Mark V. Flinn and Carol V. Ward
The human family seems to follow a typical mammalian pattern: intense maternal care including breastfeeding of an altricial (helpless) offspring, with some support from an assortment of other relatives—siblings, aunts, fathers, and the like. Beyond the shared mammal/primate commonality, however, humans have some highly unusual traits. The extent and duration of offspring care is exceptional, and unique in the huge informational transfer via language. We are the only species characterized by the combination of stable breeding bonds, extensive paternal care in multimale groups, extended bilateral kin recognition including life-long brother-sister relationships, grandparenting, and controlled exchange of mates among kin groups. These characteristics are important for theoretical and pragmatic understanding of family relationships and child development; their evolution presents one of the great challenges for science. Here we consider clues from the fossil record and the physiological mechanisms that underpin central aspects of our sociality.
Hormones and neurotransmitters help shape important aspects of our lives, including growth, differentiation, sexuality, physiology, emotion, and cognition. From romantic thoughts to jealous rage, from the release of gametes to lactation and parent-offspring bonding, the molecules produced and released by tiny and otherwise seemingly insignificant cells and glands orchestrate our reproductive strategies in extraordinary ways (see Roney, Chapter 46, this Handbook, Volume 2).
Our endocrine and neuroendocrine systems may be viewed as complex sets of mechanisms designed by natural selection to communicate information among cells and tissues. This chapter focuses on an area of particular importance for evolutionary psychology: the behavioral endocrinology of human sociality. Steroid and peptide hormones, associated neurotransmitters, and other chemical messengers guide mating and parental behaviors of mammals in many important ways (Bridges, 2008; Curtis & Wang, 2003; Rosenblatt, 2003; Young & Insel, 2002). Cross-species comparisons among primates require careful analysis (Bercovitch & Ziegler, 2002; Fernandez-Duque, Valeggia, & Mendoza, 2009) because of the apparent rapid evolutionary changes in patterns of reproductive behaviors and increased phenotypic flexibility involving intricate mental processes. Homo sapiens present special problems in these regards (Fisher, 2004; Maestripieri, 1999; Marler, Bester-Meredith, & Trainor, 2003; Rilling, 2013; Wynne-Edwards, 2001, 2003).
Here we first provide a theoretical scenario for the evolution of human patterns of mating and parenting behaviors. We test our model by examining the phylogenetic trajectories of associated traits such as sexual dimorphism and life history stages from the hominin fossil record. We then turn to a description and functional analysis of the endocrine mechanisms that may influence these remarkable reproductive behavioral characteristics of our species.
Human childhood may be viewed as a life history stage that is necessary for acquiring the information and practice to build and refine the mental algorithms critical for negotiating the social relationships that are key to success in our species (Geary & Flinn, 2001; Hrdy, 2009; Konner, 2010; Muehlenbein & Flinn, 2012). Mastering the social environment presents special challenges for the human child. Social competence is difficult because the targets (other children and adults) are constantly changing and similarly equipped with theory of mind and other cognitive abilities (Flinn, 2006c; 2013b; Flinn & Alexander, 2007).
The family environment is a primary source and mediator of the ontogeny of information processing abilities, including social competencies and group cooperation. Human biology has been profoundly affected by our evolutionary history as unusually social creatures, immersed in networks of family, kin, and dynamic, intercommunity coalitions. Human kinship systems appear unique in the universal recognition of both bilateral (maternal and paternal) and multigenerational structure, with a general trend for co-residence of male kin, but a dozen or more major variants (Chapais, 2008, 2013; Davis & Daly, 1997; Flinn & Low, 1986; Walker, Flinn & Hill, 2010; Walker, Hill, Flinn & Ellsworth, 2011). These aspects of human kinship link families into broader cooperative systems, and provide additional opportunities for alloparental care during the long social childhood. Five species-distinctive characteristics stand out as unusually important in this regard: (1) fathering, that is, extensive and specific parental investment by males; (2) life-long sibling bonds, including sister-brother relationships; (3) complex pair-bond relationships between mates; (4) grandparenting; and (5) networks of kinship that extend among communities and involve affinal (ties by marriage) and consanguineal (ties by blood) relationships. These five distinctive social characteristics appear linked with a suite of other unusual human traits, including concealed (or “cryptic”) ovulation, physically altricial but mentally precocial infants, lengthy child development, female orgasm, and menopause (Flinn, Quinlan, Ward, & Coe, 2007).
Hormones are involved in the development (ontogeny) and regulation of these and other components of reproduction, including the neurobiology that underpins the associated psychological competencies (e.g., Bartels & Zeki, 2004; Donaldson & Young, 2008). Understanding the proximate causes, phylogenetic relations, and adaptive functions of the hormonal and neurotransmitter mechanisms may provide important steps toward reconstructing the evolutionary history of our (human) unusual patterns of mating and parenting and their variability in different environmental contexts.
