Conjugality,Colonialism and the ‘Criminal Tribes’ in North India

Introduction

The Criminal Tribes Act (CTA) of 1871 was a project to geographically redistribute and immobilize criminalized populations on the basis of family units. The British colonial regime in India applied the term ‘criminal tribe’ to a diverse collection of marginalized peoples whom they regarded as criminals by hereditary caste occupation, including low-caste groups and people who were marginal to sedentary rural society. Family ties were a key site of contestation between criminalized people and the colonial state. Colonial officials attempted to manage familial relationships to create new social geographies. Provincial governments policed ‘criminal tribes’ through varying forms of spatial governance—such as mobility restrictions, forced transportation or confinement in a settlement—most of which spatially fixed or redistributed families. Authorities were often concerned with the feelings and sentiments that bound parents and children, husbands and wives, and wider kinship networks. ¹ People labelled ‘criminal tribes’ actively resisted the regulation of their familial relationships, in particular, parent–child ties, and frequently disagreed with colonial officials’ perspectives on what constituted a ‘family’. Yet, familial ties were not only a zone of conflict between ‘criminal tribes’ and the state but also of cooperation, or at least, situationally coinciding interests.

Numerous historians have suggested that sedentarization was key to the CTA, ² and some existing studies have examined the CTA’s spatial dimensions. ³ However, my contention that the ‘criminal tribes’ campaign repositioned families in space goes against the grain of existing literature, which has examined two main questions. First, what factors caused the CTA: colonial anxieties about mobile populations, the dismantling of indigenous policing systems ⁴ or a wider range of situational anxieties? ⁵ Second, was the ‘criminal tribe’ stereotype a colonial invention ⁶ or did it have roots in indigenous notions of criminality? ⁷ Studies of the CTA have neglected some important questions of social history, among them: what do the experiences of people labelled ‘criminal tribes’ tell us about the history of the family in South Asia? ⁸ Historians have taken it for granted that the CTA usually policed entire families. ⁹ However, the familial aspects of the CTA were not merely a product of colonial ideologies of hereditary criminality. ¹⁰ Rather, family—meaning both ‘a set of lived social relations’ and a ‘normative ideal’ ¹¹ —was at the centre of the CTA regime. I view the CTA as a project to colonize subaltern families and analyse how certain peoples and familial arrangements were produced as subaltern. ¹² It is particularly surprising that historians have not scrutinized the CTA’s familial dimensions, given that numerous scholars have shown that colonial population management (or ‘governmentality’) depended on the regulation of households and familial relationships. ¹³

In this article, I explore a particular type of family tie—marriage. Several historians, including Mytheli Sreenivas, have argued that in the 1800s, ‘middle-class’ Indians constructed a ‘conjugal family ideal’ that positioned the husband–wife relationship as the family’s ‘central axis’. The British regarded marriage as ‘too important to remain outside the purview of state control’ but Indians bent colonial laws to their interests. ¹⁴ Historians have explored these tussles around marriage largely in the context of elite social groups. ¹⁵ However, some Dalit studies scholars—including Charu Gupta and Anupama Rao—have explored shifts in marriage practices. In the late 1800s, colonial laws encouraged or enforced caste endogamy in Dalit communities, which had more diverse marriage practices. ¹⁶ Dalit men asserted patriarchal control over women and ‘created fresh norms of conjugal relations, embedded in endogamy and containment of sexuality’, ¹⁷ even as some twentieth-century Dalit leaders advocated for inter-caste marriages in order to dismantle caste. ¹⁸ However, Dalit studies scholars have not examined in depth those groups labelled ‘criminal tribes’, many of which were low-caste and socially marginal. It is unclear how Dalit conjugalities differed from, or intersected with, the politics, practices and state regulation of marriage in so-called ‘criminal tribes’.

This article examines how marriage became a site of contestation and cooperation in criminalized peoples’ interactions with the colonial state, providing a new vantage point on the minutiae of everyday life under the CTA, while also shedding light on the history of conjugality in modern South Asia. It focuses on the criminalization of Sansiyas (who are also known as Sansis) in the North Western Provinces and Oudh (NPWO) in the 1890s. ¹⁹ The anti-Sansiya campaign was the NWPO’s first attempt to police ‘nomadic’ people who appeared to not have permanent residences. In 1890, the administration forcibly migrated Sansiyas up to 900 kilometres from the western to the eastern parts of the province and placed (or ‘planted out’) ‘isolated’ families in villages. In addition, the NWPO detained the apparently ‘more criminal’ Sansiyas in a penal settlement at Sultanpur.

