By Sarah Blaffer Hrdy
（In a silvered langur group in New York City’s Bronx Zoo, a baby’s golden hair—a distinctive juvenile trait—attracts attention from females in the group, which will take the baby from its mother and carry and groom it until the mother retrieves it.
Photo by Noel Rowe/The Pictorial Guide to the Living Primates）
We cram our bodies into the plane’s narrow seats, elbow-to-elbow, makingeye contact with nods and resigned smiles as we yield to latecomerspushing past. Most ignore the crying baby, or pretend to. A few of useven signal the mother with a sideways nod and a wry smile. We want herto know that we know how she feels, and that the disturbance she thinksher baby is causing is not nearly as annoying as she imagines—eventhough we can tell (as can she) that the young man beside her, eyesdeterminedly glued to the screen of his laptop, does indeed mind everybit as much as she fears.
Thus does every frequent flier employ our species’ peculiarlyempathetic aptitude for intuiting the mental states and intentions ofother people. Cognitive scientists and philosophers have long calledthis awareness of others’ inner life “theory of mind,” but manypsychologists now refer to it as “intersubjectivity,” a broader conceptthat roots our sophisticated skill at mind reading in the capacity toshare in the emotional states and experiences of others. Whatever wecall it, this ability to divine and care about the mental experiencesof others makes humans more adept at cooperating than other apes are.
Imagine what would happen if one were traveling with a planeload ofchimpanzees. We would be lucky to disembark with all our fingers,testicles, and toes attached, and with the baby still breathing andunmaimed. But human passengers fill some 2 billion airline seats everyyear and submit to being compressed and manhandled, with nodismemberments reported yet! Along with our 1,350-cubic-centimeterbrains and capacity for language, such unusually well-developedimpulses to cooperate have helped propel our success as a species. Butwhy did humans become such “other-regarding” apes?
（A female Francois’s langur nurtures what look like twins, but only one ofthe infants (which are ninety-one and seventy-four days old) is hers.In this species, females other than the mother (allomothers) routinelycare for, carry, and may even suckle any infants in their group.Normally, the infant’s coat changes into full black within 130 to 140days, but recent data suggests that the first-born of young femalestake a bit longer to mature and still have golden fur on the head at150 days old.
Photo by Gang Hu）
Although the genus Homo arose before the beginning of the Pleistocene epoch (1.8 million to 12,000 years ago), H. sapiens—anatomically modern humans with upright bodies and big brains—evolved only within the last 200,000 years. And behaviorallymodern humans, capable of symbolic thought and language, emerged morerecently still, within the last 80,000 years. Most evolutionists haveassumed that our unusually sophisticated capacities for attributingmental states and feelings to others coincided with thoselate-Pleistocene behavioral transformations, and corresponded with theneed for members of one group to get along so as to outcompete anddefend themselves against other groups.
Butthere are difficulties with that scenario. There is abundantarchaeological evidence for early warfare, but none dates back muchbefore 12,000 years ago, when people began to settle down and live inmore complex societies with property to protect. Moreover, geneticevidence suggests that our foraging ancestors in the Pleistocene livedat low densities. Although individuals no doubt fought and sometimeskilled one another, there is no evidence that whole groups fought. Moreto the point, if the drive to outcompete members of opposing groups wasthe source of our hypersocial tendencies, why didn’t selection favoreven greater and more Machiavellian intelligence, better mind reading,and better capacities to cooperate against hostile neighbors among theancestors of today’s chimpanzees? Chimpanzees are competitive,dominance-oriented, aggressive, and reflexively xenophobic: wouldn’tthey have benefited just as much, or more, from being able to cooperateto wipe out competing groups?Consider, however, an alternative explanation, the possibility thatour empathetic impulses grew out of the peculiar way that children inthe genus Homo were reared. I believe that at an early stage in humanevolution, our bipedal ape ancestors were increasingly cared for andprovisioned not just by parents but also by other group members, knownas alloparents.
