Evolutionary Anthropology 22:294–302 (2013)

ARTICLE

Grandmothers and the of Longevity: A Review of Findings and Future Directions

KRISTEN HAWKES* AND JAMES E COXWORTH

Women and female great apes both continue giving birth into their forties, but the biggest genetic contribution to not beyond. However live much longer than other apes do.1 Even in future generations. That makes selec- hunting and gathering societies, where the mortality rate is high, adult life spans tion stronger on traits expressed in average twice those of chimpanzees, which become decrepit during their fertile young adults and weaker on traits years and rarely survive them.2,3 Since women usually remain healthy through expressed at older ages. How fast the and beyond childbearing age, human communities include substantial propor- force of selection declines across tions of economically productive postmenopausal women.4–7 A grandmother adulthood depends on two things: hypothesis8–12 may explain why greater longevity evolved in our lineage while the chances of surviving to older female fertility still ends at ancestral ages. This hypothesis has implications for ages and what those survivors to the evolution of a wide array of human features. Here we review some history of older ages can do for their own the hypothesis, recent findings, and questions for ongoing research. . Since natural selection spreads only traits that increase their own 13 THEORETICAL FOUNDATIONS and physiological systems. Natural reproduction, it should not favor a selection favors more investment in postreproductive period in the nor- Aging rates vary widely among liv- somatic maintenance only when it mal life span of any organism. What 14 ing things. Evolutionary explanations raises lifetime inclusive fitness. Two then, Williams asked, of meno- for varying rates assume key theoretical contributions to the pause, which was then assumed to that organisms trade off investment evolutionary study of senescence spe- be a uniquely human trait? His in current reproduction against cifically mentioned humans, dwelling answer began with the important repair and maintenance of their cells on the midlife end of women’s fertil- point that the end of fertility is not ity as a crucial clue about our evolu- the end of reproduction. As long as tion. Their inferences represent two postmenopausal women contribute Kristen Hawkes’s hunter-gatherer eth- different ways of posing the evolu- to the welfare of their kin, they affect nography drew her attention to unex- tionary riddle of mismatch between the successful reproduction of their pected sex and age differences in genes. Williams hypothesized that foraging strategies. including the crucial rates of ovarian and somatic senes- productivity of grandmothers, which cence in women. evolved when other prompted further comparisons of human changes in our lineage made late and chimpanzee life histories. births riskier and infants more James E. Coxworth’s central interest is Stopping Early dependent. Older mothers would be the application of evolutionary tools to 14 describe and explain male competitive George Williams elaborated on more likely to die in childbirth, leav- strategies, with particular emphasis on the consequences of the declining ing orphans unable to survive with- . This interest has led to force of natural selection across statistical and modeling contributions out them. He surmised that these and to projects investigating human life- adulthood. Even in populations of a circumstances would favor tenden- history evolution. hypothetical organism that did not cies to stop fertility early and for become increasingly frail with age, mothers to invest more in previously Key words: life history evolution; senescence; some individuals would die in acci- born offspring than in risky new cooperative child rearing; infant psychology; dents or from illness or predation, ones. male-male competition so sets of age-mates necessarily Subsequent work showed that 15 shrink with time. Adult cohorts are menopause is not uniquely human.16 largest at the beginning, the age of VC 2013 Wiley Periodicals, Inc. Menstrual cycling ends before death DOI: 10.1002/evan.21382 first reproduction. Thus, features in other primate females as well, if Published online in Wiley Online Library 7,17 (wileyonlinelibrary.com). expressed at early adult ages make they live long enough. In all ARTICLE A Review of Findings and Future Directions 295 mammals, including primates, a but, of the hunter-gatherer girls who selection must weaken as adult finite stock of oocytes develops survive to adulthood, about three- cohorts age, providing the founda- around or before birth and then is quarters live past the age of 45.29 At tion for Williams’14 arguments. Med- continuously depleted, mostly by cell any one time, a third or more of the awar also anticipated both death18; cycling stops when stocks adult females in human populations Hamilton’s 196431 explanation of the fall too low to support . Fol- are beyond the childbearing ages.7,27 importance of inclusive fitness and licle stock sizes and rates of atresia Hamilton’s 196630 suggestions about 19 vary across mammals, with some grandmother effects by saying, species still ovulating and giving Increased Longevity As a “Grandparents, though no longer fer- birth at older ages than humans do. Consequence of Grandmother tile, may yet promote (or impede) For example, elephants continue to Effects the welfare of their grandchildren, give birth into their sixties20 and 30 and so influence the mode of propa- Antarctic fin whales into their William Hamilton’s attention gation of their genes.”15 21 eighties. Yet there are clear simi- was drawn to grandmother effects Hamilton30 mathematically mod- larities in the oldest ages of parturi- when he used demographic data eled Williams’ verbal arguments tion among all living hominids and from a human population with mor- about the evolution of senescence. in rates of decline in follicle stocks tality even higher than that typical of Then, considering humans, he from birth to the late forties in hunter-gatherers to evaluate the fit 22 pointed out that the high rate of wom- humans and chimpanzees. These of his mathematical model of the en’s postfertile survival “inevitably commonalities among members of evolution of senescence. Evidence of suggests the special value of the old the hominid radiation are inconsis- woman ... during a long ancestral tent with the proposition that period.”31 Like Williams, Hamilton humans had an ancestor with older lacked information on the other great ages at last birth. It is likely that Hunter-gatherer life apes or human populations that were ancestral ovarian aging remained the expectancies are less not dependent on agriculture. Subse- same while increased longevity quent evidence from the Hadza, a evolved in our lineage.23 than half those of group of East African