Farewell to the ‘childhood of man’
Ritual, seasonality, and the origins of inequality

Abstract

Evidence of grand burials and monumental construction is a striking feature in the archaeological record of the Upper Palaeolithic period, between 40 and 10 kya (thousand years ago). Archaeologists often interpret such finds as indicators of rank and hierarchy among Pleistocene hunter-gatherers. Interpretations of this kind are difficult to reconcile with the view, still common in sociobiology, that pre-agricultural societies were typically egalitarian in orientation. Here we develop an alternative model of ‘Palaeolithic politics’, which emphasizes the ability of hunter-gatherers to alternate – consciously and deliberately – between contrasting modes of political organization, including a variety of hierarchical and egalitarian possibilities. We propose that alternations of this sort were an emergent property of human societies in the highly seasonal environments of the last Ice Age. We further consider some implications of the model for received concepts of social evolution, with particular attention to the distinction between ‘simple’ and ‘complex’ hunter-gatherers.

Epigraph

If we seek to know about the past, a field of study that has never seemed dishonourable to any discipline other than social anthropology, the point of departure should be hunter-gatherers in favourable regions, hunter-gatherers who might not have been such and probably remain such only by reason of restrictive social forms that for them are quite possibly a distant and glorious heritage (Testart 1988: 13).

Introduction: the ‘sapient paradox’

The Henry Myers Lecture was endowed seventy years ago to promote new perspectives on the ‘the place of religious belief in human development’. Only recently, however, two leading anthropological theorists concluded that, to all intents and purposes, ‘religion’ does not exist – at least not in the sense of a discrete analytical category that we can expect to find and study across the whole range of human societies. We are referring here to Marshall Sahlins’ assertion that ‘the elementary forms of kinship, politics, and religion are all one’ (2008: 197), and to Maurice Bloch’s (2008) conclusion that what we now term ‘organized religion’ is a historical residue, left over from the collapse of Bronze Age states where sacred and political power were initially fused.

If they are right, then a Myers Lecture on human prehistory could in theory be about almost anything. By choosing to discuss the origins of social inequality – our main topic – we will also find ourselves talking about religion and, probably, economics and politics as well: a position that resonates with the kind of language used today by archaeologists and evolutionary theorists, who no longer talk about the origins of ‘religion’ or ‘politics’, but rather speak of ‘behavioural modernity’ or ‘cultural complexity’. This is precisely to indicate that the earliest evidence for what we might now distinguish as ‘religious’, ‘political’, or, for that matter, ‘artistic’ behaviour is all of a piece, appearing together in striking configurations in the archaeological record of the last Ice Age. The main problem vexing prehistorians concerns the timing of that appearance.

To summarize briefly, the genetic and anatomical foundations of our species were established between 200 and 160 kya (thousand years ago); but evidence for complex modes of symbolic communication – in other words, for typically modern human behaviour – becomes widespread in the archaeological record only tens of thousands of years later. First glimmerings appear at Blombos Cave, on the southern tip of Africa, where evidence for the use of ochre-based pigments (at 100 kya) and shell ornaments (at 70 kya) is found across a series of deposits dating to the Middle Stone Age (Henshilwood 2007; Henshilwood et al. 2011). But it is only after around 45 kya, when our species was busily colonizing Eurasia, that evidence for cultural complexity becomes more widely attested: an efflorescence that has sometimes – and contentiously – been termed the ‘Upper Palaeolithic Revolution’ (Mellars, Boyle, Bar-Yosef & Stringer 2007).

None of these novel activities are exclusive to Upper Palaeolithic Europe and it is, indeed, unlikely that any of them originate there (see McBrearty & Brooks 2000). Nevertheless, it is across the southern and central parts of that continent that they are currently documented with the greatest frequency and intensity. The activities in question include the use of advanced toolkits for hunting and handicrafts, the transformation of diverse materials (e.g. bone, clay, fibre) into durable images and structures, new ways of clothing and decorating the body, the use of musical instruments, the exchange of raw materials over impressive distances, and also what are generally taken as the earliest proofs of social inequality, in the form of grand burials and – after the Last Glacial Maximum (c. 20 kya) – monumental dwellings as well. It is this apparent lack of synchrony between the ticking of our genetic and cultural clocks that Colin Renfrew (2007) provocatively calls the ‘sapient paradox’.

In seeking to resolve the paradox, prehistorians have so far offered two explanations. The first – which remains more of a supposition – is that a late but significant mutation took place in the human brain between c.70 and 50 kya, generating new cognitive resources that made possible the heightened cultural creativity of the Upper Palaeolithic (Mithen 1996; cf. Klein 2001). The second concerns demography. It predicts that, where critical population thresholds were reached, the transmission of complex cultural traits became incremental in an unprecedented way owing to the greater density of human interactions (Powell, Shennan & Thomas 2009). This latter view has the advantage of explaining why so much of the earliest evidence for behavioural modernity appears in Europe, in what were then the game-rich valleys and steppe between the tundra and forest zones.[1]

These parklands – seasonally traversed by migrating herds of deer, bison, and mammoth – were distributed unevenly between the western Mediterranean and the south Russian Plain. As ice sheets advanced over the continent, they acted as refugia for both human and nonhuman populations (Hewitt 2000; Stewart & Stringer 2012). Prehistorians have argued for some decades that the humans in question had nothing in common with those blissfully simple and egalitarian hunter-gatherer bands once imagined to be our remote ancestors (see, e.g., Price & Brown 1985). Yet the continued popularity of books (e.g. Diamond 2012; Fukuyama 2011) that preserve Jean-Jacques Rousseau’s vision of humanity in its original state of nature – innocent of power and complexity – suggests a reluctance to bid farewell to the ‘childhood of man’, and to embrace a new age of cynicism, where inequality is considered not only natural but also a primordial feature of human society.

Ambivalence about the social organization of Pleistocene hunter-gatherers can also be found in the more specialized literature on social evolution (e.g. Flannery & Marcus 2012) and human origins. In Hierarchy in the forest, Christopher Boehm (1999) notes how sociobiologists and evolutionary psychologists have tended to characterize humans as innately either egalitarian or hierarchical, perpetuating an ‘endless debate’ between the positions of Rousseau and Hobbes. He himself argues that this is a false dichotomy. What makes us distinctly human is instead the inherent complexity of our political repertories, and in particular the range of strategies for resisting domination, which far outstrip those available to other primates. At the psychological level these include ridicule, moral censure, and ostracism; at the social level they involve complex institutional arrangements to limit or subvert the exercise of power.

Yet, following Bruce Knauft (1991), Boehm is also willing to make ‘the major assumption that humans were egalitarian for thousands of generations before hierarchical societies began to appear’, a development that he places around 5 kya:

At that time, people were beginning to live increasingly in chiefdoms, societies with highly privileged individuals who occupied hereditary positions of political leadership and social paramountcy. From certain well-developed chiefdoms came the six early civilizations, with their powerful and often despotic leaders. But before twelve thousand years ago, humans basically were egalitarian. They lived in what might be called societies of equals, with minimal political centralization and no social classes. Everyone participated in group decisions, and outside the family there were no dominators (Boehm 1999: 3–4; and see also 5, 207).

