E

Evolution of Long-Term et al. 2005; Geary and Flinn 2001; Hill Pair-Bonding in Humans et al. 2011). Most species living in groups com- posed of multiple males and multiple females are Kelly Rooker and Sergey Gavrilets promiscuous, while most pair-bonded species live University of Tennessee, Knoxville, TN, USA in groups comprised of only one male and one female. There are no primates other than humans where multiple reproductive pairs live together Synonyms (de Waal and Gavrilets 2013). Nuclear and extended families, as well as parental investment Evolution of mating systems; Human evolution; in offspring, are characteristics found in all human Social monogamy societies (Geary and Flinn 2001); such character- istics clearly set humans apart from their African ape ancestors (de Waal and Gavrilets 2013; Flinn Definition et al. 2005). The question then is how could humans have evolved such a unique combination Long-term pair-bonding occurs when one breeding of social organization and mating system. male and one breeding female share a common In fact, long-term pair-bonding in humans has territory, associating with each other for more always been considered an evolutionary mystery. than just one breeding season regardless of whether Why is this practice so common in human socie- or not they currently have any offspring together. In ties around the world? How did long-term pair- humans, long-term pair-bonding can occur via bonding first evolve in humans? What were the monogamous or polygamous relationships. evolutionary pressures making it beneficial to humans? Although these questions have all been asked repeatedly for decades, even centuries, Introduction there has yet to be any real consensus on their answers. Still, much progress has been made. Humans have previously been called the Monogamy and pair-bonding often get used “uniquely unique species” (Alexander 1990). interchangeably, but “monogamy” can mean Despite there being many different factors con- many different things. For instance, cultural tributing to this uniqueness (e.g., the large brain), monogamy refers to human institutions such as one of the most important is humans having a marriage. Sexual monogamy is in reference only multi-male, multi-female social organization and to sexual relations; one male and one female will a long-term pair-bonding mating system (Flinn mate exclusively with one another, but have no or

# Springer International Publishing Switzerland 2016 T.K. Shackelford, V.A. Weekes-Shackelford (eds.), Encyclopedia of Evolutionary Psychological Science, DOI 10.1007/978-3-319-16999-6_99-1 2 Evolution of Long-Term Pair-Bonding in Humans few interactions outside of that sexual intercourse. Boesch 2003). However, despite the widespread Similarly, genetic monogamy is when all offspring prevalence of extra-pair matings, there still has produced by one male or female in that pair have yet to be a human society with sexual promiscuity that same mother and father. Rates of genetic as the main form of mating system (Chapais 2008). monogamy can be tested in the field by looking Hence, for the rest of this entry, “long-term pair- at the genetic makeup of all offspring produced bonding” will be synonymous with “social monog- (Reichard and Boesch 2003). amy,” thus including both monogamy and polyg- Finally, there is social monogamy or what is amy. For cases where one male and one female do more commonly called pair-bonding.Insocial remain exclusive with each other, this entry instead monogamy, one male and one female exclusively refers to these as strict monogamy. assort with each other. This assortment is not just In addition to all these different types of in terms of copulations or the genetics of their monogamy, the word “monogamy” can also offspring, but also in terms of with whom they imply differences in duration. Long-term monog- live, where they sleep, and most generally where amy implies the one male and one female which they spend their time (Reichard and Boesch 2003). form a pair will remain together across many In other words, under social monogamy, one breed- breeding seasons. Life-long monogamy is then a ing male and one breeding female will share a different term used to describe when a male and common territory, associating with each other for female’s long-term monogamy in fact lasts for the more than just one breeding season, regardless of rest of their lives. Conversely, short-term monog- whether or not they currently have any offspring amy refers to monogamy that typically lasts for together (Lukas and Clutton-Brock 2013). only one breeding season or year; one male and It should be noted that many researchers do not one female will be monogamous to each other, but restrict the word “pair-bonding” to involving only only within that limited amount of time. Cases one male and one female. Rather, the term is often where pairs practice short-term monogamy used more broadly to describe any lasting repro- throughout their lives, but in each instance of ductive relations between particular males and short-term monogamy have a different partner, females. For example, in polygyny where one are instead referred to as practicing serial monog- male is having such lasting reproductive relations amy (Reichard and Boesch 2003). with multiple females, some would say that male For example, many species of birds will be is just having multiple distinct pair-bonds occur- monogamous throughout each breeding season ring simultaneously (Reichard and Boesch 2003). (or other period of time), but then each individual Using this terminology, humans as a whole are will have a different mate come next breeding very much considered having long-term pair- season (Reichard and Boesch 2003). This means bonds since every known human society has prac- such birds are monogamous within breeding sea- ticed either polygamy or monogamy (and hence sons, but not monogamous between breeding sea- long-term pair-bonding) (Chapais 2008). Consid- sons (i.e., serial monogamy). In addition, human ering only the raw number of human societies societies where divorce is common, as well as throughout the world (vs. also the populations of remarriage for both men and women following each of these societies), only about 17 % of such divorce, have mating systems resembling serial societies practice social monogamy. The vast monogamy (Reichard and Boesch 2003). majority of the rest (>80 %) practice polygyny. Monogamy from a female’s perspective often Even among the polygynous societies though, makes a lot of sense. In eutherian mammal repro- only a small minority of the men in each are duction, only females must go through the costly actually polygynous; most males lack sufficient periods of internal gestation and lactation (Opie resources needed to support multiple female et al. 2013). Not only are such activities quite mates (Reichard and Boesch 2003). energetically expensive for females, they make True genetic monogamy is rare in humans, even predation, death, and other injuries to the female in the socially monogamous cases (Reichard and more likely. In addition, due to relatively long Evolution of Long-Term Pair-Bonding in Humans 3 gestation and lactation periods, female eutherian social monogamy (Lukas and Clutton-Brock mammals are also much more limited than males 2013) pales in comparison to the 90 % of birds are in the