As with everything he writes, James C. Scott’s The art of not being governed is original, articulate, ambitious, provocative, iconoclastic.

Earlier scholars often viewed the forested highlands of mainland Southeast Asia and the interior of many Southeast Asian islands as substantially self-sufficient worlds. Of course, they knew that, for centuries, upland interiors had traded forest exotica to the plains and coasts in return for salt, iron, and coastal imports. They also appreciated that tribal peoples redeployed religious and secular motifs from their literate neighbours in silent homage to the power of plains/coastal civilization. And they recognized that, in favoured areas of Southeast Asia in the early and mid first millennium CE, sedentary states had emerged from a tribal milieu.

Yet if upland/interior peoples and lowland/coastal populations shared a common ancestry and interacted periodically, such features paled besides what was assumed to be their growing divergence between c.500 and 1800. Whereas lowland societies, battening on maritime trade, agrarian intensification, and imported cultures, grew ever more wealthy, specialized, and cosmopolitan, tribal populations essentially stagnated. This was true, some argued, of languages and social forms; but atavism also characterized interior swiddening (shifting agriculture), illiteracy, and statelessness. In a word, interior peoples remained outside history, museum reminders of how the ancestors of modern lowlanders probably lived two thousand years ago. Many scholars further assumed that, in genetic terms also, physical isolation ensured that populations in the mountains of the mainland and in island interiors remained substantially self-contained.

All of this Scott, expanding on the insights of anthropologists such as E. R. Leach and F. K. Lehman, has sought to turn on its head. His test case is that vast upland zone – the area above 300 m – that extends from the central highlands of Vietnam to north-east India and that includes parts of all five mainland Southeast Asian states (Burma, Thailand, Laos, Cambodia, and Vietnam). Following Willem van Schendel, Scott terms this inland massif ‘Zomia’.Footnote ¹ In particular, he focuses on the historic relation between the kingdom of Burma and the highland horseshoe that encompasses the Irrawaddy basin on the west, north, and east.

Scott’s basic thesis is that highland societies, far from living in isolation, have been profoundly and continuously moulded by their relation to plains-based kingdoms. Anything but innocent or unfamiliar with state projects of control and extraction, hill peoples experienced those projects in the most brutal and direct fashion – and, in response, self-consciously pursued defensive strategies of avoidance, dispersion, economic minimalism, and political decentralization. Set against the overwhelming literature on state-making, this may be the first extended study, not merely for Southeast Asia but for any world region, of the reverse phenomenon: ‘deliberate and reactive statelessness’.Footnote ²

Ultimately, Scott argues, valley–upland interactions were governed by Southeast Asia’s chronically low population densities (perhaps only one-sixth those of premodern China or India). Of necessity, the political economy of Burmese and other mainland states focused on securing not land or commercial wealth but manpower. The practical implications of this preoccupation were threefold. First, each kingdom strongly encouraged irrigated rice (padi) at the expense of dry field and swidden crops, because the unique productivity of wet rice per land unit allowed the state to maximize the number of subjects residing in a given territory. Burmese, Siamese, Khmer, and Vietnamese capitals typically lay in lowland districts with the most intensive padi cultivation, hence the most concentrated population. Second, precolonial administrations tried to ensure labour services and taxes by maintaining elaborate records and by inhibiting movement from authorized to unauthorized service units and from districts around the capital to less easily monitored areas. While at the outset such systems proved moderately effective, over the long term they tended to decay because the state lacked adequate local information and because political rivalries led to excessive and insufficiently flexible demands.Footnote ³ In periods of military crisis, the gradual seepage of manpower became a torrent, as frantic subjects sought to escape escalating royal demands by any means possible. Third, to preclude catastrophic losses of this sort and to correct for constant erosion, mainland states sought to augment their manpower base by encouraging voluntary immigration from afar and, more especially, by acquiring slaves and prisoners for resettlement. Ultimately, Scott argues, no kingdom flourished except by slave-raiding and deportations.Footnote ⁴ The Southeast Asian padi state was therefore a ‘centripetal population machine’,Footnote ⁵ but at the same time it was ‘a fragile and evanescent affair’, subject to continuous oscillation between phases of manpower dispersion and concentration.Footnote ⁶

This peculiar political economy transformed the mountainous interior through the twinned processes of culling and extrusion. On the one hand, Scott argues, chronic manpower hunger led Burmese and Tai states to target hill populations, directly through imperial expeditions, and indirectly by providing a ready market for slaves. Typically strong hill peoples, such as the Karenni and Kachins, enslaved weaker groups, such as the Akha and Lisu. They sold some of their victims to Tais in upland valley kingdoms, who in turn resold them to the Siamese and Burmese farther south (one is reminded of indigenous intermediaries in the African slave trade). In the late 1800s, fully three-quarters of Chiang Mai’s population, for example, were said to consist of war captives, a significant proportion of whom originated in the hills.Footnote ⁷

