The janitor at the Middleton Theater, Sam, was in his thirties and drove a Firebird with the window open in any weather. Consequently, his long black hair always looked feathered. Sam would come in for work in the afternoon, before the bars opened, wearing white pants, silk shirt, and sunglasses, disappear into the theater for a while, and then emerge without a spot. He spent most of that time searching the place for wallets and loose change, tossing out bins of popcorn as he went. When he was done, Sam would put himself down on the time sheet for about five hours of work. That’s because Sam spent most of that time chatting to people working there, meaning Alex and the manager, about all sorts of things, including the Shroud of Turin, in which he had considerable interest. Not a lot of knowledge but a lot of interest.
Whether telling stories, singing songs, or texting, humans have always liked to chat with each other at work. Traditionally, humans are viewed as belonging to a unique species because we are large-brained, bipedal, and highly social—not just a smart species that can also be social, but a species that evolved to be social. Neuroscience is confirming the essential social function of the brain: functional magnetic resonance imaging scans show that social exclusion, bereavement, and being treated unfairly can activate the pain network of the brain. Conversely, having a good reputation, being treated fairly, cooperating, giving to charity, and even schadenfreude (getting pleasure from another’s misfortune) can all activate the brain’s reward network.
Large human brains evolved to be social, but why? The evolutionary costs of a large brain are considerable. According to a twentieth-century theory called the obstetrical dilemma, for example, a large cranium increases the risk to a mother giving birth, which is part of the reason hundreds of thousands of women die each year from pregnancy or childbirth-related causes. In terms of evolution, a large risk must be justified by some comparably large benefit. The benefit that arguably most makes up for these costs is social cooperation, because groups can survive better through the internal cooperation of individuals. Where this cooperation is lacking, due to a lack of health care and/or poverty, maternal death rates are higher. Prehistorically, what distinguished humans from primates was the social context of childbirth, especially midwifery. Many causes of maternal death, such as hemorrhage, were avoided through “folk” medicine. Goddesses from the Aztecs to the ancient Egyptians were depicted as powerful women giving birth, in upright positions, while assisted by other women.
The earliest form of cooperation among our ancestors 1.5 million years ago likely was in terms of hunting for and sharing meat. More meat in the diet was essential to the evolution of the large human brain, as Leslie Aiello and Peter Wheeler contended. Because humans could get calories more easily, they evolved larger brains and smaller guts. Once our early ancestors could control fire, Richard Wrangham observed, they could cook food and afford to have a smaller gut because cooked starches are much easier to digest.
Being social also means competing socially, which requires brain power. Primates “waste” a great deal of time grooming each other because the close communication helps them get “gossip” about potential mates, according to anthropologist Robin Dunbar, thus helping them reproduce those social genes. The larger the group, the more cognitively demanding this becomes, and Dunbar argued that human language evolved in place of grooming. In the early 1990s, Aiello and Dunbar famously compared the average group size among primates to brain size—technically the size of the neocortex—and found a clear relationship. The correlation was slightly different for monkeys than for apes, but in both, the larger the neocortex, the larger their typical social group. Back then, Aiello and Dunbar’s aim was to extrapolate this curve so that they could estimate the typical group size of our hominid ancestors, such as Homo habilis and Homo erectus, using the volume of the brain that can be measured from fossil remains.
Modern humans have a brain size of about 1,400 cubic centimeters, and this is the number that Dunbar used to arrive at his estimate of 150 as being the typical limit of real—defined as meaningful—social relationships that a person will have. Dunbar’s number has proven so prophetic that a 2015 study of US teenagers showed that a typical Facebook user has 145 friends and a typical Instagram user about 150 followers. Think about this for a minute. In the early 1990s, before most of us had even heard of the Internet, an anthropologist compared the brain size of different species of apes to their typical group sizes observed in the wild. He then extrapolated the correlation out to the brain size of humans. Twenty-five years later, this extrapolation predicts the typical number of Instagram followers among US teenagers. This is pretty amazing.
Social learning has become the focus of study for behavioral scientists from a range of fields, including psychology, anthropology, and economics. Our success as a species, wrote economist Samuel Bowles, relies on the correct social and networking skills of knowing who, what, and when to copy. Whereas earlier research on conformity focused on adults, recent psychology experiments demonstrate conformity among children or even infants, who will learn from adults whom others are seen to be looking at and learning from. Other kinds of learning are biased toward natural categories, such as plants. Whereas infants will put most any plastic toy in their mouths, they will hesitate when given a plant, watching first for cues from an adult as to whether the plant is edible or poisonous, and then proceeding accordingly.
