It is true, I admit, that [the Siamese Kingdom] is of greater extent than mine, but you must agree that the King of Golconda [India] rules over men, while the king of Siam only rules over forests and mosquitoes.
—King of Golconda to Siamese visitor, circa 1680
The concentration of manpower was the key to political power in premodern Southeast Asia. It was the first principle of statecraft and the mantra of virtually every history of precolonial kingdoms in the region. Creating such state space was easiest where there was a substantial expanse of flat, fertile land, watered by perennial streams and rivers and, better yet, not far from a navigable waterway. Tracing the far-reaching logic of state spaces will help distinguish the fundamental differences between manpower-poor, land-rich political systems on the one hand and land-poor, manpower-rich systems on the other.
In its crudest version, the formula goes something like this: Political and military supremacy requires superior access to concentrated manpower close at hand. Concentrated manpower, in turn, is feasible only in a setting of compact, sedentary agriculture, and such agro-ecological concentrations are possible, before the twentieth century in Southeast Asia, only with irrigated rice. These relationships are, however, not deterministic. Padi fields are easier to create and maintain in river valleys and well-watered plateaus. But they can and have been created, through prodigious feats of terracing, in steep mountainous areas where we might least expect them, such as among the Hani along the upper reaches of the Red River in Vietnam, among the Ifugao in northern Luzon, and in Bali. Similarly, there are ecological settings suitable for padi fields where they have not been developed. Nor, as we have seen, is the link between padi fields and states invariable. States are easier to create around a wet-rice core, but there are wet-rice cores without states and, occasionally, states without wet-rice cores. Irrigated rice, then, is best understood politically as the most convenient and typical means of concentrating population and foodstuffs. Without a substantial wet-rice core, such concentration would have had to be achieved by other means—by slavery, for example, or by tolls on trade routes, or by plunder.
The need to concentrate population and, at the same time, the difficulty of doing so was inscribed in the demographic given that Southeast Asia’s land mass was only one-seventh as populated as was that of China in 1600. As a consequence, in Southeast Asia control over people conferred control over land, while in China control over land increasingly conferred control over people. The abundance of arable land in Southeast Asia favored shifting cultivation, a pattern of farming that often yielded higher returns for less labor and produced a substantial surplus for the families practicing it. What constituted an advantage for the cultivators, however, was profoundly prejudicial to the ambitions of would-be state-makers. Shifting cultivation requires far more land than irrigated rice and therefore disperses population; where it prevails, it appears to “impose an upper limit of population density of about 20–30 per square kilometer.”1 Once again, concentration is the key. It matters little how wealthy a kingdom is if its potential surplus of manpower and grain is dispersed across a landscape that makes its collection difficult and costly. “Effective strength often came down to a polity’s core, not the realm’s total size or wealth,” as Richard O’Connor has put it. “Irrigated wet-rice created stronger heartlands.… It not only supported a denser population, but grain-supported villagers would have been easier to mobilize.”2 The very name of the northern Thai kingdom—Lanna, “one million padi fields”—amply reflects this fiscal and manpower obsession.
Conditions in a flourishing wet-rice heartland, then, were favorable to the development of what might be called the premodern state’s ideal subjects. That ideal is represented by densely packed cultivators of permanent grain fields who produce a considerable annual surplus. Having put considerable labor into their padi fields, over generations, perhaps, they are reluctant to pack up and leave. They and their rice fields are, above all, fixed in space, legible, taxable, conscriptable, and close at hand. For the court and its officials, the advantages are obvious.3 It is in recognition of this process of “ingathering” that Georges Condominas coined the term emboîtement (“containerization” or “bundling” might be the best translation) to describe the evolution of the Tai muang.4 For the “ideal subject,” unlike the shifting cultivator, living in a “state space,” being emboîtée, meant an additional and often unpredictable claim on his labor, his grain, and, in the case of war, his very life.
The successful premodern Southeast Asian state strove constantly to assemble the population it needed and to hold it in place. Demography was not on its side. Natural disasters, epidemics, crop failures, war, not to mention an ever-beckoning frontier, constantly threatened a tenuous state. A Chinese manual on governance, from more than a millennium earlier, when China’s demography, too, was unfavorable to state-making, put the danger starkly: “If the multitudes scatter and cannot be retained, the city-state will become a mound of ruins.”5 Archeologists working in Southeast Asia find no shortage of such mounds.
Discerning the precise balance of social and economic forces holding such agglomerations of power together and those tearing them apart is exceptionally difficult for two reasons. First, the balance was exceptionally volatile from year to year and from region to region. A war, an epidemic, a run of good harvests, a famine, the collapse of a trade route, a mad monarch, a civil war among claimants to the throne could tip the balance one way or the other. Second, we must be exceptionally wary of the written royal court records and even local chronicles that are strong on dynastic self-idealization and weak on hard information.6 To take these last at face value would credit “the king’s peace,” prosperity, religious patronage, and divine providence with the power to attract and bind a critical mass of people around the state core. Taken with a large grain of salt, this image is not entirely false. There is repeated evidence of kings and their officials enticing settlers to open padi fields by providing working capital of grain and draft animals and waiving taxes for a time. Thus a Burmese official near Pegu boasted in his 1802 revenue report that he “fed and supported those who were pleased to come from distant towns and villages in the desert[ed] places of high jungle and tall grass.”7 A peaceful and prosperous reign did, in fact, draw in migrants fleeing unsettled conditions elsewhere and hoping to farm, work, and trade near the capital. It is this depiction of a largely peaceful and gradual ingathering of hitherto stateless peoples, attracted to a luminous and thriving court center, that is the narrative conjured by dynastic histories and contemporary schoolbook idealizations of the precolonial state. As such, it is a wildly distorted narrative. It mistakes the exception as the rule; it fails utterly to explain the frequent collapse of precolonial kingdoms; it ignores, above all, the essential role that war, slavery, and coercion played in the creation and maintenance of these states. If I slight the occasions on which a Whiggish account of the flourishing dynasties might be tenable, it is because such moments are already much storied, are comparatively rare, and grossly distort the basic characteristics of state-making in mainland Southeast Asia.
If demography and an open frontier limited the effectiveness of pure coercion, it is nevertheless abundantly clear that the use of force was instrumental in creating and maintaining the “thickly-settled clumps” of people on which the state depended.8 The accumulation of population by war and slave-raiding is often seen as the origin of the social hierarchy and centralization typical of the earliest states.9 Most powerful kingdoms constantly sought to replenish and enlarge their manpower base by forcibly resettling war captives by the tens of thousands and by buying and/or kidnapping slaves. Just as the key measure of a state’s power was the manpower it could muster, so was manpower the key measure of the comparative standing of officials, aristocrats, and religious orders that competed for dependents, bondsmen, and slaves. The context of many royal decrees betrays both the effort to force the core population to stay put and, if we read between the lines, a hint of failure. If a majority of tsarist decrees of the eighteenth century concern runaway serfs, then we can safely surmise that the flight of serfs was a common problem. Similarly, the number of royal orders forbidding subjects to flee, to change residence, or to cease cultivating is a fair indication that absconding subjects were a constant preoccupation of rulers. Throughout much of the mainland, subjects were tattooed and sometime branded to indicate their status and their master. How effective such measures were is hard to tell, but they do reflect the attempt to hold the heartland population in place by force.
This overwhelming concern for obtaining and holding population at the core is shot through every aspect of precolonial statecraft. What Geertz says about Balinese political rivalries—that they were “a struggle more for men than for land”—could apply equally to all of mainland Southeast Asia.10 This principle animated the conduct of warfare, which was less a grab for distant territory than a quest for captives who could be resettled at the core. Wars were, for this reason, not particularly sanguinary. Why would one want to destroy the main prize of victory? The logic was most powerful for the inland agrarian states, which relied more on core agricultural production than on the profits of long-distance commerce. But even the raiding-and-trading states of peninsular Southeast Asia were preoccupied with seizing and holding manpower. Early European officials were frequently astounded by the extremely vague demarcations of territories and provinces in their new colonies and puzzled by an administration of manpower that had little or nothing to do with territorial jurisdiction. As the British surveyor James McCarthy “noted with puzzlement: ‘It was a particular custom [of the Siamese] in which power over individuals and land was separated.’” As Thongchai Winichakul’s insightful book shows, the Siamese paid more attention to the manpower they could summon than to sovereignty over land that had no value in the absence of labor.11
The primacy of population control was embedded in the terminology of administration. Thai officials bore titles that referred directly to the number of people they could, in theory, muster: Kun Pan meant “Lord of a Thousand Men”; Kun Saen meant “Lord of a Hundred Thousand Men,” not “Duke of such-and-such a place,” which would have been the case in Europe.12 The territorial designations that did exist in the area ruled by Bangkok in the late eighteenth century were essentially classed by their effective manpower rating. Thus provinces were ranked by the degree of power Bangkok exercised over them in declining order, fourth class corresponding to direct power and first class corresponding to weakest power (for example, Cambodia at the time). A province’s size was then calibrated by a standardized total of manpower it could plausibly be expected to muster when summoned. Distant provinces, where Bangkok’s sway was weak, tended to be both sparsely populated and much larger; the idea being that each province would yield a roughly equal number of conscripts for work or for war.13
The paramount importance of manpower rested, in the final analysis, on military considerations. Occupation of a fertile rice plain, of an important temple complex, of a choke point along vital trade routes was of little avail if it could not be successfully defended. This homely fact goes to the very heart of the analysis of power in such premodern political systems. Rather than wealth begetting power, as it might in Lockean systems, where the state’s first duty is to defend citizens’ life and property, in premodern systems only power can guarantee property and wealth. And power, before the technological revolution in warfare, was largely a matter of how many men a ruler could field; power, in other words, boiled down to manpower.
