugh

  Each evening toward the expected birth date, someone from the Georgia State University Language Research Center would drive to the field station at Lawrenceville, twenty-seven miles northeast of Atlanta, to be on hand to witness the birth and call the veterinary staff in case of complication. Rose Sevcik was on duty on the twenty-eighth, and she called me as soon as she was certain that the birth had started. By the time I arrived, Kanzi had been born, a tiny ball of black fur with spindly arms and legs. The situation with the bonobo group seemed a little tense. Instead of resting with her baby, Lorel was pacing about the cage looking tired and bewildered. The other bonobos were fascinated by the new infant and constantly pestered Lorel with requests to look at, prod, and hold the new baby. Finally, exhausted, she lay down and closed her eyes, with her arms around Kanzi as he clutched tightly to her waist, known as the ventrum, with his small hands.

  Matata, one of the original three wild-caught bonobos in the Yerkes colony, approached Lorel and sat down quietly beside her, looking fondly at the new baby while cradling her own infant, Akili, in her lap. As Lorel peeked out of sleepy eyes, Matata gently caressed Kanzi’s tiny hands, face, and feet. Lorel saw that Kanzi was not objecting and her eyes closed with heavy fatigue. Noting this, Matata leaned down next to Lorel and slowly slipped Kanzi’s long, thin arms across her own ventrum. Lorel did not notice, so Matata carefully tugged on Kanzi’s spindly leg and also slipped it onto her tummy. Kanzi’s hand and leg reflexively gripped onto Matata. Lorel, at that point, seemed to nod her head as though passing from light sleep into a deeper stage and Matata, who is a keen observer, took account of this and used that moment quickly to pull the rest of Kanzi onto her ventrum. Kanzi, at once, clung to Matata, just as though she were his own mother. Lorel opened her eyes, looked at Matata, then down at her own ventrum and at once realized what had happened. She cried out, in a manner that could only be described as “plaintive” and tried to take Kanzi back from Matata. But Matata had already moved her own son onto her back and covered Kanzi so thoroughly by wrapping her legs and arms about him that Lorel could hardly see him at all. Lorel began following Matata around, tugging on her, trying to get Kanzi back. But she was hesitant to bite or attack Matata, perhaps because she feared that others would side with Matata, or perhaps because she did not want to chance hurting Kanzi. Matata was clearly determined to keep Kanzi, and Kanzi, less than thirty minutes old, had not had time to determine who his mother was, and so he clung to Matata and resisted the tugs and pulls of Lorel as well.

  After several hours, Lorel gave up her attempts to pull Kanzi back and simply sat and watched Matata with her infant. Once Lorel was no longer trying to retrieve Kanzi, Matata began to hold him away from her body, as though she desired to look him over carefully. She held Kanzi about fifteen inches out from her ventrum and gazed intently at him as he vigorously moved his arms and legs, seeking to regain the comfort of her lap. Matata seemed to be somehow testing Kanzi, both by holding him out when he was clearly uncomfortable and whimpering, and by biting down gently but firmly on his fingers until he stopped clinging to her. It began to appear as though she were vacillating between “playing” with Kanzi as if he were an interesting toy, and caring for him, as if he were her own infant. Or perhaps she was simply doing this to frustrate and test Lorel. Lorel certainly became agitated with each of Kanzi’s whimpers during these “inspections.”

  Just then the veterinarian, Dr. Brent Swenson, arrived. The bonobos fear the veterinarian as it is his job to anesthetize them, and he uses a blowgun to do so. This darting is an unpleasant experience for all concerned and the bonobos attempt both to avoid and threaten him. When Matata saw Brent, she tried to get as far away as she could and hid Kanzi in her lap to protect him. Thereafter she treated Kanzi as her own son, never again exhibiting the unusual inspecting and toying behavior she had engaged in just prior to the appearance of the veterinarian. She was a devoted mother to both Kanzi and Akili, nursing them jointly with skill and adroitness. Lorel continued to sit by Matata and paid close attention to Kanzi for several days, but after that she gave up and treated him as though he were Matata’s infant.