In my view, cooperative breeding (as sociobiologists term thereproductive strategy in which alloparents help both care for andprovision young) came before big brains. I believe it first emergedamong upright apes that were only beginning to look like us, andfurther evolved during the Pleistocene in African H. erectus(also called H. ergaster)—creatures that did not think or use languageto communicate the way we do. Alloparental care and provisioning setthe stage for children to grow up slowly and remain dependent on othersfor many years, paving the way for the evolution of anatomically modernpeople with even bigger brains. It was not the other way around: biggerbrains required care more than caring required big brains.
Comparisons across cooperatively breeding speciesshow how nonessential a sapient mentality is for shared care, andprovide our best hope for understanding what selection pressures induceindividuals to help rear someone else’s young. Insights from suchcomparisons help explain why mothers among highly social apes living inAfrica about 1.8 million years ago might have begun to abandonmother-only care, setting our ancestors on the road to emotionalmodernity.
Although at first caring for and provisioning someone else’soffspring seems to defy evolutionary logic, cooperative breeding hasevolved many times in a taxonomically diverse array of arthropod,avian, and mammalian species. It occurs in 9 percent of the 10,000living species of birds and in perhaps 3 percent of mammals. Theadvantages for parents are well documented, with significantdemographic consequences.
（Photographed near Mount Abu in India, a female Hanuman langur mother allows herinfant to be taken and temporarily cared for by another adult female, aclose relative. Such shared care keeps the infant safe while the motheris free to forage. The practice, widespread in primates, tends toevolve in species where females remain in their natal groups close tomatrilineal kin and where, because of their feeding ecology, the femaledominance hierarchy tends to be relaxed and relatively flexible. (Formore about infant-sharing in langurs, see The Langurs of Abu: Female and Male Strategies of Reproduction, by Sarah Blaffer Hrdy (Harvard University Press, 1977.)
Photo by Sarah Hrdy）
Mothers able to confidently entrust helpless offspring togroupmates’ care conserve energy, stay better nourished, and remainsafer from predators and other hazards, leading longer lives withgreater reproductive success. Because mammal mothers that have aid alsowean babies sooner, many reproduce again sooner, and so give birth to agreater number of young over their lifetimes. More important, the extrahelp ensures the young have a better chance of survival. Certainspecies therefore spread successfully thanks to cooperative breedingand, with it, a faster pace of reproduction and the flexibilitypermitting young to survive in a wide range of habitats.
But how could natural selection ever favor caring for someone else’syoung? Why would young magpie jays in Costa Rica, ones that have neverreproduced, bring back beakful after beakful of food to beggingfledglings? Those allomothers often provide more food than the chicks’own parents do. Ornithologists J. David Ligon of the University of NewMexico and D. Brent Burt of Stephen F. Austin State University in Texaspropose a two-step process for such development. Start with a specieswith particularly helpless, slow-maturing young, in which selectionwill favor high sensitivity to the cues emitted by needy babies as aparental trait. Then add some special benefit that encourages maturingindividuals to linger in their natal place, such as defensible andheritable resources. As a result, group members will be exposed tosensory cues from chicks (or pups) and will be primed to respond. This“misplaced parental care” hypothesis helps explain why cooperativebreeding is three times more likely to evolve in taxa that producealtricial (helpless) young rather than precocial young (those that aresoon able to survive on their own).
Not all such caretaking is as self-sacrificing as it may appear.Often, alloparents only babysit when no more self-serving option isavailable. They may proffer food only when they do not actually need itthemselves. They may volunteer only when they have energy to spare, orwhen they are still too young or lack the opportunity to reproducethemselves. Or if two cohabiting mothers are reproducing, as occursamong lions, ruffed lemurs, bush babies, and some mice, they may taketurns as alloparents. One mother may suckle the other’s offspring whilethe other mother is “at work” foraging. And where practice is criticalfor learning how to parent, as is the case for many primates,babysitters derive valuable experience by first caring for another’syoung.
In other cases, however, helping is more of a one-way street—and byno means entirely voluntary. Subordinate meerkat, wild dog, and wolffemales that have never conceived (and may never do so) sometimesundergo a “pseudopregnancy,” developing a swollen belly and mammaryglands. Then, once the alpha female’s pups are born, the nonmotherssecrete milk for the alpha’s pups. By becoming a wet-nurse,a subordinate may increase her chances of being tolerated in the group.Had she given birth herself, her young might have been killed by thealpha female.
