I

By the early 1870s, European-owned tea gardens had spread beyond the administered limits of Assam in north-eastern British India, and inhabitants of these outlying areas orchestrated a series of attacks on the gardens and on colonial subjects. In response, the government of India set a new boundary, beyond which colonial subjects required a licence to travel or settle. This was the ‘Inner Line’, which despite its apparently straightforward intentions proved a complex and fragmented act of bordering.Footnote ¹ The government of Bengal considered that governmental intervention would continue beyond the Line in ‘indefinite fashion’, but its subordinate officials based in Assam took the Line to mark the outer limits of state authority.Footnote ² Furthermore, the senior British official in Assam remained sceptical of the possible effect of any bordering project in the vicinity of ‘wholly savage tribes’. He feared that the survey and demarcation parties necessary to lay down the border would be attacked. He continued to associate the Inner Line with a definite limit on governmental authority, which went against his desire to retain an indeterminate zone in which the ‘tribals’ could be dealt with ‘in the simple and summary manner applicable to their rude condition’.Footnote ³ The Inner Line, as with many other bordering efforts at the fringes of British India, served to generate confusion and anxiety as much as clarity and security.

Despite a substantial recent turn towards analysing the frontier regions of colonial India, especially in the north-west (where present-day Pakistan abuts Afghanistan), very little attention has been paid to the conjoined processes of bordering and the making and remaking of frontier zones.Footnote ⁴ A number of historians have contended that the outskirts of Britain's South Asian possessions were anything but peripheral to colonial rule in the region.Footnote ⁵ This case can be extended to include the capacities and governmental techniques of the colonial state: it was in the hills, deserts, and jungles at the state's fringes that the limitations, controversies, and some of the potentialities – including violence – of colonial rule were often most clearly evident. Frontier regions were crucial to the cultural and political development of British India as the East India Company expanded its territorial dominion across northern India from its acquisition of Bengal in 1757 to the defeat of its last major South Asian competitor, the Marathas, in 1818.Footnote ⁶ The subsequent four decades, up to the outbreak of the Indian Rebellion of 1857, saw further major additions to British possessions. These included the annexation of three areas – Assam (in 1826), Sind (in 1843), and Punjab (in 1849) – that brought the colonial state into contact with the outlying regions which this article discusses. With these acquisitions, British India came to assume what would essentially remain its territorial extent until its demise in 1947.Footnote ⁷ Nonetheless, the colonial state's spatial projects were anything but finished with these annexations: through into the twentieth century, a bewildering variety of further bordering and territory-making efforts were initiated.

The north-western and north-eastern frontiers of British India – the former comprising the external boundary regions of Punjab and Sind, the latter the fringes of Assam (and, sometimes, of eastern Bengal) – are usually treated separately by historians. Such an approach ignores that many colonial officials drew comparisons between these areas and their supposedly ‘tribal’ populations, and that a number of scholarly-administrative works considered them collectively as sub-sections of ‘the Indian Frontier’.Footnote ⁸ In considering them together, this article will therefore adopt an actor-led definition of the frontier. A comparative approach also enhances the conclusions that can be drawn from studying these regions, since similar governmental processes – including bordering, the definition and redefinition of zonal frontiers, notions of ‘tribe’, and concerns over international relations – were at play in both areas. Furthermore, the reiteration of unitary and separate north-eastern and north-western frontiers has served to generate an impression that these areas were isolable, fixed, and uniform regions, which they were certainly not.Footnote ⁹ Rather, as a focus on bordering and territory-making schemes will show, they were contested and fragmented zones of variable state penetration, knowledge, and interest, which only intermittently and in limited respects held together as coherent wholes. In adopting a comparative and integrative approach, I am therefore not seeking to bring about a universal and standardized frontier, but rather to highlight the fractured nature of British Indian frontier zones.

In frontier areas, the colonial state's complex forms of modernity were especially apparent.Footnote ¹⁰ As James C. Scott has shown in the case of upland South-East Asia, a key means of states’ self-definition as ‘modern’ was their development of an ingrained trope of ‘barbaric’ tribal society in fringe regions.Footnote ¹¹ Along with notions of ‘backward’ others, spatial forms – particularly linear boundaries and unitary territory under exclusive sovereignty – were supposed to be key distinguishing elements of modern states (and often continue to be taken as such).Footnote ¹² The British Indian state was no exception in this respect: boundary-making in British India was often described as part of the larger attempt (including most notably the Scramble for Africa) to establish the scientificity of European-led efforts to configure global space as the discipline of geopolitics emerged in the late nineteenth century.Footnote ¹³ For instance, in his 1907 lecture on frontiers, the recently resigned Viceroy Lord Curzon asserted that the ‘modern’ process of demarcating imperial boundaries using ‘expert labour and painstaking exactitude’ was an increasingly, although not yet absolutely, ‘scientific’ process.Footnote ¹⁴

However, a vast gulf separated such confident global claims, which are among the most important distinguishing features of the ‘high imperial’ era of the later nineteenth century, from the myriad localized practices of bordering, which exhibited significant continuities with an earlier period of more precarious colonial state control.Footnote ¹⁵ Actual processes of bordering, even – indeed especially – at international boundaries, remained piecemeal and provisional in many areas until well into the late nineteenth and early twentieth centuries.Footnote ¹⁶ Setting specific international boundaries was, in fact, not the exclusive or, in most cases, primary bordering concern of the colonial state. Bordering efforts were often directed towards internal variegation within spaces over which the colonial state claimed at least nominal sovereignty: processes of frontier-making. Even international borders were substantially shaped through the involvement of local inhabitants.Footnote ¹⁷ Rather than converging towards a single, definitive line, borders in many areas at the fringes of British India instead proliferated as the nineteenth century progressed. As David Ludden has commented, ‘on imperial margins, the complexity and ambiguity of imperial time and space appear more clearly’.Footnote ¹⁸

Bringing the north-eastern and north-western frontiers of British India into the same field of analysis also enables robust engagement with the burgeoning political geography literature on bordering and territory.Footnote ¹⁹ Recent works in this field have highlighted the ‘equivocal’ nature of any border,Footnote ²⁰ and the fact that ‘the border, far from being the same phenomenon for all for whom it is significant, is a focus for many different and often competing meanings’.Footnote ²¹ Borders and the spaces they enclose are, like any attempted project of power, best seen as proposals which the intended audience may subvert or renegotiate. As Michel Foucault stated in one of his later works, ‘There is no relationship of power without the means of escape or possible flight.’Footnote ²² This perspective complicates the traditional way in which Foucault's insights have been taken up by scholars of cartography and boundary-making.Footnote ²³ Clearly, the task of effectively undermining a border is significantly harder if that border's material instantiation is rigorous, a truth to which the Berlin Wall, for example, amply attests. But despite the use of material strategies to inscribe boundaries, the notion of fixed borders omits the diverse ways in which boundaries affect, and are affected by, engagements undertaken by various (in)subordinate individuals and groups.