Why, then, should our species’ engrained capacity for political complexity have been held in suspense for the greater part of human (pre)history? Sociobiology poses the question, but offers no clear answer. Moreover, broad-brush characterizations of a deep egalitarian past – before the emergence of farming and states – sit uneasily with the content of the archaeological record. That evidence, discussed further below, leads prehistorians in a very different direction, towards the identification of ranked societies and institutional hierarchy among Pleistocene hunter-gatherers.

A question can be posed at the outset: do we really have to choose between these starkly opposed views? In what follows, we propose an alternative to the characterization of Palaeolithic societies in binary terms (‘complex’ versus ‘simple’, ‘hierarchical’ versus ‘egalitarian’). Our model posits that Pleistocene hunter-gatherers alternated – consciously and deliberately – between contrasting modes of political organization. Dual structures of this kind are found across a range of historically documented societies. They were widely reported in early twentieth-century ethnographies of hunting and foraging groups, some of which we revisit below; and have since been discussed for a variety of agricultural and urban societies. Randall McGuire and Dean Saitta’s (1996) characterization of Pueblo political organization in the American Southwest as alternating, routinely and strategically, between ‘communal’ and ‘hierarchical’ modes of governance is an excellent example (and for further discussion and examples, see also Ehrenreich, Crumley & Levy 1995; McGuire 1983; McIntosh 1999).

As yet, however, such models have been little applied to the greater part of our species’ history. In extending them to the Palaeolithic past, we propose a relationship between seasonality and the conscious reversal of political structures. To date, and with some exceptions (discussed, again, below), research on seasonal variability in Palaeolithic archaeology has tended to focus on issues of subsistence and long-term environmental change, rather than social organization. For the Upper Palaeolithic, in particular, coping with ever more seasonal environments has been identified as a key factor in hominin adaptation and colonization, especially of the world’s northern latitudes (Gamble 1998: 19). It has been widely noted that the specialized hunting of migratory game – practised throughout Europe by early human, and perhaps also Neanderthal,[2] populations – implies a high degree of logistical planning (Mellars 1998: 61; Nitecki & Nitecki 1987). Here, however, we will consider how seasonal variations might be relevant to a much broader set of issues concerning the nature and expression of inequality in Palaeolithic societies.

Linking changes in social organization to seasonal variations in climate and resources might seem to evoke the type of ‘fission-fusion’ systems found in certain nonhuman species, such as chimpanzees and bonobos (see Dunbar, Lehmann, Korstjens & Gowlett 2014). The alternations that concern us here are, however, of a categorically different kind. Changes in the physical constitution of chimpanzee groups reflect the variable distribution of resources throughout the year, and often involve the renegotiation of social alliances. Human hunter-foragers also move regularly between groups of varying size and density, often on a seasonal basis. But uniquely for humans, with their particular type of social cognition (Bloch 2008), such alternations involve corresponding changes in moral, legal, and ritual organization (as first pointed out by Mauss & Beuchat 1979 [1904–5]; and cf. Bailey 1978). Not just strategic alliances, but also entire systems of roles and institutions are periodically disassembled and reconstructed to allow for more or less concentrated ways of living at different times of year. Here we revisit and develop this theme with particular reference to the political aspects of seasonal variation, exploring its implications both for Palaeolithic archaeology, and for general theories of social evolution.

Some problems with ‘complex hunter-gatherers’

In broaching these issues, we begin with the phenomenon of ‘rich’ hunter-gatherer burials. Such burials are sporadically attested from Upper Palaeolithic rock shelters and open-air settlements across much of western Eurasia, from the Dordogne to the Don. Some of the earliest instances come from the eastern end of this distribution, at sites such as Sungir (in northern Russia) and Dolní Věstonice (in Moravia), where they date to between 26 and 30 kya, before the Last Glacial Maximum. They comprise isolated interments of individuals or small groups, whose bodies were placed in striking postures and decorated – or, in some cases, virtually saturated – with ornaments. In the case of Sungir these included many thousands of mammoth ivory beads and perforated fox canines, originally attached to items of clothing. Some of the most lavish ornamentation at this site was associated with the conjoined burials of two children – a boy and girl – whose bodies were flanked by great lances made on straightened mammoth tusk (Bader & Mikhajlova 1998; Trinkaus, Buzhilova, Mednikova & Dobrovolskaya 2014).

At Dolní Věstonice a triple burial contained two young males with elaborate headdresses, posed on either side of a female, all of them lying on a bed of ochre- stained soil (Klíma 1988). Of similar antiquity is a group of cave burials unearthed on the coast of Liguria, near the modern border between Italy and France. Complete bodies of young or adult males (including one particularly lavish burial known as Il Principe) were again laid out in striking visual arrangements and suffused with decorative objects, here including beads made on marine shell and deer canines, as well as blades of exotic flint (Henry-Gambier 2003). Further west, on the Dordogne, the 16 kya burial of a young woman – known as a the ‘Lady of Saint-Germain-la-Rivière’ – contained a rich assemblage of stomach and pelvic ornaments, made on shell and on the teeth of young stags hunted some 300 km away, in the Spanish Basque country (Vanhaeren & d’Errico 2005).

Spectacular burials of this kind have been taken as evidence that – many thousands of years before the origins of farming – highly developed systems of ranking existed among at least some Upper Palaeolithic societies. Attention has focused on the extraordinary outlays of labour involved in making the grave goods (some 10,000 work hours are estimated for the Sungir beads alone); the highly advanced and standardized methods of craft production; the inclusion of exotic (and therefore prestigious) raw materials; and the association of wealth with young individuals, taken to imply ascribed rather than achieved status. On such grounds we are asked to abandon the idea that Palaeolithic hunter-gatherers were uniformly simple or egalitarian in their social arrangements, and to accept the fundamentally complex and hierarchical nature of their social systems (e.g. Vanhaeren & d’Errico 2003; 2005; White 1999).

A second category of evidence, from which similar conclusions have been drawn, is monumental architecture. In Old World prehistory, the most famous and widely discussed examples are currently the stone buildings of the Gemus Mountains, overlooking the Harran Plain in southeast Turkey. These lie outside the main chronological focus of this article, but are nevertheless relevant to any wider discussion of hunter-gatherer complexity, and can therefore be briefly mentioned. Around twenty years ago, on the plain’s northern frontier, German archaeologists began to uncover prehistoric remains at a place known locally as Göbekli Tepe. What they found has since come to be regarded as an evolutionary conundrum. The main source of anxiety is a group of twenty megalithic enclosures, raised there at a time – around 9000 BC – when the surrounding plain was woodland-steppe, teeming with wild plant and animal life that colonized the Taurus piedmont after the end of the Pleistocene. Scientific dating places these structures within the ‘Pre-Pottery Neolithic A’ period but, on current evidence, the groups responsible for their creation lived by hunting and foraging alone (Schmidt 2006).