hunter- Williams14:407 also surmised that Western nations, but, of gatherers, highlighted a particular women’s notably long postfertile sur- value of old women: They acquire and vival was an “artifact of civilization.” the hunter-gatherer girls process foods that youngsters cannot In this, he was misled by the incorrect who survive to handle for themselves.32 inference that skeletal assemblages demonstrate the rarity of surviving adulthood, about elders in ancient human populations. three-quarters live past While debate continues,24–26 paleode- 29 A GRANDMOTHER HYPOTHESIS mographers have shown that the age the age of 45. At any Those Hadza observations, com- structure of past populations cannot one time, a third or more bined with Eric Charnov’s formal be retrieved by estimating ages of of the adult females in models of life-history evolution in specimens in skeletal assemblages. female mammals,33 prompted a Biases are imposed by comparative human grandmother hypothesis to explain collections; preservation varies with populations are beyond age at death; and standard measures the evolution of human life histories. widely misestimate adult ages. Even if childbearing ages. In Charnov’s models, adult mortality taphonomic and age estimation determines an optimal age at matu- biases could be controlled, skeletal rity. When mortality is high and assemblages are not random samples adult life spans are short, selection of the deaths that occurred in differ- favors early maturation because ent sexes at different ages.27 the substantial fraction of women those who wait are more likely to die The deep antiquity of human lon- who survived the childbearing years before reproducing. When mortality gevity is also obscured by wide- led him to point toward their role in is low, the cost of waiting goes spread assumptions that because the evolution of increased longevity. down; net advantage goes to delaying national life expectancies began to An earlier version of the idea that maturity to reach a larger adult size. exceed 50 years only in the twentieth selection could favor increased lon- Larger mothers can allocate more to 34 century,28 our postfertile life stage is gevity through grandmother effects offspring [see Hawkes for more an artifact of recent history. Life- was mentioned by Peter Medawar in discussion of the modeling assump- span averages are misleading guides a footnote to his 1951 lecture, “An tions]. Rates of growth are governed to adult survivorship because high Unsolved Problem in Biology,” pub- by trade-offs the models do not infant and juvenile mortalities have lished in 1952.15 In that paper Meda- address, but they vary among mam- strong effects on them.27 Hunter- war documented and began to solve malian orders. Primate growth rates gatherer life expectancies are less the puzzle of the evolution of senes- are notably slower than those of 35 than half those of Western nations, cence by showing why the force of most nonprimate mammals. 296 Hawkes and Coxworth ARTICLE

Looking just at primates, Char- those elders subsidized their depend- dependencies point to pathways nov,33 Fig. 5.6 illustrated the wide ent grandchildren, the grandchil- whereby selection would increase lon- variation in age at maturity (first dren’s mothers could have their next gevity as longer-lived grandmothers left parturition) and average adult life babies sooner. Hadza foraging more descendants.9, 11 span with a dataset that included pointed directly toward links among humans. While age at first birth and ecology, size and strength constraints average adult life spans are notably on youngsters’ foraging, grand- Mathematical Formalizations higher and longer in humans, Char- mothers’ subsidies, and birth spacing. nov’s plot showed that the relation- This is because Hadza grandmothers’ Formal models of grandmothering ship between these traits is the same tuber digging and processing sup- depend on general principles, such as for us as for other primates. Yet, ports the continuing growth of Hadza thoseinvestigatedbyRonaldLee.38 unlike other primates, human adults children when their mothers are His one-sex model showed that inter- have life spans that include a post- occupied with newborns.32 generational transfers of support fertile stage. If postfertile grand- These connections between ecology maintain a human-like age structure. mothers subsidize the childrearing of and juvenile foraging competence sug- This extended his earlier model,39 their daughters, who then produce gested how ancestral grandmothering which demonstrated that when inter- more surviving offspring, the postfer- would have favored shifts in mecha- generational transfers are crucial for tile females, in consequence, have nisms of cell and molecular mainte- successful reproduction, it is age- more descendants. Helpful grandmo- specific productivity, not fertility, that thering could therefore explain the sets the strength of selection against shorter birth intervals of humans senescence. Hillard Kaplan and Arthur than the other great apes, the evolu- Unlike other primates, Robson40 formalized a different model tion of our exceptional longevity human adults have life for the evolution of longevity, high- with its postfertile stage, and the spans that include a lighting increased investment in evolution of our later age at first brains and skill learning, especially for birth.34 The grandmother hypothesis postfertile stage. If hunting. Their one-sex model ignored can explain why we mature so late postfertile grandmothers questions about the postfertile longev- by explaining why adult life spans ity of females. But subsequently, increased in our lineage.8–12,36 subsidize the Kaplan and coworkers41 drew on it childrearing of their and ethnographic data to show that households could not meet consump- The Ecological Scenario daughters, who then tion requirements without the produc- As climates dried in Plio- produce more surviving tivity of postfertile elders. Pleistocene Africa, the spread of It was Frederike Kachel and col- grasslands restricted the availability offspring, the postfertile leagues42 who first built a two-sex of foods that nonhuman apes depend females, in model of the grandmother hypothe- on. Observation of Hadza foragers sis. They concluded, from simulations exploiting a modern analog of those consequence, have of their agent-based model, that even environments found young children more descendants. when grandmother effects were so to be surprisingly active foragers and, strong that they guaranteed the sur- like other juvenile apes, to be compe- vival of grandchildren, those effects tent at picking soft fruits and ber- had little impact on longevity. To ries.37 But youngsters are too small to understand why, we investigated their be effective at digging and processing nance that slowed aging and model and found that simulations, the deeply buried tubers that are lengthened