On the other hand, state oppression and warfare – coupled with periodic famines and epidemics characteristic of thickly settled lowland zones – drove subjects from imperial cores into the hills. Again, state evasion may have been a chronic low-level phenomenon but, in periods of military and dynastic crisis, it reached substantial proportions. Over the long term – that is to say, during much of the late first and second millennia – Scott claims that these processes transformed Southeast Asia’s mountainous interior into a vast ‘shatter zone’, an area of flight, a sphere of asylum and marronnage for runaways from state-making projects in the plains.Footnote ⁸ Meanwhile, on Zomia’s north-eastern flank, the Chinese empire also operated a vast winnowing machine. Unlike their Southeast Asian counterparts, late Ming and Qing imperial cores typically had not a deficiency but a surplus of population, which encouraged systematic frontier colonization in what is now south-west China. However, by displacing to higher elevations and more distant locales indigenes who were unable or unwilling to submit to state control and to become Chinese, Han colonization – like Southeast Asian state projects – encouraged flight from ‘civilization’ and state authority into the hills.Footnote ⁹ Thus Zomia was peopled through simultaneous pressures from the south and the north-east. While the mountainous areas that refugees entered were by no means empty, Scott argues that the cumulative demographic weight of state-evading migrants and their descendants overwhelmed earlier inhabitants.Footnote ¹⁰

Obviously, in a crude physical sense, violent deportations from the hills and large-scale immigration to the hills were contradictory. In a deeper social and psychological sense, however, these processes were entirely complementary. Both dynamics ensured that hill populations were constantly losing members, assimilating newcomers, fissioning, fusing, shifting location, shedding familiar ethnic labels, and adopting new identities. In such conditions, the old scholarly image of stable, bounded, self-replicating ‘tribes’ is untenable.Footnote ¹¹ At the same time, Scott argues, both deportations and immigration combined to instil throughout Zomia an abiding suspicion of state projects of all sorts. Traumatized survivors of both upland slaving expeditions and lowland crises (and descendants of those survivors) vowed to keep their distance from valley polities. Indeed, Scott suggests, this attitude underlay a widespread hostility to chiefly pretensions and to institutionalized inequality within hill societies themselves. Of course, avoidance of lowland dangers was not the same as isolation. We should recall that hills and valleys engaged in mutually profitable trade and that lowland cultural motifs circulated freely in the mountains. Scott carefully explores the attraction that valley rituals and religious symbols held for upland peoples and the ways in which those imports were refined to fit local understandings. Moreover, in periods of lowland quietude and prosperity, some hill peoples, attracted by the wealth and cultural éclat of the valleys, voluntarily left the hills. Yet those who remained carefully sought to structure lowland contacts so as to maximize their autonomy and minimize valley impositions.

In Scott’s formulation, therefore, an abiding vulnerability, a visceral angst governed diverse aspects of life across Zomia. Social and cultural strategies sought one end: to avoid being governed. In residential and geographic terms this meant not only that plains refugees streamed into the hills but that, within that rugged terrain, people scattered so as to minimize exposure to lowland military forces and private slavers. Friction of terrain, as Scott terms it, provided the surest protection against state capture.Footnote ¹² In turn, physical dispersion favoured economic regimes of hunting-foraging and, more especially, slash-and-burn cultivation, which involved regularly clearing new land and occasionally shifting settlement sites. The attraction of swidden crops – including sweet potatoes, cassava, yams, taro, and maize, all of which Scott terms ‘escape crops’ – derived from their low labour requirement per land unit, from the relative invisibility of root crops to would-be attackers, and from the fact that some of these crops could remain in the ground for relatively long periods until harvested. Here was an agricultural surplus that, compared to wet rice, was exceedingly difficult for raiders to find, much less assess, much less appropriate. Swidden was not an atavistic survival. Hill peoples knew how to grow wet rice and, in a few unusually valuable and/or isolated locales, did so. Rather, Scott argues, swidden agriculture was primarily a self-protective political choice.Footnote ¹³

The political counterpart to state-repelling dispersion and escape agriculture was a radically decentralized, egalitarian social order, over which predatory states could not easily exert control by enlisting local chiefs and headmen. Seeking to emulate nearby Shan polities, with their hereditary kingship, corvées, and taxes, some ambitious Kachin (and less frequently Chin and Karen) chiefs did attempt to create stratified, state-like structures in the hills. But, over the long term, such projects invariably collapsed in the face of assassination, revolt, and flight by defiant subjects who were unwilling to provide their chiefs with corvée labour or to accept social subordination. This too, Scott claims, was a political choice, rooted in historically grounded fear of external domination. Anarchic, acephalous, Lilliputian republics offered outsiders no handle with which to gain control or impose tribute.Footnote ¹⁴