The fact that many human cultures share in the parenting of children underlies anthropologist Sarah Hrdy’s belief that sharing and cooperation, more than competition, is what makes us human. Some years ago, Alex and his family stopped at a family-owned burrito place in San Bernardino, California, and after the order was placed, the woman behind the counter asked, “Can we hold your baby?” Without much hesitation, Alex and his wife handed him over, and the woman cuddled him for several minutes while a couple of the other staff came over, and then handed him back, along with the burritos.
This simple scene would be impossible to duplicate with any other primate, as all other primate mothers would fight pretty much to the death if you tried to take their infants away. A wild ape mother will not let others hold or carry her baby, wrote Hrdy, adding that the only other primates who share infant care are marmosets and tamarins. Helping in other ways with child-rearing, however, can be seen among macaques, squirrel monkeys, meerkats, and scrub jays. This improves infant survival and later reproduction. Before all the social structures that affect human child-rearing—family structure, wealth inheritance, religion, and so on—what made humans unique, said Hrdy and others, was the sharing of food, long life of females after menopause (so grandmothers can help their daughters raise children), and fact that human infants could form attachments with multiple caregivers. All this requires a social brain, alongside emotion, empathy, and a theory of mind. In Hrdy’s terms, humans are evolved to be cooperative breeders, which means that the isolation of women in the home is not a helpful state.
The act of sharing food, such as tubers dug from the ground or the meat brought back from a hunt, is probably as old as our species or even genus. The Latin origin of companion is “one who breaks bread with another.” To solidify relationships between families, the Bantu of Zimbabwe would exchange food in a “clanship of porridge,” for example. At family dinner tables across the globe, a child in a snit can often be brought back into the conversation with an offer of more food, especially dessert.
It is no surprise, then, that sharing is what so many people do in their electronic relationships. Well ahead of the curve, anthropologists Heather Horst and Daniel Miller recorded aspects of cell phone usage in Jamaica in the early 2000s. Their research was a combination of listening, observing, and interviewing, and then collecting information about the contacts individuals had on their phones. This was back in a setting where a mobile phone was mainly a device for calling or texting friends and family. Horst and Miller found that family and kin were most important to women, who led relatively sheltered lives and, on average, had fewer than thirty numbers saved in their phones. House phones were great for long, deep, and protracted relationships, but not so well suited to maintaining connections over time with acquaintances, as Horst and Miller showed, quoting one woman who hadn’t called her male friend for a while: “Him ask if mi get rich an switch, that’s what he call mi an’ ask mi,” asking whether she was too good for him now.
Computer simulations allow scientists to consider many interactions and large numbers of agents at the same time. Back in 2010, St Andrews University’s Kevin Laland and his colleagues hosted a computer-algorithm tournament that was rooted in Robert Axelrod’s notable 1984 computer contest that pitted different repeated prisoner’s dilemma algorithms against each other. Laland’s “Social Learning Strategies Tournament” was built around proposals for the most successful default strategy for an agent in a large social environment. The contest entries consisted of software code that instructed an agent on how to interact with other agents it met over a number of rounds. The tournament creators expected the winner would have a superior social-learning strategy about whom and when to copy. Mere random copying was not seen as likely to win because information can be wrong as well as outdated.
The entries reflected the broad interest in the topic, with biologists, anthropologists, psychologists, economists, and mathematicians all jumping in. The winners, as much a surprise to themselves as to the expert panel overseeing the tournament, were two Canadian postgraduate students, Dan Cownden (a neuroscientist) and Tim Lillicrap (a mathematician), neither of whom were social-learning experts. They labeled their entry “discount machine.” Its basic instruction was to copy—and copy often—and bias that copying in favor of recent successful strategies, thus “discounting” older information. It was not quite random copying, but close: copy any success, just as long as it is a recent success.
