CHAPTER 6
State Evasion, State Prevention The Culture and Agriculture of Escape

I magine, once again, that you are a Southeast Asian counterpart to Jean-Baptiste Colbert. This time, however, your task is not to design an ideal state space of appropriation but, rather, the precise opposite. How would you go about designing a topography, a subsistence strategy, and a social structure that was as resistant to state formation and appropriation as possible?

Much of what you would design, I believe, would be an inversion of how the padi state was sculpted. In place of a flat, relatively frictionless alluvial plain, you would conjure up a rugged landscape where the “friction of terrain” was forbiddingly high. In place of concentrated grain crops that ripen simultaneously, you would prefer shifting, diverse, dispersed, root crops of uneven maturation. In place of permanent settlement and fixed political authority, you would devise a scattered, mobile pattern of residence and a fluid, acephalous social structure capable of easy fissioning and recombination.

In broad strokes, this is what one finds throughout much of Zomia, a pattern of settlement, agriculture, and social structure that is “state repelling.” That is to say, it represents an agro-ecological setting singularly unfavorable to manpower- and grain-amassing strategies of states. The pattern is state repelling in two distinct ways. The first and most obvious is that an existing state will hesitate to incorporate such areas, inasmuch as the return, in manpower and grain, is likely to be less than the administrative and military costs of appropriating it. Tributary status might be plausible, but not direct rule. The second state-repelling feature of this social landscape is that it makes the rise of an indigenous state in this space exceedingly unlikely. The critical mass of concentrated manpower, wealth, and grain on which a state must rest is essentially lacking. Furthermore, demography and agronomy unfavorable to state appropriation are, it turns out, proof against other forms of appropriation as well: in particular, raiding. Slave-raiding expeditions, marauding armies, bandits, starving would-be pillagers of foodstuffs will, like states, find “state spaces” more lucrative for raiding than the slim pickings in sparse, mobile, root crop-growing societies with no permanent structure of authority. Such hill societies are, in this sense, not simply state repelling but appropriation resistant in general.

I have used the device of a Colbertian strategist and the idea of “design” quite deliberately. Much of the history and ethnography of the hill peoples in mainland Southeast Asia tends, implicitly or explicitly, to naturalize their location, their settlement pattern, their agriculture, and their social structure, to treat these as givens, dictated, as it were, by traditional and ecological constraints. Without gainsaying the existence of some constraints, I wish to emphasize the element of historical and strategic choice. What is striking, on any long historical view, is the great flux and variety in patterns of hill and valley residence, in social structure, in forms of agriculture, and in ethnic identity. Patterns that may appear static, even timeless, at first sight, display a remarkable plasticity if one steps back and widens the historical lens to a span of a few generations, let alone a few hundred years or a millennium. The evidence, I think, requires that we interpret hill societies—their location, their residence pattern, their agricultural techniques, their kinship practices, and their political organization—largely as social and historical choices designed to position themselves vis-à-vis the valley states and the other hill peoples among whom they live.

An Extreme Case: Karen “Hiding Villages”

A limiting case can often, by its very starkness, illustrate the basic dynamics of a social process. The draconian counterinsurgency strategy of Burma’s military rulers in largely Karen areas is a case in point. Here, the “state space” around the military base is less a mere zone of appropriation than a full-fledged concentration camp. “Nonstate space,” by contrast, is not so much an area outside the effective realm of the taxman as a refuge to which people run for their lives.¹

In the Orwellian euphemism of the Burmese army, the civilian zones that they control in Karen areas are called peace villages, while the zones sheltering those who have escaped beyond their reach are called hiding villages. The official description represents “peace villages” as ones whose headmen have agreed not to assist the insurgents and to provide free labor to the military camp on a rotating basis, in return for which the villagers will not have their houses burned or be forcibly relocated. Peace villages are, in fact, frequently relocated by force to the border of the military camp itself, where they provide a ready pool of laborers and hostages. Their inhabitants are registered and given identity cards. Their agricultural land, betel nut trees, and cardamom bushes are assessed for the purpose of military taxation and requisitions. In a miniature—and militarized—version of the padi-state cores we examined in Chapter 3, the base commanders tend, in fact, to extract most of the labor, cash, and food they require from the peace villages closest to headquarters. The villagers implicitly understand the connection between the concentration of population and forced labor. In one of the many cases documented, seven villages had been forcibly consolidated into two, Kler Lah and Thay Kaw Der, nearby the barracks. As one resident said, “When they can’t find people to be porters, they take all the villagers from Kler Lah and Thay Kaw Der. They don’t mind if they are male or female, they take them.… The SPDC [government] forced them to relocate there in 1998. That is why it is very easy to force them to do forced labor or porter [since they are in one place].”² In a comparable relocation area, a villager also noted how concentration near the military base exposed them to exploitation. “In my opinion, they asked the villagers to move to these places so they could make them work.… If the villagers stay in one place, then it is easy for the Burmese to make them work.”³

Forced to provision themselves from local sources, and with a tradition of corruption and plunder, military units have transformed relocation areas into zones of hyperappropriation. The “ideal type” of military space is a flat, open terrain (no ambushes!) along a major road, surrounded by a registered, relocated civilian population growing crops in easily monitored fields, who serve as trip wires and hostages, as well as a source of labor, cash, and foodstuffs. In an amplified version of the padi state, the Burmese army presses so severely on the manpower and resources of its captive population that a large proportion of them eventually flee in desperation.⁴

Just as the villagers sequestered around a military base represent a virtual parody of state space, so are the state-repelling techniques of those who flee its burdens an exaggeration of the strategies to be examined in this chapter. Briefly put, such strategies include fleeing to inaccessible areas, scattering and breaking up into smaller and smaller groups, and pursuing subsistence techniques that are invisible or unobtrusive.

The quickest available refuge lies, generally, farther up the water courses and higher in the hills. “If we have to run, we will run up into the hills,” reports a Karen village elder. If they are pursued, they retreat still farther upstream to higher altitudes. “Then they came and looked for us so we fled upstream.” And: “The third time they came we fled up here.”⁵ The advantage of such refuges is that they are not very far, as the crow flies, from one’s village and fields but are nevertheless far from any road and virtually inaccessible. As the degree of military pressure increases, such so-called hiding villages (ywa poun— ) split into smaller units. Whereas the small villages from which they come may have fifteen to twenty-five households, hiding villages seldom comprise more than seven households and, if still endangered, split up into small family groups. The greater the degree of disaggregation, the less visible any particular group is, and the less likely to be pursued and captured or killed. In the final analysis, in this case, villagers may hazard the trek to the Thai border and to the refugee camps there—altogether outside the jurisdiction of the Burmese state.

Those who choose to remain in the hills adopt subsistence strategies designed to escape detection and maximize their physical mobility should they be forced to flee again at a moment’s notice. Foraging for forest foods is the ultimate in unobtrusive subsistence; it leaves no trace except for the passage of the forager. But pure foraging is rarely sufficient.⁶ As one villager concealed in the hills explained, “The people in the village have to eat roots and leaves just like I was eating in the forest. I had to live on roots and leaves for four or five days at a time.… For one year I’ve lived in the forest in a hut because I was too afraid to stay in the village. I planted banana trees and ate roots and some vegetables.”⁷ Many who fled to the forest brought as much rice as they could carry, which they hid in small lots. But those who stayed any length of time cleared very small plots to grow maize, cassava, sweet potatoes, and a few cardamom bushes. The pattern was to open many small, scattered, unobtrusive plots; the same principles of dispersal and invisibility governing the behavior of human refugees also governed their agricultural choices. Where possible, they chose crops needing little care, crops that matured quickly, root crops that could not easily be destroyed or confiscated and which could be harvested at leisure. People, fields, and crops were each deployed to evade capture. Villagers were well aware of what they were sacrificing in the interest of bare survival. Village rituals, schooling, sports, trade, and religious observances were all curtailed if not eliminated solely to avoid what amounted to military serfdom in hyperstate space.

The techniques of evasion practiced by desperate Karen villagers represent an extreme instance of strategies that characterize much of the history and social organization of Zomia as a whole. A good deal of what we have come to consider “hill” agriculture, “hill” social structure, and “hill” location itself is, I would argue, largely defined by patterns of state evasion (and prevention). Such strategies have been devised and elaborated over many centuries in constant “dialogue” with lowland padi states, including the colonial regime.⁸ This dialogue is, in important respects, constitutive of both hill societies and their padi-state interlocutors. Each represents an alternative pattern of subsistence, social organization, and power; each “shadows” the other in a complex relationship of mimicry and contradiction. Hill societies operate in the shadow of lowland states. By the same token, the lowland states of Southeast Asia have been surrounded, for the whole of their existence, by relatively free communities in the hills, swamps, and labyrinthine waterways that represent, simultaneously, a threat, a zone of “barbarism,” a temptation, a refuge, and a source of valuable products.

Location, Location, Location, and Mobility

Inaccessibility and dispersal are the enemies of appropriation. And for an army on the march, as for a state, appropriation is the key to survival. “The whole army continued the pursuit of the flying [sic] king, but, as the marches were rather forced and the villages few and far between in a tract scarcely populated, sufficient provisions to feed this army of men and animals could not be obtained, with the result that they were not only fatigued with continual marching but half starved from want of regular meals. Many died of disease, starvation, and exhaustion from want of food, but the pursuit was still persisted in.”⁹

The first principle of evasion is location. Owing to the friction of terrain, there are locations that are virtually inaccessible even to a nearby (as the crow flies) state. One could, in fact, calculate something of a gradient of relative inaccessibility for different locations from any particular padi state. Such a gradient is implicit in Clifford Geertz’s description of the reach of what he terms the “theatre-state” in Bali. He notes that “upland lords,” because they were located in more rugged country “had a natural advantage in resisting military pressure.”¹⁰ Even farther uphill, “at the highest altitudes, a few usually dry-farming communities existed beyond the effective reach of any lords at all.” Within Zomia itself, most of the southwest province of Guizhou was perhaps the most forbidding, inaccessible area in purely geographical terms. A standard saying about Guizhou had it that “no three successive days are clear, no three square feet are level, and no one has more than three cents in his pocket.” One late-nineteenth-century traveler noted that he had not seen even a single cart during his whole time in Guizhou—trade, “such as it is, being conducted on the backs of bipeds and quadrupeds.” Many places, reputed to be accessible only to monkeys, were in fact zones of refuge for bandits and rebels.¹¹ Location, in this context, is but one of many possible forms by which marginality to state power finds expression. As we shall see, physical mobility, subsistence practices, social organization, and settlement patterns can also be deployed, often in combination, to place distance between a community and state appropriation.