The altricial (helpless) infant is indicative of a protective environment provided by intense parental and alloparental care in the context of kin groups (Alexander, 1987; Chisholm, 1999; Flinn, 2004, 2006b; Flinn & Ward, 2004; Hrdy, 1999, 2004). The human baby does not need to be physically precocial. Rather than investing in the development of locomotion, defense, and food acquisition systems that function early in ontogeny, the infant can work instead toward building a more effective adult phenotype. The brain continues rapid growth, and the corresponding cognitive competencies largely direct attention toward the social environment. Plastic neural systems enable adaptation to the nuances of the local community, such as its language (Alexander, 1990a; Bjorklund & Pellegrini, 2002; Bloom, 2000; Geary & Bjorklund, 2000; Geary & Huffman, 2002; Small, 1998, 2001). In contrast to the slow development of ecological skills of movement, fighting, and foraging, the human infant rapidly acquires skill with the complex communication system of human language (Pinker, 1994) and other social competencies such as facial recognition (de Haan, Johnson, & Halit, 2003), eye contact (Farroni, Mansfield, Lai, & Johnson, 2003), and smiling (Bornstein & Arterberry, 2003). The extraordinary information-transfer abilities enabled by linguistic competency provide a conduit to the knowledge available in other human minds. This emergent capability for intensive and extensive communication potentiates the social dynamics characteristic of human groups (Dunbar, 1997, 2004) and provides a new mechanism for social learning and culture. The recursive pattern recognition and abstract symbolic representation central to linguistic competencies may facilitate the open-ended, creative, and flexible information processing characteristic of humans—especially of children (Flinn & Ward, 2004; cf. Ranganath & Rainer, 2003).
The advantages of intensive parenting, including paternal protection and other care, require a most unusual pattern of mating relationships: moderately exclusive pair bonding in multiple-male groups. No other primate (or mammal) that lives in large, cooperative multiple-reproductive-male groups has extensive male parental care, although some protection by males is evident in multimale troops of baboons (Buchan, Alberts, Silk, & Altmann, 2003), and extensive care is provided by males in small monogamous family groups in indris, marmosets, tamarins, night monkeys, titi monkeys, and, to a lesser degree, gibbons (Fernandez-Duque et al., 2009). Although some group-living species of birds have paternal care, there appear to be special mechanisms enhancing confidence of paternity (e.g., mate guarding and the lack of long gestation periods), and they lack the coalitionary cooperation characteristic of humans. Among primates, competition for females in multiple-male groups usually results in low confidence of paternity (e.g., chimpanzees). Males and females forming exclusive pair bonds in multiple-male primate groups would provide cues of nonpaternity for other males and hence place their offspring at higher risk for infanticide (Hrdy, 1999). Paternal care is most likely to be favored by natural selection in conditions where males can identify their offspring with sufficient probability to offset the costs of investment (Alexander, 1974; see Geary, Chapter 20, this volume), although reciprocity with potential mates is also likely to be involved (Buss, 1994; Smuts, 1985). Humans exhibit a unique “nested family” social structure, involving complex reciprocity among males and females embedded in kin networks that restricts direct competition for mates among group members. It is difficult to imagine how this system could be maintained in the absence of the unusual human trait of concealed or “cryptic” ovulation (Alexander, 1990b; Alexander & Noonan, 1979). Although many other primates lack estrus swellings and other obvious visual signals of female reproductive condition (Pawlowski, 1999; Sillén-Tullberg & Møller, 1993), humans appear especially oblivious to the timing of ovulation, although frequency of intercourse (Wilcox et al., 2004), mate-guarding activities (Flinn, 1988), and mate choice discrimination (Gangestad, Simpson, Cousins, Garver-Apgar, & Christensen, 2004; Gangestad, Thornhill, & Garver-Apgar, Chapter 14, this volume) may be higher during midcycle in some conditions.
Human social relationships are especially complex because they involve extensive coalitions. We are extraordinarily cooperative, most exceptionally and importantly in regard to competition with other groups (Alexander, 2006; Bowles, 2009; Flinn, Geary, & Ward 2005). Humans are unique in being the only species that engages in group-against-group play (Alexander, 1990b) including team sports. This trait is cross culturally universal, emerges early in child development, and often is the object of tremendous collective effort. Human groups tend to be male philopatric (men reside in the group in which they were born, although they may also emigrate), resulting in extensive male kin alliances, useful for competing against other groups of male kin (Chagnon, 1988; Flinn, Ponzi, & Muehlenbein, 2012; LeBlanc, 2003; Macfarlan, Walker, Flinn, & Chagnon, 2014; Wrangham & Peterson, 1996). Patterns of kinship residence, however, are variable (Murdock, 1949) and associated with different aspects of mating and marriage systems (Flinn & Low, 1986; Rohner & Veneziano, 2001; Walker et al., 2013). Females also have complex social networks, but usually are not involved directly in the overt physical aggression and alliances characteristic of intergroup relations (Campbell, 2002; Geary & Flinn, 2002; for an insightful case of indirect competitive activities by females, see Biella, Chagnon, & Seaman, 1997; for comparisons with other primates, see Bissonnette et al., 2015).
