Ernest Renan was right when he wrote over a century ago: “Forgetting, and I would even say historical error, are an essential factor in the creation of a nation, and so it is that progress in historical studies is often a danger to nationality.” That is, I believe, a fine task for historians: to be a danger to national myths.
—Eric Hobsbawm, Nations and Nationalism since 1780
There is nothing more universally modern than purveyors of ancient identities.
—Charles King
The Tribes are by no means to be found together in separate colonies, but are mixed up quite indiscriminately. Moreover, among their [Kachin] villages are also to be found Palaung, “La,” Wa, Chinese, and a few Shans.
—J. G. Scott, Gazetteer of Upper Burma and the Shan States
Like any heir who had come into a large estate, the British proceeded, as elsewhere, to take inventory of their new possessions in Burma. If the cadastral survey was the instrument for inventorying the real estate, the census was the instrument for enumerating the peoples inherited through conquest.
Once they came to the hills, the administrators in charge of the 1911 and subsequent censuses were faced with a baroque complexity that all but defeated their mania for classificatory order. How to proceed when most of the “tribal” designations were exonyms applied by outsiders and not used at all by the people so designated? Different outsiders seldom agreed on a common term. Furthermore, such exonyms were either derogatory (“slaves,” “dog-eaters”) or generic in a geographical sense (“hill people,” “upstream people”). Take, for example, the Marus along the border with China in the Northern Shan states described by J. G. Scott in the Gazetteer of Upper Burma. They did not call themselves Kachin, nor did the “authorities,” but their neighbors insisted that they were indeed “Kachin.” “They dress and intermarry with ‘Kachin’ [Jingpo] but their language is closer to Burmese than Jingpo.”1 What should they be called in the census?
The operationalization of “race” in the 1911 and 1931 censuses was, in fact, language. According to the linguistic theories of the time, it was “accepted as dogma that those who speak a particular language form a unique, definable unit and that this unit had a particular culture and a particular history.”2 With the equation of “tribe” or “race” (the two were used interchangeably in the census) and “mother tongue,” the fun had just begun. The trainers of the enumerators had to point out very carefully that the “mother tongue” was the language “spoken from the cradle” and might “not be the language spoken ordinarily in the home”; this last was to be recorded as a subsidiary language. Except for Burmans at the padi-state core who were generally monolingual, bilingualism among minority hill peoples was the rule rather than the exception. Those with Karen, Shan, or other (non-Burman) Tibeto-Burman languages as their mother tongues were typically bilingual, and often trilingual.3 Nor did the complexity dissolve at the microcosmic level of a single village. In a single “Kachin” village of a mere 130 households were found no fewer than six “mother tongues,” though the villagers used Jingpaw as a lingua franca on the order of Malay or Swahili.4
The equation of mother tongue with tribe and history assumes implicitly that the spoken language is the constant, steady thread that knits together a people. And yet the authors of the census go out of their way to note the “extreme instability of language and racial distinctions in Burma.” In an exasperated and instructive, not to say entertaining, appendix to the census titled “A Note on Indigenous Races in Burma,” J. H. Green goes further:
Some of the races or “tribes” in Burma change their language almost as often as they change their clothes. Languages are changed by conquest, by absorption, by isolation, and by the general tendency to adopt the language of a neighbor who is considered to belong to a more powerful, more numerous, or more advanced tribe or race.… Races are becoming more and more mixed, and the threads are more and more difficult to untangle.
The unreliability of the language test for race has again become apparent in this census.5
The conclusion, obvious to Edmund Leach if not to the designers of the census, is that language groups were not established by heredity, nor were they stable over time. This use of language to infer history was therefore “nonsense.” It is a conclusion shared by most who have considered the issue closely.6
Again and again one encounters the frustration of those taking the inventory of peoples but, to their dismay, confronting an apparent mess. A tidy, objective, and systematic classification of tribes would require, they believed, a stable trait, or traits, that all members of a tribe shared and that were not to be found outside that tribe. If mother tongue would not serve this purpose, neither would, it turned out, most other traits. It was clear enough that there were “Kachin,” and Karen, and Chin; what was not clear was where one began and another ended, and whether they had been Kachin, Karen, and Chin in the previous generation or whether they would remain so in the next.
The large group (7.5 million in China alone) known as Miao and the related Hmong in Thailand and Laos present a case in point. They speak three major languages, and within each of those languages there are dialects that are mutually unintelligible. Beyond that, most Miao men and many Miao women can speak three languages or more. There are self-identified Miao groups who dwell in valleys growing irrigated rice, and others at higher altitudes swiddening (opium, maize, buckwheat, oats, potatoes), foraging, and hunting. There are Miao who have largely adopted Chinese dress, rituals, and language and others in remote, isolated locations, who, while still bearing often archaic features of Sinitic culture, have held themselves apart from lowland practices. At the micro level of an individual village, the same cultural “sprawl” is evident. Intermarriage between Miao and other groups (Khmu, Lisu, Chinese, Tai, Karen, Yao, and so on) is extremely common, as is adoption from other groups.7 Even those characteristics that many Miao believe distinguish them culturally, such as cattle sacrifices and reed organ pipes, are in fact shared with others. Not a little of this sprawl derives from the fact that in parts of China, Han officials used the term Miao to designate any groups that were rebellious and refused to submit to Han administration. Over time, this appellation, reinforced by state routines of administration, has stuck; the “Miao” are often those people called Miao by powerful others—by those, that is, who can to a great extent impose their categories.
The diversity of the Karen is no less daunting. No single trait such as religion, clothing, burial rituals, or even a shared intelligible language applies to all of them. Each of the subdivisions of the Karen also displays an amazing variety. As Martin Smith notes, “the term ‘Sgaw Karen’ could today apply equally to a Burmese-speaking, University of Rangoon graduate, born and brought up in Bassein in the Delta, and an illiterate, animist, hill-tribesman in the Dawna Range [near the Thai border] who has never even met an ethnic Burman.”8 The Karen practices of adoption and intermarriage across ethnic groups, though perhaps less widespread than among the Miao or Yao, are nevertheless quite common. Nor does it appear that “Karenness” is necessarily an exclusive ethnic identity. In Thailand, at least, according to Charles Keyes, it is possible for someone to be “Karen” in a domestic, village, and church setting and to be “Thai at the market, in politics, and in interaction with Thais. Much the same is true, he believes, for the routinely frictionless movement between Thai and Chinese identities, between Thai and Khmer, Thai and Lao. Karen and many other minorities often appear to be ethnic amphibians, able to pass between such identities without a sense of conflict. Living in close symbiosis with another cultural complex, ethnic amphibians learn, often consummately, the performance required in each setting. Keyes points also to the Lua/Lawa, swiddeners and animists, who speak a Mon-Khmer tongue at home but who are so versed in the Thai language, lowland cultivation techniques, and Buddhism that they can become Thais virtually overnight when they move to the valley. Faced with the cultural plenitude of Karenness, Keyes takes the next logical step of minimizing the importance of shared cultural traits and declares that ethnicity is a self-making project. As he puts it, “ethnic identity itself [that is, its assertion] provides the defining cultural characteristic of ethnic groups.” Those who adopt this identity—assuming that other Karens accept it—become ipso facto Karen.9
What is clear beyond any reasonable doubt is that the trait-based desiderata of ethnicity find little traction in highland Southeast Asia. Even the desideratum of hilliness per se and the swiddening and dispersal that go with it are not completely satisfactory. Most of the Kachin, the Miao, and the Karen, it is true, are associated with hill residence and swiddening. Many of their rituals derive from “fire-field farming” and from hunting and foraging. But at the same time, a considerable number of self-identified Kachin, Miao, and Karen have come to occupy different subsistence niches, including fixed-field, irrigated padi cultivation, and have adopted other features of lowland residence, including the language of the padi core.
A major reason why trait-based designations of ethnic or tribal identity fail utterly to make sense of actual affiliations is precisely that hill groups themselves, as manpower systems, absorbed whomever they could. This absorptive capacity led to great cultural diversity within hill societies. The adoption of newcomers, rapid social mobility for captives, and genealogical sleight of hand fostered hill systems that were culturally accommodating. Even the presumably distinctive ranked segmentary lineage system of the Kachin, far from being rigid, was deployed to incorporate neighboring Lisu and Chinese. It was, as François Robinne has argued, a striking example of pluriethnic inclusiveness.10
The failure of any trait or complex of traits to draw a bright boundary around a “tribe” is reflected in the bewilderment of those whose job it was to create a Linnaean classification of peoples. Thus one could write of the struggle of colonial officials, confronted with the diversity of the Naga Hills (now on the Burma-India border), “to make sense of the ethnographic chaos they perceived around them: hundreds, if not thousands, of small villages seemed to be somewhat similar to each other but also very different, by no means always sharing their customs, political system, art or even language.”11 The confusion was genuine and four-barreled. First, any particular trait was likely to be a gradient and occasionally a seamless continuum from one village or group to another. Lacking sharp, discontinuous changes in ritual, dress, building styles, or even language, any line of demarcation was arbitrary. Second, if one did in fact meticulously chart small variations and then try, conscientiously, to justify a particular trait boundary, another nearly insurmountable problem arose. The boundaries for trait A, B, and C did not map onto one another; each yielded a different line of demarcation, a different classification of “ethnicities.” The third and fatal difficulty was that any such trait-mapped ethnicities were unlikely to coincide with the phenomenological understandings of the tribal people whose life-world was being mapped. The colonial ethnographers’ map said they were A, but they said they were B and had always been. How could that not matter? And if the classification exercise somehow survived these blows, a fourth difficulty, time, would surely be the coup de grâce. Those with some sense of historical change understood that in terms of traits or in terms of self-identification, the A had, not so long ago, been B and seemed, alarmingly, now on their way to becoming C. How could an ethnic group, a tribe, be so radically unstable over time and still be a people?
The radical flux of identity should, in one respect, occasion little surprise. Zomia is and has been what might be called a “fracture zone” of state-making, much like the Caucasus and the Balkans. It has been peopled for two millennia at least by wave after wave of people in retreat and flight from state cores—from invasion, slave raids, epidemics, and corvée. There, in this zone of refuge, they joined a hill population located in a geographical setting of such ruggedness and relative isolation that it encouraged the drift of dialects, customs, and identity. The great variety of subsistence techniques, keyed in large part to altitude, only encouraged this diversity. Add to this the exchange of populations within the hills by slaving, raiding, intermarriage, and adoption, and the complexity of identities is compounded, helping to explain the crazy-quilt pattern encountered by the colonizers. The conclusion reached by Geoffrey Benjamin and Cynthia Chou, contemplating an analogous flux in the Malay Peninsula, applies equally to Zomia: “The interflow of genes, ideas, and languages has been so intensive and multidirectional as to render futile any attempt to delineate the various ‘peoples’ in terms of completely distinct bundles of geographical, linguistic, biological, or cultural-historical features.”12 This radical undermining would seem to leave us at an impasse. There is, clearly, no such thing as a “tribe” in the strong sense of the word—no objective genealogical, genetic, linguistic, or cultural formula that will unambiguously distinguish one “tribe” from another. But, we might well ask, who is confused? The historian and the colonial ethnographer might be mystified. The mixed villages in northern Burma were “anathema to the tidy bureaucratic officials” who, until, the last moment of imperial rule, were still trying in vain to draw administrative lines between the Kachins and Shans.13 But hill people were not confused; they were in no doubt who they were and who they were not! Not sharing the researcher’s or administrator’s mania for mutually exclusive and exhaustive categories, hill people were not paralyzed by identities that were plural and variable over time. On the contrary, as we shall see, the ambiguity and porosity of identities was and is, for them, a political resource.