The logic of manpower operated at every level of the precolonial political systems of Southeast Asia. Princes, aristocrats, merchants, officials, village headmen, all the way down to the heads of households held their positions by virtue of the allies whose labor and support they could rely upon when challenged. The logic is captured well by Anthony Reid: “The political context made it dangerous for a small man to show his wealth unless he had sufficient dependents to defend and legitimate it.… Capital, therefore, had first to be deployed in obtaining people—through buying slaves, lending to those in need, marital and military alliances, and feasting.”14 Anyone bent on accumulating power in this context would, perforce, engage in behavior that would seem anomalous or profligate in a Lockean system. The Machiavellian strategy under these conditions is to surround yourself with the largest number of allies who are obligated to you; this requires a judicious liberality with gifts, loans, and feasting. Some allies can literally be bought. As a sixteenth-century visitor reported, the people of Melaka believed “that it is better to have slaves [‘bondsmen’ would be a better translation] than to have land, because slaves are a protection to their masters.”15
My assertion here is not so much that manpower was wealth as that it was the only means by which wealth could be securely held. In fact, one could argue, as Reid convincingly does, that maritime and overland trade were far more lucrative than squeezing the surplus out of a sedentary peasantry, even in the sixteenth and seventeenth centuries. Even the largely agrarian state of upper Burma depended quite heavily on the taxes and tolls its strategic position on the Irrawaddy allowed it to levy on precious commodities destined for markets in China, India, and beyond.16 Such goods were often easy to store, and their high value per unit weight and volume (like opium today) more than offset transportation costs. To reap the rewards of such trade, however, required that a kingdom defend its monopoly position on a river or athwart a mountain pass, or to enforce its claim to tribute by force if need be; in that case, the main currency of competition was, again, manpower.
It is this decisive advantage in manpower, Victor Lieberman argues, that over the long haul favored the hegemony of Southeast Asia’s “agrarian” kingdoms over maritime kingdoms. “In an era of limited military specialization when the number of conscripted cultivators offered the best single indication of military success, the north [of Burma] was the natural center of political gravity,” Lieberman writes. “In the central mainland and Java as well, we shall find easily cultivated dry, but irrigable, areas enjoyed an early demographic advantage over wetter, maritime districts.”17 Viewed in a crude synoptic fashion, over time a handful of larger maritime powers (Srivijaya, Pegu, Melaka) eclipsed their smaller maritime rivals; they were in turn eclipsed by more manpower-heavy agrarian states (Mataram, Ayutthaya, Ava), which had likewise eclipsed their smaller agrarian rivals (Vientiane, Lan Na, Chiang Mai). Everything we know about statecraft in Ava and Ayutthaya suggests a constant effort, by no means always successful, to hold a dense population at the core and augment it when possible.18
The process described above accords nicely with much of the literature on European state-making and political consolidation. Here, too, what Charles Tilly aptly called the “coercion-rich and capital-poor,” “landward” agrarian states and empires (for example, Russia, Brandenburg Prussia, Hungary, Poland, and France) enjoyed an advantage in manpower, usually a decisive one, over their maritime rivals (Venice, the Netherlands, Genoa, Florence). Less reliant on volatile trade, more hierarchical, more insulated from food-supply crises, and capable of feeding quite massive armies, these agrarian states might lose a battle or even a war, but their staying power over the long haul tended to prevail.19
Population as the principal measure of statecraft finds abundant expression in the sayings and admonitions that infuse the courtly literature of Southeast Asia. Nowhere is the relative weight of manpower vis-à-vis territory more evident than in this epigram from the Early Bangkok Period in Siam: “To have too many people [as subjects to a lord] is better than to have too much grass [uncultivated land].”20 The epigram is echoed, almost precisely, by one from the Burmese Glass Palace Chronicle, compiled at about the same time: “Yes, a soil, but no people. A soil without people is but a wilderness.”21
Two additional Siamese sayings emphasize that wise rule requires both preventing people from fleeing the heartland and attracting new settlers to cultivate the land:
In a large house with many servants, the door may safely be left open; in a small house with few servants, the doors must be shut.
The governor should appoint loyal officials to go out and persuade them to come and settle down in an inhabited area so that the area will be wealthy.22
The collapse of a kingdom was, in turn, seen as a failure of the monarch to husband his population wisely. Queen Saw’s admonition to the Burmese King Narahihapate is a dramatic illustration: “Consider the state of the realm. Thou hast no folk or people, no host of countrymen and countrywomen around thee.… Thy countrymen and countrywomen tarry and will not enter thy kingdom. They fear thy dominion, for thou, oh King Alaung art a hard master. Therefore I, thy servant, spoke to thee of old but thou wouldst not hearken.… I said bore not thy country’s belly, abase not thy country’s forehead.” Warfare as a contest for control over cultivators rather than arable land is also evident in this praise for a Siamese military commander who not only put down rebellion but delivered his captives to the crown: “From that day forward, he sent Anantathuriya, whenever there were thieves, cutthroats, rioters, or rebels in border places and purlieux. And wheresoever he went, he caught numbers of his foes alive and brought them back to the king.”23 Even in the absence of such explicit statements, the centrality of manpower is everywhere evident in the constant emphasis on what might be called “entourage politics.” It is common, whenever an official is mentioned in the court histories, to list the size and distinction of his followers.24 Whenever a victorious military campaign is reported, it is generally the number of surviving captives rounded up and marched back to the capital that receives most attention. Although I have concentrated here on the evidence from the mainland, the same preoccupation with manpower is, if anything, more striking in peninsular Southeast Asia and the Malay world in particular.25
The imperative of concentrating population and grain production, in fact, confronts all would-be state-makers who must operate in an environment where open land is abundant and military technology simple. Some means must be devised to counteract the tendency of the populace to disperse widely so as to take full advantage of the hunting, foraging, and less labor-intensive farming techniques available to them. A range of incentives—from commercial exchange to reliable irrigation to participation in military plunder to the desire for sacred knowledge—might be at work. Such advantages, however, had to outweigh the burdens of taxation, conscription, and epidemics always associated with state space if the concentration was to be achieved without resistance. They seldom did. The use of force to supplement these advantages, and often to replace them altogether, was ubiquitous.
Here it is worth recalling that the political systems of classical antiquity in the West were, manifestly, coercive systems of this kind. Athens and Sparta, Thucydides tells us, fought not over ideology or ethnicity but over tribute. That tribute was measured in grain and, above all, in manpower. The populace of a surrendered town was rarely butchered; rather, its citizens and slaves were taken into captivity by the victors and by individual soldiers who had captured them. If their fields and homes were burned, it was largely to prevent them from returning.26 The major tradable commodity on the Aegean—more valuable than grain, olive oil, or wine—was slaves. Athens and Sparta were both slave societies, although in Sparta, which was more agrarian, helots accounted for more than 80 percent of the population. In imperial Rome as well, the most important commodity transported along its fabled road system was slaves; they were bought and sold under government monopoly.
China and India, well before they became so populated that control of arable land alone ensured control of their land-starved subjects, faced similar problems of statecraft. At about the same time as the Peloponnesian War, the early Chinese state was doing everything in its power to prevent the dispersal of population. Manuals of statecraft urged the king to prohibit subsistence activities in the mountains and wetlands “in order to increase the involvement of the people in the production of grain.”27 The subtext of this and other pronouncements was that, given a choice in the matter, the king’s subjects would abandon sedentary agriculture and strike out on their own. Such resistance was seen as a moral failing. If the state has “the unique power over the mountains and marshes, then the common people who detest farming, are lazy, and want doubled profits, will have nowhere to find something to eat. If they have nowhere to find something to eat, they will be obliged to engage in the cultivation of the fields.”28 The objective of this policy was, it seems, to starve the population into grain farming and subjecthood by separating them from the open commons. Somehow, the shrill tone of the advice suggests that the policy was not a complete success.
