Against the Grain: A Deep History of the Earliest States

by James C. Scott

Sedentism arose in very special and circumscribed ecological niches, particularly in alluvial or loess soils. Later—much later—the first centralized states arose in even more circumscribed ecological settings where there was a large core of rich, well-watered soils and navigable waterways, capable of sustaining a good number of cereal-growing subjects.

this respect, the “Uruk world system,” despite the grandiosity of the term, may well have prefigured, on a smaller scale, the integration of the Chinese, Indian, and Mediterranean disease pools around the year 1 BCE that is seen to have touched off the world’s first devastating pandemics, such as the sixth-century CE Plague of Justinian, which killed between thirty million and fifty million people.

There is no way of knowing for certain how frequently epidemics brought down the earliest states, but, amplified by warfare, invasions, and trade, diseases were a prominent cause of deurbanization in late Imperial Rome and in medieval Europe. In 166 CE Roman troops returning from a campaign in Mesopotamia brought home an infectious disease that may have killed a quarter to a third of Rome’s population.

One last and more speculative consequence of deforestation and siltation is its role in the propagation of malaria. It has been suggested that malaria is a “disease of civilization,”

“vacant” periods represented a bolt for freedom by many state subjects and an improvement in human welfare.

We should, I believe, aim to “normalize” collapse and see it rather as often inaugurating a periodic and possibly even salutary reformulation of political order.

Above all, the well-being of a population must never be confounded with the power of a court or state center. It is not uncommon for the subjects of early states to leave both agriculture and urban centers to evade taxes, conscription, epidemics, and oppression.

States, being largely agrarian phenomena, would, with the exception of some intermontane valleys, have looked like small alluvial archipelagoes, located on the floodplains of a handful of major rivers. Powerful as they might become, their sway was ecologically confined to the well-watered, rich soils that could support the concentration of labor and grain that was the basis of their power.

hinterland was not simply an ungoverned—or better put, a not-yet-governed—zone, but rather a zone governed, from the perspective of the state center, by “barbarians” and “savages.” Though hardly precise Linnaean categories, “barbarians” often denoted a hostile pastoral people who posed a military threat to the states but who might, under certain circumstances, be incorporated; “savages,” on the other hand, were seen as foraging and hunting bands not suitable as raw material for civilization, who might be ignored, killed, or enslaved.

But I wish to argue here that the threat posed by barbarians was perhaps the single most important factor limiting the growth of states for a period measured more in millennia than in centuries.

The period between the first appearance of states and their hegemony over nonstate peoples represented, I believe, something of a “golden age of barbarians.”

He points to state walls and fortification against nonstate peoples springing up from western Europe through central Asia into China, and lasting until the Mongol invasions of Europe in the thirteenth century. It seems a rather extravagant claim, but, coming as it does from Lattimore, it merits pondering. “There was a linked chain of fortified northern frontiers of the ancient civilized world from the Pacific to the Atlantic. The earliest frontier walls appear to have been in the Iranian sector. The walled frontiers of the western Roman Empire in Britain and on the Rhine and Danube faced forest, upland, and meadow tribes, now pastoral nomads.”

States were likely to arise only in rich, well-watered, bottomland soils. Until the last half of the first millennium BCE, when larger, sail-driven ships could transport larger cargoes longer distances, states had to hug the grain core quite tightly. Barbarian geography and ecology is, on the other hand,

great many apparently ethnic names turn out to be, when translated literally, a description of a people’s geography, applied to them by state discourse: “hill people,” “swamp dwellers,” “forest people,” “people of the steppes.” The only reason pastoral nomads of the steppe, mountain people, and sea people figure so prominently in state discourse about barbarians is that such peoples were not only out of reach but were also the most likely to pose a military threat to the state itself.

core. In a Sumerian myth, the goddess Adnigkidu is admonished not to wed a nomad god, Martu, as follows: “He who dwells in the mountains . . . having carried on much strife . . . he knows not submission, he eats uncooked food, he has no house where he lives, he is not interred when he dies . . .” One can scarcely imagine a more telling mirror image of life as a grain-producing, domus-based state subject.13 The Record of Rites (Liji) of the Zhou Dynasty contrasts the barbarian tribes who ate meat (raw or cooked) instead of the “grain food” of the civilized. Among the Romans, the contrast between their diet of grain and the Gallic diet of meat and dairy products was a key marker of their claim to civilized status.