（Perhaps after mating and undergoing a pseudopregnancy, a Jack Russell terrierbelonging to the author’s neighbor chased away a mother cat and nursedthe litter herself. Babies can be magnetically attractive to others,even members of a different species. It has been hypothesized thatunder some circumstances such misplaced parental care could lead to theevolution of cooperative breeding.
Photo by Sarah Hrdy）
Of course, it makes good evolutionary sense for individuals toenhance the reproductive success of relatives with whom they sharegenes. But helpers are not always kin, and even kin can be less thankind: some meerkat and marmoset alphas eliminate their own daughters’offspring—the grandmothers from hell.
In roughly half the 300-odd species of livingprimates, including all four great apes and many of the best-knownspecies of Old World monkeys, such as rhesus macaques and savannababoons, mothers alone care for their infants. A chimpanzee, gorilla,or orangutan mother will be literally “in touch” with her infant foralmost every moment during its first six months of life, and theorangutan nurses her baby for up to seven years. Such continuousmaternal care cannot be attributed to lack of interest from would-bebabysitters, however. In all primates, babies are a source ofattraction, most often to subadult females. The mother’s possessivenessis the determining factor. A wild ape mother is adamant that otherswill not hold or carry her baby.
Elsewhere in the primate order, mothers are more tolerant ofallomaternal overtures. Shared care with at least minimal provisioning(often no more than one female allowing another female’s infant tobriefly nurse) is found in some 20 percent of primate genera. But onlyamong marmosets and tamarins, members of the family Callitrichidae, dowe find shared infant care combined with extensive alloparentalprovisioning, such as we also see in humans. In that respect, thosetiny-brained South American monkeys, which last shared a commonancestor with humans more than 35 million years ago, may provide moreinsights into the early evolution of human family life than do moreclosely related species such as chimpanzees.
Marmoset and tamarin mothers tend to produce twins (togetherweighing up to 20 percent of the mother’s body weight) as often astwice a year. But the social arrangements lighten the load. Usually,only the group’s most dominant female breeds, although groups with twobreeding females sometimes occur. Fathers and alloparents of both sexesare unusually eager to help mothers rear their young. Babies arecarried throughout most of the day by one or more adult males, whichexpend so much energy doing so that they actually lose weight. Otherhelpers, typically but not exclusively kin, voluntarily deliver evenprized animal prey to youngsters.
Group members are also unusually tolerant of one another duringforaging. Observing moustached tamarins in the wild, University ofIllinois primatologist Paul A. Garber recorded only one aggressive actfor every fifty-two cooperative ones he saw, such as collaborating tognaw open hard fruits. When tested in the lab, cotton-top tamarinsstudied by psychologist Marc D. Hauser’s team at Harvard, and marmosetsstudied by evolutionary anthropologists Judith M. Burkart and Carel P.van Schaik at the University of Zurich, turn out to be unusuallyattentive to the needs of others. They are far more willing to deliverfood to individuals (including nonrelatives) in an adjacent cage thanare chimpanzees in comparable experiments. Marmosets go out of theirway to provide food to others, and tamarins even keep track of andreciprocate generosity. Burkart argues that the combined mutualtolerance and spontaneous generosity of cooperative breeders areconducive to social learning, in particular to the ability ofyoungsters to glean information from and about their caretakers.
In every human hunting-and-gathering societyabout which we have information, mothers allow others to hold newborns.But how could selfish apes ever make the transition from mother-onlycare to such cooperative breeding? At some point in the emergence ofthe genus Homo, mothers must have become more relaxed about handingeven quite young infants over to others to temporarily hold and carry.No infant is more costly than a human one, and a growing body ofevidence from traditional societies makes clear that wherever rates ofchild mortality were high, children with alloparental provisioning weremore likely to survive. I believe that was the case among our ancestorsin the Pleistocene.
（Pen and ink drawing of golden lion tamarins by the artist Sally Landrybeautifully illustrates the patterns of child-rearing found amongmarmosets and tamarins belonging to the subfamily Callitrichidae.Typically, a single alpha female in the group will give birth to twinsthat are then carried most of the day by one or more of the adult malesthat the mother has mated with (they are passed back to the mother whenthey need to suckle). Other group members, like the subadult male inthe foreground, catch beetles, frogs and other small prey to feed theinfants around the time they are weaned, a time when youngsters areespecially vulnerable to malnutrition.