The dilemma for statecraft in settings of low population densities finds a more contemporary and instructive parallel in Africa south of the Sahara. In 1900 the population density there was not much greater than that of Southeast Asia in 1800, and, as a consequence, the problem of bundling a population at the state core was the crux of precolonial politics.29 The theme of manpower concentration permeates the literature on indigenous politics: “The drive to acquire relatives, adherents, dependents, retainers, and subjects and to keep them attached to oneself as a kind of social and political ‘capital’ has often been remarked upon as characteristic of African political processes.”30 The similarities are so striking that many of the adages of rule can be transposed to Southeast Asia with no loss in intelligibility. As a Sherbro proverb has it, “One cannot be a chief and sit alone.” The link between cleared, permanent fields and the foundation of a kingdom was also invoked in this advice to an ancient Malian king: “Cut the trees, transform the forests into fields, for then only will you become a true king.”31 As in Southeast Asia there was little emphasis on sharp territorial boundaries, and the important rights were rights over people, not over places, except for particular ritual sites. The competition for followers, kinsmen, and bondsmen operated at every level. Given demography so favorable to potential subjects, they had more often to be enticed rather than coerced to settle under a ruler. The relative autonomy of subjects found expression in the proliferation of titles, in feasting, in rapid assimilation and mobility for captives and slaves, in special paraphernalia and medicines to bind retainers, and above all, in the flight of unhappy subjects. This balance of power, according to Igor Kopytoff, gave subjects the distinct sense they it was they who had created the ruler and not the other way around.32
Taxes ate the valleys, honor ate the hills.
—Afghan proverb
A premodern ruler in mainland Southeast Asia would have been less interested in what today would be called the gross domestic product (GDP) of his kingdom than what we might call its “state-accessible product” (SAP!). In a premonetary setting, products that come from any considerable distance would have to be quite valuable per unit weight and volume to justify the transportation costs. Such products existed—for example, aromatic woods, tree resins, silver and gold, ceremonial drums, rare medicines. The greater the distance they traveled, the greater the likelihood they were part of a gift or voluntary trade, for the court’s capacity to compel such goods diminished more or less geometrically with distance. What mattered most were the food, livestock, and manpower—including skilled manpower—that could be conveniently seized and put to use. The state-accessible product had to be easy to identify, monitor, and enumerate (in short, assessable), as well as being close enough geographically.
State-accessible product and gross domestic product are not simply different; they are, in many respects, at odds with each other. Successful state-building is directed toward the maximization of the state-accessible product. It profits the ruler not at all if his nominal subjects flourish, say, by foraging, hunting, or shifting agriculture at too great a distance from the court. It similarly profits the ruler little if his subjects grow a diverse suite of crops of different maturation or crops that spoil quickly and are therefore hard to assess, collect, and store. Given a choice between patterns of subsistence that are relatively unfavorable to the cultivator but which yield a greater return in manpower or grain to the state and those patterns that benefit the cultivator but deprive the state, the ruler will choose the former every time. The ruler, then, maximizes the state-accessible product, if necessary, at the expense of the overall wealth of the realm and its subjects. So it is that the premodern state attempts to so arrange its subjects and to sculpt the landscape around it in order to make it a legible field of appropriation. When successful, the result in mainland Southeast Asia has been the social creation of a uniform agro-ecological landscape based on irrigated wet rice: what Richard O’Connor has called the “paddy-state.”33
The chief advantage of padi rice is that it makes possible a concentrated density of both population and of grain. Here it is worth emphasizing the way irrigated rice fixes people in space. No other cultivar could have placed so many people within a three- or four-day march of the court center. The superior productivity of wet rice per unit of land permits enormous population densities, and the relative permanence and reliability of padi rice, so long as the irrigation system is functioning well, helps ensure that the population itself will remain in place. Each padi field—representing as it does many years of “sunk” labor costs in bunding, leveling, terracing, weir and channel construction—is not lightly abandoned. “One major problem” of the Kon-baung kings, writes Thant Myint U, “was the difficulty of the central state in gaining accurate information on the number of households in a particular locality.”34 This might be called the problem of “legibility,” which was the indispensable condition of making resources accessible.35 Compared with subsistence patterns that favored dispersal and autonomy, the social ecology of wet rice greatly simplified this problem by placing a relatively stable and dense population at the doorstep of the tax man and the military press-gang.
Fixed-point production by a sedentary peasantry in the padi state meant that the ruler and his entourage of specialists and officials could also remain in one place. Without a reliable, accessible surplus of food, fodder, and firewood, the court would have had to shift to another site, just as the English and French courts of the thirteenth century did, once they had exhausted the food supply (and forbearance!) in any one district. The size of the noncultivating elite was, of course, constrained by the size of the grain surplus; the larger the core, the more numerous and well provisioned the entourage. Only padi cultivation on a substantial scale gave an agrarian state a sporting chance at persisting.
The cultivation of a single staple grain was, in itself, an important step in legibility and, hence, appropriation. Monoculture fosters uniformity at many different levels. In the case of irrigated rice, cultivators were bound to roughly the same rhythm of production. They depended on the same, or comparable, sources of water; they planted and transplanted, weeded, cut, and threshed their crop at roughly the same time and in roughly the same way. For the maker of a cadastral survey, a tax map, the situation was nearly ideal. Most land values could be calibrated to a single metric; each harvest both was compressed in time and involved a single commodity; the mapping of open fields demarcated by bunds was relatively straightforward, although matching the land to the appropriate taxpayer was not quite so straightforward. The uniformity in the field, in turn, produced a social and cultural uniformity expressed in family structure, the value of child labor and fertility, diet, building styles, agricultural ritual, and market exchange. A society shaped powerfully by monoculture was easier to monitor, assess, and tax than one shaped by agricultural diversity. Imagine, once again as an Asian Colbert, organizing a tax system for a diverse polyculture of, say, several grains, fruits, nuts, root crops, livestock, fishing, hunting, and foraging. Such diversity would give rise, at a minimum, to different land values, family arrangements, work cycles, diets, domestic architecture, dress, tools, and markets. The existence of so many products and “harvests” would, by itself, make far more intractable the creation of any tax system, let alone an equitable one. I have, for the purposes of analytical clarity, drawn the comparison too sharply; none of the mainland agrarian states was a pure monoculture. But to the degree that it approximated one, it radically simplified the consolidation of a manageable state space.
It is in this context that the strenuous efforts of successful Burmese dynasties to maintain and extend the irrigated riceland within the dry zone should be understood. Outside these padi cores lay a less productive and more diverse agricultural landscape that posed difficulties for the tax man. The district revenue reports (sit-tàns) invariably list the padi land of the district first and make it clear that the revenue from nonpadi lands—millet, sesamum, cattle, fishing, coconut palms, and handicrafts—was both more difficult to collect and, compared with padi income, negligible.36 Collecting revenue from a population that was more dispersed, that was generally poorer, and whose subsistence routines were much more varied was singularly unrewarding. What was collected, moreover, was more easily concealed from crown officials and monopolized by local strongmen. The colonial regime in Burma was no less dependent on irrigated ricelands, even when the tax on them was collected in cash. John Furnivall sees it as the “staple” of the colonial fiscal diet: “What rice is to the mild Hindoo, and to the anything but mild Burman, what macaroni is to the Italian, beef and beer to the Englishman, all that and more than that, is land revenue to Leviathan Indicus, the species of Leviathan that inhabits India; it is his victuals, his sustenance. Income tax, customs duties, excise receipts and so on … he could in a pinch do without them, but without land revenue he would starve to death.”37
Here again the distinction between gross domestic product and state-accessible product is at work. As a general rule, the agriculture organized by and for states and enterprises with appropriation, above all, in mind, is likely to bear the marks of legibility and monocropping. Monoculture plantations, the now defunct collective farms of the socialist bloc, cotton share-cropping in the postbellum U.S. South, not to mention the coercive agricultural landscapes created by counterinsurgency campaigns in Vietnam or Malaya, are cases in point. They are rarely models of efficient or sustainable agriculture, but they are, and they are intended to be, models of legibility and appropriation.38
The policy of encouraging or imposing legible, agrarian landscapes of appropriation seems hard-wired to state-making. It was only such landscapes that were directly beneficial and accessible. Little wonder, then, that the efforts to sedentarize populations through fixed (usually rice) cultivation represents a striking continuity between the precolonial states and their contemporary descendants. The Vietnamese Emperor Minh Mang (1820-41) “used every method available to encourage the cultivation of new fields of rice. These included permitting a man who cleared land himself to use it as his own private field, or encouraging rich people to come forward and recruit tenants to set up new village settlements. The state’s principal aim was the maintenance of control over the population. Vagrancy was discouraged and displaced people were fixed to a plot of land where they could be turned into reliable sources of taxation, corvée labour and military service.”39
For its part, the French colonial regime was interested in turning open land into revenue-yielding, legible crops, in particular, rubber grown on plantations. The French desired to transform the fiscally sterile hills into a space that would be rentable and utile. Socialist Vietnam, up to the present, remains devoted to “fixed cultivation and fixed settlement” (dinh canh dinh cu), with the emphasis returning to wet rice, even in places where it is ecologically unsound. An older vision of the padi state has been married to a utopian view of the conquest of nature by heroic socialist labor. It looks lyrically forward to “a tomorrow [in which] Tay Bac’s forested hills and grassy expanses will be flattened and immense fields of rice, fields of corn will be opened up.” As another brave slogan had it, “With the strength of the people, even stones turn into rice.”40 One aspect of this vast policy of resettlement has been the understandable desire of the lowland Kinh to reproduce the agricultural landscape and human settlements with which they are familiar. As is so often the case, the attempt by migrants to apply techniques of cultivation that are utterly ill-suited to their new setting has resulted in ecological damage and human suffering. The other aspect of this utopian aspiration, however, has been the attempt by the Vietnamese state to re-create the landscapes of legibility and appropriation that had sustained its precolonial ancestors since at least the Le Dynasty.