quantity of offspring they are able to estimated to practice social monogamy. However, have in a lifetime. While males are limited only in these 3–9 % of mammal species are found in the number of copulations they are able to obtain many different clades of mammals, and long- with fertile females, females have their main con- term pair-bonding is believed to have evolved straint as time (Kappeler 2013). independently upward of sixty different times in In a monogamous mating system, having a mammals (Lukas and Clutton-Brock 2013). male remain with/around a female could help the Such pair-bonding in mammals may appear to female in any number of ways, be it protection of be strictly behavioral, but neurophysiology has her and/or her offspring/resources, help in secur- also been shown to play a role. Certain neuropep- ing meat or other valuable resources, and/or help tide pathways (e.g., oxytocin, arginine vasopres- in caring for any offspring produced from the sin, and prolactin) have been shown to have been union. All of these would help loosen the female’s co-opted to form pair-bonding reward pathways main constraint (time) either by allowing a in several species of monogamous mammals, female’s interbirth interval to be shortened (e.g., including primates (Lovejoy 2009). Compared to by a male helping with parental care of the older birds, social and genetic monogamy in mammals offspring), improving her health which leads to are tightly correlated, meaning mammals’ rates of longer times of fertility (e.g., by provisioning the extra-pair copulations are relatively low (Lukas female with valuable foods), and/or lengthening and Clutton-Brock 2013). her life and hence her fertile period (e.g., by Rates of long-term pair-bonding of course vary protecting her from other males or predators) significantly among mammal species. While some (Kappeler 2013). mammalian orders contain zero socially monoga- On the other hand, of critical importance when mous species, about 29 % of primate species are considering the evolution of long-term pair-bond- socially monogamous (Lukas and Clutton-Brock ing is why males would ever play their role in a 2013). In addition, primates are unique among socially monogamous mating system. In eutherian mammals in that social monogamy has evolved mammals, males have much greater potential independently in every major primate clade, and for producing more offspring throughout their there are relatively few transitions away from lives than females (Kappeler 2013), due to the monogamy once it has evolved (Opie et al. 2013). aforementioned female traits of internal gestation Among just anthropoid primates, long-term and lactation (Opie et al. 2013). This means that pair-bonding is believed to have evolved indepen- males mating with only one female throughout dently seven different times, three in Platyrrhini their lives are sacrificing all that “extra” reproduc- (New World monkeys) and four in Catarrhini tive potential. Thus, for long-term pair-bonding (apes and Old World monkeys). Indeed, socially to evolve, there would have to be some fitness monogamous species among anthropoid primates compensation present to make up for that loss today include species as varied as New World in male reproductive potential (Kappeler 2013). monkeys (e.g., marmosets, tamarins), Old World The rest of this entry is primarily concerned with monkeys (e.g., De Brazza’s monkeys, Sumatran considering the many possible forms of such surilis), and apes (e.g., gibbons, humans) (Sillen- “fitness compensation.” Tullberg and Moller 1993). Despite social monog- Although certainly not prevalent, long-term amy not being ubiquitous among primates, its pair-bonding is widespread throughout the animal widespread prevalence indicates there must be kingdom. Long-term pair-bonding is most nota- some significant benefit to species practicing bly found in birds and mammals. Among mam- social monogamy. Although this entry focuses mals, about 3 % of species practice long-term on humans, many of the hypotheses and ideas pair-bonding (Kleiman 1977). Even the newer, presented below could similarly apply to many upper estimate of 9 % of mammals practicing of these other primate species as well. 4 Evolution of Long-Term Pair-Bonding in Humans

It is the aim here to first discuss the transition to female, potential mates are always nearby and long-term pair-bonding: what were human socie- mating attempts likely easier to facilitate (Flinn ties like prior to long-term pair-bonding becoming et al. 2005). the norm? Second, the many hypotheses which On the other hand, group living can also come have been given to explain the evolution of long- with potential costs to all group members, primar- term pair-bonding in humans will be discussed. ily from ecological competition (for resources, Then, both empirical and theoretical work will be mates, etc.) (Hill et al. 2011). Living in groups looked at as they pertain to each of these hypoth- can result in conflict between group members. eses. Finally, other work which takes a broader Such conflict can be energetically costly to those look at the question of long-term pair-bonding in group members, even resulting in injury or death humans will be discussed. if fighting is the outcome. Living in close prox- imity to others can increase theft of resources, food, mates, etc., making group members expend Pathways for the Transition more energy and time guarding such resources to Monogamy in Groups and again potentially fighting over them. Finally, visibility by a predator, disease, and other para- Part of the debate surrounding the evolution of sites becomes more prevalent when living in the long-term pair-bonding in humans is over the higher population densities that come with ancestral social organization and mating system groups. Such risk would be absorbed by all of genus Homo. Were human ancestors monoga- group members, not only potentially resulting in mous? Or promiscuous? Did they live solitary? In increased susceptibility to predation, lasting dis- one-male groups? Or in multi-male, multi-female ability, and death but also potentially in loss of groups? These ancestral-state questions are still resources and other vulnerabilities (Flinn very much debated, with many researchers having et al. 2005). wildly different, and often directly contradicting, Humans’ closest living relative, the chimpan- views (Duda and Zrzavy 2013). zee, is certainly a group-living species. Chimpan- Humans, and many other primate species, are zees live in highly promiscuous multi-male, group living. Such groups of individuals come in multi-female societies. Most evidence does not various sizes, but all with their associated benefits point to monogamy being the ancestral state of and costs (Hill et al. 2011). For instance, groups the chimpanzee-human last common ancestor. afford members protection against both predators Rather, their ancestral state was likely some type and other groups. Such protection could come of promiscuity, with monogamy in humans being from group members’ alarm calling, alerting a derived trait. However, this broad framework other group members of some impending danger, still leaves much room for debate in terms of or in the form of an alliance against a predator or what that ancestral human social organization, invading group. Groups can also facilitate and