Kanzi: The Ape at the Brink of the Human Mind

by Sue Savage-Rumbaugh

  Kanzi was aware that we employed the keyboard as a means of communication and apparently felt keenly motivated to do so as well. He also felt motivated to cooperate with toilet training, though he had been indifferent to this activity while Matata was present.

  Kanzi’s communications soon began to revolve around his daily activities, such as where we were going to travel in the forest, what we would eat, the games we wanted to play, the toys Kanzi liked, the items we carried in our backpacks, television shows Kanzi liked to watch, and visits to Sherman and Austin. We found that the computerized keyboard was impractical for such outdoor use, and instead used a board on which photographs of the lexigrams were arrayed. Whenever Kanzi’s caretakers talked among themselves or directly with him, they combined spoken English with pointing to the appropriate symbols. They also treated Kanzi’s utterances as though they were intentional and conveyed what they appeared to convey, even if there was no “proof.”

  From Sherman and Austin, I had learned that the attribution of meaning comes from the behavior of the receiver of the utterance—and that the behavior of the receiver is critical to the maintenance of veridicality as words are being acquired. For example, when Sherman was learning the word “wrench,” he often selected the symbol for “key,” even though I knew he meant to say “wrench.” Instead of telling him what he should have said, I acted as though he had intended to say “key” and responded by looking for the key and then giving it to him. In so doing, I let my behavior define for Sherman the meaning of his utterance. Thus the symbol “key” meant the object key to me, and that was what I focused my attention upon, even though I knew Sherman was wrong and that he needed to use another word. I did not tell him so directly. I let my behavior speak for itself to clarify for Sherman what “wrench” meant to me, thereby preserving the veridicality between the word and its meaning. If Sherman wanted me to act differently, for example, to pick up a wrench and give it to him, then he had to request that using different symbols—ones that “meant” wrench to me.

  By maintaining a veridical response to each other’s words, we establish a joint meaning for each word that is no more and no less than the sum of all such veridical responses within a community of common speakers. Therefore, the goal was not to determine whether Kanzi fully understood what a word meant when he said it, but rather to treat his saying it as a meaningful utterance and to respond as we would to any other party who used the word. Such responses, we had learned from Sherman and Austin, play a critical role in establishing “meaning.”

  Within four months, Kanzi’s vocabulary rose from the original eight symbols to more than twenty. These symbols referred to foods, locations in the forest, activities, and people. By this time, too, he had learned the location of all seventeen food sites in the forest. He could announce his wish to go to any one of them, either by using the appropriate symbol or by pointing to a food that corresponded to a particular site (for example, at Treehouse there were always bananas and juice, at Crisscross Corners, blackberries, cheese, and orange juice were to be found, and so on); and he could guide us to any of the locations we named. After he had indicated his wish to go to, say, Treehouse, by pointing to a photograph of a banana, Kanzi might carry the photograph on the journey, frequently pointing again to the picture and vocalizing. Soon the symbols replaced the photos.

  These symbols were added to Kanzi’s keyboard at five years of age. Many are more abstract than some of his earlier symbols because we began to realize by that time that our expectations of him needing “concrete” symbols were mistaken.

  The first symbol here is “bad,” one that Kanzi acquired right away and often used to announce both his displeasure with us and his intent to do something that we were not likely to approve. Although Kanzi used “bad” only to refer to himself, his sister Panbanisha recently repeatedly used this symbol to comment upon Kanzi’s action of biting someone she liked.

  The next symbol is “now.” We use this to tell Kanzi when we are ready to do something “right now” as opposed to “later” (also on the keyboard). Kanzi, however, rarely uses this keyboard symbol. He has devised a vocalization to communicate the same message and thus does not need the printed symbol.

  The third symbol is “dessert,” one of Kanzi’s favorite things. He enjoys helping construct desserts and has come up with many of his own food mixtures.

  The other symbols in the top row are “tummy,” “bowl,” “monster,” “coconut,” and “towel.”