Across extant primates, a long developmental period and intensive parenting are associated with a long life span (Allman & Hasenstaub, 1999; Leigh, 2004; van Schaik & Deaner, 2003). One unique feature of the life history and long life span of women is menopause. Menopause results in an extended period during which women can invest in the well-being of their later born children as part of a potential adaptation that enables the long-term investment in a smaller number of children and other relatives such as grandchildren. It allows them to focus on children they have already produced, avoiding the costs of additional pregnancies at a time when their health and the likelihood of their survival to the end of later-born children's dependency are diminishing (Alexander, 1974; Hawkes, 2003; Williams, 1957). The increasing probability of mother's death with age has especially significant effects on the reproductive value of later-born children if long-term maternal investment is important. Orphans have low reproductive value in many societies. A parallel is found in some preindustrial societies, whereby parents sometimes commit infanticide to reduce the risks to their older children (Daly & Wilson, 1988; Hill & Hurtado, 1996). Infanticide, as well as reduced fertility associated with breastfeeding and increasing age (Ellison, 2001) enables parents to reduce the number of dependent offspring and direct more parental investment to older children. When this pattern is combined with a substantial increase in the length of the developmental period, menopause follows as a logical evolutionary adaptation that serves the same function, that is, to reduce the number of dependent children and thus free parental resources that can be invested in a smaller number of children and other kin. Empirical tests demonstrating such advantages, however, have proven difficult (Hill & Hurtado, 1996; Hill & Kaplan, 1999; cf. Hawkes, 2003; Hawkes, O'Connell, Blurton Jones, Alvarez, & Charnov, 1998).
Men, with different, less risky parental activities, would not have been subject to the same selective pressures for terminating reproductive potential, although they, too, may have been selected to adjust reproductive behavior from mating to parenting with increased age (Draper & Harpending, 1988). From this perspective, older females may have had important effects on the success of their developing children, perhaps in part because of the importance of their accumulated knowledge for negotiating the social environment. Socially skilled and well-connected older mothers and grandmothers may have been especially valuable teachers of social and political wisdom, with associated reproductive benefits (Alexander, 1990b; Caspari & Lee, 2004; Coe, 2003; cf. O'Connell, Hawkes, & Blurton Jones, 1999). In short, the doubling of the maximum life span of humans, involving an increased period of prereproductive development on the one hand and an increased period of postreproductive parental and kin investment on the other, suggests the importance of parent-offspring relationships for acquiring and mastering sociocompetitive information (Bjorklund & Pellegrini, 2002; Flinn & Ward, 2004; Geary, 2005).
These characteristics of the human family—extensive biparental and kin care, physically altricial but linguistically and cognitively precocial infants, lengthy childhood and adolescence, concealed ovulation, variably exclusive pairbonds in multiple-male coalitionary groups, and menopause—are a unique combination of traits with associated morphological, physiological, and psychological mechanisms (Flinn, Geary, & Ward, 2005). In the following section, we review the paleontological evidence of the selective pressures that produced this complex set of adaptations.
The temporal sequence of changes in hominin anatomy documented in the fossil record provides evidence of the sequence of morphological changes that occurred in human evolution. Unfortunately, it is difficult to directly infer hominin social structures and associated neurobiological and endocrinological mechanisms from fossils. Some evidence comes from changes in the pattern of human sexual dimorphism and shifts in life-history strategies that would impact social interactions, in particular, reduction in the magnitude of body size sexual dimorphism, threefold increase in brain volume, near doubling of the length of the developmental period, and disappearance of related species of hominins. Covariation among these variables and social and ecological differences across living primates provide data from which inferences can be made about the nature of social dynamics in human evolution (Alexander, Hoogland, Howard, Noonan, & Sherman, 1979; Anton, 2003; McHenry, 1994a, 1994b; Dunbar, 1998; Foley, 1999; Plavcan, van Schaik, & Kappeler, 1995), although associated models may not be definitive (Plavcan, 2000; 2012a).
The best indicators of the increasing stability of male-female pair bonds and associated male coalitionary behavior in the fossil record are sexual dimorphism and life-history patterns. Reduced body size dimorphism is associated with both monogamy (Plavcan, 2000, 2001) and male coalitionary behavior (Pawlowski, Lowen, & Dunbar, 1998; Plavcan & van Schaik, 1997; Plavcan et al., 1995) in extant primates. Although the large canine size dimorphism that characterizes all living and fossil great apes had greatly diminished in Australopithecus (Manthi, Plavcan & Ward, 2012; Ward, Leakey, & Walker, 2001; Ward, Walker, & Leakey, 1999), the reduced body mass dimorphism typical of modern humans did not occur until sometime during the evolution of Homo erectus (Antòn, 2003; McHenry, 1992a, 1992b, 1994a, 1994b; Plavcan, 2012a, 2012b; Rightmire, Van Arsdale, & Lordkipanidze, 2008).