Hostile nature, obstinate and fundamentally rebellious, is in fact represented in the colonies by the bush, by mosquitoes, natives and fever, and colonization is a success when all this indocile nature has finally been tamed.
—Franz Fanon, The Wretched of the Earth
My only quarrel with Franz Fanon’s acid insight into the colonial project is that his observation, at least with respect to “the bush” and “natives,” might as easily be applied to the precolonial and postcolonial eras.
The expansion and peopling of legible state space was intrinsically difficult, given the open frontier. If it was occasionally achieved, it was due as much to the closing off of alternatives as to the inherent attractions of state space. The major alternative to irrigated padi cultivation in mainland Southeast Asia, historically, and even today in much of the region, is shifting agriculture (also known as swidden or slash-and-burn agriculture). Inasmuch as it involves population dispersal, mixed cropping (including roots and tubers), and periodic opening of new fields, swiddening has been anathema to all state-makers, traditional or modern.
As by far the most precocious state in the region, the Chinese state since at least the Tang Dynasty has been stigmatizing shifting cultivation and eradicating it when it could. Though shifting-cultivation agriculture might provide a higher return to the cultivator’s labor, this was a form of wealth that was inaccessible to the state. And, especially if it was advantageous to the cultivator, it represented, at the fringes of an often heavily taxed, rice-growing peasantry, a constant temptation as an alternative subsistence. Along China’s southwestern frontier, shifting cultivators were encouraged, and sometimes forced, to abandon their shifting agriculture for sedentary grain production. The Chinese seventeenth-century euphemism for incorporation into state space was “to enter the map.” It meant becoming a subject of the emperor, proclaiming loyalty, and setting out on a cultural journey that would, in Han eyes, eventually lead to assimilation. Above all, however, the passage from nomadic agriculture to sedentary grain growing meant becoming a registered household that now figured in the official tax rolls.41
The state fiscal imperatives that lay behind the desire of Vietnam’s Emperor Minh Mang or China’s officials to eradicate shifting cultivation have been reinforced in the modern era by two further considerations: political security and resource control. Because shifting cultivators were not incorporated into the state administration, because they spilled promiscuously across national boundaries, and because they were seen as ethnically distinct, they were seen as potentially subversive. In Vietnam this has led to vast campaigns of forced resettlement and sedentarization. Another contemporary reason given for prohibiting shifting cultivation is that it is environmentally unsound, destroying soil cover, promoting erosion, and wasting valuable timber resources. To some considerable degree this rationale was a direct policy inheritance from the colonial period. Its premise, we now understand, is wrong, except under special circumstances. The overriding reason behind such policies, it appears, has been the state’s need to use such land for permanent settlement, to realize for itself the revenue from the extraction of natural resources, and to bring such nonstate peoples finally to heel. As one government ethnographer informed a foreign colleague, the purpose of his study of the hill economy “was to see how ‘nomadic’ slash-and-burn agriculture can be eradicated among the minorities.”42 The “campaign to sedentarize the Nomads,” begun in 1954, has, in one guise or another, remained a steadfast policy.
Much the same continuity in policy, if not in consistent enforcement, has typified the Thai state for much of its existence. Nicholas Tapp, an ethnographer of the Hmong, claims that the policies of sedentarization, permanent agriculture, political control, and “Thaiization” “represent highly conservative strategies which have characterized relations between state populations and upland developing minorities of the region for several centuries.”43 The attempts to stop swiddening became significantly more brutal at the height of the cold war in the 1960s, after a Hmong rebellion was crushed by General Prapas by means of artillery, military assaults, and napalm. Despite the fact that Vietnam and Thailand feared subversion from diametrically opposite points of the ideological compass, their policies were remarkably similar. The Hmong were to stop shifting cultivation and, as a policy document noted, officials were to “persuade the hill tribes, living scatterdly [sic] to move into project areas and settle down permanently.”44 State space had, under the circumstances, acquired additional meaning, but the new meaning only strengthened the reasons to eradicate shifting agriculture.45
Perhaps the most long-standing violent campaign against shifting cultivation, designed to force the people who practice it into what amount to concentration camps around military bases or, failing that, to force them over the border into Thailand, is the Burmese military regime’s campaign against the Karen. Armed columns are sent out before swidden fields are harvested to burn the grain or to beat it down and to lay land mines in the fields. Knowing how crucial a successful “burn” is to the swidden harvest, the army also sends units to burn the slash prematurely so as to ruin the chances for a good crop. By eradicating shifting cultivation and not a few of its practitioners, they minimize the chances for survival outside state spaces.46
Such coincidence of policy across several centuries and, in the modern period, across very different types of regime is prima facie evidence that something fundamental about state-making is at work.
Whatever concentration of population around the court the padi state managed to achieve was a hard-won victory against considerable demographic odds. Statecraft with both eyes fixed on the accumulation of manpower could hardly be particular about whom it incorporated. A “manpower state” in this sense is, in principle, the enemy of hard and fast cultural distinctions and exclusiveness. Put more accurately, such states had great incentives to incorporate whomever they could and to invent cultural, ethnic, and religious formulas that would allow them to do so. This fact, true of all the padi states of mainland and maritime Southeast Asia, is rich in consequences for each of these lowland civilizations.