On any long historical view, location at the periphery of state power must be treated as a social choice, not a cultural or ecological given. Location, just like subsistence routines and social organization, is variable. Over time such shifts have been observed and documented. Most frequently they represent a “positionality” vis-à-vis forms of state power.

Recent scholarly research has served, for example, to undermine naturalized understandings of such “nonstate” peoples as the so-called orang asli (“original people”) of Malaysia. They were previously understood to be the descendants of earlier waves of migration, less technically developed than the Austronesian populations which succeeded and dominated them on the peninsula. Genetic evidence, however, does not support the theory of separate waves of migrating peoples. The orang asli (for example, Semang, Temuan, Jakun, Orang Laut) on the one hand and the Malays on the other are best viewed not as an evolutionary series but as a political series. Such a view has been most convincingly elaborated by Geoffrey Benjamin.¹² For Benjamin, tribality in this context is simply a term applied to a strategy of state evasion; its polar opposite is peasantry, understood as a system of cultivation incorporated into the state. On his reading, most of the “tribal” orang asli are nothing more and nothing less than that fraction of the peninsular population that has refused the state. Each “tribe”—Semang, Senoi, and Malayic (Temuan, Orang Laut, Jakun)—represents a slightly different state-evading strategy, and anyone adopting one such strategy in effect thereby becomes Semang, Senoi, or whatever. Similarly, such nonstate peoples have always had, even before Islam, the option of becoming Malay. Many have in fact done so, and Malayness bears traces of this absorption. At the same time, all orang asli are, and have always been, linked to lowland markets by exchange and trade as well.

For our purposes what is significant is that a peripheral location with respect to the state is a political strategy. As Benjamin puts it,

First, … tribality has resulted largely from choice and, second, … the presence of state-based civilization (both modern and premodern) has figured largely in that choice.…

All the more reason, then, for us to remember that many tribal populations have been living in geographically remote regions out of choice, as part of a strategy to keep the state off their backs.¹³

The second principle of evasion is mobility: the ability to change location. The inaccessibility of a society is amplified if, in addition to being located at the periphery of power, it can easily shift to a more remote and advantageous site. Just as there is a gradient of remoteness from state centers, so also might we imagine a gradient of mobility from a relatively frictionless ability to shift location to a relative immobility. The classic example of physical mobility is, of course, pastoral nomadism. Moving with their flocks and herds for much of the year, such nomads are constrained by the need for pasture but are unmatched in their ability to move quickly and over large distances. Their mobility is at the same time admirably suited to the raiding of states and of sedentary peoples. And indeed, pastoral nomads aggregated into “tribal” confederations have often posed the most serious military threat to sedentary grain-producing states.¹⁴ For our purposes, however, what is important are the evasive strategies vis-à-vis state power that nomadism makes possible. Thus, for example, Yomut Turkmen, located on the periphery of Persian state power, have used their nomadic mobility both to raid grain-growing communities and to escape the taxes and conscription of the Persian authorities. When large military expeditions were sent against them, they would retreat to the steppe-desert, beyond reach, with their livestock and families. “Thus, mobility provided their ultimate defense against effective control over their political affairs by the Persian government.”¹⁵ In a setting where other forms of subsistence were readily available, they chose to retain their nomadism for its strategic advantages: political autonomy, raiding, and the avoidance of the taxman and the military press-gang.

Highland Southeast Asia has, for ecological reasons, no substantial groups of herding peoples. The nearest equivalent, in terms of ease of movement, are nomadic foragers. Most hill people pursue livelihoods that incorporate a certain amount of foraging and hunting and can, when pressed, rely heavily upon it. But those groups specialized to foraging both are located in areas far from state power and have a mode of subsistence that requires physical mobility—a habit that serves them well when they are threatened. Such people have been typically understood by historians and lowland populations alike as remnants of distinct and, in evolutionary terms, more primitive “tribes.” Contemporary scholarship has overturned this judgment. Far from a response to having been left behind, foraging in the modern era is seen as a largely political choice or adaptation to evade capture by the state. Terry Rambo, writing about the foraging Semang of the Malay Peninsula, clearly states the new consensus: “Thus the Semang appear to be very primitive not because they represent a surviving Paleolithic stratum that has been pushed into an isolated, marginal refuge area, but rather because a nomadic, foraging adaptation is both the most profitable and safest strategy for a defensively weak minority ethnic group living close to military-dominant, and often hostile, agriculturalists.… From the standpoint of security, the adaptation also makes sense because nomads are much harder to catch than settled farmers.”¹⁶

It does not follow, however, that the extreme forms of dispersal are the safest. To the contrary, there is a small minimum group size below which new dangers and disadvantages loom. There is first the need to defend against raiding, especially slave-raiding, which requires a small community. A single isolated swidden field is also far more exposed to pests, birds, and other wild animals than a group of swiddens ripening together. Pooling the risks of illness, accident, death, and food shortages also argues for a minimum group size. Thus the atomization of Karen refugees fleeing the Burmese military is a limiting case, sustainable only for a short period. Even for fugitive peoples, then, long-run self-protection requires groups of at least several families.

Once we view subsistence strategies more as political options from among a range of livelihood alternatives, the mobility that any particular form of subsistence provides must enter the calculation. Foraging, along with nomadic pastoralism, affords the greatest mobility for groups wanting to give the state a wide berth. Shifting cultivation (swiddening) affords less mobility than foraging but much more mobility than fixed-field farming, let alone irrigated rice padis. For the architects of state space, any substantial move from wet rice at the core toward foraging at the remote periphery is a threat to the manpower and foodstuffs underwriting state power.

There is no reason, then, to assume that hill swiddeners and foragers are isolated in the hills by default or by virtue of their backwardness. On the contrary, there is ample reason to assume that they are where they are and do what they do intentionally. This is, in effect, the historic choice made by many former plains-dwellers who fled to the hills when oppressed by ruinous taxation or threatened with servitude by a more powerful people. Their intentions are inscribed in their practice, in the sense that they have not chosen, as have others, to assimilate into lowland societies. One of their intentions, it appears, is to avoid capture, as slaves or subjects, by states and their agents. As early as the ninth century a Chinese official in southwest China observed that it was impossible to resettle “barbarians” around centers of Han power because they were scattered in forests and ravines and “therefore managed to evade capture.”¹⁷ Nor should we overlook the attraction of the autonomy and the relatively egalitarian social relations prevailing in the hills, as important a goal as evading corvée and taxes.

Neither does the desire for autonomy exhaust the positive reasons why hill peoples might prefer their situation to the alternatives. We know from both contemporary and archeological data that foragers, in all but the most severe environments, are more robust, healthier, and freer from illnesses, particularly epidemic zoonotic diseases, than the population of more concentrated sedentary communities. All in all, it seems that the appearance of agriculture initially did more to depress standards of human welfare than to raise them.¹⁸ By extension, shifting agriculture, by virtue of its diversity and dispersal of population, is likely to favor a healthier population so long as sufficient land is available. Hill livelihoods, then, may be preferred for reasons of health and leisure. Mark Elvin’s account of the early Chinese state prohibiting its subjects from foraging and swiddening may reflect this preference, as does the widespread belief of hill peoples that the lowlands are unhealthy. This last belief may rest on more than the fact that malaria-bearing mosquitoes historically have rarely been found above nine hundred meters.

Premodern populations, despite their ignorance of the means and vectors of disease transmission, always understood that their chances of survival were improved by dispersal. In his Journal of the Plague Year, Daniel Defoe recounts that those with the means left London for the countryside at the first sign of the black plague. Oxford and Cambridge universities dispersed their students to sanctuaries in the countryside when the plague struck. For much the same reason, William Henry Scott reports, in Northern Luzon both lowlanders and “submitted” Igorots went to the hills and scattered to escape epidemics. Igorots already in the hills knew that they should disperse and close off the passes to the hills to avoid the contagion.¹⁹ There is, then, every reason to believe that the threat posed by the lowland state was not confined to slavers and tribute-takers but extended to invisible microbes as well. This would represent, by itself, another powerful reason to choose to live beyond the range of the padi state.

Escape Agriculture

Do not cultivate the vineyard; you’ll be bound

Do not cultivate grains; you’ll be ground

Pull the camel, herd the sheep

A day will come, you’ll be crowned.

—Nomad poem

New World Perspectives

Any effort to examine the history of social structure and subsistence routines as part of a deliberate political choice runs smack against a powerful civilizational narrative. That narrative consists of a historical series arranged as an account of economic, social, and cultural progress. With respect to livelihood strategies, the series, from most primitive to most advanced, might be: foraging/hunting-gathering, pastoral nomadism, horticulture/shifting cultivation, sedentary fixed-field agriculture, irrigated plow agriculture, industrial agriculture. With respect to social structure, again from the most primitive to most advanced, the series might read: small bands in the forest or savannah, hamlets, villages, towns, cities, metropolises. These two series are, of course, essentially the same; they chart a growing concentration of agricultural production (yield per unit of land) and a growing concentration of population in larger agglomerations. First elaborated by Giovanni Battista Vico at the beginning of the eighteenth century, the narrative derives its hegemonic status not only from its affinity with social Darwinism but from the fact that it maps nicely on the stories most states and civilizations tell about themselves. The schema assumes movement in a single direction toward concentrated populations and intensive grain production; no backsliding is envisioned; each step is irreversible progress.

As an empirical description of demographic and agricultural trends in the now-industrialized world for the past two centuries (and the past halfcentury in poorer nations), this schema has much to be said for it. Europe’s nonstate (“tribal”) populations had, for all practical purposes, disappeared by the eighteenth century, and the nonstate population of poorer countries is diminishing and beleaguered.

As an empirical description of premodern Europe or of most poor nations until the twentieth century, and as an empirical description of the hilly areas of mainland Southeast Asia (Zomia), however, this narrative is profoundly misleading. What the schema portrays is not simply a self-satisfied normative account of progress but a gradient of successive stages of incorporation into state structures. Its stages of civilization are, at the same time, an index of diminishing autonomy and freedom. Until quite recently, many societies and groups have abandoned fixed cultivation to take up shifting agriculture and foraging. They have, by the same token, altered their kinship systems and social structure and dispersed into smaller and smaller settlements. The actual archeological record in peninsular Southeast Asia reveals a long-term oscillation between foraging and farming depending, it would seem, on the conditions.²⁰ What to Vico would have seemed to be lamentable backsliding and decay was for them a strategic option to circumvent the many inconveniences of state power.