Through a series of case-studies, this article explores various ways in which the borders at the frontiers of mid- to late nineteenth-century British India were shifting and contested objects. It shuttles back and forth between the north-eastern and north-western outskirts of colonial South Asia, incorporating various sections of the Assam, Punjab, and Sind frontiers, and indicating equivalences and linkages between bordering processes in these regions. The case-studies also have differing, although overlapping, conceptual foci; taken together, they indicate that multiple agencies – material factors, local inhabitants, and officials – reworked borders and territories at the outskirts of British India. The complexities of the bordering projects considered in the article point not only towards the specific limitations and contradictory tendencies of colonial rule in South Asia, but also towards the dispersed agency and slippages involved, in varying ways and degrees, in all bordering or territory-making efforts.

II

The British Indian state did not simply create the partially and irregularly governed ‘frontier’ regions at the outskirts of their South Asian territories from scratch. These areas had been zones of variable and indeterminate state penetration and control prior to the expansion of the colonial state. The Sikh state in Punjab had mixed aloofness with violent reprisals in their dealings with the inhabitants of the mountains beyond the River Indus during the early nineteenth century, while the Afghan rulers of Kabul also had a fraught, albeit more intimate, set of relationships with the Pashtuns of this region stretching back to the sixteenth century.Footnote ²⁴ Many of the Baluchis to the west of Sind retained a significant degree of independence from surrounding states including the Khanate of Kalat, which during the seventeenth and eighteenth centuries expanded to claim authority over an area of 30,000 square miles.Footnote ²⁵ The Ahom kingdom in Assam had, since its advent in the thirteenth century, exercised a combination of discretion and occasional punitive forays in dealing with the residents of the hilly regions to the north, east, and south of the Brahmaputra Valley. When the British assumed governmental charge of Assam, Sind, and Punjab, their stated border policy was to retain the spatial divide they claimed had existed between upland frontiers and settled administered areas. Continuities between the frontier regions of the colonial state and those of its predecessors resulted not only from colonial intentions but also from the intractability of the terrain and of some of the inhabitants of these regions, which made lasting penetration by any state exceptionally difficult.Footnote ²⁶

However, the colonial state's bordering and territory-making activities also featured significant, albeit sporadic and uneven, breaks from pre-colonial spatial engagements. In Punjab, for example, British officials had been posted to frontier regions during the three years prior to annexation in 1849 and had attempted various measures, including regularizing revenue demands in administered areas, which served to sharpen the divide between the assessed plains and non-assessed hill regions. In many areas, the colonial state simply lacked the means of obtaining an accurate understanding of the relations that had previously existed between frontier tribes and lowland states.Footnote ²⁷ In addition, colonial conceptions of state space, while less exclusivist than some scholars have claimed, were in certain respects more unyielding than those of the states that preceded them, especially compared with the Ahom state in Assam.

The posa relationship between hill-dwellers to the north of the Brahmaputra Valley and plains inhabitants at the foot of the hills is a prime example of both the type of changes to the border between administered and unadministered territory, and the limitations of these changes, that accompanied the advent of colonial rule in many areas. Posa was a form of tribute payment that Ahom rulers had allowed to exist, in which inhabitants of the duars (outlying valley regions under limited control of the Ahom state) gave gifts to headmen who descended from the hills on an annual basis to collect them. Although some later British officials admitted the intricacy, exactitude, and consistency of the posa system, in the years immediately following the annexation of Assam the colonial state looked on the practice with great suspicion.Footnote ²⁸

In 1834, the first extensive report on posa reached the government of India. It reported differences of opinions between colonial officials over the coercion involved in the process: some said that the inhabitants of the duars gave the goods up without complaint while others claimed the dues were taken with ‘conduct…such as is most naturally to be expected from a rude people invested with so singular a power over the inhabitants of another country’.Footnote ²⁹ Robertson, the leading British official in Assam, insisted that the practice should be commuted to a money payment to the tribes through the colonial state, since ‘it seems to me quite impossible for Government long to tolerate so barbarous an interference with its own territory’.Footnote ³⁰ The government of India agreed that the practice constituted ‘blackmail’ and was ‘rude and complicated’ (criticisms that continued to resonate fifty years later),Footnote ³¹ but only tentatively agreed to Robertson's suggestion as it felt ill-equipped to coerce the tribes and, somewhat embarrassingly, it seemed that Robertson's predecessor had formally recognized the practice as legitimate.Footnote ³² Accordingly, the shift to money payments mediated by the colonial state proceeded cautiously. Robertson's successor, Jenkins, reported in 1837 that many sections of the tribes had submitted to the new arrangement, but reported that ‘in effecting the arrangements now submitted, my Assistants have labored [sic] under much difficulty from the ignorance, and suspicions of the Hill Chiefs, and probably from the misrepresentations of our own [native] subordinate officers, who had an interest in upholding the ancient system, and preventing any change being effected’.Footnote ³³ In 1849, Jenkins reported that some tribal chiefs continued to collect posa in the ‘traditional’ manner, but in general the shift to money payments had occurred by this time.Footnote ³⁴

Colonial actions connected to posa were inherently spatial. They aimed, through attempting to stop a practice that involved ‘tribals’ annually crossing what colonial officials conceived as a boundary, to change the very character of the border between the administered valley and the unadministered hills beyond. The reason for stopping existing posa practices was, in Robertson's words, that they constituted ‘barbarous…interference’ with colonial ‘territory’. All border-making efforts intend to allow certain flows to pass while blocking others:Footnote ³⁵ the British decided that the traditional method of collecting posa was an unacceptable form of border crossing. By commuting this to a money payment arbitrated by its officials, the colonial state instituted a new practice of border-crossing that it deemed to be acceptable: headmen now crossed the border on an annual basis to meet with the local British official and collect a money payment in lieu of the goods they had previously received directly from duar-dwellers. In fact, the occasional failure of chiefs to cross the border in this authorized fashion disappointed officials and seemed to suggest discontent on the part of the absent groups. In 1872, one noted that various clans of ‘Abors’ – he was unsure which ones exactly – had come in to collect their posa from him, but one that had previously come in did not on this occasion: ‘I cannot say what may be the reason of this, they seem to be tired of coming down to take the subsidy.’Footnote ³⁶

The example of posa is indicative of the centrality of performance to acts of bordering, especially those that sought to define frontier regions. As a group of geographical theorists have recently suggested, a border is not a ‘territorially fixed, static line’ but ‘a series of practices’.Footnote ³⁷ The boundary separating the frontier region from administered Assam to the north of the Brahmaputra was, until a major survey party passed through the area in the 1870s, not a cartographically precise object. (Indeed in many sections it remained imprecise even after the survey.) But the fact that the border remained an object of the imagination rather than inscription did not make it any less real for British officials. Perhaps the best representation of this was a British map showing the route of a tour undertaken by Jenkins in the mid-1830s, the time that British interference in posa began in earnest.Footnote ³⁸ The map, despite pretensions to cartographic accuracy and scale in the Brahmaputra Valley, represents the boundary between the valley and the frontier hills with a series of scalloped lines, giving the approximate effect of a cartoon cloud. This seemingly farcical lack of precision was of limited importance to officials in Assam at the time, for whom the border was primarily defined by actions rather than specific geographic location. British interference with posa did not change the indistinct location of the border, but it did alter the set of practices that could legitimately take place across it, rendering it a different spatial object than it had been previously.

