Why this sudden restlessness, this confusion?
(How serious people’s faces have become)
Why are the streets and squares emptying so rapidly?
Everyone’s going home lost in thought
Because night has fallen and the barbarians have not come
And some who have just returned from the border say
There are no barbarians any longer
And now, what’s going to happen to us without barbarians?
They were, those people, a kind of solution.
—C. Cavafy, “Waiting for the Barbarians,” 1914
In effect the essential thing is to gather into groups this people which is everywhere and nowhere; the essential thing is to make them something we can seize hold of. When we have them in our hands, we will then be able to do many things which are quite impossible for us today and will perhaps allow us to capture their minds after we have captured their bodies.
—French officer, Algeria, 1845
This people have never turned their attention to agricultural pursuits, nor can it be expected of them until they are placed upon a reservation.… If they are not provided with such a home, they are destined to remain outside of those influences which are calculated to civilize or Christianize them … [and] render [them] useful members of society. Wild Indians, like wild horses, must be corralled upon reservations. There they can be brought to work.
—Bureau of Indian Affairs agent to the Shoshone, 1865
The permanent settlement of populations is, along with taxes, perhaps the oldest state activity. It has always been accompanied by a civilizational discourse in which those who are settled are presumed to have raised their cultural and moral level. While the rhetoric of high imperialism could speak unself-consciously of “civilizing” and “Christianizing” the nomadic heathen, such terms strike the modern ear as outdated and provincial, or as euphemisms for all manner of brutalities. And yet if one substitutes the nouns development, progress, and modernization, it is apparent that the project, under a new flag, is very much alive and well.
What is striking about this civilizational discourse is its staying power. Its permanence is all the more remarkable in the light of evidence that ought to have shaken it to its very foundations. It survives despite our awareness that people have been moving, for millennia, back and forth across this semi-permeable membrane between the “civilized” and the “uncivilized” or the “not-yet-civilized.” It survives despite the perennial existence of societies that occupy an intermediate position socially and culturally between the two presumed spheres. It survives despite massive evidence of cultural borrowing and exchange in both directions. And it survives despite an economic integration driven by complementarity that makes of the two spheres a single economic unit.
Much of the actual content of what it means to be “civilized,” to be “Han,” to be a proper “Thai” or “Burman” is exhausted by being a fully incorporated, registered, taxpaying subject of the state. Being “uncivilized” is, by contrast, often the converse: to live outside the ambit of the state. Much of this chapter is devoted to examining how state formation creates, in its wake, a barbarian frontier of “tribal peoples” to which it is the pole of comparison and, at the same time, the antidote.
Legitimating the classical state in Southeast Asia was, in modern parlance, a “hard sell.” The very idea of the classical state, far from being an organic elaboration of indigenous concept of rule, was, like the modern nation-state, largely a cultural and political import. The Hinduized concept of the universal monarch provided the ideological apparatus to support a claim to ritual supremacy in a political context otherwise characterized by contending, and presumptively equal, strongmen. With the help of court Brahmins, ambitious courts from the tenth to the fourteenth century set about making large cosmological claims and incorporating local cults of their outlying provinces under an imperial ritual umbrella.1 The effect was something like the eighteenth-century Russian court at St. Petersburg mimicking the manners, language, and rituals of the French court at Versailles. Making such a claim “stick” required not just convincing theater, as Geertz has shown, but a core population whose manpower and grain could reinforce the court’s claim. And this, in turn, required coercion in the form of slaving expeditions and a system of unfree labor. The classical state, in short, was anything but self-legitimating. It was perhaps for this reason that the cosmological bluster of such states tended to compensate for their relative weakness politically and militarily.2
Given the fact that such states were created by an ingathering of various peoples living outside state structures, it is not surprising that the major elements representing a “civilized” existence happen to coincide with life in the padi state: living in permanent villages in the valleys, cultivating fixed fields, preferably wet rice, recognizing a social hierarchy with kings and clerics at its apex, and professing a major salvation religion—Buddhism, Islam, or, in the case of the Philippines, Christianity.3 Nor is it surprising that each of these characteristics should be the mirror image of the surrounding societies remaining outside the sphere of the padi state: the hill peoples.
Viewed from the court center of the padi state, the thinner the air you breathe, the less civilized you are. It is no exaggeration to say that the presumptive level of civilization can, from a valley perspective, be often read as a function of altitude. Those on the mountaintops are the most backward and uncivilized; those living midslope are slightly more elevated culturally; those living in upland plateaus and growing irrigated rice are, again, more advanced, though certainly inferior to those at the core of the valley state, with the court and king at its apex, who represent the pinnacle of refinement and civilization.
“Hilliness” per se is disqualifying. Thus many of the Palaung in Burma are Theravada Buddhists, dress like Burmans, and speak fluent Burmese. They are not, however, considered civilized so long as they live in the hills. A contemporary version of this correlation was expressed by a Vietnamese ethnographer, Mac Durong, famous for his pathbreaking and sympathetic studies of minorities, many of whom, he believes, were long ago driven into the hills solely because they arrived on the scene after the Vietnamese had occupied the bottomland. The logic behind his understanding of why such groups were called Man (a term that has come over the centuries to mean “savage”) was clear, as Patricia Pelley noted: “There were legitimate reasons to refer to the highland peoples as savage, and the rationale, while not explicitly stated, was apparent: civilization could be gauged by geography, and more especially, by elevation. The people in the lowlands [ethnic Vietnamese] were fully civilized; those dwelling in the midlands were partially civilized; but highlanders were still savage, and the higher the elevation the greater the degree of savagery.”4 It does not suffice to terrace land and create wet-rice padis to qualify as civilized. The Hani, along the upper reaches of the Red River in northern Vietnam, do just that and are still seen as Man.
This inversion of elevation and civilizational standing works much the same way in Thailand. As a lifelong student of the Akha (linguistically related to the Hani), the late Leo Alting von Geusau observed that the Akha, as a “mid-slope” people, are stigmatized as uncivilized, though not so uncivilized as groups living at the highest elevations. “This situation,” he wrote, “is structured … in an inverse way to the sakdina [lowland Thai ranking] type of hierarchy, with the lowest class the highest up [Mon-Khmer groups such as Wa, Bulang, Khmu, Htin, and Dulong] and the socially highest situated lowest, in the valleys and plains.”5
Linguistic usage in Burmese and Chinese reflects the way in which lowland centers of civilization are elevated symbolically. Thus to go to the capital city or to school is generally to “go up” or “climb” or “rise” (téq—
). Even if one lives on a mountaintop, one still goes “up” to Mandalay. Similarly, when one goes to rural villages or to the hills, one goes “down” or “descends” (s ’in— 
), even if the place in question is thousands of feet above the capital in altitude. Here, as in some Western contexts, up and down have nothing to do with altitude and everything to do with cultural elevation.6
If living at high elevations was coded “barbarian” by the padi state, so too was physical mobility and dispersal. Here again there are strong parallels with the history of the Mediterranean world. Christian and Muslim states regarded mountain dwellers and nomadic peoples—precisely those peoples who had thus far eluded the grasp of the state—as pagan and barbarian. Muhammad himself made it abundantly clear that nomads who embraced Islam must, as a condition of their conversion, settle permanently or pledge to do so.7 Islam was the faith of a sedentary elite, and it was assumed that one could not be a satisfactory Muslim without being settled. Bedouins were regarded as “wild men,” the precise opposite of the Meccan, urban ideal. In civilizational terms, nomadism was to the Arab state what elevation was to the padi state.
In Southeast Asia as well, the idea of civilization was in large measure an agro-ecological code. Peoples who appeared to have no fixed abode, who moved constantly and unpredictably, were beyond the pale of civilization. Here the condition of remaining “legible” to the state and producing a surplus that is readily appropriable is embedded in the concept of civilization. A similar stigma has been applied in the West as in Southeast Asia to subjects, even if ethnically and religiously part of the dominant society, who have no permanent residence: variously termed vagrants, homeless, vagabonds, tramps. Aristotle thought famously that man was by nature a citizen of a city (polis); people who chose consciously to not belong to such a community (apolis) were, by definition, of no worth.8 When whole peoples, such as pastoralists, gypsies, swidden cultivators follow, by choice, an itinerant or semiitinerant livelihood, they are seen as a collective threat and are collectively stigmatized.