The emphasis on inclusion and absorption was such that it would surely be mistaken to see the classical Burman or Thai state as endogenous, monoethnic expressions of cultural development. It is closer to the mark to see each such state core as a social and political invention, an alloy, an amalgam that bears traces of ingathering from many diverse sources. The culture of the center was a provisional work in progress, a kind of contingent vector sum of the various peoples and cultures who chose to identify with it or who were incorporated by force. Many of the formulas of incorporation were, one might say, “on loan” from the Indian subcontinent, in the form of Shaivite cults, Brahminical rituals, Hindu court rituals, and Buddhism—first Mahayana and then Theravada. Their value, as Oliver Wolters and others have suggested, lay in the fact that they both reinforced the claims to supernatural powers and legitimacy of local power holders and provided a universalizing framework for creating a new state-based identity from many ethnic and linguistic shards.47
If this explicitly political perspective has any merit, its effect is to radically decenter any essentialist understanding of “Burmanness” or “Siameseness” or, for that matter, “Hanness.”48 Identity at the core was a political project designed to weld together the diverse peoples assembled there. Bondsmen of allied strongmen, slaves captured in warfare or raids, cultivators and merchants enticed by agricultural and commercial possibilities: they were in every case a polyglot population. The premium on incorporation meant that assimilation, intermarriage, and social mobility across permeable social barriers were relatively easy. Identity was a matter more of performance than of genealogy.49 Each of the numerous padi states that came and went in the classical period represented something of a “career open to talent.” The culture each of them codified over time varied with the largely imported cultural and human material it had to work with. If there was a cultural attractiveness to the precolonial court centers, it was surely this capacity to absorb migrants and captives and, in two or three generations, to fashion them and their practices into an encompassing Burman or Thai cultural amalgam. A brief look at this process of amalgamation in the Thai padi state, in the Malay zone, and in classical Burma will sharpen our appreciation of the hybridity of the manpower state.50
The central plain of what would become Siam was, in the thirteenth century, a complex mix of Mon, Khmer, and Tai populations who were an “ethnicity-in-the-process-of-becoming” Siamese.51 Victor Lieberman claims that by the mid-fifteenth century in the Ayutthaya Period, a distinctive “Siamese” culture had emerged among an administrative elite (munnai)—and, it appears, only among them. Although their courtly culture drew on Khmer and Pali texts, the commoners, when the Portuguese Tome Pires wrote about them in 1545, were speakers of Mon dialects, not Tai, and cut their hair like the Mons of Pegu. The ingathering practices of the manpower state were very much in evidence at the end of the seventeenth century, when it is claimed that more than a third of the people in central Siam were “‘foreigners’ descended chiefly from Lao and Mon captives.”52 Again in the early nineteenth century, the court redoubled its efforts to make good the massive losses of population during the Burmese wars. As a result, “All told, Laos, Mons, Khmers, Burmese and Malays may have equaled the number of self-identified Siamese in the central basin. Phuan, Lao, Cham, and Khmer peasant units formed the backbone of the standing army and navy around Bangkok. On the Khorat plateau after Anuvong’s revolt of 1827 so many Lao deportees were resettled that they may have been as numerous as Siamese speakers within the Kingdom as a whole.”53
What was true for the Chao Praya basin was also true for the veritable archipelago of smaller Tai/Shan padi states scattered here and there farther north among the hills. The consensus is that the Tai/Shan statelets were a politicomilitary invention—Condominas’s système à emboîtement—in which the Tai were rather thin on the ground. This view matches the evidence that the Burmans were, themselves, also thin on the ground and that they constituted a pioneer, military elite with experience and skills at state-building. That the conquering lords should have been few in number and yet ultimately hegemonic should not surprise those familiar with British history, inasmuch as the conquering Norman elite that came after 1066 to dominate Britain consisted of a mere two thousand families.54 The Tai/Shan conquerors grew by virtue of a talent for allying, absorbing, adapting, and syncretizing a confected kingdom out of the peoples available to them. The process involved incorporating remnants of preexisting political systems (Mon, Lawa, Khmer) and, above all, absorbing large numbers of upland peoples. Condominas argues that captured hill peoples might start out as bondsmen but, over time, became Tai commoners entitled to hold padi land. Those who were lucky or skilled enough to seize the muang itself would be expected to adopt a Tai noble name, thus bringing their genealogy retrospectively in line with their personal achievement.55 A majority of the population of most such states was composed of non-Tai peoples, and even many of those who had become Tai and Buddhist continued to use their own languages and customs.56 Although today it is customary to believe that many Kachin are becoming Shan, the Kachin graduate student who attempted to show that most Shan were at one time Kachin was probably not far off the mark.57 Shan society, Edmund Leach believed, was not so much a “‘ready-made’ culture sweeping down from southwest China, as an indigenous growth resulting from an economic interaction of small-scale military colonies with an indigenous hill population over a long period.” He adds: “There are various other kinds of evidence which support the view that large sections of the people we now know as Shans are descendants of hill tribesmen who have in the recent past been assimilated into the more sophisticated ways of Buddhist-Shan culture.”58 Similarly designed, but on a far smaller scale, these padi states were ethnically plural, economically open, and culturally assimilationist. Shan identity is, in every case, tied to padi cultivation and, in turn, to being a subject of a Shan state.59 Through the medium of wet-rice cultivation, “Shanness” and stateness are firmly linked. It is wet-rice cultivation that ensures a fixed, sedentary population which is the basis for military superiority, for an accessible surplus, and for a political hierarchy.60 Shifting cultivation, by contrast, implies non-Shan identity and, virtually by definition, living at a distance from the state.61
The Burman states which have arisen since the eleventh century in upper Burma have almost defined the contours of the classical agrarian, manpower state. Their agro-ecological settling was (along with the Red River in Vietnam) perhaps the most favorable location for the concentration of manpower and grain production. The state-building core, which each dynasty had to control, consisted of six districts, four of which (Kyaukse, Minbu, Shwebo, and Mandalay) had perennial streams allowing extensive, year-round irrigation. Kyaukse, whose very name implies wet-rice cultivation, was the richest of these districts. As early as the twelfth century, there were areas where three crops could be grown annually.62 By the eleventh century, Lieberman estimates, there were several hundred thousand people within an eighty- to one hundred-mile radius of the court.63
Like the Tai kingdoms, Pagan was a political mechanism for the accumulation of manpower and grain production. As such, it welcomed, or seized, settlers wherever they could be found and tied them to the court as subjects. Mid-thirteenth-century Pagan, inscriptions suggest, was an ethnic mosaic including, in addition to Mons, Burmans, Kadus, Sgaws, Kanyans, Palaungs, Was, and Shans.64 Some were there for the opportunities that a growing empire provided and, for some, perhaps, the move represented a “voluntary assimilation of bilingual peoples eager to identify with the imperial elite.”65 There was, however, little doubt that a considerable portion of the population, particularly the Mon, was the “prize” of raiding, war, and forced resettlement.
Holding together a state core of this magnitude, given the demography, was a tenuous enterprise. The temptations of a land frontier coupled with the burdens of life in state space (taxes, conscription, bondage) meant that the inevitable leakage had constantly to be replenished by military campaigns for captives and forced migration to the center. If the center, once established, managed to hold its own demographically until the mid-thirteenth century, the exodus after that—perhaps because the rice plain offered such a concentration of booty to the Mongol invaders—became a hemorrhage and the empire collapsed.
The last (Kon-baung) dynasty before British rule was, like the earlier dynasties, a manpower-obsessed state. It was understood by its rulers to be a polyglot kingdom in which an oath of loyalty and payment of tribute signaled incorporation. Like the Burmans, the Mon, Siamese, Shan, Laos, Palaungs, and Pa Os were Theravada Buddhists. But judging from the Muslim and Christian communities with their own quarters, mosques, and churches, religious conformity was not a condition of political affiliation. Estimates of what proportion of the early Kon-baung (late-eighteenth-century) population consisted of captives and their descendants necessarily involve a great deal of guesswork. Nevertheless, it seems that they represented between 15 and 25 percent of the realm’s roughly two million subjects.66 The bondsmen, as one might expect, were heavily concentrated near the court center and organized into royal service corps responsible for, say, boat building, weaving, infantry duty, arms making, cavalry, artillery. As ahmudan (literally, “task-carriers”), they were distinguished from the athi, or commoners, on the one hand, and from the personal bondsmen of private individuals on the other. In the immediate environs of the court, the royal service population (many of them Manipuri) amounted to at least a quarter of the population.
The undifferentiated term manpower doesn’t begin to do justice to the discriminating accumulation of captives and “guests” selected with an eye to their usefulness. To cite the most celebrated instance, Hsinbyushin, after sacking Ayutthaya in 1767, brought back as many as thirty thousand captives, including officials, playwrights, artisans, dancers, and actors, much of the royal family, and many of the court’s literati. The result was not just a renaissance in Burmese art and literature but the creation of a new hybrid court culture. The court entourage included a cosmopolitan collection of valuable specialists: surveyors, gun founders, architects, traders, shipwrights, and accountants, as well as drillmasters from Europe, China, India, the Arab world, and the rest of Southeast Asia. When it came to skilled manpower as well as foot soldiers and cultivators, the need for their services precluded rigid cultural exclusion.
The combination of an elaborate status hierarchy and rapid mobility and assimilation was as characteristic of Kon-baung Burma as it was of early Bangkok Siam. Most Burmans were at some point in their recent ancestry hyphenated Burmans of Shan, Mon, Thai, Manipuri, Arakanese, Karen, or other hill-group backgrounds. If one goes back far enough, one could make a credible argument that many, if not most, Burmans were a cultural amalgam of the earlier Pyu-Burman encounter. Slaves, both debt-bondsmen and captives, were, as elsewhere in Southeast Asia, likely over time to become commoners. Father Sangermano, who lived for more than twenty-five years in Ava and Rangoon at the turn of the nineteenth century, reported the great elaboration of Burmese legal codes covering different forms of bondage. He observed, however, that “this slavery is never perpetual.”67
Perennial manpower concerns favored easy assimilation and rapid mobility and, in turn, made for very fluid, permeable ethnic boundaries. Lieberman makes an utterly convincing case that what is often seen as a Burmanversus-Mon war between Ava and Pegu was no such thing. In bilingual areas of lower Burma, ethnic identity was more a political choice than a genealogical given. A change of dress, hairstyle, and perhaps residence, and, voilà!—one’s ethnic identity had also shifted. Ironically, the force the Burmese court at Ava sent against Pegu had more Mons than Burmans and those sent later (1752) by the Peguans against Alaungpaya were largely Burman. The Ava-Pegu war is thus best understood as a regional conflict in which loyalty to the kingdom trumped all other considerations and in which, in any case, identity was relatively negotiable.68
In each of the three kingdoms we have examined, considerations of religion, language, and ethnicity did play a role in stratification within the political system. What is crucial for our purposes, however, is that they were no barrier to membership in the political system. Each of these criteria was, furthermore, subject to transformation rather rapidly and invariably over two generations. The manpower imperative was, everywhere, the enemy of discrimination and exclusion.69
The precolonial state we have been examining is, in a sense, a special, limiting case of state-making in challenging demographic and technological conditions. If a state was to arise at all, its rulers would have to concentrate its subjects within a relatively narrow geographical area. Such principles of state space—namely, legibility and appropriateness—are at work in virtually all projects of rule, whether by states or by nonstate institutions. Plantation agriculture with its monoculture and workers’ barracks and the mission station with its belltower and the congregants settled in its shadow are distinct forms of control, but each requires legibility and monitoring. It is a rare development project that does not also reshape the landscape and the patterns of residence and of production to achieve a greater degree of legibility and control. The early colonial regimes, in their pacification campaigns, used forced settlement, the destruction of swiddens, and the concentration of subjects. It was only gradually that all-weather roads, railroads, telegraph lines, and a reliable currency allowed a greater dispersal of population and production with little loss of control. Only in counterinsurgency strategies do we see, in miniature, the attempt to closely concentrate a feared population in legible space, occasionally to the point where it comes to resemble an actual concentration camp.