There are, of course, “tribes” in the lived experience of hill peoples. Self-identified Karen, Kachin, Hmong, and others have fought and died for identities that many believe have a deep and continuous history, a belief that would probably not stand up to critical scrutiny. Such powerful identities are, in this respect, no less fictitious and constructed than most national identities in the modern world.
The only viable analytical alternative is to take such self-identifications as our point of departure. As was proposed nearly forty years ago, we must treat tribal divisions as “essentially political in origin.” Ethnic identity on this reading is a political project. When, as Michael Moerman noted, hill peoples in Thailand such as Karen, Tai, Lawa, Palaung, and T’in are in an ecological situation that allows a choice “among major bases and symbols of ethnicity [such] as religion and type of farming, as well as among such emblems as dialect, diet, and dress,” the key question becomes what calculations govern that choice.14
The perspective adopted and elaborated here is a radical constructionist one: that ethnic identities in the hills are politically crafted and designed to position a group vis-à-vis others in competition for power and resources. In a world crowded with other actors, most of whom, like modern states, are more powerful than they, their freedom of invention is severely constricted. They craft identities, but not in circumstances of their own choosing, to paraphrase Marx. The positioning in question is above all a positioning vis-à-vis the lowland state and other hill peoples. That is the function of hill identities. Those who, over many centuries, migrated into Zomia were, in effect, refusing assimilation into the lowland states as peasants. In coming to the hills, they joined a population that had never been incorporated in a lowland state or one that had left it a long time ago. The basic choice was between statelessness and incorporation. Within each of these choices there were, of course, several possible calibrated variations. This perspective has been persuasively articulated by Hjorleifur Jonsson, who, in his important study of the Mien in northern Thailand, relates “how people have moved among the categories of state-subjects and non-subject clients in the forest, as well as from being autonomous uplanders dividing into two social directions, some becoming state clients and others abandoning village life for foraging in small bands. This relates back to the general case of a shifting social landscape, and how people move among structural categories, in and out of particular relationships, repeatedly reformulating the parameters of their identities, communities, and histories.”15 On this view, what imperial administrators and census takers saw as an exasperating confusion could more appropriately be seen as evidence of how subsistence routines, social structures, and identities could be deployed to express a “positionality” vis-à-vis the major lowland states.
Ethnic and “tribal” identity, in the nineteenth century and much of the twentieth, has been associated with nationalism and the aspiration, often thwarted, to statehood. And today, the utter institutional hegemony of the nation-state as a political unit has encouraged many ethnic groups in Zomia to aspire to their own nation-statehood. But what is novel and noteworthy for most of this long history in the hills is that ethnic and tribal identities have been put to the service not merely of autonomy but of statelessness. The paradox of “antistate nationalism,” if it might be called that, is typically overlooked. But it must have been a very common, perhaps the most common, basis for identity until, say, the nineteenth century, when, for the first time, a life outside the state came to seem hopelessly utopian. E. J. Hobsbawm, in his perceptive study of nationalism, took note of these important exceptions: “One might even argue that the peoples with the most powerful and lasting sense of what might be called ‘tribal’ ethnicity not merely resisted the imposition of the modern state, national or otherwise, but very commonly any state: as witness the Pushtun-speakers in and around Afghanistan, the pre-1745 Scots highlanders, the Atlas Berbers, and others who will come readily to mind.”16
The most significant others who ought to come readily to mind are, of course, the innumerable hill peoples of Zomia who have been avoiding states for more than a millennium. It is perhaps because they have fought and fled under so many names, in so many locations, and against so many states, traditional, colonial, and modern, that their struggle lacked the single banner that would have easily identified it.
The founders of the early padi state had to assemble their subjects from among hitherto stateless populations. And when, as so often happened, a state disintegrated, subsequent state-makers had to reassemble subjects from among the shards available by raiding other states or incorporating stateless hill peoples. Earlier “wave” theories of migration supposed that large numbers of Burmans and Tais swept down from the north into the alluvial plateaus suitable for irrigated rice and defeated or chased off the earlier inhabitants. This view, now discredited for lack of evidence, implicitly assumed that the Burmans and Tais came as whole societies, rulers and subjects, to establish themselves as conquerors. It now appears far more plausible that the Burmans and Tai who appeared were a kind of militarily and politically adept pioneer elite with the skills to organize and dominate a padi core. Their subjects, on this reading, were gathered in from the surrounding stateless hills and confected into the node of power we have called the padi state. Providing that we take the long view, then, most of the people today known as Shan are ex–hill people who have, over time, been completely incorporated into the Shan valley polity. Most of those who appear today as Burman are the descendants—recent or long ago—of non-Burman populations from both hill and valley (Shan, Kachin, Mon, Khmer, Pyu, Chin, Karen, and so on). Most Thai are, in the same fashion, ex–hill people, and, on the longest view, the creation of “Hanness” itself should be seen as the most successful, longterm, state-based ingathering of all time. The early need for manpower was such that none of these states could afford to be too particular about where its subjects came from.
The most carefully observed and elaborated model of such ingathering comes to us from the studies of the Malay maritime trading center—the negeri— largely because Europeans have been describing it since before the sixteenth century. Designed as an “interstitial” polity to mediate between hill collectors and international trade, and to defend its strategic location, the negeri amassed manpower both by force and by the attraction of commercial gain. Its water-borne slaving expeditions cast a wide net, and the captives thus assembled were incorporated into the negeri. The formula for full incorporation was minimal: becoming a retainer of a Malay chief, professing Islam, and speaking Malay, the lingua franca of trade in the archipelago. A negeri was less an ethnicity than a political formula for membership in the polity. As a result of the vagaries of trading and raiding, each Malay trading negeri took on a different cultural complexion depending on the mix of peoples it had incorporated over time: Minangkabau, Batak, Bugis, Acehnese, Javanese, Indian, and Arab traders and so on. At its most successful, a negeri such as Melaka would become a magnet for traders from far away, rivaling Venice. Its radical dependence on fluctuating overseas trade, however, made the negeri a very fragile affair.
The small Tai or Shan muang (polity), though somewhat less vulnerable to radical oscillations in trading fortunes, resembled the negeri in many respects.17 It was in constant competition with its neighbors and larger states for manpower. It too captured or admitted its subjects without regard to their background. It too was highly stratified but open to rapid social mobility. The terms of full membership were the profession of Theravada Buddhism—another universal religion—padi cultivation, fealty to a Tai lord, and the ability to speak the local Tai language.
The small Tai and Shan statelets spread all the way from northern Vietnam across Zomia to northeastern India. Because they were so numerous and so small, they represent a much-observed laboratory for many padi-state processes that have also operated historically on the Burmese, Thai (the most successful of the Tai states!), and even Han states.
It is a rare historian or ethnographer who has not remarked on the degree to which many—in some cases most—Tai and Shan commoners were, as Leach puts it, “descendants of hill tribesmen who have in the recent past been assimilated into the sophisticated ways of Buddhist-Shan culture.”18 Georges Condominas much later echoed the view that “especially in the Shan states and other Tay [Tai] principalities, the bulk of the population remained composed of non-Tay.”19 In keeping with the openness of the muang, the non-Tai who lived there could use their own language as well as Tai and follow their own customs.
The combination of slavery and fluid social mobility—and the alchemy with which ex–hill people became valley Shan or Tai (or Burman or Thai)—was such that historians have been careful to distinguish these forms of bondage from those of their New World counterparts.20 Leach describes a typical process of entry through bondage. Kachins, as individuals or as groups, enter the service of Shans as laborers or soldiers and, in recompense, acquire Shan wives. By settling in the valley and adopting the rituals of the new place (his Shan wife’s local spirit guardians or nats), he severs his link to his Kachin kinsmen and enters the Shan system of stratification at the bottom. Shan terms for Kachin in general incorporated a prefix (kha) meaning serf, and Leach estimates that “nearly all low class Shans” in the Kachin Hills area were “of slave [captive] or commoner Kachin origin.”21 Observing over a slightly longer time frame, Condominas shows that ex–hill people entering the Tai system as slaves became, before long, commoners with other Tai. And if, in the course of a power struggle, one of them managed to seize power, he was given a noble Tai name and his genealogy was reworked to bring his origin into line with his power.22 So despite a Tai saying declaring that “a Kha [serf] is as different from a monkey as a Tai is from a Kha,” the formulas for citizenship in this intensively competitive polity avoided practices that would drive its population away.
But the generalized practice of raiding and slavery could lead to a more rapid transformation of a polity. A visitor to Chiang Mai in 1836 described a Shan chief who had twenty-eight wives, all taken as captives, while his subordinates had themselves seized women as well. With the majority of wives of alien origin, J. G. Scott reports that “the physical features of the inhabitants of a locality might completely change in a couple of generations, and the language as well, for the mothers teach the children.” He adds that for years it had been the Shan custom for Shan chiefs to have “Chinese, Burmese, Karen, and Kachin wives, sometimes captured, sometimes bought, sometimes received as presents. Occasionally, the issue of such unions took power, with the result that often a Sawbwa is of a different race than the bulk of his subjects.”23
Another route to “Shanness” or “Tainess,” which is to say to hierarchy and stateness, had a more wholesale quality to it. This was for a successful Kachin chief to transform his more “open-rank” egalitarian realm into a petty Shan-style kingdom. Much of Leach’s classical work is devoted to this theme. Typically, the strategy involved a powerful Kachin chief taking a wife from an aristocratic Shan lineage. This marriage, while it transformed the Kachin chief overnight into a Shan prince, at the same time barred him from continuing to give wives to Kachin lineages without prejudicing his status as a Shan prince. The effect was to prevent his compatriot Kachin lineages from achieving status by contracting marriages with his (chiefly) lineage. His Kachin followers then had a choice between acquiescing in the transformation and, in effect, becoming Shan commoners; revolting and killing or driving off the chief; or leaving and founding a new community. It is this logic that Leach traces out so brilliantly.24 Wherever a Kachin chief found himself in a position to collect tribute regularly from traders and lowlanders, he inevitably tried to set himself up as a petty Shan Sawbwa, not always successfully.
This process of petty state-making, of turning hill people into valley people, could, it seems clear, just as easily be reversed. The Tai/Shan statelets were at least as vulnerable as the larger padi states to the decomposing forces of invasion, famine, tyrannical rule, raids, and civil wars of succession. What became of the dispersed population of a collapsed Shan statelet? Many, it seems from the evidence, may have moved to more hospitable Shan states within striking distance. Many others, perhaps especially the “recently” ex-Kachin and ex-Lisu, must have returned to the hills, resumed swiddening, and readopted their prior identities. This would have been a relatively easy and familiar option and would not have prevented them from returning to the padi core when it became safe to do so. It makes sense to see this ethnic transformation, before the twentieth century at least, as a two-way street and ethnic identities as dual or amphibious.