Scott goes so far as to argue that the uplanders’ preference for non-literate, oral cultures and for fluid ethnic identities was cut from the same cloth as physical dispersion, escape agriculture, and escape political structures. Unconfined by written texts or recorded precedents, hill populations – ranging from fissiparous confederations to fragile hamlets to footloose families – could alter their histories, genealogies, languages, and group identities as shifting circumstances demanded. In thus maximizing their room for cultural manoeuvre – their freedom to define and redefine themselves – interior populations again frustrated the state’s aspiration for legibility, fixity, and routine. In this polymorphous fluidity, hill peoples, in Scott’s phrase, created ‘jellyfish’ societies whose motto, in effect, was ‘Divide that ye be not ruled.’Footnote ¹⁵

Scott’s central achievement, then, is to bring hill peoples into the mainstream of regional history by uncovering their relation to lowland states and societies. For several centuries, he argues, intertwined forces of repulsion and attraction influenced both lowlands and uplands and led to a continuous circulation of genes and to a constant refashioning of hill society. In exploring these dynamics, Scott has rescued hill peoples from assumptions of stasis, primitivism, essentialism, and isolation. He has given them voice, agency, and rationality. Thus he has sought to do for hill peoples what his earlier seminal research did for Southeast Asian peasants. Previous scholars of upland tribes had anticipated some of his theoretical concerns, as Scott freely acknowledges. But he is the first to offer a regional overview, to conceive of Zomia as a vast shatter zone formed by a historic pincer movement from both Chinese and Southeast Asian states, and to analyse ostensibly unrelated economic and social phenomena in the hills as common responses to lowland pressure.

Finally, so far as I know, Scott alone has analysed external state-repelling, and internal state-avoiding, strategies as generic phenomena characteristic of frontiers around the globe. Others have described Paris’ cultural incorporation of its provincial periphery as a form of ‘internal colonialism’, but Scott is perhaps the first to apply that concept to areas beyond the state’s formal administrative control.Footnote ¹⁶ Thus he likens the peoples of Zomia to such refugees from internal colonialism as peasants on the Cossack steppe; Berbers in the Atlas Mountains; and fugitive slaves and/or indigenous peoples in interior Brazil, Surinam, and Jamaica, in the Great Lakes region, in the swamps of Florida, and along the Virginia–North Carolina border.Footnote ¹⁷ Not unlike E. R. Leach’s Political systems of highland Burma,Footnote ¹⁸ which helped to inspire Scott, this book will influence Southeast Asian and Developing World anthropology and historiography for years to come. The reader’s enjoyment is enhanced by Scott’s lucid, vigorous, often amusing style (reading his Preface, I found myself laughing aloud).

But if The art of not being governed has many obvious and outstanding virtues, it seems to me that it also has serious flaws. Its evidential base is often too weak to support its theoretical superstructure. Despite or because of its claims to be ‘an anarchist history’, its view of the Southeast Asian state is curiously ahistorical. Most critically, in analysing upland behaviour, the book leans to monocausal explanations focused on lowland provocation, to the virtual neglect of dynamics within the highlands themselves. I will examine these issues in turn.

Scott’s documentation on pre-1850 Southeast Asia is generally limited, but this is particularly troubling for Burma, because he chose that realm as his primary case study. Although his Preface shows that he studied Burmese, I see no indication that he has read Burmese-language historiography or precolonial Burmese documents.Footnote ¹⁹ Nor is there evidence that he used more than a tiny fraction of Burmese documents available in translation,Footnote ²⁰ or consulted certain valuable Western-language studies of precolonial Burmese/Southeast Asian demography and institutions.Footnote ²¹

The resultant problems are of three types. First, factual errors that do not touch Scott’s main thesis directly nonetheless erode his credibility as a guide to precolonial society and history. Some examples: he writes that literacy in precolonial mainland Southeast Asia and China, as in all ‘premodern societies was, under the best of circumstances, confined to a miniscule portion of the population, almost certainly less than 1 percent’.Footnote ²² In fact, male literacy in Burma in 1800 was at least 50%; in Siam, 20–50% or more; in Vietnam, perhaps 20%; and in China, between 15% and 50%.Footnote ²³ Chronological mistakes are also surprisingly frequent.Footnote ²⁴ He tends toward an Orientalist, dismissive view of the indigenous state, denying it a linear institutional dynamic and reversing the proper relation between regna and interregna.Footnote ²⁵ His claim that Srivijaya and Melaka yielded to ‘more manpower-heavy agrarian states’ is manifestly inaccurate.Footnote ²⁶ And so on.Footnote ²⁷