Over time an increasingly large proportion of nonstate peoples were not “pristine primitives” who stubbornly refused the domus, but ex–state subjects who had chosen, albeit often in desperate circumstances, to keep the state at arm’s length. This process, detailed by many anthropologists, among whom Pierre Clastres is perhaps the most famous, has been called “secondary primitivism.”14 The longer states existed, the more refugees they disgorged to the periphery. Places of refuge where they accumulated over time became “shatter zones,” as their linguistic and cultural complexity reflected that they were peopled by various pulses of refugees over an extended period.

A great many barbarians, then, were not primitives who had stayed or been left behind but rather political and economic refugees who had fled to the periphery to escape state-induced poverty, taxes, bondage, and war. As states proliferated and grew over time, they ground out ever greater numbers who voted with their feet. The existence of a large frontier—rather like migration to the New World for poor Europeans in the nineteenth and early twentieth centuries—provided a less dangerous avenue of relief than rebellion.

Tribes are, in the first instance, an administrative fiction of the state; tribes begin where states end. The antonym for “tribe” is “peasant”: that is, a state subject. That tribality is above all a relationship to the state is captured nicely by the Roman practice of reverting to the use of former tribal names to describe provincial populations that had broken away and rebelled against Rome.

A great many of the so-called tribal names were simply place names: a particular valley, a range of hills, a stretch of river, a forest. In some cases the term might designate the character of the presumed group—for example, a group the Romans called Cimbri, which means “robbers” or “brigands.”

Knowing this, raiders are most likely to adjust their strategy to something that looks more like a “protection racket.”

Protection rackets that are routine and that persist are a longer-run strategy than one-time sacking and therefore depend on a reasonably stable political and military environment.

The arrangement would be sealed by a treaty in which the nomads appear as tributaries and make the appropriate performance of allegiance in return for large subsidies. The “reverse” tribute was enormous: one-third of the annual government payroll went to buying off the nomads. Seven centuries later, under the Tang, officials were delivering half a million bolts of silk to the Uighurs annually on similar terms. On paper it may have looked as if the nomads were tributary inferiors to the Tang emperor, but the actual flow of revenue and goods suggests the opposite in practice. The nomads were, in effect, collecting bribes from the Tang in exchange for not attacking.

The early agrarian states and the barbarian polities had broadly similar aims; both sought to dominate the grain-and-manpower core with its surplus. The Mongols, among other raiding nomads, compared the agrarian population to ra’aya, “herds.”

Both were slaving and raiding states in which the major booty of war and the major commodity in trade were human beings. In this respect they were competing protection rackets.

Suddenly the periphery and semiperiphery of the early states were the sites of valuable commodities for which there was now an appreciable market. Foraging, hunting, and marine collecting became lucrative commercial activities.

The most tendentious of these pairs, the civilized-barbarian pair, are born together as twins. Lattimore has articulated this “dark twin” thesis most clearly:

Although Lattimore ignores the millions of nonstate foragers, shifting cultivators, and marine collectors who were not pastoralists, he does capture the parallel evolution of nomadism and states. These nomads, most especially those on horseback who “plagued” state centers, are best seen simply as the strongest competitors of the state for control of the agrarian surplus.

mounted pastoralists were designed to extract wealth from sedentary states; they were a “state in waiting” or, as Barfield puts it, a “shadow empire.”

the most robust cases, such as the itinerate state founded by Genghis Khan, the largest contiguous land empire in world history, and the “Comanche Empire” in the New World, we would be better advised to think of them as “horseback states.”

Furthermore, according to this view, barbarian confederations operate as “shadow empires” adjacent to and parasitic on large sedentary polities. Their quasi-derivative status is emphasized by the fact that they tend to disappear when their host collapses. As Nikolay Kradin puts it, “The degree of centralization