Illustration by Sally Landry）
Among ethnographically recorded hunter-gatherers, provisioning byallomothers starts early and goes on for years, beginning with“kiss-feeding” of unweaned infants with saliva sweetened by honey orwith premasticated mouthfuls of other food. That encourages infants topay attention to others, including their own mothers, with whom theyare eager to maintain visual and vocal contact. An infant temporarilyout of its mother’s arms will spend more time monitoring herwhereabouts and looking at her face. Youngsters also have a bigincentive to learn who else might be available and willing to care, andchildren with several trusted attachment figures learn to integratemultiple perspectives. In the words of pioneering child psychologistsTed Ruffman of the University of Otago and Josef Perner of theUniversity of Salzberg, “theory of mind is contagious”—you catch itfrom older siblings and other caretakers.
Among our Pleistocene ancestors, infants with multiple caretakerswould have been challenged in ways that no ape had ever been before.The needy youngster would have had to decipher not only its mother’scommitment but also the moods and intentions of others who might beseduced into helping. How best to attract care in varied circumstances?Through crying? With smiles, funny faces, gurgling, or babbling? Theyoungster best at mind reading would be best cared for and best fed.Such novel (for an ape) selection pressures favored a very differenttype of ape—one that we might call emotionally modern.
（Artist’s rendering of Homo erectus,ancestral humans who lived in Africa in small groups as early as 1.8million years ago: The woman in the center is cracking open nuts, whileparents and other group members (so-called alloparents) help take careof infants. Shared care coevolved with empathetic awareness; mothersand infants benefited from intuiting who would help and who would hurt.
Illustration by Patricia J. Wynne (www.patriciawynne.com)）
Almost all primates live in social groups, and itis generally advantageous for a mother to be in a group that includesclose kin. Their help is especially critical when an inexperiencedyoung female first gives birth. In most social mammals, and in themajority of monkeys, females remain with the group where they are born,and maturing males strike out to make their fortunes. But among ournearest living relatives, the great apes, only a tiny minority of newmother apes ever have matrilineal kin nearby. Evolutionary biologistshave taken for granted that, like other apes, our female ancestors musthave left their natal groups to breed in another community. There theywould have encountered unrelated females, possibly competing mothers,who might be not only unsupportive but actually infanticidal.
Until recently, in fact, evolutionary biologists assumedhunter-gatherers followed a similar pattern of female dispersal. But in2004, in an exhaustive review of ethnographic studies, University ofUtah anthropologist Helen Alvarez concluded that mothers living inhunting-and-gathering groups were likely to have their mothers andother kin nearby when they gave birth.
For example, Stanford University anthropologists Brooke A. Scelzaand Rebecca Bliege Bird found that among the traditionally polygamousMardu hunter-gatherers of Australia’s Western Desert, older motherswould relocate to be near daughters of childbearing age, especially ifthe daughter lacked an older cowife to advise and help her. Motherswere also eager to join a daughter if she was married to the same manas her sister. In consequence, half of married Mardu women between theages of fourteen and forty had a mother in the same group, while manyhad sisters or cousins as well, often as cowives. On average, femalegroup members had an 11 percent chance of sharing a gene by commondescent—just as do females of some of the nonhuman primate species thatpractice infant-sharing.
Something happened in the line leading to H. sapiens that encouraged female relatives to stick together. The impetus, I believe, had to do with food.
（Chimpanzee mother with her baby in Gombe National Park, Tanzania: althoughconsidered the closest living relatives of humans, in the Great Ape(chimpanzees, gorillas, orangutans) mothers are highly possessive oftheir infants and do not allow other group members to share in theircare.