It would be incorrect, however, to suggest that the spatial shift engendered by colonial meddling in posa were either wholly intended or complete. Captain Matthie, the official ‘on the spot’ who led efforts to change posa from material exchanges to money payments, reported his confusion over the proper collection methods or monetary value of existing posa allowances, since

The implementation of border-making or border-redefining schemes was never merely defined by a decisive pronouncement from the colonial state. Instead, such efforts were often piecemeal and partial, for a variety of reasons including, perhaps most importantly, local incomprehension and resistance. For instance, when colonial officials first prevented headmen of the Dafla tribe from taking posa in traditional fashion in 1835, some Daflas responded by attacking villages in colonial territory. Matthie was only able to conclude formal agreements with the Dafla chiefs (and even then, only some of them) after the colonial state had violently retaliated to the attack. Despite this attempt to institute the border through what, following Walter Benjamin, might be termed law-making (or boundary-making) violence,Footnote ⁴⁰ some Daflas continued to take posa by traditional means into the 1850s. After yet another set of Dafla border infractions in the early 1870s, to which the colonial state responded with another bout of violence, the famous administrator-chronicler of north-eastern frontier regions, Alexander Mackenzie, stated that ‘the Duphlas [sic] have not yet been brought to see that they are not at liberty to attack men of their own race living within our territory’.Footnote ⁴¹ In other words, at least some Daflas failed to accept the colonial meaning of the border between the Dafla's country and normally administered British Indian territory. The semantics of bordering partially failed, and colonial officials resorted to another show of violence in an effort to provide a foundation for local acceptance of their definition of the border.

III

The Durand line, the boundary between Afghanistan and British India which came into existence with the agreement of a treaty in 1893 between the foreign secretary of India and the amir of Afghanistan, might be assumed to have been a far more authoritative imposition than the posa boundary in northern Assam. Unlike the internal posa boundary, which had to be negotiated with numerous tribal groups, the Durand line was ostensibly an international boundary agreed with the Afghan state. However, this border still proved deeply problematic: the Afghan and post-colonial Pakistani states have both refuted its legitimacy, the former from the early twentieth century onwards. However, the original resistance to the Durand line came from Pashtuns residing in the mountainous area it bisected; Pashtun actions reshaped and undermined even this international and supposedly precise border in ways comparable to the posa boundary in northern Assam.

Despite making some significant departures from the boundary line that had been agreed with other senior members of the government of India prior to his departure, Henry Durand, the foreign secretary of British India, returned from Kabul in 1893 satisfied that he had agreed ‘a well-defined’ border.Footnote ⁴² With the benefit of hindsight, we can see that Durand's self-congratulation at the conclusion of his mission was, from the perspective of grand strategy, justified. The Indian–Afghan border he negotiated sufficiently fulfilled its international purpose of delineating an Afghan buffer state separating the British and Russian players as they played out what would be the last moves of the ‘Great Game’. But if we examine its local impact, it becomes clear that the Durand line was not such a straightforward success. The purely theoretical delineation that had taken place in Kabul had no material reality whatsoever until Boundary Commission demarcation parties, comprising British and Afghan members, took the field from early 1894. And as Thomas Holdich, the lead surveyor involved in demarcating the Durand line, later commented, ‘The process of demarcating a boundary…is the crux of all boundary making…It is in this process that disputes usually arise, and weak elements in the treaties or agreements are apt to be discovered.’Footnote ⁴³

In the case of the demarcation of the Durand line, such disputes arose with both Afghan members of the Boundary Commission and groups residing in the vicinity of the border. There were persistent disagreements between British and Afghan surveyors over the actual location of the frontier, which were further complicated by numerous inaccuracies in the maps that Durand had used in his negotiations with the amir (and which he had insisted were correct in the face of the amir's numerous objections).Footnote ⁴⁴ The result of these confusions during demarcation was that, as Holdich himself admitted, ‘no part of the boundary defined south of the Hindu Kush [i.e. along the entire Punjab and Baluchistan frontiers] was the actual boundary of the agreement’.Footnote ⁴⁵ These major alterations directly contravened the intention set out in the formal agreement between the amir and Durand, which stated that the Boundary Commission was to lay down ‘a boundary which shall adhere with the greatest possible exactness to the line shown in the map attached to this agreement’.Footnote ⁴⁶ A pertinent example of the failures of demarcation was the bizarre and antagonistic performance of contradictory boundaries that continued to take place every time a caravan passed through the Khyber Pass: British and Afghan guards accompanied the caravans to the different villages where their respective governments believed the border to be, so for a distance the caravans travelled with two military parties, locked in a tense impasse.Footnote ⁴⁷

The thinly veiled hostility and confusions between British and Afghan boundary commissioners was widespread and caused British officials significant concern. Of far greater and longer-lasting importance to the colonial state, however, was the open enmity of numerous groups that lived in the zone between fully administered British India and Afghanistan. Many of the Pushtun communities who inhabited the mountains in this area had a long history of contesting the administrative boundary that separated the fully administered sections of Punjab from the ‘independent tribes’ (whose jurisdictional status was in fact perennially uncertain). Prior to the Durand agreement this border was, like the administrative boundary in Assam before and after the introduction of the Inner Line Regulation mentioned in the introduction, intermittently and often contradictorily defined, and material demarcation was infrequent.