Without slavery, no Greek state, no Greek art and science; without slavery, no Roman Empire.… We should never forget that our economic, political and intellectual development presupposes a state of things in which slavery was as necessary as it was universally recognized. In this sense we are entitled to say: without the slavery of antiquity, no modern socialism.
—Karl Marx
Where did the people living in “state space” in the precolonial era come from? Earlier theories, increasingly discredited, held that massive numbers of Tai and Burmans came from the north to displace earlier populations. Instead, it appears that rather modest numbers of Tai and Burmese established their political hegemony over the wet-rice zones that suited them.70 These padi states undoubtedly absorbed preexisting populations such as the Pyu and Mon as well as, in times of peaceful expansion, attracting immigrants seeking position, work, and trading opportunities. What is most striking, however, is that none of these padi states flourished except by slave-raiding on a substantial scale. Formulaically, and paraphrasing an observation by Karl Marx about slavery and civilization, there was no state without concentrated manpower; there was no concentration of manpower without slavery; hence all such states, including especially the maritime states, were slaving states.
Slaves, it is fair to say, were the most important “cash crop” of precolonial Southeast Asia: the most sought-after commodity in the region’s commerce. Virtually every large trader was, simultaneously, a slave-raider or a buyer. Every military campaign, every punitive expedition was, at the same time, a campaign for captives who could be bought, sold, or held. So familiar was the pattern that when Magellan was murdered on his second voyage, the Filipinos responsible rounded up what was left of his crew and sold them off in the islands. When the Burmese captured the port city of Syriam from Portuguese adventurers in the early seventeenth century, they grabbed the surviving Europeans and forcibly resettled them in villages near the capital, Ava. Southeast Asian kingdoms were remarkably broad-minded when it came to the acquisition of manpower.
Only by scouring its periphery was a growing padi state able to achieve the concentration of population required to dominate and defend its core. The process of scouring was a pan-Southeast Asian phenomenon with systematic characteristics. Anthony Reid, author of the most important analysis of slavery, explains the pattern: “Before indentured labour was developed in the nineteenth century, the movement of captive peoples and slaves was the primary source of labour mobility in Southeast Asia. Typically it took the form of transferring people from weak, politically fragmented societies to stronger and wealthier ones. The oldest, and demographically most important, form of movement was the border raiding against animist swidden cultivators and hunter-gatherers by stronger wet-rice cultivators of the river valleys.”71 Another way of describing the process is the systematic removal of captives from nonstate spaces, particularly the hills, in order to deposit them at, or nearby, state spaces. The pattern could be discerned in Cambodia in 1300 and continued in some areas (for example, Malaysia) well into the twentieth century. Gibson claims that until roughly 1920, the majority of the urban Southeast Asian population were either captives or their descendants (often in the past two or three generations).72
The evidence is pervasive. In the Tai world, by way of illustration, fully three-quarters of the kingdom of Chiang Mai’s population in the late nineteenth century consisted of war captives. Chiang Saen (Kaing Hsen), another Tai statelet, had nearly 60 percent slaves; in Lamphun seventeen thousand of a total of thirty thousand subjects were slaves. Rural elites also held slaves as part of their labor force and entourage. Such slaves were captured directly in war or were purchased from slave-raiding parties that combed the hills kidnapping whomever they could.73 To read any of the court histories or chronicles of Tai or Burman kingdoms is to be treated to a long series of accounts of raids whose success is typically measured by the number and skills of the captives. Rebellion, or failure to send the appropriate tribute, was frequently punished by the sacking and burning of the disobedient district and the deportation of its subjects to the victor’s court center. When the ruler of Songkhla, after first refusing, finally came to Ayutthaya to present tribute, the king arranged for all of Songkhla’s inhabitants to be carried off into slavery nearer the capital. The magnitude of slaving is less obscure than many other subjects to historians precisely because the taking of captives was the public purpose of statecraft.
There were, of course, many other ways than by capture to become a slave in these political systems. Debt bondage was common, in which the debtor and/or members of his family became “slaves” of the creditor until the debt was acquitted. Children were sold into bondage by their parents, and convicted criminals were condemned to slavery as punishment. If, however, such mechanisms constituted the major social origin of slavery, one would expect that most slaves would be culturally similar to their masters, except for the formal status difference. But this was not the case, as Katherine Bowie shows for northern Thailand. The majority of slaves, here and elsewhere, appear to come from culturally distinct hill populations and to have been taken in slave raids as prizes of war.74
The scale of slaving and its effects are hard to imagine.75 Slaving expeditions were a regular, dry-season commercial venture in much of the mainland. Between freebooting expeditions, small-scale kidnapping, and the larger-scale deportations (for example, the six thousand families removed forcibly to Thailand after the Siamese capture of Vientiane in 1826), whole regions were largely stripped of their inhabitants. Bowie quotes the late-nineteenth-century observer A. C. Colquhoun, who captures something of the extent and human impact:
There is little doubt that the sparsity of hill tribes in the hills neighboring Zimme [Chiang Mai] has been chiefly caused by their having been, in the olden time, systematically hunted like wild cattle, to supply the slave market.…
The slaves who are captured become slaves in the fullest sense of the word; they are carried off with no hope of deliverance save death and escape. Trapped by ambush, and driven off after capture, like fallow-deer, by the man-hunters, they are torn from their forests, chained, and taken to the chief places of the Shan country [Chiang Mai], Siam, and Cambodia for disposal.76
As different agro-ecological zones, the hills and lowland valleys were natural trading partners. Alas, however, the most important hill commodity for expansive valley states was its manpower.77 This kind of manpower hunting and gathering was so lucrative that hill people and not a few hill societies as a whole became deeply implicated in the trade. Alongside war captives and punitive resettlement, lowland populations were augmented by what were essentially commercial slaving expeditions. Societies in the hills could often be classified into weak, fragmented societies that served as the sources of slave raids, the prey, and highland groups that organized slave raids and frequently held slaves themselves, the predators. The Akha, the Palaung, and the Lisu, for example, seem to fit in the first category and, at times, the Karenni and Kachin in the second. Capturing and selling slaves was such a mainstay of the Kachin economy that an early colonial official could declare that “slavery is a national custom among the Kachins.”78 The Karen, by contrast, appear sometimes as prey and sometimes as predator.79
As is so often the case with a major commodity, slaves became virtually the standard of value by which other goods were denominated. A slave in the Chin Hills in the late nineteenth century was worth four head of cattle, a good gun, or twelve pigs. “Slaves were current coin in the hills, and passed from hand to hand as easily as a bank-note in more civilized regions.”80 The tight association between hill peoples and the social origin of most slaves is strikingly indexed in the fact that the terms for slaves and hill peoples were often interchangeable. At the bottom of a Tai kingdom, Condominas reports, were the Sa or xa in Vietnamese, like its equivalent kha in the Lao and Siamese languages. The term “can be either translated as ‘slave’ or ‘mountain tribe,’ according to the context.”81 Similarly, the term for savage or barbarian in Vietnamese, moi, has indelible servile connotations and, in precolonial time, the Central Highlands were called rung moi or “forests of the savages.” The Khmer term for barbarian, phnong, has similar connotations.82
The memory of slave raids permeates many contemporary hill societies. It is present in legend and myth, in accounts of abduction that the present generation has heard from parents and grandparents, and, in some cases, in personal memories of the elderly. Thus the Pwo Karen tell of repeated abductions from the area around Mawlamyine and their forced relocation as slaves in Tai kingdoms. When Karens want to make their children behave, they scare them by saying that a Thai will come and carry them off.83 The Lamet, in what is now Laos, have a collective memory of Burmese slave raids in which their hair was colored with lime to make them easily identifiable. In response, they remember retreating to ridge villages surrounded by pits to avoid being taken.84 The culture of some groups, it appears, has been fundamentally shaped by a fear of slavery and by the measures taken to avoid it. Leo Alting von Geusau makes this case convincingly for the Akha of the Thai-Yunnan border region, whose curing rituals recapitulate the experience of lowland captivity and eventual freedom. They, like the Lamet, see themselves as a relatively powerless group which must live by its wits and stay well clear of lowland power centers.85
The difference between slave-raiding and warfare, in this context, becomes almost a theological issue. Large-scale warfare was conducted against other kingdoms in which the existence of the realm and its dynasty were often at stake. In the smaller wars, less was at stake, but in each case the losing side could expect to have much of its population swept up and carried back to the victor’s core region. In the case of expeditions into the hills to hunt slaves, warfare gave way to something like a manhunt directed at less organized peoples whose only option was guerrilla-style self-defense or flight. The prize in all three cases was manpower: a war was a dangerous wholesale gamble for manpower; slave raids were a less dangerous, though armed, retail enterprise. Just as one could fairly call the Burmese and Tai states “warfare states,” it would be just as accurate to describe them as “slaving states.”