We have come to appreciate only very recently the degree to which many apparently more primitive peoples have deliberately abandoned settled agriculture and political subordination for a more autonomous existence. Many of the orang asli of Malaysia provide, as we have noted, a case in point. It is in the post-Conquest New World, however, that some of the more striking cases have been documented. The French anthropologist Pierre Clastres was the first to argue that many of the hunting-and-gathering “tribes” of South America, far from being left behind, had previously lived in state formations and practiced fixed-field agriculture. They had purposely given it up to evade subordination.²¹ They were, he argued, quite capable of producing a larger economic surplus and a larger-scale political order, but they had chosen not to so as to remain outside state structures. Termed disparagingly by the Spaniards as peoples “without God, law and king” (unlike the Inca, Maya, and Aztecs), they were, Clastres saw, rather peoples who had elected to live in a relatively egalitarian social order with chiefs who had little or no power over them.

The precise reasons why such groups would have taken to foraging in small bands is a matter of some dispute. Several factors, however, played a role. First and foremost was the catastrophic demographic collapse—as great as 90 percent mortality in many areas—due to European-borne diseases. This not only meant that established social structures were devastated but that the land available to the survivors for foraging or shifting agriculture was vastly expanded.²² At the same time, many were fleeing the Spaniards’ infamous reducciones, designed to turn them into indentured laborers, as well as the epidemics that characterized such concentrations of population.

A paradigmatic case is that of the Siriono, of Eastern Bolivia, described initially by Allan Holmberg in his anthropological classic Nomads of the Longbow. Apparently lacking the ability to make fire or cloth, living in rude shelters, innumerate, having no domestic animals or developed cosmology, they were, Holmberg wrote, Paleolithic survivors living in a veritable state of nature.²³ We now know beyond all reasonable doubt that the Siriono had been crop-growing villagers until roughly 1920, when influenza and smallpox swept through their villages, killing many of them. Attacked by numerically superior peoples and fleeing potential slavery, the Siriono apparently abandoned their crops, which, in any event, they did not have the numbers to defend. Their independence and survival in this case required then to divide into smaller bands, foraging and moving whenever threatened. They would occasionally raid a settlement to take axes, hatchets, and machetes, but at the same time they dreaded the illnesses that the raiders often brought back with them. They had become nonsedentary by choice—to avoid both disease and capture.²⁴

Clastres examines many such instances of previously sedentary peoples who, threatened by slavery, forced labor, and epidemics, adopted nomadic subsistence strategies to stay out of harm’s way. The Tupo-Guarani groups in particular were, it is clear, populous agricultural peoples who in the seventeenth century, by the tens of thousands, fled the triple threat of Jesuit reducciones, Portuguese and mestizo slave-raiders intent on sending them to plantations at the coast, and epidemics.²⁵ They appeared to the ahistorical eye, much later, as a backward, technologically simple people—an aboriginal remnant. In reality, they had adapted to a more mobile life as a means of escaping the servitude and disease that civilization had to offer.

There is still another New World case of escape agriculture closer to hand. That is the study of maroon communities—of African slaves who had escaped and established communities outside the easy reach of slavers. These communities ranged in size from Palmares in Brazil, with perhaps twenty thousand inhabitants, and Dutch Guiana (Surinam), with that many or more, to smaller settlements of escapees throughout the Caribbean (Jamaica, Cuba, Mexico, Saint-Domingue), as well as in Florida and on the Virginia-North Carolina border in the Great Dismal Swamp. I shall elaborate a theory of “escape agriculture,” but here we may simply note the overall pattern of the agricultural strategies employed in maroon communities.²⁶ We shall, in the context of describing upland peoples in Southeast Asia, encounter practices that bear a strong family resemblance to those of the maroons.

Runaway slaves clustered in precisely those out-of-the-way places where they could not easily be found: swamps, rough mountain country, deep forests, trackless wastes. They chose, when possible, defensible locations accessible by only a single pass or trail that could be blocked with thorns and traps and observed easily. Like bandits, they prepared escape routes in case they were found and their defenses failed. Shifting cultivation, supplemented by foraging, trade, and theft, was the commonest maroon practice. They preferred to plant root crops (for example, manioc/cassava, yams, and sweet potatoes), which were unobtrusive and could be left in the ground to be harvested at leisure. Depending on how secure the site was, they might plant more permanent crops, such as bananas, plantains, dry rice, maize, groundnuts, squash, and vegetables, but such crops could more easily be seized or destroyed. Some of these communities were short-lived, others survived for generations. Outside the law by definition, many maroon communities lived in part by raiding nearby settlements and plantations. None, it seems, were self-sufficient. Occupying a distinctive agro-ecological zone with valued products, many maroon settlements were closely integrated into the larger economy by clandestine and open trade.

Shifting Agriculture as “Escape-Agriculture”

Rather than being dictated by necessity, then, the adoption of shifting agriculture may have been part of a distinctive politics.

—Ajay Skaria, Hybrid Histories, 1999

Shifting cultivation is the most common agricultural practice in the hills of mainland Southeast Asia. Those who practice it are rarely understood to have made a choice, let alone a political choice. Rather, the technique is seen by lowland officials, including those in charge of development programs in the hills, as both primitive and environmentally destructive. By extension, those who farm this way are also coded as backward. The implicit assumption is that, given the skills and opportunity, they would abandon this technique and take to permanent settlement and fixed-field (preferably irrigated rice) farming. Again, movement from swiddening to wet rice was seen as unidirectional and evolutionary.

Contrary to this view, my claim is that the choice of shifting cultivation is preeminently a political choice. This claim hardly originates with me, and, in the argument that follows, I shall rely on the judgment of many historians and ethnographers who have examined the issue closely. The foremost Chinese specialist on swiddening techniques and swiddening peoples in Yunnan rejects outright the claim that it is an earlier or more primitive technique form of cultivation, bound to be abandoned once its practitioners master irrigation techniques: “But it must be stressed here that it is incorrect to take Yunnan swidden agriculture as a representative of such a primitive ‘stage’ of agricultural history. In Yunnan, swiddening, knives and axes, coexist with hoes and plows, and have their different uses and functions. It is difficult to say which came earlier and which later.… But the crux of the matter is that there is no basis for taking our ‘pure’ swidden agriculture as the original state of affairs.”²⁷

To choose swiddening or, for that matter, foraging or nomadic pastoralism is to choose to remain outside state space. This choice has historically been the bedrock of freedom enjoyed by Southeast Asian commoners. The subjects of the small Tai statelets (muang) in the hills, Richard O’Connor points out, always had two alternatives. One alternative was to shift residence and affiliation to another muang where conditions were more advantageous. “Yet another escape was to farm the hills rather than paddy land.” O’Connor points out: “A hill farmer had no corvée obligations.”²⁸ More generally, swiddening facilitated physical mobility and, on that account alone, according to Jean Michaud, could also “be used as an escape or survival strategy by groups needing to move such as the Hmong or Lolo from China.… These once sedentarized groups were set in motion by adversity, wars, climatic change, or untenable demographic pressure in their homelands.”²⁹ Shifting cultivation was understood to be outside the fiscal and manpower apparatus of even the smallest states. It is for precisely this reason that the representatives of historical states in mainland Southeast Asia have spoken unanimously in discouraging or condemning swidden cultivation. Shifting cultivation was a fiscally sterile form of agriculture: diverse, dispersed, hard to monitor, hard to tax or confiscate. Swiddeners were themselves dispersed, hard to monitor, hard to collect for corvée labor or conscription. The features that made swiddening anathema to states were exactly what made it attractive to state-evading peoples.³⁰

Irrigated rice and shifting cultivation are not a temporal, evolutionary sequence, nor are they mutually exclusive alternatives.³¹ Many hill populations practice both irrigated-rice cultivation and shifting cultivation simultaneously, adjusting the balance according to political and economic advantage. By the same token, valley populations have in the past replaced irrigated rice with swiddening, especially when epidemics or migrations suddenly made more land available. In a great many geographical settings, shifting, dry cultivation, or irrigated rice is possible. With terracing and the availability of reliable springs or streams, irrigated rice can be grown at relatively high altitudes and in steep terrain. The sophisticated rice terraces of the Hani in the upper reaches of the Red River in Vietnam and of the Ifugao in Northern Luzon are cases in point. Spring- and stream-fed, terraced rice fields are also found among the Karen and Akha. The earliest archeological remains of rice cultivation in Java and Bali come not from the lowlands but from the midslope uplands skirting mountains and volcanoes, where perennial springs and a pronounced dry season made it practical.³²

Lowland officials, both colonial and contemporary, have seen shifting cultivation not simply as primitive but as inefficient in the strict sense of neoclassical economics. To some degree this is an unwarranted deduction from the apparent disorder and variety of a swidden as compared to a mono-cropped rice padi. At a deeper level, it represents a misunderstanding of the concept of efficiency. Wet rice is, to be sure, more productive per unit of land than shifting cultivation. It is, however, typically less productive per unit of labor. Which of the two systems is the more efficient depends mainly on whether land or labor represents the scarcer factor of production. Where land is comparatively plentiful and labor scarce, as has been the case historically in most of mainland Southeast Asia, shifting cultivation was more labor saving per unit of output and hence more efficient. The importance of slavery in state-making is evidence that coercion was required to capture shifting cultivators and move them to the labor-intensive padi fields, where they could be taxed.