For a great many hill people in Southeast Asia, the lowland state identity closest at hand was the Tai muang, a “low church” version of the more distant, “high” models available from the Han, Burman, and Thai courts. A successful valley state would attract Kachin, Lisu, Akha, Wa, Khmu, Lue, Mien, and many other hill peoples to a new identity that was, one suspects, less definitive or permanent than becoming a subject of a much larger kingdom. For some ethnic groups, located in the interstices of two or more padi states, there were more risks and more choices. This seems to have been the situation of the Karen, especially the Pwo Karen, located between the Mon, Burman, and Thai padi states. Much of the Burman population of lower Burma was undoubtedly of Mon or Karen origin, and a convincing case has been made that the Karen are culturally, and strategically, medially positioned between these three valley states, able, with little friction, to move between one identity and another.25 The Karen have both exploited this medial position and suffered grievously from it. While they have boasted of their role as agents and representatives of the Thai court in exacting tribute from other hill peoples in elegiac terms—“Our administration was so oppressive and the taxes so heavy that the tumplines of the carrying baskets rang like guitar strings when the people were bringing in their tribute”—they were the first to pay a heavy price as a suspected Thai “fifth column” during the devastating Burmese invasions of Siam.26
If, as seems evident, group boundaries are porous and identities flexible, then we would expect that they would shift over time as one identity became increasingly advantageous and another less so. Such seems clearly to have been the case recently with the Zhuang, a Tai language–speaking group and, today, one of the largest official minority groups in China. Perhaps driven, like other groups, into hilly southwest China by Han expansion, the Zhuang considered themselves a valley people like other Tai. They settled at generally lower altitudes below the swiddening Yi, Miao, and Yao. They came to occupy—or better stated, to create—a cultural niche intermediate between the Han and the highlanders above them. The saying was: “The Miao live at the head of the mountain; the Zhuang at the head of the river, and the Han at the head of the street.”27 Over time, the hitherto stigmatized Zhuang invented a mythical Han origin for themselves and, in effect, “played Han” to the minorities above them. The revolutionary government’s new delineation of nationalities according to Stalinist criteria, however, identified them, mostly on the grounds of language, as “Zhuang.” This seemed at first a new stigmatization, and most putative Zhuang resisted the classification, claiming instead that they were “Han who can speak the Zhuang language.” In the 1953 census most Zhuang-speakers did not self-identify as Zhuang. The party created Zhuang administrative districts despite vernacular, folk understandings at variance with their categories.
Under the new minority policy, however, there were substantial new advantages to a “Zhuang” identity: new political and administrative posts, preferential access to technical schools and higher education, and exemption from the “one-child” policy. Suddenly, the cash value of an official Zhuang identity skyrocketed to a point where it more than compensated for the possible stigma, and the new identity took hold. The “Zhuang-Han,” of course, were still, in a vernacular sense, Zhuang-and-Han, but one side of their Janus-faced identity, the official side, had become more valuable. The new dispensation reversed what had apparently been a process of gradual Sinicization of the “Zhuang.”
To see the growth of the valley states as the product of certain techniques of cosmopolitan ingathering, some more coercive and some less so, is decentering. It serves to correct the more ethnic understanding of state-making that characterized both early historiography and modern nationalist histories. Shan culture, Leach concluded, “is not to be regarded as a complex imported into the area ready-made from somewhere outside, as most authorities seem to have supposed. It is an indigenous growth resulting from the economic interaction of small scale military colonies with an indigenous hill population over a long period.”28 Much the same could be said for the Burmese and Siamese precolonial states at their most successful. Each was an effective political formula for the gathering and holding of populations from various linguistic and cultural origins at the padi core—for creating a concentrated, productive force suitable for state-building. As suggested earlier, the Burmese and Siamese were to the numerous and varied populations they incorporated as the two thousand conquering Norman families were to the indigenous peoples of Britain.
Burmese and Thai polities are, then, best seen as recipes for state-making rather than as ethnic projects. First, there were, it appears, no massive wholesale invasions from the north eliminating or replacing the previous inhabitants. Second, if we squint a bit at the cultural foundations of the precolonial padi state, they make best sense as frameworks for state space rather than as ethnicized markers. The keystone, of course, is the technique of irrigated rice cultivation that makes possible a padi core in the first place. It was not, however, a technique exclusive to the Burmese and Tai, inasmuch as it had been the basis of the Khmer, Pyu, and Mon courts earlier. The cosmology and architecture of the Indic court center is, as it were, the ideological superstructure of the divine monarchy and was adapted for that purpose. Theravada Buddhism, another import, served as a universal domain, to assemble “ethnic deities and spirits” under a new hegemon, much as the padi state’s subjects were gathered around the court. Local spirits (nat, phi) were accommodated as subsidiary deities much as Catholicism accommodated pagan deities under the rubric of the saints. Even the languages of the state-builders, Burmese and Thai, were, in their written form (from Sanskrit via Pali), linked to the legitimating cosmology of Buddhism and the Indic state. Much of what passes as the ethnic particularity and distinctiveness of the Shan, Burmese, and Thai cultures is closely tied to the basic devices for state-building. Put another way, “stateness” is built into the foundations of ethnicity. Reciprocally, in the Shan, Burmese, and Thai view, much of the ethnicity of those populations in the hills, those not-yet-gathered-in, consists precisely in their statelessness.
The great distinction, culturally, between the valley kingdoms and the hills is the striking uniformity of valley society religiously, linguistically, and, over time, ethnically as well. While the historical process of ingathering was a cosmopolitan enterprise, the population thus assembled came to share a set of common cultural practices and institutions. One could travel hundreds of miles within the padi state and still encounter religious practices, architecture, class structures, forms of governance, dress, language, and ritual that were strikingly similar from one end to the other. By contrast, traveling even a short distance in the hills would expose one to a crazy-quilt of languages, rituals, and identities. Valley systems, in the words of Hjorleifur Jonsson, are centripetal, while hill systems are centrifugal. This “sharp contrast between the rather uniform lowland areas and a bewildering variety of the upland areas” is not, he insists, the result of different migrations but rather the systematically different social outcomes of the centripetal tendencies of closed-rank systems, on the one hand, and the centrifugal tendencies of open-rank systems on the other.29
This cultural distinction not only marks off the hills from the great valley kingdoms of the Burmese, Thai, Han, and Vietnamese; it is just as notable in the contrast between the Shan statelets and their hilly neighbors. Leach highlighted this discrepancy a half-century earlier and pointed to the probable cause:
The hill people who are neighbors to the Shans are astonishingly varied in their culture; the Shans, considering their wide dispersal and their scattered form of settlement, are astonishingly uniform. My argument is that this uniformity of Shan culture is correlated with a uniformity of Shan political organization which is in turn largely determined by the special economic facts of the Shan situation. My historical assumption is that the valley Shans have everywhere, for centuries past, been assimilating their hill neighbors, but the unchanging economic factors in the situation have meant that the pattern of assimilation has everywhere been very similar. Shan culture itself has been modified relatively little.30
The uniformity of the petty Shan kingdoms lies, as Leach implies, in the fact that they are miniature state spaces geographically, economically, and politically.
Each of the Shan states lies between six hundred and nine hundred meters in a valley or plain, “some long and narrow, some rounded like a cup, some flattened like a saucer, some extensive enough to suggest the Irrawaddy Valley on a miniature scale.”31 As in the larger valleys, each of these areas was suitable for wet-rice cultivation, and Shanness became synonymous with padi planting. The compacting of people and grain in the small core area made possible state formation on a correspondingly small scale. Padi planting, however, had other decisive social effects. Strong reliance on a single crop, as in the larger valleys, comes to dominate the work routines and social organization of much of the population. Each household planted, transplanted, weeded, and harvested the same crop at virtually the same time, and in much the same way. Coordination of water use required a certain level of institutionalized cooperation and dispute settlement. Agricultural uniformity, in turn, facilitated ritual uniformity around the padi plant itself, its harvest rituals and the control of water. A padi-planting society also shaped a common material culture—diet, cuisine, farm implements, plow animals, household architecture, and so on.32
Permanent padi-field cultivation also leads to systems of landed property and inheritance and the social class distinctions they foster. Inequalities per se do not distinguish the valleys from the hills. Status differences and inequalities abound in the hills, but in contrast to inequality in the padi state, they are not underwritten by inherited inequalities of property enforced, if need be, by the coercive power of a rudimentary state. The homogenizing effects of a common agrarian regime and a class system were frequently punctuated by rebellion, which typically reproduced the previous social order under new management. The only structural alternative was flight to the common-property regime and the open-ranked systems of the hills.
Social and cultural homogeneity in the valley state was also an artifact of the political sway possible over a padi zone in which the friction of terrain was low. It permitted the creation and maintenance of a common institutional order, as well as the density of trade and exchange that fostered cultural integration. Power could be projected far more easily across this geographical space than across the heterogeneous hills. Precisely because it exercised many of the same functions, albeit on a miniature scale, as the great valley states, the Shan realm’s palace, ritual, and cosmology constituted a provincial imitation of the palace, ritual, and cosmology at Ava, Amarapura, and Mandalay.
The process of valley homogenization throughout Southeast Asia, Victor Lieberman claims, was much advanced by increasing state centralization between 1600 and 1840. A combination of imitative state-building modeled on the West and enlarged revenues from international trade allowed the mainland state to stamp out religious heterodoxy, create a more uniform and efficient tax regime and administration, and promote kingdomwide economic integration and militarization.33 Advances in firearms, military organization, land surveying, record keeping, and the dissemination of texts had, on a modest scale, the same distance-demolishing effects as railroads, steam power, and the telegraph were to have later in the nineteenth century. While the valley states were busy fabricating more uniform Burmese, Siamese, Vietnamese, and Shan, the hills continued to fabricate further differences, heterogeneity, and new identities.
Most hill people in mainland Southeast Asia, before the colonial state insisted on classifying them, did not have what we might consider “proper” ethnic identities. They often identified themselves by a place name—the people of X valley, the people of Y watershed—or by kin group or lineage. And to be sure, their identity would vary depending on whom they were addressing. Many names were implicitly relational—the “uphill” people, the people of the “west ridge”—and their designation made sense only as one element in a relational set. Still other names were exonyms used by outsiders—as was often the case for the Miao—having no further meaning except in that context. Identities were, to complicate matters further, plural; most hill people had a repertoire of identities they could deploy in different contexts. And such identities as there were, were subject to change: “Ethnic identity and language differences have tended to be fluid in mainland Southeast Asia. A group could change both in a relatively short period of time as a result of being in close contact with other peoples.”34 A certain plasticity of identity was built into precolonial power relations. Many valley peoples, as well as hill peoples, found themselves located between two or more power centers whose waxing and waning influence shaped their world. Before the advent of modern statecraft, with its practices of territorial administration and mutually exclusive sovereignties and ethnicities, such ambiguities were common.