  To test Kanzi’s capacity to state his intentions to travel to various locations, we enlisted the aid of Mary Ann Romski, who works with children at the Language Research Center. Mary Ann and Kanzi had developed a close relationship during Kanzi’s frequent visits to the “childside,” but as Mary Ann was somewhat afraid of the snakes and creatures in the forest, she had never accompanied Kanzi into the woods. Mary Ann was therefore the perfect person for this experiment. She did not know the trails nor the locations where food was to be found. Moreover, Kanzi seemed anxious to show her around in his world, which led me to suspect that he would announce his intentions and lead Mary Ann to the locations he spoke of. I knew that Mary Ann would be completely lost in the woods without Kanzi’s help. If Kanzi said he was going to the Treehouse, for example, and then led Mary Ann there, his behavior could not be explained on the basis of subtle cues from Mary Ann. Moreover, Mary Ann, being somewhat fearful of the forest, would take great comfort in being with Kanzi, a companion who was totally at ease in this element. She would trust him and his intuitions, and Kanzi would do his best to make Mary Ann feel “at home” in his world and to show her around.

  Mary Ann recorded Kanzi’s utterances and the route he led her along in going from one food site to another. During the test, Kanzi used photographs on five occasions and lexigrams on seven to announce a proposed destination, with 100 percent accuracy. Moreover, on all but one journey between two sites, Kanzi chose the most direct route possible. On that one occasion, he exploited the opportunity of being accompanied by someone who did not know that where they were to go was a part of the forest he normally is not allowed to go. Kanzi therefore demonstrated not only what he knew about himself, but also what he knew about Mary Ann.

  Kanzi’s capacity to use lexigrams to state his intentions was clearly evident in the test in the forest. Sometimes it took twenty minutes for him to lead Mary Ann to the destination he had earlier specified, and along the way he often talked of other things, without forgetting where he was going.

  Baby Kanzi shortly after he arrived at the Language Research Center. Kanzi arrived accompanied by his mother, Matata, when he was six months old. (Photograph by Sue Savage-Rumbaugh)

  Kanzi at two years of age. (Photograph by Sue Savage-Rumbaugh)

  Matata walks in the woods with baby Kanzi on her back. Foods were hidden under and around certain trees with markers on them. We found that we needed the markers to recall where we had hidden the food, but Matata did not. She could remember individual trees with uncanny precision. (Photograph by Sue Savage-Rumbaugh)

  A bonobo in the wild carries her infant similarly. She must remember not only specific trees, but when they are fruiting as well. Bonobo mothers carry their infants up to five or six years of age, but only very young infants, less than five or six months, are carried continually. (Photograph by Frans Lanting)

  Kanzi shows me where he wishes to go by gesturing. He began to gesture at about twelve months of age and continued to use gestures intermingled with lexigrams after he began mastering the formal symbols. Bonobos do not typically direct others by such gestures in the wild. (Photograph by Elizabeth Pugh)

  Kanzi at four years of age talking to himself on the keyboard. He began accompanying people on daily outings in the woods slightly before his third birthday. By four years of age he knew the forest far better than we did. He hated the cold months when he had to stay inside, so in the fall we let him wear sweaters to extend the time he could be outside. (Photograph by Elizabeth Pugh)


  By three years of age, it became apparent that Kanzi was learning lexigrams readily all on his own. He especially liked traveling out-of-doors, so we developed a portable keyboard. Here is our earliest attempt, a computer in a suitcase in 1983, long before the appearance of the first commercial portable computers. When Kanzi touched a symbol, it lighted up and the computer kept a record. (Photograph by Elizabeth Pugh)

  Kanzi uses the computer to comment “apple” after seeing someone take an apple out of the backpack. Kanzi often commented on things when he was small, whether he wanted them or not. (Photograph by Elizabeth Pugh)

  Between four and five years of age, Kanzi loved to take the keyboard aside and say things to himself. If we approached and tried to see what he was saying, he would pick up the keyboard and scurry further away. He could communicate by pointing to the symbols with the keyboard laid flat on the ground, but he preferred to prop it up in the vertical position as he saw us do. Here he struggles to properly stand his keyboard up. (Photographs by Elizabeth Pugh)

  Kanzi watching one of his many favorite television programs. We set aside a scheduled time each day when Kanzi could watch TV. He selected his favorite programs by pointing to photos on the outside of the various tape cassette boxes. Often we made videotapes for him of things that were familiar from around the lab, using a “story line” to keep his attention. Kanzi typically prefers our “bonobo-oriented” videos, but also likes movies such as Greystoke, Iceman, and Quest for Fire, in which the actors portray “primitive man.” (Photograph by Sue Savage-Rumbaugh)