Second, and more germane to Scott’s central theses, his emphasis on manpower as the sole factor in an unchanging political economy is overdrawn. He claims that, because large armies were the only means by which wealth could be acquired, manpower was the only independent political variable.Footnote ²⁸ Surely, however, it is more accurate to see the relation between labour and commercial wealth as reciprocal: powerful armies could seize trade outlets, but trade revenues and contacts – by providing skilled mercenaries, more effective weapons and ships, and the wherewithal to support forces in the field – could enhance military strength. To some degree, both resources were always attractive. But as international trade expanded in the sixteenth century and again in the late 1700s and 1800s, the strategic value of maritime resources and the appeal of maritime strategies grew apace, along with the need for institutional experiment. Thus between c.1620 and 1771 maritime networks and administrative innovation helped the Nguyen state in southern Vietnam to withstand the Trinh state in the north – even though the Trinh was at least seven times more populous. From 1771 to 1802 first the Tayson movement and then the revived Nguyen movement – both based in the least populated sector of Vietnam, but both benefiting once again from maritime contacts – went on to subdue the entire country.Footnote ²⁹ In fact, those victories had been adumbrated as early as the fourteenth and fifteenth centuries, when maritime contacts allowed the coastal-oriented realms of Pegu, Ayudhya, Phnom Penh, and Champa to defy, and in some cases ravage, the far more populous, but less commercially developed empires of Upper Burma, Angkor, and Dai Viet.Footnote ³⁰ In their epic wars of 1548–99 and 1759–1811, the Burmese and Siamese, as Scott notes, sought to weaken one another through large-scale deportations. But the combatants also had strong commercial objectives, while Siam’s ultimate success over its more populous rival reflected, most basically, its creative institutional and fiscal response to the lure of Chinese trade.Footnote ³¹ Likewise, the nineteenth-century surge in Siamese slaving expeditions in the hills of Laos responded less to an age-old imperative than to novel economic conditions, namely, the opening of Singapore and rising Chinese prices for forest and plantation goods, whose procurement required additional labour.Footnote ³²

Lest one exaggerate the political import of population deficiency per se, note finally that imperial collapse on the mainland frequently reflected not loss of population but rather overpopulation relative to land. To follow Aung-Thwin’s classic work on Pagan and recent research on Angkor, both of those great charter polities collapsed in the late thirteenth and fourteenth centuries, in large measure because they ran out of good cultivable land.Footnote ³³ Political crises in northern Vietnam in the fourteenth, sixteenth, and late eighteenth centuries were also caused or intensified by severe land shortages.Footnote ³⁴

In short, Scott’s vision of the Southeast Asian state as a ‘centripetal population machine’ following an iron imperative to augment and replenish scarce manpower minimizes or ignores four issues: the inconstant, situationally specific nature of demographic pressures; commercial spurs to aggression that were substantially independent of domestic demography; the often critical role of non-demographic factors in determining military outcomes; and sustained changes in institutional as well as economic context. He strives, with sensitivity and sympathy, to introduce agency, nuance, and contingency to the history of hill peoples; for their putative nemesis, the lowland state, however, he performs no similar service.Footnote ³⁵

These considerations of political economy lead to a third, yet more basic, set of problems. Evidence for massive, sustained flight from the lowlands to Zomia is thin. Evidence for deportations from Zomia to the lowlands is better, but still limited. And the implications of both types of displacement are ambiguous.

Scott’s thesis that the so-called hill tribes of mainland Southeast Asia ‘are best understood, at least as a first approximation, as a fugitive population that has come to the hills over the past two millennia’Footnote ³⁶ makes most sense for those groups, organized into lineage societies, that fled south and west to escape Chinese settlement and state pressure. The most recent examples are the Mien, Lolo, Lahu, Lisu, Akha, Hui, and many in the Hmong subgroup of the Miao, who are now spread across the uplands of northern Thailand, Vietnam, and Laos, where they comprise, respectively, about 44%, 27%, and 19% of minority highland peoples.Footnote ³⁷ The problem, however, is that, while non-Tai migrations from what is now China into the Southeast Asia massif may well be quite ancient, major movements seem unlikely before the mid or late Ming, when Chinese settlement of the south-west, aided by the sixteenth-century introduction of New World maize, accelerated sharply. Such colonization, and attendant Ming and then mid Qing ‘pacification’ campaigns, must have forced local peoples unwilling to accept assimilation, subordination, or annihilation to move to higher elevations or farther south into southern Yunnan and perhaps into northern upland Southeast Asia. But in Western literature at least, such movements into Southeast Asia are poorly documented.Footnote ³⁸ Scott provides no hard evidence before 1807 of flight from south-west China to the Southeast Asian uplands, as opposed to the uplands of China itself;Footnote ³⁹ and, in fact, the earliest large-scale, well-documented non-Tai exodus to the Southeast Asian massif occurred in the second half of the nineteenth century, in the aftermath of the Taiping (1851–64), Miao (1854–73), and Panthay (1855–72) rebellions against the Qing.Footnote ⁴⁰ Almost certainly, before 1850 the proportion of identifiable refugees from South China among the hill peoples of what are now Thailand, Vietnam, and Laos was dramatically smaller than it is now. What is more, in upland Burma and Cambodia such refugees seem to have been far fewer than in the rest of the mainland. Today they are only 3% of highland minorities in Burma, and are virtually nonexistent in Cambodia.Footnote ⁴¹

