Social evolution in Plio-Pleistocene hominins: Insights from hamadryas baboons and paleoecology

Abstract

Reconstructions of hominin evolution have long benefited from comparisons with nonhuman primates, especially baboons and chimpanzees. The hamadryas baboon (Papio hamadryas) is arguably one of the best such models, as it exhibits both the male kin bonding and the cross-sex pair bonding thought to have been important in hominin evolution. Here we link processes of behavioral evolution in hamadryas baboons with those in a Plio-Pleistocene hominin, provisionally identified as Homo erectus (sensu lato) – a pivotal species in that its larger body and brain size and wider ranging patterns increased female costs of reproduction, increasing the importance of sociality. The combination of these higher costs of reproduction and shifts in diet and food acquisition have previously been argued to have been alleviated either via strengthening of male-female bonds (involving male provisioning and the evolution of monogamy) or via the assistance of older, post-reproductive females (leading to post-reproductive longevity in females, i.e., the grandmother hypothesis). We suggest that both arrangements could have been present in Plio-Pleistocene hominins if they lived in multilevel societies. Here we expand on our earlier scenario with two sets of recent data in support of it, (1) archaeological data from the 2 million year old Oldowan site of Kanjera South, Kenya and other sites that are suggestive of tool dependent foraging on nutrient dense resources (animal carcasses and plant underground storage organs), cooperation, and food sharing; and (2) a pattern of genetic variation in hamadryas baboons that suggests the operation of kin selection among both males and females at multiple levels of society. Taken together, these two sets of data strengthen our model and support the idea of a complex society linked by male-male, male-female, and female-female bonds at multiple levels of social organization in Plio-Pleistocene hominins.

Introduction

Several key changes in biology and behavior characterized the evolutionary trajectory of hominins from the late Miocene onwards. These include postcranial adaptations for bipedality, decreased anterior-posterior tooth ratio, increased brain and body size, and changes in diet and social organization. We focus here on the changes in social behavior that may have accompanied the increase in body size and brain size in one or more hominin lineages of the early Pleistocene, particularly those characterized as Homo erectus sensu lato. We focus our analysis on parallels between behavioral evolution in H. erectus and that in hamadryas baboons, Papio hamadryas. We have previously proposed the utility of such an analogy (Swedell and Plummer, 2012), and here we expand on this approach by drawing on more recent behavioral and genetic data from wild hamadryas baboons as well as paleoecological and archaeological data from the 2 million year old (Ma) Oldowan site of Kanjera South, Kenya and other sites that have emerged since our previous paper was published.

Biological anthropology encompasses a long history of the use of nonhuman primates to help understand the evolution of hominin biology and behavior (Washburn, 1951, Washburn, 1973a, Washburn, 1973b). These lines of inquiry have typically been characterized by one of two approaches, the first based on homology and the second on analogy (Jolly, 2013a). The first approach is phylogenetic in that it carries the assumption that the underpinnings of shared features are synapomorphic; this approach thus necessarily relies on comparisons with the closest relatives of hominins, chimpanzees and bonobos (Susman, 1987, Wrangham, 1987, Muller et al., 2017). The second approach, by contrast, is referential in that it focuses on ecologically relevant comparisons across more distantly related taxa. This approach relies on a similar basic template, e.g., that of an anthropoid primate, and draws analogies across convergent evolutionary processes between hominins and some other primate taxon deriving from that similar basic template.