It is tempting to assume that the behavioral characteristics of the ancestor common to the australopithecine species and humans were similar to those observed in modern chimpanzees or bonobos (de Waal & Lanting, 1997; Kano, 1992; Wrangham, 1999; Wrangham & Peterson, 1996; Zihlman, Cronin, Cramer, & Sarich, 1978). This appears a reasonable assumption in some respects, as relative brain sizes of chimpanzees, bonobos, and australopithecines are very similar (McHenry, 1992a, 1992b). In addition, sexual dimorphism in body weight is about 20% for chimpanzees and bonobos (Goodall, 1986; Kano, 1992), as are lean mass and skeletal size dimorphsim in humans (Plavcan, 2012a). Thus, it might appear that the large multimale, multifemale group structures characterizing all three species would have been found in the last common ancestor and thus in earliest hominins. Chimpanzees and humans display coalitional aggression (Wrangham, 1999), and although this is not documented for the less studied bonobos, it has been hypothesized to be a homologous trait shared with the common ancestor of chimpanzees and humans (Wrangham & Peterson, 1996).
Size dimorphism was substantially greater in Australopithecus than in Pan or Homo, although less than in gorillas and orangutans (Gordon, Green & Richmond, 2008; Harmon, 2006; Kimbel & Delezene, 2009; Lockwood, 1999; Lockwood, Richmond, Jungers & Kimbel 1996; Lockwood, Menter, Mogg-Cecci & Keyser, 2007; McHenry, 1992b; Plavcan, 2012b; Richmond & Jungers, 1995; Ward et al., 1999, 2001; Wood & Constantino, 2007; but see Reno, Meindl, McCollum, & Lovejoy, 2003; Reno, McCollum, Meindl & Lovejoy, 2010 for an alternate interpretation of A. afarensis,). The contrast suggests that reproductive strategies of australopithecines may have differed in important respects from that of male chimpanzees, bonobos, and humans. Australopithecus body mass dimorphism suggests that these early hominins were polygynous, as significant mass dimorphism is not associated with monogamy in any extant primate (Plavcan, 2001). Body mass dimorphism is inconsistent with both monogamy and extensive coalitionary behaviors in extant primates (Plavcan, 2000; Plavcan & van Schaik, 1997). Therefore, the social structure of Australopithecus was unlikely to have been characterized by either monogamy or extensive male coalitions (Plavcan, 2012b).
At some point during the evolution of Homo erectus, body-size sexual dimorphism became reduced to near-modern human levels. The reduction in sexual dimorphism resulted in spite of a slight increase in male size, because of an even more substantial increase in female body size (McHenry, 1994a; Plavcan, 2012a). Body mass dimorphism in early H. erectus is difficult to estimate accurately, but disparities in size and robusticity among even early H. erectus crania are less than in australopithecine species, signaling a reduction in body size sexual dimorphism. By the early mid-Pleistocene (approximately 800 k), body mass dimorphism was similar to that found in modern humans (McHenry, 1994a; Ruff, Trinkaus, & Holliday, 1997), consistent with either an increase in pair bonding and/or male coalitionary behaviors.
Changes in social behavior accompanying the shift in mating and parenting strategies are likely to have presented novel cognitive challenges involving complex reciprocity among coalition members. Unlike gorillas, with one-male breeding groups, and chimps, with promiscuous mating and little male parental behavior, at some point, the evolving hominids were faced with the difficulties of managing increasingly exclusive pair bonds in the midst of increasingly large coalitions of potential mate competitors. These behavioral changes would be consistent with the documented decreases in dimorphism.
Prolongation of childhood, including secondarily altricial infants born early in their ontogenies coupled with extended juvenile periods, an adolescent growth spurt, and delayed maturation relative to apes (Bogin, 1991, 1999), seems to have broadly coevolved with changes in sexual dimorphism and reproductive behaviors. The first major changes in hominin infant altriciality probably occurred in later Homo erectus, possibly concurrent with changes in sexual dimorphism and cranial capacity—that is, more recently than 1.5 mya (Antòn & Leigh, 2003; Nelson, Thompson, & Krovitz, 2003). Female pelvic dimensions are constrained by mechanical-locomotor as well as thermoregulatory constraints, so birth canal size was not greatly expanded over australopithecine levels (Begun & Walker, 1993; Ruff, 1995), yet adult brain sizes were nearly doubled even in early Homo. This means that, to have appropriate neonatal proportions relative to the size of the mother's pelvic inlet, infants must have been born at a relatively small size and were relatively altricial early (Martin, 1990; Portman, 1941; Rosenberg & Trevathan, 1996) with rapid rates of brain growth (Antòn & Leigh, 2003; Martin, 1983). They do not appear to have attained large adult brain size simply by prolonging overall growth (Deacon, 1997; Dean et al., 2001; Leigh, 2004). Increasingly altricial infants would have required more intensive parenting by the mother, and, given the decrease in sexual dimorphism occurring at this time, which may indicate pair bonding, perhaps parental care by the father and/or alloparents (Flinn & Ward, 2004; Rosenberg, 1992; Rosenberg & Trevathan, 1996).
