Kanzi: The Ape at the Brink of the Human Mind

by Sue Savage-Rumbaugh

  Many Oldowan tools have just two or three flakes removed, while few have more than a dozen. Once the first flake is removed, the direction of subsequent flaking appears to be fixed. The complexity of the tools, in relation to spatial concepts of their manufacture, is therefore rather limited. What of the spatial concepts involved in the natural tool-making of chimps?

  Most of the tools chimps make are of raw material other than stone, and so a direct comparison is impossible. And many of the tools’ characteristics are influenced by the nature of the material. For instance, the position of leaves to be stripped from a twig is a property of the plant, not a decision of the ape. However, suggest Tom and Bill, it is possible to find examples of the three spatial concepts in several realms of chimp behavior. For instance, chimpanzee “artists” do not scrawl randomly on blank pieces of paper. And if they are given a sheet of paper on which a geometric shape, such as a square, has been drawn, chimps tend to mark on or near the shape, thus displaying the concept of proximity. They exhibit the property of boundary in the preparation of sedge stems as termite fishing-probes. The stem is triangular in cross section, and the chimps carefully remove one of the ridges by longitudinal stripping. Finally, the concept of order is to be found in chimpanzees’ nest building, a skill that is learned rather than innate, as it is in birds. When a chimp builds a nest, the choice of the first major branch to be bent provides the foundation of the subsequent pattern of interweaving of smaller branches and twigs.

  For Tom and Bill, the conclusion was clear: “All the spatial concepts for Oldowan tools can be found in the minds of apes. Indeed, the spatial competence described above is probably true of all great apes and does not make Oldowan tool-makers unique.”7 And they respond to the positing of the “using a tool to make a tool” as a Rubicon, by pointing out that Richard Wright’s orangutan learned to do just that, so the skill is within apes’ cognitive realm. Moreover, they make the reasonable point that chimps have sharp teeth, and use them to fashion wooden tools; they don’t need stone flakes for the job. “We cannot fault apes for not employing unnecessary techniques in making tools needed in their subsistence,” they add.8

  When they examined ape and early human subsistence strategies, including tool-use, Tom and Bill again see no great cognitive divide. For instance, anthropologists often point to two aspects of early human strategies as evidence for advanced mental powers. One is the fact that several different types of raw material may be used for tool manufacture at one site. And the second is that raw material for tool-making is often transported from as far as ten miles distant. But, say Tom and Bill, the chimpanzees of the Taï Forest use granite rocks to crack the hard Panda nuts while using softer rock or even a wooden hammer for Coula nuts, which are easy to crack. Moreover, the chimps often carry their rock hammers a third of a mile to the nut-cracking stations. This latter point is therefore a quantitative not a qualitative difference, and surely has to do with a bipedal animal’s greater ease of carrying things.

  Modern chimpanzees are known to be occasional hunters, whereas early humans are inferred to have hunted, although some anthropologists suggest they were exclusively scavengers. Early humans used stone tools to gain access to meat on a carcass, whereas chimpanzees use their teeth to the same effect. It is true, however, that early humans sometimes processed the carcasses of large animals, whereas chimpanzees’ meat-eating is exclusively of small-to medium-sized creatures, such as small antelope and monkeys. The use of sharp cutting tools allows for butchery of large carcasses, an activity for which canine teeth are not suited.

  Judging from the evidence of fossilized animal bones on archeological sites, early humans frequently broke open bones to gain access to the marrow. Archeologists characterize this as an “extractive” process, a term that can also be applied to the removal of meat from a carcass. This does not distinguish early humans from apes. For instance, probing for termites, breaking open nuts, and using leaf sponges to soak up fluids or brain material are all extractive processes, and each is part of the subsistence repertoire of chimpanzees.