Vietnamese, however widely they may range to find work and land, think of themselves, nevertheless, as having an ancestral place to which they will, or might, return.9 Those without an ancestral place are stigmatized as “people of the four corners of the world.”10 Hill peoples are, by extension, whole societies of vagrants, at once pitiable, dangerous, and uncivilized. The state-sponsored “Campaign to Sedentarize the Nomads” or the “Campaign for Fixed Cultivation and Fixed Residence,” designed to curtail shifting cultivation and resettle highland peoples away from the frontiers and to “teach” them wet-rice cultivation, had a deep resonance among the Vietnamese population. It seemed to them and their officials that they were engaged in a magnanimous effort to bring backward and uncouth peoples into the fold of Vietnamese civilization.
Burmans, while less concerned than Vietnamese with ancestral tombs per se, have a comparable fear of and contempt for wanderers with no fixed abode. Such people are called lu lè lu lwin (
), literally, “a person blown about by the wind,” which could variously be rendered as vagrant, tramp, or wanderer, with the connotation of one going to waste.11 Many of the hill peoples are seen in the same light as backward, unreliable, and without culture. For the Burmese, as for the Chinese, itinerant peoples were civilizationally suspect by definition. These stereotypes persist and plague hill peoples in Burma today. Thus a Catholic student of Padaung-Karen (both hill peoples) parentage hesitated when fleeing the military repression of the 1988 democratic movement on account of the stigma attached to taking refuge in the forest:
I was afraid, quite simply of being branded as a jungle fugitive by my fellow countrymen. The word “jungle” [táw—
] still carried pejorative overtones in the speech of urban Burmese. Anyone taking refuge with the ethnic insurgents was called a jungle child [táw ka lé—
] which implied primitiveness, anarchy, violence, and disease, as well as the unpleasant proximity of wild animals which the Burmese detested. I had always been painfully sensitive about being regarded as part of a primitive tribe, and much of my ambition in Taunggyi and Mandalay had been to escape into civilization.12
The Qing general Ortai, who described the hill peoples of Yunnan as “barbarian nomads who were the antithesis of civilized ideals,” was not only being redundant; he was expressing an equation to which all padi-state rulers might subscribe.13
For the Chinese, Burmese, and Siamese states, certain modes of subsistence and the agro-ecological niche in which they were practiced were irretrievably barbaric. Hunting and gathering as well as shifting cultivation were necessarily practiced in the forests.14 This in itself was outside the pale. A seventeenth-century Chinese text describes the Lahu of Yunnan as “people of the mountains, forests and streams.”15 It claims that they eat everything raw and do not bury their dead, comparing them to apes and monkeys. Far from entertaining the possibility that, as Anthony Walker believes, they became hill swiddeners only when they fled the valleys, they were assumed to be true aboriginals. The proof of their primitive, ur- condition was precisely the list of their customs and practices—dwellings, clothes (or their absence), footwear (or its absence), diet, burial practices, and demeanor—that contravened every ideal of Confucian civilization.
Reading some of the reports of Han officials on the multitudinous and confusing hill peoples of the southwest frontier, one is left with two impressions. The first is that of an ethnographic “field guide to the birds”—the Lahu wear such-and-such colors, can be found in such-and-such a habitat, and subsist in such-and-such a manner—allowing administrators to recognize them as they “fly by,” so to speak. The second impression is that they are all being placed in an evolutionary and civilizational sequence in which the ideals of Han civilization are the metric. Viewed this way, the hill tribes are ranged from very “raw” (primitive) to very cooked. Thus we get a series of the following kind: “almost Han,” “on-their-way-to-being-Han,” “could-eventually-become-Han-if-they-wanted-to (and if-we-wanted-them-to!),” and, finally, a category (for the “wildest” of the Lahu, for example) “uncivilizable,” which meant, of course, “not-really-human.”
It is a rare term for people at the periphery of state power—swiddeners, hill people, forest dwellers, or even peasants in the “deep” countryside—that does not carry stigmatizing connotations. For the Burmans, the term for villagers far from cultural centers is táw thà (
), literally, forest dweller, with the connotation of rustic, wild, and uncouth (yaín— 
).16
The indelible association of the valley state with fixed-field grain agriculture, and hence with a quasi-permanent social order of aristocrats and commoners that represented “civilization,” had ironic consequences. Those who chose to leave the realm of inequalities and taxes for the hills placed themselves, by definition, beyond the pale. Altitude could then be coded “primitive.”17 In addition, to the degree that irrigated padi cultivation massively alters the landscape, while hill agriculture appears less visually obtrusive, hill peoples came to be associated with nature as against culture. This fact enables the following false but common comparison: the civilized change the world; the barbarians live in the world without changing it.
For Thai and Burman states, the profession of Theravada Buddhism was also a necessary, but not a sufficient, condition for the inclusion of hill peoples in the charmed circle of civilization. The importance of a major salvation religion would, as with Islam in the Malay world, seem to mark off such societies sharply from Han civilization, which had no such religious test.18 The markers for levels of civilization, even among the party ethnographers classifying the “tribes” of Guizhou and Yunnan in the 1950s, were essentially Han technologies and customs. Did they plant in irrigated fields? Did they plow and use agricultural tools? Did they live in fixed settlements? Could they speak and write Chinese? Before 1948 they would have received extra marks for erecting temples to Han-style deities—most particularly, to the God of agriculture.19 Even today the popular Han characterizations of “minorities” codes in an identical way for “civilization.”20
Despite many superficial differences, the religious test for civilized behavior in Thai and Burman culture was also closely linked to the technologies and customs associated with wet-rice cultivation. In strictly religious terms, adherence to Theravada Buddhism did not require (though it might provoke) momentous ritual changes; pre-Buddhist animist practices (nat worship and propitiation in Burma, phi worship and propitiation in Siam) were readily accommodated, even doctrinally, within a syncretic Buddhism. Buddhism was, however, closely associated with a shift in religious and ethnic identity. As Richard O’Connor observes in the Tai context, “Mainlanders link religion to agriculture, agriculture to ritual, and ritual to ethnic identity. When hill farmers like Karen, Lawa, or Kachin take up valley wet-rice, they find its proper cultivation requires Tai rituals. In effect, agricultural choices are bundled into ethnic wholes that function as competing agro-cultural complexes. A pragmatic shift between complexes thus begins ritual adjustments that may end as an ethnic shift.”21 O’Connor writes “ethnic shift,” but he might as well have written “religious shift,” inasmuch as the two are inseparable in this case. Thus we arrive at something of a civilizational paradox in this, as in the Chinese case. Conversion to Buddhism per se, when combined with attributes of “hilliness”—for example, shifting cultivation, residential mobility—is, as we saw in the case of the Palaung, not convincingly civilizing, although it is a step in the right direction. That step, however, not only makes religious conversion that much more likely but is also historically associated with becoming Tai or Burman—that is, a subject of the padi state. Thus becoming fully civilized, in the valley view, is nearly indistinguishable from becoming Han, Thai, or Burman and, in turn, by definition, being incorporated as a state subject.22 Remaining outside the state is, as we shall see, coded “uncivilized.”
Valley states, large and small, though they looked contemptuously down on their uphill neighbors, were bound to them by powerful ties of economic dependence. Their indissoluble mutuality was underwritten by the complementarity of the agro-ecological niches each occupied. Economic partners and frequently political allies as well, valley and hill peoples, state cores and hinterlands, provided essential goods and services to each other. Together they represented a robust and mutually beneficial system of exchange. If anything, the valley centers were even more dependent on products and especially manpower from the hills than vice versa. But each was economically impoverished without its natural trading partner.