Despite these ontogenetic shifts associated with the timing of birth, delayed maturation does not appear to have occurred until later in human evolution (summary in Nelson et al., 2003). Dental development is coupled to life history variables such as age at sexual maturity, and thus can be used to infer the timing of important life history stages. Early Homo erectus appears to have had relatively rapid development, similar in rate to Australopithecus and great apes, whereas that of modern humans is much slower (Dean et al., 2001). Rates of development of the dentition as determined by studies of tooth calcification patterns support the hypothesis that all Pliocene and Pleistocene hominin growth was rapid, more similar to that of apes than of modern humans (reviews in Lacruz & Ramírez Rossi, 2010; Smith, 2008, 2012), until very recently, possibly even until the appearance of anatomically modern Homo sapiens (Smith et al., 2010). However, it is important to note that the relation between dental development and development of the brain and other aspects of hominin biology are uncertain (Smith, 2012). Coincident with its rapid rate of development, early H. erectus is predicted to have lacked a humanlike adolescent growth spurt, based on the fact that the single known juvenile skeleton, KNM-WT 15000, appears to have had a more rapid rate of dental development than that of his postcranial skeleton when compared with humans (Antòn & Leigh, 2003; Smith, 1993). There are no comprehensive data on rates of child development for hominins between 1.6 million and 60 thousand years ago, but the single Neandertal specimen examined by Dean and colleagues (2001) was modern in its developmental trajectory, indicating a humanlike extended childhood had occurred by this time. A modern human pattern of dental development was present by 800 k (Bermudez de Castro, Rosas, Carbonee, Nicolás, Rodríguez, & Arsuaga, 1999; Bermudez de Castro, Ramírez Rossi, Marinón-Torres, Sarmiento Pérez, & Rosas, 2003), but this may or may not imply a similar rate (Dean et al., 2001). Relatively large brains in some Neandertals compared to their dental development stages (Dean, Stringer, & Bromage, 1986) may reflect the overall larger brains of at least some individuals, rather than significant maturational differences. If it does, it might be reasonable to hypothesize that the human adolescent growth spurt was already in place by this time as well (Bermudez de Castro et al., 2003). Neandertals and modern humans probably shared similar stages of development, including an adolescent growth spurt that would have been present in their mutual ancestry, perhaps by 500 kya (Krovitz, 2003). Still, their dental development patterns are more similar to those of earlier hominins than modern humans (Smith, 2008, 2012; Smith et al., 2010). Longevity appears to have gradually increased from Australopithecus to modern humans with a higher proportion of individuals living to old age in the past 50,000 years (Caspari & Lee, 2004). If ecological dominance reduced mortality from extrinsic causes, this would allow for selection for delayed reproduction and extended life histories (Chisholm, 1999; Stearns, 1992; Williams, 1957). Taking all the data together, it appears that the evolution of altriciality may have begun after the initial brain expansion but that delayed maturation and an adolescent growth spurt may have evolved later in human evolution, perhaps as brain size increase continued throughout the Pleistocene.
Thus, it appears that modern human social structures, and likely human family structures, developed gradually during the early to mid-Pleistocene. Integrated adaptations included more altricial infants, delayed maturation, increasingly stable mating relationships between males and females, increasing paternal and alloparental care of offspring, and more significant nonkin coalitionary behaviors. All of these changes roughly co-occurred with brain size expansion, which began increasing with early Homo and continued through the mid-Pleistocene where it reached modern human levels (Lee & Wolpoff, 2003). Evidence for coevolution among all these variables broadly supports a model in which increasing social complexity favored sociocognitive competencies, necessitating a longer childhood and more parental care of children (Figure 24.1).
Figure 24.1 Cranial Capacity (cubic centimeters) and Ratio of Male-to-Female Body Mass in Hominins.
In the following sections, we examine the hormonal mechanisms that may be involved with the ontogeny and regulation of this unique combination of life history, reproductive, and social traits suggested by the fossil and comparative primate evidence.
The constellation of behaviors associated with the human family and the dynamics of social competition described in previous sections require complex regulatory systems. In this section, we first briefly review the potential mechanisms for human pairbonding, maternal and paternal attachment to offspring, kin attachment, and male coalitions. We then turn to a more detailed analysis of how the neuroendocrine stress response system functions to enable acquisition of social competencies during childhood in the context of the human family environment.
The chemical messenger systems that orchestrate the ontogeny and regulation of sexual differentiation, metabolism, neurogenesis, immune function, growth, and other complex somatic processes tend to be evolutionarily conservative among primates and more generally among mammals. Hence rodent and nonhuman primate models provide important comparative information about the functions of specific human neuroendocrine systems, for which we often have little direct empirical research. It is the particular balance of human mechanisms and abilities that is unique and reflects the history of selection for complex social interactions that shaped the human lineage.
Some of the most precious of all our human feelings are stimulated by close social relationships: a mother holding her newborn infant for the first time, brothers reunited after a long absence, or lovers entangled in each other's arms. Natural selection has designed our neurobiological mechanisms, in concert with our endocrine systems, to generate potent sensations in our interactions with these most evolutionarily significant individuals. We share with our primate relatives the same basic hormones and neurotransmitters that underlie these emotional gifts. But our unique evolutionary history has modified us to respond to different circumstances and situations; we are rewarded and punished for somewhat different stimuli than our phylogenetic cousins. Chimpanzees and humans delight in biting into a ripe, juicy mango. But the endocrine, neurological, and associated emotional responses of a human father to the birth of his child (e.g., Storey, Walsh, Quinton, & Wynne-Edwards, 2000) are likely to be quite different from the responses of a chimpanzee male. Happiness for a human (Buss, 2000) has many unique designs, such as romantic love (Fisher et al., 2002), that involve shared endogenous messengers from our phylogenetic heritage.