If we are going to substantiate the hypothesis that for 2,000 years the hills of Southeast Asia became a shatter zone for waves of subjects fleeing valley oppression, the inadequacy of current Chinese evidence obliges us to turn to Southeast Asia. Scott provides the following evidence from that region. First, between c.900 and 1400 the Ganan people, who may have been part of the Pyu kingdom in the lowlands of Upper Burma, may have fled Nanzhao and Burman attacks by entering the hills in the upper reaches of the Mu River system.Footnote ⁴² Second, in the early nineteenth century a Burmese invasion of Assam led Shans and Kachins around Mong Hkawn, a small town in the upper Chindwin valley, to move to higher ground.Footnote ⁴³ Third, Siamese invasions apparently pushed lowland farmers from the plains around Kengtung to become swidden cultivators in the surrounding hills.Footnote ⁴⁴ Lastly, Siamese incursions in the early 1800s also drove a Karen group to an inaccessible mountain gorge between Saraburi and Khorat.Footnote ⁴⁵

To this modest inventory of flight to the hills I would add the following. Along the eastern littoral in the eighteenth century, Vietnamese colonization obliged minority peoples to enter, or retreat further into, the western mountains. Vietnamese pressure in turn continued Cham expansion, which in earlier centuries had forced Mon-Khmer speaking Ban-na (Mada) peoples to scatter into the highlands.Footnote ⁴⁶ In Laos also, some 200 years ago a Tibeto-Burman-speaking group known as the Phu Noi reportedly sought refuge in the hills from feuds between lowland powers.Footnote ⁴⁷ Indirect linguistic evidence and oral tradition (although the latter’s credibility has been questioned) suggest that many inhabitants of the northern Chin hills in western Burma were ‘pushed’ up from the plains at some uncertain time.Footnote ⁴⁸ In the 1590s, as royal tax and service demands at Pegu escalated, Jesuits reported that desperate subjects sold themselves as slaves or ‘went to hide in the forests’.Footnote ⁴⁹ Again at Pegu, in 1740, an Englishman noted that people were ‘chusing to live among the Wild Beasts, [rather] than be at the Mercy of the … tyrannical Government’.Footnote ⁵⁰ (Yet it must be said that neither of the latter references specifies upland as opposed to lowland jungle.)

All in all, outside central Vietnam perhaps, this remains a rather limited record of displacement and flight. Of course, not only is my secondary survey far from systematic, but the primary record is fragmentary. Local conflicts in which the losers fled to the hills often lay outside the centre’s official gaze, while valley disdain for upland peoples also circumscribed official reportage. But as I have shown elsewhere,Footnote ⁵¹ in Burma – whose situation I know best – such factors hardly eliminated official reportage on the uplands, because those areas keenly interested the crown as a site of potential disturbance as well as of refuge. The fact that the uplands did not appear in official documents more often than they did must reflect, at least in part, the fact that the hills did not attract large or continuous refugee flows. The Burmese court was indeed obsessed with population movements.Footnote ⁵² However, the vast majority of unauthorized shifts by people seeking easier conditions were from one service category to another, from royal service to private service or slavery, or from one lowland district to another. Very few, on present evidence, fled to the hills.

Note, finally, that among those who did flee, many were members of ethnic minorities who sought to escape not state pressure per se – typically they lived in marginal districts where state authority was minimal – but encroaching settlement by members of the Burman majority, who often operated without direct state support.Footnote ⁵³ In his preoccupation with state domination, Scott conflates official impositions with popular conflicts. Conceptually and in practice, the two could be quite distinct.