  Since Matata now had two infants, Yerkes offered to trade Akili for Kanzi and the San Diego Zoo agreed. They removed Akili and placed him in the San Diego group, while Matata and Kanzi traveled to the Language Research Center together. We kept Kanzi with his adopted mother to foster appropriate orientation toward his own species. Had Matata not “adopted” Kanzi in the way she had, the course of ape-language studies would have been entirely different, as Kanzi would have remained with Lorel and not been exposed to language during his infancy.

  I had always been impressed by Matata’s evident intelligence and eagerness to communicate. She developed ways of letting me know quite clearly what she wanted. For instance, she would hand me her food bowl and push me toward the refrigerator when she was hungry, or point to the lock on the door when she wanted to go out. Matata had been part of the language program for one year prior to giving birth to her son Akili. During that time, she proved to be a willing and interested, though incompetent, study. She quickly understood that Sherman and Austin used the keyboard to communicate and that pressing the lexigrams was what achieved this feat. However, the idea that a specific lexigram was used in specific ways eluded her. Matata would take my hand and lead me to the keyboard, fully cognizant of what she wished to convey. However, once she began to use the keyboard, she would press any lexigram and then look at me as though I should now know her wish. Consequently, her desires and the lexigrams she chose to use to express them did not correspond on a reliable basis. Sometimes she pressed “juice” when she really wanted a banana, other times she pressed “groom” when she really wanted to go outdoors. Often, she did not reject the juice or the grooming, even if her real intent was different, as there was no reason not to accept the juice and the grooming that I offered her. How then did I know that the symbols she selected did not match her real intent? Generally, either the nature of her glance, or the events that she was attending to gave her away. For example, if Sherman and Austin had asked to go outdoors and were getting ready to do so, Matata often wanted to go with them. She would vocalize and look in their direction, and if I permitted, she would rush over to them. Consequently, in such a situation, pressing “groom” did not really seem to indicate what she wanted to say.

  Perhaps our hopes had been too high for Matata. Her early training had not been as systematic as that received by Sherman and Austin. She had not spent the long months learning to discriminate lexigrams, to request food items from a dispenser, to comprehend and carry out requests, or to differentiate between naming things and asking for them. Matata seemed so intelligent that we assumed she would be able to tell the lexigrams apart and utilize them for communicative ends.

  Prior to this time, no bonobo had been language trained, and there were many reasons to suspect that Matata, as a bonobo, should do even better than Sherman and Austin. Robert Yerkes had observed that bonobos appeared to be intelligent and humanlike in many ways, and my earlier investigations of the gestural communicative skills of Matata, Lokelema, and Bosondjo had confirmed that.

  Bonobos manifest a more intricate socio-communicative repertoire, including the use of more gestures and more vocalization, than common chimps do. These gestures are natural—that is, they are not trained but rather are reflective of the bonobo’s own inclinations to communicate. In a paper with my colleagues Kelly McDonald, Rose Sevcik, William Hopkins, and Elizabeth Rubert, I expressed my expectations this way: “Because elaboration of the gestural, visual, and vocal domains of communication must have occurred in evolution before the emergence of speech proper, the more extensive development of these skills in the pygmy chimpanzee, in contrast to other apes, suggests that they might be better prepared to acquire language.”1 Matata was to be the test of that proposition.

  Psychologists’ understanding of language acquisition in children had advanced significantly by the late 1970s. E
arlier, the simplistic concept of language had prevailed, with its extraordinary emphasis on syntax as the sine qua non of language. The newer concept, which was being developed by ape-language studies as well as detailed investigations of parent-infant patterns of interaction, was beginning to place language in a social context. Here, language is viewed as “a highly complex set of behaviors that is acquired through joint interactions that involve the intertwining of words and actions between two or more individuals.”2

  Both the results of the research with Sherman and Austin and new studies of children were suggesting that it was out of the combination of context, social interaction, and social expectancy that the child or ape steadily built up an understanding of the words that were being used around him or her. This understanding included the discovery that words could be used to refer to things, events, feelings, and so on. That is to say, words could serve as arbitrary replacements for objects and events in conveyance of meaning. This new account of language acquisition for the first time placed a major emphasis on comprehension of words and sentences as a key component of language skill in apes.