Attachments are central in the lives of the social mammals. Basic to survival and reproduction, these interdependent relationships are the fabric of the social networks that permit individuals to maintain cooperative relationships over time. Although attachments can provide security and relief from stress, close relationships also exert pressures on individuals to which they continuously respond. It should not be surprising, therefore, that the neuroendocrine mechanisms underlying attachment and stress are intimately related to one another. And although more is known about stress response systems than affiliative systems, some of the pieces of the puzzle are beginning to fall into place.
The mother-offspring relationship is a core aspect of mammalian social life. The biochemistry of this intimate bond was also selected to serve in primary mechanisms regulating bonds between mates, paternal care, the family group, and even larger social networks (Fisher et al., 2002; Hrdy, 1999). Although a number of hormones and neurotransmitters are involved in attachment and other components of relationships, the two peptide hormones, oxytocin (OT) and arginine-vasopressin (AVP), appear to be primary (Carter, 2002; Curtis & Wang, 2003; Lim et al., 2004; Young & Insel, 2002), with dopamine, cortisol, and other hormones and neurotransmitters having mediating effects.
The hypothalamus is the major brain site where OT and AVP, closely related chains of nine amino acids, are produced. From there they are released into the central nervous system (CNS) as well as transported to the pituitary where they are stored until secreted into the bloodstream. OT and AVP act on a wide range of neurological systems, and their influence varies among mammalian species and stage of development. The neurological effects of OT and AVP appear to be key mechanisms (e.g., Bartels & Zeki, 2004; Donaldson & Young, 2008) involved in the evolution of human family behaviors. The effects of OT and AVP in humans are likely to be especially context dependent, because of the variable and complex nature of family relationships.
Along with OT and AVP, prolactin, estrogen, and progesterone are involved in parental care among mammals (Insel & Young, 2001). The involvement of these hormones varies across species and between males and females. The effects of these hormones are influenced by experience and context. Among rats, for example, estrogen and progesterone appear to prime the brain during pregnancy for parental behavior. Estrogen has been found to activate the expression of genes that increase the receptor density for OT and prolactin, thus increasing their influence (Young & Insel, 2002).
Experience also influences parental behavior and the hormonal activity associated with it. In animal studies, a significant body of evidence demonstrates that early life experience influences later parental behavior (Champagne & Meaney, 2001; Fairbanks, 1989). And a number of studies demonstrate that this experience influences the neurohormonal biology involved in the expression of maternal care (Barrett & Fleming, 2010; Fleming, O'Day, & Kraemer, 1999). The hypothalamic-pituitary-adrenal (HPA) system of offspring during development is influenced by variation in maternal care, which then influences their maternal behavior as adults. Such changes involve the production of, and receptor density for, stress hormones and OT.
HPA-modulated hormones and maternal behavior are related in humans during the postpartum period (Fleming, Steiner, & Corter, 1997). During this time, cortisol appears to have an arousal effect, focusing attention on infant bonding. Mothers with higher cortisol levels were found to be more affectionate, more attracted to their infant's odor, and better at recognizing their infant's cry during the postpartum period.
fMRI studies of brain activity involved in maternal attachment in humans indicate that the activated regions are part of the reward system and contain a high density of receptors for OT and AVP (Bartels & Zeki, 2004; Fisher, 2004). These studies also demonstrate that the neural regions involved in attachment activated in humans are similar to those activated in nonhuman animals. Among humans, however, neural regions associated with social judgment and assessment of the intentions and emotions of others exhibited some deactivation during attachment activities, suggesting possible links between psychological mechanisms for attachment and management of social relationships. Falling in love with a mate and offspring may involve temporary deactivation of psychological mechanisms for maintaining an individual's social “guard” in the complex reciprocity of human social networks. Dopamine levels are likely to be important for both types of relationship but may involve some distinct neural sites. It will be interesting to see what fMRI studies of attachment in human males indicate because that is where the most substantial differences from other mammals would be expected. Similarly, fMRI studies of attachment to mothers, fathers, and alloparental caretakers in human children may provide important insights into the other side of parent-offspring bonding.
Paternal care is not common among primates or mammals in general (Fernandez-Duque et al., 2009; Geary, Chapter 20, this volume). It is, however, found among some rodent and primate species, including humans. The extent and types of paternal care vary among species. The hormonal mechanisms for parental care among males appears to differ somewhat from that found among females. Vasopressin appears to function as a co-factor to OT in males (Young & Insel, 2002). Along with prolactin and OT, vasopressin prepares the male to be receptive to and care for infants (Bales, Kim, Lewis-Reese, & Carter, 2004; Bridges, 2008; Rilling, 2013).