The taking of captives was not just the major strategic objective of warfare; it was the personal objective of the officers and men serving in the army. These were armies with both eyes fixed on the loot. Of the spoils, only elephants, horses, arms, and ammunition were reserved explicitly for the Burmese crown. The rest—children, women, men, cattle, gold, silver, apparel, and foodstuffs—was the property of the soldiers who had seized it, to dispose of as they pleased. The Glass Palace Chronicle of the Kings of Burma reports that during the late-eighteenth-century attack on Linzin (Vientiane) an infantryman leading back forty captives as his personal booty sold one of them to the king, who thought the man would make a good soldier.86 One must see these armies not as unified bureaucratic organizations collectively obeying the will of their commander but rather as something like a joint trading venture, albeit a dangerous one, from which the various investors and participants expect to make a profit. The pattern conforms to Max Weber’s description of certain forms of premodern warfare as “booty capitalism”: a speculative, for-profit war in which there is an understanding among the investors about how the proceeds will be distributed if the enterprise succeeds. When we consider that such armies also had to provision themselves en route to their military objective, we can appreciate how destructive and feared they must have been. The armies, some of which were apparently quite large, would require carts, oxen, water buffaloes, porters, rice, meat, and recruits (to replace deserters!) all along the way. When plunder is added to the requirements of “living off the land,” and the need to destroy the crops and dwellings of captives to discourage them from returning, one appreciates that this sort of warfare could be utterly ruinous without necessarily being very sanguinary.87
A certain proportion of the captives forcibly taken back to the victor’s territory thus came as personal property rather than crown property. Manpower was not simply an end of statecraft; it was also an important mark of status, reflected in the size of one’s personal entourage. Elites jockeyed to accumulate a critical mass of dependents through debt bondage and purchase that would ensure their status and wealth. The crown, prominent families, and religious establishments (for example, Buddhist abbeys) all competed against one another for the available manpower resources. At a higher level, the padi states were in competition with one another for population that represented the only guarantee of their power. Thus the Siamese and Burmans were, once Pegu had fallen, in constant conflict over which of them would come to monopolize the Mon and Karen populations that lay between them. Ava and Chiang Mai competed over the Lawa and Karen who lay between them. Their competition was not always warlike. From time to time, like real estate agents in a buyer’s market with a high vacancy rate, they would offer favorable terms to those who would agree to settle under their wing. Thus Northern Thai leaders offered the Lawa and Karen exemption from corvée and taxes as long as they would permanently settle in a designated area and provide annual tribute in valuable mountain products. In the teeth of rapacious district officials, military commanders, and slave-raiders, however, even the well-intentioned ruler was unlikely to be able to keep such a promise. In this context, the Chiang Mai ruler’s curse—“May those who oppress the Lawa be destroyed”—may fairly be read against the grain to indicate how relatively powerless he was to enforce his wish.88
When the manpower machine was working well, when a dynasty was attracting or, more likely, capturing population at a rate that far exceeded its losses, it was perforce becoming more cosmopolitan at the same time. The greater the diversity of peoples it absorbed, the more its metropolitan culture bore the linguistic and cultural traces of its hybridity. In fact, such cultural hybridity was a condition of its success. Just as the Malay coastal state, underneath its shared Malay language and the profession of Islam, was appreciably different depending on the cultural streams it had incorporated, so did the Tai and Burman padi states reflect the cultures of those they had accepted or seized under a cultural portmanteau of Theravada Buddhism and a dominant language.
The padi states’ project of amassing population was a perilous, shaky enterprise for several reasons. First, of course, the demography was working against it. Population was always, for reasons we shall explore in more detail, leaking away. Much of the history of any particular padi state could well be written in terms of the oscillation over time between ingathering and exodus. Whenever the crown was unable to replenish its population through a combination of capture by warfare, slaving expeditions, and the attractions of commerce and culture at the center, it risked a fatal erosion of its demographic and military strength. The decline of the Restored Taung-ngu Kingdom after 1626 and that of the Kon-baung Kingdom after 1760 can be traced to similar disequilibria. Following the conquests of the early Restored Taung-ngu kings, an extended period of peace meant that there were not enough new captives to offset the losses of subjects fleeing the “over-exploitation” of the “nuclear area.” In the 1780s the unraveling of the Kon-baung Kingdom under King Bò-daw-hpaya was less due to passivity than to failed invasions of neighbors and unprecedented labor drafts for public works that turned the normal trickle of subjects leaving the core into a crippling mass exodus.89
The second obstacle was simply that, looked at comprehensively, the mad scramble for manpower was essentially a zero-sum game. This was painfully obvious in the case of wars between padi states in which the gains of the victor tended to equal the losses of the defeated. Even in the case of slave-raiding expeditions to the hills, a small number of petty kingdoms were competing for the same limited pool of captives. Finally, the rulers of the padi states were systematically losing much of their accessible grain and population through their inability to overcome the combined fiscal resistance and evasion of their own elites and commoners. We turn then to this last dilemma of rule and the paradox that, when such resistance was crushed, it provoked a massive exodus, with consequences for the state that were typically even more catastrophic.
An efficient system of taxation requires, first and foremost, that the objects of taxation (people, land, trade) be made legible. Population rolls and cadastral maps of productive land are the key administrative tools of legibility. As in the case of our earlier distinction between gross domestic product and state-accessible product, there is an important distinction to be made here between the total population and what James Lee calls the “fiscal population”—the population which is administratively legible.90 A similar distinction might be made between actual cultivated land and total trade on the one hand and “fiscal landholding” and “fiscal trade” on the other. It is, of course, only the registered (“fiscal”) land and population that are assessed and, hence, accessible. The degree of slippage between fiscal resources and off-the-books resources is a rough measure of the efficiency of a tax system. In premodern political systems that slippage was substantial.
An effort at fine-mesh record keeping was made on Burmese King Thalun’s order in the early seventeenth century, “to list the land under cultivation and thus taxable; the people’s names, ages, sex, birthdays, and children; the members and lands of the various crown service groups; the local officials and their service lands, and the boundaries of their jurisdiction.”91 The king wanted, in effect, a complete inventory of his taxable resources. Like all such records, even if it was accurate when compiled, it was a static snapshot that was soon overtaken by land transfers, population movement, and inheritance, among other things. Other decrees aimed at preserving the validity of the records by prohibiting certain kinds of social change that would make them invalid. Subjects were forbidden to move without explicit permission, and were barred from changing their civil status from commoner or royal servicemen to bondsmen. The relative permanence of irrigated padi fields and a standardized “fiscal” family under a male head of household were also aids to legibility at the core.92
Above and beyond the inherent difficulties of premodern fiscal administration, the monarch faced a more systematic and intractable structural problem. He was in direct competition with his own officials, nobles, and clerics for manpower and grain. Although the royal crown service population (ahmudan) was the most accessible manpower base for the crown, its ranks were being constantly eroded. It was in the interest of such servicemen to change their status to a less onerous and less legible one. Several options were open: to be a commoner (athi), a private client to a powerful patron, or a debt bondsman, or to join a large, undocumented “floating” population. It was, at the same time, very much in the interest of the king’s officials and prominent nobles to abet, in every way possible, this shift in fiscal status, for it allowed them to sequester these resources for their own entourages and tax bases.93 Many of the Kon-baung legal codes are devoted to thwarting this drift toward fiscal invisibility and hence into the hands of other elites. Reading this litany of prohibitions again against the grain suggests that the crown was less than completely successful.