survival. Ape-like aging, either with or without grandmother- Hadza staples.32, 37 Geophytes like previously maintained by stabilizing ing, never reached an equilibrium, those relied on by the Hadza became selection, would now be subject to because of their assumptions about more abundant as ancient grasslands directional selection. Where increased males.43 Longer lives had no cost for spread,9 presenting valuable resource somatic maintenance had been males, so the perpetual advantages opportunities to those able to extract selected against because it meant less they gained from increased longevity and process them. Young ancestral allocation to current reproduction and obscured the selection consequences juveniles, like Hadza children, would reduced offspring survival, greater of grandmother effects.43 not have been able to do it. maintenance was now favored. Building on lessons from Kachel’s Ancestral mothers who relied on Because grandmothers subsidized work and taking advantage of the those foods would have had to subsi- juveniles, mothers who shifted more life-history relationships assumed in dize their weaned offspring. That resources to their own physiological Charnov’s formal models,33,35 Peter would have opened a novel fitness maintenance, leaving less for current Kim constructed an agent-based window for the few elder females reproduction, did not suffer higher off- model and used deterministic differ- whose declining fertility meant they spring loss. The more robust the elders, ence equations to simulate the verbal had no newborns of their own. If the more they could help. These inter- grandmother hypothesis.12 Starting ARTICLE A Review of Findings and Future Directions 297 with an ape-like life history and ation in the sources of childrearing are paternal or maternal.49 Harmful assuming that the end of female fer- help is as wide as the variation in paternal grandmothers50,51 and ten- tility remained fixed, while increased other aspects of socioecology.44 sions between mothers and daughters- longevity delayed maturity and Researchers looking for grandmother in-law are not uncommon, but this lengthened juvenile dependence, effects usually find them, 45,46 varies with details of socioecology.52,53 Kim’s simulations showed that very although exceptions have been As Rebecca Sear54 has shown, mater- weak grandmothering was enough to noted.47 Often measured by the gold nal grandmothers can also be harmful, drive the ape-like equilibrium to a standard of grandchild survival, as can mothers,55 when scarce resour- human-like one. The ancestral ape- grandmother effects might be under- ces must be carefully allocated. like condition included very few estimated where women can choose Surprisingly, measured effects also females who were eligible to grand- where to direct their help. If they allo- sometimes differ depending on a mother, but in 60,000 years or less, cate it to those that need it most, as grandchild’s sex. Grandmothers share postfertile grandmothers expanded Hadza grandmothers do,48 they one X chromosome with their sons’ to a proportion of the population reduce differences between those daughters and none with their sons’ similar to that observed among with and without grandmothers’ help. sons; there is a 50% chance that they hunter-gatherers. In this model, We assume that selection propelled share an X chromosome with their trade-offs for increased longevity dif- by grandmother effects was stronger daughters’ children. Molly Fox and fer between the sexes, although through help to daughters’ children, colleagues,56 analyzing this variation grandmothering makes increased not only because of the assurance of across several populations, found cor- longevity advantageous through both shared genes with the grandchild, but relations between the strength and males and females. The model direction of grandmother effects and includes no skill learning, no brain the likelihood of a shared X chromo- expansion, no hunting, and no pater- some. Further investigation of these nal provisioning. It demonstrates Even though model intriguing differences must continue. that, at least in principle, grandmo- Ethnographic and historical varia- thering alone could have been the grandmothers help any tion in the strength and direction of foundation for subsequent evolution dependent juvenile old grandmother effects reveals the com- of other distinctly human features. plexities of reproductive competition. Kim36 then used the agent-based enough to survive Michael Cant and Rufus Johnstone57 model to build probabilistic simula- without its mother, showed that if females disperse and tions. Adding stochasticity slowed increased longevity then mate locally, fitness benefits the transition to a human-like life from helping coresident kin change history by about an order of magni- evolves nevertheless. with age. They modeled an ancestral tude. It even, in some cases, pre- population in which fertility in vented escape from the ancestral females continued to late ages. Repro- condition. But multiple simulations ductive conflict over care of depend- confirmed that our assumptions lead also because a main effect is shortened ents led to an equilibrium in which to two life-history equilibria, a great- birth intervals for the childbearer elder females stopped childbearing ape-like one and a human-like one. whose offspring are subsidized. Yet, in early. This stopping-early scenario Once the advantages of grandmo- our small initial Hadza sample, help- could, Cant and Johnstone suggested, thering begin to spread, selection ers were not always maternal grand- explain the evolution of human meno- moves inexorably to the human equi- mothers. Peter Kim’s simulations12,36 pause. However, in addition to the librium. The absence of intermedi- include no restriction to helping problems with stopping-early scenar- ates is consistent with the possibility daughters’ children. Even though ios noted earlier, the assumption of that the evolution of this distinctively model grandmothers help any depend- ancestral female dispersal is disputa- 58 human feature, helpful grandmother- ent juvenile old enough to survive ble. That assumption had been ing, is the foundation for genus without its mother, increased longev- influentially justified by male philopa- , long antedating the appear- ity evolves nevertheless. More investi- try in the genus Pan and longstanding 9 ance of our species. gation of this unexpected outcome characterizations of hunter-gatherer 59 and more modeling to evaluate the bands as patrilocal. However, such characterizations of hunter-gatherer BEHAVIORAL FINDINGS consequences of different family grouping patterns on the evolution of residence patterns have now been The grandmother hypothesis is spe- longevity are obvious next steps. repeatedly falsified by evidence indi- cifically directed at the evolution of cating high mobility and regularly 8, 9 