The effect that the Durand line was intended to have on the existing administrative boundary and the intervening area between governed British Punjab and sovereign Afghanistan was also the subject of much debate among colonial officials, before, during, and after the parties of boundary commissioners took to the field. For example, provincial officials and members of the government of India were unable, despite prolonged debate, to agree unanimously on the implications of demarcating the section of the Durand line that ran through the area inhabited by Waziri tribes. Richard Bruce, a protégé of Robert Sandeman (whose own frontier-making activities in Baluchistan are discussed later in this article) and the official charged with control over relations with the hill inhabitants along this section of the frontier at the time of demarcation, suggested that the imposition of the Durand line committed the British to intervene more actively in Waziri affairs. Durand himself agreed with Bruce that the ‘natural consequence’ of demarcation in this area was that the colonial government should tax and give service to the Waziris living between the cultivated plains of south-western Punjab and the newly defined sovereign territory of Afghanistan. ‘We should’, Durand wrote, ‘without annexing, bring the whole [Waziri] tribe into order, and insist upon the stoppage of raids and fighting, and gradually make the country free to our political officers.’ Other members of the Punjab government and the government of India dissented from these interpretations of the border in various ways; the ensuing debates touched, sometimes explicitly, on interrelated uncertainties over the sociological constitution of Waziri communities (which were renowned as especially ‘democratic’ – that is, without authoritative headmen) and the implications of tribal structures on colonial spatial policies.Footnote ⁴⁸

Bordering and the reworking of frontier spaces always entailed entanglements with other complex considerations, including the nature of the ‘tribe’ and the limitations of the colonial state in the light of the apparent fissiparousness of certain hill populations. In the face of such difficulties, the government of India wavered when delivering its verdict on the implications of demarcation ahead of the Waziri Boundary Commission party taking the field. It was not prepared to back Bruce's proposals in their entirety and avowed that officials were ‘to interfere as little as [possible] in [the Waziri's] internal affairs’. But the government acknowledged that laying down the international boundary meant it had ‘assumed a measure of responsibility for the peace of the Afghan border which has not hitherto been ours’, meaning that a military post should be established beyond the existing administrative border and headmen should be given allowances to try to ensure peace.Footnote ⁴⁹ However, three members of the government dissented from this policy, claiming that positioning a post in Waziri country amounted to ‘practically assuming…administrative control’ and, given the loose social constitution of the tribes, was ‘eminently calculated to set the whole frontier in a blaze’. This minority instead advocated a continuation of the policy prior to Durand's trip to Kabul, which they described succinctly as ‘resting on our present frontiers and influencing the Waziris from without’.Footnote ⁵⁰

The lack of clarity among British officials on the Durand line's intended impact on local inhabitants and tentative talk of extending British authority into Waziri areas meant that the heightened suspicions of the residents of the irregularly governed hills were wholly understandable. When the Boundary Commission finally assembled in the Waziri hills in October 1894, locals undertook probably the most spectacular immediate contestation of the Durand line, assembling a party of over 1,200 and attacking the Commission's party at the village of Wana before it had even begun to undertake its task of materially inscribing the border by erecting boundary posts or piling up stones.Footnote ⁵¹ Even Holdich, who suspected Afghan complicity in almost every instance of tribal resistance to British governmental projects, admitted that ‘there was no suspicion of concerted action about this affair at Wana. The Waziris had heard about the boundary that was to be placed between them and their refuge.’Footnote ⁵² Along the border's entire length, the parties of surveyors, political officers and civil administrators were accompanied by a substantial armed force to fight off anticipated resistance. Following the attack at Wana, a full-scale military ‘expedition’ was sent to subdue the groups responsible and clear the way for the return of the demarcation party for a second attempt to establish the border in material reality.Footnote ⁵³ It was also deemed prudent to maintain indefinitely a permanent garrison of troops at Wana in case of continued unrest; in fact, the new garrison served primarily to inflame discontent further.Footnote ⁵⁴

Initial British estimates of when the demarcation parties would complete their work were hopelessly inaccurate.Footnote ⁵⁵ Indeed, the fundamental idea of ‘completing’ the Durand line proved overly optimistic, and the demarcation parties left the field – often with fewer men than they had contained at the outset – with large sections of the border without material markers. Disagreements with their Afghan counterparts, the difficult landscape – which led to what Holdich termed ‘the geographical impossibility’ of inscribing some sections of the agreed boundaryFootnote ⁵⁶ – and above all extensive resistance from local inhabitants overcame the capacity of the British to transfer the boundary from the (often inaccurate) map to the less tractable territory.

The apparent impression of most frontier officers and subsequent scholars that the completion, or rather cessation, of demarcation represented an end to the unrest induced by the Durand line is wholly inaccurate.Footnote ⁵⁷ For many people who resided in its vicinity, the border was indicative of increasing British interference and was therefore a continuing source of irritation. Holdich wrote, sympathetically if simplistically, that from the tribals’ perspective, ‘a boundary line indicated by piles of stones had been drawn across their hills to show that theoretically they were shut, and that beyond that line they might appeal no more to people of their own faith and their own language in times of difficulty and disaster’.Footnote ⁵⁸

There may have been widespread awareness of the Durand line among those who resided nearby, but such cognisance did not translate into either widespread practical acknowledgement of the border as a limit that should not be transgressed, or recognition of the right of the British Indian government to impose such a limit. One British officer's analysis of the north-western frontier, written in 1939, bluntly stated that even then, ‘Many [locals] would not admit that the British had any right to control their actions across the Durand Line.’Footnote ⁵⁹ Acts of border subversion among the frontier inhabitants continued apace after 1895, ranging from quotidian practices of ignoring the colonial border, which concerned the British most when they involved illegal acts such as trading stolen British guns with Afghans, to much grander rejections.Footnote ⁶⁰ From the perspective of colonial officials, this afterlife of the Durand line came into abrupt focus in 1897 with a series of outbreaks of unrest, which collectively constituted the largest uprising the British faced in their century of frontier administration in the north-west. Contrary to many historical accounts that follow most colonial officials in attributing the uprising to religious or ‘fanatical’ elements, the 1897 uprising should be seen primarily as a series of border contestations.Footnote ⁶¹ Even the secretary of state for India acknowledged the centrality of the border in provoking the uprising in a letter to the viceroy:

The 1897 uprising also indicated that the border was problematic in another, somewhat contradictory, way. Many who were involved in the uprising crossed the border into Afghanistan to avoid British military retaliation, then recrossed to take part in violence against British infrastructure and subjects, thereby rendering the border thoroughly ineffective as a barrier against movement into Afghan territory. Large numbers of men residing in Afghanistan also transgressed the border in order to join relatives and associates who now nominally resided in colonial territory in the fight against the British.Footnote ⁶³ The border was a significant irritant but a flimsy deterrent. British officials were aware of the absurd situation that throughout the uprising they were the only people observing the Durand line. At one point, the Punjab government noted: ‘A punitive expedition will not touch many of those who joined the raid, a large proportion of the raiders having come from [Afghanistan].’Footnote ⁶⁴ In short, in 1897, the Durand line clearly became detached from its supposed function as constituent and visible marker of state power and unitary sovereignty. Instead, even with the attendant problems of demarcation noted previously, it acted first as a provocation to refuse state authority, and then as an obstacle to attempts to reinstitute that authority.