Flexible identities also characterized systems of social stratification in which those of lower status sought to emulate, or at least to defer to, powerful others of higher status. In his analysis of a Tai area of northern Vietnam, Grant Evans notes the duality of identities and how they are deployed.35 Thus a low-status group, the Sing Moon, considered serfs by the Black-Tai, speak Tai in addition to their own language, have a Tai name as well as their own “ethnic” name, and generally emulate the Tai. The Black-Tai, for their part, in the past emulated Vietnamese mandarin dress and incorporated Vietnamese vocabulary, while the higher-status White-Tai went so far as to adopt Vietnamese funeral ritual and to assimilate into Vietnamese society through intermarriage. Tai elites in general, Evans shows, are culturally amphibious; they deploy the strongly Tai side of their identity when exercising power over their nominally Tai equals and inferiors and the Vietnamese side when dealing with those above them. His point is that identities are plural and that they are systematically structured by relations of power and prestige. This general theme is elaborated with great subtlety by Jonsson in his analysis of the Mien (Yao) of northern Thailand, who embody several registers of self-representation, each of which can be strategically enacted, depending on the context.36
Quite apart from the confusion of colonial officials and census takers, later ethnographers and historians of Burma have only confirmed Leach’s earlier judgment that ethnic boundaries are labile, porous, and largely artificial. Thus, for example, different observers could classify the same group of people as “Karen,” “Lawa,” or “Thai” depending on the criteria used and the purposes of the classification. Where different peoples had long lived in close proximity, they often merged seamlessly into one another so that demarcating a boundary between them seemed both arbitrary and futile.37 And, as we noted earlier, the Lua/Lawa, a swiddening, Mon-Khmer-speaking animist people, are so intimately familiar with the Thai language, padi cultivation, and Buddhism that it is no exaggeration to say that they can be a convincing Lua on Monday and a convincing Thai on Tuesday. To relegate them to a single ethnic category makes little sense. A better description might be to say that X had a bandwidth of traits or identities that could be deployed or performed as the situation required. A person’s ethnic identity in this sense would be the repertoire of possible performances and the contexts in which they are exhibited.38
Another way of recognizing the bandwidth or repertoire of identities available to many actors is to recognize, as Leach did, that they have, as it were, status positions in several different social systems at the same time. So common was this that F. K. Lehman believed that throughout much of the region ethnicity, far from being an ascriptive “given,” was a choice. “Whole communities might be faced with a conscious choice about what group to belong to.”39 This seems a useful way to conceive of such plural identities, providing that we keep three qualifications in mind. The first is that powerful outsiders, especially states, constrain the identity choices of most actors. Second, movement toward one of an array of identities does not exclude the possibility that, should circumstances change, that movement might be reversed. Finally, and surely most important, we must never confuse what an outsider might perceive as a momentous shift in identity with the lived experience of the actors involved. Leach points out in this context that “Kachin” communities, whether egalitarian or stratified, and Shan communities share much of the same ritual language, although they interpret it differently. When an economically well-situated small Kachin polity becomes part of a Shan muang, it will appear to an external observer that the Kachin have become Shan. This is true enough, but to the actor, “this change may be hardly noticeable. In becoming ‘sophisticated,’ the individual merely begins to attach Shan values to the ritual acts which previously had only a Kachin significance.” It is only the external observer “who tends to suppose that shifts in the culture and social organization of a group must be of shattering significance.”40
There is surely something defective about any analytical understanding of identity change that is so radically at variance with the experience of real actors. Ethnic change can, I believe, be differently formulated so as to accommodate better the vernacular understandings of local actors. If we assume, for many hill people, a plurality of identity repertoires, then it follows, as we have seen, that various portions of that repertoire will be elicited by a particular social context of action. Performed identity will, in other words, be situationally cued. Someone with, say, a broad Karen-Thai repertoire will dress, speak, and behave differently in the Thai marketplace than in the context of the Karen village festival. There is, of course, no reason at all to suppose one part of the repertoire is more authentic or “real” than any other. To a great extent, then, the expressed or enacted identity is a function of the relative frequency of the social context in which it is appropriate. If, say, the Karen-Thai in question moves into a predominantly Thai, lowland settling and plants irrigated rice, then the frequency of the Thai social and cultural context will prevail, and with it the Thai portion of the enacted repertoire. What looks to an outside observer like a change in ethnic identity is no more and no less than a shift in the relative frequency with which the Thai-coded part of the repertoire is performed. This could occur gradually and, in any event, if conceived in this fashion need not imply any momentous subjective sense of displacement or loss.
The historical relationship between the Mon and the Burmans, both lowland padi-planting peoples, not only provides an illustration of identities as plural and performed but also suggests the strategic value of having a range of identities at one’s disposal. At the start of the eighteenth century, both Mons, who slightly predominated, and Burmans shared the Irrawaddy Delta. The main differences between them had to do with body markings (Burmans tattooed below the waist), hair style (Mons cut their hair round in front, while Burmans had long hair gathered in a top knot), clothing, and language. A change in identity was a matter of substituting these codes, and in those sections of the Delta where Mons and Burmans lived adjacent to each other and were bilingual, this was a comparatively simple matter. As Ava’s power waxed, so did the proportion of the population adopting Burman cultural codes—speaking Burmese and tattooing their thighs. As Pegu’s power waxed, Burmans in its orbit cut their top knot and spoke Mon. Independent municipalities, nominally tributary to Ava or Pegu, switched their loyalties. Although it seemed clear that the conflict itself served to crystallize and politicize certain cultural markers that later came to be seen as ethnic, the origin of the conflict was by no means ethnic.
One sees in this context the adaptive value of certain identities.41 Being able to represent oneself, as the circumstances dictate, as Burman or Mon must have been the salvation of many a bystander or captured combatant in these wars between Pegu and Ava. It is tempting to see this command of a “mixed portfolio” of identities as a cultural insurance policy, an escape social structure. Like a chameleon’s color adapting to the background, a vague and shape-shifting identity has great protective value and may, on that account, be actively cultivated by groups for whom a definite, fixed identity might prove fatal. Like the “jellyfish” tribes described earlier, such plasticity affords outsiders no easy institutional access.
“Tribes” in the strong sense of the word have never existed. By “strong sense,” I mean tribes as conceived as distinct, bounded, total social units. If the test of “tribeness” is that the group in question be a genealogically and genetically coherent breeding population, a distinct linguistic community, a unified and bounded political unit, and a culturally distinct and coherent entity, then virtually all “tribes” fail the test.42 As noted earlier, the actual disposition of cultural practices, social integration, languages, and ecological zones rarely offers a sharp break, and when it does, one such break almost never maps on another. Nor is “tribe,” as it was once imagined, part of an evolutionary series of some kind—for example, band-tribe-chiefdom-state, or, alternatively, tribe-slavery-feudalism-capitalism.
States and empires have been founded by peoples conventionally understood as tribes—Jinggis Khan, Charlemagne, Osman, the Manchu. And yet, it would be far more correct to say that states make tribes, than to say tribes make states.
Tribes are what have been called a “secondary form,” created in two ways and only in the context of a state or empire. The antonym or binary to “tribe” is “peasantry.” The difference, of course, is that the peasant is a cultivator already incorporated fully as a subject of a state. Tribes or tribals, on the other hand, are those peripheral subjects not (yet?) brought fully under state rule and/or those who have chosen to avoid the state. Colonial empires and the modern state have been most prolific at creating tribes, but the designation of a tribal periphery was just as common for such earlier empires as the Roman, Tang China, and even the small Malay trading state.
The “tribe” might be called a “module of rule.” Designating tribes was a technique for classifying and, if possible, administering the non- or not-yet-peasants. Once a tribe and its tribal area had been marked off, it might be used as a unit for tribute in goods and men, as a unit over which a recognized chief could be appointed and made responsible for its conduct, and as a military zone of pacification. At the very least, it created, however arbitrarily, a named people and their supposed location for purposes of bureaucratic order where an otherwise indistinguishable mass of settlements and peoples without structure had often prevailed.
States and empires create tribes precisely to cut through the flux and formlessness that characterize vernacular social relations. It is true that vernacular distinctions were made between, say, swiddeners and foragers, between maritime and inland populations, between grain growers and horticulturalists. Such distinctions, however, crisscrossed many other distinctions of language, ritual, and history; they were typically gradients rather than sharp discontinuities and rarely became the basis for political authority. At some level, it simply did not matter how arbitrary the invented tribes were; the point was to put an administrative end to the flux by instituting units of governance and negotiation. Thus the Romans insisted on the territorialization of named barbarians under chiefs who, in principle, could be held responsible for their conduct. The bureaucratic grid was necessary “because there was so much fluidity in social bonds and internal barbarian politics.”43 Whether the designations made vernacular sense to the natives was largely beside the point. In late imperial and Republican China, on the southwest frontier, the names for the subdivision of the troublesome “Miao” were arbitrary designations based loosely on women’s dress and bore no relation to the vernacular terms of self-identification.44
Colonial rulers faced much the same “anarchy” of vernacular identifications and resolved it by decreeing administrative tribes that were no less arbitrary. Armed with ethnographers and deterministic theories of social evolution, the French in Vietnam not only drew boundaries around the tribes they dimly discerned and appointed chiefs through whom they intended to rule but placed the peoples so designated on a scale of social evolution.45 The Dutch accomplished much the same administrative alchemy in Indonesia by identifying separate indigenous customary law (adat) traditions which they proceeded to codify and use as a basis for indirect rule through appointed chiefs. As Tanya Li puts it, “The concept of ‘adat community’ assumed, as it simultaneously sought to engineer, a rural population separated into named ethnic groups with ‘traditions’ stable enough … to serve as definitions of group identity and centralized political structures with recognized leaders.”46
This technology of rule, at a single stroke, not only proposed new and sharp identities but assumed a kind of universal, hierarchical, chiefly order. Acephalous, egalitarian peoples without chiefs or any permanent political order beyond the hamlet or lineage had no place in the new dispensation.47 They were hustled willy-nilly into a world of chiefs by fiat whether they liked it or not. Peoples whose vernacular order was egalitarian lacked the institutional handles by which they could be governed. Those institutions would have to be provided, if necessary, by force. In what became known as the Shan states of eastern Burma, the British were confronted with a population roughly half of whom were acephalous and egalitarian (gumlao Kachin, Lahu, PaO, Padaung, Kayah). Seeking the hierarchical institutions that would provide a political entry point for indirect rule, the British naturally chose to rely instead on some forty or so Shan sawbwas who, more often in theory than in practice, laid claim to rule their local domains. Though this choice sparked resistance then and later, it was the only institutional transmission belt available to the British.
Once invented, however, the tribe took on a life of its own. A unit created as a political structure of rule became the idiom of political contestation and competitive self-assertion. It became the recognized way to assert a claim to autonomy, resources, land, trade routes, and any other valuable that required a statelike claim to sovereignty. The recognized idiom for making a claim within a state was an appeal to class or to estate—the peasantry, the merchants, the clergy. The recognized idiom for claims outside state space was an appeal to tribal identities and entitlements. Nowhere was this more apparent than in the white settler colony of North America. As Alfred Kroeber aptly writes, “The more we view aboriginal America, the less certain does any consistently recurring phenomenon become that matches with our conventional concept of tribe; and the more largely does this concept appear to be a White Man’s creation of convenience for talking about Indians, negotiating with them, administering them—and finally impressed upon their own thinking by our sheer weight… the time may have come to examine whether it is not overwhelmingly such a construct.”48
In this secondary sense, named tribes with self-consciousness of their identities do most certainly exist. Rather than existing in nature, they are a creative human construction—a political project—in dialogue and competition with other “tribes” and states. The lines of demarcation are essentially arbitrary at the outset, given the great variety of ethnographic difference. The political entrepreneurs—officials or not—who endeavor to mark out an identity based on supposed cultural differences are not so much discovering a social boundary as selecting one of innumerable cultural differences on which to base group distinctions. Whichever of these differences is emphasized (dialect, dress, diet, mode of subsistence, presumed descent) leads to the stipulation of a different cultural and ethnographic boundary distinguishing an “us” from a “them.” This is why the invention of the tribe is best understood as a political project.49 The chosen boundary is a strategic choice because it organizes differences in one way rather than another, and because it is a political device for group formation. The only defensible point of departure for deciding who is an X and who is a Y is to accept the self-designations of the actors themselves.
States fabricate tribes in several ways. The most obvious is to create them as a template for administrative order and political control. But it is striking how often a tribal or ethnic identity is generated at the periphery almost entirely for the purpose of making a political claim to autonomy and/or resources.
The creation of the Cossacks as a self-conscious ethnicity, conjured up out of thin air, so far as origins are concerned, is particularly instructive for understanding ethnogenesis in Southeast Asia. The people who became the Cossacks were runaway serfs and fugitives from all over European Russia. Most of them fled in the sixteenth century to the Don River steppelands “to escape or avoid the social and political ills of Muscovite Russia.”50 They had nothing in common but servitude and flight. On the vast Russian hinterland, they were geographically fragmented into as many as twenty-two Cossack “hosts” all the way from Siberia and the Amur River to the Don River basin and the Azov Sea.
They became a “people” at the frontier for reasons having largely to do with their new ecological setting and subsistence routines. Depending on their location, they settled among the Tatars, Circassians (whose dress they adopted), and Kalmyks, whose horseback habits and settlement patterns they copied. The abundant land available for both pasture and agriculture meant that these pioneer settlers lived in a common property land regime where each family had its own independent access to the means of subsistence and complete freedom of movement and residence. An ethos of independence and egalitarianism, desired by a people who had known servitude, was underwritten by the political economy of frontier ecology.