Just a few of the enclosures known to exist at Göbekli Tepe have been excavated. Each comprises pillars – some over 5 m high, and weighing up to a ton – that were hewn from the site’s limestone substratum, raised into sockets, and linked by walls of rough stone. Each pillar is a unique and remarkable work of sculpture, carved with images from the world of dangerous carnivores and poisonous reptiles, as well as game species, waterfowl, and small scavengers. Animal forms project from the rock in varying depths of relief, some hovering coyly on the surface, others emerging boldly into three dimensions. They follow divergent orientations, sometimes marching to the horizon, sometimes working their way down into the earth. And in certain cases the pillar itself becomes a sort of standing body, with human-like limbs and clothing. Daron Acemoğlu and James Robinson (2009) argue, on the basis of these structures, that ‘hunter-gatherer societies had evolved institutions to support major public works, projects, and monumental constructions, and thus had a complex social hierarchy prior to their adoption of farming’ (2009: 679; see also Dietrich, Heun, Notroff & Schmidt 2012; Flannery & Marcus 2012: 128–31; and for critical discussion of the evidence for institutional hierarchy at Göbekli Tepe, see Banning 2011).

Evidence for monumental construction among early hunter-gatherers – implying sophisticated design and the large-scale recruitment and co-ordination of labour – is not confined to the Middle East, or to the onset of the Holocene. Between 18 and 12 kya, along a transect of the glacial fringe reaching from Kraków to Kiev, people lived in impressive circular houses that Olga Soffer (1985b) describes as the Pleistocene’s version of ‘public works or monumental architecture’. Each such dwelling was built on a framework of mammoth tusks and carefully selected mammoth bones, arranged in alternating sequences and (sometimes) in rhythmic patterns that go beyond the merely functional. Wooden versions – of which only the post-holes and sunken floors remain – are likely to have existed at other open-air sites such as Pavlov and Kostenki. These were settlements of considerable scale whose inhabitants exchanged amber, marine shells, and animal pelts over impressive distances (see also Soffer 1985a); and they find their western European counterparts in the large rock-shelter occupations of southern France, such as La Madeleine and Abri Pataud (Mellars 1998: 61–3).

Based on evidence of this kind, archaeologists can now claim to have pushed back the record of institutionalized inequality to a very early phase of human prehistory (cf. Flannery & Marcus 2012). We also note the suggestion, in a recent review of ‘complex hunter-gatherers in evolution and history’, that recognizing institutions of rank among non-farming populations constitutes one of ‘the most significant advances in anthropological research in the last thirty years’ (Sassaman 2004: 228). Taking a longer-term view of research on this topic, we would strike a less triumphal note. The existence of ranking and other hierarchical structures among non-farming societies was common knowledge for much of the twentieth century, both for anthropologists and for archaeologists (see, e.g., Childe 1954: 41–2; and we discuss some well-known examples below); but, more importantly, we would contend that simply observing the existence of inequality in certain aspects of social life and material culture, in certain times and contexts, says little about social evolution in general.

Current definitions of ‘complexity’ in the Upper Palaeolithic, while accepting the cognitive modernity of prehistoric hunter-gatherers, often continue to ascribe them a classically primitive type of social intelligence. Rather than being aware of multiple social possibilities, early Homo sapiens appear as effectively (or perhaps stereotypically) childlike, living the only lives they were able to imagine. Instead of experimenting consciously with different social strategies in different contexts, they are cast back into a single evolutionary stage, albeit a slightly more advanced one. Robert L. Kelly offers a clear statement of the problem, urging a study of ‘hunter-gatherer prehistory in terms other than broad typological contrasts such as generalized versus specialized, simple versus complex, storing versus non-storing, or immediate versus delayed return’ (2013: 275; and see note 4 below). Still, in his seminal definition of ‘the foraging spectrum’, Kelly himself maintains a broad dichotomy between ‘egalitarian’ and ‘non-egalitarian’ hunter-gatherers as distinct types of society with stable internal characteristics (tabulated as a binary contrast between ‘simple versus complex’ forms; Kelly 2013: 242, Table 9–1; and for the application of a similar dichotomy in archaeological interpretation, see, e.g., Hayden 1990; 2009).

Revisiting an earlier ethnography

To substantiate these criticisms, and suggest alternative ways forward, we want to revisit an earlier tradition of anthropological research, linking the work of Marcel Mauss (Mauss & Beuchat 1979 [1904–5]), Robert Lowie (1948), and Claude Lévi-Strauss (1967 [1944]; and for the relationships between them, see also Lévi-Strauss 1949). What interests us about this group of studies – aside from their broad comparative scope – is their attentiveness to the institutional plasticity of groups that exhibit pronounced seasonal variations in their economic pursuits.[3]

Our starting-point is a 1944 study of chieftainship by Claude Lévi-Strauss, which centres on the Nambikwara, a small tribe inhabiting the resource-starved savannah of northwest Mato Grosso (Brazil). It is worth noting, in the context of the present discussion, that Lévi-Strauss began his essay by pointing out some of the obvious limits of ethnographic analogy for archaeological reconstruction (e.g. the fact that, unlike Pleistocene hunter-gatherers, modern peoples who practise hunting and foraging do so on the margins of agro-industrial states, and frequently supplement these activities with various forms of low-level farming; cf. Kelly 2013; Testart 1988). The point, for him, was to use ethnographic accounts not as proxies for particular stages of past life (as defined, for example, by modes of subsistence), but rather as a source of insight into features of the human condition that might be considered of general evolutionary significance.

Precisely because of their material impoverishment and aversion to competition, Lévi-Strauss felt that a study of chieftainship among the Nambikwara could expose ‘some basic functions’ of political life that ‘remain hidden in more complex and elaborate systems of government’ (under cover of complexity, as it were). In particular he argued that the role of ‘chief’ seemed analogous – in its social and psychological aspects – to that of a national politician or statesman. It also attracted similar kinds of people who ‘unlike most of their companions, enjoy prestige for its own sake, feel a strong appeal to responsibility, and to whom the burden of public affairs brings its own reward’ (Lévi-Strauss 1967 [1944]: 61). Maintaining the role of chief also had everything to do with the way that the Nambikwara shifted back and forth between two different modes of social and economic organization: the hilltop villages of several hundred people, occupied mainly in the rainy season when they practised horticulture, and the small foraging bands into which they dispersed for the rest of the year. Chiefs made or lost their reputations by offering guidance during the nomadic adventures of the dry season. And with the greater abundance of the wet season, a chief who had performed this task well could attract large numbers of followers to settle in villages, where he directed the construction of houses and tending of gardens.

Neither patriarchs, nor petty tyrants, nor mystical healers, Lévi-Strauss’s chiefs were truly and fully holders of public office: the pivot of something like a small-scale welfare state. They were also mature and self-conscious politicians, capable of moving regularly back and forth between what other anthropologists at the time were inclined to see as different phases of evolutionary development (band/tribe/chiefdom), and developing careful strategies to do so. It was their skill at guiding small bands of hunter-gatherers that qualified them later to play the role of mediator and representative in the village plaza. For Lévi-Strauss it was precisely this quality that made the Nambikwara chief seem so peculiarly familiar as a political figure: the calm sophistication with which he shifted between different social arrangements, all the time balancing a sense of individual ambition with the common good.