IV

Demarcating borders at the outskirts of British India with a regular series of man-made markers was the exception rather than the rule. In some instances in the north-east and north-west, border posts manned by troops or local militias were established to mark the administrative boundary; although these were generally set back some distance from the border drawn on maps, they were often the first signifier of the boundary that a crosser would meet. The material form (or lack thereof) of the border was vital, especially where the colonial state attempted to change the significance of the border to people whose understandings of boundaries greatly differed from those of colonial officials. In this section, I will examine an instance in which the colonial state did attempt to inscribe borders materially, and highlight the difficulties and limitations of these demarcation practices.

In 1873–4, during the annual cold season when most colonial governmental activity relating to frontier areas in the north-east took place, the northern boundary of Darrang District in Assam was surveyed and partially demarcated. To a greater extent than even the troubled demarcation of the Durand line, the process in Darrang was haphazard and riven with confusion. As mentioned at the beginning of this article, different colonial officials had inconsistent understandings of the intended nature of this section of the border: some thought it formed the ‘Inner Line’ (across which colonial subjects could not pass), others believed it marked the outer limit of colonial government. Furthermore, the boundary was something of an afterthought, constructed by a survey party primarily tasked with laying down the border between the sovereign territory of Bhutan and that of British India.Footnote ⁶⁵ Certainly, this extension into Darrang did not obtain the wholehearted approval of the Assam government: Keatinge, the chief commissioner, noted that ‘to survey and define a boundary already laid down by Treaty is one thing, to lay down a new boundary between ourselves and savage tribes who are controlled by no central government, and with whom Treaties are impossible, is quite another thing’.Footnote ⁶⁶ Keatinge felt that although the boundary with the Bhutias may have been moderately worthwhile, since these people ‘are not wholly savage, and appear to have understood the meaning of our proceedings’, this was not necessarily the case further to the east. He feared for the safety of the boundary survey party, as ‘what the Akas think of these proceedings may be doubtful…[and] there must in any case be a risk that the tribes will look upon the Surveyor and his guard as people who are defining the boundary in their own interests, and who, being few, may safely be opposed’.Footnote ⁶⁷ It is clear from these concerns that colonial officials understood bordering to be a politically charged intervention that entailed a serious threat of provoking resistance, especially among groups that seemed to the British to lack leadership structures.

Local resistance was certainly among the problems faced by the survey party, particularly since it was tasked with a dual role to demarcate the boundary while also helping to enforce a blockade of the Daflas, who resided to the east of the Akas. The tensions between the two roles meant that the leader of the mission, Lieutenant-Colonel Graham, had to omit one border pillar since it was to be placed at too great a distance from the blockaded portion of the boundary for him to attend to the other aspect of his party's duties.Footnote ⁶⁸ Another pillar was left unconstructed on the initial expedition as Graham's subordinate officer, Captain Cowan, ‘fear[ed] a collision with the Akas…although he marked the site, carried stones, sand, &c.’.Footnote ⁶⁹ When this particular marker was eventually put in place almost two years after the main expedition, colonial officials evidently considered it an event of some importance: the ‘substantial pillar, measuring 9 × 7 × 5 feet…on a high spot’ was erected ‘in presence of the Akha [sic] chief’.Footnote ⁷⁰ Despite this ceremony, however, many of the officials involved in the demarcation process doubted the effectiveness of the material instantiation of the border as a means of effectively communicating its meaning to the population whose territorial limits it ostensibly defined. Even Graham admitted:

In light of Graham's comments and Keatinge's similarly sceptical outlook on the tribals' understanding of the pillars, the government of India agreed that demarcation should not be extended further eastwards along the northern boundary of Assam.

Clearly, then, resistance and incomprehension, or apprehensions thereof, were significant aspects of the shortcomings of demarcation. Equally important in certain instances was material resistance. Even in areas where inhabitants seemed amenable to the survey party's project, efforts to materialize the border met with serious difficulties resulting from difficult terrain. For instance, Graham reported that along the portion of the boundary between Assam and Dafla country, ‘no objections were made to our proceedings, the Dufflas [sic] merely remarking that “the plain belonged to the sirkar [British], and the hills to them”’. However, he continued, ‘an endeavour made by me to get to the hills in this direction failed, owing to the nature of the jungle, which was simply a mass of creepers, thorn-bushes, and broken ground’.Footnote ⁷² The party encountered similar problems along the entire length of the Darrang District's northern boundary. Graham wrote,

The apparently intractable terrain thoroughly confounded Cowan's original instructions ‘to place each pillar, if possible at such a height on the first low range as would enable it to be seen from the pillars on either side of it, and give us at the same time a fairly straight boundary line, which would take in all the plain’.Footnote ⁷⁴ Contrary to these orders (and contrary to broader notions of a ‘natural’ and clear boundary between hills and plain), the party found that

After three months of toil against seemingly impossible broken hills and forest and inclement weather, only fifty-two miles of border were demarcated by twenty-four pillars. The material struggles to convey the stones, lime, and sand to hilltops where clearly visible pillars could be located complemented concerns over the limited communicative potential of these markers, rendering the whole enterprise relatively futile.

V

Whereas in the cases of the Durand line and Darrang District the colonial state sought, but in some degree failed, to impose a clear boundary, in some other areas at the outskirts of British India, officials came to acknowledge and even embrace the idea of an amorphous, unsettled border (and its corollary concept of territory that was ‘full of holes’).Footnote ⁷⁶ The Naga Hills District to the south of the settled agricultural regions of the Assam Valley provides a clear example of this trend. The colonial state's engagement with this largely hilly and forested area inhabited by the (ever-changing) set of tribes it classified as ‘Nagas’ was tumultuous, variable, and often incoherent. Having first encountered Nagas in the early 1830s, when British officials met with significant resistance while travelling through the hills between Assam, Burma, and Manipur, relations for the next twenty years centred on tours by colonial officials that mixed efforts at reconciliation with displays of violence.Footnote ⁷⁷ A brief attempt to institute a limited form of government in the hills in the late 1840s ended in disaster, and a policy of non-interference in Naga affairs was instituted in 1851 when the state withdrew its minimal territory-making technologies (essentially, a detachment of soldiers) from the hills for the following fifteen years.