Cossack society was, at this stage, something of a mirror image of tsarist Russian servitude and hierarchy. All three of the great peasant uprisings that threatened the empire began in the Cossack lands. Here, as in Zomia, the stateless frontier also attracted religious dissidents, most prominently the so-called Old Believers, who associated religious reforms with servitude.51 After the defeat of the Bulavin uprising (1707–8), the autonomy of the Cossacks was made contingent on supplying fully equipped cavalry units to the tsar’s forces. And after the ruthless military campaign of suppression against those Cossacks most closely associated with the Pugachev rebellion (1773–74), the rudimentary local democratic assemblies of the Cossacks were supplanted by a titled, landholding Cossack nobility with its own serfs—mostly from the Ukraine.
Not by any stretch of the imagination a coherent “people” at the outset, the Cossacks are today perhaps the most solidaristic “ethnic” minority in Russia. To be sure, their use as a “martial minority”—like the Karen, Kachin, Chin, and Gurkha levies in South and Southeast Asia—contributed to this process of ethnogenesis.52 It did not, however, initiate it. As an invented ethnicity, Cossackdom is striking, but it is not unique. Cases of essentially maroon communities that became distinctive, self-conscious, ethnic formations are reasonably common. In place of the Cossacks, the case of the maroons of Surinam—who developed into no fewer than six different “tribes,” each with its own dialect, diet, residence, and marriage patterns—would have served just as well.53 The Seminoles of North America or Europe’s Gypsies/Roma are also cases of ethnicities that were fused from unpromising, disparate beginnings, by a common ecological and economic niche as well as by persecution.
All ethnicities and tribal identities are necessarily relational. Because each asserts a boundary, it is exclusionary and implicitly expresses a position, or a location, vis-à-vis one or more other groups falling outside the stipulated ethnic boundary. Many such ethnicities can be understood as asserted structural oppositions between binary pairs: serf-versus-free Cossack, civilized-versus-barbarian, hill-versus-valley, upstream (hulu)-versus-downstream (hilir), nomadic-versus-sedentary, pastoralist-versus-grain producer, wetland-versus-dryland, producer-versus-trader, hierarchical (Shan, gumsa)-versus-egalitarian (Kachin, gumlao).
The importance of “positionality,” and often agro-economic niche, is so common in the creation of ethnic boundaries that what begins as the term for a location or a subsistence pattern comes to represent ethnicity. For Zomia and the Malay world it is striking how frequently a term merely designating residence in the hills of, for example, Padaung, Taungthu, Buikitan, Orang Bukit, Orang Hulu, Mizo, Tai Loi, has become the actual name for a tribe. Many such names surely began as exonyms applied by valley states to the hill people with whom they traded, and connoted rudeness or savagery. Over time, such names have often taken hold as autonyms carried with pride. The frequent coincidence of ecological and occupational niches and ethnic boundaries has often been noted by anthropologists, and Michael Hannan has gone so far as to claim that “in equilibrium, ethnic group boundaries coincide with niche boundaries.”54
The most essentialized distinction of this kind is perhaps that between the barbarians and the grain-growing Han people. As the early Han state grew, those remaining in, or fleeing to, “the blocks of hilly land, marsh, jungle, or forest” within the empire became known by various terms but were, as we have seen, collectively called “the inner barbarians.” Those extruded to the steppe fringe, where sedentary agriculture was impossible or unrewarding, were “the outer barbarians.” In each case, the effective boundary between different peoples was ecological. Baron von Richtofen in the 1870s vividly described the abruptness of the boundary between geologies and peoples: “It is surprising, after having crossed over several [patches of loess soil], to see, on arriving on the summit of the last, suddenly a vast, grassy plain with undulating surface.… On the boundary stands the last Chinese village; then follows the ‘Tsauti’ [grassland] with Mongol tents.”55 Having shown that “the Mongols” were not some ur-population, but instead enormously diverse, including many ex-Han, Lattimore saw the hegemony of ecology: “The frontiers between different types of soil, between farming and herding, and between Chinese and Mongols coincided exactly.”56
Ecological niche, because it marks off different subsistence routines, rituals, and material culture, is one distinction around which ethnogenesis can occur. But it is neither a necessary nor a sufficient condition for ethnic or tribal formation. Inasmuch as the creation of such markers is a political project, it follows that they can also be attached to distinctions of no intrinsic importance in order, for example, to stake a claim to a valuable resource. The invention of the Kayah/Karenni “tribe” as a people distinct from other nearby Karennic people seems a clear case of this kind.57
The birth of the Karenni in the early nineteenth century is recent enough to allow us some grounded speculation about the origin of the “tribe.” It appears to be associated with the arrival of Karen millenarian pretenders claiming to be princes on the Shan model among this otherwise egalitarian, non-Buddhist community in the 1820s. As we shall see in the next chapter, millenarian movements have played a disproportionate role in the genesis of new communities in the hills. The creation of a Shan-style kingdom with its own Sawbwa “succeeded in transforming one of a mere congeries of central Karen dialects into a very distinctive Kayah social and cultural system.”58 Imitative state-making of this kind is not so unusual. What is unusual is its success both politically and culturally. That success, in turn, was contingent on the happy fact that the newly minted Karenni statelet was home to the most valuable stands of teak in the country.
The assertion of a new tribal identity outfitted with a small-scale state apparatus had the effect of establishing a local monopoly over the trade in teak. Charismatic leadership served to fuse these loosely related Karennic communities into something like a joint stock company “in order to wrest control of the increasingly lucrative teak trade from the Shan for whom they had been working.”59 The Karenni ethnic entrepreneurs borrowed the statecraft model closest at hand: that of the Shan padi state, itself borrowed in turn from the Burmese monarchy, which would serve to make a sovereign claim to the teak and defend it. As an identity-creating and resource-controlling strategy, it succeeded admirably.
Many identities are crafted, it is clear, with similar purposes in mind: to defend a strategic trade route location, to assert an exclusive claim to water, minerals, or valuable land, to claim ownership of a particular commodity, to defend fishing or hunting rights against competition, to guard entry into a lucrative occupation, or to claim ritual privileges. The creation of tribes and ethnicities in this sense might be termed the standard mode of claim-making by stateless people who interact with states. It serves, for such societies, essentially the same purpose that the formation of a trade union, a corporation, or a craft guild would serve in a more contemporary society.60 Those who successfully stake a claim to resources on this basis acquire a powerful reason for embracing the new identity. By the same token, they exclude others from access to these same resources. Those thereby excluded and forced into a less desirable niche are often reciprocally ethnicized.61
The demarcation of tribes in Africa was, as in all colonies, an official imperial project as well. A small army of specialists was busy drawing ethnic boundaries, codifying customs, assigning territories and appointing chiefs to create manageable units of imperial rule, often over stateless peoples. Some grid of classification had to be imposed on a bewildering cultural variety so as to yield named units of tribute, taxes, and administration. Driving the enterprise, Wilmsen writes, was “that self-fulfilling prophecy that foretold the existence of tribes and, through administrative order, created what it could not discover.” The tribe, once established as the only social form proper to the representation of stateless peoples, became quickly hegemonic. Arbitrary or fictitious as it might be, “natives understood that they must draw themselves up into tribes” in order to function within the colonial framework.62 The enterprise of dividing up the natives into mutually exclusive, territorially delimited tribes was not an administrative mania peculiar to Cartesian Enlightenment thinking or, for that matter, Anglo-Saxon, Calvinist tidiness. And one need only read Caesar’s Gallic Wars to notice a tribal order of the same kind which, however confounded it might have been by facts on the ground, was a gleam in the eye of every Roman governor. The Han-Chinese imperial project, with its appointed tributary chiefs (tusi) and named barbarians, bears the marks of a comparable administrative exercise. The tusi system (ruling barbarians with barbarians) was devised in the Yuan Dynasty (1271–1368) and flourished until the eighteenth century in those areas of the empire where direct control was either impossible or fiscally unprofitable.63
Underneath the arbitrary exercises of classification elaborated by states lies the ever-changing tumult of local struggles over resources, prestige, and power. These struggles constantly produce new social and cultural cleavages—quarrels over ritual, factions striving to control the best land, lineages vying for marriage alliances, succession struggles for local leadership. The potential basis of new and competitive social units, in other words, is being reproduced daily.
Following Max Gluckman’s original insight, we can roughly distinguish between centripetal and centrifugal conflict.64 When factions battle over a chieftainship, implicitly agreeing on what the prize is and so reaffirming the importance of the unit itself, their conflict is centralizing. When a faction splits off or secedes to found another unit, such conflict is decentralizing or centrifugal. The demographic and geographic setting of Zomia, in this context, promoted centrifugal conflict. It was a relatively simple matter, say, for the losing faction in a leadership dispute to hive off, open new swiddens, and found a new settlement. It was a relatively egalitarian alternative to what otherwise might have become subordination and permanent hierarchy. The friction of distance also meant that separate communities often remained in relative isolation, especially as compared with valley society. Just as rugged topography and relative isolation, given time to percolate, promote growing differences in speech dialects—a kind of linguistic speciation—so, too, the conditions of Zomia encourage the multiplication and solidification of cultural difference. These processes of cultural drift and differentiation are often the raw material for “tribal” distinctions. But we must not confuse cultural difference with tribal or ethnic group identity. The creation of named tribes or ethnic groups is a political project that sometimes makes use of cultural differences. Many striking cultural differences, on the other hand, are never politicized, and by the same token, striking cultural differences are frequently accommodated within the same tribal polity.
Once launched, the “tribe” as a politicized entity can set in motion social processes that reproduce and intensify cultural difference. They can, as it were, create the rationale for their own existence. Political institutionalization of identities, if successful, produces this effect by reworking the pattern of social life. The concept of “traffic patterns” used by Benedict Anderson to describe the creation by the Dutch colonial regime in Indonesia, virtually from thin air, of a “Chinese” ethnic group, best captures this process.65 In Batavia, the Dutch discerned, according to their preconception, a Chinese minority. This mixed group did not consider itself Chinese; its boundaries merged seamlessly with those of other Batavians, with whom they freely intermarried. Once the Dutch discerned this ethnicity, however, they institutionalized their administrative fiction. They set about territorializing the “Chinese” quarter, selected “Chinese” officials, set up local courts for customary Chinese law as they saw it, instituted Chinese schools, and in general made sure that all those falling within this classification approached the colonial regime as Batavian “Chinese.” What began as something of a figment of the Dutch imperial imagination took on real sociological substance through the traffic patterns of institutions. And voilà!—after sixty years or so there was indeed a self-conscious Chinese community. The Dutch had, to paraphrase Wilmsen, through an administrative order, manufactured what they could not discover.
Once a “tribe” is institutionalized as a political identity—as a unit of representation with, say, rights, land, and local leaders—the maintenance and reinforcement of that identity becomes important to many of its members. Geoffrey Benjamin has shown that highland groups such as the Senoi and Semang responded to the lure and danger of the colonial and Malay state by becoming more “tribal,” by instituting a set of marriage practices that promoted dispersal and foraging. A cultural taboo against the use of the plow further discouraged sedentism.66 The more successful the identity is in winning resources and prestige, the more its members will have an interest in patrolling its borders and the sharper those borders are likely to become.67 The point is that once created, an institutional identity acquires its own history. The longer and deeper this history is, the more it will resemble the mythmaking and forgetting of nationalism. Over time such an identity, however fabricated its origin, will take on essentialist features and may well inspire passionate loyalty.
