Mothers and Others

by Sarah Blaffer Hrdy

  Yet careful demographic analysis revealed that a forager who survived to age 15 had about a 60 percent chance of living to 45. And those who made it to 45 had a good chance of surviving into old age.23 Consider the !Kung during the period when they still lived as hunter-gatherers. The average life expectancy was only 30. But for those who survived childhood, the odds improved. Of the girls who survived to age 15, the majority (62.5 percent) survived to age 45. About 8 percent of the population lived on to 60 or older.24 Today, there is a remarkable convergence among demographic anthropologists, evolutionary-minded historians, and human biologists who study life history patterns across primates that the bodies of Homo sapiens are “designed” to last about 72 years.25

  Hunter-gatherer women who survive to middle age have a reasonable chance of surviving past reproductive age. Like this !Kung grandmother, they continue to interact lovingly with children and grandchildren. (Peabody Museum/Marshall Expedition image 2001.29.414)

  Once behavioral ecologists recognized that substantial numbers of women were living long lives, efficiently gathering and processing food for decades after menopause, it became important to explain why creatures who could no longer directly contribute to the gene pool of the next generation would do so. In early versions of “the grandmother hypothesis,” evolutionary biologists George Williams and William Hamil ton proposed that postreproductive lifespans are favored when the mother’s continued survival enhances the survival of her last-born offspring.26 A “prudent” mother could not afford to die before her last child was independent. Struck by how especially hardworking old females were, Hawkes proposed an alternative version of the grandmother hypothesis, arguing that the reason women lived longer than other apes after they ceased to ovulate had to do with their impact on grandchildren. But even with compelling evidence for the longevity and industry of grandmothers, a seemingly insurmountable obstacle remained: Even if she were still alive, a hunter-gatherer woman’s mother was unlikely to live in the same group as her daughter—or so it was thought.27

  THE ALVAREZ CORRECTIVE

  Not until near the end of the twentieth century did accumulating information from long-term studies of Great Apes in the wild prompt primatologists to reassess their assumptions about residence patterns. These field observations revealed that the breeding systems of chimpanzees and gorillas were more flexible, and the apes themselves more opportunistic, than previously supposed.

  If they could do so and still be safe and find enough fruit to eat, some females (like Gremlin’s daughter Gaia) did remain in their natal place. Fifi, daughter of Jane Goodall’s famous chimpanzee matriarch Old Flo, provides a case in point. Born to a locally dominant mother, Fifi remained in Flo’s relatively secure and food-rich home range, within her mother and brothers’ sphere of influence. Advantaged by this legacy, Fifi went on to produce nine offspring, the all-time record for a wild chimpanzee, and almost all of them survived. In 2004 when Fifi together with her last infant daughter disappeared and were presumed dead, a few of her older daughters and all of her sons still resided at Gombe in the same community as their mother and grandmother.28 Several of Flo’s daughters and some of her granddaughters continue to live in their natal place, enjoying what sociobiologists refer to as “the benefits of philopatry.”

  In addition to these field studies, new DNA evidence revealed that co-resident males who jointly defended their community against neighboring males were not necessarily close kin. Thus, even though females were more likely than males to migrate, community males—even males who were close allies—were on average no more closely related to one another than females were.29 Gorillas too were turning out to be more flexible, with both sexes routinely transferring between groups, often more than once.30 Then came a reanalysis of the ethnographic evidence for hunter-gatherers which suggested that they too were more flexible in their residence patterns than previously assumed.

  Undeniably, Murdock’s early efforts to make cross-cultural comparisons more evidence-based and amenable to statistical analysis represented a tremendous advance. From the 1960s onward, he and his followers strove to lay empirical foundations for “the science of human behavior.” But the devil was in the details, in translating complex, often very incomplete published records into simple codes that accurately reflected the complex realities of people’s residence decisions. When the University of Utah anthropologist Helen Alvarez went back to the original ethnographies for a painstaking reexamination of how Murdock had determined hunter-gatherer residence patterns, her reassessment came as a shock.

  Murdock had set up strict criteria for assigning each culture to a particular residence category. For example, a specified proportion of couples had to conform to particular residence rules in order to be assigned as either patrilocal, bilocal (or what Murdock called “ambilocal”), with residence established optionally near parents of either husband or wife and perhaps alternating over time, or matrilocal (“uxorilocal” in Murdock’s terminology). Yet as Alvarez reread the ethnographies, she realized that the numerical census data needed to meet Murdock’s criteria were rarely there. His exacting, explicit specifications not withstanding, residence patterns were often assigned on the basis of hunches. When Alvarez recoded the ethnographies, this time using only the 48 hunter-gatherer societies for which empirical evidence on residence patterns was actually available, she found that only 6 of the 48 (12.5 percent) were patrilocal. The majority, 26 of 48 (54 percent), were bilocal.31

  Notwithstanding dogmatic pronouncements about how humans “tend to be patrilocal” because “in traditional societies sons stay near their families and daughters move away,” this underlying assumption about human nature is not supported by evidence from people actually living as hunter-gatherers.32 Rather than being naturally patrilocal, most hunter-gatherer societies have remarkably flexible and opportunistic residence patterns as couples move between the woman’s natal group and the man’s.33 Furthermore, various customs increase the likelihood women will have matrilineal kin nearby when they first give birth. The same pattern we saw among the !Kung can also be found a continent away, among such bilocal foragers as the Pomo Indians of northern California: “The married couple kept moving from one family to the other . . . [but] when a child was expected they always went to live with the wife’s family.”34

  Even among unequivocally patrilocal peoples such as the Maidu foragers of northern California, the ethnographer specifically noted that “before residing permanently in the husband’s village, the married couple lived for a time with the wife’s family, and the new husband rendered service to them by providing food.”35 Murdock along with early ethnographers even had a name for it: “matri-patrilocal.”