The most commonly used referential model for hominin evolution is the genus Papio, which arguably shares more features with hominins than any other primate genus (Strum and Mitchell, 1987, Alberts and Altmann, 2006, Elton, 2006). The use of baboons as an analogue for human evolution derives from their broad ecological similarities with hominins, including their contemporaneous adaptation to open and variable terrestrial habitats during the Plio-Pleistocene, their behavioral and ecological flexibility, and their lack of ecological specialization – ways in which hominins share more ecologically and evolutionarily relevant adaptations with baboons than they do with chimpanzees and bonobos. Examples of this approach include Washburn and DeVore (Washburn and DeVore, 1961a, Washburn and DeVore, 1961b, DeVore and Washburn, 1963) and Strum and Mitchell (1987), who focused on behavioral and ecological links between baboons and hominins; Jolly (1970), who drew an analogy between differences between Papio and Theropithecus and those between hominins and chimpanzees; Jolly, 1970, Jolly, 2001 and Harvati et al. (2004), who used variation in Papio baboon biology and bony morphology to shed light on hominin evolution and speciation; and Codron et al. (2008) and Sponheimer et al. (2009), who used baboon dietary variability and its correlates to elucidate hominin feeding ecology. Here we focus specifically on the use of baboons to elucidate processes of social and behavioral change during human evolution.

The genus Papio consists of six distinct allotaxa (cf. Jolly, 2001) of baboons that can be considered either six full species (Groves, 2001) or six subspecies of Papio hamadryas (Thorington and Groves, 1970, Frost et al., 2003) depending on whether the phylogenetic or biological species concept, respectively, is followed (Jolly, 1993, Jolly, 2013b). The subspecies classification is consistent with the biological species concept based on the well documented hybridization between hamadryas and olive (P. anubis) baboons in Ethiopia (Phillips-Conroy and Jolly, 1981, Phillips-Conroy and Jolly, 1986, Phillips-Conroy et al., 1991, Phillips-Conroy et al., 1992), as well as between olive and yellow (P. cynocephalus) baboons in Kenya (Maples and McKern, 1967, Alberts and Altmann, 2001) and chacma and Kinda (Papio kindae) baboons in Zambia (Jolly et al., 2011). To be consistent with other contributions to this special issue we follow the phylogenetic species concept here and consider all baboons to be separate species.

One of the first explicit parallels between baboons and humans was made by Sir Solly Zuckerman (Zuckerman and Parkes, 1930, Zuckerman and Parkes, 1939, Zuckerman, 1932), who drew a connection between the heightened levels of aggression and sexuality in hamadryas baboons at the London Zoo and fundamental aspects of human nature. Zuckerman's conclusions were discredited by further research on olive baboons under natural conditions, which suggested that the primary behavioral motivation underlying the evolution of primate behavior was neither aggression nor sexuality, but instead sociality (Washburn and DeVore, 1961b). While Washburn and DeVore's theorizing on sociality focused largely on dominance relationships, recent research on baboons and other primates has suggested that it is not necessarily dominance relationships but, more broadly, social relationships as a whole and their adaptive benefits that drive the evolution of primate sociality and the variety of primate social systems that exist today (Silk, 2007a, Silk, 2007b).

An emphasis on underlying social relationships rather than the more ephemeral emergent properties of primate social organization was the focus of Hinde, 1975, Hinde, 1983 and Kummer, 1970, Kummer, 1978 in their early theoretical treatments of primate sociality, as well as, more recently, Chapais (2008) in his groundbreaking interdisciplinary analysis of human social evolution. Chapais linked the importance of social bonds to kinship, noting that humans are the only primates that maintain kin bonds across multiple generations through both the maternal and paternal lines, combining bilateral kinship and reciprocal exogamy. Such a configuration is not the norm for nonhuman animals, as typically one sex undergoes locational dispersal and loses all (or most) contact with same-sex kin; the advent of a reciprocal exogamy configuration is thus a unique development in the human lineage. Important to our discussion here is that in hamadryas baboons, unlike any other nonhuman primate, we see an incipient version of this configuration: neither hamadryas males nor females disperse voluntarily or frequently and both sexes therefore have opportunities to interact with their same sex kin as adults (Sigg et al., 1982, Swedell et al., 2011, Staedele et al., 2015). A situation in which neither sex undergoes significant dispersal is only possible – i.e., evolutionarily stable without leading to inbreeding depression – in a social system with multiple levels of organization, as this subdivides social groups into smaller units between which breeding usually does not occur (Swedell and Plummer, 2012). In the case of hamadryas, this means that the movement of females predominantly among one-male units (and less often among the larger social units within which they are nested) is sufficient to prevent inbreeding because the one-male unit is the reproductive unit and a female will thereby reproduce in a different one-male unit than her father (and likely her brothers as well), albeit within the same larger social grouping (Hapke et al., 2001, Hammond et al., 2006, Swedell et al., 2011, Staedele et al., 2015). The evolution of a similar type of multilevel social system in a cognitively sophisticated hominin would likely have provided the raw material for Chapais's (2008) bilateral kinship and reciprocal exogamy to evolve in the context of regular exchange of females among reproductive units.