In the egalitarian societies usually designated tribal, I would argue that the most prevalent mode of reckoning descent is by stipulation.
—Morton Fried, The Notion of Tribe
Early colonial officials could, in a sense, be forgiven for “finding” tribes in the hills.68 Not only was it their expectation, but the self-representation of many hill peoples reinforced that expectation. Where there were no states to speak of, the principles of social cohesion were formally ordered by kinship, genealogy, and lineage. These principles were, of course, exactly what the colonizers expected to find in the tribal zone. While the political realities of competition, usurpation, rebellion, migration, social fission, and fluctuating, plural identities were dizzyingly complex and in constant motion, what Marxists would call the ideological superstructure maintained the appearance of a well-ordered, historically coherent descent group. The formal roles of succession, descent, and precedence were saved by a kind of genealogical and historical legerdemain. If a sense of continuity and symbolic order are important for a given state of affairs, then this process makes eminent sense. After all, contingencies of a fragmented and tumultuous hill politics could not be foreseen, let alone ordered. It was far easier to adjust and interpret the outcome of, say, a successful usurpation or an anomalous marriage in a way that artfully demonstrated that, after all, the rules had been satisfied and were still intact.
All hill peoples, without exception, as nearly as I can tell, have had long experience in reworking their genealogies to absorb strangers. Hill societies were also manpower systems and sought to augment their numbers by absorbing new migrants, by adoption, by marrying and incorporating outsiders, by purchase, and by slave-raiding expeditions. The new manpower was welcome not only for opening new swiddens but also to improve the political and military weight of the receiving group. Valley societies, seeking manpower as well, readily absorbed its newcomers into a class order, typically at the bottom. Hill societies, by contrast, attached newcomers to descent groups or kindreds, often the most powerful ones.
The Karen have something of a reputation as less likely to intermarry than many other hill peoples, but, living in the interstices of many more powerful neighbors, they are in fact prodigious at incorporating new members. A partial list of those absorbed would include Chinese, Shan, Lamet, Lisu, Lahu, Akha, Burmans, Mon, Lao, and Lue.69 The Akha, also an “in-between” people in the sense of being a midslope society with valley peoples below them and other hill peoples above them, have a long-established system of assimilating people. A man marrying into this society, known for its exceptionally long oral genealogies, is accepted as the founder of a new (junior) lineage so long as he practices ancestor service, has a son, and speaks Akha. Many of the lineages with comparatively short genealogies (a mere fifteen or twenty generations) actually recall their ancestors’ previous identity: Yunnan-Chinese, Wa, Tai. A captured slave is incorporated into his master’s lineage or one genealogically close to it. According to Leo Alting von Geusau, the repertoires of incorporation are so long-standing and routine that it is clear, over time, that the Akha, while remaining the Akha, have been massively replenished, in a genetic sense, by infusions of newcomers.70 Genealogically speaking, however, the tumult disappears inasmuch as all of the immigrants are in short order “naturalized” by being worked into the warp and woof of existing descent patterns.
The same genealogical sleight of hand operates to devise chiefly lineages that reconcile the facts of power with the ideology of descent. A chiefly Kachin lineage can produce a genealogy forty or more generations deep. Leach regards these genealogies as “fictional” and “of no value as evidence of historical fact.” Once a lineage has become influential, it in turn inspires commoner lineages to rewrite their own genealogies in order to stress, to their advantage, their closeness to the powerful lineage. Actual power in the Kachin lineage system turns on lavish feasting while meeting ritual obligations, thereby establishing a claim to the loyalty and labor (manpower) of all those in the sponsor’s debt. Anyone who successfully fulfills what are seen as the obligations of an aristocrat will be accepted as an aristocrat, virtually no matter what the facts of descent may be.71 Leach makes it abundantly clear that a genealogy can be devised to throw a mantle of legitimacy over virtually any set of actual power relations: “Social climbing then is the product of a dual process. Prestige is first acquired by an individual by lavishness in following ritual obligations. This prestige is then converted into recognized status by validating retrospectively the rank of the individual’s lineage. This last is largely a matter of manipulating the genealogical tradition. The complicated nature of Kachin rules of succession makes such manipulation particularly easy.” Furthermore, “Any influential aristocrat” can “reconstruct remote sections of his genealogy in his own favor.” Despite the fact that in principle “a man’s rank is … precisely defined by his birth, there is almost infinite flexibility in the system as actually applied.”72
The “tribality” of the hills was thus abetted by elaborate genealogical myths reconstituted as necessary so as to legitimate, by descent, the actual distribution of power in society. Saga tellers were the specialists whose job it was to bring the myths in line with the facts. Driven by some combination of myopia and administrative convenience, colonial officials sought out hierarchy and tribality, and the elaborate myths of descent in the hills were there to encourage them. They established a Shan state in an area where less stratified people were just as numerous. Among the Kachin, they much preferred the aristocratic, chiefly, autocratic variety and showed great distaste for the “anarchic,” democratic, gumlao Kachin.73
For the Karenni as well, recruitment to leadership depends largely on charisma, feasting, and political and military skills. The importance of heredity and genealogical standing lies in the way in which political success must be reconciled with the ideological rule of rightful descent. “Even a usurper,” Lehman explains, “upon seizing power, attempted to show that one or more of his ancestors was of the ‘royal bone’ even though the usurper may have been an ordinary villager.” Inasmuch as the Karenni, and Karen society in general, follow cognatic kinship, in which male and female lines are reckoned, finding the requisite connection was even easier than with the Kachin. If, then, the colonizers were looking for tribes constituted by orderly rules of descent and long histories, the uplanders they encountered were only too happy to oblige them with retrospective genealogical order in which they themselves clothed their turbulent politics.74
The tribe, as a formal social institution, appears, in other regions as well, to be more in evidence as a kind of ideological exoskeleton than as a useful guide to political realities. One of the most famous tribes in history—that of Osman, founder of the Ottoman Empire—was in fact a motley collection of different peoples and religions collaborating for political purposes. This was not an exception. Surveying the evidence, Rudi Paul Lindner claims that “Modern anthropologists’ field studies [in the Middle East] show that tribal, clan, and even camp membership are more open than the tribal idiom or ideology might indicate.”75 The tribe was, in Osman’s case, a useful vehicle to bring together Turkish pastoralists and Byzantine settlers. And if blood ties were deemed desirable in keeping with the idea of the tribe as consanguines, then clan genealogies could be recalled to forge distant relationships. The segmentary lineage model is, without doubt, a common tribal ideology, but it is not common tribal practice except insofar as it is necessary to keep up appearances.76
The hegemony of blood ties and rules of genealogical descent, as the only legitimate foundation for social cohesion, though at variance with the facts, was so powerful as to dominate self-representations. It was, for hill people, the only way to justify actual power. Like Charles Dickens’s couple the Veneerings, who were quite as determined to keep up appearances to themselves as to their neighbors, the mimicry is not a cynical ploy but a mode of reasoning about social relations. One can even think of it as a democratic mechanism in which the members of a community confer retrospective legitimacy on leaders who observe the obligations of ritual and generosity incumbent on chiefs. Tribal thinking in this special sense was as deeply embedded in highland ideology as it was in the imagination of the colonizers.
Given what we know of the ambiguity of ethnic identities, the porosity of their boundaries, the creation and demise of different identities, and the constant “power politics” taking place beneath the comparatively placid surface of genealogical continuity, a radical constructionist position on upland identities seems inescapable. At the very least, as Leach has shown for the Kachin, any hill population will have a range of social forms it can assume. Much of the determinate form applied to this flux was, as we have seen, an artifact of the imperial imagination. Thomas Kirsch, following Leach’s lead, directed attention to the flux of social organization itself as an important phenomenon to be explained. He argued that “none of these various upland peoples enshrined in the ethnographic tradition of Southeast Asia has (or had) any permanent or immutable ethnographic status. Rather they are all undergoing a continuous process of change.”77
The very indeterminacy of social forms in the hills, the pliability of histories and genealogies, the baroque complexity of languages and populations is not just a puzzle for rulers, ethnographers, and historians; it is a constitutive feature of hill societies. First, it is what one might expect in a zone of refuge, peopled, as in parts of Latin America, by a multitude of migrants, deserters, ruined peasants, rebels, and a preexisting, variegated hill society. The topography itself conspired to promote and preserve the cultural and linguistic pluralism of the hills. But it also seems reasonable to see this indeterminacy as an adaptive response to a context subject to radical, sudden, and unpredictable change. Noting how the Karen are spread throughout many ecological zones and adjacent to several more powerful valley kingdoms, Renard believes that the remarkable suppleness of their social structures, their oral histories, kinship patterns, subsistence techniques, cuisine and architecture is adapted for travel and change. If necessary, most Karen groups can turn on a dime. It is a quality that has great adaptive advantages and has served them well.78
Never quite knowing what role they will be called on to play, what situations they will have to adapt to, hill peoples find it in their interest to develop the widest cultural repertoire possible. Jonsson refers to much of this repertoire as “tribal identity formation” and notes in particular that “tribality,” being virtually the only idiom for extravillage action, was one element of that repertoire. He convincingly places the range of social and economic practices at the disposal of hill peoples at the center of his analysis: “People have moved among the categories of state subjects and non-subject clients in the forests, as well as from being autonomous uplanders, dividing in two social directions, some becoming state clients and others abandoning village life for foraging in small bands. This relates back to the general case of a shifting social landscape, and how people move among structural categories, in and out of particular relationships, repeatedly reformulating the parameters of their identities, communities and histories.”79
We can, I think, discern two axes along which these options are arrayed; they are all but explicit in Jonsson’s analysis. One axis is that of equality-versus-hierarchy and the second is statelessness-versus-“stateness,” or state subjecthood. The foraging option is both egalitarian and stateless, while absorption into valley states represents hierarchy and subjecthood. In between are open-ranked societies with or without chiefs and hierarchical chiefly systems sometimes tributary to states. None of these quasi-arbitrarily defined locations along these axes is either stable or permanent. Each represents, along with others, one possible adaptation to be embraced or abandoned as the circumstances require. We now turn finally to the structure of these choices.
In puzzling through how the Karenni/Kayah came into being as an identity and a petty state, Chit Hlaing [F. K. Lehman] concluded that they could be understood only as a position or strategic relationship within the larger constellation of other Karennic-speaking groups and the adjacent padi states, especially the Shan and Burman. It followed that when that constellation was transformed or disrupted, it would prompt the Karenni to adjust their social structure or even their very identity accordingly.80
If we then think metaphorically of the system Lehman points to as a solar system, we can talk broadly of the mass of different bodies that make it up, the relative distances between them, and the gravitational pull each exerts on the others. The largest planets in these systems are, to pursue the metaphor, the padi states. They wax and wane, they may, by their rivalry, limit one another, and the smallest of them may be hostage to its hilly neighbors, but by and large their concentration of manpower, material culture, and symbolic centrality make them centers of gravity.
Here, however, the metaphor breaks down, inasmuch as the padi state can be a repelling force as well as an attractive one, and it exerts several different kinds of influence. Its cultural charisma, its symbolic reach, is greater than any other force it exerts. Even in the most remote hill settlements one encounters symbols of authority and tokens of power that seem to float up in fragments from the valley states: robes, hats, ceremonial staffs, scrolls, copies of court architecture, verbal formulas, bits of court ritual. There is hardly any claim to extravillage authority in the hills that does not deploy some cosmopolitan trapping to enhance its assertion of authenticity. In the hills where Han and Theravada symbolic penumbras overlap, fragments from both lowland systems mingle promiscuously. These feather-light symbolic shards travel easily to the hills because they are largely ornamental—a form of cosmological bluster. They recapitulate, in miniature, the journey made by the ideas and symbolic technology of divine kingship from South India to the classical courts of Southeast Asia.
