Primates and the Origin of Culture

W C McGrew. Encyclopedia of Life Sciences: Supplementary Set. Volume 25. Chichester, UK: Wiley, 2007.

Clues to the origins of culture may lie in the diversity of behaviour seen in living largebrained creatures with complex cognition. Comparison across apes, monkeys and other species may help us to understand what is unique and universal about human culture and what is not.

Introduction

Human culture emerged some time in the distant past. Given that it is now universal among modern ‘Homo sapiens‘, then culture must have emerged in our ancestors, but the question remains as to when? The most probable strategy is to seek for cultural origins in the evolutionary lineage of our forebears, after they separated from the apes. Thus, culture could have emerged as early as about 7 million years ago or as late as 200 000 years ago, if the older date is about when the human line split off and the younger date is about when anatomically modern human fossils first appeared on the scene.

Other living species show behaviour that resembles human culture, and this sometimes yields products (e.g. artefacts), but most of these will not become part of the archaeological record, because of their ephemeral or perishable nature. Vocalizations and other behavioural patterns leave no tangible trace, and items of organic materials will turn to dust. Recently, however, the first archaeological study of nonhuman primates showed that stone artefacts can be recovered and analysed, at the nut-cracking sites of wild chimpanzees in West Africa (Mercader et al., 2002).

Lacking a time machine, we cannot do ethnography directly on our ancestors, but we can gain indirect knowledge by studying living creatures. That is, we can use the accessible anatomy, behaviour and artefacts of extant species to infer the inaccessible counterparts of extinct ones, always keeping in mind some sensible caveats, such as no living species can be an ancestor.

Humans are primates, so our nearest living relations are the apes, monkeys and prosimians. But which ones to choose as models of culture? By descent (homology), the closest to us are the African apes, especially the chimpanzee (Pan troglodytes) and the bonobo (Pan paniscus). By ecological convergence (analogy), the closest appear to be the baboons (Papio spp.), who are terrestrial, adaptable (from rain forest to desert), and widespread across African habitats. Apart from these, there are more distantly related and distributed forms of primates that show relevant special abilities, such as capuchin monkeys (Cebus spp.) or Japanese macaques (Macaca fuscata).

Ideas and Definitions

Culture is the core concept in anthropology but the discipline has no agreed definition of the phenomenon. Instead, there is a wide variety of views: some definitions stress ‘behaviour‘ (which is observable), others focus on underlying knowledge (which is inferred), while still others insist on belief and meaning (which are ascribed). Behaviourally, culture is characterized by social learning that leads to multifaceted, shared patterns that persist in time and space. Thus, culture is transmitted by a variety of mechanisms, from mimicry to mentoring, within or across generations, with diffusion within or across populations. Cognitively, knowledge and associated belief systems underpin behaviour but these cannot be directly observed. Instead, the content of cultural minds is indirectly inferred from artefacts or from verbal report.

For example, if we find that baby girls are wrapped in pink blankets and baby boys in blue, then we have both material embodiment and customary behaviour available to study culture. However, to understand ‘why‘ this occurs, that is, what it means, we typically turn to language. We question the blanket-wrappers and try to make sense to their answers. Both methods have their constraints: What to do with a baby wrapped in a green blanket, or not wrapped in a blanket at all? What if our informants refuse to answer, or lie to us about their motives and intentions?

Even if we define culture operationally (that is, in terms that are testable by quantitative or qualitative means), there are other explanations for observed or inferred diversity. Variation may be induced directly by genetic or environmental factors, or by individuals learning in parallel but without interaction. (Similar arguments apply to observed similarity, even to uniformity, as well as to diversity.) What we eat may be a result of what we can digest, what is available or what individuals have put into our mouths, regardless of what others eat. Diet need not be cuisine.

Instead of seeking a simple, consensual definition of culture, perhaps we should use clear and explicit working definitions that are apt for the particular question at hand. This is especially important outside anthropology, as when psychologists or biologists tackle culture from the different viewpoints of their disciplines.

Is human culture unique? Yes, by definition, because our species is unique, as are all others. Is culture uniquely human? Yes, if we define it so, but, not necessarily, depending on how we define it. What is not acceptable is to pretend to answer the question with relabelling. To call what humans do ‘culture‘, versus what other species do ‘proto-culture‘ begs the question. To call what living humans do ‘culture‘ versus what our ancestors did ‘pre-culture‘ merely restates the issue. Such ‘weasel words’ shed no light. Consider another familiar natural phenomenon, digestion: we are happy to apply the same term to both humans and apes, while we examine the relevant variables of enzymes, metabolism, etc. Why should it be different for culture?