  “What we know about Oldowan foraging seems to be within the capabilities of apes,” conclude Tom and Bill. “The details differ, but this might be expected given local differences in habitat. Oldowan foraging appears to have been that of a hominoid who lived in semi-arid, open grassland and who combined scavenging of carcasses and hunting small game. There is no evidence for a dramatic reordering of general hominoid foraging, nor for an evolutionary leap in the cognitive capabilities underlying it.”9 No one who spends many hours in close contact with chimpanzees, as I have, can fail to be impressed with their keen cognitive abilities. Tom and Bill’s assessment, though regarded by some as an extreme statement of ape abilities, is surely plausible. Still, when we began the tool-making project with Kanzi, I wondered whether we were embarking on a task that required more patience and motor control than he possessed. He might understand why he needed a tool and even how to make one. Indeed, I thought this would be the easy part for him. But I was not at all certain that he could actually do so since stone tool construction requires a great deal of bimanual eye-hand motor coordination, as well as proper timing and orientation. It seemed to me that the mechanical demands of flake construction went far beyond those required for termite fishing or Panda nut cracking.

  Nick Toth was to be Kanzi’s model, as I was too incompetent a flaker to teach Kanzi anything properly. But Kanzi considered Nick a stranger, and, as an adult male, felt it his duty to frighten away all “outsiders.” Thus it was not surprising that he was initially aggressive toward Nick: He puffed himself up in a great display, repeatedly rushed at Nick as he stood safely on the other side of the caged enclosure, and threw handfuls of cedar chips. Although Kanzi is gentle and even flirtatious with females, he is often aggressive to males at first encounter. I told Kanzi that Nick was there to help us, and that he should stop being aggressive. When Kanzi began to see for himself that Nick was indeed being helpful, he did calm down, but still did not fully trust Nick. When Nick got too close to the wire that separated them, Kanzi would stop watching how he made tools and begin to concentrate instead on opportunities for grabbing his shirt.

  At first, we had set up the tool site outside Kanzi’s cage, so that Nick could show Kanzi how it was possible to gain access to the baited box. He struck a cobble with a hammerstone, selected a sharp flake, and then cut the string securing the lid of the box. Kanzi got the treat that was inside. Nick did this several times, after which we put the tool site inside Kanzi’s enclosure. Nick knelt outside, making flakes. He handed sharp ones to me while I was inside with Kanzi, and I encouraged Kanzi to use them to cut the string. He very soon realized the utility of the sharp flake, and eagerly took it from whoever was in with him. He then quickly went to the tool site to open the box. He even knocked two rocks together on several occasions, but in a rather desultory way, and without producing flakes. Nevertheless, he was clearly emulating Nick.

  During that first afternoon, and throughout the project (which is continuing), Kanzi was never required to perform a task, but merely provided with the opportunity to participate if he wanted to. We wanted to motivate him to make and use flakes, and we hoped he would learn by example. As the days and weeks passed, he became more and more determined and proficient, and he displayed a degree of persistence at the task that exceeded anything I’d seen him do. It seemed we had a potential tool-maker in our midst.

  Kanzi very quickly learned to discriminate between sharp flakes and dull ones, using visual inspection and his lips. On the second afternoon, for instance, we gave him a series of ten trials, in which we provided him each time with five flakes. For the first few trials, he seemed to be working from trial and error, but in the final five he unerringly picked the sharpest flake. He developed a very keen eye for good flakes as they were being produced. On one occasion, about three weeks into the project, Rose Sevcik was striking a rock when, for the first time for her, it split and several flakes flew off
in different directions. Kanzi was watching closely and seemed to know which was the best flake, even before they hit the ground. He let out a bonobo squeal of delight, rushed to pick up the sharpest flake, and was off to the tool site with it, all in one fluid motion.