The relative efficiencies of each agricultural technique varied not only with the demography but also with agro-ecological conditions. In areas where annual river flooding deposited fertile silt that could be easily worked, flood-retreat farming of irrigated rice was far less labor intensive than where elaborate irrigation works or ponds (tanks) were required. Where, on the contrary, the terrain was steep and the water supply unreliable, the labor cost of irrigated rice would be nearly prohibitive. Such evaluations of relative efficiency in terms of factor costs, however, entirely miss the determining political context. Despite the enormous amounts of labor involved in their construction and maintenance, elaborate irrigated rice terraces have been created in the hills against any plausible neoclassical logic. The reason, it appears, is largely political. Edmund Leach wondered about terracing in the Kachin hills and concluded that it took place for military reasons: to protect a key pass and to control its trade and tolls, which required a concentrated and self-provisioning military garrison.³³ Such an enterprise was, in effect, an effort to sculpt a miniature agro-ecological space in the hills that might support a statelet. In other instances it seems that terracing, like the fortified ridge settlements reported by early colonial travelers, were necessary for defense against raiding by lowland states and the slaving expeditions that fed their manpower needs. Here again, the logic was political, not economic. A successful defense against slave raids required both a relatively inaccessible location and a critical mass of concentrated defenders who could prevail against any but the largest and most determined foes.³⁴ Michaud suggests that the highland wet-rice terraces of the Hani in northern Vietnam are the work of a people who wish to be sedentary and at the same time well away from state centers.³⁵

Under most conditions, however, shifting cultivation was the most common agropolitical strategy against raiding, state-making, and state appropriation. If it makes sense to think of rugged terrain as representing a friction of distance, then it may make just as much sense to think of shifting cultivation as representing, strategically, the friction of appropriation. The decisive advantage of swiddening is its inherent resistance to appropriation, a political advantage that, in turn, pays economic dividends.

To illustrate this political advantage, let us imagine a demographic and agro-ecological setting in which either swiddening or wet rice is possible and in which neither technique is markedly superior to the other in terms of efficiency. The choice in this case becomes a political and sociocultural one. The great political advantage of swiddening is population dispersal (favoring escape rather than defense), poly-cropping, staggered maturities of crops, and an emphasis on root crops that can remain in the ground for some time until harvested. For the state or a raiding party, it represents an agricultural surplus and population that is difficult to assess, let alone seize.³⁶ It is an agricultural technique that, short of foraging, maximizes the friction of appropriation. If, on the other hand, the population chooses to grow padi rice, they represent an easy target for a state (or raiders), who know where to find them and their crops, carts, plow animals, and possessions. The likelihood of having oneself and one’s crops confiscated or destroyed is greatly increased; the friction of appropriation is reduced.

Thus even a purely economic evaluation of shifting cultivation must allow for the political advantage it offers in evading taxes and corvée and in making raiding less lucrative. If the gross return from padi farming were more or less equivalent to the return to shifting cultivation, its net return would still be inferior because, with padi farming, the farmer must surrender “rents” in the form of labor and grain. There are, then, two advantages to swiddening: it offers relative autonomy and freedom (though not without its own dangers), and it allows the farmers to dispose of their own labor and of the fruits of their labor. Both are essentially political advantages.

To practice hill farming is to choose a social and political life outside the framework of the state.³⁷ The element of deliberate political choice is emphasized most eloquently by Michael Dove in his analysis of Javanese states and agriculture: “Just as cleared land became associated with the rise of Javanese states and their cultures, so did the forest become associated with uncivilized, uncontrollable, and fearful forces.… The historical basis for this fear was empirical, since the swidden cultivators of ancient Java were neither part of a reigning court culture nor—and this is most important—under its control.”³⁸ It is but a small step to suggest, as Hjorleifur Jonsson has in his study of the Yao/Mien on the Thai-Chinese border, that swiddening is practiced in large part because it is beyond the reach of the state. It is the state’s identification with wet rice, he suggests, that gives rise to the political meaning of what might otherwise be a more politically neutral choice among agricultural techniques. “The two agricultural methods may historically have been practiced in conjunction, but the state’s issue of control forces people to stand with the state as wet-rice farmers, craftspeople, soldiers or whatever, or they stand without, as swidden farmers.”³⁹

The Hmong/Miao provide an instructive case. They are typically considered an emblematic highland ethnic group living above nine hundred meters by swidden farming of opium, maize, millet, root crops, buckwheat, and other highland cultivars. But in fact, Hmong can be found practicing a great variety of agricultural techniques. As one farmer put it, “We, Hmong, some of us only cultivate [dry] fields, some of us only cultivate wet rice, and some of us do both.”⁴⁰ What appears to be operating here is a political judgment about how much distance a community should put between itself and the state. Where the state is not a clear and present danger—or, more rarely, an irresistible temptation—the choice is not so politically freighted. But where the state looms over the choice both culturally and politically, agricultural technique comes to represent a decision between being a state subject or a “hill tribe”—or, still more precarious, straddling the divide. Of the subsistence alternatives available to cultivators, shifting agriculture, by virtue of the obstacles (friction) it places in the way of appropriation, is the most common state-repelling option.

Crop Choice as Escape Agriculture

The logic of escape agriculture and the friction of appropriation apply not only to a technical complex as a whole, such as shifting cultivation, but to particular crops as well. Of course, the overall resistance of swiddening to state appropriation lies both in its hilly location and dispersal and in the very botanical diversity it represents. It is not uncommon for shifting cultivators to plant, tend, and encourage as many as sixty or more cultivars. Imagine the bewildering task facing even the most energetic tax collector attempting to catalogue, let alone assess and collect taxes, in such a setting.⁴¹ It is for this reason that J. G. Scott noted that the hill peoples were “of no account whatever in the state” and that “it would be a sheer waste of energy in the eyes of an official to attempt to number the houses or even the villages of these people.”⁴² Add to this the fact that nearly all swidden cultivators also hunt, fish, and forage in nearby forests. By pursuing such a broad portfolio of subsistence strategies they spread their risks, they ensure themselves a diverse and nutritious diet, and they present a nearly intractable hieroglyphic to any state that might want to corral them.⁴³ This is a major reason why most Southeast Asian states were reduced to capturing the swiddeners themselves and removing them forcibly to an already established, state space.⁴⁴

Particular crops have characteristics that make them more or less resistant to appropriation. Cultivars that cannot be stored long without spoiling, such as fresh fruits and vegetables, or that have low value per unit weight and volume, such as most gourds, rootcrops, and tubers, will not repay the efforts of a tax gatherer.

In general, roots and tubers such as yams, sweet potatoes, potatoes, and cassava/manioc/yucca are nearly appropriation-proof. After they ripen, they can be safely left in the ground for up to two years and dug up piecemeal as needed. There is thus no granary to plunder. If the army or the taxmen wants your potatoes, for example, they will have to dig them up one by one. Plagued by crop failures and confiscatory procurement prices for the cultivars recommended by the Burmese military government in the 1980s, many peasants secretly planted sweet potatoes, a crop specifically prohibited. They shifted to sweet potatoes because the crop was easier to conceal and nearly impossible to appropriate.⁴⁵ The Irish in the early nineteenth century grew potatoes not only because they provided many calories from the small plots to which farmers were confined but also because they could not be confiscated or burned and, because they were grown in small mounds, an [English!] horseman risked breaking his mount’s leg galloping through the field. Alas for the Irish, they had only a minuscule selection of the genetic diversity of New World potatoes and had come to rely almost exclusively on potatoes and milk for subsistence.

A reliance on root crops, and in particular the potato, can insulate states as well as stateless peoples against the predations of war and appropriation. William McNeill credits the early-eighteenth-century rise of Prussia to the potato. Enemy armies might seize or destroy grain fields, livestock, and aboveground fodder crops, but they were powerless against the lowly potato, a cultivar which Frederick William and Frederick II after him had vigorously promoted. It was the potato that gave Prussia its unique invulnerability to foreign invasion. While a grain-growing population whose granaries and crops were confiscated or destroyed had no choice but to scatter or starve, a tuber-growing peasantry could move back immediately after the military danger had passed and dig up their staple, a meal at a time.⁴⁶

Other things equal, crops that will grow on marginal land and at high altitudes (for example, maize) favor escape because they allow their cultivators more space to disperse in or flee to. Cultivars that require little attention and/or that mature quickly are also state repelling inasmuch as they afford more mobility than labor-intensive, long-maturation crops.⁴⁷ Unobtrusive crops of low stature that mimic much of the natural vegetation around them thwart appropriation by being easy to overlook.⁴⁸ The greater the dispersal of the crops, the more difficult they are to collect, in the same way that a dispersed population is more difficult to grab. To the degree that such crops are part of the swiddener’s portfolio, to that degree will they prove fiscally sterile to states and raiders and be deemed “not worth the trouble” or, in other words, a nonstate space.

Southeast Asian Swiddening as Escape

Once we have shed the erroneous idea that shifting cultivation is necessarily historically prior to, more primitive than, and less efficient than fixedfield cultivation, there remains one further illusion to shed. That illusion is that it is a relatively static technique that has not changed much in the past millennium. On the contrary, one could argue that swiddening and, for that matter, foraging have undergone far more transformation in that period than has wet-rice cultivation. Some scholars claim that the shifting cultivation with which we are familiar was essentially a product of iron and, later, steel blades, which massively reduced the labor required to clear swiddens.⁴⁹ What is certain, however, is that the steel axe made escape through shifting agriculture both possible in previously hard-to-clear areas and less onerous generally.

At least two other historical factors worked to transform swiddening. The first was international trade in high-value goods that had, at least since the eighth century, linked both swiddeners and foragers to international markets. Pepper, which was the most valuable commodity in world trade between 1450 and 1650, save gold and perhaps slaves, is the most striking example. Before that medicinal herbs, resins, animal organs, feathers, ivory, and aromatic woods were much sought after in the China trade. One Bornean specialist goes so far as to argue that the very purpose of shifting cultivation was to sustain a population of traders scouring the forest for valuable trade goods.⁵⁰ The final factor transforming shifting cultivation was the arrival of an entire suite of New World plants from the sixteenth century on that vastly extended the scope and ease of swiddening. Quite apart from its margin of political autonomy, then, the comparative economic advantage of swiddening vis-àvis irrigated rice would only have improved from the sixteenth century to the nineteenth, while affording, as it always had, access to international trade goods.

How decisive such factors were in the massive flight and movement to shifting cultivation by Burmese living in the state core region in the early years of the nineteenth century is difficult to gauge. Nevertheless, the event is diagnostic for our purposes. Swiddening is typically seen as a practice confined to ethnic minorities. Here, however, we have a case of a putatively Burman padi-state population turning to it. The circumstances of their departure from the core approximate a limiting case of crushing taxes and corvée. As noted in Chapter 5, King Bò-daw-hpaya’s early-nineteenth-century ambitions for conquest, pagoda building, and public works caused massive destitution among his subjects. The response was rebellion, banditry, and above all, headlong flight. Core land was abandoned by cultivators to such an extent that officials began to record large tracts of abandoned farmland. “In the face of these exactions, many families decamped to less accessible rural locales.” This prompted, as William Koenig notes, a wholesale “movement to shifting cultivation.”⁵¹ A massive reapportionment of population ensued, with the king’s subjects fleeing out of range and/or practicing a form of agriculture far more impervious to seizure.