60–62 human life history. When research- Variation in Grandmother changing group composition. ers look at contemporary and histori- Effects cal human populations with diverse PHYSIOLOGICAL FINDINGS social organizations and subsistence Ethnographically and historically regimes, they always find characteris- measured effects sometimes vary The hypothesis that grandmother- tically human age structures, but vari- depending on whether grandmothers ing slowed somatic aging in our 298 Hawkes and Coxworth ARTICLE lineage draws attention to physiolog- late sixties do levels fall to the chim- Telomere length in replicating tis- ical puzzles. How do we do it? Estro- panzee maximum.68 Consistent with sues varies with age, and shortening gen plays a key role in the the proposition that increased pro- rates correlate with rates of aging maintenance of many physiological duction of DHEAS is one mechanism both within and between species.72 systems in both sexes, but that ques- that slowed aging in our lineage, cir- On these grounds, we expected the tion has special force when it is culating DHEAS levels are even greater longevity of humans to be focused on women. Although men lower in the other great apes.69 Add- associated with slower telomere produce gonadal steroids that, ing to this notable and surprising shortening rates and hypothesized throughout life, can be converted to variation among the hominids, corre- that attrition rates would be twice as estrogen in peripheral tissues, ovaries lations found between adrenal ste- fast in chimpanzees as humans. To secrete estrogen mostly during roid levels and menopausal status in find out, Justin Tackney73 extracted cycling; circulating levels drop below individual women70 pose questions, DNA from our chimpanzee blood detection after menopause.63 Never- as yet unexplored, about the roles of samples and, using the same proto- theless, postmenopausal women other adrenal androgens in hominid col, assayed telomere repeats in the maintain competence in most physio- physiological maintenance. chimpanzee sample and an age- logical systems aside from fertility. matched sample of women. Instead The maintenance puzzle is under- of faster shortening rates in chim- scored by comparison with other great panzees, results showed similar rates ape females, which usually die while Hormonal assays of attrition in the two species. Even they are still cycling. Evidence of geri- showed that circulating more surprising, telomeres in atric morbidities in these years is best humans are about half as long as documented in chimpanzees.2,17,64 DHEAS declines are those in chimpanzees.73 slower, not faster in Gomes and colleagues74 have pro- vided the phylogenetic framework Adrenal Androgens chimpanzees. But young for interpreting these initially aston- The lack of correspondence adult DHEAS levels are ishing results. Examining cell lines between somatic maintenance and cultured from more than 60 mam- ovarian estrogen production in homi- three times as high in mal species, they found telomere nids is reason to suspect that periph- women as in chimpan- lengths to be negatively correlated eral estrogen must have nonovarian with life span. Their analysis indi- sources. Consistent with that, studies zee females. Only as cated that ancestral mammals had of women’s health report that adrenal women reach their late relatively short telomeres. This, they steroids account for most of women’s hypothesized, was a response to the peripheral estrogen, even during their sixties do levels fall to higher load that accompa- cycling years.65 That evidence, com- the chimpanzee nied the evolution of homeothermy bined with nomination of the adrenal maximum. and made the protective advantage steroid dehydroepiandrosterone sul- of short telomeres and telomerase fate (DHEAS) as a biomarker of aging suppression outweigh the costs of in primates,66 led us to expect higher replicative aging. Subsequently, telo- DHEAS levels in women than in merase expression and longer telo- Telomeres female chimpanzees. If that steroid meres evolved in many smaller, plays an important role in somatic Another mechanism to retard shorter-lived species. maintenance, then higher levels in somatic aging in humans might be Comparative primate data are women could help explain the persis- slowed telomere attrition. Telomeres sparse, but human telomeres appear tent physiological health of women are the noncoding DNA repeats at to be the shortest in the order.73 Our well beyond menopause. the ends of chromosomes; they very short telomeres may have The biomarker of aging hypothesis maintain chromosome integrity at evolved as greater longevity raised predicts that DHEAS levels decline each cell replication. Telomeres lose cancer dangers, increasing the twice as fast in chimpanzees as they sequence repeats at each replication, advantages of protection against do in women.67 To find out, we arresting the cell cycle when they them. Our comparison of chimpan- requested that blood be drawn from become too short. This attrition zee and human telomeres implicates captive female chimpanzees when decreases an aging organism’s the evolution of less sensitivity to they were anesthetized for reasons capacity to regenerate tissues. While cell death, increased DNA repair, unrelated to our project. Hormonal telomeres can be rebuilt by telomer- decreased cell division rates, and/or assays showed that circulating ase, one advantage of suppressing other mechanisms to maintain func- DHEAS declines are slower, not telomerase expression in replicating tion with shorter telomeres in the faster in chimpanzees. But young somatic tissues is that shortened human lineage. This surprising find- adult DHEAS levels are three times telomeres block the growth of malig- ing underlines the indispensable role as high in women as in chimpanzee nant cells. Thus, telomere shortening of great ape comparisons for under- females. Only as women reach their may protect against cancer.71 standing human aging.73 ARTICLE A Review of Findings and Future Directions 299