  IF DAUGHTERS HAD MOTHERS NEARBY AFTER ALL . . .

  In less than a decade, the starting assumptions of evolutionary-minded anthropologists studying societies who still subsist (at least partly) by gathering and hunting have changed. Fieldworkers take seriously the proposal that humans evolved as cooperative breeders and so include information on available alloparents in their censuses and record the effects of their presence on child survival. Thus, when Brooke Scelza and Rebecca Bliege Bird recently went back to study the Mardu, a traditionally patrilineal and patrilocal people who still actively hunt and gather wild foods in the Western Desert of Australia—albeit these days with trucks and government food subsidies—the researchers specifically asked women how much they were able to rely on matrilineal kin to help rear their children. Following the lead of researchers working in Africa among Aka and Hadza, they also wanted to know how grandmothers and sisters strategized so as to be nearby when help was needed.

  Even though the Mardu are, like many Australian Aborigines, traditionally patrilocal, women still manage to line up matrilineal assistance. In particular, women will urge their husbands to take a kinswoman as their second or third wife. As in earlier times, sororal polygyny (when the man marries his wife’s sister) was a preferred form of marriage. Marriage to the wife’s cousin was
also common. Fifty-one percent of women were in polygynous unions with co-wives who were close relatives. Usually, polygynous marriage with more than one wife favors the husband’s reproductive interests. Several wives bear him children, but competition between wives for limited family resources can undermine child well-being. Rivalry is less pronounced when wives are related. In line with this logic, elsewhere in Australia, among the Aborigines of Arnhem Land, child survival to age five was significantly higher for polygynous families where co-wives were close relatives. In search of social support and help, Aboriginal wives actively lobby husbands to marry their sisters, and in the interests of harmony (and perhaps child well-being) men oblige.36

  Among the Mardu, 68 percent of polygynously married women were in sororal unions. Mardu mothers also obtained help from their own mothers, who often relocated to be near daughters of childbearing age, especially if the daughter was monogamously married and lacked an older co-wife to advise and help her. Mothers were especially eager to join a daughter if she was married to the same man as her sister. Traditionally patrilocal or not, half of married Mardu women ages 14 to 40 have a mother in the same group. Between footloose mothers-in-law and related co-wives, average degree of relatedness between females in a Mardu band is high, with women related to each other on average as closely as cousins and having an 11 percent chance of sharing genes by common descent. This average degree of relatedness turns out to be virtually the same as that found among infant-sharing matrilocal monkeys like langurs.37

  Did ancestral hunter-gatherers likewise have matrilineal kin nearby? We cannot know for sure, but post-Alvarez, long-standing barriers against thinking this was possible have disappeared. Instead of some highly conserved tendency, the cross-cultural prevalence of patrilocal residence patterns looks less like an evolved human universal than a more recent adaptation to post-Pleistocene conditions, as hunters moved into northern climes where women could no longer gather wild plants year-round or as groups settled into circumscribed areas. In the Middle East, people began to herd livestock and became increasingly dependent on growing crops, storing the surplus, and accumulating property. As group sizes along with population densities increased, people adjusted their behavior to these new demographic, dietary, epidemiological, and social realities.

  For settled people, shorter birth intervals and faster population growth, along with the accumulation of resources and the emergence of social stratification, brought with them the need to protect livestock and cultivable land as well as wives and children. Protecting such valuable resources became a higher priority than maintaining cordial and reciprocal exchange with neighbors. As outside invasions became more routine, men needed allies they could count on. Who better to rely on than close male kin? Increasingly, men sought to remain near fathers and brothers, obtaining wives from other groups. Only in the past 10,000 or so years has interclan warfare become an integral part of human lives, necessitating patrilocal residence patterns and in the process changing the way that children are reared.

  As in all primates, mothers without support from matrilineal kin lost some of their autonomy. The reproductive interests of patrilines increasingly took priority. With both patrilocal living arrangements and shorter birth intervals, the alloparents at hand were more likely to be older siblings of the current infant than maternal grandmothers or great aunts, with mixed results for children, not always good.

  GENETIC EVIDENCE ABOUT RESIDENCE IN THE RECENT PAST

  Based on genetic evidence from the past few thousand years, after the introduction of herding, horticulture, and social stratification, we know that women in many parts of the world were marrying out and moving between groups. But so far, genes cannot tell us much about residence patterns during the Paleolithic when our ancestors still lived exclusively by gathering and hunting. Let me explain.