  The first attempts at language training with apes had ignored comprehension, assuming that production of symbols implied comprehension. Our work with Sherman and Austin had demonstrated the fallacy of that assumption. It made clear that while apes could come to comprehend symbols, the skill had to be put in place through experiences designed specifically to foster understanding. Simply teaching an ape the association between a word and a thing did not always result in the capacity to comprehend the intentions of others when they utilized the same word. The associations between word and referent too often ran in only one direction; the ape declared what it wished to have happen and expected the proper events to follow. When the tables were turned, and someone else announced what they wanted to have happen, it was found that these associations were not always reversible. Thus much of the ape’s understanding of language was limited to the ape as speaker.

  Comprehending and producing language proved to be very different sorts of affairs. When apes produce symbols, they are attempting to affect the behavior of others—for example, to ask for a banana. When apes comprehend symbols directed toward themselves, they are expected to bring about the effect intended by the user of the symbols. Consequently, by focusing on the ability of Sherman and Austin to comprehend symbols, we were forced to develop paradigms in which the execution of the symbol and the ape’s receipt of some object or activity associated with that symbol became completely detached. This marked a dramatic break with all other ape-language efforts, and it led to the apes recognizing that symbols can be used to communicate information about a specific object, event, or whatever without being tied to the occurrence of that event.

  When Sherman and Austin reached this level of understanding of symbol use, we observed the spontaneous emergence of the capacity to use language to express future intentions. For the first time, it seemed that they really “had words”; that is, they understood that words could be used to express future intentions and thereby coordinate actions, rather than simply as a mechanism to get others to do something for them. Their ability to produce statements regarding their future intentions represented a profound advance in ape-language studies. Once apes could make statements about their intentions, and then carry out such statements appropriately, their behavior could no longer be explained by condition-response chains.

  The work with Sherman and Austin had therefore set the stage for what we were to see in Kanzi in four ways: first, we had learned that “words” are more than simple associations between object and referent; second, we had learned that apes can appear to be able to produce relatively complex utterances without comprehending such utterances in the speech of others; third, we had discovered the importance of concentrating on language comprehension and the breaking down of the stimulus-response chains this entails; and fourth, we had found that comprehension leads to the ability to use language to make accurate and experimentally verifiable statements about intentions of future behavior.

  We had thought that by working with a different species of ape we would gain further insights. Our initial results with Matata had not been encouraging, and Matata was fast approaching maturity. We hoped that perhaps, at some future time, we would be able to work with a young bonobo.

  When Kanzi was six months old, this opportunity arrived. The Yerkes Great Ape Committee, which oversees the assignment of apes, assigned Kanzi to the Language Research Center as long as he could remain with Matata while he participated in language studies. I was looking forward to working with Matata again, because I had formed a close and trusting relationship with her. I knew that she would be excited to be at the new and spacious laboratory that we had recently been blessed with—fifty-five acres of primary forest on Georgia State University land. I anticipated that Matata would welcome a chance to be in the forest once again and that our relationship could be readily renewed. However, I was uncertain as to how Kanzi would react.

  Unlike Matata, Kanzi had never interacted with me before, and all of his friends had been bonobos. Shortly after Matata and Kanzi were settled in their new quarters, Matata gestured to ask me to come into the cage. I entered rather circumspectly, hoping not to startle Kanzi. As soon as I was within a few feet of his mother, though, Kanzi emitted a piercing scream and leapt into my arms from Matata’s ventrum, doing a midair acrobatic twist as he went. With his lips pulled back across both sides of his face, he looked directly at me, screaming with all the power his young six-month-old lungs could muster.