In 1891, the NWPO launched a province-wide matchmaking campaign in which district Magistrates became marriage brokers who located suitable brides and grooms for Sansiyas. This matchmaking scheme emerged for two main reasons. First, Sansiyas’ strategies and demands propelled colonial involvement in marital matters. Many parents could not find suitable marriage partners for their children, due to the fracturing of their social networks, or afford marriage expenses, due to underpaid forced labour. Thus, criminalized people persistently lobbied the government for assistance in arranging and celebrating marriages under conditions of criminalization. Second, under the CTA, provincial governments attempted more intensified marital interventions in populations designated ‘nomadic’ or ‘wandering criminal tribes’ (like Sansiyas), than in so-called ‘settled criminal tribes’—defined as groups in which women and children were sedentary and men were seasonally mobile—even as this nomadic/settled distinction was unstable and shifting. Of course, in other settings, colonial governments also regulated specific marriage practices. Anti-infanticide campaigns, for instance, included limits on marriage expenses. But the colonial governance of conjugality was highly contextual. In the policing of ‘criminal tribes’, provincial governments showed an uneven concern with marriage practices, which varied between criminalized communities and over time. In the case of ‘nomadic’ criminalized groups, the state attempted more significant transformations in the relationships between individuals, families, social groupings and space. Since officials aimed to build these new social geographies from family units, marriage and other kinds of familial relationships took on a magnified importance. Separating ‘isolated’ Sansiya family units from wider kinship and social ties required officials to define the boundaries of the family, chart changes in family membership and, thus, manage marital relationships.

This article primarily draws upon: official correspondence and reports authored by British administrators and a few Indian officials; petitions from criminalized people summarized in provincial proceedings volumes; colonial ethnologies; and English and Urdu newspaper articles. ²⁰ These historical sources describe the political economy of households, as well as conjugal and familial norms, in an impressionistic, criminalizing and often contradictory manner. While I tease out these ambiguities, the source material imposes certain limits on the analysis of criminalized communities. Moreover, the NWPO classified a wide variety of peripatetic and socially marginalized groups as ‘Sansiah’ (the main colonial spelling of Sansiya/Sansi). It is impossible to determine what proportion of the official Sansiya population identified as such or pin down the full range of social groups included under the Sansiya umbrella. Thus, when I use the term Sansiya below, I am primarily referring to the colonial population category.

Marriage, Prostitution and the Making of the Sansiya Category

Between 1850 and 1870, the colonial ‘criminal tribe’ stereotype solidified. Colonial officials believed that from ‘time immemorial’, certain castes had been ‘destined by the usage of caste to commit crime’. ²¹ The criminal tribe stereotype emerged from metropolitan British anxieties about gypsies and notions of ‘habitual’ criminality; tropes of savagery, which the use of ‘criminal tribe’ rather than ‘caste’ evoked; ²² and pre-colonial Indian representations of ‘robber castes’ that generationally transmitted expert knowledge of crime. ²³ Under the CTA, the authorities could either restrict registered people to their existing residence or detain them in a penal settlement. Between 1871 and 1890, the NWPO favoured the first, less expensive option and only applied the CTA to people with permanent residences. ²⁴ The NWPO classified a diverse mix of socially marginalized people as criminal tribes, including: so-called ‘vagrant tribes’, which in fact had varied patterns of migration that were often seasonal and/or gendered; ‘hunting tribes’, which commonly combined cultivation, hunting and gathering of forest products; ²⁵ a number of communities described as ‘untouchable’; ²⁶ and groups that policed a ruler or landlord’s domain and raided rivals’ areas of influence, which Piliavsky has recently termed ‘raider-protectors’. ²⁷

The colonial Sansiya category was especially amorphous. In 1887, A. Ollivant claimed that ‘[t]he names they go by differ … in different localities’, including Kanjars, Rudhua Kanjars, Berias (or Bedias), Bhatus (or Bhantus), Gidhias, ‘Chandu-Sansiahs’, Ghurkals and ‘Bans-Sansiahs’. ²⁸ Ollivant classified all these groups under the term ‘Sansiah’, since this was what ‘these people’ were known as in Aligarh, where they were ‘most turbulent’. ²⁹ Crooke’s authoritative 1896 Tribes and Castes volumes—which drew together archived NWPO records and Crooke’s correspondence with British officials and elite Indian men—listed Sansiyas, Kanjars, Bhatus and Berias as separate castes and mentioned ‘Gidiyas’ briefly in the Sansiya entry. But Crooke claimed that all these groups were ‘vagrants’ who were ‘very closely allied if not identical’. ³⁰ The incorporation of various communities into the official Sansiya category rested on the ‘intermarriage’ thesis: since many jatis (birth groups or ‘castes’) were endogamous, administrators reasoned that marriage connections would show which caste affiliations were ‘disguises’ for Sansiya-hood. The archived evidence for Sansiya conjugal ties to other groups was flimsy. ³¹ Nevertheless, the Inspector-General of Police claimed that ‘the facts of intermarriage’ proved that the Aligarh Sansiyas were ‘absolutely one and the same people’ as Kanjars, Bhatus and Berias. ³²

In 1890, Sansiyas were proclaimed a ‘criminal tribe’ in Muzaffarnagar, Meerut, Aligarh and Mathura districts. Gidhias, Kanjars and Radhua Kanjars were included in the 1890 proclamation, and were thus officially defined as Sansiyas, but Bhatus and Berias were not. ³³ In practice, the police paid little attention to the ethnographic distinctions of the proclamation and indiscriminately rounded up a variety of people who appeared (to police) to be Sansiyas due to their mobility, livelihoods or marginality. ³⁴ The NWPO later concluded that the authorities had registered large numbers of Bhatus and Berias, who were not included in the 1890 proclamation, as well as Nats, whom the NWPO did not even consider a ‘criminal’ group at this time. ³⁵ At any particular point in the 1890s, the precise mix of communities classified as Sansiyas is unclear.