The essay on Nambikwara chieftainship was written quite early in Lévi-Strauss’s career; but it received little attention even at the height of his fame. In emphasizing continuities between the political lives of hunters, horticulturalists, and modern industrial democracies it cut against the grain of an emerging evolutionism: not only the formal distinction between ‘bands’, ‘tribes’, ‘chiefdoms’, and ‘states’ laid out by Elman Service (1962), but also the larger research agenda on hunter-gatherers set out in the 1966 Chicago symposium ‘Man the Hunter’ (Lee & DeVore 1968a), to which Lévi-Strauss offered a forlorn and now equally forgotten epilogue. Instead it was behavioural ecology, and rigorously quantified studies of African savannah and rainforest groups – the Kalahari San, Eastern Hadza, and Mbuti Pygmies – that provided the basis for a new characterization of hunter-gatherers.

As summarized by Richard Lee and Irven DeVore (1968b: 11), foraging peoples could be assumed – by virtue of their fragile and unstable mode of subsistence – to ‘live in small groups’, ‘move around a lot’, and follow egalitarian principles, resolving conflicts by ‘fission’ rather than arbitration or violence.[4] This quickly became self- evident wisdom, such that it is still commonplace for sociobiologists to remark that, prior to the invention of farming, most humans lived in small bands with little social structure or internal differentiation, other than distinctions of age and gender (e.g. Boehm 1999: 3–4; Dunbar 1996: 69–70; and for a critical review of the concept of the hunter-gatherer ‘band’, see Ingold 1999). Here we consider how a return to the ethnographic tradition of Lévi-Strauss, which flourished between the abandonment of Victorian evolutionism and the neo-evolutionary theory of the 1960s, might generate more fertile perspectives on the archaeological record.

Individuality and egalitarianism in the Upper Palaeolithic

The first of Lévi-Strauss’s points that we wish to develop is a relatively simple one. It is generally acknowledged that egalitarian societies of the Americas were typically marked by an ethos of extreme individualism. Far from encouraging a stifling conformity, they emphasized individual autonomy and self-realization. In practice this meant that even in these least materialistic and competitive of societies, individual differences – whether of psychology and personality, or for that matter physical capacities and appearance – were treated with respect, and even valued in and of themselves. This ethos existed in tension with egalitarianism, and such societies were also marked by mechanisms (e.g. mockery of proficient hunters) that seem designed to prevent extraordinary individuals from undermining the fundamental principles of the group.

Similar tensions might account for one startling feature of those Upper Palaeolithic burials that have been interpreted as the earliest material expressions of hierarchy or ranking in human societies. In a remarkable number of cases the bodies of these individuals bear evidence of striking physical anomalies that could only have marked them out dramatically from their social surroundings (see Cowgill, Mednikova, Buzhilova & Trinkaus 2015, with further references; Formicola 2007). They include pronounced congenital deformities (the adolescent females of Sungir and Dolní Věstonice) and examples both of dwarfism (the Romito Cave of Calabria) and extreme height (Grimaldi Cave). This leaves one to wonder if the anatomically typical skeletons similarly treated may have been those of individuals with qualities – physical or otherwise – that just as readily differentiated them from their kin, but left no traces in their skeletal remains. We can know little of the day-to-day status of those buried with rich grave goods; but in such cases we can at least suggest that they would have been seen as the ultimate individuals, about as different as it was possible to be.

What does this really tell us about the origins of social inequality? It seems unlikely that Palaeolithic Europe produced a stratified elite that just happened to comprise a high proportion of physically anomalous people. On the other hand, the ethnographic literature is full of examples of anomalous beings – human or otherwise – who are treated simultaneously as exalted and profoundly dangerous, or who alternate between the two. A being revered in life might well prove dangerous in death, or vice versa.[5] It may be relevant in this context that the very practice of burying corpses intact, and clothed, appears to have been socially anomalous in the Upper Palaeolithic. The majority of corpses seem instead to have been subject to various processes of defleshing, fragmentation, and curation. Human teeth, for instance, were made into jewellery and modified crania circulated as relics and containers (see Gambier 1992).

Palaeolithic people seem to have been very much at home with human body parts, which (properly cleansed and processed) formed an integral part of their material world. If so, then the human corpse in its complete and articulated form – and the clothed corpse perhaps even more so – was something quite unusual and, one would presume, inherently marginal and strange: incompletely absorbed either into the community of the dead or that of the living. In many of these cases, as Paul Pettitt (2011: 213) observes, an effort was clearly made to contain the bodies of the deceased by covering them with heavy mammoth scapulae, pinning them down with wood, tightly binding them, or weighing them down with stones. Saturating bodies with clothing, weapons, and ornaments may extend these concerns, celebrating but also containing the dangerous powers of extraordinary individuals.

Clearly there is no single interpretation that accounts for the full range of Upper Palaeolithic burial practices, which are both diverse and widely separated in time and space. But seeing them as evidence for hereditary systems of social ranking – as has generally been done – seems to us the most improbable interpretation of all. If anything, the ostentatious display of personal wealth was ritually associated with the same kind of ‘otherness’ seen as inherent in anomalous or exceptional individuals, and extended by the unusual practice of decorating, displaying, and burying articulated corpses. Such burials were exceptional in every sense, and can hardly be interpreted as simple proxies for social structure among the living.

They do, however, reveal the existence of elaborate and creative ritual practices, for which little evidence exists in earlier periods of human prehistory. This takes us back to the larger question of the ‘sapient paradox’. If the efflorescence of cultural creativity and symbolic expression in Upper Palaeolithic Europe is not the reflection of some new and complex form of social stratification, then how should we understand it? Here, we think, the Nambikwara example – and the larger body of ethnographic literature on which it draws – points in a very different and promising direction; one that hinges on the reversible nature of authority in societies with marked seasonal variations.

The politics of reversal: seasonal variations, social consciousness, and institutional change

In the Nambikwara case, as outlined above, top-down leadership and ingenuity were highly valued as chiefly traits during the mobile foraging season, when group sizes were small and resources scarce. By contrast, in the densely settled villages of the wet season, the chiefly role was largely one of arbitration and diplomacy. If Lévi-Strauss did not draw special attention to these seasonal variations in Nambikwara political life,[6] we suggest it was largely because early twentieth-century studies of hunter-gatherers took for granted this kind of institutional plasticity. The groundwork was laid with Marcel Mauss’s (1904–5) Essai sur les variations saisonnières des sociétés eskimo, written in collaboration with Henri Beuchat (English trans. 1979). There they defined the ‘double morphology’ of hunter-gatherer societies in the circumpolar North. Mauss, in his own later words, believed he had shown that ‘the Eskimo, and likewise many other societies ... have two social structures, one in summer and one in winter, and that in parallel they have two systems of law and religion’ (cited in W. James & Allen 1998: 37).

