COMPARATIVE PSYCHOLOGISTS HAVE IDENTIFIED DIFFERENT types of social learning that may be involved in behavioral traditions. On the perhaps most primitive level there is contagion. As we all know, yawning is contagious. Seeing someone yawn substantially increases the likelihood that you will yawn as well. Such contagious behavior occurs in other primates, too. Yawning might have evolved as a means of a tired individual influencing the group to stop moving and rest. On a slightly more complex level, any interaction with an object can draw others’ attention. The effect is regularly observable when you have more than one child in the house. As one of them starts to play with a toy, that toy can suddenly acquire a new degree of attraction for the other child, no matter how long it has been ignored before. Animals that see another animal intently exploring something can, by the same token, also become interested in examining that object. Even an octopus that sees another octopus selecting one of two objects is afterwards more likely to pick the same object. It makes evolutionary sense to attend to what others find rewarding—after all, there might be a reward for you, too.
Learning by copying others is a more complex form of social learning.14 There has been considerable debate about what qualifies as imitation learning among comparative psychologists. To rule out previous associative learning, some have advocated considering only copying of novel behaviors as evidence. Others are happy to refer to imitation whenever any behavior is copied. Like human newborns, chimpanzee and macaque infants are more likely to poke their tongue out when a human faces them and pokes the tongue out than in control conditions. One explanation for this behavior is commonly assumed to be a functioning mirror neuron system, which, as discussed earlier, links what is seen with one’s own action. In fact, the mirror neuron system was first discovered in monkeys. Yet, surprisingly, there is little evidence to suggest that monkeys imitate in other ways, in spite of their reputation. Monkey see; monkey don’t do.
Yet other animals do—at least in some contexts. Imitation of sound is widespread in the animal kingdom. Many songbirds imitate the songs of their parents, and thus song traditions are maintained. Imitation of sound may be substantially easier than imitation of action. In copying sound, one needs to match one’s own sounds to those heard, whereas one’s own and observed actions look different from each other. Observation of another’s, say, dance move, cannot be directly compared to doing it yourself. To imitate actions you may need to mentally assume the perspective of the model. Nonetheless, sound imitation can be quite complex. The most impressive sound imitators I know are the lyrebirds in Queensland. Male lyrebirds sing and dance to attract females. They can copy the sounds of other birds and animals but are also proficient at imitating the sounds of chainsaws, didgeridoos, camera shutters, and even the opening of beer cans.
Some mammals copy elaborate songs to attract the opposite sex. Male humpback whales are renowned for their songs. All males in a population sing the same song, although it changes over time. My colleague Mike Noad observed how the song sung on the east coast of Australia underwent a radical transformation. In 1996, two of eighty-two whales were recorded singing an utterly different song from the other whales. Their song was typical of another population of whales migrating up the western coast of Australia—as if this pair had taken the wrong turn on their way back from the Antarctic. The next year, some 40 percent of males adopted the new song. It was a hit. On the subsequent southward migration, virtually all males sang the new song. This rapid spread of the new song must have depended on imitation. Such transmission of songs between unrelated individuals has since been documented in other humpback populations.
Humpbacks may copy other actions. Off the coast of Brazil they have been observed displaying a strange behavior: prolonged floating vertically in the water with tail in the air. These observations have increased over time, suggesting possible social transmission. In captivity, some cetaceans have indeed demonstrated a capacity for bodily imitation. The comparative psychologist Louis Herman has shown that dolphins can imitate on command and even copy the behavior of human models. When the human turns and flaps his arms, the dolphin must somehow map its different body plan onto that of the model. These marine mammals are the exception, not the rule. There is little evidence that other mammals so readily imitate.
The other notable exceptions are the great apes. Orangutans at a reintroduction center, for instance, have been observed copying human activities such as hanging a hammock, applying insect repellent, and sweeping with a broom. One, you might recall, was even observed trying to imitate the lighting of fire. Great apes, moreover, appear to recognize when a human copies them. We first found evidence for this by asking Emma Collier-Baker to mirror everything the chimpanzee Cassie did. He would repeat behaviors more frequently in this than in various control conditions, and he even appeared to test her resolve with unusual sequences of actions. These responses to copying have since been documented in other great apes. Furthermore, the ape language pioneers Keith and Katherine Hayes taught the chimpanzee Viki to copy whatever they did on the command “do this.” Viki understood the idea sufficiently to later imitate entirely novel behaviors. This “do as I do” paradigm is perhaps the most direct way to assess understanding of imitation, and great apes have been shown to do well. Similar attempts with monkeys have so far remained fruitless.15
Do great apes use imitation to learn new tricks from one another? Andrew Whiten and colleagues have conducted numerous studies on this question. For instance, they presented chimpanzees with a puzzle box—an artificial fruit with a treat inside—that could be opened in two ways. Participants in one group saw a human demonstrator pull and twist out bolts, turn and remove a pin, and then turn a handle to get at the treat. The other group saw a model poke out the bolts with a finger, twist and remove a pin, and then pull the handle up to release the lid. Chimpanzees opened the box primarily in the way they were shown, suggesting that they acquired their knowledge from the social model. However, in other experiments chimpanzees did not learn socially, acted inconsistently, or demonstrated social learning other than imitation. The emerging picture suggests that chimpanzees can but often do not imitate.