  Making flakes for himself, however, proved difficult. At first, he was extremely tentative in the way he hit the rocks together. Almost always he used his right hand to deliver the hammer blow. He held the core in his left hand, often cradled against his chest, or sometimes braced against the floor, with his foot adding further support. Sometimes he put the core on the ground and simply struck it with the hammerstone. No one had demonstrated this “anvil” technique to him. No matter how he held the core, however, he seemed unable or unwilling to deliver a powerful blow. Bonobos are three times stronger than a human of the same size, so there was no doubt that Kanzi had the muscle power to do the job. We wondered whether he was nervous about hitting his fingers; perhaps he lacked the correct wrist anatomy to produce a “snapping” action; or perhaps he was reluctant to deliver a hard blow, because throughout his life we had discouraged him from slamming and breaking objects.

  Then, one afternoon eight weeks into the project, I was sitting in my office, which is close to the inside room where the tool site was set up. I was suddenly assailed with the sound of a BANG … BANG … BANG. It kept on and on, and I wondered what on earth had happened. I rushed to the tool-site room, and there was Kanzi, stone knapping with tremendous force. He had finally learned how to fracture rocks to make sharp flakes, albeit small ones.

  During the first three months of the project Kanzi became steadily more proficient at producing flakes, in part because he seemed to have learned to aim the hammer blows at the edge of the core. But despite his willingness to deliver harder hammer blows than he had initially, he still wasn’t hitting hard enough to produce flakes bigger than about an inch long. Nevertheless, he persisted with his newfound concentration, and we in turn made the string that secured the tool site thicker and thicker, so that small flakes would wear out before they cut the string.

  One day during the fourth month, I was at the tool site with Kanzi, and he was having only modest success at producing flakes. He turned to me and held out the rocks, as if to say, “Here, you do it for me.” He did this from time to time, and mostly I would encourage him to try some more, which is what I did that day. He just sat there looking at me, then at the rock in his hand, then at me again, apparently reflecting. I wondered what he was thinking, because he did seem to be pondering weighty matters as he gazed at the rocks. Suddenly he stood up bipedally and, with clear deliberation, threw a rock on the hard tile floor with a tremendous amount of force. The rock shattered, producing a whole shower of flakes. Kanzi vocalized ecstatically, grabbed one of the sharpest flakes, and headed for the tool site.

  There was no question that Kanzi had reasoned through the problem and had found a better solution to making flakes. No one had demonstrated the efficacy of throwing. Kanzi had just worked it out for himself. I was delighted, because it demonstrated his ingenuity in the face of a difficult problem. I quickly telephoned Nick, and told him what had happened. I was so excited by the event that I didn’t give a thought to the fact that Nick might not be delighted too. He wasn’t. He was disappointed. “The Oldowan tool-makers used hard-hammer percussion, not throwing,” he said. “If Kanzi throws the rocks, the percussion marks will be random, and we won’t learn anything.” Our different reactions reflected, I suppose, the different interests of the psychologist and the archeologist. Nick said I had to discourage Kanzi from throwing, and I pointed out that that would be difficult. After all, he had found an efficient method to get what he needed. “Try,” said Nick. I agreed to try.

  Rose Sevcik came up with the obvious suggestion, which was to cover the floor with soft carpeting. The first time Kanzi went into the carpeted room, he threw the rock a few times and looked puzzled when it didn’t shatter as usual. He paused for a few seconds, looked around until he found a place where two pieces of carpet met, pulled back a piece to reveal the concrete, and hurled the rock. We have assembled a videotape of the tool-making project, which I show to scientific and more general audiences. Whenever the tape reaches this incident there is always a tremendous roar of approval as Kanzi—the hero—outwits the humans yet again.

  By this time, spring was approaching, and we decided to take the tool site outdoors, where Kanzi would have to resort to hard-hammer percussion once again, as there was no hard floor to throw against outdoors. Forced to abandon his throwing technique, Kanzi steadily became more efficient at hard-hammer percussion, delivering more forceful and more precisely aimed blows. Very consistently now, Kanzi was hitting the edge of the core and was more successful at producing flakes. The resulting cores sometimes were very simple, with just a couple of flakes removed, or, if Kanzi had persistently hammered at them, they had many small flake scars and steep, battered edges, some of which resembled the “eoliths,” or dawn stones, found in Europe in the decades around the turn of the century. There had been great controversy about these objects, with some arguing that they were true artifacts. They turned out to have been the product of natural forces, such as wave action or glaciation.