  Having never been greeted by a bonobo infant before, I was uncertain how to interpret Kanzi’s actions. Because his screaming was so intense, I worried for an instant that Matata might view this excitement as a sign that I was scaring Kanzi. Even though I counted Matata as a friend, I expected her to be fiercely protective of Kanzi, and to bite me if she thought my behavior potentially warranted any such action. I stood there, with Kanzi’s arms and legs wrapped securely around my waist, trying to use all the nonverbal skills at my disposal to indicate to Matata that I was doing nothing bad to Kanzi in spite of the intense noise emanating from his small body.

  I should not have worried. Matata understood Kanzi’s emotion far better than I and thought it perfectly acceptable that Kanzi should scream and leap away from her to give me such a boisterous greeting. She continued to observe us as Kanzi gradually calmed down and began to explore my face, hair, and clothing. He was particularly fascinated by my nose, which he seemed to note was distinctly different from those he had been accustomed to seeing. This highly demonstrative greeting marked the beginning of what came to be a long and deep friendship between a human being who wanted to understand what it meant to be an ape and a bonobo who would strive with equal intensity in his own way to understand what it meant to be a human being.

  The research plan laid out for Matata had been to pick up where her training had stopped several years back, when she was in our Yerkes laboratory with Sherman and Austin. We did not intend to work with Kanzi for some time, as six months was too young to begin any sort of systematic training. Consequently, Kanzi simply spent all day with Matata, doing whatever he could do to entertain himself.

  Although Matata’s previous progress had been disappointing, she had learned how to tell one lexigram from another, which was an important first step toward using them for communication. She could also make a few requests, but she still did not respond appropriately to usages by others. Thus she could not select an apple from the refrigerator if asked to, but she could herself ask for apples in a limited sense.

  We began a systematic teaching program using essentially the same computerized keyboard equipment we had previously employed with Sherman and Austin, and we introduced new symbols to Matata very slowly, so as not to confuse her. Matata was eager to learn. She appeared to recognize that we were attempting to use the symbols for the purpose of communication. She was also patient, which proved quite a valuable trait,
as Kanzi was a hyperactive infant. He ran around the test room, jumped on Matata’s head, pushed her hand away from the keyboard as she tried to select the correct symbol, and stole the food she earned as a reward.

  At times, Kanzi became mesmerized by the keyboard, staring at the symbols as they flashed onto the projectors at the top of the keyboard. He tried to grab each one just as it lit up, as though it was somehow crucial to catch it at just the moment it appeared. So occupied would he become with this activity that for ten or fifteen minutes nothing mattered to him except catching the symbols as they appeared. Then, he would turn and suddenly ignore them entirely, as if they did not exist.

  Matata, like many other bonobo mothers, indulged Kanzi to the extreme. She made it plain that she did not approve if I or anyone else attempted to discipline Kanzi, regardless of the gravity of his mischievous behavior. She permitted him to behave in ways toward me that she would never consider doing herself. For example, if Kanzi was angry about anything that happened around, near, or with me or another person—regardless of whether we had anything to do with the situation—he was permitted to bite us. And we were supposed to ignore him, or better yet, distract him. It seemed that in Matata’s view, Kanzi, as an infant, could not and should not be held responsible for his own actions—much as we, in our culture, also do not hold children or animals responsible for their actions.

  When he was about fourteen months of age, Kanzi began occasionally to press keys on the keyboard and then run to the vending machine as though he had grasped the idea that hitting keys produced food. However, he gave no discernible indication of understanding the relationship between specific keys and specific foods. Rather, his use of the symbols was generally sporadic and playful. By the time he was two years old, he started deliberately to select the “chase” symbol. He would look over the board, touch this symbol, then glance about to see if I had noticed and whether I would agree to chase him. If I answered yes, either by smiling, nodding my head, or saying “yes” on the keyboard, he would run off, looking back with a big play grin on his face. He also began to use the “chase” gesture (a hand clap), which he had seen Sherman and Austin use between themselves. We were pleased with Kanzi’s use of chase, but at that point did not recognize what this spontaneous learning portended for Kanzi’s future.