One reason to separate out human culture is language. Humans use syntactic and semantic communication to report their thoughts and feelings. Nonhumans may have some such communication but we have yet to decode it. We can discern that different alarm calls uttered by vervet monkeys (Cercopithecus aethiops) refer to different predators such as leopard, snake or eagle, but we do not know if they ever refer to tomorrow’s eagle or to the leopard seen at the waterfall in the next valley. (It would be wondrous to interview a vervet monkey in ‘vervetese’, but this is presently impractical.)

Finally, a word of warning: We should not fall into the ‘space shuttle’ fallacy. Do not forestall the argument by snorting in derision ‘What nonsense! Whatever monkey or ape ever designed a space shuttle?!’ (Or composed a symphony, or invented calculus or cooked a souffle´, etc.) Remember that hundreds of traditional human societies have never done these things either. Many lack the wheel, numbers or even cooking vessels. Do you wish to deny them humanity? Even in industrialized society, how many of us have ever designed a space shuttle, etc.? The toughest challenge for cultural comparison across species is the minimal human example versus the maximal nonhuman one (McGrew, 1987). There is another good reason not to be fixed on language: we can decipher a lot of what goes on in another human culturewithout knowing the language, even when the people are nonliterate. Then we use instead the ethology of facial expression, gesture, posture, etc., and the paraverbal signals of intonation, inflection, etc. as embedded in recognizable contexts. We can do the same across species, with appropriate caution; lack of a common language is a handicap but not an impossible impediment, at least for species with whom we share some common features: apes’ vocalizations are easier to study than bats’, but bats’ are easier than bees’.

Nonprimates

Humans are primates, so it makes sense to look at our nearest living relations as candidates to model the origins of human culture by homology. We ask them to ‘stand in’ for our missing ancestors. However, if culture emerged only recently in the human line, then the best models may be cultured nonprimates, by analogy. But where to start?

If social learning is a necessary but not sufficient condition for culture, then most invertebrates need not apply. (Perhaps the best such case can be made for octopus, but it is not compelling.) If proven performance of socially learned symbol use is the criterion for a culture-bearer, then only the marine mammals are worth a look, outside the primates. These two hurdles, one easy (social learning) and one hard (symbol use), illustrate the issues. Guppies (Poecilia reticulata) show socially mediated learning, but what they learn is trivial by primate cultural standards (Laland and Hoppitt, 2003). Humpback whales have no known technology, but their song-learning traditions are astonishing and unrivalled.

Consider the black rats (Rattus rattus) that have cannily occupied the empty squirrel niche in Israeli conifer plantations. These normally terrestrial rodents become arboreal processors of pine cones and consumers of pine seeds. How they came to be that way shows innovation, dissemination, standardization, tradition and diffusion. A clever combination of experimentation and observation by Terkel and his colleagues shows that there are alternative techniques for extracting the seeds, passed down from mother to offspring, and that these can be explained only by social learning.

Among the birds, the best example of an all-rounder may be the Caledonian crow (Corvus moneduloides) of the Pacific islands of New Caledonia. It shows several types of tool-use, the best-documented example of which is their insect-extracting probes made from the leaves of Pandanus palms. The data came almost entirely from artefacts compared across 21 populations. Little is known of their social learning, but the form of the tool has shown cumulative change over time, from simple to complex. (Not developed here is the huge literature on the learning of bird song, a fascinating but specialized form of vocal learning that does not generalize to other modalities.)

Dolphins are hardly equipped for handling objects, with appendages that lack grasping digits, much less opposable pads and nails, which primates use constantly (Rendell and Whitehead, 2001). Yet even wild dolphins show some elementary technology, e.g. using sponges worn on the rostrum to protect it when rooting for prey in the sandy seabed. Some of the best evidence for teaching by animals comes from killer whales (Orca orcinus), in which mothers teach their young the risky technique of temporarily beaching themselves in order to snatch and then dispatch sea lion pups from a Chilean strand. But the cultural forte of whales, dolphins and seals is vocal communication: bottlenosed dolphins have individual ‘signature whistles’. Humpback whales have song fashions that evolve on an ocean-wide scale, and may even be ‘subverted’ by unexpected migrations of immigrants.