This pattern of economic mutuality has been most elaborately described in the Malay world, where it typically takes the form of exchange between upstream (hulu) and downstream (hilir) zones of a watershed. Huluhilir systems of this kind are based on the products each zone, owing to its agro-economic location, can supply the other. Many are of great antiquity. The lowland center in the Malay case is, as we have seen, typically located near the mouth of a river or at the confluence of two rivers. Its position, like that of a settlement dominating an important mountain pass on a trade route, is something of a natural monopoly, allowing it to dominate the trade along the entire watershed from this choke point. The lowland center functions as an entrepôt, exchanging lowland and overseas products for the upriver and forest products coming down the watershed.23
The lowland center, despite its positional advantage, did not hold the whip hand in dictating the terms of exchange. Highly mobile communities, particularly at the upper reaches of the watershed, were frequently close enough to an alternative watershed so that they could, if they chose to, shift their trade to a different entrepôt on an adjacent river system. Failing that, upstream groups were seldom so dependent on trade goods from the lowland polities that they couldn’t substantially withdraw from downstream markets if they found the terms of trade too onerous politically or economically. Nor could the rulers of the entrepôt polities impose themselves militarily on a recalcitrant hinterland. The dispersal and mobility of the upstream population made them virtually immune from punitive expeditions, let alone systematic coercion. Port polities were, as a result, in competition with one another to acquire hinterland allies and the profits that trade with them made possible. Lacking the means to simply impose themselves, they were impelled to solicit loyalty by redistributing many of the gains of trade in the form of prestige goods, jewelry, and lavish gifts that upstream leaders could, in turn, redistribute to their followers to further encourage loyalty and trade.
For mainland Southeast Asian states, particularly smaller states in or near the hills, the same symbiosis between hill and valley prevailed, though it might not be so neatly mapped onto a single watershed. It is no exaggeration to say that the prosperity of such states was largely dependent on its capacity to attract to its markets the products of the surrounding hill peoples, who often outnumbered the population at the state core. Any reasonably comprehensive account of the commodities brought down (for sale, barter, debt repayment, and tribute) by hill peoples would require many pages. Here I can only suggest something of their extraordinary variety, keeping in mind that the composition of the trade shifted over time, occasionally dramatically, with changes in overland (to China) and overseas trade routes and the demand for particular commodities.
Hill people had, from at least the ninth century, been scouring the hills for commodities they knew could be traded advantageously at valley markets and at the coast. Many such products were part of an extensive international luxury trade. Among the naturally occurring forest products that could be gathered were rare and/or aromatic wood (for example, gaharu, sandalwood, sappan, and camphor woods); medicinals (rhinoceros horn, bezoar stones, dried organs of forest fauna, aloe wood); various resins (tung oil); and latexes (guta percha) from forest trees, as well as rare hornbill feathers, edible birds’ nests, honey, beeswax, tea, tobacco, opium, and pepper. All these products had high values per unit weight and volume. That meant that they repaid the effort, even if they had to be carried on foot along mountainous paths to market. During the extraordinary long pepper boom from 1450 to 1650, when pepper exceeded all other commodities traded internationally in value save gold and silver, bringing a head-load of peppercorns to a coastal market could make a young man’s fortune. Precious metals and gems (and, in the twentieth century, opium) provided an even more striking case of high value joined to portability. In light of the physical mobility of highland peoples, such goods could easily be carried to another market in another polity if the potential sellers were dissatisfied.
Other hill products were bulkier and less valuable. They could not be taken long distances to other markets except where water transportation was easily available. Such products included rattan, bamboo, timber, logs (all of which float), cattle, hides, cotton, and hill fruits, as well as such staples as hill (unirrigated) rice, buckwheat, maize, potatoes, and sweet potatoes (these last three from the New World). Many of these products can be left to grow or stored for long periods, allowing their sellers to withhold or sell them, depending on what price they would fetch.
Even quite large kingdoms in precolonial Southeast Asia were strikingly dependent for their prosperity on export goods from the hills. The first Thai trade mission to Beijing of Rama I (Chulalongkorn), in 1784, calculated to dazzle the Chinese, included luxury products that were almost entirely provided by the hill-dwelling Karen: elephants, eaglewood, ebony, rhinoceros horn, elephant tusks, bastard cardamom, long peppers, amber, sandalwood, peacock feathers, kingfisher feathers, rubies, sapphires, cutch, gamboges (a gum resin), sappanwood, dammar, krabao seeds, and a variety of spices.24 Precolonial exports from Cambodia were similarly hostage to the Jarai hill people. Most of what these lowland states sold abroad “consisted of forest products from the highlands, as can be gleaned from Vietnamese and Cambodia annals and documents as well as travel accounts by Chinese and European authors.”25 The smaller Shan states were dependent on the hill peoples surrounding them both for the wealth of hill products necessary for valley life and for important export goods. One cannot see the cornucopia brought down to the five-day rotating markets in the Shan states today without appreciating how much the Shan diet, its building materials, its livestock, and its trade with the wider world—its prosperity in general—depends on abundant trade with its hinterland. With respect to the hill Kayah and the Shan, F. K. Lehman goes so far as to suggest that the main purpose of a Shan ruler is to manage this trade and profit from it.26 Both the Shan and the Kayah had much to gain by exploiting the comparative advantage afforded by their respective ecological niches, but it seems clear that such states were at least as dependent on products from the hills as hill peoples were dependent on valley products.
Valley markets supplied hill populations with desired products unavailable in the hills. Foremost among these products were salt, dried fish, and ironware. Ceramics, pottery, and porcelain, manufactured cloth, thread and needles, wire, steel implements and weapons, blankets, matches, and kerosene were among the most important commodities eagerly sought by hill traders.27 Under what hill peoples saw as favorable terms of exchange, a brisk trade knitted hill and valley economies together, facilitated by a host of intermediaries—traders, peddlers, brokers, creditors, speculators—not to mention various forms of tribute. Under unfavorable circumstances, however, the valley polities had no way of compelling the delivery of hill products. The valley polities, especially the smaller ones, being more fixed geographically and heavily reliant on hill trade, were more fundamentally threatened by the defection of their hill trading partners.
A mere list of commodities, however, misses the decisive hill product on which the valley centers absolutely depended: its population. The nucleus of irrigated padi fields and concentrated manpower of Tai and Burman court centers was, on a long view, forged by the assimilation, with varying proportions of compulsion and choice, of hill peoples. What the valley polities needed most from the hills were people. Those it could not attract by the advantages of trade and cultural opportunities, it tried to seize, as we have seen, through slaving expeditions and wars. Thus, of all the commodities that the hill societies could deny the valleys, their trump card was manpower. It was the flight of hard-pressed valley subjects from the state core and the migration of hill peoples beyond the range of easy capture that was the Achilles’ heel of valley states.
Under favorable circumstances, the symbiosis of hill and valley peoples was so durable and mutually recognized that the two “peoples” could be thought of as an inseparable pair. The economic interdependence was often reflected in political alliances. This pattern was strongly evident in the Malay world, in which most trading ports, large and small, were associated with “hilly” or seafaring, nonstate peoples who provided most of the trade goods on which the Malay state relied. Although these people were not normally considered “Malays”—they did not profess Islam or become direct subjects of the Malay Raja—it is clear that much of the population of Malays had derived historically from these groups. By the same token, commercial collecting from the hinterland and from the sea for such trading centers was also fostered by the opportunities it presented. That is, much of the population in the hinterland had moved there or stayed there by choice either because of the economic advantages it offered in specialized collecting or because of the political independence it afforded—or both. Abundant evidence suggests human movement back and forth across these categories and indicates commercial gathering is a “secondary adaptation” (rather than some primitive condition). We would do better, conceptually, to consider the upstream population as the “hilly” component of a composite economic and social system.28 And yet from the valley perspective, such people were considered essentially different, less civilized, living outside the religious pale.