Mauss and Beuchat observed, for example, how the congregation of Inuit families in the long winter months was much more than an adaptive response to the presence of walrus and seal on the Arctic coast (cf. Bravo 2006). Winter aggregations brought together an extended society of both the living and the recent and remote dead, who were inaccessible to the living for much of the year. The winter houses gave expression – in wood, whale-rib, and stone – to time-transcendent principles of Inuit social life that endured even through those summer months, when groups dispersed under the authority of a single male elder in pursuit of freshwater fish, caribou, and reindeer. But many aspects of winter life also reversed the values of summer. In the summer, for instance, property rights were clearly asserted and sometimes physically inscribed onto personal objects, especially hunting weapons. But in the communalistic atmosphere of the winter house, generosity trumped accumulation as a route to personal prestige. The right of male patriarchs to coerce their sons (and indeed the group as a whole) was acknowledged only in the summer months. It had no place around the winter hearth, where the principles of Inuit leadership were turned on their head. Legitimate authority became a matter of charisma rather than birthright; persuasion instead of coercion.

In their conclusion, Mauss and Beuchat (1979 [1904–5]) drew a contrast with the tribes of the American Northwest Coast. For the Kwakiutl, inequality was most dramatic in the winter settlements, when society became structured around religious confraternities in which nobles and commoners form a hierarchy, only to give way again in the summer to smaller clan formations which, though still ranked, were less formal and coercive. What remains consistent – whether we are talking about Inuit, Nambikwara, or Kwakiutl – is the oscillation of social life between two clearly distinct systems, which accompanied seasonal changes in the material form and composition of groups. The ‘complexity’ of their moral, religious, and political systems cannot be measured on a single scale, just as their demographics – while perhaps reducible to raw population figures – are more accurately expressed as an alternation or flux between different types of mass, volume, and density.

Mauss and Beuchat’s observations, we suggest, have political implications that warrant further discussion. The different seasonal modes of existence typically involved different forms of political organization and different ways of exercising authority. What’s more – and this, for us, is the really crucial point – everyone was quite self-conscious about these differences. Among the Kwakiutl, for instance, individuals adopted different names in summer and winter seasons, literally becoming different people, depending on the time of year (Boas 1966). As a result, social structures not only became more visible as subjects of reflection; they were regularly assembled and disassembled, created and destroyed. It is surely no coincidence that much of Kwakiutl art plays visually on the relation of name, person, and role – relations laid open to scrutiny by their seasonal practices (Lévi-Strauss 1982 [1976]).

Much of this could be said to be implicit in Mauss and Beuchat’s essay; but it was not the aspect they chose to emphasize. Their own analysis tended instead to contrast the relatively pragmatic and secular existence of the summer with the intense ceremonialism of winter life:

Winter is a season when Eskimo society is highly concentrated and in a state of continual excitement and hyperactivity. Because individuals are brought into close contact with one another, their social interactions become more frequent, more continuous and coherent; ideas are exchanged; feelings are mutually revived and reinforced. By its existence and constant activity, the group becomes more aware of itself and assumes a more prominent place in the consciousness of individuals (Mauss & Beuchat 1979 [1904–5]: 76).

One can already see here the kind of language that Durkheim (under Mauss’s influence) was to use in Les formes élémentaires de la vie religieuse (1912), juxtaposing the ordinary economic life of Australian bands – concerned mainly with obtaining food – with the ‘effervescence’ of their seasonal gatherings. It was there, in the excitement of the corroboree, that the power to create society appeared to them, as if it were an alien force projected into totemic spirits and their emblems. In this account, the potential for self-conscious social transformation is never actually realized: ‘[S]ocial action follows ways that are too circuitous and obscure, and employs psychical mechanisms that are too complex to allow the ordinary observer to see when it comes’ (Durkheim 1915 [1912]: 209).

In the sociological tradition of Mauss and Durkheim, seasonality was of interest because it lay bare the mechanisms of human sociality, not so much to the participants themselves as to the outside observer. There was, however, a different strain of thought emerging from this tradition, which took a more explicitly political direction. In a largely forgotten Huxley Lecture, Robert Lowie (1948) extended his own work on the Crow to consider more general features of political organization in Great Plains societies. There, during the late summer months, small and highly mobile bands of Cheyenne and Lakota congregated in large settlements to make logistical preparations for the buffalo hunt, and for subsequent collective rituals. Lowie’s conclusions were startling, and are worth citing at some length:

In order to ensure a maximum kill, a police force – either coinciding with a military club, or appointed ad hoc, or serving by virtue of clan affiliation – issued orders and restrained the disobedient. In most of the tribes they not only confiscated game clandestinely procured, but whipped the offender, destroyed his property, and, in case of resistance, killed him. The very same organisation which in a murder case would merely use moral suasion turned into an inexorable State agency during a buffalo drive. However, ... coercive measures extended considerably beyond the hunt: the soldiers also forcibly restrained braves intent on starting war parties that were deemed inopportune by the chief; directed mass migrations; supervised the crowds at a major festival; and might otherwise maintain law and order (Lowie 1948: 18).

The ‘unequivocal authoritarianism’ that prevailed before a bison drive, and during the later Sun Dance rituals, was kept in check by the dispersal of sovereignty among tribal chiefs and police squads (‘soldiers’), and also by the ‘seasonal rhythm’ of social life on the Great Plains. ‘During a large part of the year’, as Lowie (1948: 19) noted, ‘the tribe simply did not exist as such; and the families or minor unions of familiars that jointly sought a living required no special disciplinary organisation. The soldiers were thus a concomitant of numerically strong aggregations, hence functioned intermittently rather than continuously’. Their sovereignty was no less real for its periodicity; and we must therefore accept that the Plains Indians knew something of state power (in Weber’s sense of Gewaltmonopol; see Gerth & Mills 1946: 78), without ever having developed a state. In more recent evolutionary parlance, they were a kind of band/state amalgam.

Even more critically, Lowie observed that the Plains nations – like almost all societies of the Americas – were quite self-conscious about the dangers of authoritarian power. They created explicit mechanisms to limit its abuse, rotating the clan or warrior societies that held office so that anyone holding coercive powers one year would be subject to them the next. Much of the rest of Lowie’s essay focuses on the role of chiefs, arguing that the power of political leaders over the largely ‘anarchic’ societies of the Americas was so carefully circumscribed as to exclude the internal emergence of permanent structures of coercion. Insofar as states – or indeed any peacetime powers of command – emerged in the Americas, he concludes, it could only have been through the power of prophecy, with religious figures claiming direct inspiration from the divine.

This is, of course, precisely the argument developed a generation later by Pierre Clastres in his famous (1974) essay La Société contre l’État. Clastres’s essay follows Lowie’s so closely[7] that it can only have been directly inspired by it. His argument – that stateless societies do not represent an evolutionary stage, innocent of higher organization, but are based on self-conscious rejection of the principle of coercive authority – has been enormously influential. Still the one element not carried over by Clastres from Lowie is that of seasonal variations in modes of authority; and this despite the fact that many of the Amazonian societies he discusses did have very different structures at different times of year (cf. Maybury-Lewis 1979).[8]

The result of all this, we suggest, is that the promise of Mauss and Beuchat’s early essay as a contribution to political anthropology has never been fully realized. Their insights are known to us today largely through Durkheim, who stressed the dual seasonal structure of hunter-gatherer societies but turned away from the notion of political self-consciousness; or through Clastres, who embraced the notion of political self-consciousness but disregarded the role of seasonality in structuring hunter- gatherer social organization. What, then, does this foray into early twentieth-century ethnography imply for the Upper Palaeolithic, and for social evolution in general?