In the mid-1860s, a combination of occasional Naga incursions across the putative border of British Assam and colonial officials' desire to interfere in episodic violence between Naga villages tipped debates over the proper extent of British sovereignty and responsibility back in favour of expanding colonial authority into the hills. Despite being garlanded with the title of the Naga Hills District – suggestive of the imposition of uniform authority over a bounded area – the District was initially a set of claims and intentions of the part of colonial officials with an extremely limited material instantiation. A British officer, Lieutenant Gregory, and a party of troops were established in the Angami Naga village of Samaguting, tenuously linked to the nearest subdivisional station in Assam by a thirteen-mile road (that had to be built by the soldiers en route to their post) and a river (a further sixty miles). Despite a theoretical claim to hold territorial sovereignty over a large swathe of the Naga-inhabited region,Footnote ⁷⁸ Gregory was initially instructed not to exercise ‘direct control’ over Nagas except those residing in Samaguting, although he was allowed summary punitive powers against ‘any village proved to have been concerned in any gross outrage’.Footnote ⁷⁹ In this initial form, then, the Naga Hills District was not a clear territory delineated by a border from outside areas: its limits were explicitly amorphous, ebbing and flowing with the intermittent dispatch of punitive and exploratory expeditionary parties.

Colonial territorial power was severely circumscribed by perceived difficulties of terrain, limits on effective knowledge gathering, and tactics of flight and avoidance among the locals. Partly because of these limitations, the British Indian state's power to define territory in the Naga Hills District consisted more of destructive interludes than entrenched and lasting structures. Probably the most significant spatial intervention in the years immediately after the foundation of the District was Gregory's destruction of the refractory village of Razepemah and subsequent prevention of the villagers from rebuilding on the same site. Gregory's superior, the commissioner of Assam, felt it ‘essential’ that

At this stage, the colonial state was generally limited to spectacular shows of force that exceeded the acknowledged limits of ‘direct control’ in the Naga Hills rather than consistent and normalized efforts to construct stable state space.

Although the state, or more accurately its representative stationed in the hills, gradually gained traction in the area during the early 1870s through schemes such as receiving delegates from increasing numbers of the principal villages in the vicinity of Samaguting, there was no corresponding creation of a uniform territory.Footnote ⁸¹ From the mid-1870s, a creeping expansion of formal British authority took place, with the colonial official on the spot intermittently taking villages under formal British protection in return for securing villagers' assent to pay a house tax. Commenting on the first two villages taken into British protection through this process, the government of India explicitly stated that there was no need for a general principle to be laid down to dictate similar events in the future; rather, a ‘roug[h] and indefinit[e]’ process of expansion was sufficient to fulfil the primary purpose of ‘keeping order on the frontier [i.e. the border between Assam and the Naga Hills]’.Footnote ⁸² In other words, hazy and variable influence in the hills – which sometimes involved a form of administrative control, as in the case of levying a house tax – was paradoxically thought to be a necessary part of the colonial state's attempt to enforce a definite administrative border.

Despite the amorphous nature of the Naga Hills District, the location of its nominal boundaries still greatly exorcized colonial officials. In 1875, Butler, the political agent in the Naga Hills, claimed that the ‘theoretical boundaries of this district’ laid down in 1867 were ‘absurd and impossible’, being in many places ‘utterly wrong, or so extraordinarily defined as to be quite impossible of identification’.Footnote ⁸³ Although his superiors accepted Butler's alternative boundaries, their practical significance was limited, as ever more villages came under British protection and a series of military and survey expeditions performed territory in ways that differed substantially from its fixed representation in maps and in official discussions on bordering. The patent absurdity of these bordering practices was summed up in the comment by one member of the government of India that they led to ‘a perpetually disturbed frontier line’.Footnote ⁸⁴ Another member, Steuart Bayley, accurately identified that the territorial and border confusions arose in significant part because the mainspring of actions by local colonial officials was ‘no longer…protecting our own frontier or our settled districts…it is really…extending our authority, village by village, over the whole tract of country’. Bayley continued, ‘The Naga Hills District…is a geographical expression, not an administrative fact…Consequently, the present boundary has no special or intelligible meaning from an administrative point of view.’Footnote ⁸⁵

The indeterminate and irregular territorial nature of colonial administrative space in the Naga Hills continued long after the suppression of a major rebellion at the British post at Kohima in 1879, which is generally seen as a major turning point in the normalization of colonial authority in the hills.Footnote ⁸⁶ Robert Reid, the governor of Assam in the 1940s, wrote in his administrative history of the Assam frontier that from 1881 colonial ‘penetration went inexorably if irregularly. It was impossible to draw a line as the boundary of our area of control and to say that we should be blind and deaf to all that went on across that line.’Footnote ⁸⁷ The continuation of the border's provisional and flimsy status is clear from the fact that in 1903, Captain Howell, the new deputy commissioner of the Hills District, complained that the supposed, and ever-shifting, boundary had little effect on inhabitants of the area: ‘Whatever a red line on the map may mean to a civilized man, it has no meaning to these people, who entirely fail to see, where all men are alike, by what principle Government should make such important distinctions.’Footnote ⁸⁸ Having discussed the various merits and drawbacks of physical boundaries such as hill crests and rivers, Howell concluded,

Although the government of India did not explicitly ratify Howell's suggestion (which pointed towards an explicitly borderless administrative territory), it acquiesced to a series of extensions to the eastern boundary of the District throughout the 1900s and 1910s that essentially amounted to the same thing.

In the case of the Naga Hills District, the notion of territory as a uniform, even coverage of a defined portion of the earth's surface is inadequate to describe both the intentions of officials and the actualization of the colonial state in this area. State space in the Naga Hills was patchy, and had the character of an ever-shifting network of nodes and routes under varying degrees of control which frequently rendered the very idea of a clear border meaningless. As Howell's words indicate, some British officials came to believe that a border was worthless in an area in which the inhabitants seemed not to comprehend the concept of a fixed boundary and appeared to engage in social and economic relations that confounded efforts to bound homogeneous territories. These officials instead developed a notion that effective control depended on imprecision, meaning that the Naga Hills District remained amorphous and patchy rather than developing into a clearly bounded and fixed territory.Footnote ⁹⁰

VI

The preceding case-studies have made it clear that the subversion of colonial borders by inhabitants of the very regions they sought to enclose significantly inflected the meaning and location of these boundaries. In the case of the Naga Hills, some colonial officials in frontier regions even came to see fixed boundaries as inimical to their administrative roles. Colonial boundaries could also be fundamentally reshaped by violations from British officials as well as ‘tribals’.