Similar allied pairs were, and are, common in mainland Southeast Asia. Thus the Pwo Karen in lower Burma were allied with the Mon padi states. Living interspersed with the Mon, but generally in more forested upstream areas, they represented, as a pair with the Mon, a successful circuit of economic exchange. The Mons, judging from the chronicles, appear to have thought of them less as a sharply demarcated ethnic group than as a continuous gradient of customs and practices from pure padi planters at one pole to pure swiddeners and foragers at the other.29 Virtually all the Tai/Shan kingdoms exhibit an analogous symbiosis between a padi core and an adjacent hill people with whom they trade, from whom they draw population, and with whom they are frequently allied. Such alliances, when solemnized by documents (invariably lowland documents), appear as tributary relations in which the hill ally is seen as the junior partner. In practice the hill peoples often held the upper hand, extracting, in effect, tribute or “protection payments” from the valley courts. Where the lowland court was dominant, as in the case of the Vietnamese and the Jarai, the hill peoples were no less essential to the court’s prosperity, and their ritual role in appeasing the capricious spirits of the natural world was acknowledged.30
The reliance of the smaller valley states on hill trade and forest collecting was so pronounced that it acted occasionally to restrain efforts to assimilate hill peoples to lowland culture. The fear was that if, indeed, hill people took on valley religion, dress, and settlement patterns and began to cultivate wet rice, they would perforce leave off playing the valuable but stigmatized role of supplier of hill products. Cultural difference, along with the economic specialization that it fostered, was the basis of comparative advantage. Though the lowland states might poach slaves from the hills, they had every incentive to ensure that the hill-trading niche on which they depended was always occupied.31
If semiotics has taught us anything, it is that linguistic terms are inherently relational. They can be “thought”—let alone understood—only in relation to their implicit exclusions and contrasts.32 So it is with the terms civilized and barbarian.
The social production of “barbarians” in classical China, as Owen Lattimore has explained, was integrally tied to the rise of specialized, irrigated rice cores in the valleys and the state structures associated with them. Irrigation was “spectacularly rewarding” in the loess cores of ancient China, and this agro-political complex which concentrated production and population, and hence military might, spread farther and farther wherever the terrain was suitable. In the course of its expansion, this complex absorbed some neighboring populations and extruded others, which moved to higher ground, forests, marshlands, or jungles and maintained their less specialized, extensive, dispersed forms of subsistence. In short, the rise of irrigated-rice state cores created by definition a new demographic, ecological, and political frontier. As the padi state increasingly coded itself as “Han-Chinese”—as a unique culture, a civilization—it coded those who were not incorporated, or who refused to be incorporated, as “barbarians.” Those barbarians still living within what the Chinese state saw as its frontiers were termed “inner” barbarians, and those who “detached themselves from the old matrix to become one of the components of the pastoral nomad society of the steppes” became the “outer” barbarians. By roughly the sixth century CE, “the Chinese were in the plains and in the major valleys, the barbarians in the hilly country with smaller valleys.” In southwest China, in what we have called Zomia, a similar process was at work, where, to repeat Lattimore’s formula, “the influence of ancient high civilizations of China and India reach[ed] far out over the lower levels where concentrated agriculture and big cities are to be found, but not up into the higher altitudes.”33
What Lattimore calls the Chinese matrix of concentrated agriculture and state-making created, as a condition of its existence, an ecological and demographic frontier. In time this frontier became both a civilizational and an ethnic border where before there had been no sharp demarcation. The early Chinese state had ample strategic reasons to mark this new boundary with a sharply etched civilizational discourse and, in some cases, with physical barriers such as the Great Wall(s) and the Miao walls of the southwest. It is easy to forget that until roughly 1700, and later in frontier areas, the Chinese state itself faced the classical problem of Southeast Asian statecraft: sequestering a population in state space. Thus the walls and the rhetoric were calculated as much to keep a tax-shy Chinese peasantry from “going over to the barbarians” as to keep the barbarians at bay.34
The process by which state formation in the valleys generates a civilizational frontier which is typically, then, ethnically coded, is not confined to the Han state. Siamese, Javanese, Vietnamese, Burman, and Malay valley polities exhibit the same forms, although the cultural content is different. Writing of the Mien (Yao) of northern Thailand, Jonsson suggests that the social construction of “hill peoples” as a category is based on the hold that states established over valley agriculture and its peoples. Referring to the Indic polities of Siam, particularly Haripunyai (northern Thailand, seventh to tenth centuries CE), he notes that its cosmologically universal claims generated a barbarian periphery: “The making of polities involves the take-over of the lowland areas for intensive agriculture which, hierarchized with a court, regional towns, and farming villages constitute a universal domain. The universal domain is imagined in part by what lies beyond it: the forested wilderness, and the people who live in the latter domain are imagined by those in the former as living like animals.”35 In quite the same fashion, as cleared padi land became the basis for the elaboration of Javanese states and their cultures, so did uncleared forestland and its people become associated with an uncivilized, barbarian frontier.36 The orang asli (usually translated as “aboriginal”) population of Malaya came into being only as an antonym of “Melayuness.” The new element, Geoffrey Benjamin and Cynthia Chou note, was Islam, and Islam created “tribals”: “Previously there would have been no legal reason to define ‘Malay’ at all, and many of the non-Muslim populations of the time were as ‘Malay’ as the Muslims.… The post 1874 notions of Malayness, however, had the effect of converting these populations, virtually overnight, into the ‘aborigines’ they are considered to be today.”37
All the classical states of Southeast Asia conjured up a barbarian hinterland just out of reach in the hills, forests, and swamps. The friction between the need—both semiotically and economically—for a barbarian frontier and, on the other hand, the impulse of universalizing cosmologies to absorb and transform that frontier is the subject to which we now turn.
The earliest court centers in Cambodia and Java, and later in Burma and Siam, were, ritually and cosmologically speaking, luxury imports from the Indian subcontinent. Using the ritual technology afforded them by Indian merchants and the court Brahmins who came in their wake, small lowland courts ratcheted up their ritual status vis-à-vis potential rivals. In a process Oliver Wolters has called “self-Hinduization,” local rulers introduced Brahminical protocol and ritual. Sanskritized personal and place names were substituted for the vernacular. Monarchs were consecrated by magical Brahminical rites and given mythical genealogies tracing a divine origin. Indian iconography and epics were introduced, along with the complex ceremonies of South Indian court life.38 It appears that this Sanskritization did not penetrate very deeply into lowland cultures beyond the immediate precincts of the court. According to Georges Coedès, it was a “veneer,” “an aristocratic religion which was not designed for the masses.”39 Wolters, in much the same vein, calls the Sanskritic flourishes of early royal texts—and the Chinese flourishes in the Vietnamese texts—“decorative effects” intended to add an air of solemnity and erudition to otherwise vernacular practices.40 Another interpretation, favored by M. C. Ricklefs, is that ideas about the indivisibility of the realm in fact constituted a kind of ideological makeweight—what I have earlier called cosmological bluster—against the reality that power was inevitably fragmented.41
Though such mimicry may have done little to improve the day-to-day power of the lowland courts, it did have an important bearing on the texture of hill-valley relations. First, it connected the lowland courts and their monarch to a universalizing, ecumenical, and charismatic center. Much as the Romans used Greek, the early French court used Latin, the Russian aristocracy and court used French, and the Vietnamese court used Chinese script and Confucianism, so the use of Sanskritic forms staked a claim to participation in a transethnic, transregional, and indeed, transhistorical civilization.42 Even when vernacular scripts made an appearance shortly after the first millennium CE, the Sanskritic flourishes remained, and translations of the cosmopolitan classics of the Sanskritic and Pali world preceded even the translation of the Buddhist canon. Unlike the court cultures (as in southern India) that are largely elaborations and refinements of rituals and beliefs that already exist in the local vernacular tradition, the Indic courts of Southeast Asia were self-consciously modeled on an external, universalizing center.