  Bonobos frequently express pleasure and happiness by smiling just as we do. Unfortunately, many people mistake Kanzi’s smile for an expression of aggression because his teeth look frightful to them. Kanzi, however, uses exactly the same facial muscles to produce a smile as we do and his lips and eyes assume precisely the same countenance as ours when we smile. (Photograph by Sue Savage-Rumbaugh)

  Lana, the first ape to use a keyboard system, constructs one of a number of complex “stock sentences” she was taught; here, “?You put piece of bread in machine.” When Lana touched the symbols on the keyboard, the image was projected above the board in the order of her selection. After food was loaded into the machine at her request, Lana then had to extract each piece from the dispenser using other stock sentences, such as, “Please machine give piece of bread.” (Photograph by Frank Kiernan)

  A computer “read” Lana’s sentences, and if they were grammatically correct and corresponded to the food that was actually in the machine, the computer vended Lana a piece of the requested food. Lana revealed to us that she could go beyond the “stock” sentences to construct novel appropriate requests. For example, when the machine was broken and food could not be loaded, Lana was able to ask, “?You move food into room?” (Photograph by Frank Kiernan)

  Austin (on the right) has just watched as a particular type of food was hidden in the plastic container. Austin has learned that he can have this food only if both he and Sherman ask for it correctly. Sherman, however, does not know what food to request, since he did not see the food as it was placed in the container. After realizing that Sherman does not know what to ask for, Austin uses the keyboard to identify the food he saw hidden. Sherman watches, then asks for the correct food. (Photograph by Sue Savage-Rumbaugh)

  Austin and Sherman sharing food at a picnic using their version of the portable computer keyboard. By ages eight and nine, they had mastered ninety-two symbols and could readily use them to communicate with each other as well as with people. (Photograph by Elizabeth Pugh)

  One day Kanzi wanted to go for a ride in the lab van but found it locked. When no one would give him a key, he tried the next best thing—a screwdriver that the shop foreman had left nearby. (Photograph by Elizabeth Pugh)

  Kanzi’s younger sister, Panbanisha (one year old here), studies a photograph of one of the places where food is hidden in the forest. When the apes were too young to understand symbols, we often showed them photos of the places we planned to go in this way. They liked this practice immensely and studied the pictures intently, as Panbanisha does here. (Photograph by Sue Savage-Rumbaugh)

  Feral born and reared, Matata studies her mirror image. Although she has not been able to learn language, like most apes she quickly recognized herself when given the opportunity to gaze into a mirror. Matata uses the mirror to facilitate the extraction of tiny hairs from her chin, so she will not appear to have the least sign of a “beard.” We do not know where she acquired this practice, because it was not demonstrated for her; perhaps bonobos have something akin to vanity in the wild. (Photograph by Elizabeth Pugh)

  In this important experiment, Sherman has been shown some bananas being put into a box on his side c a divided room with a window separating him and Austin. Austin does not se this happening. The box has been locked, so that Sherman will need a key to open it. Here Sherman goe to the keyboard and state “give key.” (Photographs b Sue Savage-Rumbaugh)

  Austin watches from his side of the room and reaches into the tool kit for the key.

  Sherman (left) approaches and Austin hands him the key.

  Sherman goes back to the box, inserts the key into the padlock, and twists the key to open the lock.

  Austin watches carefully as Sherman opens the lock.

  Sherman (left) takes half of the food back to Austin (he tastes Austin’s portion on the way).

  Sherman passes Austin’s portion to him through the opening in the window.