Robert Sandeman's repeated violations in the decade from 1866 of the administrative boundary between Dera Ghazi Khan District in southern-western Punjab and the areas to the west which were nominally under the control of the Khanate of Kalat, a large but unstable subsidiary ally of the colonial state, were perhaps the most famous, and almost certainly the most consequential, border crossings undertaken by a British Indian official.Footnote ⁹¹ Sandeman's role in engendering a wholesale shift in frontier policy in the north-west from a defensive ‘close-border policy’ to an active ‘forward policy’ has been exaggerated by both some of his contemporaries and a number of subsequent historians.Footnote ⁹² Contrary to such assertions of the uniform implementation of a ‘close border’ policy along the Sind and Punjab frontiers until a sudden break at some point between the late 1860s and mid-1870s, colonial interventions and interactions with the tribes beyond the administrative boundary were volatile and changeable. For instance, in the mid- to late 1840s John Jacob famously innovated and implemented a policy of patrols and pursuits well beyond acknowledged colonial territory in Upper Sind, which reached a crescendo when approximately 600 Bugtis were killed by a detachment of the Sind Irregular Horse in what the commissioner of Sind termed ‘a sanguinary engagement…in which one of the hill tribes has been almost annihilated’.Footnote ⁹³ Nevertheless, Sandeman's border violations during the late 1860s undoubtedly played a vital role in British expansion into Baluchistan through the stationing of a force at Quetta from 1877 and the assumption of ever more of the Khan of Kalat's territory and governmental responsibilities, culminating in Baluchistan's formal incorporation into British India by the late 1880s.

Although Jacob's interventions achieved some success in securing administered Upper Sind from tribal infractions, outbreaks of inter-tribal violence and machinations against the Khan of Kalat continued.Footnote ⁹⁴ In fact, British ambivalence towards Kalat's rulers and repeated interference were major drivers of this instability, and had been since the time of the First Anglo-Afghan War when the British attacked the town of Kalat, deposed the existing Khan, and replaced him with another chief of their choice.Footnote ⁹⁵ The Bugtis also began to conduct raids on the administered areas in Dera Ghazi Khan District, the section of south-western Punjab that bordered Upper Sind. Whereas the material instantiation of the Upper Sind boundary, with its row of manned fortresses, presented a formidable obstacle to would-be transgressors, the Dera Ghazi Khan boundary was sparsely patrolled and frequently ineffective. Sandeman took up his post as deputy commissioner of Dera Ghazi Khan in 1866 against a backdrop of a series of sporadic and generally ineffective measures undertaken by colonial officials in Sind and Punjab to try to put a stop to border violations and an uncertain policy regarding the unstable Kalat Khanate. According to Sandeman's first biographer T. H. Thornton, a key problem facing officials on the Punjab frontier relative to their colleagues in Sind was the retention of the policy stating, in Thornton's words, ‘District officers were never (without special sanction) to risk their lives beyond the border, or to dream of its extension beyond present limits.’Footnote ⁹⁶ In fact, this both oversimplified the actual practices that occurred among Punjab officials near the frontier and exaggerated the fixity and precision of the border between administered and independent territory. It is, however, true that Sandeman's border crossings stood out owing to their regularity and apparent surety of purpose: he ‘broke the border rule repeatedly and successfully’.Footnote ⁹⁷

From the outset of his tenure in Dera Ghazi Khan, Sandeman spoke out against the existing location of the administrative border, claiming that since it divided unified tribal groups it represented a ‘cruel injustice’.Footnote ⁹⁸ He later stated that the boundary line adopted by the Sikhs and purportedly maintained by the British ‘has proved disastrous to our subjects, and to the frontier clans, and rendered the administration of the frontier tract itself a matter of extreme difficulty’.Footnote ⁹⁹ In the longer term, Sandeman's actions undermined the original location of the boundary through extending colonial control into Baluchistan; but their most immediate effect was to challenge it as a performed object.

During his first year in Dera Ghazi Khan, Sandeman took his first tour beyond the administrative boundary accompanied by his assistant Bruce, four Tumandars (tribal chiefs), and approximately 300 followers. Beyond the border, he met with the Bugti Tumandar, Gholam Mortaza Khan, and a number of men subordinate to him who had been among the chief raiders of colonial territory.Footnote ¹⁰⁰ In undertaking this border crossing, Sandeman directly contravened the established chain of command: all dealings with nominal Kalat territory should have gone through the superintendent of the Upper Sind Frontier or the British resident at the court of Kalat, who was also under the auspices of the Sind and Bombay governments. Sandeman, as a Punjab official, had no official right to cross the border as he did. In the name of the colonial state, but without authorization from any senior officials, Sandeman dictated terms to the headmen, including the restoration of stolen property, and in return offered employment to a number of men in irregular levies and payments to particular Baluchi chiefs.Footnote ¹⁰¹ Although the colonial state employed many hill inhabitants in the north-east in a variety of roles, including Shans in militias in the Naga Hills and Singphos in expeditions north of the Brahmaputra,Footnote ¹⁰² taking tribals into colonial service in Baluchistan was, as Sandeman admitted, an ‘entirely…experimental measure’.Footnote ¹⁰³ The advent of tribal service as a means of refashioning the frontier was convenient both for the chiefs who benefited from additional income and status, and for Sandeman as the system served to justify additional, generally unauthorized, trips beyond the border throughout the remainder of the 1860s and into the 1870s. During this period, Sandeman also established, without express permission, a summer camp twenty-five miles beyond the acknowledged limits of British territory at which he and Bruce spent a few months each year in contact with the tribes of the surrounding hills.Footnote ¹⁰⁴

Although Sandeman's repeated border violations in the late 1860s and early 1870s had the important consequence of reinvigorating attacks on colonial subjects in Upper Sind, causing consternation among some officials in Sind, especially the commissioner, William Merewether, the government of India generally backed Sandeman.Footnote ¹⁰⁵ By 1874, against a background of continued upheavals in Kalat, which were by this stage inflamed by the ongoing lack of clarity over which colonial officials had powers to act beyond the administrative border, the government of India pronounced that ‘the time has now arrived…[to] make our own arrangements direct with the frontier tribes and without reference to the Khan’.Footnote ¹⁰⁶ This marked a significant breakthrough for Sandeman's border crossings and his advocacy of direct relations with headmen, especially in the form of payments and service, which opposed Merewether's insistence on operating through the Khan. Sandeman was chosen to lead two expeditions in November 1875Footnote ¹⁰⁷ and April 1876, which culminated with the establishment of the Baluchistan Agency and the permanent occupation of a British military post at Quetta.Footnote ¹⁰⁸ The agent to the governor-general in charge of the Baluchistan Agency, a post first occupied by Sandeman, also formally assumed powers of intervention between the Khan of Kalat and the Sirdars (chiefs) ostensibly under the Khan's authority, thereby essentially completing the process started during the First Anglo-Afghan War of reducing the Khanate to little more than a ceremonial authority controlled by British officials.Footnote ¹⁰⁹ Once again, Sandeman's individual initiative was key to these outcomes: the authorized aim of his second mission was to secure the Bolan Pass from tribal depredations, but the government of India pronounced after the mission that ‘Whilst…we were fully alive to the difficulties and responsibilities of the permanent intervention advocated by Major Sandeman, we could not disguise from ourselves the greater difficulties of renouncing the position in which the success of his mediation had conspicuously placed us.’Footnote ¹¹⁰