The lowland elite, having thus vastly elevated itself via the ritual helium from south India, left its earthbound commoners and hinterland far below. As Wolters puts it, they “defined the hinterland’s lowly status in the world order from the perspective of those who saw themselves at the center of a civilized ‘Hindu’ society.”43
Sanskritization thus engendered the invention of barbarians by those who had, not long before, been, well … “barbarians” themselves. Khmer culture, originally tied to the forested uplands, now propagated, once Indic court centers were formed, “a polarity of wild and tamed, of dark, haunted bushland versus inhabited open spaces [that] runs like a leitmotiv through Khmer cultural consciousness.”44 The cultural distance between a refined, settled court center on the one hand and a rude, uncultured zone of forests and hills outside its reach was maximized, and civilization became, as David Chandler aptly puts it, “the art of remaining outside the forest.”45
Much the same process of symbolic hyperventilation and validation of hierarchy through external referents can be observed in the smaller realms and in the hills. Just as by 1300 every coastal plain had its miniature kingdom based on Indian conceptions of royalty, so too was the formula followed meticulously by petty chiefs who had even the slightest pretensions.46 One could argue that such expansive ritual trappings were even more necessary in the hills than in the valleys. Largely populated by dispersed, mobile, swiddening peoples who, owing to a common property frontier, had little in the way of inherited inequalities, the hills also had little in the way of indigenous traditions that would legitimate any supravillage authority. Confederations of villages did exist for trade and warfare, but they were limited associations of nominal equals rather than permanent claims to authority. If models of supravillage authority were to be deployed at all, they would have to be borrowed from the lowland Indic courts, or, alternatively, from the Han-Chinese imperial order to the north. Claims to charismatic, personal authority were indigenous to the hills, but the universalizing, Indic, state-making formula represented an attempt to make it a permanent institution and to turn a leader with followers into a Ruler with Subjects.
The idea of the Indic or Chinese state has long had great currency in the hills. It floats up, as it were, in strange fragments from the lowlands in the form of regalia, mythical charters, kingly dress, titles, ceremony, genealogical claims, and sacred architecture. Its attractiveness derives, it seems, from at least two sources. The first and most obvious is that it provides virtually the only grounds—the only cultural format—for a successful and ambitious hill chief to transform his sway from primus inter pares into something like a petty state with a monarchy, aristocrats, and commoners. Such a move, as E. R. Leach has persuasively shown, was likely to be opposed by flight or rebellion on the part of those fearing permanent subordination. Sometimes, however, an upland chief, even a nominal one, might serve a useful purpose as an intermediary for negotiating compacts with lowland powers, organizing tribute and trade, or protecting against lowland raids for slaves. Successful diplomacy of this kind could prove decisive in the competition between hill groups as well.47
From a valley perspective—whether precolonial, colonial, or postcolonial—structures of stable authority in the hills were greatly to be preferred. They provided a fulcrum for indirect rule, a negotiating partner, and someone who might be held responsible (or held hostage) if there was trouble. For this reason, valley authorities, including colonizers, have had something of a “hill-chief fetish.” They have seen such chiefs where they did not exist, have exaggerated their power when they did exist, and have striven to create both tribes and chiefs, in their own image, as units of territorial rule. The state’s desire for chiefs and the ambitions of upland local strongmen coincided often enough to create imitative state-making in the hills, though such achievement was seldom durable. Local chiefs had ample reason to seek the seals, regalia, and titles conferred by a more powerful realm; they might overawe rivals and confer lucrative trade and tribute monopolies. Recognition of a lowland realm’s imperial charisma was, at the same time, entirely compatible with remaining outside its administrative reach and with a disdain for the subject populations of these lowland realms.
The charisma of state effects in the hills is striking. J. George Scott, on military campaign in the Shan states in the 1890s, encountered a number of Wa “chiefs” who had come with tribute. They urged Scott, now presumptively their ally, to join them in sacking some nearby Shan villages. Failing that, they “clamored for some visible token that they were British subjects.… I gave each of them a strip of paper with the name of the district written across it and my signature on a half-anna stamp.… They were impressed and went off to cut bamboos to store the papers.… They told me that Monglem had steadily taken territory from them for the last ten or twelve years.”48 Scott was seeking submission and tribute in the unruly hills; these Wa “chiefs” were seeking an ally, in the service of their own political ends. Much the same theater of tribute and alliance is reported by Leach in the Shan hills when the area was still, in 1836, nominally under Burman administration. A Burman official was received; a ritual meal was prepared; the solidarity of ten Kachin and Shan chiefs attending was dramatized; and the rule of the kingdom of Ava was acknowledged. But Leach notes that several of the chiefs attending were at war with one another. He cautions us to read the ceremony as a state effect:
All my example really shows is that the Burmese, the Shans, and the Kachins of the Hukawng Valley … shared a common language of ritual expression; they all knew how to make themselves understood in this common “language.” It doesn’t mean that what was said in this “language” was “true” in political reality. The statements of the ritual in question were made in terms of the supposition that there existed an ideal, stable, Shan state with the saohpa [ruler] of Mogaing at the head of it and with all the Kachin and Shan chiefs of the Hukawng Valley his loyal liege servants. We have no real evidence that any real saohpa of Mogaing ever wielded such authority, and we know for a fact that when this particular ritual took place there had been no genuine saohpa of Mogaing at all for nearly 80 years. At the back of the ritual there stood not the political structure of a real state, but the “as if” structure of an ideal state.49
The “as if” structure of an ideal state was incorporated into the architecture of actual and “would-be” states in the hills. The Shan statelets, with their own, albeit modest, wet-rice core population and professing the same Theravada Buddhism as their neighboring Siamese and Burman states, copied their architecture as well. Visiting a Shan palace [haw in Shan] in Pindaya, Maurice Collis noted that it was a replica, in miniature, of the Burmese royal capital: “a wooden house of two stories, with a pillared hall on the first floor, over which was a turret or pya-that, five little roofs piled on top of one another and ending in a gilded finial.” Collis observed, “It was the style of the palace of Mandalay in little.”50 Mimicry of the same kind characterized monastery architecture, funeral processions, and regalia. The more negligible the kingdom, the ruder and smaller the imitation, right down to very minor Kachin chiefs (duwa) with pretensions to Shan-style power, whose haws were as Lilliputian as their actual power. In this connection, Leach claims that the Kachin see the Shan not so much as a different ethnic group as the bearers of a hierarchical state tradition—one that they might, under the right circumstances, emulate.51 It is from the Shan that the Kachin borrow their state effects.
The Kayah, a Karennic people in the Shan hills, have, in the course of asserting their autonomy, copied their political system from what they conceived to be the Shan and Burman models. In this case, since the Kayah were not, by and large, Buddhists, the Theravada elements of the mimicry were omitted. All Kayah leaders, Lehman notes, whether usurpers, rebels, ordinary villagers, or millennial prophets, adhere to the state forms derived from Shan lowland courts: titles, paraphernalia, the concoction of royal genealogies, and architecture.52 One way or another, such leaders always claimed their authority by virtue of its connection to the “as if,” ideally unified, Burmese state. This symbolic subordination, while it might be compatible with actual rebellion, is a sign that such symbolism is the idiom, the singular language, of stateness—of any claim to power in the supravillage sphere. As often as not, given the very limited power of most Kayah chiefs, this language is radically at variance with the practical realities of power.
There are, however, two very different models of state authority available to hill peoples: the Indic courts to the south and the Han-Chinese court to the north. Thus a good many Kachin chiefs with aspirations model their “palaces,” their ceremonies and dress, and their cosmology after the Shan pattern, which the Shan, in turn, modeled on the Chinese. In the Kachin chiefly tradition, both the symbolic spaces and the ceremonies devoted to “heavenly spirits” and to “earth spirits” bear a striking resemblance to older imperial ceremonies in Beijing.53 The Akha, a nonstate people if there ever was one, are hardly influenced at all by Tai/Shan models of authority, but instead take their bearings from Taoist, Confucian, and Tibetan models of genealogy, authority, and cosmology—with the Buddhist elements more or less discarded.54 Where the two state traditions were both available, they made possible quite exotic hybrids of state mimicry. In each case, however, they put the conceptual and symbolic language of a divine, universal monarch into the mouth and ritual conduct of rulers whose actual sway might not have extended beyond the borders of their own hamlets.