There are also good reasons to believe that much of the Mon population, previously sedentary, Theravada, wet-rice cultivators, abandoned their padi fields as a consequence of a series of wars, punctuated with revolts, against the Burman court at Ava in the mid-eighteenth century. Their flight, along with many of their Karen allies, from the chaos and defeat appears to have been accompanied by a retreat to shifting agriculture to protect their food supply as well.⁵²

Flight and shifting cultivation were not uncommon as a response to the colonial state when its claims, too, became intolerable. Georges Condominas notes that French colonial officials in Laos complained frequently of “seeing whole villages move when their responsibilities became too burdensome; for example, their village was situated near a road which they were constantly expected to maintain.”⁵³ Such movement was typically associated with swiddening since the Laotian, Thai, and Vietnamese peasantry knew that their swiddens were illegible and hence likely to evade appropriation.

The resort to shifting cultivation and foraging as a means of escape from the deadly perils of warfare is not merely of antiquarian interest. During World War II and the subsequent counterinsurgency warfare in Southeast Asia, retreat up the watershed and out of harm’s way was often an option. The Punan Lusong in Sarawak had begin to grow rice before 1940, but with the Japanese invasion they returned to the forest as foragers and swiddeners and did not return to fixed cultivation until 1961. In this they were not unlike neighboring Kenyah and Sebop farmers, who may leave their fields to range the forests for two or three years at a time, subsisting on sago palm and game. Nor did this adaptation necessarily signal penury, although during the war the usual trade outlets were closed, since sago palm has at least double the caloric return to labor as hill-rice swiddening.⁵⁴ On the peninsula in Western Malaysia, the Jakun (Orang Malayu Asli) fled to the upper reaches of the Sungei Linggui (Linggui River) to avoid contact with or capture by the Japanese forces. Prized for their knowledge of the forest, they were liable to be pressed into service as guides and porters by the Japanese and, afterward, during the Emergency, by British forces or Communist rebels. They lived on the run on cassava, sweet potatoes, bananas, some vegetables, and a small amount of rice for the old people and children. They ate their noisy roosters lest the crowing betray their whereabouts.⁵⁵

Southeast Asian Escape Crops

“Escape crops” may have one or more of several characteristics that facilitate evasion of raiding either by states or by freebooters. In many cases they simply qualify by being well adapted to environmental niches that are difficult to map and control: high, rugged mountains, swamps, deltas, mangrove coasts, and so on. If, in addition, they are of staggered maturity, fast growing, and easily hidden, if they require little care, are of little value per unit weight and volume, and grow below ground, they acquire greater escape value. Many such cultivars are ideally adapted to swiddening routines, in which case their escape value is still further enhanced.⁵⁶

Before the introduction of New World crops, a few high-altitude grains offered those seeking autonomy from the state a certain amount of running room. Oats, barley, fast-growing millets, and buckwheat were tolerant of poor soils, high altitudes, and short growing seasons, as were cabbage and turnips, and allowed people to settle at higher altitudes than hill rice would permit. Old World roots and tubers, taro and yams, as well as the sago palm, were also favored by nonstate peoples.⁵⁷ Taro could be grown at relatively high elevations, though it required wet, fertile soils. It could be planted anytime; it ripened quickly; it required little care or preparation before eating; and once ripe it could be left in the ground and dug up as needed. Yams, which also grew wild, had many of the same advantages, and then some. Though yams required more labor and had to be planted at the end of the rainy season, they were less susceptible to insect and fungal attack, would grow under a greater variety of conditions, and could be sold as a cash crop in markets. Until both were overtaken by New World cultivars, yams tended to replace taro because, Peter Boomgaard believes, much of the land suitable for taro was increasingly planted to irrigated rice, while yams were more suited to the drier hillsides. The sago palm (not a true palm) and the powdery starch derived from splitting its trunk, crushing, kneading, washing, and grating its pith also qualifies as an escape food. It is naturally occurring and fast growing, involves less work than hill rice or perhaps even cassava, and will thrive in swampy environments. Its starchy powder can be sold or bartered, as can yams, but it will not grow at altitudes above nine hundred meters.⁵⁸ All these foods were known as “famine” foods. Even wet-rice growers often depended on them during that hungry time before the new rice crop was gathered. For others, however, they were the basis of a diet that could be shielded from state appropriation.

Escape agriculture was radically transformed beginning in the sixteenth century with the introduction of New World plants. Maize and cassava played such a decisive role in this transformation that each merits its own discussion. Some of the generic characteristics of New World crops, however, stand out. Above all, like many “exotics” taken to completely new ecological settings, they initially had no natural pests and diseases, as they had at home. Hence they tended to thrive in the new environment. This advantage, as much as any, explains why they were adopted with great alacrity in much of Southeast Asia, especially by those who wished to live beyond the reach of the state. The sweet potato was a striking example. Georg Eberhard Rumphius, the great Dutch botanist and illustrator, was amazed to discover how swiftly its cultivation had spread throughout the Dutch East Indies by 1670. Among its advantages were high yields, disease resistance, nutritional value, and tastiness. Its value as an escape crop, however, rested on three characteristics: it matured quickly, it had a higher caloric yield for the labor than indigenous edible roots and tubers, and, perhaps most decisive, it could be grown successfully at higher elevations than yams or taro. Boomgaard implies that the sweet potato may have aided flight by raising the population of highland areas where it was often (as in New Guinea) combined with pig husbandry. Its cultivation had also spread to nomadic and semisedentary populations in such inaccessible places as the island of Buru.⁵⁹ The sweet potato’s status as an escape crop was even more evident in the Philippines, where the Spaniards blamed it for the nomadism of the Igorot, whom they could neither count nor settle: “[They move] from one place to another on the least occasion for there is nothing to stop them since their houses, which are what would cause them concern, they make any place with a bundle of hay; they pass from one place to another with their crops of yames and camotes [sweet potato] off of which they live without much trouble, pulling them up by the roots, since they can stick them in wherever they wish to take root.”⁶⁰ Any crop that allowed people to move to hitherto inaccessible areas and to provision themselves successfully there was, by definition, a crop stigmatized by the state.

Amid our discussion of food-crops, it is important to recall that no matter how isolated a hill people or maroon community was, they were never entirely self-sufficient. Virtually all such groups grew, hunted, or foraged for valuable trade goods that could be bartered or sold in lowland markets. They aimed to have the advantages of trade and exchange while remaining politically autonomous. Historically such trade crops included cotton, coffee, tobacco, tea, and, above all, opium. These crops required more labor and had sedentarizing features, but if the communities that grew them were beyond the state’s range, they were compatible with political independence.

For any particular crop, it is possible to estimate roughly how suitable it is for the purpose of state evasion. Table 3 is confined, with the exception of opium and cotton, to a comparison of food crops along these dimensions.⁶¹ An ordinal scale of “escapability” is unrealistic inasmuch as considerations of labor intensivity, hardiness, and storability admit of no comprehensive metric. Given a specified agro-ecological niche, however, nominal comparisons are plausible. The examination of how two such crops, maize and cassava (also known as manioc or yucca), both New World cultivars, came to be valued for their escape characteristics will provide the historical context that the more global comparisons in the table necessarily lack.

MAIZE

Brought by the Portuguese to Southeast Asia in the fifteenth century, maize spread rapidly.⁶² It was firmly established throughout maritime Southeast Asia by the late seventeenth century, and by the 1930s it counted for roughly a quarter of smallholder cropping. So firmly had it become established and worked into local cosmologies that, along with the chili pepper, another New World cultivar, it was considered an indigenous crop by most Southeast Asians.

If one were designing an escape grain, one could hardly do better. Maize had many advantages over hill rice. Not only did it have higher caloric yields per unit labor and per unit land than hill rice, but its yields were more reliable; it could survive more erratic weather. Maize could easily be intercropped with other cultivars; it matured quickly; it could be used as fodder; it stored well if dried; and it was nutritionally superior to hill rice. For our purposes, however, what mattered most is that it “could be grown in areas that were too high, too steep, too dry, and too infertile for hill rice.”⁶³ These virtues allowed both hill peoples and valley peoples to colonize new zones that had previously been forbidding. They could settle farther up a watershed, at an elevation of twelve hundred meters or more, and still have a reliable staple. They could, in steep, inaccessible places where the friction of distance provided some security, establish a quasi-sedentary existence outside the ambit of the state. In upland plateaus where irrigated rice had been grown for a long time, it allowed communities to colonize the nearby hills outside the padi core.

With maize, an autonomous existence outside the padi state suddenly became far easier and more tempting. The opportunity was seized by so many people that it prompted a significant redistribution of population. As Boomgaard puts it, “Maize, then, may have enabled groups or individuals, who, for political, religious, economic, or health reasons, wanted to leave the population centers in the lowlands or the upland villages to survive and even flourish in hitherto sparsely populated mountain areas.”⁶⁴ A stronger claim has been made that the availability of maize was instrumental in the constitution of upland, nonstate societies. In the case of the Hindu-Javanese living in the Tengger uplands of East Java, Robert Hefner believes that maize may well “have played a role in facilitating the slow retreat of the Hindu farmers upslope into the less accessible terrains of the Tengger highlands in the aftermath of the Muslim conquest of Hindu Majapahit.”⁶⁵ Elsewhere, it appears that maize and other upland crops (potatoes, cassava) were often critical in the creation of upland populations and in codifying their political and cultural distinctiveness from the lowland state. The reasons for moving away from state space could vary dramatically—religious division, war, corvée, forced cultivation under colonial schemes, epidemics, flight from bondage—but the availability of maize was a new and valuable tool for potential runaways.⁶⁶

Table 3 Escape Characteristics of Crops

The highland Hmong living in or near Thailand and Laos have, for the past two centuries, been fleeing both from Han military pressure and from the aftermath of failed rebellions against the Han and, later, the French in Tonkin. Living at altitudes generally above one thousand meters and growing maize, pulses, root crops, gourds, and the opium poppy, they are very much a nonstate people. It is maize, in particular, that has been instrumental in making good their escape. Hill rice will generally not grow above one thousand meters; the opium poppy, on the other hand, thrives only above nine hundred meters. If the Hmong were to rely on hill rice and opium as their main crops, they would be confined to the narrow band between nine hundred and one thousand meters. With maize, however, they can range another three hundred meters higher, where both maize and the opium poppy thrive and where they are even less likely to attract the attention of the state.