resulting mating system, specifically looked resource sharing, whether it be with meat, other like (Chapais 2013). The following are the four valuable foods, or other resources. Individuals in main pathways which have been proposed for groups are better guaranteed a steady supply of such an ancestral state. nourishment, with food sharing in groups espe- cially helping with food/resource scarcity in the Promiscuity event of a drought or other severe ecological con- First, the Pan-human last common ancestor could ditions. In the case of species with alloparenting, have been very much like Pan (chimpanzees and group living can help with child-rearing and bonobos). In this case, the ancestral state for allowing females to have more offspring than Homo would be individuals living in fully pro- they would be able to otherwise. Finally, group miscuous (i.e., with no stable mating bonds) living may also help with acquiring mates. By multi-male, multi-female groups. From this back- living in groups with at least one male and one drop, humans would have retained this multi- Evolution of Long-Term Pair-Bonding in Humans 5 male, multi-female social organization, but then female. If the size of the male harem was reduced evolved to a socially monogamous mating system to only female per male, strict monogamy would from the promiscuous Pan-like one (Chapais be the effective result (Chapais 2013). 2013). Strict Monogamy Harems Finally, although considered unlikely by many, it An often-favored alternative to the above is if the is also possible that strict monogamy was the Pan-human last common ancestor had a harem- ancestral mating system for humans. Similar to style mating system, similar to that of gorillas. In gibbons and other primates with such monogamy, this case, the social organization would have monogamous pairs would have been spaced out consisted of groups made up of one male with geographically and not had any kind of group many females (his “harem”) and then lone social organization. What evolved from this back- roaming males outside of any group. The one drop would then be the group social organization: male in the harem would mate with all of the monogamous pairs moving closer to each other, females in the harem and also provide protec- ultimately forming multi-male, multi-female tion/guard them, while the roaming males got groups while still retaining their monogamous few, if any, copulation opportunities. From this mating system. Again, although this pathway is backdrop, humans would have reorganized into certainly a logical transition, there is little evi- a social organization consisting of multiple males dence supporting the chimpanzee-human last and multiple females making up any one group common ancestor being strictly monogamous while still keeping their long-term pair-bonding (Chapais 2013; Sillen-Tullberg and Moller 1993). structure (Chapais 2013). Evidence for this path- way includes humans’ australopithecine ances- tors having large body size sexual dimorphism (larger than either chimpanzees, modern-day Hypotheses for the Evolution of Social humans, or later human ancestors), which is indic- Monogamy ative of fierce one-on-one male competition, as seen in gorillas (Geary and Flinn 2001). In addi- Regardless of from where exactly it came, there is tion, one mathematical modeling approach found no doubt that long-term pair-bonding evolved at this pathway to be the most likely one leading to some point in humans’ past. Similar to the debate humans (Nakahashi and Horiuchi 2012). over the ancestral state of Homo, there is also much debate over how long-term pair-bonding Multilevel Societies first evolved in humans. Again, there have been A third alternative discussed is the idea of many different hypotheses proposed, some multilevel societies. Here, the chimpanzee- directly contradicting each other, and relatively human last common ancestor would have lived little headway made in finding concrete answers with a social organization and mating system sim- as to which hypothesis is the most likely. Despite ilar to hamadryas baboons. Under this framework, no consensus after decades of research, progress males still have harems (i.e., groups made up of has still been made. Note some hypotheses one male and multiple females, all of whom that presented below are unique to humans, while one male mates with and protects), but then mul- others are more widely applicable to many pri- tiple harems live together in larger groups, often mates or even mammals in general. Some hypoth- referred to as “clans.” Under this backdrop, these eses are also highly dependent on one or more of larger groups would already be practicing long- the pathways discussed above, while other term pair-bonding. However, from this humans hypotheses could still be possible regardless of could have even evolved strict (one-male, the specific pathway from which early humans one-female) monogamy, since “clans” could con- evolved. Also note that some hypotheses below tain “harems” of one male and precisely one only work with regard to strict monogamy (these 6 Evolution of Long-Term Pair-Bonding in Humans will be noted), while most do not specify the type intolerant of sharing their ranges. Under these of long-term pair-bonding needed. conditions, males would likely be unable to In general, all the hypotheses discussed below defend access to more than one female. Among can be grouped into four main categories. The first all nonhuman mammals, socially monogamous is ecological. This hypothesis has gained much species today do in fact live at significantly support for the evolution of social monogamy lower densities than solitary species, even after among nonhuman mammals, but is widely controlling for body size differences and despite discredited as being a factor in the evolution of the fact that both socially monogamous and soli- social monogamy in humans. The second group tary species share similar home-range sizes of hypotheses include those having to do with (Lukas and Clutton-Brock 2013). protection, of either the offspring, the mate, or Although considered by many to be a favored one’s self. The third group of hypotheses relates hypothesis for the evolution of social monogamy to food: provisioning of the mate, the offspring, or in nonhuman mammals, this hypothesis is much in cooperation. Finally, the fourth group of less accepted with regard to human evolution hypotheses involves conflict, either strictly since humans did not likely come from solitary between males fighting over mates, group-related, ancestors. Each of the three pathways which or among siblings produced by the same male. It assume humans had non-strictly monogamous should be noted that most of these hypotheses are ancestors (i.e., three of the four discussed above) not mutually exclusive, but rather two or more all centers on early humans living in groups of could have simultaneously contributed selection more than two individuals (Kappeler 2013). How- pressure to the evolution of long-term pair- ever, it is certainly possible that changes in early bonding in humans. human dietary patterns (even within groups) lowered female densities, making it much harder Ecological Hypothesis for one male to guard multiple females. In this The first hypothesis discussed here is the “ecolog- way, a transition from polygyny to strict monog- ical” hypothesis (Komers and Brotherton 1997). amy could still have been facilitated in humans Females are assumed to be solitary and relatively under the ecological hypothesis (Lukas and spaced out geographically (due to ecological com- Clutton-Brock 2013). petition), while males are roaming, looking for mates. However, because of the females’ geo- Protection Hypotheses graphic spacing, it may become increasingly The following three hypotheses all deal with pro- lucrative for a male to stick to just one female’s tection as the main selection pressure on early range. Doing so will save males’ energy resources humans for the evolution of long-term pair- expended in searching for females and also better bonding. guarantee males a mate. If they instead wandered, The first hypothesis under this category deals males would have to travel long distances with protection of one’s offspring, namely, between females with no guarantee of, first, even protecting them against the risk of infanticide. finding another female, and, second, her desiring Infanticide, the practice of killing an infant not to mate with said male once found (Komers and one’s own prior to it being weaned, has been Brotherton 1997). widely documented across primates (van Schaik A recent study found the most likely ancestral and Dunbar 1990). In most instances, infanticide state for all nonhuman mammals is solitary is committed by a new alpha male who has just females and roaming males with overlapping taken over a group. By killing any infants not his ranges (Lukas and Clutton-Brock 2013). They in the group prior to their weaning, he is forcing then found social monogamy to have evolved in their mothers to return to fertility sooner, increas- nonhuman mammals under conditions involving ing the probability of him having offspring with intense female-female competition, low female one or more of those ex-mothers and hence density, and where breeding females were increasing his overall fitness. Such evolutionary Evolution of Long-Term Pair-Bonding in Humans 7 benefit for males has already been well- her gestation period. However, it is in cases where established (van Schaik and Dunbar 1990). such lactation duration does exceed gestation Having an offspring killed in its first year of length that a benefit would be expected for males life is clearly very costly; the biological mother killing other males’ offspring (Lukas and Clutton- and father would each get zero fitness benefit from Brock 2013). This is because female mammals the said offspring, despite the mother (and possi- will not typically have two unweaned offspring bly the father as well) investing much time and concurrently. Hence, if lactation duration exceeds energy into the pregnancy and early life of the gestation length, there is benefit to the male in offspring. As such, it may behoove the father to killing one offspring, so the next (his) can be stick around that offspring up until the point it gets born sooner and the female will still only have weaned in order to protect it from any infanticidal one unweaned offspring at any given time (Opie attacks by other males. In doing so, such a male et al. 2013). has effectively formed the groundwork for having For example, in modern-day humans, one a pair-bond with the mother of his offspring (van study found 50 % of nursing women to return to Schaik and Dunbar 1990). normal estrous cycling within 10 months of giv- Infanticide is quite common among primates, ing birth and 100 % of the women by 20 months, in large part due to altriciality (and in turn long while the cross-cultural average age of weaning in durations of lactation), which makes the infants humans is 30 months (Quinlan 2008). In chim- more susceptible to having infanticide committed panzees, offspring are typically weaned around against them (Opie et al. 2013). These facts would 4 years of age, while females on average only hold for humans as well. In addition, Opie have one offspring every 5 years, meaning female et al. found that in primate phylogeny, infanticide chimpanzees will also only have one offspring to by males typically precedes social monogamy care for at a time (Pillsworth and Haselton 2006). (Opie et al. 2013). They concluded that not only The second hypothesis under the category of is social monogamy in primates much more likely protection deals with mate guarding or a male to evolve in the presence of high male infanticide, protecting his mate. In most formulations, this once social monogamy does evolve, it is more hypothesis centers on the idea of paternity cer- likely for there to be a subsequent decline in the tainty. After mating with a female, males can amount of male infanticide than an increase in the either stay around that one female (“guarding” level of polygyny (Opie et al. 2013). her) or leave her in search of new mates. By There is still much debate over the role of guarding her, a male can prevent other males infanticide in the evolution of long-term pair- from mating with that female (which can increase bonding in nonhuman mammals. Contrary to the his fitness by reducing or eliminating completely Opie et al. work discussed above, other prominent any sperm competition), as well as offer protec- researchers (Lukas and Clutton-Brock 2013) tion to her physical well-being (discussed later). found “the available evidence suggests that male Roaming males attempting to copulate with a infanticide is unlikely to be the principal mecha- guarded female normally face a high cost-to- nism for the evolution of social monogamy in benefit ratio in any fight with a male guard, and mammals.” Despite finding rates of male infanti- the guards typically win these fights (Reichard cide among socially monogamous species (9 %) and Boesch 2003). to be lower than rates among solitary species In general, a male must choose between acquir- (27 %), they concluded from their analysis that ing more mates and increasing his paternity cer- this difference was most likely due to “an inde- tainty with one mate. If conditions make the latter pendent evolution of the two traits” (Lukas and more important to a male’s fitness than the former, Clutton-Brock 2013). one could expect pair-bonding to evolve, as the One key factor considered there was that in male constantly remains near his one mate to help most socially monogamous mammals, the dura- ensure paternity certainty, which would increase tion of female lactation does not typically exceed his fitness (Hawkes 2004). However, note that 8 Evolution of Long-Term Pair-Bonding in Humans mathematical models of this hypothesis have not early humans, as group size and population den- supported mate guarding as a strong enough selec- sity increased (i.e., as humans became increas- tion pressure for long-term pair-bonding to evolve ingly ecologically dominant). As these STDs (e.g., see Gavrilets (2012)). became more widespread and lethal, the costs of Among socially monogamous species, levels having sexual relations with multiple partners of mate guarding vary drastically. Kokko and greatly increased. If such costs increased suffi- Morrell (2005) constructed a theoretical model ciently enough, it could have been better in and found that males will evolve to guard little if terms of a male’s fitness to have fewer mates females are either very faithful or very unfaithful. rather than more. If that was