The cultural achievements of nonprimates are impressive, but none of these as well as none of the other classic examples, e.g. the colourful constructions of bowerbirds, really compares well with primate culture.

Monkeys

There are two main families of monkeys, the Old World Cercopithecidae and the New World Cebidae. Each has species that show cultural behaviour (Perry and Manson, 2003).

In the Old World, Japanese macaques yielded data on culture long before any of the ongoing great ape field studies began (de Waal, 2001). From 1948, the monkeys of Koshima, an offshore island, have been studied by Imanishi and his students. In addition to basic socioecological data, they acquired and propagated new behavioural patterns, some of which were wholly spontaneous and unexpected. They invented the washing of sweet potatoes, and the sluicing of wheat grains, which led to sea bathing and adding marine resources to the diet. Elsewhere, on the Japanese mainland, monkeys added hot spring bathing to their repertoire, as a thermo-regulatory strategy in their snowy winter habitat.

Such patterns are not always so obviously functional. Various populations of wild macaques around Japan engage in stone handling, a complex of at least 17 motor patterns such as rubbing, cuddling, clacking, etc., which achieves no apparent goal. Stone handling, like the other patterns listed above it, is sometimes questioned for being shown by monkeys provisioned with artificial foods by humans. However, similar behavioural variation, e.g. lizard-eating by the monkeys of Yakushima, is shown by Japanese macaques that have never been fed by humans.

In the New World, capuchin monkeys are an opportunistic evolutionary radiation that ranges from Mexico to Argentina, and from arid scrub to rain forest. All Cebus species seem to have a penchant for percussive technology, either by pounding hard-shelled items on the substrate (i.e. anvil-use) or by hammering at the item with a hand-held stone or stick (i.e. hammer-and-anvil use). In many cases, capuchin monkeys seem to show uncanny parallels with chimpanzees, e.g. various strategies for hunting mammalian prey, whether this be division of labour in hot pursuit of squirrels or ‘smash-and-grab’ in the canopy to get coati pups from their nests.

One area of cultural behaviour in which capuchins seem to go beyond apes is in their conventionalized games. These bizarre dyadic interchanges have been studied in Costa Rican white-faced capuchins, from inception to extinction. Examples are: putting fingers over or into another’s nostrils; sucking another’s body-parts such as ear or tail; tugging out another’s hair, so that the detached tuft then becomes a toy. These habits may last for years, and researchers hypothesize that they may act as tests of bonding between individuals (Perry et al., 2003).

There may be other species of monkeys that show cultural capacity and performance, but which have not yet been studied from the viewpoint of cultural primatology.

Apes

The great apes are the biggest-brained and brightest of the nonhuman primates, yet the four living species present very different profiles with regard to culture.

The gorilla (Gorilla gorilla) is the biggest species of ape, largely confined to evergreen forests of equatorial Africa. The best-known population, the mountain gorillas of the Virunga Volcanoes, made famous as Dian Fossey’s ‘gorillas in the mist’, is also the most anomalous. At such high altitudes, the apes are almost entirely foliage-eaters and so live in large, cohesive groups that travel little and rest a lot. They seem to have almost no technology, but their techniques for handling spiny, stinging vegetation show cognitive complexity and some regional variation.

Most gorillas live in lowland forest and few populations have been habituated to the extent of allowing the closerange behavioural observation needed to record cultural variation. There are tantalizing hints: in one population in Congo (Brazzaville) that frequents marshes, males show a dramatic splash display. Such candidate patterns remain to be described (ethnography) and then systematically compared (ethnology) by cultural primatologists.

The bonobo, or pygmy chimpanzee, lives only in the war-torn Democratic Republic of Congo and research on this ape species has been handicapped accordingly. The first attempt to compare populations across five study sites with differing degrees of habituation raises intriguing possibilities for culture (Hohmann and Fruth, 2003). At least 14 behavioural patterns resemble those of their close relation, the chimpanzee (see below), such as branch drag, hand clasp, leaf sponge, etc. Another nine patterns seen in the Lomako bonobos are rarely, if ever, seen in chimpanzees: wade in foraging, slap in grooming, scratch with a stick, etc. Notably absent is any kind of elementary technology applied to extractive foraging of either plant or animal prey. The close phylogenetic relationship of bonobo to chimpanzee is not reflected in their ‘extended phenotypes’ of cultural adaptation. However, if research on bonobos were resumed and extended at several study sites, with detailed crosscultural comparison in mind, the picture for this enigmatic species could change quickly.