Paternal care is more common in monogamous than polygamous mammals and is often linked to hormonal and behavioral stimuli from the female. In the monogamous California mouse, disruption of the pair bond does not affect maternal care but does diminish paternal care (Gubernick & Alberts 1989). In other species with biparental care, however, paternal care is not as dependent on the presence of the female (Young & Insel, 2002). Experience also plays a role in influencing hormonal activation and paternal behavior. Among tamarins, experienced fathers have higher levels of prolactin than first-time fathers (Ziegler & Snowdon, 1997).
Like male parental care, bonding between mates is also uncommon among mammals but has been selected for when it has reproductive advantages for both parents (Carter, 2002; Clutton-Brock, 1991; Young, Wang, & Insel, 2002). Monogamy is found across many mammalian taxa, but most of the current knowledge related to the neuroendocrine basis of this phenomenon has been obtained from the comparative study of two closely related rodent species. The prairie vole (Microtus ochrogaster) mating pair nest together and provide prolonged biparental care, whereas their close relatives, the meadow vole (Microtus pennsylvanicus), do not exhibit these behaviors (Young et al., 2002). As with other social behaviors in rodents, OT and AVP have been found to be central in the differences these related species exhibit with respect to pairbonding.
The receptor density for OT and AVP in specific brain regions might provide the basis for mechanisms underlying other social behaviors. Other neurotransmitters, hormones, and social cues also are likely to be involved, but slight changes in gene expression for receptor density, such as those found between the meadow and prairie voles in the ventral palladium (located near the nucleus accumbens, an important component of the brain's reward system), might demonstrate how such mechanisms could be modified by selection (Lim et al., 2004). The dopamine D2 receptors in the nucleus accumbens appear to link the affiliative OT and AVP pair-bonding mechanisms with positive rewarding mental states (Aragona, Liu, Curtis, Stephan, & Wang, 2003; Wang et al., 1999). The combination results in the powerful addiction that parents have for their offspring.
Given the adaptive value of extensive biparental care and prolonged attachment found in the mating pair and larger family network, it is not surprising that similar neurohormonal mechanisms active in the maternal-offspring bond would also be selected to underlie these other attachments. Though there is some variation among species and between males and females, the same general neurohormonal systems active in pair bonding in other species are found in humans (Wynne-Edwards, 2003). The challenge before evolutionary psychologists is to understand how the general systems have been modified and linked with other special human cognitive systems (e.g., Allman, Hakeem, Erwin, Nimchinsky, & Hof, 2001; Blakemore, Winston, & Frith, 2004; Feldman et al., 2012; Gordon, Zagoory-Sharon, Leckman, & Feldman, 2010) to produce the unique suite of human family behaviors and sociality.
The evolutionary scenario proposed in previous sections posits that the family is of paramount importance in a child's world. Throughout human evolutionary history, parents and close relatives provided calories, protection, and information necessary for survival, growth, health, social success, and eventual reproduction. The human mind, therefore, is likely to have evolved special sensitivity to interactions with family caretakers, particularly during infancy and early childhood (Baumeister & Leary, 1995; Belsky, 1997, 1999; Bowlby, 1969; Daly & Wilson, 1995; Flinn, 2011a; Geary & Flinn, 2001).
The family and other kin provide important cognitive “landmarks” for the development of a child's understanding of the social environment. The reproductive interests of a child overlap with those of its parents more than with any other individuals. Information (including advice, training, and incidental observation) provided by parents is important for situating oneself in the social milieu and developing a mental model of its operations. A child's family environment may be an especially important source and mediator of stress, with consequent effects on health.
Psychosocial stressors are associated with increased risk of infectious disease (Cohen, Doyle, Turner, Alper, & Skoner, 2003) and a variety of other illnesses (Ader, Felten, & Cohen, 2001). Physiological stress responses regulate the allocation of energetic and other somatic resources to different bodily functions via a complex assortment of neuroendocrine mechanisms. Changing, unpredictable environments require adjustment of priorities. Digestion, growth, immunity, and sex are irrelevant while being chased by a predator (Sapolsky, 1994). Stress hormones help shunt blood, glucose, and so on to tissues necessary for the task at hand. Chronic and traumatic stress can diminish health, evidently because resources are diverted away from important health functions. Such diversions may have special significance during childhood because of the additional demands of physical and mental growth and development and possible long-term ontogenetic consequences.
Physiological response to environmental stimuli perceived as stressful is modulated by the limbic system (amygdala and hippocampus) and basal ganglia. These components of the CNS interact with the sympathetic and parasympathetic nervous systems and two neuroendocrine axes, the sympathetic—adrenal medullary system (SAM) and the HPA. The SAM and HPA systems affect a wide range of physiological functions in concert with other neuroendocrine mechanisms and involve complex feedback regulation. The SAM system controls the catecholamines norepinephrine and epinephrine (adrenalin). The HPA system regulates glucocorticoids, primarily cortisol (for reviews, see McEwen, 1995; Sapolsky, Romero, & Munck, 2000).
Cortisol is a key hormone produced in response to physical and psychosocial stressors. It is produced and stored in the adrenal cortex. Release into the plasma is primarily under the control of pituitary adrenocorticotropic hormone (ACTH). The free or unbound portion of the circulating cortisol may pass through the cell membrane and bind to a specific cytosolic glucocorticoid receptor. This complex may induce genes coding for at least 26 different enzymes involved with carbohydrate, fat, and amino acid metabolism in brain, liver, muscle, and adipose tissue (Yuwiler, 1982).