the case, individuals The intuition behind this finding is again that in would have switched to a strategy favoring gaining paternity, males face a trade-off between remaining faithful to one or two mates rather having more mates and increasing paternity cer- than promiscuity, and hence long-term pair- tainty in their one mate. Hence, if a female is bonding could have evolved (Kokko et al. 2002). highly faithful, her mate would be able to gain In addition, using the same reasoning, STDs more mates by guarding little, with little risk to the could influence female choice. A female may paternity certainty of his mate’s offspring. Simi- view a dominant male, or a male who has mated larly if a female is highly unfaithful, her mate with many females, as too risky to mate with due would likely have little to lose by trying to gain to the threat of getting an STD from him. Hence, more mates by guarding little since even if he females may actively begin rejecting mating with spent his time/energy instead guarding her, she such males. Mathematical models have supported would likely still sneak copulations outside this this hypothesis as a way for mating systems to mating bond. In addition, Kokko and Morrell change and long-term pair-bonding to become found that attractive males are predicted to guard more prevalent (Kokko et al. 2002). less, with the intuition being they are more likely to find more mates, and hence paternity certainty Provisioning Hypotheses becomes less crucial in the trade-off (Kokko and These next three hypotheses all fall under the Morrell 2005). category of food or other provisionings as the A related hypothesis to mate guarding, involv- key selection pressure on the evolution of long- ing the direct physical protection of one’s mate term pair-bonding in early humans. mentioned earlier, is the bodyguard hypothesis The first hypothesis under this category, food- (Chapais 2008). Rather than paternity certainty for-sex, deals with males provisioning their mates being the driving selection force behind the evo- or, in particular, males provisioning females in lution of long-term pair-bonding, it could instead exchange for copulations. As hunting became be physical protection of the female. If a group is more important in early human history, and also such that a female must be constantly worried the nutrients obtained from meat more valuable, about harassment/violence from other males, her females could increase their fitnesses by mating (and her future offspring’s) fitness would be with males in order to obtain more high-benefit increased by having the male stick around her. foods or other valuable resources. Females were However, in doing so he would also be increasing able to increase their fitnesses through such ben- his paternity certainty, especially if the harassment eficial provisionings, while males increased their she was receiving was sexual in nature, meaning fitnesses by increasing their likelihoods of the mate guarding and bodyguard hypotheses are obtaining mates. Provided such benefits were largely indistinguishable (Chapais 2008). strong enough, and these exchanges consistently The last hypothesis in this category deals with occurring between the same male and female, protection of one’s self, namely, against the threat long-term pair-bonding could ultimately be of sexually transmitted diseases (STDs) (Kokko selected for (Reichard and Boesch 2003). et al. 2002). This hypothesis hinges on the fact Such exchanges of important foods for copu- that STDs were increasingly problematic among lations are routinely seen in the genus Pan Evolution of Long-Term Pair-Bonding in Humans 9

(chimpanzees and bonobos). “Important foods” day in humans) lactation period in a female mam- include meats and fruits high in fats and/or pro- mal’s life (Quinlan 2008). teins, as well as foods requiring increased search In addition, some researchers believe that as time (Lovejoy 2009). Looking at primates, there is the human brain got bigger and human infants an association between the evolution of social became increasingly dependent on others monogamy and a reliance on relatively scarce (altricial) during the early years of their lives, foods (Lukas and Clutton-Brock 2013). Fruit, mothers were no longer able to care for their compared to foods like gum, bark, and fungi, offspring alone (Benshoof and Thornhill 1979). offers much higher nutritional value while also For instance, children cross-culturally consume being a more limited resource. Indeed, fruit con- more calories than they produce until at least age stitutes a main part of the diet in about 90 % of fifteen; contrast this with chimpanzees who typi- socially monogamous primate species and only cally produce more calories than they consume by 28 % of solitary primate species. Conversely, age five (Conroy-Beam et al. 2015). Over the past foods like gum, bark, and fungi constitute the 4 million years, human ancestors underwent a main part of the diet in about 40 % of socially doubling of the length of their developmental monogamous primate species and 78 % of solitary period (Geary and Flinn 2001). Among modern- primate species (Lukas and Clutton-Brock 2013). day hunter-gatherers, females on average have In addition, a hypothesis related to both the one child every 3–4 years, meaning most women above and the ecological hypothesis, the resource will have multiple dependent offspring to care for defense hypothesis, posits that social monogamy at the same time, unlike in chimpanzees and other could have evolved from a male defending a primates (Pillsworth and Haselton 2006). In addi- highly valued (e.g., due to food resources) terri- tion, parental investment in offspring does not just tory, able to keep all intruders out. In this case, a stop; rather, it often continues all the way through female may want to remain in that territory per- to become “grandparental investment” (i.e., manently, not so much because of the male him- alloparenting) (Conroy-Beam et al. 2015). self, but because of the valuable territory he is able If more than one adult was needed to care for to control (Reichard and Boesch 2003). an offspring, fitness-wise it made sense that that The next hypothesis in this group, paternal “other adult” would be the offspring’s biological care, looks at a different type of male provision- father, since no other adult (besides the mother) ing, this time to his offspring (Benshoof and would have as much genetic material invested in Thornhill 1979). Parental investment is any the offspring. As hunting became more important expenditure benefiting an individual’s offspring in the early human diet, males (the hunters) were at some cost to that individual (even if just an increasingly needed as meat providers. Hence it inability to invest that energy into other fitness could have been some combination of increas- components) (Quinlan 2008). Offspring fitness is ingly altricial offspring, the importance of hunt- increased when a male is around to help provision ing, and bipedality (meaning humans were more the said offspring. Provided this helping male is easily able to carry offspring and/or the provision- indeed the offspring’s biological father, the male ings) that could have helped long-term pair- will be helping his own fitness as well bonding to evolve, provided paternity certainty (Wittenberger and Tilson 1980). 59 % of socially was sufficiently high (Benshoof and Thornhill monogamous mammal species have been found 1979). However, two separate phylogenetic ana- to have regular carrying or provisioning of off- lyses suggest that paternal care is more likely to be spring by males, whereas only three a secondary adaptation occurring either with or nonmonogamous mammal species have such car- after the evolution of long-term pair-bonding, rying or provisioning (Lukas and Clutton-Brock rather than a main driver of this evolution 2013). Mammals in particular may require the (Lukas and Clutton-Brock 2013; Opie extra help from offspring provisioning because et al. 2013). of that special energy-intensive (i.e., 670 kcal/ 10 Evolution of Long-Term Pair-Bonding in Humans

