The task I’ve set myself here is to think about the history of civilisation before the state. Why should this be so difficult? And why is it nevertheless a task worth pursuing? The answers lie partly in a set of widespread assumptions about what constitutes a civilisation in world history. It is common, for example, to group a whole series of ancient societies together under the banner of ‘early civilisation’. The ones that usually get included are ancient Mesopotamia, ancient Egypt, the Classic Maya, the Aztecs, Shang China, the Inca Empire and the Yoruba kingdoms of West Africa. But when it comes to defining what these particular societies have in common, the concept of civilisation seems to drop out of the equation. Suddenly the focus shifts, without explanation, to other factors that can be easily described without any use of the term civilisation – factors such as class stratification, urbanisation, centralised (and often literate) administration, sacred kingship, economic exploitation of the many by the few and so on. ‘Civilisation’, by this point, has simply become an umbrella term for a whole cluster of other cultural attributes that are basically to do with the effective exercise of power by a small and determined elite. In other words, ‘early civilisation’ and ‘state formation’ have become indistinguishable from one another as historical and descriptive categories.

Sensing the trap of a circular argument, some anthropologists have tried, albeit tentatively, to separate these two things out. For instance, in cases of early civilisations that are made up of politically independent city-states – such as the Classical Maya or ancient Mesopotamia – the term ‘civilisation’ is sometimes used to refer to the shared cultural and cosmological milieu within which multiple states exist: a kind of overarching set of guidelines about the proper moral relationships among mortals, kings and gods that encompasses the strategic rivalries of political factions. Yet the more fundamental equation between state and civilisation as coeval stages of social development stands largely unquestioned. This might be defensible were it possible to argue that only the power of centralised states is capable of generating such large-scale patterns of cultural uniformity and moral consensus. But as we shall see, this is very far from being the case.

Also largely unquestioned are the wider historical implications of a term such as ‘early civilisation’, which must then imply that there is also such as a thing as ‘late’ or ‘developed’ civilisation. Where then would we locate the transition from one to the other? Was medieval Europe – with its sacred kingdoms and administrative elites – an early or a late civilisation? Have we, in fact, until very recently – say, until the political revolutions of the eighteenth to twentieth centuries – been living under social and political conditions that are basically analogous to those of ancient Egypt or Mesopotamia? How different in structure were the ‘old regimes’ of Europe from what are sometimes described as ‘archaic states’ or ‘early civilisations’? Fencing off early civilisation as a distinct stage in human history can quickly become an arbitrary and subjective affair.

No less arbitrary, it seems, are the initial criteria on which current definitions of ‘early civilisation’ are grounded. Urbanism, for instance, was sometimes characteristic of societies that actually exhibit very little evidence for pronounced class stratification. To continue describing the appearance of 250-hectare settlements in fourth millennium BC Mesopotamia as an ‘urban revolution’ begins to look a little strange when we realise that, during the same period, the prehistoric societies of Eastern Europe – from the Carpathian Mountains to the Dniepr River – were forming settlements of over 400 hectares. The Ukrainian site of Talianki, for instance, is thought to have held something in the order of 10,000 inhabitants. What is so striking about these earliest European cities is that they achieved such immense sizes with no apparent need for centralised government, bureaucracy or a political elite. And their ground plans – forming concentric rings of similarly designed households – suggest a robustly egalitarian ethos. Yet they are usually excluded from the roster of ‘early civilisations’ and their (potentially enormous) implications for how we understand the root causes of social inequality go largely unrecognised. Studies of human political evolution seem to remain, for the most part, oblivious to the whole phenomenon and happily continue to assume that urban egalitarianism was something confined to a very brief period of mid-twentieth-century Catalonian history.

Similar points could be made about bureaucracy. Although often included as a key component within the list of attributes shared by ‘early civilisation’, complex administration in fact turns out to predate the emergence of cities and kingship by thousands of years. In Upper Mesopotamia – what is now inland Syria and northern Iraq – there is clear evidence for the use of complex bureaucratic devices, such as commodity seals and economic archives, in small-scale farming communities as far back as the 7th millennium BC. Why such devices were adopted in what must essentially have been face-to-face societies remains something of a mystery. And, again, the writers of general sourcebooks on world history seem largely content to ignore the facts and continue to describe specialised administration as a distinct evolutionary feature of urban civilisations.

It is worth adding at this point that the invention of the first writing systems was – in many parts of the world – a gradual development from these much earlier systems of bureaucratic notation, rather than a revolutionary innovation. This is perhaps most clearly evident in the case of Mesopotamian cuneiform, which has been shown to evolve from systems of numerical representation that can be traced back – on at least one interpretation – almost to the origins of farming itself.