  This is one of the indoor rooms where Kanzi now plays and works when he cannot go outside. It includes a television with an attached joystick and two keyboards mounted in study racks. On the floor is a bin of objects and photographs for Kanzi. (Photograph by Steve Winters)

  Austin watches a videotape of himself. (Photograph by Elizabeth Pugh)

  Working with Kanzi at the keyboard. (Photograph by Steve Winters)

  In one of the early blind test formats presented to Kanzi, experimenter A (with back turned) places three photographs inside a gray plastic “text booklet” and then passes the closest booklet to experimenter B (facing Kanzi). Experimenter A then plays a tape-recorded word to Kanzi. The word is randomly chosen from Kanzi’s suspected vocabulary. When Kanzi hears the word, he gesturally asks experimenter B to open the test booklet and show him the photographs. Kanzi looks over the options, selecting the one that corresponds to the word he heard on his headphones. After Kanzi makes his selection, experimenter B tells Kanzi whether or not he is correct. In this example, Kanzi has selected a photograph of fire, and when experimenter B turns over the answer card, she finds the word fire. (Photographs by Nicholas Nichols)

  Kanzi often enjoys drawing and painting. Sometimes he labels his drawings, but as yet he has not achieved sufficient skill to produce recognizable images—a skill that also eludes most children under the age of three. (Photograph by Steve Winters)

  Kanzi plays grab, tickle, and chase with Jeannine’s son, Nathaniel, who has learned lexigrams himself. Yerkes veterinarians do not permit apes to interact directly with children for fear of transmission of childhood diseases, so they are separated here by a plastic barrier. (Photograph by Nicholas Nichols)

  Kanzi bashes one stone against another to produce a small sharp-edged flake which he will use to cut through a thick nylon rope holding tight a cover on a box filled with a favorite food. (Photograph by Nicholas Nichols)

  This panel represents one quarter of the symbols on Kanzi’s keyboard. By the time Kanzi was seven years of age, he comprehended three-quarters of the symbols on his 256-symbol keyboard, though he regularly used only about half of them. Symbols he did not learn were very abstract, such as “away,” or for things that he rarely encountered, such as pomegranate. From left to right, the symbols stand for: Bad, Now, Dessert, Tummy, Bowl, Monster, Coconut, Towel, Taco, Chicken, Lettuce, Noodles, Sugar, Bunny, Burrito, Butter, Away, Slap, Salt, Observation room, Trash, Perrier, Bottom, Strawberry, Kiwi, Pillow, Pomegranate, Grapes, Privet berries, Panzee, Pinke
y, Yogurt, River, Jello, Lana, Backpack, Koolaide, Wipie, Noise, Popsicle, Swimming pool, Panbanisha, Cold, Draw, Middle test room, Carry, Shop, Thank you, Karen, Honeysuckle, Toy, Dan, Gorilla, unassigned, Book, unassigned. (Photograph © Language Research Center, Georgia State University)

  Here Kanzi and I are at Lookout Point in November when it is starting to get fairly chilly. Kanzi has just told me he wants to go on to Flatrock to look for M&Ms and for a ball. (Photograph by Nicholas Nichols)

  Matata’s breeding sojourn was successful and we were pleased to find that she would soon return to the laboratory, pregnant. Despite the joy that another bonobo would bring, however, we wondered if Kanzi would continue to use the keyboard when Matata returned. After all, he had rarely used it before she left. Maybe he would simply want to stay with her again and the importance of his “human friends” would fade into the background. We talked endlessly about what should be done. Some suggested that we should not let Kanzi rejoin Matata, for fear we would never learn what he might be capable of communicating. Keeping a young bonobo away from his mother, when both lived in the same laboratory, seemed to me to be an impossible feat, as well as an unbearable cruelty for the sake of science. We reached a compromise in which Kanzi would be allowed to interact with Matata through the bars of a cage for the first few days. We would then evaluate how we should proceed, based on Kanzi’s interest in continuing to communicate.

  Matata arrived at the center one day in September, drugged, while Kanzi was in the forest. I stayed with her in the colony room as she came out of the anesthesia. She was delighted to see me, and vocalized happily when she heard Sherman and Austin’s calls from nearby. Eventually, Kanzi returned, hot and tired after a long day in the forest. I sat down with him and told him there was a surprise in the colony room. He began to vocalize in the way he does when expecting a favored food—“eeeh … eeeh … eeeh.” I said, No food surprise. Matata surprise; Matata in colony room. He looked stunned, stared at me intently, and then ran to the colony room door, gesturing urgently for me to open it. When mother and son saw each other, they emitted earsplitting shrieks of excitement and joy and rushed to the wire that separated them. They both pushed their hands through the wire, to touch the other as best they could.