Because of its solitary nature, the orangutan (Pongo pygmaeus) was thought not to be a promising candidate as a culture-bearer. It is hard to imagine how social learning can be of much use to an arboreal fruit-eater living on its own most of the time. However, some orangutans turned out to be more social than others, and they led the way to a joint article by nine field workers from six sites in Borneo and Sumatra (van Schaik et al., 2003). This collaborative effort echoed that of Whiten et al. (1999) for chimpanzees, and so allows direct comparison. It was only the revelation of customary extractive foraging by the orangutans of Suaq Balimping that pushed the ‘orangologists’ over the threshold for this analysis. Overall, cultural behavioural patterns in the Asian apes are heavily biased towards self-maintenance, however, rather than subsistence, by comparison with chimpanzees. Surprisingly, the proportion of cultural behavioural patterns reported for the two species with such different social lives is much the same (McGrew, 2004).

Although uncommon, the material culture of subsistence in orangutans is just as compelling as that of chimpanzees (see below). For example, orangutans at Suaq use stick tools held in hand or mouth to dislodge the seeds of Neesia fruits. This allows them to avoid the plant’s painfully stinging hairs. The seeds are highly nutritious, so their exploitation is no surprise, and the sticks fall to the forest floor below, to be found by cultural primatologists. On another front, the same population of Asian apes uses green leaves as noise-making devices, which yields a spluttering sound of as yet unknown function. Why is there so much cultural activity at this site? Perhaps because the orangutans at Suaq Balimping live at the highest known density of any wild population of the species. At about 7 apes km-2, the density easily exceeds that of several chimpanzee populations. This suggests social tolerance, which enables the spread of new habits, either ones invented by residents or introduced by immigrants.

The ethnographic record for the chimpanzee is staggering, with at least some data on elementary technology available from over 50 populations, across Africa from Uganda to Senegal (McGrew, 1992, 2004). Furthermore, all chimpanzee populations studied in the long term (for more than a year) are tool-users, showing that for these apes, material culture is a species-universal. However, the bulk of the behavioural data comes from the seven populations that are fully habituated to human observation at close range: Bossou (Guinea), Budongo (Uganda), Gombe (Tanzania), Kanyawara (Uganda), Ngogo (Uganda), Mahale (Tanzania) and Tai (Ivory Coast). Other sites in markedly different types of habitat have yielded indirect data, through ethoarchaeological data, when habituated subjects were lacking: Assirik (Senegal, savanna) and Lope´(Gabon, rain forest).

Unlike all other living species of nonhumans, chimpanzees have extensive ‘tool-kits‘ (a repertoire of elementary technology serving social, subsistence and maintenance functions) and ‘tool-sets‘ (a sequence of different tools used to achieve a single goal). In making tools, chimpanzees may use the same raw material for several functions (a leaf may be a sponge, napkin or container) or the same function may be served by several types of raw materials (a termitefishing probe may be made of twig, bark, vine, grass or leaf midrib). In chimpanzee material culture, an artefact may be discarded after making but before use if it is substandard, reconfigured after use or retained for future reuse. Both raw materials and fabricated tools may be transported both before and after use, or passed to another individual, most commonly between kin.

Jane Goodall was the first to record wild chimpanzees making and using tools in nature, although this ability had been studied in captives decades before by Köhler. Goodall (1986) described how the apes at Gombe fished out underground termites from their earthen mounds, using their flexible probes that had to be skillfully threaded in and then withdrawn. She also described how the chimpanzees tackled biting swarms of driver ants, by going bipedal and using a two-handed technique of swiping the attacking ants off a wand. Nearby, also on the eastern shore of Lake Tanganyika, the chimpanzees of Mahale daily fish for arboreal wood-boring ants, using flexible probes up in the canopy.

Elsewhere, in West African forests, chimpanzees attend less to insects but use percussive technology to crack open nuts with hammer and anvil. At Bossou and Tai, this use of hammers of wood or stone and anvils of living roots or embedded stones is well studied, from its handedness to efficiency to energetics. Chimpanzees typically use only one hand to hold the hammer, maximize their caloric intake and use harder types of stone to crack tougher-shelled types of nuts. Emigrating females spread information on nut cracking to neighbouring or nearby groups. Nut cracking is the prime example of lithic technology in apes, but stones are not flaked to produce a cutting tool.