Primal heterarchy

As Gregory Monks (1981) pointed out some decades ago, the full implications of seasonality studies for archaeology may only be realized if the concept is extended from its traditional focus on environmental adaptation and subsistence to include a broader array of human activities, including ritual and trade. But the more fundamental break with established theories of social evolution comes, we suggest, when we begin to consider the significance of seasonal variations for modes of social organization in their totality: in other words, cases where the same population might experience entirely different systems of economic relations, family structure, and political life at different times of year.

It is simply not possible to have an evolutionary progression such as ‘band’-‘tribe’- ‘chiefdom’-‘state’ if your starting-point is a society that moves effortlessly between institutions deemed exclusive to one category or another; or that experiences – as aspects of contemporary reality – what are supposed to be discrete stages of evolution, moving back and forth from bands to tribes or even organizations with elements of the state (such as a legitimate monopoly on the use of violence within a given territory). This may come as no surprise, and we are certainly not the first to critique the use of such models in archaeology and anthropology (see, e.g., Sherratt 1995).[9] But, more worryingly, seasonal dualism also throws into chaos more recent attempts to classify hunter-gatherers as either ‘simple’ or ‘complex’, since it assumes that supposedly diacritical features – like territoriality, social ranking, material acquisitiveness, or competitive display – will be put into effect at certain times of year, but then effectively reversed at others, routinely, within the same population.

What specific bearing do these observations have on the archaeological record of Upper Palaeolithic Europe, with its sporadic but striking evidence for social inequality? There can be little doubt that humans inhabiting the northern latitudes of the Pleistocene world experienced much sharper seasonal variations than their contemporaries elsewhere. But to what extent does archaeological evidence support the idea that their social structures alternated in harmony with such variations, for example through patterns of regular aggregation and dispersal, linked to the seasonal predation of large migratory game? And if so, how might such alternations be associated with conscious changes in political organization? The archaeological literature on this topic is voluminous, and here we can offer only a brief – but, we hope, representative – summary of the major points.

Seasonality and social evolution in the Upper Palaeolithic

The identification of aggregation sites, and their relationship to seasonal variations in hunter-gatherer ecology, is in fact a long-standing methodological problem in Palaeolithic archaeology. Attempts to tackle this problem have nearly always been informed by the ethnography of recent hunter-foragers. Lewis Binford’s (1978) pioneering ethno-archaeological work, undertaken among the Nunamuit during the 1960s, was exemplary in this respect. Motivated in part by the desire to understand such variations through their material traces, Binford (2001: 11–31) later acknowledged Mauss and Beuchat’s essay as a guiding influence. Mauss’s work on seasonality was also a stimulus for early studies of Upper Palaeolithic settlement patterns by Margaret Conkey (1980) and Randall White (1985).

Focusing on the valley systems of the French Périgord, White identified a close spatial association between the larger sites and natural ‘choke points’ along the Dordogne and Vézère, such as fords or meanders: ideal locations for intercepting herds of reindeer on their seasonal migrations. Close analysis of site size and location, combined with seasonality studies on reindeer tooth and antler (Delpech 1978), led him to propose that Magdalenian hunter-gatherers in southwest France (one of the most densely populated areas of Palaeolithic Europe) followed a ‘cycle of annual [winter] aggregation and [summer] dispersion’ – prompting him to draw direct comparisons with Mauss and Beuchat’s description of the Inuit.

Turning to northern Spain, the famous cave sites of Altamira and Castillo were identified long ago as aggregation locales based on their topographical location, the dominance of seasonally available resources (deer, ibex, shellfish) in associated faunal assemblages, and the sheer density of painted and engraved imagery within them (Straus 1977). In her 1980 study, Conkey added a detailed analysis of decoration on portable bone and antler objects, identifying patterns of spatial and stylistic variation that, in her view, supported an aggregation/dispersal model of late Upper Palaeolithic settlement on the Cantabrian coast. She further proposed a link between episodes of aggregation and heightened levels of artistic and ritual activity; but like earlier hypotheses, those of Conkey remained limited by the rudimentary excavation methods of the caves’ original investigators in the early twentieth century.

Olga Soffer’s comprehensive (1985a) analysis of Upper Palaeolithic remains on the Central Russian Plain led her to interpret differences of site-scale and complexity as evidence of seasonal variability, reflecting the sharply uneven distribution of animal resources on the periglacial ‘mammoth steppe’. Spectacular settlements such as Mezhirich and Mezin – with their mammoth-bone dwellings, abundant portable art, fixed storage installations, and imports of amber and marine shell – were suggestively aligned on major river systems (Dnepr and Desna), which also channelled the annual north-south movements of steppe bison, horse, reindeer, and mammoth. Sites lacking those features typically occurred at higher elevations, away from the floodplains, forming ‘seasonal and occupational variants of the same settlement system’ (Soffer 1985b: 238). Intriguingly the pattern here seems to have been one not of aggregation and dispersal over long distances, but of more limited oscillations between warm- and cold-weather base camps, with the latter exhibiting a greater density of trade items, personal ornaments, and elaborate architecture: a process of flux that Soffer (again echoing Mauss and Beuchat) sees as driven less by environmental pressures than by social and ideological factors (see also Soffer 1985b).

Among the most richly documented areas of Upper Palaeolithic habitation in Europe are the Pavlov Hills of southern Moravia. Prior to the Last Glacial Maximum, this region formed part of a narrow belt of forest-steppe vegetation, linking the valley of the Danube and the northern European plain, and bridging the non-glaciated zones of eastern and western Europe (Svoboda, Klíma, Jarašová & Škrdla 2000). The largest Moravian settlements, such as Dolní Věstonice I and II, are characterized by planned dwellings, extensive cooking areas, diverse craft activities, and also elaborate burials, figural art, and evidence of long-distance trade in the form of exotic stone, shell, and pigments. An abundance of plant and wetland resources, combined with exploitation of both large and small game, made year-round habitation a possibility at such locations (Mason, Hather & Hillman 1994; cf. Svoboda 2001). This possibility does not in itself preclude marked seasonal variations in the density of human activity and occupation, as indicated by impressive accumulations of mammoth remains at the majority of Moravian sites. It is still debated whether these accumulations result from large-scale, co-ordinated hunting (Musil 1994), or simply from the location of settlements adjacent to available carcasses (Soffer 1993). Either way, it is clear that seasonal abundance of bone, ivory, and frozen meat provided opportunities for social gatherings of considerable scale and intensity (Svoboda, Péan & Wojtal 2005), and various other lines of evidence support the current interpretation of these hunter-gatherer ‘mega-sites’ as aggregation points ‘where sizeable groups of people gathered between early autumn and the spring months’ (Soffer 2000: 59).