Hamadryas baboons are unique among primates – and different from other baboons – in their combination of strong and differentiated social and kin-based bonds among males, stable and semi-permanent bonds between the sexes outside of the reproductive period, and multiple nested tiers of social organization that cleave and coalesce in consistent ways over time (Kummer, 1968a, Abegglen, 1984, Swedell, 2006, Swedell, 2011). The hamadryas social system is hierarchical and centers on “one-male units”, or OMUs, that are formed via takeovers of females, usually only one at a time, by “leader” males (Kummer, 1968a, Abegglen, 1984, Pines et al., 2011). Leaders maintain continuous social and sexual associations with these females until they are taken over by another male, and this bond between the leader and each of his females underlies the cohesion of each OMU. OMUs may also contain “follower” males, males that are attached to an individual OMU and have social but not sexual access to its females. Above the level of the OMU are larger associations called “clans”, defined by patterns of spatial associations among leader, follower, and “solitary” (bachelor) males (Schreier and Swedell, 2009). Multiple clans coalesce as larger groupings called “bands”, which are the most commonly visible social entities and the closest analogs to the “groups” or “troops” of other papionin monkeys (Kummer, 1990, Swedell, 2011). Finally, bands aggregate at sleeping sites to form “troops”, which are not consistent social groupings. Most social interactions occur within OMUs, either between the leader male and his females, among the females, or between the females and follower males. Also common within clans and bands are social relationships among males, which consist of close proximity and grooming or – especially in the case of leaders, who focus most overt social efforts on maintaining the integrity of their OMU – ritualized greetings called “notifications” (Kummer, 1968a, Colmenares, 1991b, Colmenares, 1991c). These notifications appear to play a role in the hamadryas “respect for possession” of females, which inhibits males from attempting takeovers of females despite their close proximity within clans and bands (Kummer, 1968b, Kummer, 1971, Kummer et al., 1974). Crucial to our discussion here, hamadryas social organization includes social bonds at multiple levels of society, the two most important being male-female bonds within OMUs and male-male bonds across OMUs within clans and bands (Kummer, 1968a, Abegglen, 1984, Staedele et al., 2016).

Hamadryas share some of the above features with Guinea baboons (Papio papio), which are also characterized by a multilevel social system with four detectible tiers of society (Galat-Luong et al., 2006, Fischer et al., 2017), albeit more fluid and egalitarian than in hamadryas (Patzelt et al., 2011, Goffe et al., 2016). The social dynamics of Guinea baboons differ from hamadryas in the reduced longevity and stability of male-female bonds, the greater degree of female social flexibility and movement among OMUs, and the reduced importance of male kinship in structuring male relationships (Patzelt et al., 2014, Kopp et al., 2015, Goffe et al., 2016). It is as yet unclear how or to what degree the multiple levels of the Guinea baboon social system correspond to ecological or social pressures.

Gene flow in both hamadryas and Guinea baboons occurs mainly via females rather than males, an unusual pattern for cercopithecoid primates (Hapke et al., 2001, Hammond et al., 2006, Lawson Handley et al., 2006, Kopp et al., 2014, Kopp et al., 2015, Staedele et al., 2015). Unlike in Guinea baboons and other primates with female-biased dispersal, however, hamadryas females do not transfer frequently or willingly among social units, but are instead coercively transferred by males (Kummer, 1968a, Swedell and Schreier, 2009, Swedell et al., 2011). This pattern, because it leads to only a limited degree of female dispersal – two-thirds of female transfers within bands and only one-third between bands (Swedell et al., 2011) – results in an unusual degree of bi-sexual philopatry (Staedele et al., 2015).