All court cultures on the periphery of Zomia developed more or less sharp distinctions between what they considered “civilized” people and “barbarians”—variously termed raw, hill people, forest people, wild people, people of streams and grottoes. The terms civilized and barbarian are, as we have seen, inseparable, mutually defining, traveling companions. Like dark and light, each can scarcely be said to have an existence at all without its contrasting twin. One can usually be inferred from the other. Thus in Han dynasty times, when the Xiongnu people were described as “having no written language, family names, and respect for the elderly,” nor cities, permanent dwellings, or fixed agriculture, the list of what the Xiongnu lack is nothing but a summary of what the civilized Han have.55 Of course, as with most binaries, those seeking to apply them in practice encounter many cases that do not admit of easy classification. Such ambiguities did not threaten the hold of binary thought in matters of civilization any more than they have in matters of race.
The standard civilizational narrative for Siamese, Burmese, Khmer, Malay, and, especially, Chinese and Vietnamese court cultures was that, over time, the barbarians were gradually assimilating to the luminous, magnetic center. Incorporation would never be total, for then the very concept of a civilizing center would cease to have any real meaning. There would always be a barbarian frontier.
Civilizing the people of the hinterland was conceptually more plausible if the barbarians were considered to be essentially “like us,” only more backward and undeveloped. In the case of the Vietnamese, the Muong and Tay were literally considered to be “our living ancestors.” As Keith Taylor and Patricia Pelley point out, the Muong “were popularly regarded (and are still regarded today) as the pre-Sinitic version of the Viet.”56 Muong totems, dwellings, agricultural practices, languages, and literature were minutely scoured, not so much in their own right as for the light they could shed on the origins and development of the Viet people.57
Recognition of the barbarians as an earlier, but not irremediably different, people led, in principle, to the assumption that they were capable of eventually becoming fully civilized. This was Confucius’s belief. When asked how he could possibly consider living among barbarians, he replied, “If a gentleman lived among them, what uncouthness could there be.”58 The civilizational discourse in this case is notably singular; it is a matter of ascent to a single cultural apex. Other, different, but equally worthy civilizations are not generally recognized and hence a (civilized) biculturalism is inconceivable.
The early-nineteenth-century Vietnamese emperor Minh Mang exemplified, in his rhetoric if not in his actions, a magnanimous version of this philosophy of a civilizing mission:
This land [of the Jarai and the Rhadé] is a distant and remote place. It is a land in which they tie knots in strings to keep records. It is a land in which people make swidden fields and harvest rice for a living and a land in which the customs are still archaic and simple. However, their heads have hair, their mouths have teeth, and they have been endowed with innate knowledge and ability by nature. Therefore, why should they not do virtuous things. Because of this, my illustrious ancestors brought the civilization of the Chinese to them in order to change their tribal customs.59
Having annexed eastern and central Cambodia, a people themselves heir to classical Khmer civilization, Minh Mang urged his officials to teach them Vietnamese customs and language, show them how to grow more rice and mulberry trees, and to raise livestock and poultry; finally, the officials were to simplify and repress any barbarous customs. “[It is] like bringing the Cambodian people out of the mud into a warm feather bed.”60
Neither in its Chinese nor its Vietnamese guise was this vision of comfort and luxury awaiting those who chose civilization incompatible with merciless repression of those who resisted by force. Before the great midnineteenth-century uprisings in Guizhou, the largest of the military campaigns were those led by Han Yong (1465) and, sixty years later in 1526, by the renowned Ming scholar-general Wang Yangming to put down the great Miao-Yao uprisings. The first victory of the Ming forces in a climactic battle at Great Vine Gorge resulted in at least sixty thousand deaths, of which eight hundred were victims sent to Beijing for public beheading.61 Later, the victorious Wang Yangming helped re-create the (in)famous tusi system of “ruling barbarians with barbarians” but nevertheless held to the view that the barbarians were “like unpolished gems,” capable, if carefully shaped and burnished, of becoming fully civilized.62 His explanation of why direct rule of such a rude people would create havoc is both memorable and diagnostic: “To institute direct civil administration by Han-Chinese magistrates would be like herding deer into the hall of a house and trying to tame them. In the end they merely butt over your sacrificial altars, kick over your tables, and dash about in frantic flight. In the wilderness districts, therefore, one should adapt one’s methods to the character of the wilderness.… [Those doing so] are adapting themselves to the wild nature of these people.”63
The pieties of civilizational discourse propagated by the imperial center are one thing. Reality was something else. These self-idealizations had little to do with life in the imperial capital and even less to do with the rough-and-tumble of the imperial frontier. In place of the Analects was a pandemonium of adventurers, bandits, speculators, armed traders, demobilized soldiers, poor migrants, exiles, corrupt officials, fugitives from the law, and refugees. A 1941 report from the southwest frontier identified three sorts of Han people: displaced and desperate refugees, petty artisans and traders, described as “speculators looking for lucky breaks,” and finally, officials: “The higher ranks … lived indolently, often were overbearing opium-smokers, negligent of government orders.… The lower ranks indulged in petty graft and collected money from fines while illegally trafficking in opium and salt. There was not a lucrative crack into which they did not pry. These activities were bound to lead to enmity between them and the frontier tribesmen who suffered from their oppression.”64 As in any colonial or imperial setting, the experience of the subject was wildly at odds with the ideological superstructure that aimed at ennobling the whole enterprise. The pieties must, in this case, have seemed to most subjects a cruel joke.65
The civilizational project is alive and well in twentieth-century mainland Southeast Asia. Following a Hmong/Miao rebellion in northern Thailand in the late 1960s, General Prapas not only deployed all the counterinsurgency techniques at his disposal—including napalm and aerial bombing—but undertook to “civilize” the rebels with schools, resettlement, clinics, and sedentary agricultural techniques. The cultural campaign, Nicholas Tapp observes, was virtually a carbon copy of the Republican Chinese government’s program in the 1930s in Guangdong, carried out by the “Bureau for Civilizing the Yao.”66 In contemporary China, although the stigmatizing names for minority peoples have been sanitized, the great divide remains between the Han and the many enumerated minorities. The euphemisms of “development,” “progress,” and “education” have replaced “raw” and “cooked,” but the underlying assumption is that minority societies and cultures are “social fossils” whose days are numbered.67
Depending on the culture of the court center, the content of what it meant to be civilized—and reciprocally, what it meant to be stigmatized as barbarian—varied. Each represented, metaphorically, a ladder of ascent, but many of the rungs were unique and particular. In Siam and Burma, Theravada Buddhism was a key marker of civilized status.68 In Vietnam and China literacy and, beyond that, familiarity with the classics was crucial. In the Malay world, upstream populations were, much as Wang Yangming described the Yao, “unfinished Malays.” An essential rung on the way to being “finished” (the Chinese term would be cooked) was the profession of Islam. All these ladders, however, had at least two rungs in common, despite their cultural particularities. They stipulated, as a condition of civilization, sedentary agriculture and residence within state space.
This centripetal narrative of civilization in which nonstate peoples gradually move downhill, adopt wet-rice agriculture, and assimilate linguistically and culturally is not inherently mistaken. It describes a historical process. The Shan people—the sedentary subjects of Shan statelets—are, Leach and O’Connor agree, largely the descendants of hill peoples who have adopted valley ways.69 Malayness was similarly confected by a process of nonstate peoples becoming subjects of the small port polities. It is similarly clear that the first Burmese kingdom at Pagan was itself an amalgam, an ingathering of many peoples.70 This narrative is, then, not so much mistaken as radically incomplete; it records only the events that fit into the imperial self-description of court centers.