Similar games shed light on how the success of a social-learning strategy depends on what other strategies are being played out in a group. Mike and his colleague Alex Mesoudi created an experiment where the participants played a computer game in which they “made” stone projectile points designed to hunt bison. The participants were allowed to change aspects of the stone points—for instance, the length and width—and then see how well their points would perform (based on archaeological knowledge) on bison hunts. After each round, the hunters could see their own scores (in caloric return) compared to the scores of the other hunters, and they could also see the different designs that others were using. Each hunter could invent new shapes or copy others whose hunting success scores they could see. In all runs of the game, social learners scored better than those who refused to copy the success of others.
But it really is a bit more complicated because people tend to work in groups, and if everyone always copies someone else—meaning there is no innovation—then the group runs the risk of extinction. Mike and Alex were able to see what the proportion was of copying and inventing within the entire group of projectile point makers. They then charted the group’s success as a function of this ratio and found that there seemed to be an optimal mix of new information created by the minority “producers” and the majority of people who copied them—“scroungers.” We would expect that over time, or in a well-governed community, a nice mix would settle out, with some producers and a majority of discerning scroungers.
What is a “nice mix,” though? It turns out that a number of studies have shown that somewhere around 5 percent producers is ideal. The Max Planck Institute’s Ian Couzin and his colleagues, for example, showed that within a flock of birds, it takes only a small fraction of individual learners (producers) to impart a coherent direction to the entire group, as the majority copies the travel direction of its neighbors. Individual thinking and accurate information among these few producers are critical, because false alarms can spread and amplify across a flock through bad information. Couzin and his colleagues demonstrated how ignorance or ambivalence among the majority allows the consensus to be controlled by a small but determined minority. There is a rapid transition from the ambivalent majority rule to the determined minority rule as either the minority is made more determined or ignorance and/or ambivalence among the majority is increased.
Scroungers can proliferate when there isn’t a lot of change in the environment, especially when it is costly or risky to produce new ideas, but having too many scroungers compared to producers leaves you with a noisy echo chamber. As Mesoudi pointed out, when copiers become too numerous, they will be copying from each other so much that the quality of their information deteriorates in the echo chamber. We can think of Internet availability as raising the number of information producers not to 5 percent of the community but rather to 5 percent of the world. The obsession with novelty and individualism is not only a disadvantage; it is WEIRD, as in the acronym for “Western, educated, industrialized, rich, and democratic” societies, coined by Joe Henrich and his colleagues. When Mesoudi and his colleagues took the projectile point experiment to a non-WEIRD community in a provincial Chinese city, they found that the inhabitants tended to copy more than did the British nationals, Chinese immigrants in Britain, or residents of Hong Kong. As a result, the Chinese scored better with their projectile points, both in groups and individually. Westerners tended to persist in their individual learning, and their scores suffered for it. Differences in cultural individualism may take centuries to develop. The history of rice farming in southern China, for instance, underlies a more interdependent and holistic culture compared to the wheat-growing populations of northern China, according to a psychological study led by Thomas Talhelm of the University of Chicago.
Evolutionary anthropologists have argued that we are wrapped in culture, or addicted to culture, in that we inherit our instructions for how to cope in the world from past generations. Under this cultural-intelligence hypothesis, the cumulative cultural toolbox contains a different set of skills than just individual intelligence and problem solving. Humans are evolved not to solve problems on their own, or even solve them cooperatively on the fly, but rather to accumulate knowledge collectively, over generations, and apply that knowledge to the same environment (roughly) in which previous generations lived. The environment, however, has two components—cultural as well as physical—and as we will see throughout the book, humans are now living in cultural environments that are radically different from those of their ancestors—even recent ancestors. The fast pace of cultural change is having enormous impacts on human evolution.
Ironically, no matter the pace at which it changes, culture still provides the basis for humans to exist. It has to. We are so far into the game that without culture—literally, the high capacity for cooperation and learning—humans would be lost. Consider hunter-gatherers, who live in small, highly mobile bands and exploit wild resources over a wide area. Although both chimpanzees and hunter-gatherers share food—for example, meat—only humans exchange nonfood items as gifts to maintain social relationships between groups. Because their population density is low, this exchange has benefits. In the Kalahari Desert, a Ju/’hoansi band could use the water holes of their exchange partners when theirs dried up in a drought. Until the mid-twentieth century, the Ache hunter-gatherers of eastern Paraguay hunted mammals with bows and arrows and foraged for plant foods. A regional group of Ache was something over 500 people living in twenty or so residential bands at least ten kilometers apart.