The evolution of the hamadryas social system may have begun during the northward expansion of baboons from their southern African origins in the late Pliocene, leading to the northern clade of extant baboons (hamadryas, Guinea, and olive) that exist today (Newman et al., 2004, Zinner et al., 2009, Rogers et al., 2019). In this scenario, the ancestral system of female philopatry and male dispersal that characterizes most baboons today (the “COKY” species, cf. Jolly, 2019 In press) would have evolved into a system of male philopatry via selection for “philopatry at the frontier,” whereby males that remained in their natal groups would have achieved higher fitness than those that dispersed out of the group due to increased costs of dispersal at the expanding frontier of baboon distribution (Jolly, 2009, In press). This would have provided the basic template for both hamadryas and Guinea baboons, which are alone among Papio in being characterized by male rather than female philopatry. Alongside the evolution of male philopatry, the cohesive multi-male multi-female groups of the ancestral baboon would have evolved into an incipient multilevel social system that was then retained in both hamadryas and Guinea baboons. For hamadryas, the evolution of their male-dominated multilevel social system likely occurred on a backdrop of drier, more heterogeneous, and more seasonally limited habitats in the Horn of Africa that favored fissioning into small foraging groups (Jolly, 1963, Kummer, 1968a), similar to the “grooming clusters” of female olive baboons (Washburn and DeVore, 1961b). In this socioecological milieu, males that chose to remain with – and defend – these small female groups on a semi-permanent basis would have been at a selective advantage due to increased paternity certainty and infant survival. The resultant skewed sex ratio within these social units would have increased the power of selection for male defense and herding due to the increased risk of infanticide by bachelor males (Henzi and Barrett, 2003, Bergman, 2006, Swedell and Saunders, 2006, Grueter et al., 2012, Swedell and Plummer, 2012). This would have led to a system comprising multiple layers: OMUs as reproductive units, clans and bands as ecological units connected by male kinship, and troops as facultative gatherings at shared resources.

Our use of hamadryas baboons as a model for our arguments here does not preclude the utility of other primate taxa living in multilevel societies, which appear to have evolved independently in at least three clades: Asian colobines, Theropithecus, and Papio (Kirkpatrick and Grueter, 2010, Grueter et al., 2012). In theory, principles derived from the evolutionary ecology of each of these primate multilevel societies can be applied to the relationship between ecology and social evolution in hominins to add additional insight. Snub-nosed monkeys (Rhinopithecus spp.), for example, are found in relatively stable OMUs that congregate into larger associations called bands and in which gene flow occurs via both male and female dispersal (Zhang et al., 2006, Qi et al., 2009, Qi et al., 2014, Kirkpatrick and Grueter, 2010); like hamadryas, the fissioning of these bands into smaller social units appears to be at least in part driven by shortages of seasonal foods (Kirkpatrick and Grueter, 2010, Schreier and Swedell, 2012a). Similarly, geladas (Theropithecus gelada) live in multilevel societies over which males compete to reside in and defend OMUs, which coalesce into teams, bands, and communities; unlike Papio, however, geladas retain the ancestral female philopatry that characterizes most cercopithecines and thus OMUs are structured by female kinship (Snyder-Mackler et al., 2014). Geladas are dietarily more specialized than baboons in their focus on grasses and herbs, which Jolly (1970) argued to be analogous to small object feeding in the earliest hominins compared to Pan-like apes (Jolly, 1970). Neither Rhinopithecus nor Theropithecus, however, is likely to have evolved from an ancestral multi-male multi-female group, as both appear to be outcomes of the coalescence of one-male units into larger social units through increased tolerance, mitigation of bachelor threat, and/or predator protection (Bergman, 2010, Grueter and van Schaik, 2010, Pappano et al., 2012, Qi et al., 2014).