FURTHER IMPLICATIONS OF THE cult for infants and juveniles to advantage. Hrdy84 explicitly links our 81 GRANDMOTHER HYPOTHESIS handle on their own. The processing cooperative childrearing patterns and steps of cooking vastly expand oppor- the selection pressures they impose on The grandmother hypothesis tunities for resource pooling.82,83 infants to the evolution of social and addresses the evolution of human life Wrangham’s enumeration of the ana- emotional capacities that Michael history. A distinctive postfertile life 7,75 tomical characteristics of Homo erec- Tomasello and collaborators charac- stage is a characteristic of our spe- tus that indicate dependence on terize as shared intentionality.85,86 cies. The grandmother hypothesis pro- cooked food76,77 identifies evidence of Hrdy87 highlights evidence of the poses that grandmothering drove the reliance on processing that young ingratiating sensibilities of human evolution of the genus Homo.Thefirst juveniles cannot manage. This directly infants who are especially prone to expansion of our genus out of Africa implicates helpful grandmothers in detect and repeat what appeals to implies a life history that allowed the evolution of our genus.9 caregivers. These abilities and the mothers to colonize habitats in which motivation to please others and estab- just-weaned juveniles could not feed lish joint attention and mutual under- independently. Grandmothers’ subsi- standing are the foundation for our dies can explain how our genus spread The size and nutritional distinctively human cultural lives. so quickly through much of the previ- richness of meals make Hrdy’s hypothesis draws attention ously unoccupied temperate and tropi- to the evolutionary importance of the cal Old World. This scenario them attractive very early neurological development implicates cooking in the evolution of opportunities for and social sensitivity of human our genus; it entails shifts in maternal 88 potential consumers babies, which has been obscured by trade-offs that impose novel selection longstanding characterizations of pressure on infant sociality; and those other than the humans as “secondarily altricial.” social capacities, as well as the sub- processor.77,79 Babies are physically helpless, but this stantial shift in operational sex ratio, is also true of newborn nonhuman have large consequences for competi- Youngsters are apes. Human neonates smile and coor- tion among men. especially likely dinate mutual gaze with caretakers, an initial interactivity not evident in appropriators. wild chimpanzees,89 but seen in cap- Grandmothers’ Cooking tivity.90 The attention of multiple care- Ethnographic lessons from the takers in captivity evokes these modern Hadza link the evolution of Selection on Infants and responses from chimpanzees under helpful grandmothering to the eco- circumstances similar to those of the Distinctively Human Prosociality logical circumstances of the African first ancestral infants arriving in a savannah and the likely importance Novel interdependencies entailed in world where grandmothering was dis- of cooking in the evolution of genus a grandmothering life history have placing independent mothering.87 Homo.9 Richard Wrangham76, 77 and enormous consequences for human Such responses hint that this tendency collaborators78 have marshaled and psychology. Sarah Hrdy55,84 has was also present in our shared ances- elaborated compelling morphological focused attention on the new prob- tors. It appears to be a trait that was and physiological evidence of the lems that cooperative childrearing exaggerated by social selection as a dependence of our genus on cooked posed for both ancestral mothers and consequence of the change in rearing food. They have also recognized that infants. Unlike other ape mothers, environments.84 The drive for shared cooking increases the vulnerability of who rear offspring one at a time, help attention quickly fades in chimpanzee resources to expropriation.79 Cooks allows human mothers to have next babies,90 but persists and expands in gain notable economies of scale as babies sooner without necessarily humans. This is consistent with Hrdy’s items are accumulated, fuel gath- reducing the chances for survival of hypothesis of a history of strong selec- ered, and the nutritional utility of their previous child. Mothers’ repro- tion on the social and emotional the meal increased with processing. ductive success then depends on capacities of human infants. With These steps concentrate feeding into enlisting assistance and allocating their cooperative rearing, engaging care- long handling stages followed by attention among multiple dependents. takers’ commitment became a matter bursts of highly efficient consump- These trade-offs set up challenges of life and death, leading to greater tion.80 The size and nutritional rich- for human babies not faced by other survival for ancestral infants more ness of meals make them attractive ape babies. Unlike other apes, human motivated to coordinate with others opportunities for potential consum- babies cannot count on their mothers’ and more effective at doing so. Thus, ers other than the processor.77, 79 undivided commitment even though “emotionally modern hominins,” to Youngsters are especially likely attention from their caregivers has life use Hrdy’s label, evolved well before appropriators. Some food sharing and death consequences for them. the appearance of Homo sapiens. between mothers and offspring is Thus, infants who are more successful Human language, cultural learn- widespread among primates and at actively engaging their mothers and ing, and cooperative activities more likely with foods that are diffi- grandmothers have a strong selective depend on the distinctively prosocial 300 Hawkes and Coxworth ARTICLE motivations apparent in prelinguistic and it notably contrasts with the history expands both the ages and the infants. Recent techniques for prob- opposite preference in other prima- relative number of men in competition ing the social sensitivities of very tes, especially well documented in for paternities.95 The consequences of young human babies show them chimpanzee males, which prefer old these age-structure changes for mate evaluating interactions and discrimi- females.94 But the human reversal guarding and other male strategies are nating between helpful and harmful might well be a consequence, not a overdue for attention. Notably, com- actors during the first postnatal cause, of our midlife menopause. parison of four hunter-gatherer soci- weeks.88 If the advantages of lan- With the evolution of human longev- eties with wide variation in pair-bond guage and cultural cooperation are ity and its postfertile stage, many stability96 showed that father effects subsequent to prosocial capacities, healthy and competent females are cannot explain the variation. Pair- those subsequent advantages cannot not fertile. Under these circumstan- bond stability depends instead on vari- explain why our unique psychology ces, males that favored mates with ation in the operational sex ratio. evolved in the first place. Hrdy’s markers of youth might have left When there are fewer fertile females child rearing hypothesis might. Con- more descendants. relative to the number of competing tinuing work to assess the social males, pair bonds last longer.96 motivations and abilities of infants Grandmothering contributes to the in all the hominids will help clarify evolution of shared intentionality85 these early differences. Postnatal Hrdy’s framework makes through its effects on infants whose brain imaging and gene expression mothers have overlapping dependents. studies are also promising lines of precocious social Resulting capacities and motivations comparative