Cortisol modulates a wide range of somatic functions, including: (a) energy release (e.g., stimulation of hepatic gluconeogenesis in concert with glucagon and inhibition of the effects of insulin), (b) immune activity (e.g., regulation of inflammatory response and the cytokine cascade), (c) mental activity (e.g., alertness, memory, and learning), (d) growth (e.g., inhibition of growth hormone and somatomedins), and (e) reproductive function (e.g., inhibition of gonadal steroids, including testosterone). These complex multiple effects of cortisol muddle understanding of its adaptive functions. The demands of energy regulation must orchestrate with those of immune function, attachment bonding, and so forth. Mechanisms for localized targeting (e.g., glucose uptake by active versus inactive muscle tissues and neuropeptide-directed immune response) provide fine-tuning of the preceding general physiological effects. Cortisol regulation allows the body to respond to changing environmental conditions by preparing for specific short-term demands.
Further complications arise from interaction between HPA stress response and a wide variety of other neuroendocrine activities, including modulation of catecholamines, melatonin, testosterone, serotonin, β-endorphins, cytokines, and enkephalins (de Kloet, 1991; Saphier et al., 1994; Ponzi, Muehlenbein, Sgoifo, Geary, & Flinn, 2014). Changes in cortisol for energy allocation and modulation of immune function may be confused with effects of psychosocial stress. As discussed in the previous section, OT and vasopressin intracerebral binding sites are associated with familial attachment in mammals and may influence distress involving caretaker-child relationships. Other components of the HPA axis such as corticotropin-releasing hormone (CRH) and melanocyte stimulating hormone have effects that are distinct from cortisol.
Relations between family environment and cortisol stress response appear to result from a combination of factors including frequency of traumatic events, frequency of positive “affectionate” interactions, frequency of negative interactions such as irrational punishment, frequency of residence change, security of “attachment,” development of coping abilities, and availability or intensity of caretaking attention. Probably the most important correlate of household composition that affects childhood stress is maternal care (Flinn, 2009, 2010). Mothers in socially “secure” households (i.e., permanent amiable co-residence with mate and/or other kin) appeared more able and more motivated to provide physical, social, and psychological care for their children. Mothers without mate or kin support were likely to exert effort attracting potential mates and may have viewed dependent children as impediments to this. Hence co-residence of father may provide not only direct benefits from paternal care but also may affect maternal care. Young mothers without mate support usually relied extensively on their parents or other kin for help with childcare (Flinn & Leone, 2006, 2012/2009).
Children born and raised in household environments in which mothers have little or no mate or kin support were at greatest risk for abnormal cortisol profiles and associated health problems. Because socioeconomic conditions influence family environment, they have consequences for child health that extend beyond direct material effects. And because health in turn may affect an individual's social and economic opportunities, a cycle of poor health and poverty may be perpetuated generation after generation (Flinn, 2006b, 2011b; Flinn, Nepomnaschy, Muehlenbein, & Ponzi, 2011).
People in difficult social environments tend to be less healthy in comparison with their more fortunate peers (e.g., Cohen et al., 2003; Flinn, 2008; Wilkinson, 2001). Social support may often have reproductive consequences. If the brain evolved as a social tool, then the expenditure of somatic resources to resolve psychosocial problems makes sense. Relationships, especially family relationships, are of paramount importance. They have been a key factor affecting human reproductive success at least for over half a million years, and selection has shaped our hormonal, neural, and psychological mechanisms to respond to this critical selective pressure. Children elevate their stress hormone (cortisol) levels much more frequently and extensively in response to psychosocial stimuli than to challenges associated with the physical environment. The adaptive effects of the major stress hormones (Huether, 1996, 1998) and affiliative neurotransmitters on neural reorganization are consistent with the observation that children are especially sensitive to their social worlds (Flinn, 2006b, 2013a).
Social competence is extraordinarily difficult because the target is constantly changing and similarly equipped with theory of mind and other cognitive abilities. The sensitivity of the stress-response and affiliative systems to the social environment may enable adaptive neural reorganization to this most salient and dynamic puzzle. Childhood is necessary and useful for acquiring the information and practice to build and refine the mental algorithms critical for negotiating the social coalitions that are key to success in our species. The human family provides critical support for the developing child in this regard. Traumatic early environments may result in diminished abilities to acquire social competencies as a consequence of glucocorticoid hypersensitivity disrupting neurogenesis, particularly in the hippocampus (Mirescu, Peters, & Gould, 2004; Weaver et al., 2004). An improved understanding of the hormonal and neurological mechanisms that facilitate the intensive and extensive relationships involved with human families and broader kin coalitions (e.g., Carter, Grippo, Pournajafi-Nazarloo, Ruscio, & Porges, 2008; De Dreu, 2012; Flinn et al., 2012), including comparisons between humans and our close primate relatives, may provide important insights into the selective pressures that shaped human psychology.