While general principles that can be derived from the inferred evolutionary histories in each of these clades are certainly useful in thinking about hominins, we would argue that the genus Papio is the most useful as it is most ecologically analogous with hominins, as discussed above. Within Papio, we view hamadryas baboons as the most informative due to the importance of male-female bonds, the role of male kin ties (described further below), and the demonstrated utility of their multilevel society in coping with the dynamic ecological and social pressures that they experience (Schreier and Swedell, 2009, Schreier and Swedell, 2012a, Schreier and Swedell, 2012b).

In our previous iteration of these ideas (Swedell and Plummer, 2012), we explicitly linked behavioral processes in hamadryas baboons with those in Plio-Pleistocene hominins, particularly H. erectus. H. erectus has been argued to be a pivotal species in human evolution in that its larger brain and body size as well as more extensive ranging patterns increased the costs of reproduction for females (Zihlman, 1978, Parker, 1990, Foley and Lee, 1991, Leonard and Robertson, 1997, Aiello and Key, 2002, Aiello and Wells, 2002, Antón et al., 2002, Ellison, 2008). These higher costs of reproduction, exacerbated by increasingly difficult to acquire, nutrient-dense foods, may have been alleviated via a strengthening of male-female bonds, i.e., male provisioning in the context of a sexual division of labor, a “sex contract”, and the evolution of monogamy, a suite of features that have been proposed to have evolved at various points during human evolution (Darwin, 1871, Isaac, 1978, Lovejoy, 1981, Lovejoy, 2009, Fisher, 1983, Foley and Lee, 1989, Kaplan et al., 2000). An alternative set of ideas instead posits that H. erectus females were assisted by older, post-reproductive females who were in effect benefiting via kin selection by provisioning their own maternal grandchildren; this has become known as the “grandmother hypothesis” and has been used to explain the evolution of post-reproductive longevity (Hawkes et al., 1997, O'Connell et al., 1999, Hawkes et al., 2000, O'Connell et al., 2002). We have previously suggested (Swedell and Plummer, 2012) that in fact both of these arrangements could have been present in Plio-Pleistocene hominins – but only if they lived in multilevel societies. The key to such a model is that multilevel societies allow all three types of social bonds – male-male, female-female, and male-female – to operate simultaneously while also allowing sufficient outbreeding, by virtue of the fact that each set of bonds operates at a different level of social organization.

While current evidence still suggests that H. erectus was a pivotal species in human evolution, additional fossil and archeological finds may modify this interpretation. Since our 2012 publication, Oldowan tools have been excavated at ca. 2.4 Ma at Ain Boucherit, Algeria (Sahnouni et al., 2018) and ca. 2.1 Ma in Shangchen, China (Zhu et al., 2018). The Ain Boucherit occurrence may indicate that the Oldowan was more broadly distributed across Africa close to its 2.6 Ma first appearance datum (FAD) than was previously appreciated. Significantly, these artifacts are associated with bones showing butchery damage, indicating that hominins at this early date were at least occasionally procuring meat and marrow from animals ranging in size from small antelopes to megafauna. The Shangchen locality in the southern Chinese Loess Plateau has a long stratigraphic sequence with multiple archeological levels, the oldest of which is ca. 2.1 Ma (Zhu et al., 2018). It is unclear what the 2.1 Ma artifacts were used for, and whether these early tools reflect a hominin occupation of Asia that was persistent and widespread or one that was ephemeral in nature. The hominin species producing the tools at both localities is unknown. It seems likely that these sites reflect hominin dispersals across and ultimately out of Africa prior to the Dmanisi H. erectus at ca. 1.85 Ma. These new findings exemplify the point that as the record of hominin fossils and activities expands temporally and geographically, our hypothesized scenario may apply to a taxon other than H. erectus. Either way, we expect that the general principles and processes that we describe here will be useful in enhancing our understanding of the evolution of sociality during the history of the hominin lineage. Our aim is not to present a rigid reconstruction by analogy, but to point out the potentially significant implications of a multilevel society in the context of the discussion of the evolution of pair bonding and food sharing during hominin evolution.