What, then, about class or caste distinctions and other forms of social inequality? Can these be more legitimately regarded as defining criteria of ‘early civilisation’? Surely not, since archaeologists have carefully traced clear and pronounced evidence of status differentiation back to the period of Palaeolithic hunter-gatherers, long before the origins of sedentary life, let alone cities or literacy. Consider, for example, the exceptionally preserved site of Sungir in Russia, which offers striking evidence for the concentration of material wealth – prestigious weapons and thousands of finely worked body ornaments – within the burials of a middle-aged man and two children; or, indeed, the monumental stone temples of Göbekli Tepe in southern Turkey, built over 10,000 years ago by hunter-gatherers, who were contemporaries and neighbours of the first farmers in the Fertile Crescent.

Deducing the political structures of prehistoric societies from their material remains is no simple task. But at the very least, such dramatic findings should lead us to question the persistence – even in quite specialised literature – of a Rousseau-like vision of the childhood of man, in which hunter-gatherer societies are still portrayed as invariably small and essentially egalitarian bands. It is worth noting in this context that, when disturbingly hierarchical features are found among hunter-gatherer societies, the tendency has been to compare them with the playing out of dominance hierarchies among chimpanzees and other higher primates, as opposed to viewing them in the perspective of other types of human political systems, such as those of early states and ‘civilisations’. The implication, whether or not the authors of such views realise it, seems to be either that hunter-gatherers are in some sense closer to chimpanzees than they are to other human beings, or that politicians as a distinct social class have more in common with chimpanzees than they do with most other human beings. I personally find the latter interpretation to be the more plausible. At a more general level, we might also ask why some people find it necessary – even in the face of contradictory evidence – to equate ‘civilisation’ with the growth of authoritarian, agro-industrial states? What is really at stake here?

The answers may often have as much to do with the recent past – and the present contexts in which historical research is conducted – as they do with remote antiquity. Let me offer some examples from my own research, which falls quite squarely within what is often called the comparative study of ‘early civilisations’ and, in particular, those of Egypt and Mesopotamia, often deemed to be the earliest known examples of this distinct type of society. In studying their development from the end of the last Ice Age to the period of the first monarchies, I have always struggled with terms such as ‘early state formation’ and ‘early civilisation’. I think the reason for my unease is that such terms seem to imply that, whatever is said must be conceptually divorced from an understanding of modern state formation and, indeed, from the kind of modern civilising processes that were described, respectively for northern and Mediterranean Europe, by the sociologist Norbert Elias and the historian Fernand Braudel. In reality, however, the arguments we make about the emergence of states and civilisations in early human history are always informed by our grasp of contemporary political life, just as they are limited by our inability to envisage alternative forms of government and civil society to our own.

For example, it was argued for a long time that ancient Egypt – the world’s first known territorial state and an indisputable claimant to the title of ‘early civilisation’ – must have been established on the basis of a Neolithic farming economy, essentially like that of Mesopotamia; in other words, an agrarian economy based mainly on the cultivation of cereal crops in settled villages. This interpretation – which is enshrined in the historical sociology of Marx and Weber – has always seemed odd to me, for reasons that I can briefly summarise. First, the Neolithic sites in the Nile Valley that are often described as ‘villages’ resemble nothing more closely than the cumulative remains of seasonal camps, of a kind found until recent times in mobile, pastoral societies throughout much of East Africa and with prehistoric parallels in areas such as the Deccan Plain of southern India. They present no clear evidence for durable domestic architecture, or for the characteristic accumulation of mud-brick remains that, in Mesopotamia, gives rise to the mounds known as tells. Assemblages of animal remains from these sites suggest the adoption of a mixed herding economy some centuries before the use of domestic cereals, a phenomenon that has now been documented across much of northern Africa. Moreover, the material culture of the Neolithic Nile Valley, dating to the fifth millennium BC, has a strikingly mobile and portable character. It includes a complex array of tools for cosmetic adornment, such as stone palettes used to grind and prepare body paint, bone and ivory combs, and a stunning profusion of coloured beads and other personal ornaments, which are found in impressively similar forms throughout both Upper Egypt and Sudan.

Cereal farming and settled village life would indeed become crucial elements in the emergence of the early Egyptian state. But they were grafted onto an earlier form of Neolithic civilisation that was markedly different in character: oriented more strongly towards a mobile pastoral economy, combined with seasonal hunting and foraging, and culturally focused – less upon the household and village than upon the elaboration of the body, in life and in death, as a way of forming enduring attachments between person and place. As we will see later on, this distinctive form of early pastoral civilisation left a clear imprint on the types of political organisation that were later to emerge in Egypt. Why, then, was it excluded for so long from our understandings of Egyptian state formation?