In Thornton's adulatory rendering, which might be understood as a particular form of what Mary Louise Pratt has termed colonial ‘anti-conquest’ narrative,Footnote ¹¹¹ Sandeman's journey into Baluchistan in 1876 became a self-renunciating pilgrimage:

Thornton also described Sandeman's previous border crossings in similarly self-sacrificial terms, claiming they were ‘particularly hazardous proceeding[s], because he not only risked his life, but his career’.Footnote ¹¹³ Regardless of their dangers, Sandeman's frontier-redefining tours should be seen as the acts of a relentless self-promoter, who successfully sought to accumulate power for himself within the structures of the colonial state. His strategy of border violations as a means of self-aggrandizement was eminently successful: he was knighted and given charge of the Baluchistan government. Sandeman's reconfiguration of the border may have been in certain respects exceptional at the outskirts of British India in the later nineteenth century, but it also points to certain important elements in the practices of bordering and frontiering in nineteenth- and early twentieth-century colonial South Asia. It was indicative of the fact that borders and frontiers were shaped by local actions and tribal concerns at least as much as by the high strategic motivations that many historians claim undergirded frontier policies.Footnote ¹¹⁴ The primary rationale for advancing into Baluchistan in 1877 was not grand strategy but long-standing concerns over tribal control and communication, concerns that were both partially engendered and purportedly addressed by Sandeman's tours beyond the border.

In his 1960 article, John Galbraith famously claimed that the ‘turbulent frontier’ was the major imperative behind British imperial expansion in India, Malaya, and South Africa.Footnote ¹¹⁵ ‘Governors, charged with the maintenance of order’, Galbraith argued, ‘could not ignore turbulence beyond their borders, turbulence which pulled them toward expansion.’Footnote ¹¹⁶ Sandeman's involvement in Baluchistan in the decade or so before the move to Quetta shows that turbulence was not merely something experienced by colonial officials: it was also produced and encouraged by them. As was the case for a number of officials who played a key role in the frontier areas in the north-east, for Sandeman the very idea of fixed borders – the supposedly foundational spatial principle of the ‘modern’ state – was anathema: a block on personal ambition and on the form of free-ranging governmental intervention he deemed appropriate at the fringes of British India.

Some administrators at the upper echelons of the colonial state in India came to acquiesce in, and even celebrate, such notions of amorphousness (and their corollary of administrative anti-modernity) as defining principles of frontier spaces. Shortly before he became viceroy, Curzon indicated the symbiosis of spatial indeterminacy and an explicitly atavistic form of government wrote while praising Sandeman's achievements: ‘It is no good to have a “Warden of the Marches” unless you give him a comparatively free hand.’Footnote ¹¹⁷ In invoking as a model of the colonial frontier official the wardens of the marches – officials with wide-ranging discretionary powers who controlled the frontier areas between England and Scotland and England and Wales during the late medieval period – Curzon, presumably inadvertently, showed the extent to which high-level officials came to support the initially subversive tendencies of their subordinates. Some fifty years earlier, Sandeman's role-model John Jacob had defended his repeated and violent crossings of the boundary between British India and the Baluchi tribes by claiming that on this frontier,

Curzon's statement, then, indicates the traction that local colonial officials' initially subversive engagement with the Baluchistan border eventually gained with their superiors. It also points towards the enormous importance of such boundary violations – by turbulent officials as well as turbulent locals – which at some times and in some places occurred so frequently as to render the official fixed border partially meaningless and subsequently lead to changes in its location, meaning, and material form.

VII

The multiple spaces that collectively constituted the frontiers of the colonial state in India were consistently mutating and never definitively settled. Here, the limitations of the state's ability to fashion and impose coherent and lasting governmental territories were often laid bare. The examples examined in this article collectively suggest that the interwoven colonial projects of bordering and territory-making were defined by multiple slippages rather than being assured projections supposedly characteristic of modern state authority. Borders and territories are ever-shifting and protean combinations of practices and ideas, far more complex and less fixed than the areas and lines on a map. They are performed, subverted, and reformulated over time; categorically, they are not fixed receptacles, the static backdrop against which diachronic developments – the stuff of history – happen.

Bordering in colonial India was an especially fractured and variable set of processes in which the straightforwardly successful imposition and maintenance of clear unitary lines rarely occurred. Even before any attempt was made to impose a border, the intentions of most bordering projects were contested among colonial officials. Considering the prevalence of confusions and subversions among actors operating under the auspices of the colonial state, it is unsurprising that communicating the border to the local populations whose actions it was intended to inflect was fraught with difficulties. Such problems were especially apparent in areas inhabited by groups the colonial state identified as ‘tribes’, which was, in general, another way of officials saying that these groups lacked headmen with enough authority to enable a binding border agreement to be concluded with confidence.

It is important to emphasize that, as Andrew Walker has shown in the case of Laos, the agency of borderland inhabitants in terms of border-making is not only negative.Footnote ¹¹⁹ Contrary to James Scott's characterization of the hill populations of upland South-East Asia as communities seeking above all to escape government, examples from the frontiers of British India suggest that ‘tribals’ (and the terrain they inhabit) do not only oppose states and state-designed borders efforts, but also shape these efforts – and therefore also shape the states themselves – in more subtle ways.Footnote ¹²⁰ Likewise, some officials – often, but not exclusively, those ‘on the spot’ – perceived certain advantages to amorphous boundaries and patchy territories. A major effect of their efforts was to generate conceptions of borders and territory contrary to the clearly delineated and unitary spaces and boundaries that are generally considered to be integral aspects of the modern state. The haziness of the limits of colonial territory and the multiple variegations within this territory was a key and fairly ubiquitous element of the colonial state's spatial engagement, not an exceptional or unusual aberration. The advocacy of hazy borders was not necessarily linked to a straightforward territorial expansionism, although in the case of Sandeman in Baluchistan such expansion did result. More often, it coalesced with proposals that the governmental initiatives and interventions of the colonial state should be variable and locally managed, rather than definitively delimited.

The spatial engagement of colonial states during the era of ‘high imperialism’ in the mid- to late nineteenth century is most often approached through technologies – ranging from mapping to irrigation or jungle clearanceFootnote ¹²¹ – that seek to naturalize territories and render space lifeless and immutable. There is no doubt that the colonial state in India did seek, in Doreen Massey's words, to ‘tam[e] the challenge of the spatial’.Footnote ¹²² But in many cases, especially at tribal fringes, its efforts were contradictory and prone to failure. Territories, borders, and frontier zones continued to be constituted through myriad relationships that never allowed for complete or stable colonial state domination. What Massey has aptly termed ‘the chance of space’ remained, and often overwhelmed the colonial state's haphazard and sometimes ambivalent efforts to impose legibility and order.Footnote ¹²³