The quantity and quality of palaeo-environmental data to support such interpretations have increased markedly in recent decades. Inferences about prehistoric hunting strategies are now routinely made on the basis of bone, tooth, and antler from archaeological prey assemblages, which exhibit growth marks indicating the age of the animal at death and the season in which it was killed. Studies of this kind are supplemented by isotopic analyses to determine the migration patterns and diet of hunted game (e.g. Vlačiky et al. 2013). Rather than a uniform pattern of aggregation and dispersal, this growing body of information indicates a complex mosaic of seasonal hunting strategies and types of mobility across the forest, steppe, and tundra zones of southwest France (Pike-Tay & Bricker 1993), the Middle Danube (Nývltová Fišáková 2013), and central-eastern Europe (Péan 2001).[10]

Allowing for such regional and local variability, it seems reasonable to conclude that the movements, activities, and social lives of human populations in many parts of Upper Palaeolithic Europe were organized in accordance with pronounced seasonal variations in climate and resources, notably the annual or biennial migrations of large game. Furthermore, and as outlined above, evidence for heightened cultural creativity and social differentiation – including the elaborate funerary rituals and monumental dwellings of Ice Age Europe – clusters repeatedly at points of intense aggregation along natural (often riverine) corridors, from the Vézère to the Dnestr, offering seasonal access both to migrating herds and to an abundance of floodplain resources. This, we suggest, is no coincidence.

To be clear, we are not arguing that such seasonal variations actually caused changes in human social or cognitive capacities, at least not in any ‘hard-wired’ sense. What we instead propose is that strongly dualistic patterns of organization – such as seem likely to have existed along the glacial fringe of Upper Palaeolithic Europe – created particular opportunities for the conscious and reflexive elaboration of social structures. This is revealed in the archaeological record as an apparent explosion of expressive activities that address perennial problems of social life, such as the relations between men and women, people and animals, or life and death; and also in the instrumental use of symbolic resources, as groups and individuals explored new types of political arrangements – hierarchical and egalitarian – and ways of expressing them materially.[11]

Conclusion: farewell to the ‘childhood of man’

The archaeological record of Ice Age Europe is to the archaeologist, as the ethnographic record of the Inuit was to the anthropologist, a world of structured extremities where elementary features of human sociality, otherwise imperceptible, are laid open to investigation. Similarly structured variations may lie behind the much later phenomenon of Göbekli Tepe, where isotopic studies now link the construction of ‘stone temples’ with periods of annual superabundance, when large herds of gazelle descended onto the Harran Plain (Lang, Peters, Pöllath, Schmidt & Grupe 2013). It is relevant, in this context, that despite their monumentality, each of these massive structures appears to have had a relatively short lifespan, culminating in the rapid and deliberate infilling of its walls with the remains of large-scale feasting: hierarchies raised to the sky, only to be swiftly torn down again.

Viewed in a larger perspective, all this suggests new questions about the origins of agriculture, urbanism, and many other aspects of settled life. This, however, is not the place to explore them. Instead, by way of conclusion, we return to the question of self- consciousness and, in deference to Henry Myers, to the themes of ritual and religion. Mauss and Beuchat ended their (1979 [1904–5]) essay by suggesting that the seasonal ebb and flow of Inuit sociality – with its alternations between times of collective intensity and pragmatic, individualistic dispersal – is a general feature of all human societies. Simply put, we are incapable, psychologically and emotionally, of living in constant awareness of our full social universe. But Mauss and Beuchat also held that it was in the moments of effervescence, of ritual intensity, that we become most clearly aware of our social existence, and hence capable of creating new social forms, even if we are never quite conscious of how we achieve this.

Most contemporary theories of ritual follow a similar line of argument. The assumption is that ritual seasons – for instance, the period between Carnival and Lent in medieval Europe, or the Christmas/New Year ‘holiday season’ in modern Europe, or even individual rites of passage – are miniature versions of such ancient seasons of collective effervescence. As such they are often assumed to be, in one way or another, statements of unity and cohesion. Ritual is mostly presented as a celebration of cosmic order, which provides a foundation for social life. The most sophisticated and, to our minds, compelling formulations of this position are Maurice Bloch’s (2008) notion of the ‘transcendental’ versus ‘transactional’ realms; and Seligman, Weller, Puett, and Bennett’s (2008) argument that ritual creates a ‘subjunctive’ or ‘as if’ domain of order, consciously set apart from a reality that is always seen – in a contrasting light – as fragmented and chaotic.

These recent studies draw insights from cognitive and developmental psychology to argue that ritual is, in essence, an extension of the logic of etiquette. Social roles, corporate groups, and most everything we call ‘social structure’ does not really exist in this perspective; or, better, does not exist in the concrete, empirical way we like to imagine. It is all a kind of collective make-believe that we are continually bringing into existence, either in very small ways – such as everyday acts of respect towards elders, or saying ‘please’ and ‘thank you’ – or in very large ways – like collective rituals when abstractions such as ‘clans’, ‘moieties’, ‘movements’, or ‘nations’ are temporarily given physical form and expression.

Bloch (following Harris [2000]) has even suggested that this is precisely what the Upper Palaeolithic Revolution actually consisted of: the emergence of an apparently unique human capacity to create such imaginary and transcendent social realms, as reflected in the efflorescence of pictorial art, elaborate structures for dwelling, clothing and ornamentation, and burials in which the bodies of the deceased were organized into complex dioramas. This is a powerful line of argument, but it has always been confronted with a major problem: rituals do not always act to reinforce order, deference, hierarchy, or respect for social form. Sometimes they have just the opposite effect.

Even before the popularity of Mikhail Bakhtin’s (1993 [1940]) work on the ‘carnivalesque’, there was a lively literature about the subversive potential of seasonal festivals like the Roman Saturnalia, the medieval carnival, and May Day – their possibilities as ‘rituals of rebellion’ or attempts to create a ‘world turned upside down’. Such rituals would typically alternate between dramatic assertions of social and cosmic hierarchy, and apparently revolutionary moments where all eminences were toppled to the mud, intentionally cast into disarray. Were such processes genuinely subversive or, in the end, merely ingenious methods of maintaining social order? Such questions are no doubt as old as the rituals themselves.[12]

Looking back at the literature on seasonality and social structure, we find the same kind of confusion. The Durkheimian tradition suggests that times of seasonal aggregation should also be moments for the assertion of an ultimate collective authority, even the birth of religion itself. Yet Mauss and Beuchat’s own Inuit material suggested how just the opposite could be the case. With seasonal gatherings, the authority of fathers and husbands, rules of property and even sexual propriety, were more likely to be challenged, subverted, or simply melt away. The societies of the Great Plains created structures of coercive authority that lasted throughout the entire season of hunting and the rituals that followed, dissolving when they dispersed into smaller groups. Those of central Brazil, by contrast, dispersed into foraging bands as a way of asserting a patriarchal authority that was ineffectual in village settings. And the Kwakiutl of the Northwest Coast explored still other possibilities, granting effective police powers to performers in the Midwinter Ceremonial (the ‘bear dancers’ and ‘fool dancers’) that could be exercised only during the performance of the ritual itself.

There is no pattern here. Or, if there is one, it resides precisely in the fact that this shifting back and forth allowed mature and self-conscious political actors to be continually aware that no social order was immutable: that everything was at least potentially open to negotiation, subversion, and change. Are rituals and ritual seasons expressions of arbitrary authority or venues of social creativity? Are they, in essence, reactionary or progressive? Were our earliest ancestors simple and egalitarian, or complex and stratified? Are humans good or bad? Perhaps all these questions blind us to what really makes us human, which is our capacity – as moral and social beings – to negotiate between such alternatives.