FIGURE 8.1.
The chimpanzee Cassie may recognize it when his postures are being copied.
Not surprisingly, there were debates about what all this might mean. In a seminal study Victoria Horner and Andrew Whiten found data that go a long way toward an explanation. They used a version of the artificial fruit task and found a profound difference between the behavior of chimpanzees and that of human children. A model demonstrated first poking a stick into a hole at the top part of an artificial fruit and then inserting it in another hole further below. Both chimpanzees and children copied both actions of the model to retrieve the reward. In a second condition, the apparatus was made of transparent material. It became patently obvious that the first action, poking into the top, had no causal role in the opening of the box. You could simply poke the stick into the lower hole to obtain the reward, and this is what the chimpanzees subsequently did. They stopped imitating the first action and instead immediately proceeded with the second. Three- and four-year-old children, on the other hand, continued to copy the superfluous first action before putting the stick into the lower hole. Human children tend to “overimitate,” whereas chimpanzees imitated only to the extent required to achieve their immediate goal.
FIGURE 8.2.
The young male orangutan Putu trying to open one of Andrew Whiten’s puzzle boxes (photo courtesy of Mark Nielsen). Like chimpanzees, orangutans showed no signs of overimitation.
Though the chimpanzees in this experiment acted more efficiently and arguably more rationally, it is the children’s overimitation that is crucial for faithful transmission of cultural knowledge and for accumulation.16 As we saw earlier, by age two children tend to overimitate even when there is a quicker way to the goal. Mark Nielsen recently examined this in children in the Kalahari Desert and found the same patterns as that described in Eu
ropean and Australian children. Great apes, on the other hand, do not seem to overimitate (see Figure 8.2). Perhaps faithful copying is a key difference between the cultural transmission of humans and our close relatives. It enabled our cultures to gradually ratchet up the repertoire of our second inheritance system.
THERE ALSO APPEAR TO BE differences in teaching. Mammals may provide opportunities for their young to learn in a safe environment. Adults encourage or discourage certain behaviors. But do they structure learning events in light of their assessment of the pupils’ knowledge? There is no obvious evidence of animals having anything like a curriculum. Without sophisticated cognitive capacities such as foresight and theory of mind, flexible, tailored teaching would seem to be impossible. It was therefore long thought that other animals do not teach, period. However, it has become clear recently that animals can at least act in ways that function like teaching—even if the teacher might not understand it as such.
A simple definition of teaching is that a teacher modifies his or her behavior in the presence of a naïve pupil without immediate benefit, and that this behavior fosters learning in the pupil. Even using this definition there is only limited evidence of teaching in the animal kingdom. Cats can be observed throwing half-dead prey, such as mice, in front of their young in apparent attempts to get the offspring to learn a fundamental hunting skill. Similar behavior has been observed in some other carnivores. The most compelling evidence comes from meerkats. Young meerkats learn to handle dangerous foods such as scorpions within the first three months of life. Adults gradually introduce them to the prey by bringing first killed and then disabled prey to begging pups. Adults remove the sting of scorpions and so make it safe for the young to interact. As they get older, pups are increasingly given live and intact prey. This adult behavior clearly aids learning. It may justifiably be called “teaching” because the adults change their behavior in the presence of the learner, the behavior brings no immediate benefit to the teacher, and the pups can learn in ways they otherwise could not.
Yet it turns out the adult meerkat behavior is not based on assessing the skill level of the pups. Playback studies have shown what really drives the behavior is the pups’ calls. If the begging calls of old pups are played to adults, they bring live prey even if their real pups are far too young to handle them. Conversely, playing back the begging calls of very young pups to a group with old pups increases the number of dead prey brought back. Although this behavior functions like teaching, it may have little in common with the mechanisms underlying flexible teaching in humans.
With more research I suspect that we will find functional teaching in a range of other species. Some killer whales use the dangerous technique of stranding themselves to hunt elephant seals. Adult whales have been observed pushing juveniles forward in their early attacks and, importantly, helping them get back off the beach. Functional teaching may be found in unexpected places. Indeed, perhaps the second most compelling case for teaching in the animal kingdom comes from ants. Ants that know the way to a food source guide others. They can run “in tandem,” thereby slowing down progress for the leader but ensuring that the follower learns the route. Nonetheless, there are only a handful of such cases of functional teaching in the literature at present, and in each the teaching is limited to one type. They do not show the flexible, customized teaching of virtually indefinite content that is evident in humans.
Curiously, there is as yet little evidence of functional teaching in our closest relatives. As already discussed, they do not seem to declaratively point things out to each other in the wild. Indeed, there are few signs that the traditions of chimpanzees or orangutans are maintained through instruction. Where chimpanzees crack nuts with hammers and anvils, it takes the young several years to learn the process. The mothers do not typically guide the young or help them along. They usually allow the infant to observe closely and, at times, to indulge in some of the morsels of goodness that await the successful nutcracker. Active instruction could transmit this information far more quickly and effectively.
