The origins of human cumulative culture: from the foraging niche to collective intelligence

(a) Evolution of the human foraging niche: inferences from extant hunter–gatherers

From around 2 Ma, the hominin fossil and archaeological records reveal more persistent stone tool use, occupation of more spatially and temporally variable savannah environments, and exploration of more diverse resources including meat, initially obtained by australopithecines through scavenging and later evolving into high-level scavenging and eventually large-game hunting in Homo [36], and underground storage organs [49]. Such subsistence strategy implied larger home ranges and increased mobility compared to australopithecines or extant apes, and dependence on cultural buffering [50]. Those features most likely appeared gradually and were the foundation for the foraging niche of later hunting and gathering hominins including Homo sapiens. While it is not possible to reconstruct in full detail the evolution of the foraging niche from its early Homo roots (box 2), extant hunter–gatherers can provide invaluable insights into human past adaptations. Different from the fission–fusion groups of chimpanzees and bonobos, hunter–gatherers live in multilevel societies built upon nested levels of organization [86]. Households, household clusters [87], camps and the whole multi-camp structure [11] are recognizable social clusters in hunter–gatherer societies. Clusters interconnectivity is maintained by high rates of inter-camp mobility, with families moving on average every 16 days, with a range between 6 days in the South American Ache and 63 days in the Kalahari Jo/Huansi [88]. Multilevel sociality probably evolved as a consequence of adaptations still observed in hunter–gatherers. While chimpanzees and bonobos are polygynandrous and male philopatric, with cooperation predominantly among related males in chimpanzees [38], monogamy and sex division of labour in hunter–gatherers favour multilocality or dispersal of both sexes, consequently reducing hierarchies within and between sexes [89]. Sex division of labour and biparental provisioning, with consequences such as central place foraging [90], are unique to humans among apes and increase cooperation between sexes and access to resources in multiple camps [91], but also create co-residence with unrelated individuals [92,93] and the challenge of coordinating cooperation among unrelated individuals. The combination of environmental unpredictability and high reproductive costs accounts for food sharing beyond nuclear families [87] and interdependent family units. Extension of cooperative ties and sharing beyond kin may happen owing to shared reproductive interests among in-laws [94], and cooperation among unrelated or even unknown individuals owing to constant mobility and dynamic assortment [88,95]. In conclusion, the fabric of society created by the human foraging niche set the human evolutionary path apart from non-human apes by increasing cooperative ties among kin, affinal kin and unrelated individuals (box 3), as well as promoting specialization, high mobility, fluid sociality and interdependence between family units in open-ended multilevel networks [11,96].

(b) From foraging niche to human cumulative culture

We argue that the foraging niche and its components (pair bonding, multilocality, reduced hierarchies, extended kinship, high mobility, multilevel social networks and prosociality beyond kin) are the explanation for the evolution of human unique cumulative culture. Based on evidence from extant hunter–gatherers, we propose that the foraging niche created the social environment and selective pressures for the evolution of cognitive mechanisms widely recognized as underlying human cumulative culture. Such mechanisms include transmission fidelity, teaching, cultural specialization, recombination and ratcheting. Consequently collective intelligence, generally defined as the increased problem-solving abilities of groups compared to individuals [3,110,111], has been transformed in the human foraging niche. This resulted from the evolution of a ‘network memory’ and ‘social ratcheting’ or irreversible division of labour, spreading the burden of cumulative cultural knowledge across individuals.

(c) Breakdown of social hierarchies increases fidelity of cultural transmission

We propose that reduction in social hierarchies was a central factor behind the increased efficiency of cultural transmission in humans. In chimpanzees and other non-human primates with dominance hierarchies, reduced social tolerance decreases close physical proximity and opportunities for direct and extended observation of complex cultural behaviours performed by role models [19,112]. Subordinate individuals are also less likely to express learned behaviours or be copied when dominants are present [113,114]. The result is reliance on low-fidelity transmission mechanisms such as emulation [39] or reverse engineering [115]. For example, chimpanzees and bonobos acquire moss-sponging skills by reusing discarded sponges without close contact with skilled users [31]. Reliance on low-fidelity transmission does not seem to result from a lack of ability to imitate or copy actions [116]. For example, wild juvenile chimpanzees imitate infant-caring behaviour using rock-dolls, facilitated by tolerant mothers allowing for close-range observation [117]. We conclude that social hierarchies rather than intrinsic cognitive limitations are the main reason why chimpanzees rely mainly on low-fidelity mechanisms of cultural transmission. We can therefore argue that the reason for increased opportunities for more precise copying in hunter–gatherers [118] is that egalitarianism and social tolerance increase proximity and available time for direct observation of cultural role models differing by age, sex and family [119]. For example, hunter–gatherer children interact freely with unrelated individuals from different age groups and spend the most time in playgroups [102,120], where they imitate each other and practise adult skills by hunting small animals, producing toy tools or simulating rituals based on extended observations of adults [121]. In addition, the density of households in camps and cosleeping of various family members within households reduces physical proximity and enhances social learning [122]. In summary, reduced hierarchies in hunter–gatherers facilitate close-range interactions between learners and role models, providing a social context more favourable to high-fidelity cultural transmission.