Some aspects of chimpanzee culture have nothing to do with objects or subsistence, but instead occur in social life. One example will suffice: the grooming handclasp (GHC). All wild chimpanzees groom themselves and others, and such interaction goes far beyond the needs of hygiene. In basic one-to-one grooming, the groomee relaxes while the groomer employs fingers and mouth to clean the body surface. However, a few populations use the GHC, a special form of mutual grooming in which one arm is fully extended overhead to grasp the other’s arm, while the other hand is used to groom the revealed armpit of the other. This ‘A-frame’ configuration is striking; it is conventional at Mahale but totally absent at Gombe. At Mahale, its performance varies even across neighbouring groups, as K-group favours the palm-to-palm GHC while M-group prefers the palm-to-wrist version. Females emigrating from K to M groups adjust their style of GHC to their new companions.

Ancestors

What, if anything, can primate culture tell us about the evolutionary origins of culture in hominin species from prehistory? The obstacles are big: most of primate material culture, however defined, is of organic material, and so is lost to archaeology. This may well be true too for early hominins (e.g. Australopithecus spp.). Further, living primates, as with Homo sapiens, are not ancestral forms; there are huge gaps in the primate fossil record of the last 6 million years. Even when we find processed bones with toothmarks or cutmarks from stone tools from the past, we cannot be sure who made them, ape or hominin.

However, certain inferences seem likely: if chimpanzees today use percussive technology in nature that entails accurate ballistic movement (with all that this implies about eye-hand coordination and sensory-motor skills), then it is not hard to imagine nut-cracking capacities to be transferred to knapping stone in the production of cores and flakes. For other types of behaviour, there is apparent uniformity in action: both chimpanzees and earlier hominins used stones to crack bones to extract marrow. On this basis, Wynn and McGrew (1989) predicted that apes if given the chance had the cognitive ability to make simple flaked stone tools comparable to the earliest known lithic technology (Oldowan) of ancestral hominins. This has since proven to be true for captive bonobos and chimpanzees taught to make flakes with functional cutting edges, who then went on to invent other techniques for fracturing stone edges (Schick et al., 1999).

However, there are no lithic productions by apes that are comparable to the next stage of human technological evolution, the symmetrical bifaces (e.g. handaxes) of the Acheulean starting at about 1.8 million years ago.

Human Culture

So, what can nonhuman primates tell us about the emergence of human culture? Nothing directly, because no living species of monkey or ape was a human ancestor. Also, culture may have arisen several times in human evolution, sometimes in lines that died out, just as it has in primates. It may be that different forms of primates may be the best analogues for different types of cultural evolution, e.g. chimpanzee social-hunting strategy versus capuchin games.

Also, we now believe, after much scrutiny, that teaching (= active, socially targeted transmission of information, as seen in techniques or skills) is not important in nonhumans primate culture. The few anecdotal cases of apparent tuition have not been followed by cases of customary instruction. However, much social learning even in traditional human societies is done not by teaching, but through passive observation.

There has been progress in discerning social learning in the uncontrolled conditions of nature, thus satisfying some of the objections of experimentalists. Simple criteria such as behavioural diversity not readily explained by genetic or environmental influences have been superceded by more focused ontogenetic study. If within a group, individuals adopt the idiosyncratic techniques of their mothers, while the equally viable and available alternative techniques of other matrilines are ignored, then it is hard to lay this variation at the door of inheritance or environment. Equally, its arbitrariness makes coincident individual learning an unlikely explanation (Whiten et al., 2005).

However, primates can help us understand better even the basic processes of culture. For example, some modellers of cultural transmission have assumed that information spreads in a characteristic way that is a function of group size and accelerating speed. This is belied by the realities of social organization, in which some age-sex categories interact much more than others, thus influencing differential transmission within them. This was nicely exemplified in the dissemination of sweet potato washing in Japanese monkeys, which spread horizontally across peers, then vertically across generations.

Nevertheless, some commentators have tried in principle to maintain certain simplistic distinctions between human and nonhuman culture. It is not uncommon in textbooks to see the assertion that only humans depend on culture, while for nonhumans it is somehow optional. The now impressive ethnographic record for chimpanzees belies this logic. Since all known human populations are cultural, it is reasonable to infer culture as a human universal, suggesting dependence. By the same yardstick, such universality in chimpanzees suggests comparable dependence.