The rulers of the Thai kingdoms struggled against the same tendency for officials, nobles, and religious authorities to appropriate the crown’s fiscal resources for themselves. Thus the founder of the northern Thai kingdom of Lan Na, King Mangrai, declared the “deserters from the king’s service who try to avoid their obligations, should not be allowed to become slaves [of others than the king].”94
Both the Thai and Burman crowns, in the era before internal passports and identity papers, hit on the device of tattooing much of the male population to indelibly mark one’s status. Soldiers recruited—or press-ganged—into the Kon-baung army were tattooed with symbols showing that they were liable for military service.95 The Thais tattooed as well. Thai slaves and bondsmen were tattooed on the wrist with a mark indicating whether they belonged to the king or to a noble. If a slave belonged to a noble, that particular noble was indicated in the tattoo in much the same way a cattle brand is used to indicate livestock ownership.96 Karen prisoners of war were tattooed to indicate their status as war captives. The system of tattoos gave rise to bounty hunters, who coursed the forests looking for runaways to return to their “rightful” owners. Such measures not only indicated that the monitoring of manpower was, in most respects, more important than the registration of land, but that it was also more difficult.
The king’s officials and local power holders also had more banal reasons for any sleight of hand that would remove resources from the crown so that they could be “privatized” and plundered. Thus the population rolls were, as the first colonial censuses documented, greatly understated. Officials removed land from the land registers for a fee, appropriated poorly documented crown lands themselves, underreported tax receipts, and left households off the tax rolls altogether. William Koenig estimates that anywhere from 10 to 40 percent of crown revenue was lost in this fashion. He cites an instance after an 1810 fire in Rangoon when officials were directed to conduct a new housing census. They omitted one thousand of the twenty-five hundred houses from the new register.97 The net result was anything but an easing of the tax burden on commoners. Rather, it was a shift in the division of the spoils of state that was potentially ruinous for the crown and for commoners as well.
Faced with the steady evaporation of its tax base through outright flight and through the fiscal invisibility we have just described, the ruler of the padi state was hard put to hold the realm together. One of the few options open to him was a military campaign for captives to replenish those he was steadily losing. The advantage of new war prisoners was that many of them would become crown servicemen and hence owe service, at least initially, directly to the crown. This may, speculatively, help explain the tendency of such padi states to become warfare states. Only through warfare could the ruler stand a chance on making good, at one stroke, his continuing loss of manpower.
Smaller-scale slave raids into the hills and attacks on peripheral villages carried less risk, but the yield in manpower was correspondingly small. Larger-scale wars could bring in many thousands of captives. As noted, though, while this may have been a rational strategy for a particular ruling dynasty, it was systemically irrational. In a war between two padi states, the loser was likely to suffer a catastrophic diminution of its population.
The most thoughtful historians of premodern states in Southeast Asia have been struck by their fragility, by the boom-and-bust quality of their growth and collapse. Victor Lieberman has described them as having a “convulsive quality,” whereas Oliver Wolters has applied the term concertina.98 In this final section, I want to endorse and expand on Lieberman’s argument that there are systematic, structural reasons behind this fragility and oscillation.
The straightforward logic of “self-liquidation” can, for the purposes of illustration, be seen at work in the counterinsurgency policy of the contemporary military tyranny in Burma. Military units are attempting to control more of the insurgent border region while, at the same time, being told by their financially strapped commanders that they must provision themselves locally. Thus, a little like the premodern state, military units must find the labor, cash, building material, and foodstuffs to sustain themselves in a rugged and hostile environment. They do this, typically, by essentially capturing and concentrating a substantial civilian population around their base, which becomes their available pool of manpower, grain, and revenue. The civilians try to flee, first and foremost among them the poorest, who cannot buy their way out of forced labor or afford to provide the grain and taxes extorted from them. As a Karen schoolteacher put it to a human rights researcher, “Along the road … down in the plains, there used to be many villages, but the big villages have become small and the small villages have become forest. Many people have gone to the [other] towns or come up here [to the mountains], because the SPDC [military government] demands so many taxes from them and forces them to do all kinds of labour.”99 The consequences for those left behind are predictable: “As the abuses continue, focusing on fewer people, the less vulnerable become progressively more vulnerable and are gradually forced into flight themselves.”100
Variations of essentially this argument have been made about the premodern Tai and Burman states by Rabibhadana, Lieberman, and Koenig.101 The heartland or core region of the padi state is the most legible and accessible concentration of grain and manpower. Other things being equal, it is this population from which it is easiest and most efficient to extract the resources necessary to sustain the state and its elite. The fiscal temptation was to press heaviest on this core population and, as a result, it was likely to be the most beleaguered. Thus, under the Kon-baung kings, those in the Mandalay-Ava area were the most “combed over” for corvée and grain, whereas those at a greater remove were able to get away with more nominal tribute. If we recall that a considerable portion of the core population itself had been sequestered by officials and private notables, then it is clear that the burden fell disproportionately on the crown-service population, many of whom were descended from captives and commoners who figured on the tax rolls. This population operated like something of a homeostatic device for the state as a whole: the greater the pressure exerted on it, the more likely it would simply flee out of range or, in some cases, rebel.
Lieberman offers many examples of this pattern—of the padi state, in effect, killing the goose that lays the golden egg. The late-sixteenth-century king of Pegu (the son of the famous Bayin-naung), deserted by many outlying military tributaries, was compelled to press with desperation on his core population, forcing monks into the army and executing deserters. The harder he pressed the more population he lost. Cultivators disappeared en masse to become private retainers and debt bondsmen or defected to the hills and other kingdoms. Deprived of its grain producers and soldiers, Pegu was, at the end of the century, sacked by its enemies.102 Perhaps the most notable instance of near collapse was at the turn of the nineteenth century. Although the kingdom was blessed with the captives that Alaungpaya’s conquests and his dispossession of his opponents had added, the fiscal pressure on this population, aggravated by a drought and a failed invasion of Siam, resulted in a great exodus.103 The breakdown of the Trinh in the early eighteenth century fits this pattern as well. Increasingly, the autonomy of local notables had allowed them to escape taxes themselves and appropriate labor and property that would otherwise have been available to the state. As a result, “the burden was carried by ever fewer people who were at the same time the ones least able to pay.”104 Flight and rebellion on a large scale followed.
The king’s counselors surely knew, implicitly or explicitly, the structural problems they faced. The proverbs about manpower, their efforts to prevent the loss of manpower and grain to their own officials, their attempts at a more rigorous inventory of the resources they did have, and their search for other forms of revenue tell us as much. Knowing this, one might have expected statecraft to consist in sailing as close to the wind as they could: that is, in extracting resources just short of the point at which they would provoke flight or rebellion. Short of a series of very successful wars-for-captives and slave-raiding, this would be the most reasonable strategy.105
There are at least three reasons why the premodern state was unable to calibrate its extractions in this way. Their relative importance is hard to determine and, in any event, might vary from case to case. The first reason is simply that states did not have the kind of structural information that would allow them to make such a fine-grained judgment, especially since many of their officials had their own reasons to deceive the crown. The crop was in principle legible, but the officials were not. Second, the fiscal capacity of the population varied widely, as it would in any agrarian economy, from season to season depending on harvest fluctuations due to weather, pests, and crop diseases. Even theft and banditry could be a factor here: concentrated aboveground grain crops were just as big a temptation to gangs of thieves, rebels, or rival kingdoms as they were to the state. Allowing for the great variation in the cultivators’ capacity to pay year by year would have required the crown to sacrifice its own fiscal demands for the welfare of its peasantry. All the evidence suggests that, quite to the contrary, the precolonial and colonial states tried to guarantee themselves a steady take, at the expense of their subjects.106
There is evidence, which we shall explore in more detail, that the demography and agro-ecology of state space, in fact, makes it more vulnerable to instability in food supply and to illness. Very briefly, fully occupied monocultures seem less environmentally resilient than dispersed, mixed cultivation. They are more prone to crop disease; they have less of an environmental buffer in case of crop failure; and they promote the multiplication of their obligate pests. Much the same might be said about the concentration of people together with livestock and poultry as well. We know that most epidemic diseases are zoonotic, moving between domestic animals and humans; we know that urban populations in the West did not successfully replace themselves reproductively until the mid-nineteenth century; we know that the grain diet of early agrarian societies was nutritionally inferior to the mixed diet it replaced; and, finally, we have ample evidence of crop failures, famines, and cholera outbreaks in precolonial Southeast Asia. Although somewhat speculative, it is likely that the concentrations of rice and men in state spaces carried their own substantial risks.
The third reason, however obvious it might seem, was the tremendous capriciousness introduced by a political system in which the king was, at least in theory, all powerful. There is no rational accounting for Bò-daw-hpaya’s ruinous invasion of Siam on the heels of a famine, nor is there one for his massive use of corvée around 1800 to build hundreds of pagodas, including one at Mingun that was to have been the largest in the world.107 After all the structural and ecological reasons for the instability of precolonial dynasties have been accounted for, there is the added factor of arbitrary, tyrannical rule that is not institutionalized.
It is little wonder that the padi state was a fragile and evanescent affair. Given the demographic, structural, and personal obstacles in its path, what is remarkable, on a long view, is that it did occasionally coalesce long enough to create a defining cultural tradition.