The next hypothesis under this category, Once producing one’s own food became prev- related to both parental investment and food-for- alent among early human societies, there were two sex, is that of economic interdependence (Conroy- main strategies each individual could adopt, pro- Beam et al. 2015). Much research has gone into ducing vs. scrounging. An individual following the evolution of a division of labor among early the producing strategy would be focused on pro- humans, namely, why, how, and when did men ducing his/her own food and then storing such and women become solely responsible for specific food for later consumption. On the other hand, and different tasks or chores. The economic an individual following the scrounging strategy interdependence hypothesis hinges on the fact would instead not be focused on producing his/her that, on average, men have more strength/endur- own food, but rather finding already-produced ance needed for hunting large game than women food elsewhere. Since females were naturally do. In addition, during pregnancy and lactation, less mobile (due to the demands and restrictions women are already expending more energy in of pregnancy, lactation, and parental care) and their daily lives than men and would likely males naturally stronger, a logical division of encounter increased risk during large-game hunts labor resulted where females followed the produc- (Conroy-Beam et al. 2015). For example, a preg- ing strategy and males the scrounging strategy. nant female requires 8–10 % more in caloric However, this means males were often stealing intake than a nonpregnant female and a lactating already-produced foods from females, making it female about 26 % more (Pillsworth and Haselton in females’ best interests to “buy protection” in 2006). any way they could. Since one such form of pay- Under this hypothesis, males and females ment could be the female giving one male her would then engage in economic partnerships, exclusive reproductive rights, long-term pair- whereby each sex would focus on that resource bonds could have evolved under this selection acquisition they could accomplish most effi- pressure (Wrangham et al. 1999). ciently. Hence, males would provide meat for a female and her offspring, and the female would Conflict Hypotheses share with him the foods she had produced or Finally, this last category of hypotheses all deals gathered. Not only does this give females easier with conflict as the main selection pressure for the access to valuable meat, it also provides males evolution of long-term pair-bonding in early with protection in the event that he is unsuccessful humans. in hunts (Conroy-Beam et al. 2015). The first hypothesis in this category, one spe- A related idea to this involves the idea of cifically relating to humans, has to do with the cooking (Wrangham et al. 1999). Cooking food idea of leveling (Chapais 2013). As tool and is beneficial in terms of (1) calorie intake, (2) pro- weapon use became increasingly common across tection from disease or other sickness due to food early humans, male size/strength dimorphism poisoning, and (3) increasing the edible range of became less important in fights. Just because an plant foods through digestibility. While men went alpha male is bigger and stronger than a lower- out to hunt, women were left at the camp and ranked male does not necessarily give that alpha hence became in charge of cooking. While males male an advantage in a physical confrontation if relied on women then for the cooked food each the lower-ranked male has more and/or better night they came back to camp, women in turn also tools/weapons available to him than the alpha relied on the men to help protect these valuable male (Chapais 2013). This meant that any fights cooked foods as well as any foods which had between two now-evenly-matched males were already been gathered but yet to be cooked. It is much more costly to both males involved and in this way that such a relationship benefited both also more likely to end without a winner despite the male and female and could have fostered the the larger cost involved. Hence, it was harder and formation of long-term pair-bonds (Wrangham more costly for an alpha male to protect any et al. 1999). “harem” of females, making pairs of one male Evolution of Long-Term Pair-Bonding in Humans 11 and only one female increasingly common until would evolve quickly in humans’ ancestral line. only strict monogamy was left (Chapais 2008). The above fossil record provides support for Males fighting over mates can be very costly to this idea. a group as a whole. If males’ time and energy are One intriguing alternative related to this being spent in intragroup conflict, there will be hypothesis is if long-term pair-bonding evolved less time and energy available to devote to via cultural evolution (i.e., as an institution via intergroup conflict. As such, this second hypoth- marriage) instead (Henrich et al. 2012). This esis under the category of conflict predicts that means that instead of relying on the evolution of groups with more monogamous pairs will be more gene(s) relating to long-term pair-bonding, it is successful in the long term than promiscuous instead cultural traits (norms) that are important in groups which routinely have intragroup fighting this evolution. Humans and their ancestors are still over mates. This hypothesis posits that since sexually dimorphic in size (as discussed above), cooperation is key to early human survival and an indication that the species should perhaps not long-term pair-bonding supports such coopera- be considered genetically monogamous and that it tion, long-term pair-bonding could have been is culture playing the crucial role instead (de Waal able to evolve (Chapais 2013). and Gavrilets 2013). However, note these cultural There is little doubt that cooperation is in fact hypotheses would be operating on a much shorter key to early human survival. Humans are a coop- time scale than the genetic hypotheses (i.e., tens of erative group-living species, and coalitions are thousands vs. hundreds of thousands of years) often formed in human social competition (Fortunato and Archetti 2010). (Geary and Flinn 2001). Among contemporary Henrich et al. question how one-male, hunter-gatherers, cooperation is shown in ways one-female (strict) monogamy came to be when far exceeding that of other primate species. Such approximately 85 % of societies (meaning raw examples include sharing food across all members number, not considering each society’s population in the group, groups having