To conclude, we do not have to choose between an egalitarian or hierarchical start to the human story. We just have to bid farewell to the ‘childhood of man’ and acknowledge – as Lévi-Strauss insisted – that our early ancestors were not just our cognitive equals, but our intellectual and philosophical peers too. Likely as not, our Palaeolithic forebears were aware, at least in a very broad sense, of many later social possibilities. Likely as not, they grappled with the paradoxes of social creativity just as much as modern theorists, and understood them – at least the most reflexive among them – just as much, which means also just as little. Perhaps this is what being ‘intellectually modern’ actually means. If there is a riddle here it is why, after millennia of constructing and disassembling forms of hierarchy, Homo sapiens – supposedly the wisest of apes – allowed permanent and intractable systems of inequality to first take root.

Notes

This is an extended version of the 2014 Henry Myers Lecture, as given by David Wengrow.

For their constructive advice, support, and criticism in preparing this article, the authors wish to thank Manuel Arroyo-Kalin, Maurice Bloch, Ignacio de la Torre, Ewa Domaradzka, Clive Gamble, Erhard Schüttpelz, Alpa Shah, and Stephen Shennan. David Wengrow is grateful to the Council of the Royal Anthropological Institute for the opportunity to deliver the Henry Myers Lecture.

References

[1] Unlike cognitive or other biologically based explanations, the demographic model is also compatible with sporadic but widespread evidence for behavioural modernity in the African Middle Stone Age (see again McBrearty & Brooks 2000), since what it seeks to explain is not the origin of the behaviours in question, but their peculiarly dense manifestation in the archaeological record of the European Upper Palaeolithic. Gamble (2012) discusses some potential weaknesses of the model, such as genetic and linguistic evidence for high population densities in tropical regions with very different archaeological records.

[2] See Britton et al. (2011). Based on seasonality data from gazelle remains, Daniel Lieberman (1993) argues that – in the prehistoric Levant (Israel/Palestine/Jordan) – anatomically modern humans (AMH) were considerably more mobile than their Neanderthal contemporaries, with only the former practising long- range seasonal migration between habitats in pursuit of game. Steven Kuhn and Mary Stiner (2006) further suggest that only AMH regularly supplemented large game with a wide spectrum of small mammal and plant resources, developing a sophisticated division of age and gender roles in order to do so. Such diversification strategies, they propose, most likely developed in tropical or sub-tropical environments, but would have had the greatest returns – in terms of cultural and demographic expansion – in the more seasonally variable habitats of Upper Palaeolithic Europe.

[3] In a recent article the prehistorian Clive Gamble (2013) makes an eloquent case for the ongoing relevance of classic sociological theory to Palaeolithic studies and human origins. While focusing on the contribution of the Année Sociologique, Gamble does not, however, develop the specific aspects of their work on seasonality that we are concerned with here.

[4] In an influential study, James Woodburn (1982) subsequently identified an important distinction in the economic systems of recently documented hunter-gatherers. He distinguished between systems in which people receive a direct and immediate return from their labour, and those in which material and social assets are stored in order to obtain ‘delayed returns’. Woodburn further argued that strategies of ‘assertive egalitarianism’ – such as prohibitions on the monopoly of violence, or on the accumulation of wealth and technological skills – are most likely to succeed in societies of the ‘immediate-return’ type. He himself was cautious about the extension of this dichotomy to prehistoric hunter-gatherers (1982: 447), and clearly it does not allow for the kind of conscious alternations in social and moral codes that we discuss here.

[5] In Mary Douglas’s (1966) formulation, anomalous beings are more likely to be treated as sacred in social orders open to the surrounding world, and as abominations in those that emphasize group boundaries. The former seems a better fit for those Upper Palaeolithic societies where ‘princely’ burials occur, and in which long-range movements and exchanges of materials and populations seem to have been commonplace.

[6] Based on field research conducted thirty years later, Paul Aspelin (1976) argued that Lévi-Strauss and other early observers of the Nambikwara had produced an over-simplified account of their ‘dual economic system’. Focusing mainly on subsistence practices, Aspelin found considerable overlap between the activities of the dry and wet seasons. Lévi-Strauss (1976: 32) clarified his position in response, noting that the early accounts of Nambikwara economic dualism were ‘corroborated by the missionaries who, having lived for ten years in contact with the natives, had ample time to get acquainted with their seasonal moves’ (1976: 32), and suggested that the subsequent construction of airfields and highways across Nambikwara land may have considerably altered their patterns of mobility.

[7] For example, in its outline of chiefly authority as consisting of peacemaking, hospitality, and oratory. Clastres does not cite Lowie; but in general he cites only ethnographic sources and never theoretical ones.

[8] A common objection to Clastres’s argument is to ask how Amazonian societies could have consciously organized themselves against the emergence of forms of authority they had never actually experienced. But, as demonstrated by the Nambikwara example or those of the Gê and Bororo societies of central Brazil – who break up their wet season villages to form smaller ‘trekking’ bands under the authority of male elders – this is not so much of a mystery as sometimes suggested (cf. Gross 1979).

[9] The tradition of ‘oppositional thinking’ (e.g. bands versus tribes, etc.) on which such models are based has its roots in the ‘stadial’ evolutionism of the Scottish Enlightenment, which insisted both on the essential singularity of human social forms, and on their direct correlation with modes of subsistence (see O’Brien 1993).

[10] In the latter regions, seasonal mobility has been further linked to the large-scale movement of flint for manufacturing tools and weapons, notably between the Kraków Basin (in southern Poland) and the Vag River (in western Slovakia), where an overwhelming majority of stone tools are made on exotic raw materials, originating between 60 and 300 km away from their places of discovery (Kozłowski 1989).

[11] It may be significant in this respect that the much earlier human occupation at Blombos Cave – with its clear signs of cultural complexity and symbolic expression – also formed part of a seasonal migratory round, as indicated by recent studies of blue antelope dentition from Middle Stone Age deposits (Thompson, Faith & Henshilwood in press). Similar correlations between pronounced seasonal variations and evidence of ‘cultural complexity’ might also be sought in other areas of early human expansion, such as the later Palaeolithic of the Indian subcontinent (cf. H.V.A. James & Petraglia 2005) and western Asia (Maher et al. 2012).

[12] As Peter Burke (2009: 283–5) notes, the idea that rituals of rebellion were simply ‘safety valves’ or ways of allowing common folk to ‘let off steam’ is first documented only two years after the invention of the steam engine – the favoured metaphor had earlier been to let off the pressure in a wine cask. At the same time, however, medieval authorities were keenly aware of the fact that most peasant revolts or urban insurrections would begin precisely during such ritual moments (see Bercé 1976). Those who turned the world upside down were often reluctant to put it back the right way up again. Consider also Roger Caillois’s seminal essay on ‘the festival’, written for Georges Bataille’s Collège de Sociologie in the 1930s (trans. 2001 [1939]). It went through two drafts, the first holding forth the festival as a model for revolutionary social liberation, the second as a harbinger of fascism.