If we examine the centripetal narrative of civilization closely, it is striking how much of the actual meaning of “being civilized” boils down to becoming a subject of the padi state. So momentous and consequential is this distinction between being a ruled subject and remaining outside the state that it is typically marked by a shift in identity—often ethnic identity. Moving to a wet-rice core, and hence into a stratified, state-structured hierarchy, meant, depending on the context, becoming Tai, becoming Burman, becoming Malay. On China’s southwest frontier, it meant moving from the “raw” (sheng) barbarian status to “cooked” (shu) civilized status and, eventually, it was assumed, to Han identity itself.
A twelfth-century document from Hainan makes the association between subjecthood and being “cooked”—variously understood as being cultivated, domesticated, or, in the French idiom, évolué— quite clear: “Those who have submitted and are attached to the county and township administration are the cooked Li. Those who live in the mountain caves and are not punished by us or [who do not] supply corvée labor are the raw Li. These sometimes come out and engage in barter with the administered population.” The “cooked” Li occupied a liminal space. They were no longer “raw” and yet were not yet assimilated Han subjects. Officials suspected them of outward conformity while “sly[ly]” cooperating with the “raw” Li to “invade governmental lands and roam about plundering travelers.” Despite the fear of treachery from “cooked barbarians,” they are, as a category, associated with political (state) order while the “raw” are associated with disorder. Thus the “raw Wa rob and plunder,” whereas the “cooked” Wa “safeguard the road.” It would be a mistake, Magnus Fiskesjö emphasizes, to believe that, for a Han administrator, raw was simply another word for primitive or close to nature. While all “primitives” were presumptively raw, not all developed barbarians were cooked. The key was submission to Han administration. Most of the Nuosu people (now subsumed under the Yi designation) on the Yunnan-Sichuan border, who were hierarchically organized in castelike structures and boasted a writing system, were classified as raw because they had eluded political incorporation. That small portion of the same Nuosu people which had come under Chinese rule were designated cooked. In short, “The ‘raw’ barbarians were those located beyond the enforceable jurisdiction of the agents of state.”71 The antiquity of this criterion may extend, if Patricia Ebrey is correct, to the Eastern Zhou period (eighth to third centuries BCE), when the distinction between those who submitted to Zhou rule and those who did not became merged with the ethnic distinction between the Chinese [Hua or Xia] and the barbarians.72
Returning to the eighteenth-century highland Hainan and Li barbarians, those who declared their loyalty and came under Qing administration were said “to enter the map.” By doing so, they became—instantly, politically microwaved, as it were—“cooked,” though their other customs and habits remained as before: “The definition of shu and sheng had mostly political and very little cultural meaning.”73 Implicit in “entering the map,” in being incorporated into the bureaucratic system, was the idea that a people were being readied for a process of acculturation to the civilized norms of Han subjects—a process, it was assumed, they would eagerly embrace.74 The first, essential step in that process, however, was the political-administrative status of being “cooked”—of being “set on the path of becoming registered, taxpaying, and corvée-delivering ‘good’ subjects.… The category of the ‘barbarian’ can have no permanent referent apart from being ‘beyond the law.’ It simply refers to those who at any given time are made to stand for an idea, any of the peoples living on the periphery who meet (or are cast as meeting) the minimal criteria of non-subject status, ethno-linguistic difference, and location at the periphery.”75
It is in the light of administrative control, not culture per se, that one must understand the invention of ethnic categories at the frontier. The category Yao in fifteenth-century Guangdong was an artifact of civil status—of whether the people in question had entered the map or not. Those registered for tax and corvée, thereby also benefiting from settlement rights, became min (civilian, subject), while those who did not became Yao. The “invented” Yao might be culturally indistinguishable from those who had registered, but, over time, the label became “ethnicized” by Han administrative practice.76 Much the same could be said of the label Miao in Qing administrative practice. It came to be a portmanteau term covering dozens of distinct groups speaking often mutually unintelligible tongues. What characterized them all was their refusal to become part of the “fiscal population.” Over time, an expression that initially had no coherent cultural content came to represent an ethnicized identity.77
Barbarism, then, is in Ming and Qing practice a political location vis-àvis stateness—a positionality. Nonbarbarians are fully incorporated into the taxpaying population and have, presumably, adopted Han customs, dress, and language. Barbarians come in two varieties, the cooked and the raw, and these categories are also positional. The cooked are culturally distinct but now registered and governed by Han administrative norms—even if they retain their local chiefs. They have, also presumably, started their march toward cultural incorporation as Han. The raw barbarians, by contrast, are wholly outside the state population, a necessary “other,” and heavily ethnicized.
It follows that those who move beyond the reach of the state thereby cross the conceptual boundary between civilization and barbarism. Likewise, those who leave either the regimented mín or the supervised cooked for the raw periphery enter a zone of definitive ethnicization.
Historically speaking, the process of becoming a barbarian is quite common. At certain historical moments, it has been more common than becoming civilized. It suffices only to leave state space in order to become a barbarian and, usually, an ethnicized “tribal” as well. As early as the ninth century, Chinese officials report that a people called the Shang in southwestern China originally had been Han but had, over time, gradually blended in with the “Western Barbarians.”78 And the people who later became known as the Shan Yue ethnic group and thereby barbarians (sheng) were, it appears, merely ordinary mín who had fled to avoid taxes. Early-fourteenth-century administrative reports treat them as dangerous and disorderly, but without any indication that they are distinct, racially or culturally (never mind aboriginal), from the taxpaying, administered population. But over time, living outside the reach of the state, they became the ethnic Shan Yue.79 All those who had reason to flee state power—to escape taxes, conscription, disease, poverty, or prison, or to trade or raid—were, in a sense, tribalizing themselves. Ethnicity, once again, began where sovereignty and taxes ended. The ethnic zone was feared and stigmatized by officials precisely because it was beyond sovereignty and therefore a magnet for those who, for whatever reason, wanted to elude the state.
Much the same dynamic is at work elsewhere. In the Malay world, Benjamin writes of the “tribalization” or “re-tribalization” of previously nontribal peoples as they move beyond the jurisdiction of the Malay state or, as often happened, the Malay state itself disintegrated, creating an instant hinterland.80 The very terms by which nonstate peoples were stigmatized encodes the absence of effective sovereignty. For example, the Meratus people in Kalimantan, by virtue of their autonomy and mobility, are stigmatized as being “not-yet-arranged/regimented” (belum diator).81 A Spanish official in the Philippines in the mid-seventeenth century describes the Chico River hill population in terms that both stigmatize their statelessness and convey a hint of envy: “They were so free, so completely without God or law, without King or any person to respect, that they gave themselves freely up to their desires and their passions.”82 What passes, in the eyes of valley officials, as deplorable backwardness may, for those so stigmatized, represent a political space of self-governance, mobility, and freedom from taxes.