Part of the answer, I suggest, lies in an over-literal application of ethnographic analogies to the archaeological record. With varying degrees of explicitness, archaeologists have often taken anthropological descriptions of modern herding societies – the Bedouin of the Negev desert, or Nuer cattle-keepers on the marshes of South Sudan – as direct guides to the scale and organisation of pastoral societies in prehistory. Edward Evans-Pritchard’s classic ethnography of the Nuer has been particularly abused in this regard. Conducted during the period of Anglo-Egyptian colonial rule in Sudan, his research established the Nuer tribal system as a paradigmatic model of societies that resist political centralisation, responding to situations of crisis and conflict through temporary alliances, rather than through the establishment of stable leadership roles.

Comparisons of this sort, when applied to the archaeological record, create something of a paradox. If the kind of situation described by Evans-Pritchard for the Nuer under British rule really does represent the type of social organisation that preceded the emergence of ancient states, then how did such states ever emerge in the first place? It was precisely this paradox that occupied Pierre Clastres in his analysis of Amazonian tribal systems, La Société contre l’État, a study written in the 1970s that has been strangely ignored by archaeologists. How, he asked, could small-scale societies of the kind classically described by anthropologists – societies that showed no sign of containing even the germ of state formation – have given rise to the kind of hierarchical political structures that have appeared in both the Old and New Worlds over the past five thousand years?

Clastres’ radical solution was to argue that recently observed populations of foragers, pastoralists and swidden cultivators – far from being relics of Neolithic lifeways – are more often made up of groups that consciously resisted participation in settled lowland agriculture, rejecting a civilising process imposed on them from outside, together with its realities of forced labour, excessive taxation and the decimation wreaked upon them by epidemics that spread through mission stations and the mono-cropping estates of sedentary farmers. According to Clastres, the social structures of marginalised groups, such as the Tupi of Amazonia or the San of the Kalahari, were deliberately cultivated to avoid the predatory interests of agro-industrial states and to prevent the emergence of state-like structures within their own midst.

The point is amplified by James C Scott in his recent book, The Art of Not Being Governed, a history of the region called ‘Zomia’ by Willem van Schendel. Extending from the northern frontiers of India to the western borders of China, this rugged highland zone is currently thought to be home to well over 100 million people classified as minorities: an enclave of bewildering linguistic and ethnic diversity, located at the peripheries of nine different nation states. This situation, argues Scott, is the result of a 2,000-year history of relatively continuous persecution and exploitation, in which groups have looked to the mountains as a refuge from state-making projects pursued by successive lowland governments, from the Han Empire to the Cultural Revolution and, indeed, up to the present day.

The perspectives developed by Clastres and Scott encourage us to take a fresh look at the process we call ‘early state formation’ or ‘early civilisation’. It sometimes seems to be assumed that most human beings desire to live in state-like structures and that they have been working diligently towards that goal ever since the inception of farming economies at the end of the last Ice Age. Such expectations are enshrined, not only in the theoretical language of social evolution but also in the chronological terms given to those transitional phases that come directly before state formation: ‘proto-urban Mesopotamia’, ‘proto-palatial Crete’, ‘proto-dynastic Egypt’ and ‘archaic Greece’, all expressing a certain impatience with people who really ought to have figured out, by that point, what they were aiming for – namely a literate, bureaucratic state with stable mechanisms for extracting economic surplus from a large body of sedentary dependents and a stable machinery of enforcement to ensure that everything goes smoothly. In continental Europe, the phrase ‘proto-history’ continues to define a whole series of encounters in which peoples, such as the Scythians or the Celts, emerge suddenly into the light of narrative history through the writings of literate colonisers who were often trying to extract fiscal revenue from them. Post-colonial theory, perhaps because of its tendency to eschew grand narratives, has done little to revise these overarching structures of historical thought, which lead almost inevitably to a view of history as a game of winners and losers – (‘the West versus the rest’, as one British historian put it) – in which the barbarian emerges onto the scene like a shooting star, destined to fizzle out back in the darkness from whence he came.

An alternative approach to the long-term study of civilisation might begin by drawing a basic but fundamental contrast between modern and ancient geographies of state power. Modern nation states exist in a crowded world, made up of similar territorial entities, whose official borders carve a passage through even the most inhospitable deserts and mountain ranges. The kind of tribal refuge zones described by Scott, although sometimes large in scale, are now very limited in number. To comprehend the relationship between state and civilisation in earlier epochs of history, we have to perform a feat of the imagination, for which there is no theoretical shortcut. In short, we have to reverse the picture.