high levels of size) in the anthropological record allow for alloparental offspring care, cooperative hunting polygyny. They reason that strict monogamy or other food acquisition strategies, construction was favored by cultural evolution because of and maintenance of living spaces, transportation how much it benefits groups (i.e., cultural group of children or resources, contribution to public selection). Similar to the above gene evolution, goods, etc. (Hill et al. 2011). because of those benefits, groups with strict There is some evidence for this hypothesis monogamy would do better in the long term than from the palaeontological record (Lovejoy their counterpart groups with polygamy (Henrich 2009), albeit rather controversial. Fossils 4.4 mil- et al. 2012). lion years old of Ardipithecus ramidis, a species So where do such benefits come from? As occurring relatively soon after the hominin/chim- already discussed, monogamy can lower male- panzee split, indicate reduced sexual size dimor- male competition within a group; it can also phism and reduced upper canine teeth size, both of reduce the number of unmated males within the which are indicative of a decline in male-male group. This is true regardless of the mating system conflict (Lovejoy 2009). Homo erectus appeared of the comparative group (i.e., promiscuous about 1.5–2 million years ago with males about vs. polygynous). Having fewer unmated males 20 % heavier than females (similar to the sexual within a group can reduce crime rates (for murder, dimorphism seen in today’s humans). Contrast rape, assault, robbery, etc.) and other personal this with the 60 % heavier males found in the abuses which are harmful to groups. Males low Australopithecine species from up to 5 million enough in rank/status that they get no mating years ago (Geary and Flinn 2001). If reduction opportunities will have essentially zero fitness in intragroup conflict was a driving force behind and thus, evolutionarily, nothing to lose. This early human evolution, it would be expected that creates males who are more willing to take risks related characteristics like sexual dimorphism in order to get copulations and hence creates 12 Evolution of Long-Term Pair-Bonding in Humans conflicts within their groups (Henrich et al. 2012). opposed to hoe agriculture or pastoralism of Again, groups who impose long-term pair- much of sub-Saharan Africa). In intensive agri- bonding as the social order will do better than culture, land is often the limiting resource, so those that do not. Under this hypothesis, even if splitting land among multiple heirs can only go selection on the actual genes is not strong enough so far. Hence, this hypothesis would be predicted to produce long-term pair-bonding, such evolu- to have much greater selective pressure under tion is still possible, just considering instead evo- intensive agriculture than other forms of agricul- lution on the related cultural traits. ture (Fortunato and Archetti 2010). Finally, the last hypothesis considered here deals with conflict between a male’s various off- spring, especially if produced by different females Comparative Work (Fortunato and Archetti 2010). As it became the norm for humans for property and/or other Most of the research discussed above focuses on resources to be passed on to the next generation, only one of these hypotheses. There is relatively more clearly defined familial lines became little comparative work done pitting such hypoth- increasingly beneficial. A father’s fitness can be eses against each other. One exception to this is decreased by having to split his wealth (resources) Gavrilets (2012). In this paper, Gavrilets among the more offspring he would have by hav- constructed models for several of the above ing more mates. Strict monogamy would allow hypotheses in order to look at how long-term male/female mating bonds to be stable over time pair-bonding could have evolved in early humans’ and hence all the male’s resources to be given to past. In particular, he found that long-term pair- his offspring, all full siblings. In this way, strict bonding was rarely predicted to evolve in (1) the monogamy could simply be a strategic behavior baseline scenario (“communal care”; males allo- on behalf of males and females in allocating their cate some amount of effort toward helping all resources to the next generation (Fortunato and offspring in their group), (2) with mate guarding, Archetti 2010). (3) with food-for-sex, and (4) with “mate provi- In particular, Fortunato and Archetti (2010) sioning” (equivalent to the food-for-sex hypothe- constructed a mathematical model and found sis except males are only able to provision one two conditions which would make strict monog- female and females to be provisioned by one amy (or “monogamous marriage” in their words) male) (Gavrilets 2012). the most beneficial cultural mating system. First is However, long-term pair-bonding was able to when there is unigeniture, meaning only one off- evolve provided two things were present: spring inheriting all of a parent’s wealth, and (1) inequalities between males in their fighting second is when women are more faithful in a abilities (i.e., rank among males in a group) and monogamous bond than a polygynous one. If the (2) differences in females in both “faithfulness” latter is the case, while men would still have that (how faithful any given female will be to her fitness loss from not mating with more than one mate) and choosiness (for a male who will provi- woman, they would get a benefit from that one sion her more). These results imply that it is likely woman being more faithful than otherwise (i.e., that more than just mate guarding or food-for-sex increased paternity certainty). Note the female of is needed in order for long-term pair-bonding to this long-term pair-bond would also benefit, in evolve. Not only is male rank important (since that her male mate would be investing his lower-ranked males may initially be the only resources exclusively in her children for the next ones who do better pursuing a pair-bonding mat- generation (Fortunato and Archetti 2010). ing strategy), but females’ ability to choose mates This hypothesis could help explain why (namely, those who provision more and/or pro- monogamous marriage is so common in societies vide more care) is also important in allowing long- with intensive agriculture (e.g., plow or irrigation term pair-bonding to evolve (Gavrilets 2012). agriculture of much of Europe and Asia, as Evolution of Long-Term Pair-Bonding in Humans 13