Economically, the gravitational attraction of the padi-state core is nearly as broad. The lowland court centers in the mainland, as in the Malay world, were the outlets, for more than a millennium, for an international luxury trade in which hill products were the most valuable commodities. As explored earlier, the hills and valleys, as different ecological zones, were locked in mutual economic dependence. This trade could not, by and large, be coerced. Even in the case of tribute-based exchange, the hill tributary, though nominally inferior according to valley documents, often had the upper hand, and tribute relations were welcomed as opportunities for mutually advantageous exchange. The economic integration of hill and valley was extensive because it was uncoerced and mutually beneficial.
The reach of the Southeast Asia padi state was quite modest, however, when it came to direct, political-administrative control. Topography, military technology, low population, and an open frontier conspired to limit the successful application of coercion to a relatively small core area. Where coercion did come into play was in slaving expeditions (via war and intermediary slave-raiders) designed to capture and settle a large population within this narrow sphere of control. Flight could negate that achievement.
Given these sharp limitations, almost every valley state had a working alliance—sometimes formalized—with one or more adjacent hill populations. It was in the interest of some hill groups to settle close by the valley core to take advantage of the often rich ecotone zone between hill and valley in an attempt to dominate, as intermediaries, the trade between them. In the case of the Yao/Mien with the Chinese court and the Lawa with the courts of Chiang Mai and Jengtung, the alliance apparently took the form of a written decree or code.81 In essence, the “contract” depicts a kind of bargain. In return for tribute and good behavior (no rebellions!), the hill people, in the case of the Yao, are given leave to “cross the mountains” to find new swiddens, are exempt from taxes, corvée, and tolls, and will not be required to kneel before lords and officials. The documents are filled with civilizational discourse placing the Yao and Lawa well outside the magic circle of civilized life. As Jonsson astutely points out, they also have the effect of naming and perhaps stabilizing an identity in flux, implicitly abrogating to the court center the right to confer land and mobility rights, and go far toward demarcating a “tribal territory” and assuming that it will have accountable chiefs. The document could be considered, in Han terms at least, a formula for “cooking barbarians.”
Such valley charters for hill peoples can as well, I think, be read against the grain. It was vitally important for the valley courts to have nearby hill allies. They constituted a crucial buffer and early warning system between the valley core and its valley-state enemies across the mountains.82 The hill allies could guard vital trade routes and mediate trade and diplomatic relations with other hill peoples. Finally, they could become slave raiders themselves, helping to replenish the unstable core population. Though such arrangements might look like deference and submission to a valley official, they could as easily be seen as a hill achievement, insisting on their terms for an alliance, including not having to bow and scrape before valley officials. Truly “cooked” barbarians would, presumably, bow. The way in which the Yao/Mien is known to flourish the document to valley officials and outsiders suggests that it might be construed this way.83
Such arrangements are legion. Located in the interstices of several lowland kingdoms, groups of Karen became, at one time or another, allies of each. They were integral to the initial victory of Mon-Pegu over Ava in the mid-eighteenth century, contributing three thousand troops under a pretender who may have been Karen or Mon-Karen. The Pwo Karen in this area were known as the Mon-Karen (Talaing-Kariang), as distinct from the more northerly Sgaw Karen, sometimes known as the Burman-Karen. When the Pegu kingdom was crushed and much of its population dispersed, the Karen fled with the Mon, seeking Thai protection. The Thai “planted” Karen on the frontier as an early-warning system and—in Burmese eyes—a fifth column. To the Tai kingdom of Chiang Mai, Karen were considered “keepers of the forest,” ritually important as first-comers to the land and valuable allies and trading partners. The identity of different Karen groups was thus marked in each location and period by the lowland society with which it was affiliated.84
Every “civilized” lowland padi state required one or more hill-dwelling barbarian allies in a relationship that was often mutually advantageous. The Akha have been paired to the Tai states of Kengtung and Sipsongpanna, the Chin to the Burmese court, the Lawa to the Tai Yuan at Chiang Mai, the Wa to various Shan/Tai states, the Pwo-Karen to the Mon, the Lawa to La Na and earlier to Lamphun, the Jarai to the Kinh, the Palaung to the Shan, upland Tai to the Lao, Kachin to Shan; and in the northwest, the Naga are said to have been a kind of highland auxiliary to the Manipuri court.85 In each case there grew up a kind of cultural symbiosis in which the hill allies, or some of them, came more closely to resemble their valley partners.
This does sound very much like a formula for “cooking” barbarians. More than that, it was a formula for absorption and assimilation. If, as has been argued, the Thai and Burmans came in small numbers as military colonists, then these and other valley populations may well have been constituted in just this fashion.86 Close hill allies were increasingly likely to be governed by chiefs with valley connections and to be increasingly hierarchical. They would then correspond in the Han scheme to “cooked barbarians.” The Han civilization series—raw barbarians, cooked barbarians, full subjects/“entered the map”—is structurally similar to the Shan civilizational series outlined by Leach: egalitarian/gumlao, stratified gumsa, Shan.87 Leach portrays ethnic succession in what might be termed a gradient. Such succession is to be found in much the same form stretching between all padi states and their adjacent hill allies. This is, after all, the social and cultural route by which hill people become valley subjects: physical proximity, exchange and contact, linguistic integration, ritual appropriation and, in the classical case, wet-rice cultivation. It is to be stressed that this route is a gradient, not a series of abrupt, wrenching changes; it may not even be perceived at all in terms of ethnic succession!
If we can imagine this ethnic succession as a relatively seamless affair, then it follows that it could be just as seamless when the direction is reversed. The route to lowland “civilization” is also the route to highland autonomy, with innumerable way stations in between. In the case of wars or epidemics, the transition might be abrupt (though perhaps familiar), but it might just as often have been a gradual and imperceptible process as the padi state decayed, as trade routes shifted, or as taxes became more onerous. The way to the valley state was a two-way street and leaving need be no more jarring or traumatic than entering.
Blow us all up with cannons or make us all eighteen thousand of us Nawabs.
—Pashtun elders to British
The Lamet simply could not understand the concept “chief.”
—Karl Gustav Izikowitz, Lamet
Because of their savagery, the Bedouins are the least willing of all nations to subordinate themselves to each other as they are rude, proud, ambitious and eager to be leaders.
—Ibn Khaldun
A major reason why Leach’s Political Systems of Highland Burma is such a durable classic is that the opposition between what he calls the autocratic factions and democratic-egalitarian factions he finds among the Kachin travels well outside its immediate ethnographic context. For stateless peoples living on the margins of states, it seems to represent a fundamental choice about positioning. In canvassing the literature of his time, especially works on the Assam-Burma border region, Leach found many other examples of the contrast between democratic-egalitarian forms and autocratic-monarchical forms within indigenous groups. He cites work on the Chins, the Sema, the Konyak, and the Naga.88 To these examples we could add the Karen, the Lahu, the Wa, the Karenni, and perhaps a good many more if one conducted a comprehensive search of the literature.89
The British-led “pacification” forces were struck by the resistance they met in egalitarian Kachin areas. “Our opponents here, were the Kumlao Kachins, whose principal characteristic is that they do not own the authority of any chief, even in single villages.”90 They had, it was noted, “no form of salutation or obeisance.” Acephalous communities like the gumlao were subversive to British—or any other—administration; they provided no institutional levers or handles with which to enter the community, negotiate with it, or govern it. The colonial administration accordingly would only recognize “properly constituted tracts under a Duwa” and cautioned officers, even in these villages, to be alert to “this spirit of independence” and to suppress it without delay.91 Thus the compiler of the Gazetteer, proud heir to a democratic tradition himself, can write without a trace of irony, “Such republican or democratic communities are no longer permitted within the Burma administrative boundary.”92
In the case of the Kayah, as Lehman demonstrates, the democratic and autocratic principles are embedded in “two simultaneous ritual cycles and sets of personnel.”93 What one might call the aristocratic cult makes symbolic gestures beyond the locality and, in particular, adopts the trappings and symbols of the Shan kingdoms and the Burmese royal capital at Mandalay. Its ritual center is the center of the village where tall teak poles—a diagnostic feature of Kayah villages—which are equated with the flagstaff found in most Shan and Burmese pagodas (symbolizing the submission of local spirits to the Buddha) and are topped with an umbrella-form (hti) finial which decorates most Buddhist architecture. The hereditary priesthood devoted to this cult makes offerings to a high god whose name is derived from the Shan term for lord, as is the term sawbwa. The priesthood may not mix with, intermarry with, or accept offerings from the other cult’s priests, who are devoted to the local spirits (nat) of the countryside, especially the forest.
What seems important for our purposes is that the ritual complex of the Kayah appears thoroughly amphibious. Both democratic and autocratic components are present but are ritually segregated. One complex seems, by copying lowland forms, to contribute to the ideological superstructure of the sawbwas and state formation in general, while the alternative cult is purely local and makes no reference to chiefly authority. If fluidity of identity—being able to “turn on a dime”—involves shifts from hierarchical to non-hierarchical forms, then the Kayah seem fully equipped, ritually, for either eventuality.
A final comparison of egalitarian and hierarchical social formations points to some of the cultural practices that work to hinder the development of permanent hierarchies of state and power. The Lua/Lawa, a relatively stratified hill society, and the Lisu, a relatively egalitarian group, are illustrative. Among the Lua there is strong emphasis on the ritual and feasting superiority of elite (samang) lineages, which, at the same time, control access to land.94 The ruling lineages have elaborate and deep genealogies emphasizing their status and their connection with powerful lowland courts—Chiang Mai, in particular. Prominent in that connection is a “charter” somewhat like that of the Mien, exempting them from corvée, conscription, and supplying fodder for elephants and horses, and confirming their right to swidden. The Lisu, on the other hand, emphasize equal access of all lineages to competitive feasting, open access to land, and the lack of essential differences in rank or status.
For our purposes, however, two features of the egalitarian Lisu are notable. First, they have short and truncated genealogies, in what amounts to a refusal of history. The purpose, after all, of most lineage histories, oral or written, is to establish a claim to distinction and rank—to establish a “lineage” for those claims. If, then, lineage histories are abbreviated or ignored altogether, it amounts to something of a cultural discouragement, if not prohibition, of historical claims to superiority. To have little or no history is, implicitly, to put every kin group on roughly the same footing. We have seen at some length how the absence of a written textual history and written genealogies may have the same strategic and adaptive advantages for subaltern groups. Oral genealogies, however contrived and invented, are also claims of the same genre, and to repudiate them is also an egalitarian move. It is often and correctly noted that text-based civilizations have consistently seen the stateless people outside their grasp as peoples without history.95 But what we encounter here is the practice of disavowing status-building histories in the name of preventing hierarchy and its frequent companion, state formation. The Lisu are without history not because they are incapable of history but because they choose to avoid its inconveniences.
The absence of history in this sense contributes to the fact that in egalitarian groups, each lineage—or, for that matter, each family—has its own particular customs and usages. There is, however, one “tradition” to which most Lisu proudly point: namely, the tradition of murdering headmen who become too autocratic. As Paul Durrenberger puts it, “The Lisu loathe … assertive and autocratic headmen,” and the “stories Lisu tell of murdered headmen are legion.”96 Such traditions can also be found among a good many egalitarian hill peoples. How frequently these traditions are acted upon is hard to establish, although the original gumlao revolt against Kachin headmen reported in the sources appears to have been a case in which they sustained something of a political movement. Such cautionary tales are, in any event, a kind of egalitarian, structural prophylactic warning of the possible consequences to any would-be autocratic headman bent on cementing his lineage’s power.