(d) Cooperation between pairs, extended families and households explain teaching in hunter–gatherers

One of the puzzles in the evolution of cumulative culture is the origin of teaching. Although central for cultural transmission and widespread in humans, teaching is very rare in non-human primates. This is generally explained by an unfavourable balance of costs, benefits and relatedness between tutors and pupils [123]. Teaching is predicted to evolve when skills are highly valuable and difficult to acquire, and when tutors are closely related to pupils. This accounts for rare examples of teaching by related helpers, parents or siblings in cooperatively breeding meerkats and callitrichids, eusocial insects, and between mother and offspring dependent on hunting skills (meerkats, cats and cheetahs) [123]. We argue that this is why teaching in chimpanzees has so far only been observed between closely related individuals (mother and offspring) in the context of tool donation and learning facilitation of termite fishing, a complex behaviour bringing significant benefits but hard to learn without social transmission [13]. By contrast, the hunter–gatherer foraging niche has significantly shifted the balance of costs and benefits towards teaching and modulated the effect of relatedness. First, pair bonding spreads the costs of teaching between parents and significantly increases paternal contributions [124,125]. Second, shared reproductive interests further distribute teaching costs and benefits to affinal kin [94]. Finally, cooperative foraging with shared returns allows exploration of high-quality resources among multiple participants, spreading the benefits of teaching unrelated individuals (who can help to increase group production). For example, cooperative octopus hunting by Agta hunter–gatherers provides opportunities for adults to teach youngsters through cooperation and increased returns. We conclude that while teaching may be occasionally observed in other great apes, the social structure of hunter–gatherers including pair bonding, shared reproductive interests and cooperation with unrelated individuals reduces costs and increases the benefits of teaching, facilitating the learning of more complex technologies and social norms.

(e) Differentiation of network pathways favours cultural specialization within groups

Humans are the only apes where individuals perform highly specialized roles. An explanation for cultural specialization may rest in the unique structure of human social networks and their differentiated interaction channels. For example, in chimpanzees and orangutans, the main channel of cultural transmission and differentiation is transmission across matrilines [32,126]. In hunter–gatherers, social networks are multiple and polyfunctional [127], allowing individuals to interact simultaneously within various specific social groupings differentiated by sex [124], age [128] and skill levels such as female foraging groups, male hunting groups, child playgroups, household units and residential camps. For example, in BaYaka hunter–gatherers medicinal plant knowledge is transmitted between partners, their kin and affinal kin owing to shared reproductive interests, facilitating cultural specialization [10]. Specific transmission channels also stabilize cultural domains and allow for their coexistence and differentiation among social groupings. Restrictions on information flows between channels may further increase cultural diversification [129], as exemplified by BaYaka secret sex-specific rituals [109]. Therefore, we argue that the channelling of information through differentiated interaction paths can explain cultural diversification within hunter–gatherer populations and the emergence of specialized roles such as elephant hunter, honey gatherer, shaman, storyteller, dancer, ritual runner, singer, among others [128,130,131].