CASSAVA/MANIOC/YUCCA

The champion New World escape crop was, without question, cassava.⁶⁷ Like maize, it spread quickly throughout both maritime and mainland Southeast Asia. It could be grown almost anywhere under an amazing variety of conditions. So hardy and self-sufficient is this large root plant that preventing it from growing seems almost more difficult than cultivating it.⁶⁸ It is ideal for opening new land; it is drought resistant; it will tolerate soils in which virtually nothing else can be grown; it, like other transplanted New World cultivars, has few natural enemies; and, compared with taro and the sweet potato, it is less attractive to wild pigs.⁶⁹ If it has one drawback, it is that it does not flourish at the highest altitudes like maize and the potato, but otherwise, it places few restrictions on where one can settle or roam.

Cassava shares escape features with other roots and tubers. Although it does not mature as fast as, say, the sweet potato, it can be allowed to ripen and left in the ground until needed. The combination of its versatility and hardiness, together with the fact that only the aboveground foliage can be destroyed by fire, earned it the name, farina de guerra—roughly, staple or flour of war—in the Spanish-speaking world. Guerrillas represent, after all, something of a limiting case of state-evading, mobile peoples. A further advantage of cassava is that, once harvested, it can be made into a kind of flour (tapioca) which can then be stored for some time. Both the root and the flour can be sold in the market.

Perhaps the most striking advantage of cassava, however, is its undisputed status as the crop requiring the least labor for the greatest return. For that reason, it was much favored by nomadic peoples who could plant it, leave, and then return virtually anytime in the second and third years to dig it up. In the meantime, its leaves can be eaten. Cassava allows its planters to occupy virtually any ecological niche, roam more or less at will, and avoid a great deal of drudgery. On the basis of its striking advantages, it became the most common root crop, displacing the sweet potato, which had, in turn, displaced the yam.

To the padi state, whether precolonial or colonial, such easily accessible and labor-saving subsistence crops, though valued in a pinch as famine foods, were a threat to state-making. The state’s interests were best served by maximizing padi land or, failing that, other important cash, export crops such as cotton, indigo, sugar cane, and rubber, often using servile labor. Access to New World escape crops made the economics of escape as tempting as its politics. Colonial officials tended to stigmatize cassava and maize as crops of lazy natives whose main aim was to shirk work. In the New World, too, those whose job it was to drive the population into wage labor or onto the plantations deplored crops that allowed a free peasantry to maintain its autonomy. Hacienda owners in Central America claimed that with cassava, all a peasant needed was a shotgun and a fishhook and he would cease to work regularly for wages.⁷⁰

Cassava, like many root crops, has a large impact on social structure that, in turn, bears on state evasion. This impact makes for an illuminating contrast with grain cultures generally, and with wet-rice cultures in particular.⁷¹ Wet-rice communities live by a single rhythm. Planting, transplanting, and harvesting, and their associated rituals, are closely coordinated, as is water control. Cooperation in water management, crop watching, and labor exchange is rewarded if not mandated. Not so with root crops like sweet potatoes and cassava. Planting and harvesting take place more or less continuously according to the choices and needs of the family unit. Little or no cooperation is required by the agronomic characteristics of the crop itself. A society that cultivates roots and tubers can disperse more widely and cooperate less than grain growers, thereby encouraging a social structure more resistant to incorporation, and perhaps to hierarchy and subordination.

Social Structures of Escape

The padi state requires and fosters a legible landscape of irrigated rice and the concentrated population associated with it. This accessible economy and demography might be termed an appropriable landscape. Just as there are economic landscapes that lend themselves to monitoring and appropriation, so too are there social structures that lend themselves to control, appropriation, and subordination. The contrary is also true. There are, as we have seen, agricultural techniques and crop regimens that are resistant to appropriation, and hence are state repelling. By the same token, there are patterns of social and political organization that are resistant to monitoring and subordination. Just as shifting cultivation and cassava planting represent a “positionality” vis-à-vis the state, so, too, do various forms of social organization represent a strategic position with respect to the state. Social structure, like agricultural technique, is not a given; it is substantially, especially over time, a choice. Much of that choice is in a broad sense political. Here a dialectical view of social organization is necessary. Peripheral political structures in mainland Southeast Asia are always adjusting to the state systems that make up their immediate environment. Under some circumstances they, or rather the human actors who animate them, may adjust that structure so as to facilitate alliances with or incorporation into a nearby state. At other times, they may pattern themselves so as to break loose from ties of tribute or incorporation.

Social structure, in this view, ought to be seen not as a permanent social trait of a particular community but rather as a variable, one of the purposes of which is to regulate relations with the surrounding field of power. Nowhere has this position been more articulately stated than by F. K. Lehman (aka Chit Hlaing) in his study of the Kayah in eastern Burma. After noting, as had Leach before him, the oscillation of social organization over time, he directed attention to the rules of transformation by which this oscillation might be understood: “Indeed, it seems impossible to make sense of the Kayah, or any other category of Southeast Asian hill people, without thinking of the social system in approximately the foregoing terms. It appears to be an ineluctable premise of these societies that one changes one’s social structure, sometimes even one’s ‘ethnic’ identity, in response to periodic changes in ongoing relations with neighboring civilizations.”⁷²

Broadly speaking, whenever a society or part of a society elects to evade incorporation or appropriation, it moves toward simpler, smaller, and more dispersed social units—toward what we have earlier termed the elementary forms of social organization. The most appropriation-resistant social structures—though they also impede collective action of any kind—are acephalous (“headless”) small aggregates of households. Such forms of social organization, along with appropriation-resistant forms of agriculture and residence, are invariably coded “barbarian,” “primitive,” and “backward” by the lowland padi “civilizations.” It is no coincidence that this metric of more or less civilized agriculture and social organization should so perfectly map onto their suitability for appropriation and subordination, respectively.

“Tribality”

The state’s relation with tribes, though it preoccupied Rome and its legions, has long since disappeared from European historiography. One by one, Europe’s last independent, tribal peoples—the Swiss, the Welsh, the Scots, the Irish, the Montenegrins, and nomads of the south Russian steppe—were absorbed into more powerful states and their dominant religions and cultures. The issue of tribes and states, however, is still very much alive in the Middle East. Thus it is from the ethnographers and historians of tribal-state relations there that we can begin to take our bearings.

Tribes and states, they agree, are mutually constituting entities. There is no evolutionary sequence; tribes are not prior to states. Tribes are, rather, a social formation defined by its relation to the state. “If rulers of the Middle East have been preoccupied by a ‘tribal problem,’ … tribes could be said to have had a perennial ‘state-problem.’”⁷³

One reason why tribes often appear to be stable, enduring, genealogically and culturally coherent units is that the state typically desires such units and sets out, over time, to fashion them. A tribe may spring into existence on the basis of political entrepreneurship or through the political identities and “traffic patterns” that a state can impose by structuring rewards and penalties. The tribe’s existence, in either case, depends on a particular relationship to the state. Rulers and state institutions require a stable, reliable, hierarchical, “graspable” social structure through which to negotiate or rule. They need an interlocutor, a partner, with whom to parlay, whose allegiance can be solicited, through whom instructions can be conveyed, who can be held responsible for political order, and who can deliver grain and tribute. Since tribal peoples are, by definition, outside the direct administration of the state, they must, if they are to be governed at all, be governed through leaders who can speak for them and, if necessary, be held hostage. The entities represented as “tribes” seldom exist with anything like the substantiality of state imaginings. This misrepresentation is due not only to the official identities cooked up by the state but also to the need of ethnographers and historians for social identities that can serve as a coherent object of description and analysis. It is hard to produce an account of, let alone govern, a social organism that is continually going in and out of focus.

When nonstate peoples (aka tribes) face pressures for political and social incorporation into a state system, a variety of responses is possible. They, or a section of them, may be incorporated loosely or tightly as a tributary society with a designated leader (indirect rule). They may, of course, fight to defend their autonomy—particularly if they are militarized pastoralists. They may move out of the way. Finally, they may, by fissioning, scattering, and/or changing their livelihood strategy, make themselves invisible or unattractive as objects of appropriation.

The last three strategies are options of resistance and evasion. The military option has, with a few exceptions, rarely been available to nonstate peoples in Southeast Asia.⁷⁴ Moving out of the way, inasmuch as it often involves adoption of shifting cultivation or foraging, has already been examined. What remains to be explored is the final strategy of social reorganization. It involves social disaggregation into minimal units, often households, and is often accompanied by the adoption of subsistence strategies that favor small, scattered bands. Ernest Gellner describes this deliberate choice among the Berbers with the slogan “Divide that ye be not ruled.” It is a brilliant aphorism, for it shows that the Roman slogan “Divide and rule” does not work past a certain point of atomization. Malcolm Yapp’s term for the same strategy, jellyfish tribes, is just as apt, for it points to the fact that such disaggregation leaves a potential ruler facing an amorphous, unstructured population with no point of entry or leverage.⁷⁵ The Ottomans, in the same vein, found it far easier to deal with structured communities, even if they were Christians and Jews, than with heterodox sects that were acephalous and organizationally diffuse. Most feared were such forms of autonomy and dissent as, for example, the mystical Dervish orders, which deliberately, it seems, avoided any collective settlement or identifiable leadership precisely to fly, as it were, beneath the Ottoman police radar.⁷⁶ Faced with situations of this kind, a state often tries to find a collaborator and create a chiefdom. While it is usually in someone’s interest to seize this chance, nothing, as we shall see, prevents his would-be subjects from ignoring him.