Among the hierarchical Lua, lineages are ranked; they jockey for status; and part of the jockeying rests on claims to superiority based on different, and fabricated, origin myths and genealogies. The Lisu, like the gumlao Kachin, deny lineage ranking and ranked feasting, deny history, and, more directly, thwart the emergence of ambitious headmen who might take them in that direction. The egalitarian Lisu have, in effect, created a culture that is a fairly comprehensive program of state prevention.
The incorporation of egalitarian and hierarchical models of social organization within a recognizably single culture is by no means confined to the Kachin-Shan context. It is found throughout much of Southeast Asia.97 More speculatively, there is some reason to suppose that it is a structural regularity of many stateless peoples living on the borders of states. Thus Robert Montagne’s classic thesis on Berber society in Morocco proposed that “Berber society oscillates between two rival and opposed social forms, between, on the one hand, democratic or oligarchic tribal republics ruled by assemblies or hierarchies of assemblies, and, on the other hand, ephemeral tribal tyrannies, exemplified in modern times by the great Caids of the South.”98 As was the case with the Kachin, the Berbers had no indigenous model of state-making, and their form, when states did first arise among them, was based on the Hellenic model of the states they abutted. To mention just one of many parallel cases, Michael Khodarkovsky’s study of the Kalmyk nomads and the Russian state posits the same oscillation. The nominally ruling lineage, along with the clergy, was devoted to creating a dynasty by hereditary succession and centralized power. Other tribal leaders favored decentralization and “indeterminacy” of succession rules: that is, open ranks. “Two structurally antagonistic tendencies, one propelling the top of the society toward increasing centralization, the other consolidating a separatist tendency, may explain the endless cycles of civil wars so often associated with nomadic society.”99 What Khodarkovsky makes clear, however, is that the centralizing tendency was closely associated with accommodation to the adjacent state. Thus the tsarist regime promoted Kalmyk khans as a form of institutional linkage and control. As it was for the British and imperial China, tribal anarchy was anathema to the tsars. Centralization and autocracy depended on a combination of the power of the tsarist state—including the benefits it could bestow—and the political ambitions of Kalmyk khans.
Egalitarian, acephalous peoples on the fringes of states are hard to control. They are ungraspable. To the command “Take me to your leader” there is no straightforward answer. The conquest or co-optation of such peoples is a piecemeal operation—one village at a time and, perhaps, one household at a time—and one that is inherently unstable. No one can answer for anyone else. Acephaly is therefore, like the “jellyfish tribes” of the Middle East described earlier, itself something of an escape social structure. The logical corollary of acephaly is typically the inability to unite except under very special circumstances (for example, charismatic religious leadership and brief military confederacies). A social structure that thwarts incorporation by an outside state also inhibits crystallization of any internal statelike structure.
What are the material conditions that underwrite such egalitarian social structures? The circumstances of the gumlao Kachin, Lisu, Berbers, and Kalmyks are suggestive in this respect. An open common property frontier seems particularly vital. Just as fixed, inheritable property in land facilitates permanent class formation, a common property frontier equalizes access to subsistence resources and permits the frequent fission of villages and lineages that seems central to the maintenance of egalitarianism. The farther away, in terms of friction of terrain, such peoples live from state centers and the more mobile their subsistence routines—foraging, pastoralism, shifting cultivation—the more likely they are to maintain the egalitarian, stateless option. Enclosure of the commons and encroachment by the state are everywhere a threat to such arrangements.
Much of the logic behind the exceptionally complex checkerboard identities in the uplands and the movement between them is best understood as a strategic choice of position vis-à-vis lowland states. Relative altitude and agro-economic niche are often indicative of such positioning. This perspective is most obvious with identities invented by lowland states to suit their own administrative purposes. Following the mid-Ming dynasty “Yao wars,” those who collaborated and settled under imperial rule became “mín,” or subjects, and those who did not became, by definition, “Yao.”100 The ethnic term meant nothing beyond non-taxpaying hill people; it had, initially, no cultural or linguistic coherence. The term Miao, as we have seen, was similarly often applied comprehensively to all those in an area who were still defiantly beyond the state’s grasp. And, of course, the terms raw and cooked, wild and tame, jungle and house (as applied to the Karen) can be understood as references only to the degree of political submission.
Quite apart from state-applied exonyms, ethnic identities, subdivisions, and even villages—as in gumlao and gumsa Kachin—came to acquire reputations for relative degrees of hierarchy and linkage to states. The highland Akha, according to von Geusau, chose subsistence routines that maximized their autonomy and specifically chose locations that put them beyond easy reach of states and slave-raiders.101
We should distinguish here between state-repelling characteristics and state-preventing ones. They are related but not identical. State-repelling traits are those that make it difficult for a state to capture or incorporate a group and rule it, or to systematically appropriate its material production. State-preventing traits, on the other hand, are those that make it unlikely that a group will develop internally durable, hierarchical, statelike structures.
The state-repelling features we have repeatedly encountered in the foregoing analysis can be summarized in general terms. First, a society that is physically mobile, widely dispersed, and likely to fission into new and smaller units is relatively impervious to state capture for obvious reasons.102 These features are, in turn, highly correlated, not to say mandated, by the choice of subsistence routines. Foraging, hunting, and gathering (land-based or maritime) encourage mobility, dispersal, and fission. One can easily stipulate a series or gradient of declining mobility, dispersal, and fission that moves from foraging to swiddening and then to fixed-field crops and irrigated rice. For crop-planting societies, as we saw in Chapter 6, versatile, unobtrusive root crops of staggered maturity are far more state repelling than aboveground grain crops of synchronous maturity. Outside Southeast Asia, the series would also include nomadic pastoralism, with its great advantages in mobility and dispersal.
A third state-repelling feature is a highly egalitarian social structure that makes it difficult for a state to extend its rule through local chiefs and headmen. One of the key material conditions of egalitarian structure—necessary but not sufficient—is open and equal access to subsistence resources. Common-property land tenure and an open frontier are, in this respect, the material conditions that underwrite egalitarianism. In fact, the two major state-repelling subsistence routines, foraging and swiddening, both of which promote mobility and dispersal, are virtually unthinkable without an open, common-property frontier. Its disappearance is a mortal blow to autonomy.
A final state-thwarting strategy is distance from state centers or, in our terms, friction-of-terrain remoteness. Until nearly the twentieth century, remoteness alone sufficed to put some groups definitively outside the reach of the state. As a distance-making strategy, remoteness could, in fact, substitute for other state-repelling strategies. The Hani and Ifugao could safely grow irrigated rice in their remote highland terraces precisely because they were at such a great remove from state centers.
Certain peoples have for so long manifested these state-repelling characteristics that the invocation of their very name conjures up statelessness—often glossed by nearby states as “wildness” or “rawness” or “barbarity.” The Lahu, Lisu, gumlao Kachin, Akha, Wa, Khmu, and Hmong, to mention a few, largely fit this description. Providing that one allows for variation over time and the various subdivisions of many ethnic groups, one could, if so inclined, devise something of a nominal scale of state-repelling characteristics along which any particular group might be ranked.
The other pole of the scale would be anchored by what might be called “state-adapted” characteristics: densely settled, sedentary, grain-growing societies marked by property in land and the disparity of power and wealth that it promotes. Such characteristics are, of course, socially engineered into state space. The peoples manifesting these state-adapted features and therefore indelibly marked by their “stateness” are the Shans, Burmans, Thais, Mons, Pyus, Khmer, and Kinh/Viet. To paraphrase Fernand Braudel, not all the human traffic in the world moving back and forth between these poles is likely to erase these indelible associations. At the stateless extreme we get dispersed, mobile foragers or small clusters of people along remote ridges far from any state center; at the other, taxpaying, padi-planting peasants near the state core.
What is surely most important in this ethnic positioning vis-à-vis the state is the constant movement of individuals between these positions and, indeed, the shift over time in what, say, the position “Karenni,” “Lahu-nyi,” or “Kachin” might mean. At any one place and time, historically, the ethnic identities on offer might be seen as a bandwidth of possibilities for adjusting one’s relationship with the state—a gradient of identifications which may be, over time, fitted to the prevailing economic and political conditions. To be sure, it makes eminent economic sense for padi planters to drop everything and take up foraging when the price of resins, medicinal plants, or edible birds’ nests shoots up. But the move to foraging can as easily occur because it is a state-evading strategy. Similarly, the choice between padi planting and swiddening is more likely to be a political choice than a mere comparative calculation of calories per unit of labor. Insofar as the choice of subsistence routines, altitude, and social structure are associated with particular cultural identities and a “positionality” vis-à-vis the lowland state, then a change in ethnic identity may represent, first and foremost, a political choice that just happens to carry with it implications for cultural identity.103
Some Lahu, for example, have moved to remote mountains and to foraging and, on other occasions, to settled village life and cultivation. As recently as 1973 many Lahu left Kengtung, Burma, for the hills, following a failed revolt against taxation and corvée imposed by the Burmese regime.104 The Khamu have a comparable, though less rebellious, history; some have abandoned village life for foraging at times, and some have moved to the valleys to become Buddhist padi planters.105 And of course, as Leach discovered, many Kachin had been moving between different social forms, each of which expressed more of a positioning vis-à-vis the Shan valley state and hierarchy than it did any momentous cultural shift. To reiterate what should now be obvious, swiddening and foraging as practiced for the past few centuries in Southeast Asia are not prior to padi planting in some scheme of social evolution; they are, instead, “secondary adaptations,” indicative of a largely political choice.106
Jonsson has noted astutely that “ethnic distinctions may primarily have to do with lowland affiliations.” In this respect, he claims that “Ethnic groups do not have a [determinate] social organization,” by which he means to imply that a particular named identity can, in terms of subsistence, cultural affiliations, internal hierarchy, and above all its relationship to lowland states, vary widely.107 In other words, not only are individuals and groups moving between ethnic identities as a consequence of positioning themselves, but these identities themselves are labile, as the aggregate of decisions their bearers take has the effect of repositioning the very meaning of that ethnic identity.
If hill peoples have at hand a bandwidth of identities that they can take up, and if each of these identities calibrates a different relationship to lowland states, what can one say about historical trends? Here the shift in the past half-century has been momentous. Until then, as we have seen, Zomia was largely a zone of refuge for societies and fragments of societies fleeing or choosing to place themselves beyond the grasp of valley states. The mosaic of named ethnic identities is testimony to a long and complex history of migrations and remigrations marked by rebellions, war, and cultural reformulations. Originally, much of Zomia’s population came from the lowlands, particularly from China, and, however they came by their ethnic name, they kept it largely because they had left the realm of state power. Those who stayed—perhaps a majority—became part of the lowland cultural amalgam and were no longer called Miao, Yao, or Tai. This history, together with the exceptional ecological diversity and the geographical isolation of the region, has produced perhaps the largest mosaic of relatively stateless peoples in the world.
For the past half-century, however, the gradient of available identities has, as it were, been radically tilted in favor of various degrees of state control. The classical narrative of “raw” barbarian peoples being brought to civilization has been replicated by a narrative of development and nation-building. While the older narrative was, owing to the limitations of state power, more an aspiration than reality, the new narrative is more imposing. At least three factors account for this. First, the modern idea of full sovereignty within a nation-state and the administrative and military wherewithal to effect it means that little effort is spared to project the nation-state’s writ to the borders of adjacent states. Zones of overlapping, ambiguous, or no sovereignty—once virtually all of Zomia—are increasingly rare. Second, the material basis of egalitarian, acephalous societies—common property in land—is increasingly replaced by state allocation of land rights or individual freehold tenure. And finally, the massive growth of lowland populations has prompted a massive, growing, state-sponsored or abetted colonization of the hills. The colonists bring with them their crops, their social organization, and, this time, their state. The result is the world’s last great enclosure.