(f) Multilevel sociality accelerates cultural innovation and recombination

Recombination of differentiated tools or techniques is widely proposed as the source of major cultural innovations or technological leaps. Cultural recombination products are easy to observe in contemporary societies [132] and hunter–gatherer groups (as exemplified by the bow and arrow or cocktails of distinct medicinal plants). Recombination is also found in chimpanzees, as in honey gathering requiring the sequential use of three to five tools [14]. However, while Tai chimpanzees exhibit 16 tools with 30 technounits, the Hadza exemplify the simplest tool repertoire among extant hunter–gatherers with 39 tools comprising 92 technounits [14]. Hadza tools are also more complex, with clear evidence of recombination such as arrows using up to nine materials. Furthermore, only humans extensively combine tools and technological traditions to create new techniques, for instance when employing stone pounders to produce plant-based medicines, processing food with pounders and fire [133], or collecting honey with tools and smoke [134]. We argue that the reason for the higher rates of cultural recombination in hunter–gatherers is their multilevel social structuring, which is a network adaptation favouring extensive cultural exchange. Since cultural recombination events are rare and occur in evolutionary time, they are hard to observe in field studies and have been often investigated through simulation studies. A recent simulation of cultural evolution based on real hunter–gatherer social networks showed that observed levels of within-camp connectivity and between-camp mobility can significantly accelerate cultural recombination and major technological leaps [11]. As a result, Agta and BaYaka social networks exhibit ‘small-world’ features displaying both high clustering and reduced low path length, which can explain both cultural specialization among close-knit households, and cultural recombination through friendship links between households and camps [96].

(g) Network memory promotes cultural complexity and ratcheting

Human cumulative culture is characterized by a ratchet effect, whereby cultural traits survive across generations with relatively little backward slippage and continuous incorporation of innovations [7]. Although non-human primates have provided many examples of long-lasting cultural traditions, the more complex cultural traits of hunter–gatherers suggest higher rates of cultural accumulation. The fact that culture is produced by populations rather than individuals may explain why ratcheting is more efficient in humans. For example, a BaYaka population from Congo demonstrated collective knowledge of 32 medicinal plants, but no individual knew the whole medicinal repertoire [10]. Therefore, the ratchet effect implies reliance on a collective memory that distributes cultural knowledge across individuals unable to fully recreate it from scratch. We argue that a main reason hunter–gatherers build more efficient collective memories is their unique social structure. Simulations have shown that large population size and full network interconnectivity reduce the risk of cultural loss, but also wipe out diversity by homogenizing traits owing to group-wide transmission. On the other hand, fragmented groups may produce more diversity because of differentiation between clusters, but extreme fragmentation may result in isolation, reduced introduction of innovation from other clusters, loss of collective memory in smaller groups and ultimately loss of cultural complexity over time [135,136]. Therefore, the reason for cultural ratcheting in hunter–gatherer societies is that their multilevel social structuring takes advantage of both large population size and fragmentation without their side-effects [96], allowing for innovations to accumulate across generations with reduced cultural loss. By contrast, chimpanzees and bonobos live in more stable and cohesive groups, a social structure accounting for their ability to preserve innovations across generations but also for the relative rarity of major cultural leaps. In summary, multilevel sociality can explain how cultural complexity may evolve owing to a collective or network memory splitting the individual burden of storing more diverse cultural repertoires [11,137–139]. The evolution of a complex ‘network memory’ is therefore a distinguishing feature of a human collective intelligence intrinsically linked to cultural accumulation.

(a) Social ratchet and the origins of gene-culture coevolution

It has been shown that culture can relax or increase selection pressures and favour adaptations in various species [140,141]. However, the reasons for the transition from a facultative inheritance system to dependence on a system of gene-culture coevolution in hominins remain a puzzle. We propose that as a result of multiple network channels of cultural differentiation and recombination in the foraging niche, at some point hominin cultural repertoires must have grown to a point where single individuals could not master a significant fraction of accumulated knowledge and techniques, as observed in current hunter–gatherers. The expected irreversible interdependence among specialists was proposed as a feature of major evolutionary transitions [142], with the division of labour and extensive cultural exchange favouring specialization, complementary skills and increased system efficiency [21,27]. Similar to the role played by sexual reproduction in genetic evolution, cultural recombination became the main mechanism generating innovations from a pool of skill-differentiated individuals. While debates have mostly focused on cultural ratcheting, the foraging niche also set in motion a ‘social ratchet’ or trade system where specialization within populations became irreversible. This process generates storytellers and shamans in hunter–gatherers, and later medical doctors and IT specialists in industrial societies, in a process analogous to sexual reproduction and the eventual evolution of interdependent sexes. In summary, the foraging niche has created the behavioural and social conditions for the emergence of social ratcheting, or the cultural specialization and interdependence between specialists. The consequence was the transition from reliance to dependence on culture and hence the process of gene-culture coevolution itself.