Photo © Irven DeVore/Anthro-Photo）
By 1.8 million years ago H. erectus had new ways offinding, processing, and digesting food needed to support both largerbodies and energetically more expensive, larger brains. The mostplausible scenario, set forth by anthropologists James F. O’Connell andKristen Hawkes of the University of Utah, is that long-term trendstoward a cooler, drier climate leading up to the Pleistocene pressuredthe precursors of H. erectus to supplement a diet that hadconsisted mostly of fruit and occasionally meat. Game was increasinglyimportant, but its availability unpredictable. A division of laboremerged between male hunters and female gatherers, and social bondsensuring that men and women shared became increasingly essential.
O’Connell and others suggest that when other foods were scarce, ourancestors relied on the large underground tubers that plants in dryareas use to stockpile carbohydrates. Those storage organs occurthroughout the savanna, but are protected by a deep layer of sunbakedearth. Savanna-dwelling baboons dig up rhizomes and undergound stemscalled corms, both found nearer the surface, and at least one unusualpopulation of savanna-dwelling chimpanzees is known to use sticks todig out the shallower tubers, suggesting that early bipedal apes mayhave done so as well. But it takes special knowledge and equipment todig out the deeply situated larger tubers.
Tubers are not only hard to extract. They are fibrous and difficultto digest, hardly ideal food for children. Like nuts, they need skilledprocessing. To eat them, weaned juveniles would have to depend oncapable providers. Nevertheless, evidence is increasing that starchytubers were an important fallback food for African hunter-gatherers. A2007 report in Nature Genetics revealed that people like the Hadza ofTanzania, who rely on roots and tubers, have accumulated extra copiesof a gene that makes an enzyme useful in the digestion of starch,salivary amylase. While we can’t test the saliva or sequence the genesof African H. erectus, isotopic analysis of their tooth enamel yieldsresults consistent with a diet substantially reliant on undergroundroots. Once H. erectus developed the use of fire, perhaps as early as800,000 years ago, roasting tough, fibrous tubers would have renderedthem more digestible, and more useful still.
Even before cooking, the addition of tubers to nuts and other plantfoods gathered and processed by women would have provided newincentives for food sharing between hunters and gatherers, as well asnew opportunities for postreproductive women motivated to share. Intheir “grandmother hypothesis,” Hawkes and O’Connell propose thatDarwinian selection would have favored experienced, hardworking womenwho live on for decades after menopause, not just for a few more years,as in other primates. Such women could help provision younger kin,without the distraction of infants of their own.
（In central Africa, an Efe girl carries an infant, a common helper rolethat allows a mother to give birth again before her infant has attainedindependence. Ethnographic studies of such traditional hunter-gatherersliving in small, tight-knit communities, provide insights into howearly humans kept children alive in the Pleistocene.
Photo © Steve Winn/Anthro-Photo）
Across traditional societies, where it is not unusual for 40 percentor more of individuals to die prior to maturity, mortality rates dependa lot on family composition. Not surprisingly, presence of the mothermatters most. The father’s impact varies from being vitally importantto having no detectable impact, depending on local conditions and whoelse is around to help. When it comes to alloparents, older siblingsand grandmothers, especially maternal grandmothers, have the mostreliably beneficial impact. Under some circumstances, their presencecuts the chance of dying during childhood in half.
In purely practical terms, we can envision a sequence that beginswith hunters and gatherers sharing the fruits (and tubers) of foragingand then moving toward cooperative breeding. That would have allowedour Pleistocene ancestors to produce young that depended on manycaretakers for a long time. No ape produces such big babies that matureso slowly, yet not only did our ancestors manage to survive, but ourspecies eventually expanded beyond Africa and around the globe.
In terms of cognition and emotions, the transformations wrought byshared care and provisioning were even more profound. Our bipedal apeancestors were surely as clever and manipulative as are livingchimpanzees, able to manufacture and use tools; they must have been atleast as empathetic in some circumstances, and endowed with arudimentary theory of mind. But when they adopted what was, for an ape,a novel mode of rearing young, one that produced individuals moremutually tolerant and other-regarding than other apes, they laid thefoundations for ever higher levels of empathy and cooperation. In suchmodest beginnings we can identify the groundwork for spectacular laterdevelopments.
This article was adapted from Mothers and Others: The EvolutionaryOrigins of Mutual Understanding, by Sarah Blaffer Hrdy, ©2009,published by Harvard University Press. On sale in bookstores in April.