Take, for example, the situation in the Middle East during the Early Bronze Age, between 4,000–5,000 years ago. Imagine the heartlands of urban life compressed along the floodplains of the Tigris and Euphrates, where people – for the first time in history – lived within the centralised political units known as city-states: places such as Ur and Lagash with their literate bureaucracies, populations in the tens of thousands and agricultural hinterlands supplying revenue to the centre. It was in such areas that not only cities, but also the world’s first peasantries, emerged at this time. In fact, the first appearance of cities in this region dates back a full thousand years earlier than the period in question, to around 4,000 BC. Yet after a millennium or so of exposure to urban civilisation, the result is something quite different from what our evolutionary models might predict. In terms of spatial extent (as opposed to demography), the territories directly controlled by urban states in the Early Bronze Age were actually quite limited. Around them – reaching across the highlands of the Taurus and Zagros ranges – extend more thinly populated areas, occupied by relatively small-scale societies that show few archaeological signs of political centralisation, despite their proximity to urban states.

What is immediately striking about these supposedly marginal areas is the degree to which they are linked together through extensive networks of cultural interaction. Perhaps most impressive is the so-called ‘Trans-caucasian’ cultural zone, which extends all the way from the Caucasus down across the northern margins of the Fertile Crescent in both directions and, within which, relatively small-scale and egalitarian communities shared specific methods of food preparation, house construction and rituals for the dead. To refer to this entity simply as an archaeological ‘culture’ or ‘culture area’ is to elide the fact that, unlike its Neolithic predecessors, this particular network of decentralised communities took form in close proximity to urban centres: a ‘civilisation against the state’, as Adam T. Smith (similarly following Clastres) has recently put it.

Admittedly much local variation can be found within this Trans-caucasian civilisation. Nevertheless, the features shared throughout it are suggestive of societies that may have defined themselves in conscious opposition to nearby states. Intriguingly, this may even be evident in their modes of cuisine. Among the types of material culture found throughout the network, from the Caucasus to the Jordan Valley, are ceramic hearths decorated with human-like faces, on which food was prepared in highly burnished vessels, topped with purpose-made lids. This method of boiling and stewing food in closed containers stands in contrast to the roasting and baking traditions of the urban lowlands, where the ritual preparation of food was conducted in open containers, or on exposed altars, so that the upward release of fumes from a sacrificial meal could attract the attention of the gods down to their human subjects.

Such culinary contrasts may be the stuff on which civilising missions rise or fall, as Catholic missionaries to the New World discovered, when confronted with the native Tupi, whose aversion to the baked substance we call bread proved an obstacle to their acceptance of the Holy Communion, itself a ritual descendant from the wine and cereal-based rituals of the ancient Mediterranean and Middle East. In the contrasting distributions of methods for food processing and preparation, we can perhaps detect the kind of conscious differentiation between state and non-state societies discussed by Clastres and Scott for more recent situations. But in the case of the Bronze Age, the tables are turned: this was a ‘world of peripheries’, where cities and state-centres rest like small islands amid a great sea of stateless civilisations.

I have said little as yet about why urban and state-like societies ever emerged in the first place. What I will offer by way of conclusion are some very brief and admittedly broad-brush observations. The first concerns utopian visions. It is striking that each of the earliest centres of urban civilisation presents us with a scaled-up and spectacular version of cultural values that extend back, in the same regions, to much earlier periods of prehistory. I am thinking here of the first monumental precincts at the city of Uruk in southern Iraq – designed as vastly expanded versions of a common household form, found in almost every Mesopotamian village during the pre-urban period; but also of the earliest royal monuments in Egypt – ceremonial versions of personal display items, the use of which (as we saw earlier) has deep Neolithic roots in the Nile Valley. We might think in similar terms of the great bathing facilities at the heart of Mohenjo-Daro, on the plains of the Indus.

In each case, time-honoured and familiar concepts of domesticity, wellbeing, or cleanliness were reproduced on a greatly magnified scale. For all their exclusionary qualities, we can hardly doubt that these early centres offered their dependants an image of cosmological perfection. It was in this fragile world of bread and circuses that the best and the worst of human nature conspired to produce what we now recognise as states. Yet the values of civilisation in which such political projects were grounded were both older and more durable than the projects themselves and were never truly encompassed by them, even at the height of ancient empire. By reducing our definition of ‘early civilisation’ to the formation of states, we risk losing sight of these much longer and more spatially extensive trajectories of cultural change, the roots of which must be sought in the development of prehistoric societies that succeeded – for millennia – in maintaining distinct forms of civilisation, while avoiding the emergence of states.

This essay originally appeared under the title ‘Civilisation Before the State’ in ‘Civilisation: Perspectives from the Engelsberg Seminar’, Bokförlaget Stolpe, in collaboration with Axel and Margaret Ax:son Johnson Foundation, 2013.