The elementary units of the tribal structure were like bricks; they could lie scattered or in heaps without discernible structure, or they could be joined together to build large, sometimes massive, tribal confederations. As Lois Beck, who has examined this process in exquisite detail for the Qashqa’i of Iran, describes it, “Tribal groups expanded and contracted. Some tribal groups joined larger ones when, for example, the state attempted to restrict access to resources or a foreign power sent troops to attack them. Large tribal groups divided into small groups to be less visible to the state and escaped its reach. Intertribal mobility [shifting ethnic identity] was a common pattern and was part of the process of tribal formation and dissolution.” In a Middle Eastern version of Pierre Clastres’s argument for Latin America, Beck points to agriculturalists who shifted to nomadism and sees both social organization and subsistence strategies as political options, sometimes deployed in the service of illegibility. “The forms that many people identify as primitive and traditional were often creations responding to, and sometimes mirroring, more complex systems.” Beck adds: “Such local systems adapted to and challenged, or distanced themselves from, the systems of those who sought to dominate them.”⁷⁷ Social structure, in other words, is, in large measure, both a state effect and a choice; and one possible choice is a social structure that is invisible and/or illegible to state-makers.

This theme of social shape-shifting is articulated in accounts of nomadic and foraging peoples. The amorphous nature of Mongolian social structure and its lack of “nerve centers” were credited by Owen Lattimore as preventing Chinese colonization.⁷⁸ Richard White’s meticulous analysis of Indian politics in colonial North America emphasizes the radical instability of tribal structure and identity, the autonomy of local groups, and the capacity to shift to new territory and alternate subsistence strategies quickly.⁷⁹ In the ethnic, migrant shatter zones that White examines, and which characterize much of Zomia, identities are genuinely plural. Such populations do not so much change identities as emphasize one aspect of a cultural and linguistic portfolio that encompasses several potential identities. The vagueness, plurality, and fungibility of identities and social units have certain political advantages; they represent a repertoire of engagement and disengagement with states and with other peoples.⁸⁰ Studies of pastoral nomadic groups such as the Turkmen on the Iranian-Russian border or the Kalmyk in Russia emphasize the capacity of such groups to divide or segment into small independent units whenever it was advantageous.⁸¹ A historian of the Kalmyk quotes Marshall Sahlins’s general description of tribesmen: “The body politic may then retain features of a primitive organism, covered by a protective exo-skeleton of chiefly authority, but fundamentally uncomplicated and segmented underneath.”⁸²

Several features of such societies appear to foster, and in some cases may require, a social structure that can be both disaggregated and reassembled. The existence of such common property resources as pasture, hunting grounds, and potential swiddens allows groups to strike out on their own and, at the same time, impede the development of large, permanent distinctions in wealth and status characteristic of inheritable private property. Equally important is a mixed portfolio of subsistence strategies—foraging, shifting cultivation, hunting, trade, livestock raising, and sedentary agriculture. Each form of livelihood is associated with its own forms of cooperation, group size, and settlement pattern. Together, they provide a kind of practical experience, or praxis, in several forms of social organization. A mixed portfolio of subsistence techniques yields a mixed portfolio of social structures that can easily be invoked for political as well as economic advantage.⁸³

Evading Stateness and Permanent Hierarchy

Every state with ambitions to control parts of Zomia—Han administrators in Yunnan and Guizhou, the Thai court in Ayutthaya, the Burmese court in Ava, Shan chiefs (Sawbwa), the British colonial state, and independent national governments—has sought to discover, or, failing that, to create chiefdoms with which they could deal. The British in Burma, Leach noted, everywhere preferred autocratic “tribal” regimes in compact geographical concentrations with which they could negotiate; conversely, they had a distaste for anarchic, egalitarian peoples who had no discernible spokesman. “In the Kachin Hills area … and also in many other areas of low population density, there is a large preponderance of very small independent villages; the headman of every village claims to be an independent chief of full du baw status.… This fact has been noted repeatedly and is the more remarkable in that the British administration was consistently opposed to such fragmented settlement.”⁸⁴ Another turn-of-the-century British official warned observers not to take the apparent subordination of petty Kachin chiefs seriously. “Beyond this nominal subordination, each village claims to be independent and only acknowledges its own chief.” This independence, he emphasizes, anticipating Leach, characterizes even the smallest social units; it “extends down even to the household and each house owner, if he disagrees with his chief, can leave the village and set up his own house elsewhere as his own sawbwa.”⁸⁵ Accordingly, the British, like other states, tended to label the democratic, anarchic peoples as “wild,” “raw,” “crude” (yaín—) vis-à-vis their more “tame,” “cooked,” “cultured,” and autocratic neighbors, even if those neighbors shared the same language and culture. Stable, indirect rule of anarchic “jellyfish” tribes was well nigh impossible. Even pacifying them was both difficult and impermanent. The British chief commissioner from 1887 to 1890 noted that the conquest of the Kachin and Palaung areas had to be accomplished “hill by hill” inasmuch as these peoples “had never submitted to any central control.” The Chins were, in his view, at least as frustrating. “Their only system of government was that of headmen of villages or at the most a small group of villages, and, consequently, negotiation with the Chin as a people was impossible.”⁸⁶

Daunted by the recalcitrant and slippery Chin, the British set about creating a chief in the “democratic” Chin area and enforcing his writ. Colonial support allowed the chief to sponsor lavish community feasts, which in a “feasting society” enhanced his relative status vis-à-vis commoners. In reaction, a new syncretic cult arose that repudiated community feasts while continuing the tradition of individual feasts that served to increase personal, not chiefly, status. This Pau Chin Hau cult was in short order adopted by the entire Zanniat (a democratic tribal area) and more than a quarter of the Chin population in that administrative division.⁸⁷ In this, as in many instances, it appears that independent status—taking one’s distance from the state and statelike formations—“was more highly valued than economic prosperity.”⁸⁸

The Wa, seen as perhaps the fiercest of the hill peoples, with a reputation for taking heads, are, like the “democratic” Chin and the gumlao Kachin, strongly egalitarian. They emphasize the equality of access to feasting and status competition, refusing to allow those who were already prominent or too wealthy to conduct further sacrifices lest they aspire to chiefdom status. This egalitarianism is, as Magnus Fiskesjö points out, constructed as a state repelling strategy: “The Wa egalitarianism, mistakenly construed as a ‘primitive’ society in Chinese or other evolutionisms, can also be understood as a way of avoiding the collapse of autonomy in the face of threats from the greater powers that loom on the horizon: the state(s) waiting to exact tribute or to enforce taxation, as they were already doing in the intermediary buffer zone (which here, in a sense, served the role of an ‘anti-barbarian’ defensive wall system we see elsewhere in China).”⁸⁹

Another response to the pressure to create a political structure through which the state can act is to dissimulate—to comply by producing a simulacrum of chiefly authority without its substance. The Lisu of northern Thailand, it seems, do just that. To please lowland authorities, they name a headman. The Potemkin nature of the headman is apparent from the fact that someone without any real power in the village is invariably named rather than a respected older male with wealth and ability.⁹⁰ An identical pattern has been reported for hill villages in colonial Laos, where bogus local officials and notables were produced on demand while respected local figures continued to guide local affairs, including the performance of the bogus officials!⁹¹ Here “escape social structure” is not so much a social invention for state evasion as it is an egalitarian, existing social structure that is protected by an elaborate staged performance of hierarchy.

The most celebrated ethnography of hill peoples anywhere in Zomia is Edmund Leach’s study of the Kachin, Political Systems of Highland Burma. Leach’s analysis has been the subject of a nearly unprecedented volume of scrutiny and criticism by nearly two generations of scholars. It is clear that Leach deliberately disregarded the larger political and economic changes (British imperial rule and the opium economy in particular) impinging on Kachin social organization in favor of his structuralist idea of an oscillating equilibrium.⁹² He also appears to have seriously misconstrued the vernacular terms for Kachin marriage-alliance systems and their effect on the permanence of social ranking by lineage. A thorough critical examination of his contribution by contemporary ethnographers appears in a volume recently edited François Robinne and Mandy Sadan.⁹³

Nothing in this distinguished critical literature, however, questions the fact that there are important differences in the relative openness and egalitarianism of various Kachin social systems or that there was, near the close of the past century, something like a movement to assassinate, depose, or desert the more autocratic chiefs. At its core, Leach’s ethnography is an analysis of escape social structure—a form of social organization designed to thwart capture and appropriation either by Shan statelets or by the petty Kachin chiefs (duwa) who attempt to mimic Shan power and hierarchy. Leach argues, to put it very briefly and schematically, that there are three models of political organization in the Kachin area: Shan, gumsa, and gumlao. The Shan model is a statelike structure of property and hierarchy marked by a hereditary (in principle) chief and systematic taxes and corvée. At the other extreme is the gumlao model, a model that repudiates all hereditary authority and class difference—though not individual differences in status. Gumlao villages, which were unrecognized by the British, are independent and typically have a ritual organization and tutelary deities that reinforce equality and autonomy. The Shan and gumlao forms, Leach argues, are relatively stable. Here it is crucial to underline that these are not ethnic distinctions as understood phenomenologically by Leach’s subjects. To move in a “Shan” direction is to be associated more closely with the hierarchy, ritual, and opportunities of this statelike social formation. To move in a gumlao direction is precisely to take one’s distance from the Shan state and its practices. Historically, people have moved back and forth between these models and codes.