The civilizational series—mín, cooked barbarian, and raw barbarian—is at the same time a political series of diminishing state incorporation. It resembles in important respects the Arab-Berber civilizational series in which the siba is the zone outside Arab-state control and the makhazem the zone within Arab control. Those who live in the siba are, or become, Berbers. As with the raw and the cooked barbarians, the task of dynastic rule is to expand the circle of dynasty-supporting tribes (guish) and thereby expand the state’s sway. The siba is, Ernest Gellner writes, best translated as “institutionalized dissidence,” and its inhabitants are therefore looked down upon and coded “Berber.” Tribal society, virtually by definition, exists at the edge of nontribal society as its dark reciprocal twin.83 Unlike Southeast Asia, the “tribals” in the Middle East and North Africa share a common religion, though perhaps not its practice, with state populations. It becomes difficult to discern, under those circumstances, what “Berberdom” means except as an Arab designation for those who elude control by the state and incorporation into its hierarchy.84
Barbarians are, then, a state effect; they are inconceivable except as a “position” vis-à-vis the state. There is much to recommend Bennet Bronson’s minimalist definition of a barbarian as “simply a member of a political unit that is in direct contact with a state but that is not itself a state.” Thus understood, barbarians can be, and often have been, quite “civilized” in the sense of literacy, technological skills, and familiarity with nearby “great traditions”—say, of the Romans or the Han-Chinese. Consider, in this light, such nonstate peoples as the Irish or, in insular Southeast Asia, the Minangkabau and Batak. They may also be more powerful militarily than an adjacent state and, on that account, may raid or exact tribute from that state. Consider also, in this context, the Mongols under the Tang, the Moros, the Bedouins, the Scots, the Albanians, the Caucasians, the Pathans, and, for much of their history, the Afghans. The stronger such “barbarian” societies have been, the more likely they are to prey systematically on the nearby state spaces, with their lucrative concentration of wealth, grain, trade goods, and slaves. Bronson attributes the relative weakness of historical state formation in India and Sumatra—despite favorable agro-ecological settings—to the proximity of powerful nonstate predators.85
All empires, as cultural-political enterprises, are necessarily exercises in classification. Thus the Roman Empire, on a cursory reading, exhibits many of the same characteristics as those impinging on Zomia.86 Slavery was as central to Roman statecraft as it was to Burmese, Thai, or early Han statecraft. Merchants accompanied each military campaign with a view to buying captives and reselling them closer to Rome. Many of the interbarbarian wars were fought between competitors striving to control and profit from this human trafficking. Roman culture, from province to province, as distinct from the famed uniformity of Roman citizenship, looked different depending on the various “barbarian” cultures it had absorbed.
Like their Han and mainland Southeast Asian counterparts, the Romans had a barbarian chiefdom fetish. Wherever possible they created territories, promulgated more or less arbitrary ethnic distinctions, and appointed, or recognized, a single chief who was, willy nilly, the local vector of Roman authority and answerable for the good conduct of his “people.” The peoples so codified were likewise ranged along an evolutionary scale of civilization. The Celts closest to Roman power in Gaul, a stateless but culturally distinct group of peoples with fortified towns and agriculture, were comparable to cooked barbarians in the Chinese scheme. Those beyond the Rhine (the various Germanic peoples) were raw barbarians, and the mobile Huns between Rome and the Black Sea were the rawest of the raw. In the Roman province of Britain, the Picts beyond Hadrian’s Wall in the north were the rawest of the raw, or “the last of the free,” depending on one’s perspective.87
Once again, positionality vis-à-vis imperial rule was a crucial marker for a people’s degree of civilization. Administered (cooked) barbarians in Roman ruled provinces lost their ethnic designations as they became, like farmers, liable for taxes and conscription. All those beyond this sphere were invariably ethnicized, given chiefs, and made responsible for tribute (obsequium), as distinct from taxes, especially as they were seen as a non-grain-growing people. The link between direct Roman rule and barbarian status is obvious in those cases when such “provincials” rebelled against Roman rule. They were, in such cases, reethnicized (rebarbarianized!), demonstrating, in the process, that civilizational backsliding was possible and was very much a political category. Depending on the circumstances, Romans might move into barbarian territory as deserters, traders, settlers, and fugitives from the law, and “barbarians” might move into the Roman sphere, though they needed permission to do so collectively. The dividing line, despite the two-way traffic across it, was always sharply marked. Here too, “barbarians” were a state effect. “Only conquest produced real knowledge of the barbarian world, but then it ceased to be barbarian. Thus conceptually, the barbarians were forever retreating from Roman understanding.”88
As a political location—outside the state but adjacent to it—the ethnicized barbarians represent a permanent example of defiance of central authority. Semiotically necessary to the cultural idea of civilization, the barbarians are also well nigh ineradicable, owing to their defensive advantages in terrain, in dispersal, in segmentary social organization, and in their mobile, fugitive subsistence strategies. They remain an example—and thus an option, a temptation—of a form of social organization outside state-based hierarchy and taxes. One imagines that the eighteenth-century Buddhist rebel against the Qing in Yunnan understood the appeal of “barbarian-ness” when he exhorted people with the chant: “Api’s followers need pay no taxes. They plow for themselves and eat their own produce.”89 For officials of the nearby state, the barbarians represent a refuge for criminals and rebels, and an exit for taxshy subjects.
The actual appeal of “barbarity,” of residing out of the state’s reachlet alone forsaking civilization—has no logical place in the official state narratives of the four major civilizations that concern us here: the Han-Chinese, the Vietnamese, the Burman, and the Siamese. All are “predicated on irrevocable assimilation in a single direction.” In the Han case, the very terms raw and cooked imply irreversibility: raw meat can be cooked but it cannot be “uncooked”—though it can spoil! No two-way traffic or backsliding is provided for. Nor does it allow for the indisputable fact that the core civilizations to which assimilation is envisaged are, themselves, a cultural alloy of many diverse sources.90
A civilizational narrative that assumes its own cultural and social magnetism and that depicts acculturation to its norms as a much desired ascent could hardly be expected to chronicle, let alone explain, large-scale defection. And yet it is historically common. The official invisibility of defection is encoded in the narrative itself; those who move to nonstate space, who adapt to its agro-ecology, become ethnicized barbarians who were, presumably, always there. Before the decisive military victory of the Han forces over the Yao in the mid-fifteenth century, it appears that “Han people were to a large extent nominally turning into non-Han rather than the other way around.… Marginal migrants in an area under weaker government control responded to Panhu [ethnic mythology] symbols that, among other things, promised help from the Yao in the vicinity. They were the very reverse of the ‘Barbarians’ who went to court to pay tribute and express admiration for civilization. From the perspective of the state, the rebels betrayed civilization and attached themselves to the barbarians.”91 Barbarians bearing gifts had an honored place in the civilizational discourse. Han subjects going over to the barbarians did not! When they were mentioned at all in Qing literature, they were stigmatized as “Han-traitors” (Hanjian), a term that now had strong ethnic resonances.92
“Self-barbarianization” could occur in any number of ways. Han populations wanting to trade, to evade taxes, flee the law, or seek new land were continually moving into barbarian zones. Once there, they were likely to learn the local dialect, marry locally, and seek the protection of a barbarian chief. Remnants of disbanded rebels (most notably the Taiping in the nineteenth century) and deposed dynasts and their entourage (for example, Ming supporters in the early Qing) contributed to the influx. Occasionally, when a local barbarian kingdom was strong enough, as in the case of Nan Chao, Han populations might be captured or purchased and then absorbed. Nor was it uncommon for a Han military official appointed to rule a barbarian area to make local alliances, take a local wife, and, in time, assert his independence as a native chief. There was, finally, a kind of self-barbarianization that makes the association between coming under Han rule and being civilized crystal clear. When a district of cooked barbarians under Han rule revolted successfully, they were reclassified as raw and moved back into the “barbarian” column. What had changed was not their culture but only their subordination to Han rule.93
William Rowe claimed, perhaps for dramatic effect, that “going over to the barbarians” was more the norm than the exception: “The historical reality for centuries has been … that far more Chinese had acculturated to aboriginal life than aborigines to Chinese civilization.”94 Whatever a complete demographic account book would show, what matters in this particular context is that backsliding was common, even banal, and that it could have no legitimate place in the official narrative. In times of dynastic decline, natural disasters, wars, epidemics, and exceptional tyranny, what was a steady flow of adventurers, traders, criminals, and pioneers might become a population hemorrhage. One imagines that much of the population near the frontiers could see the positional advantage in being culturally amphibious and stepping to one side or the other depending on the circumstances. Even today, on China’s southwest frontiers, there are some substantial advantages to being an ethnic minority—a barbarian. One escapes the “one-child” policy, one avoids certain taxes, and one profits from certain “affirmative action” programs benefiting minorities. Han Chinese and people of mixed ancestry in these areas are known to seek registration as Miao, Dai, Yao, Zhuang, and so on.