  Just as Kanzi was becoming quite proficient at hard-hammer percussion he foiled us yet again, which again delighted the psychologist and dismayed the archeologist. Kanzi discovered that even outside on soft ground he could exploit his throwing technique. This discovery seemed to be the result of a thoughtful analysis of the problem as well: He placed a rock carefully on the ground, stepped back, and took careful aim with the second rock, poised in his right hand. His aim was true, and the rock shattered. He continued to use this technique, and there was no way of stopping him. As far as I was concerned, we had presented Kanzi with a problem and he had figured out the best way to solve it—three times.

  Kanzi had become a tool-maker. But our question was, how good a tool-maker is he? Could he have stood shoulder to shoulder with the makers of Oldowan tools, striking flakes off cores as effectively as they did? Nick’s experience as an Oldowan tool-maker offered us a way of addressing these questions.

  Through teaching himself to make the apparently simple core forms and flakes of the Oldowan, Nick came to understand both the process of flaking and the product. “The mechanics of flaking stone are not intuitively obvious,” he told me.10 When he tried to teach me and some of the other Wenner-Gren conferees how to flake that afternoon on the beach in Portugal, I could understand what he meant. I was impressed by how very difficult it is to produce flakes, and the challenge is not simply to hit the core with sufficient force. You have to know where to aim the hammer blow, and how to deliver it.

  The initial inclination of the naive stone knapper is to hit the core hard enough so that a flake will pop out of the core, as if it were being chiseled out. But, as Nick demonstrated, the flakes come from the bottom of the core, not the top. The best everyday example of the principle of concoidal fracture at work in stone tool-making is the effect of a tiny pebble hitting a window: A cone of glass is punched out of the pane, and the exact shape of the cone is determined by the direction at which the stone hits the glass.

  For effective flaking by hard-hammer percussion, three conditions have to be met. First, the core must have an acute edge (one with an angle of less than 90°). Second, the core must be struck with a sharp, glancing blow, hitting about half an inch from the edge. And third, the blow must be directed through an area of high mass, such as a ridge or a bulge. With these conditions met, and starting with suitable raw material, one can form long, sharp flakes. Oldowan assemblages were often made from lava cobbles, which had been rounded through being carried along stream or river beds. The tool-makers sometimes struck flakes from one side of a cobble, producing the unifacial form of a chopper, or along two sides, yielding a bifacial form. Whatever forms are produced, they have the appearance of great simplicity. But as Nick correctly points out, “It is the process,
not the product, that reveals the complexity of Oldowan tool-making.”11

  Nick and Kathy Schick recently drew up a list of criteria by which to assess the technological sophistication of simple cobble and flake tools. “It was necessary to get beyond relying on gut reaction for distinguishing between true artifacts and naturally fractured stone,” explains Nick.12 The criteria are as follows:

  Flake angle. This is the angle formed between the striking platform of a flake and the dorsal surface of the flake (representing the edge angle of the core before the flake was detached).

  Degree of removal of outer unusable surface. Many flint stones are encased in a tough concrete exterior that crumbles when it flakes and is thus unusable as a tool. In order to get to the flint, which is an inner stone that can readily be flaked to produce a sharp edge, it is first necessary to remove this rough outer core. The ratio of remaining usable to unusable material left in the stone after flaking provides an index of how efficiently the stone has been reduced by the activity of flaking. A good knapper will removed virtually all of the unusable material very easily.

  The size of flakes removed. The ratio of the size of the largest flake scar to the maximum dimension of the core is a partial indication of how efficiently flakes are being detached from a block of stone.