The primatologist Christophe Boesch, after years of observation, documented two instances in which the chimpanzee mother acted in a way suggestive of active teaching. One involved a five-year-old who failed to crack a nut after several minutes of trying. The mother, who had been resting, joined her, took the hammer, and slowly rotated it into a better position. She then cracked ten nuts, of which six were eaten by the offspring, before returning to rest. The daughter, adopting the better hammer grip, managed to open four nuts over the next fifteen minutes. The other instance involved a mother apparently helping her infant by putting a nut in a better position on an anvil. These behaviors look like teaching, although a skeptic might account for them in leaner ways. What is striking, though, is that this is so unusual. After decades of systematic observation, these are the only two instances of proposed ape teaching in the literature. Regardless of whether one takes a rich or lean view on the evidence, it is clear that chimpanzees do not commonly instruct each other.
Human teaching relies heavily on language, theory of mind, and mental time travel, so animal teaching may be limited for all the reasons already discussed. Just as great apes are not particularly motivated to share intentionality and experiences with others, their capacity and motivation to teach appears to be profoundly limited. In a recent study chimpanzees, capuchin monkeys, and human children were presented with a sequence of puzzle boxes. The children taught each other through language and gestures, imitated each other, and shared rewards. Children who received such social support did better on the tasks. By contrast, the chimpanzees and monkeys acted to secure rewards for themselves and showed no signs of teaching others.
Great apes have some capacity for cooperation and social learning through which they maintain behavioral traditions. However, without sufficiently sturdy pillars of imitation and teaching, neither the knowledgeable nor the ignorant are sufficiently equipped for a ratchet effect to build up an ever-increasing, cumulative cultural heritage.
Humans have a strong urge to link their minds, to overimitate, and to teach. In this way our inventions, skills, and knowledge spread and are adjusted to local conditions, fine-tuned, or optimized by others. Human social groups cooperate to accumulate cultural capital over many generations to an extent that has seriously altered the fitness of its members. As I will discuss further in the next chapter, we have developed moral norms and means of indirect reciprocity that critically facilitate reliable cooperation. With the right cultural knowledge we can live in the desert or in the Arctic. We can even leave the Earth and briefly live in space. Our second inheritance system has allowed us to gain powers vastly beyond what any previous organisms have wielded. This has confronted us with novel challenges because with great powers come great responsibilities (to paraphrase Voltaire—or Spiderman’s Uncle Ben). In spite of our many failings, most of us try to do the right thing. To the eternal struggle between right and wrong, between good and evil, we therefore turn next.
1This may be the case, regardless of what one might think about “linguistic relativity”—the idea that language determines thought. The popular example of Inuit having many different words for snow has been debunked. Nonetheless, it is clear that experts in any discipline have more words and hence conceptual categories for the things they study. Where I see some fish, my spouse, Chris, who is a marine biologist, sees dozens of different species and complex ecosystems.
2The actual genetic similarity is often significantly higher. When there is little migration, the baseline average relationship of group members can be quite high, and so the similarity of close family members must be even higher.
3Hamilton’s rule is rB > C. Individuals may act altruistically if the fitness cost to the individual (C) is less than the benefit (B) to the recipient times the degree of relationship (r) between the individual and the recipient.
4One glaring exception, at first sight, seems to be the fact that the highest number of child abuse a
nd murder cases comes from within the family. However, the evolutionary psychologists Martin Daly and Margo Wilson showed that the perpetrators of these crimes are typically not blood relatives. Stepparents are much more likely to harm the children of their spouses than biological parents are, in what has become known as the Cinderella effect. Of course, such abuse is rare, but among the parents who abuse their children, stepparents are over-represented. For instance, in Canada between 1974 and 1990, less than 3 fathers per million fatally beat their child, whereas 321 per million stepfathers did.
5The ethologist Niko Tinbergen clarified that we need to distinguish the proximate explanations for behavior, the mechanisms involved, and how they develop from the ultimate explanations for behavior, how the behavior evolved, and what its functions are. The proximate explanations of feeling empathy and having been raised to “do the right thing” may serve ultimately more selfish functions.
6The evolutionary psychologist Leda Cosmides has even argued that we evolved innate cheater-detection mechanisms. The evidence for this, however, is controversial.
7Isaac Newton famously declared, “If I have seen a little further it is by standing on the shoulders of Giants.”
8Like biological evolution, cultural evolution clearly involves variation and differential replication. But this need not mean that there are other parallels, such as cultural versions of genotype, phenotype, RNA, sexual selection, and so forth. Whether the concept of memes (the “meta-meme,” if you like) is helpful in the study of culture is unclear. Some prominent scholars find the very idea totally misguided.
9Once the island of Tasmania was physically separated from mainland Australia some ten thousand years ago, the Aboriginal Tasmanians could not share in subsequent mainland inventions such as boomerangs and in fact even lost technologies they previously possessed, such as bone tools. This is not to say that their cultural heritage was not exceedingly rich and successful in the environment that they lived in—at least until a radical change occurred with the arrival of Europeans.
The Gap Page 22