To the extent that Burmese experience was representative of other Indic societies, one must therefore treat with some scepticism Scott’s claim that in recent centuries emigration, especially state-induced emigration, from the Southeast Asian lowlands radically transformed the demography of the massif. In the view of many scholars, not only the Chin hills, the Kachin hills, the Shan hills, and the Burmese–Thai interface, but also northern Thailand, upland Cambodia, upland north Vietnam, even the Vietnamese cordillera have been reasonably well populated by ethnically fluid and fragmented swidden cultivators since at least the first millennium.Footnote ⁵⁴ In weighing the potential demographic impact of lowland emigration, we should consider that, all things being equal, the superior nutrition and per-man-hour yield of swiddening (to which Scott himself draws attention) would have supported rates of population growth in Zomia that compared favourably to those in the lowlands.Footnote ⁵⁵ Indeed, insofar as some plainsmen did enter the hills, the appeal of swidden agriculture may have been a pull factor no less influential than the push of lowland oppression. This only reinforces our doubts that flight by traumatized victims of lowland brutality in general, and of state brutality in particular, was sufficiently frequent or large-scale to transform Zomia into a vast shatter zone, a realm of refuge, gripped by a morbid dread of valley impositions.

If flight-inducing brutality and disorder in the valleys were limited, what can we say about Scott’s second grand dynamic inspiring upland alienation? As we have seen, this consisted of slave raids and other impositions on people already residing in the hills that were sponsored directly or indirectly by lowland states. Evidence that such assaults contributed to interior strategies of state avoidance is somewhat more persuasive.

Oral literature in the hills typically associated plainsmen with chicanery, violence, and despoliation. So subject were the Chins, for example, to Burmese ill-treatment that traditions accounting for this formed, in the words of one observer, ‘a part of the national religion’.Footnote ⁵⁶ Inscriptions from Pagan and Angkor and a Chinese account of Angkor show that in both charter societies the labour force included a sizeable element of enslaved ‘savages’ from the ‘mountainous wastelands’.Footnote ⁵⁷ In a later period, in Hkamti district in the far north of modern Burma, a significant portion of the Shan population were descended from captured hill peoples.Footnote ⁵⁸ For a century after 1778, Siamese wartime losses and rising demands for plantation labour led to devastating Siamese slave raids on Laotian hill peoples, accompanied by massacres of those who resisted.Footnote ⁵⁹ In Nguyen Vietnam, not only did trade in enslaved hill peoples (long a feature of the Cham political economy) continue, but disproportionately heavy taxes on uplanders helped to precipitate the Tay Son uprising of 1771 that overthrew the Nguyen polity.Footnote ⁶⁰

In many contexts, clearly, the ensuing distrust of the lowland state and of the wider lowland society impelled hill peoples to act much as Scott claims. To escape slaving expeditions that were organized either by valley kingdoms or by other hill tribes, many of whose victims were sold to such kingdoms, weak interior peoples retreated to ever higher, more inaccessible locales. They opted for swiddening and foraging at least in part because such strategies minimized visibility and exposure. Likewise upland peoples, or at least their non-aristocratic lineages, thwarted the transformation of chieftainships into supra-local principalities, partly because they feared that local rulers might collaborate with Shan kingdoms and allow the latter kingdoms to climb the hills, as it were. Through enslavement, trade, and travel, Kachins, Was, Akhas, Lahus, Chins, and Karens knew something about lowland sociopolitical systems. To join such societies, even as voluntary migrants, let alone as slaves, would oblige them to be incorporated at its lowest rungs. Better by far, many concluded, to be a successful hill-dweller than a defective Shan or Burman. Thus, even as many envied lowland wealth, uplanders typically defined themselves by the express absence of lowland traits: wet rice, literacy, kings, institutionalized hierarchy.Footnote ⁶¹

Yet if reactive strategies of avoidance and statelessness explain part of upland behaviour, they hardly explain everything. For one thing, though here Scott can hardly be faulted, the problem of quantification remains. It is very difficult to know how frequent or demographically significant slaving expeditions and other assaults on the hills were, especially in the post-charter era. The ferocious Siamese raids in Laos after 1778 may have been an atypical response to Ayudhya’s depopulation and Chinese commercial demands. I see no comparable evidence in Burma of imperial assaults on the uplands.

But the larger problem is this: Scott’s explanatory dynamic neglects the hills themselves. The great virtue of his book is its unification of lowland and upland history through a lowland-focused developmental model. Its principal weakness, in my view, is the flip side of that strength, namely, its relentless lowland-centred monocausality and functionalism. Scott sees virtually all highland behaviour – physical dispersion, economic routines, statelessness, ethnic fluidity, oral illiterate culture –through the prism of lowland state oppression. As the book’s subtitle advertises, his is an anarchist, anti-state history. But this means that ecological and cultural conditions that were intrinsic to the hills and that were substantially or completely divorced from the valleys receive little or no attention. Therein lies a certain irony. Scott implicitly applauds the uplanders’ urge to freedom from valley domination. But in tightly tethering hill society to a lowland-centred dynamic, he robs that society of functional independence.