evidence about the sensitivities a crucial shift for mutual engagement have important developmental timing of interactive consequences for all of social life, capacities.88 Hrdy’s framework in our lineage, a shift as including relationships among men.97 makes precocious social sensitivities important as the delay in When joint activities are a preoccupa- a crucial shift in our lineage, a shift tion, in part due to their jointness, as important as the delay in other other aspects of occasions for evaluating others multi- aspects of development that has development that has ply, expanding opportunities for com- made neoteny seem to be the major petition over relative social standing.97 determinant of human evolution.91 made neoteny seem to That competition and its profound con- Accelerated social cognition, as well be the major sequences can be obscured when men as delayed independence and delayed are characterized as primarily hus- maturity, would all have been determinant of human bands and fathers, with nuclear fami- favored as longevity increased with evolution lies highlighted as the basic social and ancestral grandmothering.88 economic units of human societies. The grandmother hypothesis under- scores the importance of relationships beyond nuclear families. Pair bonds Men’s Strategies In Peter Kim’s simulations,12,36 likely evolved in our lineage as addi- Our grandmothering life history adult life spans favored by female tions to other social relationships.58, 98 alters the constraints and opportuni- trade-offs alone are shorter than those Preferences for shared intentionality ties for men compared to other ape favored when males are included. The wouldhaveaffectedallofthoserela- males. Morton and colleagues92 simulations do not allow male strat- tionships, expanding both the opportu- recently modeled the effects of men’s egies to shift with increasing longevity, nities for coordination and the range of preference for young mates in a pop- however. Exploring these potential occasions to identify and impress ulation having human longevity but shifts could contribute to explanations potential allies and competitors. female fertility that continues to for the evolution of pair bonding. older ages. As noted, the similarity in Other ape males compete for pater- CONCLUDING REMARKS latest ages of parturition in all living nities within a narrower range of adult hominids make such an ancestor ages. With human longevity, a man’s Hypotheses about what happened in unlikely. Moreover, Morton and competitors and potential allies human evolution are necessarily con- coworkers’ assumptions about repro- include old men who are always ahead strained by the ancestors we assume. duction invite skepticism. In their of younger ones in establishing their Frans de Waal’s continuing documen- model, a birth occurred only after a social position. On one hand, grand- tation of emotions and empathy in death, with potential mothers then mothers’ subsidies shorten birth inter- other primates99,100 enriches Darwin’s competing for the maternity. With vals, increasing the rate of paternity case that these mechanisms for social those assumptions, men’s preference opportunities in humans; on the other, behavior also characterized our ances- for young mates could explain the men continue to compete for pater- tors.101 From an already smart, evolution of early termination of fer- nities to much older ages, while child- empathic, social ancestor, social sensi- tility. Men’s preference for young bearing ages do not increase. As a tivities evolved to develop earlier and mates does seem to be established,93 consequence, our grandmothering life extend further in humans;88 appetites ARTICLE A Review of Findings and Future Directions 301 for mutual attention distinguish us the NSF for research support, Grant 22 Jones KP, Walker LC, Anderson D, et al. 85,86 BCS-0717886. 2007. Depletion of ovarian follicles with age in from other apes. The most promis- chimpanzees: similarities to humans. Biol ing hypothesis to explain those fea- Reprod 77:247–251. tures is our distinctive habit of REFERENCES 23 Hawkes K, Smith KR. 2010. Do women stop cooperative childrearing,84 a pattern early? Similarities in fertility decline in humans 1 Robson SL, van Schaik CP, Hawkes K. 2006. and chimpanzees. Ann NY Acad Sci 1204:43–53. that, as we have argued here, likely The derived features of human life history. In: 24 Caspari R, Lee S-H. 2004. Older age began with helpful grandmothering. Hawkes K, Paine RR, editors. The evolution of becomes common late in human evolution. human life history. Sante Fe, NM: School of Proc Natl Acad Sci USA 101:10895–10900. Human appetites for shared attention, Advanced Research Press. p 17–44. as well as the relatively male-biased 25 Hawkes K, O’Connell J. 2005. How old is 2 Goodall J. 1986. The chimpanzees of Gombe: human longevity? J Hum Evol 49:650–653. operational sex ratio that evolved with patterns of behavior. Cambridge, MA: Belknap Press of Harvard University Press. 26 Caspari R, Lee S-H. 2005. Are OY ratios our life history, have consequences for invariant? A reply to Hawkes and O’Connell 3 Hill K, Boesch C, Goodall J, et al. 2001. Mor- (2005). J Hum Evol 49:654–659. men’s social strategies that we are tality rates among wild chimpanzees. J Hum Hawkes K, Blurton Jones NG. 2005. Human only beginning to explore. Evol 40:437–450. 27 age structures, paleodemography, and the The grandmother hypothesis focuses 4 Howell N. 1979. Demography of the Dobe!- grandmother hypothesis. In: Voland E, Chasio- attention on singular features of human Kung. New York: Academic Press. tis A, Schiefenhovel W, editors. Grandmother- 5 Hill K, Hurtado AM. 1996. Ache life history: hood: the evolutionary significance of the life history, including our exceptional lon- the ecology and demography of a foraging peo- second half of female life. New Brunswick, NJ: gevity, a postfertile life stage in women, ple. New York: Aldine de Gruyter. Rutgers University Press. p 118–140. and both our very short birth intervals 6 Blurton Jones NG, Hawkes K, O’Connell JF. 28 Oeppen J, Vaupel JW. 2002. Broken limits and very late ages of first parturition as 2002. Antiquity of postreproductive life: are to life expectancy. Science 296:1029. there modern impacts on hunter-gatherer post- 29 Gurven M, Kaplan H. 2007. Longevity compared to those of the other great reproductive life spans? Am J Hum Biol 14: among hunter-gatherers: a cross-cultural exam- apes. This hypothesis makes comparisons 184–205. ination. Popul Dev Rev 33:321–365. with the other living hominids a central 7 Levitis DA, Burger O, Lackey LB. 2013. The 30 Hamilton WD. 1966. The moulding of senes- human post-fertile lifespan in comparative evo- line of inquiry. The similarity in oldest cence by natural selection. J Theor Biol 12:12– lutionary context. Evol Anthropol 22:66–79. 