(b) Human cultural cognition was driven by selection for cumulative culture in the foraging niche

Comparative studies have proposed that human cumulative culture is explained by unique cognitive capacities such as theory of mind, teaching, shared goals and intentionality, or a tendency of children to overimitate role models [7]. However, most features have now been identified in other species pointing to a continuum with humans at its higher end, suggesting that variation in cognitive abilities may be the result of differences in intensity of past selective pressures. Therefore, the foraging niche perspective implies that uniquely advanced cognitive abilities in humans evolved as a consequence (or proximate mechanisms) of stronger selection for efficient transmission of cumulative culture (the ultimate or evolutionary cause). For example, the stepwise transition to the foraging niche increased interdependence and reliance on cultural transmission, causing stronger pressure for cooperation and collective problem solving. We propose that the result was the evolution of more sophisticated collective intelligence and shared intentionality [3,7,8,11]. While those cognitive abilities had often been described as causes of human culture from a mechanistic or proximate perspective, they are ultimate consequences of the selective pressures for cumulative culture in the foraging niche. The second consequence of our argument is that language may also have emerged from a cumulative cultural process owing to stronger pressure for efficient cultural transmission. This view is compatible with its possible gestural origin as a tool-making teaching aid [70]. In later hominins with larger and more interconnected social networks, language might have evolved into speech as a more complex communication technology [143,144]. Similarly to stone tool technology, speech-based language is a system of multipart tools (or sentences) built from vocalization units and could therefore have evolved through cultural recombination [145]. In summary, the foraging niche may have provided the selective context for the evolution of cognitive and cultural abilities underlying human cumulative culture.

(c) The foraging niche accelerated the genetic and cultural evolution of Homo sapiens

The foraging niche had equally important consequences for the evolution of the human lineage itself. It set some hominin taxa on a path of increasing prosociality, interdependence and cultural exchange dependent on multilevel social structuring, a process reaching its most extreme expression in the larger and fluid metapopulations of early Homo sapiens. We propose that large-scale social networks promoted the genetic, morphological and cultural evolution of modern humans by facilitating not only cultural and material exchanges but also flows of people and genes (figure 1). This would explain the accelerated pace of technological evolution in the Middle and Late Stone Ages in Africa [16]. From this perspective, ‘cultural revolutions' such as the Upper Palaeolithic in Europe (possibly incorporating elements from Neanderthal technology [149]) would represent a local case of a continuous process of cultural innovation, recombination and ratcheting within structured hunter–gatherer populations. On the genetic front, large-scale networks may also explain the emergence of modern humans from regionally differentiated early sapiens groups identified in east, south and north Africa at 300–400 ka and contributing in different degrees to current modern diversity [146]. The occasional expansion of social networks between species may have also accelerated genetic change in modern humans [150] owing to introgression of adaptive alleles from Neanderthals and Denisovans [151,152]. By the time of Neanderthal extinction in Europe, humans had lived in unrelated and interconnected bands [72] and were the outcome of a long history of cultural and genetic recombination at the continental scale in Africa [25]. By contrast, genetic data indicate that Neanderthals and Denisovans may have faced higher rates of inbreeding [153]. If the latter is an indication of reduced population connectivity, they should also exhibit lower rates of cultural exchange than modern humans. It follows that the adaptive edge of early sapiens may have resided in higher cultural recombination levels, as well as superior collective intelligence based on more sophisticated network memories and social ratcheting, rather than differences in individual cognitive ability. Therefore, social and ecological factors may explain why social networks in Neanderthals did not exhibit the levels of regional differentiation and integration observed in H. sapiens.

• Download figure
• Open in new tab
• Download PowerPoint

In conclusion, we suggest that a unique foraging niche still observed in a few extant hunter–gatherer populations provided the foundations for human cumulative culture by reducing hierarchies and increasing opportunities for social learning and high-fidelity cultural transmission; facilitating teaching and cooperative skill transfers; promoting sexual and social division of labour and skill specialization; promoting cultural recombination across multilevel social structures; and establishing network memory and social ratcheting processes spreading the burden of cultural knowledge across individuals, resulting in a human collective intelligence uniquely suited to ratcheting culture over generations. The outcome was the eventual transition of the foraging niche into a cultural niche where cumulative culture became a second inheritance system and the main driver of human evolution.