Consider, for example, the issues of swidden cultivation and dispersed residence. Scott concedes that hill residents faced a series of choices. They could leave the hills entirely to enter the Tai-, Burman-, or Vietnamese-dominated lowlands. Although, as just noted, this was instinctively unattractive to the great majority of uplanders, some did so, especially in periods of lowland prosperity when the valleys seemed to beckon. If they remained in the hills, Scott suggests, they also had a choice between wet-rice agriculture, on the one hand, and pursuing some combination of swiddening, small-scale wet-rice cultivation, foraging, and hunting, on the other. Scott points to wet-rice terraces built by the Hani in the upper reaches of the Red River in northern Vietnam, by the Ifugao in northern Luzon, and by some Karen to show that large-scale irrigated rice in the hills was a viable option. That uplanders by and large rejected wet rice in favour of dispersed fields and invisible root crops reflects, again in his view, their pervasive fear that valley foes would tax and confiscate their principal resource if it lay exposed.Footnote ⁶² In fact, however, rice terraces on steep hillsides required considerable investments of labour; the benefits per labour unit were typically less than for swiddening. Terraces made sense chiefly where population growth had reduced opportunities for extensive swiddening or where unusual commercial opportunities – such as the chance to control trade and tolls at a key mountain pass – required a sizable local military garrison that only irrigated padi could support.Footnote ⁶³ Elsewhere, swiddening was the logical choice for perfectly sound economic reasons. As just indicated, its caloric yield per man-hour often exceeded that of irrigated rice. Polycropping and staggered maturities protected against the failure of any particular crop and ensured supplies throughout the year. Crop diversity also enhanced nutrition, while population dispersal reduced human epidemics and facilitated protein acquisition through hunting. If this economic routine, then, was overdetermined, why privilege lowland-centred political pressure?

Once interior peoples had committed to swiddening, that commitment, along with the fragmentation of upland topography, tightly constricted their political options. Even more obviously than economic routine, statelessness was less a matter of strategic calculation vis-à-vis valley kingdoms (and other hill peoples), as Scott would have it, than of brute physical restriction. Without an extensive padi base, manpower simply could not be concentrated, and without a concentrated stable population, state taxation, corvées, and military forces were implausible. Only rarely in the hills could trade compensate for demography in the fashion of Malay coastal states. No less critical, without a riverine artery, central control over outlying dependencies could not be enforced with any regularity.

A similar logic influenced oral culture and ethnicity. Scott may be correct that a desire to remain invisible to valley tax collectors strengthened hill peoples’ inclination to eschew writing. But, whatever their inclination may have been, surely the primary reason that uplanders could not read is that they lacked the means, institutional supports, and incentives to literacy that we find in the lowlands. That is to say, they had no state system of record-keeping and no state-centred tradition of religious and literary patronage. They lacked the wealth and population densities needed to sustain a network of monasteries-cum-schools, on which mass male education in the lowlands rested. And the spurs to literacy and numeracy that derived from urbanism and dense commercialization were anaemic to non-existent.Footnote ⁶⁴

By the same token, although labile identities may have let hill people escape restrictions as they moved between and around states, in the most basic sense porous, fluid identities arose from the hills’ irreducible social fragmentation and illiteracy. Without a central state to define and police cultural boundaries, without written norms to codify tradition, without dense commercial and communications circuits to promote circulation, the prerequisites for historically stable ethnic expression were quite absent. Across the mountains at large this was less a matter of nimble strategy (although this certainly played a role) than of ecological imposition.Footnote ⁶⁵

Without a supra-local regulatory authority, moreover, there was no reliable way to mediate disputes, which rendered village and family vendettas the most practical means to defend local interests. By most accounts, this left hill societies subject to extraordinary levels of personal violence and vulnerability. Pace Scott’s anarchist inclination to equate violence with state impositions, recent studies argue that conflict among non-state peoples was far more lethal than among settled populations, and that, as a generic phenomenon, statism in fact tended to enhance security. Of the stateless societies of north Sulawesi it has been written that they ‘“lived more in war than in peace”, and in upland areas almost the only way to avoid regular violence was to avoid regular contact with any non-kinsmen by living a life of flight, isolation, and dishonour on the infertile frontiers of settlement’. Nuclear settlements had to be more or less permanently fortified. If violence of this sort was later aggravated by Bugis and Sulu raiders, its origins were entirely indigenous, drawing on religious sanctions often connected to headhunting.Footnote ⁶⁶ Likewise, in the hills of what is now western Burma and north-eastern India, precolonial-era feuding between families, villages, or tribal confederations, accompanied at times by headhunting, was as common as it was destructive.Footnote ⁶⁷ In the Kachin hills, homicide, unjustified enslavement, the wounding of a chief, and disputes over women were all considered justifications for inter-lineage feuds and extended marauding expeditions.Footnote ⁶⁸