The above study looked at individual fitnesses, body of literature spanning over a century, but meaning how different mating systems could have few agreed-upon answers have yet been obtained. evolved by benefiting the individuals within such Some hypotheses have certainly gained more sup- a system (Gavrilets 2012). Another theoretical port throughout the years than others, but there are study instead took a group selection approach, still so many being actively considered. The time looking at how different mating systems could is ripe for researching human origins, and pro- have evolved by benefiting groups as a whole gress will likely be made in this upcoming decade. (Nakahashi and Horiuchi 2012). These authors Being human, it is only natural to ask the question constructed a mathematical model looking at the of from where did this species come and to inves- evolution of the alpha male strategy, as well as tigate those traits that make humans the “uniquely female mating and grouping strategies. In partic- unique species” (Alexander 1990). The human ular, they found that a human-type mating system universal of forming long-term pair-bonds is just (i.e., one in which groups consist of multiple one such trait. females and multiple males, male-male competi- tion is weak, and there is no female promiscuity) can evolve when having a larger group is benefi- Cross-References cial (e.g., in defense against other groups) and when the cost of female promiscuity is large ▶ Ardipithecus Group (e.g., due to the many types of human venereal ▶ Australopithecus Group diseases) (Nakahashi and Horiuchi 2012). ▶ Biparental Care ▶ Breeding Systems ▶ Ecology of Pair-Bond Stability Conclusion ▶ Evidence of Mating Systems ▶ Genetic Polygamy Despite all the research and curiosity surrounding ▶ Harems the evolution of long-term pair-bonding in ▶ Hominid Evolution humans, few concrete answers have yet to be ▶ Homo Group obtained. In particular, there is still debate over ▶ Infanticide the ancestral mating system in early humans, as ▶ Infanticide Risk well as the main selection pressure(s) resulting in ▶ Long-Term Mating humans switching their mating system to one of ▶ Male and Female Distribution long-term pair-bonding. Indeed, researchers still ▶ Male-Male Competition Hypothesis contradict each other over whether the ▶ Male-Male Competition or Male-Provisioning? chimpanzee-human last common ancestor had a ▶ Male-Provisioning Hypothesis promiscuous mating system (like chimpanzees), a ▶ Marriage harem-style polygynous mating system (like ▶ Mating Systems gorillas), or multilevel societies (like hamadryas ▶ Mating Systems (Behavioral Ecology) baboons). Many of the hypotheses for the evolu- ▶ Monogamy tion of long-term pair-bonding in humans have ▶ Obligate Monogamy more support under only one of these mating ▶ Pair-Bonding system ancestral states, making this a related, ▶ Pair-Bonding among Non-Humans and equally important, research question. ▶ Pair-Bonding in Other Mammals (3 %) In terms of how long-term pair-bonding ▶ Paternal Care evolved and whether strict monogamy was a cul- ▶ Phylogenetic Analysis turally and/or genetically evolved trait in humans, ▶ Polyandry these are both still very much unanswered ▶ Polyandry (Behavioral Ecology) research questions. Many different researchers ▶ Polygamy have contributed much to this ever-growing ▶ Polygamy (Behavioral Ecology) 14 Evolution of Long-Term Pair-Bonding in Humans

▶ Polygamy in Humans Henrich, J., Boyd, R., & Richerson, P. (2012). The puzzle ▶ Polygyny of monogamous marriage. Philosophical Transactions – ▶ of the Royal Society B, 367, 657 669. Promiscuity Hill, K., Walker, R., Bozicevic, M., Eder, J., Headland, T., ▶ Sexual Conflict in Mating Strategies Hewlett, B.,..., & Wood, B. (2011). Coresidence pat- ▶ Sexual Size Dimorphism terns in hunter-gatherer societies show unique human ▶ Shared Resource Defense social structure. Science, 331, 1286–1289. ▶ Kappeler, P. (2013). Why male mammals are monoga- Social Monogamy mous. Science, 341(6145), 469–470. ▶ Social Monogamy (Versus Genetic Kleiman, D. (1977). Monogamy in mammals. Quarterly Monogamy) Review of Biology, 52(1), 39–69. ▶ Why Humans Are Unique Kokko, H., & Morrell, L. (2005). Mate guarding, male attractiveness, and paternity under social monogamy. Behavioral Ecology, 16(4), 724–731. Kokko, H., Ranta, E., Ruxton, G., & Lundberg, P. (2002). Sexually transmitted disease and the evolution of mat- References ing systems. Evolution, 56(6), 1091–1100. Komers, P., & Brotherton, P. (1997). Female space use is Alexander, R. (1990). How did humans evolve? Reflections the best predictor of monogamy in mammals. Proceed- on the uniquely unique species. ings of the Royal Society London B, 264, 1261–1270. Benshoof, L., & Thornhill, R. (1979). The evolution of Lovejoy, C. (2009). Reexamining human origins in light of monogamy and concealed ovulation in humans. Jour- Ardipithecus ramidus. Science, 326,74–74e8. nal of Social Biological Structures, 2,95–106. Lukas, D., & Clutton-Brock, T. (2013). The evolution of Chapais, B. (2008). Primeval kinship. Cambridge, MA: social monogamy in mammals. Science, 341(6145), Harvard University Press. 526–530. Chapais, B. (2013). Monogamy, strongly bonded groups, Nakahashi, W., & Horiuchi, S. (2012). Evolution of ape and the evolution of human social structure. Evolution- and human mating systems. Journal of Theoretical ary Anthropology, 22,52–65. Biology, 296,56–64. Conroy-Beam, D., Goetz, C., & Buss, D. (2015). Why do Opie, C., Atkinson, Q., Dunbar, R., & Shultz, S. (2013). humans form long-term mateships? an evolutionary Male infanticide leads to social monogamy in primates. game-theoretic model. In J. Olson & M. Zanna (Eds.), Proceedings of the National Academy of Sciences, (Vol. 51, p. 1–39). New York: Academic Press. 110(33), 13328–13332. de Waal, F., & Gavrilets, S. (2013). Monogamy with a Pillsworth, E., & Haselton, M. (2006). Women’s sexual purpose. Proceedings of the National Academy of Sci- strategies: The evolution of long-term bonds and ences, 110(38), 15167–15168. extrapair sex. Annual Review of Sex Research, 17, Duda, P., & Zrzavy, J. (2013). Evolution of life history and 59–100. behavior in hominidae: Towards phylogenetic recon- Quinlan, R. (2008). Human pair-bonds: Evolutionary func- struction of the chimpanzee-human last common tions, ecological variation, and adaptive development. ancestor. Proceedings of the National Academy of Sci- Evolutionary Anthropology, 17, 227–238. ences, 65, 424–446. Reichard, U., & Boesch, C. (Eds.). (2003). Monogamy: Flinn, M., Geary, D., & Ward, C. (2005). Ecological dom- Mating strategies and partnerships in birds, humans, inance, social competition, and coalitionary arms races: and other mammals. New York: Cambridge University Why humans evolved extraordinary intelligence. Evo- Press. lution and Human Behavior, 26,10–46. Sillen-Tullberg, B., & Moller, A. (1993). The relationship Fortunato, L., & Archetti, M. (2010). Evolution of monog- between concealed ovulation and mating systems in amous marriage by maximization of inclusive fitness. anthropoid primates: A phylogenetic analysis. The Journal of Evolutionary Biology, 23, 149–156. American Naturalist, 141,1–25. Gavrilets, S. (2012). Human origins and the transition from van Schaik, C., & Dunbar, R. (1990). The evolution of promiscuity to pair-bonding. Proceedings of the monogamy in large primates: A new hypothesis and National Academy of Sciences, 109(25), 9923–9928. some crucial tests. Behaviour, 115(1/2), 30–62. Geary, D., & Flinn, M. (2001). Evolution of human paren- Wittenberger, J., & Tilson, R. (1980). The evolution of tal behavior and the human family. Parenting: Science monogamy: hypotheses and evidence. Annual Review and Practice, 1(1–2), 5–61. of Ecological Systems, 11, 197–232. Hawkes, K. (2004). Mating, parenting, and the evolution of Wrangham, R., Jones, J., Laden, G., Pilbeam, D., & human pair bonds. In B. Chapais & C. Berman (Eds.), Conklin-Brittain, N. (1999). The raw and the stolen: Kinship and behavior in primates (pp. 443–473). New cooking and the ecology of human origins. Current York: Oxford University Press. Anthropology, 40(5), 567–594.