In Tanzania, the eastern Hadza still hunt and gather with bows, axes, and digging sticks. About a thousand of them live in some fifty bands, some less than a kilometer apart, and others as much as eighty kilometers apart. Anthropologist Kim Hill and his colleagues found that the average Ache or Hadza man will have interacted with about 300 to 400 different men in his lifetime. When you roughly triple this, to account for opposite-sex adults and children, the number of interactions is much higher than for chimpanzees, which is about 20, and much higher than Dunbar’s number of about 150. Remember, though, that Dunbar was referring to meaningful relationships, not the one-off kind.
We see this with the Hadza. When a group led by Coren Apicella asked more than 200 women and men, from seventeen distinct camps, to whom they would give a gift of honey, they found the average number of gift partners was about six. Despite the apparent messiness of the network, which makes it appear as if everyone is connecting with everyone else, they’re really not. Despite everyone knowing everyone else, meaningful relationships involve a much smaller number of people.
Now compare this to a Christmas card study that Dunbar did with his student Russell Hill in England in the days before social media, which found the average number of cards sent was close to 150—the Dunbar number. Not surprisingly, Hill and Dunbar also discovered that self-rated emotional closeness decreased between pairs of friends as the frequency of contact decreased. So we may say “keep your friends close but your enemies closer,” yet in reality we want to keep our best friends closest.
But do we know who those people are? Can we reliably identify our “close friends”? Hill and Dunbar’s study is interesting, but it was built on the self-rated closeness someone feels for someone else. In other words, you identify your close friends and then rank them in terms of closeness. A recent study by Alex Pentland and his group at MIT’s Media Lab, however, showed how people are not particularly good at judging their asymmetrical friendships. Like King Richard II of England and his cousin Henry IV (who imprisoned him in 1399), each would clearly rank their friendship differently. This lack of reciprocity could lead to hurt feelings—you suddenly find out that you feel closer to a person than that person does to you—but more than that, it limits a person’s ability to engage in mutually beneficial arrangements. To help sort this out, we have names.
Maybe their importance is diminished in today’s Western world, but names have traditionally played an important role in society and still do today in much of the world. Names can inform us about our inherited biological and social relationships, which were crucial for a social species that became organized by kinship systems that controlled not just marriages but also alliances, food sharing, wealth inheritance, and the specialized jobs and crafts that people engaged in that contributed to the overall success of a group. On a sparsely populated landscape, where exchange with kin is essential for survival, knowing how to interact socially with those with whom you have only occasional contact is crucial. This starts with names and how to address someone.
Hunter-gatherer societies tend to be small, but their kinship-naming systems can be complex and highly informative. In the early 1950s, Lorna Marshall of Harvard’s Peabody Museum traveled to southwestern Africa’s Kalahari Desert to study the !Kung Bushmen (now Ju/’hoansi). During a fourteen-month stay, Marshall was able to record a kinship network of about six hundred people. They did not know the proximate reasons for their kinship system looking the way it did, but they did have a keen sense of cultural transmission. The !Kung would say, “God created people and told them what terms to use for each other. Parents have taught their children ever since what terms to use.” Marshall then asked people what kinship terms they used for each other, and while she found the terms were straightforward for parents, children, and siblings, they were less so for grandparents or grandchildren, where one of two terms might be used—one for biological relatedness and the other as a way of classifying individuals with the same names.
The kin terms were intricate and precise. Terminologically, no two people were simultaneously interrelated in more than one way, even though their names may have changed on marriage. The names conveyed information about the relative biological age—nothing surprising here, as our terms “mother” and “daughter” do the same thing. There were five parent-child terms, which were never modified or used for any other purpose, and three sibling terms for elder brother, elder sister, and younger brother or sister that were never modified but occasionally were used for another purpose. On top of this, there were generational terms—one set applied to males, and another to females.
Sound confusing? In fact, individuals had their own ancestral lists of names to learn, taught to them by their parents. The !Kung did not know exactly how the list was compiled, but the parents knew exactly what the list should be for each of their children. A long individual memory was not required, however, as many of the !Kung did not know the names of their great-grandparents. This is an example of cultural inheritance doing its job: keeping things simple yet elegant. Speaking of culture as being simple yet elegant, let’s turn the page and see what cultural inheritance was like on another continent—Europe—and at another time—the eleventh century.