Thus autonomous social/cultural dynamics seem to have favoured physical dispersion, swiddening, and defensive strategies at least as persistently as did pressures emanating from the lowlands.Footnote ⁶⁹ Self-replicating local culture may also help to explain the aforementioned pattern, whereby Kachin commoners sabotaged chiefly efforts to convert themselves into Shan-style rulers. Scott sees fears of Shan incorporation at play; but since there were few, if any, precedents to justify such fears, is it not equally plausible to assume that commoners resisted chiefly ambitions because those ambitions, by violating cherished norms of affinal reciprocity, struck at the foundations of Kachin self-image?Footnote ⁷⁰ Similar egalitarian impulses prevailed in the southern Chin hills, whose people knew neither internal state formation nor regular Burmese incursions, and in Minangkabau Sumatra, which again had no experience with credible state impositions.Footnote ⁷¹

In sum, whereas Scott sees hill formations as lowland ‘state effects’, it is no less reasonable to interpret many of those same phenomena as manifestations of a self-sufficient physical and cultural environment peculiar to the uplands. How, then, can we judge between these two approaches? The obvious solution is to juxtapose societies with similar ecologies but different political exposures. That is to say, compare Zomia, subject to pressures from Chinese and Southeast Asian states, to the isolated interior of Southeast Asian islands or to New Guinea – areas that lacked regular state contact. Scott has not tried this, but the following cursory inquiry does not lend strong support to his lowland emphasis.

Consider, for example, interior Borneo/Kalimantan. Although they might have differed in demography and kinship systems, the Dusun peoples, the highland Idahan Murut, the Kenyah-Kayan-Kajang peoples, the Punan-Penan population, Ibans, and Kalimantan Dayaks all typically practised not wet-rice but swidden agriculture, focusing on a mix of dry rice, maize, root crops, and vegetables, supplemented by hunting/fishing and gathering.Footnote ⁷² They lived in forested and/or mountainous terrain, were in some cases residentially mobile, and maintained purely local sociopolitical organizations. Their cultures were exclusively oral and preliterate, and their ethnic/tribal identities were localized, fissiparous, porous, and labile. All these elements they shared with the peoples of Zomia. Yet, whereas in Zomia Scott sees such features as conscious defences against predatory kingdoms, in precolonial Kalimantan these groups were effectively insulated from state contact. The island’s chief coastal polities (Brunei, Banjarmasin, Sukadana, Pontianak) traded with the interior but, in typical Malay fashion, had minimal military reach away from the coast.Footnote ⁷³ Likewise in precolonial Alor, a small island north of Timor, although its precipitous mountain interior was effectively off limits to the narrow coastal plain, swidden mountain cultivators were divided into anywhere from fifteen to fifty ethno-linguistic groups.Footnote ⁷⁴ In northern Sulawesi, while coastal–interior contact was marginally easier than in much of Kalimantan or Alor, we find a similar interior constellation of widespread swiddening, kaleidoscopic cultural/linguistic diversity, and political localization. Confederacies of intermarrying villages were the largest units capable of political solidarity, and chronic internecine conflict, as noted, was embedded in ritual and religion.Footnote ⁷⁵ These ‘jellyfish’ societies were divided, certainly, but not because they feared being ruled from without.

New Guinea shows yet more dramatically that Zomia-like social and cultural traits could flourish without the stimulus of state predation. Inhabiting a fragmented mountain, lowland, and swamp terrain, the population of Papua New Guinea (the eastern half of New Guinea) alone still has over 700 often ill-defined ethnic groups, some 200 Austronesian languages, and perhaps 550 non-Austronesian languages. Indeed, New Guinea as a whole has by far the highest concentration of languages in the world. If some intermontane valleys in the central cordillera supported intensive sweet-potato cultivation from the same plots for relatively long periods, virtually everywhere outside the swamps swiddening, supplemented by hunting and gathering, formed the basis of the premodern economy. Taro, yams, cassava, and sweet potatoes were standard, but rice was conspicuously absent. Without stable supra-local political structures, endemic violence and low-level warfare between villages, tribes, and clans required elaborate defensive measures, all the more so after the introduction of sweet potatoes from the late sixteenth century increased population densities and resource competition. Culture throughout New Guinea remained resolutely oral and preliterate. Initially and superficially at least, we are again reminded of precolonial Zomia. In New Guinea, however, these features hardly embodied anti-state strategies, because no state impinged on what is now Papua New Guinea until Europeans arrived in the late 1800s and early 1900s.Footnote ⁷⁶

None of these considerations necessarily invalidate Scott’s central thesis that the peoples of Zomia responded continuously to pressures from valley states. But they do argue in favour of isolating those lowland pressures more systematically than Scott has done so as to weigh them against other, no less potentially critical determinants of highland behaviour.