45. ages of pregnancy among humans and 8 Hawkes K, O’Connell JF, Blurton Jones NG, 31 Hamilton WD. 1964. The genetical evolution 1 other hominids and similar rates of et al. 1998. Grandmothering, menopause, and of social behaviour. II. J Theor Biol 7:17–52. the evolution of human life histories. Proc Natl 32 Hawkes K, O’Connell JF, Blurton Jones NG. decline in ovarian follicle stocks in Acad Sci USA 95:1336–1339. humans and chimpanzees22 point to sim- 1997. Hadza women’s time allocation, offspring 9 O’Connell JF, Hawkes K, Blurton Jones NG. provisioning, and the evolution of long post- ilar ages of fertility decline in our ances- 1999. Grandmothering and the evolution of menopausal life spans. Curr Anthropol 38:551– tors. The differences in longevity between . J Hum Evol 36:461–485. 577. 10 Alvarez HP. 2000. Grandmother hypothesis 33 Charnov EL. 1993. Life history invariants : humans and other living homi- and primate life histories. Am J Phys Anthropol 1,3,11,27,29 some explorations of symmetry in evolutionary nids suggest that our common 113:435–450. ecology. Oxford: Oxford University Press. ancestor had aging rates more like theirs. 11 Hawkes K. 2003. Grandmothers and the 34 Hawkes K. 2006. Slow life histories and Unexpected differences in aging physiol- evolution of human longevity. Am J Hum Biol human evolution. In: Hawkes K, Paine R, edi- 15(3):380–400. tors. The evolution of human life history. Santa ogy between humans and chimpanzees 12 Kim PS, Coxworth JE, Hawkes K. 2012. Fe: SAR Press. p 95–126. highlight the importance of further com- Increased longevity evolves from grandmother- 35 Charnov EL, Berrigan D. 1993. Why do parisonswithgreatapesifweareto ing. Proc R Soc London B 279:4880–4884. female primates have such long lifespans and understand the evolution of human 13 Kirkwood TBL, Rose MR. 1991. Evolution so few babies? Or life in the slow lane. Evol of senescence: late survival sacrificed for repro- Anthropol 1:191–194. somatic aging. The grandmother hypoth- duction. Philos Trans R Soc B 332:15–24. 36 Kim PS, McQueen JS, Coxworth JE, et al esis continues to stimulate new compara- 14 Williams GC. 1957. Pleiotropy, natural selec- n.d. Grandmothering drives the evolution of tive questions about youngsters as well as tion, and the evolution of senescence. Evolution longevity in a probabilistic model. Submitted. 11(4): 398–411. adults of both sexes. Further, it continues 37 Hawkes K, O’Connell JF, Blurton Jones NG. 15 Medawar PB. 1952. An unsolved problem of 1995. Hadza children’s foraging: juvenile to stimulate more careful appraisal of the biology. London: HK Lewis. dependency, social arrangements, and mobility behavior and effects of those Hadza 16 Graham CE. 1986. Endocrinology of repro- among hunter-gatherers. Curr Anthropol 36: grandmothers102 who supplied initial rea- ductive success. In: Dukelow WR, Erwin J, edi- 688–700. tors. Comparative primate biology, vol 3: 38 Lee R. 2008. Sociality, selection, and sur- sons for inquiry into the distinctive post- reproduction and development. New York: Alan vival: simulated evolution of mortality with menopausal longevity that evolved only R. Liss. p 93–99. intergenerational transfers and food sharing. in our lineage of the hominid radiation. 17 Herndon JG, Paredes J, Wilson ME, et al. Proc Natl Acad Sci USA 105:7124–7128. 2012. Menopause occurs late in life in the captive 39 Lee RD. 2003. Rethinking the evolutionary chimpanzee (Pan troglodytes). Age 34:1145–56. theory of aging: transfers, not births, shape 18 vom Saal FS, Finch CE, Nelson JF. 1994. senescence in social species. Proc Natl Acad Sci Natural history and mechanisms of reproduc- USA 100:9637–9642. ACKNOWLEDGMENTS tive aging in humans, laboratory rodents, and 40 Kaplan H, Robson A. 2002. The emergence other selected vertebrates. In: Knobil E, Neill J, of humans: the coevolution of intelligence and Thanks to Jim O’Connell and Nick editors. The physiology of reproduction. New longevity with intergenerational transfers Proc Blurton Jones for longstanding collab- York: Raven. p 1213–1314. Natl Acad Sci USA 99;10221–10226. oration; Sarah Hrdy for continuing 19 Gosden R, Telfer E. 1987. Numbers of follicles 41 Kaplan H, Gurven M, Winking J, et al. 2010. and oocytes in mammalian ovaries and their allo- Learning, menopause, and the human adaptive guidance; John Fleagle for suggesting metric relationships. J Zool 211:169–175. complex. Ann NY Acad Sci 1204:30–42. clarifications; James Malcolm for call- 20 Moss CJ. 2001. The demography of an Afri- 42 Kachel AF, Premo L, Hublin J-J. 2011. ing attention to Medawar’s prescience; can elephant (Loxodonta africana) population Grandmothering and natural selection. Proc R Christopher Parker, Nicole Herzog, in Amboseli, Kenya. J Zool 255:145–156. Soc London B 278: 384–391. 21 Mizroch SA. 1981. Analyses of some biologi- 43 Hawkes K, Kim PS, Kennedy B, et al. 2011. and Earl Keefe for developing ideas cal parameters in the Antarctic fin whale. A reappraisal of grandmothering and natural about the ancestral role of fire; and Report Int Whaling Commission 31:425–434. selection. Proc R Soc London B 278:1936–1938. 302 Hawkes and Coxworth ARTICLE

44 Bentley GR, Mace R. 2012. Substitute 63 Labrie F, Martel C, Balser J. 2011. Wide dis- 83 Leonetti DL, Chabot-Hanowell B. 2011. The parents: biological and social perspective on tribution of the serum dehydroepiandrosterone foundation of kinship. Hum Nat 22:16–40. alloparenting across human societies: New and sex steroid levels in postmenopausal 84 Hrdy SB. 2009. Mothers and others: the evo- York:Berghahn Books. women: role of the ovary? Menopause 18:30– lutionary origins of mutual understanding. 45 LahdenperaM,LummaaV,HelleS,etal.2004.€ 43. Cambridge, MA: Belknap Press of Harvard Uni- Fitness benefits of prolonged post-reproductive life- 64 Finch CE. 2010. Evolution of the human versity Press. span in women. Nature 428:178–181. lifespan and diseases of aging: roles of infec- 85 Tomasello M, Carpenter M, Call J, et al. 46 Sear R, Mace R. 2008. Who keeps children tion, inflammation, and nutrition. Proc Natl 2005. Understanding and sharing intentions: alive? A review of the effects of kin on child Acad Sci USA 107(suppl 1):1718–1724. the origins of cultural cognition. Behav Brain survival. Evol Hum Behav 29:1–18. 65 Labrie F. 1991. Intracrinology. Moll Cell Sci 28:675–91;691–735. 47 Hill K, Hurtado AM. 2009. Cooperative Endocrinol 78:C113–C118. 86 Tomasello M. 2008. Origins of human com- breeding in South American hunter–gatherers. 66 Lane MA, Ingram DK, Ball SS, et al 1997. munication. Cambridge, MA: MIT Press. Proc R Soc B 276:3863–3870. Dehydroepiandrosterone sulfate: a biomarker 87 Hrdy SB. 2013. Development 1 social selec- 48 Blurton Jones NG, Hawkes K, O’Connell JF. of primate aging slowed by calorie restriction. J tion in the emergence of "emotionally modern" 2005. Hadza grandmothers as helpers: residence Clin Endocr Metab 82:2093–2096. humans. In: Meehan CL, Crittenden AN, edi- data. In: Voland E, Chasiotis A, Schiefenhovel W, 67 Ingram D, Nakamura E, Smucny D, et al. tors. Origins and implications of the evolution editors. Grandmotherhood: the evolutionary sig- 2001. Strategy for identifying biomarkers of of childhood. Santa Fe: SAR Press. nificance of the second half of female life. Piscat- aging in long-lived species. Exp Gerontol 36: 88 Hawkes K. 2013. Primate sociality to human away, NJ: Rutgers University Press. p 160–176. 