The third model, the gumsa form, is an intermediate model of theoretically rigid and stratified lineages where wife-taking lineages are socially and ritually superior to wife-giving lineages, leading to a division between commoners and aristocrats.⁹⁴ This model, Leach claims, is particularly unstable.⁹⁵ The head of a top-ranked lineage in the gumsa system is well on the way to transforming himself into a petty Shan ruler.⁹⁶ At the same time, his effort to make his status permanent and to turn lower-ranked lineages into his serfs threatens to provoke a rebellion or flight, and hence a move toward gumlao equality.⁹⁷

For our purposes, Leach’s Kachin ethnography illustrates a model of egalitarian social organization readily at hand to prevent or evade state formation. Leach writes of the oscillation between these three models as if it were a permanent feature of Kachin society. And yet the gumlao form was also, in part, the result of a specific historical revolution. The Gazetteer of Upper Burma reports that the gumlao “revolt” began when two suitors for the daughter of a chief (duwa) were refused (an acceptance would have raised their status and that of their kin group).⁹⁸ They killed that duwa and the man to whom the daughter had been given. They went on, with followers, to depose many duwa, some of whom escaped death or exile by giving up their titles and privileges. This story is in keeping with the view, expressed by Leach, that the gumsa structure, by its stratified rankings, is likely to block the status aspirations of men from lower-ranked lineages, typically expressed through competitive feasting.⁹⁹ Leach’s own account of the proximate cause of the revolt is far more nuanced and elaborate, but at its center is the refusal to provide corvée labor, which, along with the thigh of slaughtered animals, was the prerogative of a chief.¹⁰⁰

Gumlao villages come about in either of two ways. First, as just described, they are the result of small-scale leveling revolutions that establish small, commoner republics. The second and perhaps more common origin is the migration of families and lineages from more stratified villages to found new, more egalitarian villages. The origin myths of gumlao villages emphasize one or the other. At this point, Leach proposes that the gumlao itself is unstable, since, as inequalities develop, those advantaged will strive to legitimate and codify those advantages with gumsa trappings. But another interpretation is possible: that gumlao communities are typically reproduced by fission, by small groups of families of equal status striking out on their own when they find that the inequalities have become stifling. Fission, as well as small-scale revolution, was greatly conditioned by demography and developments in the larger world. Inequalities might prove far more stifling where British pressure had diminished caravan revenue and slaving. The attractions of the frontier might prove more alluring in a booming opium market. Where there was less demographic pressure and hence plenty of available swidden land, fission was probably far more likely than revolt.¹⁰¹

Gumlao areas were anathema to the state. An early British account of the Kachin areas contrasted the ease of marching through the villages of a well-disposed hereditary chief with the difficulty of traversing “a gumlao village which is practically a small republic, the headman, however well-meaning he may be, is quite unable to control the actions of any badly-disposed villager.”¹⁰² Gumlao social organization was state repelling in a number of ways. Its ideology discouraged, or killed, would-be hereditary chiefs with feudal pretensions. It was resistant to tribute or control by the neighboring Shan principalities. Finally, it presented a relatively intractable anarchy of egalitarian, Lilliputian republics that were hard to pacify, let alone govern.

I have devoted some considerable space to gumlao villages as escape social structure not simply because it is well documented, thanks to Leach. There is more than a little evidence that many, if not most, hill peoples have bifurcated or even tripartite models of social organization: one approximating the egalitarian gumlao Kachin model, one approximating the more stratified gumsa model, and, occasionally, another approximating a petty Shan kingdom. Leach notes that “contrasted theories of government of this kind are current throughout the Burma Assam frontier area” and cites studies of the Chin, Sema, Konyak, and Nagas.¹⁰³ To Leach’s list, we can add more recent studies of the Karen and the Wa.¹⁰⁴ It would seem that just as hill peoples in mainland Southeast Asia are likely to have escape crops and escape agriculture in their economic repertoire, so too are they likely to have state-thwarting social models in their political repertoire.

In the Shadow of the State, in the Shadow of the Hills

Shortly before Burmese independence, an inquiry was held, to which tribal representatives were summoned. The chief of Mongmon, in the remote Northern Wa State, was asked what kind of administration he would favor. He replied, reasonably enough, “We have not thought about that because we are a wild people.”¹⁰⁵ The point about being a Wa, he understood better than the officials questioning him, was precisely not to be administered at all.

This diagnostic misunderstanding underlines the key fact that most hill societies are “shadow” or “mirror” societies. I mean by this that they are structures of political, cultural, economic, and often religious positioning, often self-consciously contradicting the forms and values of their more state-like neighbors. This defiance may come at some economic cost, according to Leach. He concludes that “the Kachins often value independence more highly than economic advantage.”¹⁰⁶ At the same time, those who migrate to lowland states and assimilate—and historically there have been a great many—enter valley society at its lowest rungs. In short-run status terms, as Lehman explains, a Chin entering Burman society has a choice between being a defective Burman or a successful Chin.¹⁰⁷

Identity in the hills is an implicit dialogue and debate about how to live. The interlocutors are the contrasting civilizations closest at hand. For peoples such as the Miao/Hmong, whose oral history records a long running battle with the Chinese/Han state, it is that dialogue which looms largest. The story the Hmong tell about themselves is thus something of a posture, a defense, a positioning in a debate with the Han and their state. Some Hmong debating points: they have emperors and we are all (notionally) equal; they pay taxes to overlords and we pay none; they have writing and books and we lost ours while fleeing; they live crowded in lowland centers and we live free, scattered in the hills; they are servile and we are free.¹⁰⁸

One might be tempted to conclude from this way of putting it that hill “ideology” was entirely derivative of valley ideologies. That would be mistaken for two reasons. First, hill ideology is in dialogue not only with valley societies but with other adjacent hill peoples and has other weighty matters like genealogy, the propitiation of the spirits, and the origin of man to deal with—matters that are somewhat less inflected by the debate with the valley centers. Second, and perhaps more important, if hill ideologies can be said to be deeply influenced by lowland states, it is equally the case that the lowland states, themselves historical aggregates of ingathered peoples, are preoccupied in explaining the superiority of their “civilization” vis-à-vis their “ruder” neighbors.

At least three themes in this connection appear again and again in the narratives and positional self-understandings of hill peoples. They might be termed equality, autonomy, and mobility, all understood relatively. As a matter of practice, of course, all three are encoded in material life in the hills—in location well away from lowland states, in dispersal, in common property, in shifting cultivation, and in the choice of crops. By choice, as Lehman has pointed out, hill peoples have “practiced an economy that the Burman [state] institutions were not adapted to exploit and, therefore, never thought of as part of the Burman kingdom.”¹⁰⁹ Just as “wet-rice cultivation implied a subject relationship to the polity, so to engage in swiddening was to some degree a statement of political positioning within a bifurcated regional culture of universalizing polities and forested hinterlands.”¹¹⁰

The gumlao Kachin, as we have seen, have a history of enforcing egalitarian social relations by deposing or assassinating overreaching chiefs. One imagines that this history and the narratives that accompany it operate as a chilling cautionary tale for lineage chiefs with autocratic ambitions. Whole districts in Karen, Kayah, and Kachin areas are known for their traditions of revolt.¹¹¹ Where the Kachin had chiefs, they were frequently ignored and shown no special respect. Other peoples have analogous traditions. The Lisu “loathe assertive and autocratic headmen,” and the “stories Lisu tell of murdered headmen are legion.”¹¹² The strict veracity of these stories matters less than the announcement it makes about norms of power relations.¹¹³ Similar stories circulate among the Lahu. Their society is described as “extremely egalitarian” by one ethnographer, and another claims that they are, in gender terms, as egalitarian as any people in the world.¹¹⁴ The Akha, for their part, reinforce their egalitarian practices with a mythic charter in which a chief and his son, who has a shamanic horse with wings mended with beeswax, flies too high. As with Icarus, his wings melt and he falls to his death. The “‘flowery’ exaggerated way” the story is told, “clearly shows an aversion to hierarchical chiefdom and state-formation.”¹¹⁵

The autonomy of hill peoples from permanent internal hierarchy and from state formation has depended absolutely on physical mobility. In this respect, the gumlao revolt is the exception that proves the rule. Flight, not rebellion, has been the basis of freedom in the hills; far more egalitarian settlements were founded by runaways than by revolutionaries. As Leach notes, “In the Shan case the villagers are tied to their [padi] land; the rice fields represent a capital investment. Kachins have no investment in the taungya [swidden—literally “hill cultivation”]. If a Kachin doesn’t like his chief he can go somewhere else.”¹¹⁶ It is the ability and, indeed, the practice of hill peoples to move at the drop of a hat and on the slightest pretext that bedeviled both the colonial regimes and the independent states of Southeast Asia. Although much of Zomia could be aptly described as a vast zone of refuge from state-making, movement was constantly taking place within Zomia from more stratified, statelike places to more egalitarian frontiers.

The hill Karen provide a case in point. Part or all of their small settlements would move to a new location, not simply to clear a new swidden but for many nonagricultural reasons as well. An inauspicious sign, a series of illnesses or death, a factional split, pressure for tribute, an overreaching headman, a dream, the call of a respected religious figure—any of these might be enough to prompt a move. Various state efforts to sedentarize the Karen and make use of them were frustrated by the constant fissioning and mobility of their settlements. In the mid-nineteenth century, when many Karen had, along with their Mon allies, fled Burma and accepted Thai authority, they would not permanently settle, as the Thai officials desired.¹¹⁷ For their part, the British tried to settle the Karen in subsidized “forest villages” in the Pegu Yoma, where they would practice a restricted swiddening regime and, not incidentally, become the guardians of valuable stands of teak. The scheme was resisted and the Karen moved away.¹¹⁸ Everything we know about the hill Karen—their historical fear of slavery, their self-image as an orphaned and persecuted people—suggests that their social structure and swiddening were designed to keep them at a safe distance from captivity. Safety also meant adopting pliable social structures. The hill Karen are commonly described as having an autonomous and loosely structured society—one that splits easily over economic, social, political, or religious issues.¹¹⁹

The utter plasticity of social structure among the more democratic, stateless, hill peoples can hardly be exaggerated. Shape-shifting, fissioning, disaggregation, physical mobility, reconstitution, shifts in subsistence routines are often so dizzying that the very existence of the units beloved of anthropologists—the village, the lineage, the tribe, the hamlet—are called into question. On what unit the historian, the anthropologist, or, for that matter, the administrator should fix his gaze becomes an almost metaphysical issue. The lowest-status hill peoples, it appears, are especially polymorphous. They deploy a wider range of languages and cultural practices that allow them to adapt quickly to a broad range of situations.¹²⁰ Anthony Walker, ethnographer of the Lahu Nyi (Red Lahu), writes of villages that divide up, move, evaporate altogether, scatter to other settlements, and absorb newcomers, and he writes of new settlements suddenly appearing.¹²¹ Nothing appears to remain in place long enough to sit for its portrait. The elementary unit of Red Lahu society is not the village in any meaningful sense. “A Lahu Nyi village community is essentially a group of households whose members, for the time being, find it convenient to share a common locale under a common headman more or less acceptable to them.” The headman, Walker writes, is headman of a “collection of jealously independent households.”¹²²

Here we are dealing not merely with “jellyfish” tribes but with “jellyfish” lineages, villages, chiefdoms, and, at the limit, jellyfish households. Along with shifting agriculture, this polymorphism is admirably suited to the purpose of evading incorporation in state structures. Such hill societies rarely challenge the state itself, but neither do they allow the state an easy point of entry or leverage. When threatened, they retreat, disperse, disaggregate like quicksilver—as if their motto was indeed “Divide that ye be not ruled.”