  Analyses of non-recombining portions of the Y chromosome, which is passed only from fathers to sons, as well as comparative frequencies of mitochondrial DNA, which is passed exclusively from mothers to both daughters and sons, reveal that in the past five thousand years or so women were more likely to move between populations than men were. If residence was patrilocal, we would also expect reproductive behavior to have been more tightly regulated, since men living in patrilocal clans tend to guard their mates from outsiders. Such reproductive control could explain why gene flow between patrilocal populations was largely confined to women.38

  Consider an admittedly extreme but very telling case involving the recent migration of people between Africa and the Middle East. While there was little male-mediated gene flow from sub-Saharan Africa into the area around Yemen about 2,500 years ago, as evidenced by Y-chromosome data, mitochondrial DNA indicates a tremendous influx of fertile women of African origin around this time. This genetic information, combined with historical accounts, means that captured or enslaved African women bore children to Middle Eastern Arab men. African men either were not taken to the new location or, if taken, left no surviving offspring.39 Conquests yielding access to women are starkly inscribed in genetic records from other parts of the world as well. The most famous case involves genetic evidence from a particular haplotype on the Y chromosome that points to a rapid spread of genes from one particular male lineage linked to Genghis Khan. It is consistent with the dates of his army’s conquests across Asia from the Pacific to the Caspian Sea.40

  Such reproductively skewed patterns contrast with those from matrilocal societies, which tend to be more relaxed about who breeds with whom. Routinely, both sexes move around, although men usually move somewhat more. Over the past 10,000 years or so, matrilocal and matrilineal societies have increasingly given way to pressures from expansionist, patrilineal neighbors and invaders so that patterns of conquest are widely documented across Europe, Africa, Asia, and South America.41

  Genes tell us a surprising amount about patterns of conquest. They even tell us when people started to live in cities, rely on milk products, or suffer from various diseases. They can shed light on when dogs and cats began to be domesticated. Comparing the genetic histories of lice that live in body hair with lice that cling to garments even allows us to make an educated guess at when humans started wearing clothes.42 But genetic evidence tells us almost nothing about the residence patterns of men and women prior to a few thousand years ago, with one possible exception.

  In 2000 scientists working on the Human Genome Project reported that genes involved in sperm production turn out to have evolved at an unusually fast rate compared with other genes. This curious finding suggests that there may have been selection pressure on our hominin ancestors to produce quantities of competitive sperm, a trait critical for the reproductive success of males in primates where females mate with more than one male.43 Polyandrous matings would not be at all consistent with females being captured or exchanged between patrilineal clans, where reproductive access to women is closely guarded. However, occasional polyandrous matings are perfectly consistent with the more flexible breeding combinations (alternately monogamous, polyandrous, and polygynous) found in cooperative breeders.

  Even if this admittedly speculative interpretation concerning sperm-related genes holds up, genetic evidence still does not tell us whether or not matrilineal kin were on hand to help mothers rear their children among African Homo erectus 1.8 million years ago. What it does do, though, is remind us how much evolution has gone on since humans last shared a common ancestor with gorillas (whose females mate with a single alpha male and where sperm competition is virtually nonexistent) and with chimpanzees and bonobos (whose females mate with many males and where sperm competition plays an important role in reproductive fitness). Each species of ape differs from every other, and none of them breed like women do today.

  Chimpanzee females advertise ovulation with large red swellings around the time of ovulation, and they only copulate around midcycle. Bonobos, by contrast, exhibit swellings that last for weeks and copulate with multiple partners throughout most of their cycle. Gor
illas, orangutans, and women do not advertise ovulation with conspicuous swellings at all.44 In other words, reproductive traits like the sexual swellings of chimpanzees can evolve quite fast, and 5–10 million years have elapsed since humans last shared a common ancestor with primarily patrilocal African Great Apes. This is the main reason I agree with Alvarez that as far as residence patterns are concerned, the best we can do is extrapolate from people who still lived as nomadic foragers when first described.

  Granted, the residence patterns of modern hunter-gatherers may or may not resemble those of the first anatomically modern humans. As humans became behaviorally modern—armed with higher-caliber tools and weapons, with fire to cook food, and with language to communicate—their subsistence strategies would likely have diverged from those of the earliest Pleistocene hunter-gatherers. Their lifeways would have been altered further still by contacts with post-Neolithic herders and neighboring farmers, not to mention anthropologists. Yet ethnographic evidence about these people reveals the sorts of behaviors, customs, and strategic maneuvering by parents and alloparents that make it feasible for highly mobile foraging peoples to survive and rear unusually costly and slow-maturing children. This is the basis for arguing that in order to successfully reproduce, foragers needed to be, and were, opportunistic and flexible in their mating and residence patterns.

  If correct, Helen Alvarez’s revised interpretation of hunter-gatherer residence patterns removes the last barrier for taking seriously the hypothesis that maternal grandmothers and other matrilineal kin helped early hominin mothers rear their young. But even if a mother had older matrilineal kin nearby to help, would they want to?

  ON THE ALTRUISM OF AGING FEMALES