1025–1034. cooperation: why us and not them? Hum Nat 49 Voland E, Chasiotis A, Schiefenhovel W, 68 Blevins JK, Coxworth JE, Herndon JG, et al 25. DOI 10.1007/s12110-013-9184-x. editors. 2005. Grandmotherhood: the evolution- 2013. Brief communication: adrenal androgens 89 Plooij FX. 1984. The behavioral develop- ary significance of the second half of female and aging: female chimpanzees (Pan troglo- ment of free-living chimpanzee babies and life. Piscataway, NJ: Rutgers University Press. dytes) compared with women. Am J Phys infants. Norwood, NJ: Ablex. 50 Strassmann BL, Garrard WM. 2011. Alter- Anthropol 151:643–648. 90 Tomonaga M, Tanaka M, Matsuzawa T, natives to the grandmother hypothesis. Hum 69 Bernstein RM, Sterner KN, Wildman DE. et al. 2004. Development of social cognition in Nat 22:201–222. 2012. Adrenal androgen production in catar- infant chimpanzees (Pan troglodytes): face rec- ognition, smiling, gaze, and the lack of triadic 51 Voland E, Beise J. 2002. Opposite effects of rhine primates and the evolution of adrenarche. interactions. Jpn Psychol Res 46:227–235. maternal and paternal grandmothers on infant Am J Phys Anthropol 147:389–400. survival in historical Krummhorn.€ Behav Ecol 70 Lasley BL, Chen J, Stanczyk FZ, et al. 2012. 91 Hawkes K. 2006. Life history theory and Sociobiol 52:435–443. Androstenediol complements estrogenic bioac- human evolution: a chronicle of ideas and find- ings. In: Hawkes K, Paine R, editors. The evolu- 52 Lahdenpera€ M, Gillespie DO, Lummaa V, tivity during the menopausal transition. Meno- pause19:650. tion of human life history. Santa Fe: SAR et al. 2012. Severe intergenerational reproduc- Press. p 45–94. tive conflict and the evolution of menopause. 71 Shay JW, Wright WE. 2007. Hallmarks of Ecol Lett 15:1283–1290. telomeres in ageing research. J Pathol 211:114– 92 Morton RA, Stone JR, Singh RS. 2013. Mate 123. choice and the origin of menopause. PLoS 53 Leonetti DL, Nath DC, Hemam NS, et al. Comput Biol 9:e1003092. 2005. Kinship organization and the impact of 72 Haussmann MF, Marchetto NM. 2010. Telo- grandmothers on reproductive success among meres: linking stress and survival, ecology and 93 Jones D. 1996. An evolutionary perspective the matrilineal Khasi and patrilineal Bengali of evolution. Curr Zool 56:714–727. on physical attractiveness. Evol Anthropol 5: 97–109. Northeast India. In: Voland E, Chasiotis A, 73 Tackney J, Cawthon RM, Coxworth JE, Schiefenhovel W, editors. Grandmotherhood: et al. 2013. Blood cell telomere lengths and 94 Muller MN, Thompson ME, Wrangham RW. The evolutionary significance of the second shortening rates of chimpanzee and human 2006. Male chimpanzees prefer mating with old half of female life. New Brunswick: Rutgers females. Submitted. females. Curr Biol 16:2234–2238. University Press. p 194–214. 74 Gomes N, Ryder OA, Houck ML, et al. 2011. 95 Marlowe FW, Berbesque JC. 2012. The 54 Sear R. 2008. Kin and child survival in rural Comparative biology of mammalian telomeres: human operational sex ratio: effects of mar- Malawi. Hum Nat 19:277–293. hypotheses on ancestral states and the roles of riage, concealed ovulation, and menopause on mate competition. J Hum Evol 63:834–842. 55 Hrdy SB. 1999. Mother nature: a history of telomeres in longevity determination. Aging mothers, infants, and natural selection. New Cell 10:761–768. 96 Blurton Jones NG, Marlowe FW, Hawkes K, York: Pantheon Books. 75 Alberts SC, Altmann J, Brockman DK, et al. et al. 2000. Paternal investment and hunter- 2013. Reproductive aging patterns in primates gatherer divorce rates. In: Cronk L, Chagnon N, 56 Fox M, Sear R, Beise J, et al. 2010. Irons W, editors. Adaptation and human behav- Grandma plays favourites: X-chromosome reveal that humans are distinct. Proc Natl Acad Sci USA 110:13440–13445. ior: an anthropological perspective. New York: relatedness and sex-specific childhood mortal- Aldine de Gruyter. p 31–55. ity. Proc R Soc London B 277:567–573. 76 Wrangham RW. 2007. The cooking enigma. In: Ungar P, editor. Evolution of the human 97 Coxworth JE. 2013. First among equals: 57 Cant MA, Johnstone RA. 2008. Reproductive male-male competition among the Bardi of diet: the known, the unknown, and the conflict and the separation of reproductive gen- northwestern Australia and its implications for erations in humans. Proc Natl Acad Sci USA unknowable. Oxford: Oxford University Press. p 308–323. human evolution. Ph.D. dissertation. Salt Lake 105:5332–5336. City: University of Utah. 77 Wrangham RW. 2009. Catching fire: how 58 Koenig A, Borries C. 2012. Hominoid dis- 98 Chapais B. 2008. Primeval kinship: how cooking made us human. New York: Basic persal patterns and human evolution. Evol pair-bonding gave birth to human society. Books. Anthropol 21:108–112. Cambridge: Harvard University Press. 78 Wrangham R, Carmody R. 2010. Human 59 Wrangham RW. 1987. The significance of 99 de Waal F. 2011. What is an animal emo- adaptation to the control of fire. Evol Anthro- African apes for reconstructing human social tion? Ann NY Acad Sci 1224:191–206. evolution. In: Kinzey W, editor. The evolution pol 19:187–199. 100 de Waal FB. 2012. The antiquity of empa- of human behavior: Primate models. Albany: 79 Wrangham RW, Jones JH, Laden G, et al. thy. Nature 336:874–876. SUNY Press. p 55–71. 1999. The raw and the stolen. Curr Anthropol 40:567–594. 101 Darwin C. 1871. The descent of man, and 60 Alvarez HP. 2004. Residence groups among selection in relation to sex. New York: D. hunter-gatherers: a view of claims and evi- 80 Organ C, Nunn CL, Machanda Z, et al. Appleton. dence. In: Berman C, Chapais B, editors. Kin- 2011. Phylogenetic rate shifts in feeding time ship and behavior in primates. Oxford: Oxford during the evolution of Homo. Proc Natl Acad 102 Blurton Jones NG. 2013. Hadza demogra- University Press. p 420–442. Sci USA 108:14555–14559. phy and sociobiology. http://www.sscnet.u- cla.edu/anthro/faculty/blurton-jones/hadza-part- 61 Marlowe FW. 2004. Marital residence 81 Jaeggi AV, Van Schaik CP. 2011. The evolu- 1.pdf. among foragers. Curr Anthropol 45:277–284. tion of food sharing in primates. Behav Ecol 62 Hill KR, Walker RS, Bozic ˇevic´ M, et al. Sociobiol 65:2125–2140. 2011. Co-residence patterns in hunter-gatherer 82 Kramer KL, Ellison PT. 2010. Pooled energy societies show unique human social structure. budgets: resituating human energy allocation VC Science 331:1286–1289. trade-offs. Evol Anthropol 19:136–147. 2013 Wiley Periodicals, Inc.