Given the diverse people who were categorized as Sansiyas, it is difficult to determine registered peoples’ mobility patterns, social practices and livelihoods. While Sansiyas were reportedly ‘nomadic’, several people registered as Sansiyas had land and were tenant farmers. ³⁶ Despite characterizing these various groups as ‘vagrants’, Crooke noted that their migration patterns were seasonal. Berias and Kanjars were largely sedentary during the monsoon, when several gots (exogamous groups/families) gathered together. Some Kanjars and Sansiyas had apparently ‘settled down’. ³⁷ As with many communities designated as ‘criminal tribes’, Sansiyas and associated groups appear to have been socially marginalized, but the precise character of this marginality is less clear. Colonial scholar-officials believed that Sansiyas, Kanjars, Berias and Bhatus were of ‘degraded’ or untouchable caste status but differed on whether they claimed a higher status than ‘sweepers’. ³⁸ An 1871 Urdu newspaper referred to Sansiyas as a ‘vagrant tribe’ but I have not yet found explicit discussion of their caste position in the press. ³⁹ There were hugely disproportionate rates of incarceration among men from communities criminalized as Sansiyas, which were clearly the targets of systemic police harassment. ⁴⁰ Colonial accounts typically characterized Sansiyas and linked criminalized communities as thieves, beggars and prostitutes (the latter is further discussed below). But colonial commentators mentioned other forms of work, especially cultivation and gathering forest produce. Crooke wrote that Kanjars were often stone-cutters, cultivators, agricultural labourers or sold goods they made from gathered forest materials, including mats and door-screens. Sansiya women sold ‘roots and other jungle medicine’ ⁴¹ as well as grass and firewood they collected. ⁴²

Official knowledge of the marriage ‘customs’ of people classified as Sansiyas was contradictory and sketchy. In 1890, the NWPO Secretary simply assumed that ‘the Sansiah ritual … is presumably in substance the same as the Hindu ritual’. ⁴³ In contrast, Crooke’s 1896 ethnography emphasized differences between the marriage rituals of ‘respectable’ upper-caste groups and those of Sansiyas and Kanjars, whose marriages reportedly involved ‘simple’ rituals. These included the groom’s ritualized ‘pretence of seizing the bride by force’ and the circumambulation of the bride and groom seven times. ⁴⁴ Colonial commentators disagreed on whether people classified as Sansiyas practiced dowry or bride price. In the mid-late 1800s, dowry solidified as a status marker among upper- and intermediate-caste Hindus, who condemned bride price as ‘dishonourable’. ⁴⁵ Many lower castes also increasingly practiced dowry. Samita Sen has argued that the growing incidence of dowry, along with hypergamy and child marriage, was due to a ‘devaluation of women’s work’ that ‘meant … parents were less willing to undertake the maintenance of daughters for longer periods’, ⁴⁶ while in contrast, Veena Talwar Oldenburg asserts that dowry could be an ‘economic safety net’ for women. ⁴⁷ It is difficult to locate groups classified as Sansiyas within this debate, as it is unclear whether they practiced dowry or bride price. In 1896, NWPO officials claimed that all Sansiyas practiced bride price—presumably including Kanjars who were officially classified as Sansiyas. ⁴⁸ In contrast, in the same year, Crooke claimed that Sansiyas practiced bride price but, conversely, Kanjars practiced dowry. Sansiya grooms’ families reportedly paid up to ₹300 or ₹400 in bride price. Kanjar fathers of grooms paid for some expenses, including a betrothal feast, but the bride’s father gave a dowry of ‘whatever he could afford’. ⁴⁹ Moreover, in light of the registration of Bhatus, Berias, Nats and possibly other communities as Sansiyas, the marriage customs of criminalized people undoubtedly varied. Of course, this complicates the task of explaining the causes and effects of colonial interventions into criminalized peoples’ conjugal practices.

While colonial accounts of marital practices were vague and conflicting, there was a resounding official consensus that Sansiyas were ‘prostitutes’. J. Kennedy, Aligarh’s Magistrate, wrote that Sansiya ‘women … live chiefly by begging and prostitution’. ⁵⁰ Kanjars, who were officially classified as Sansiyas, also apparently practiced sex work: ‘The women are usually prostitutes and the men thieves’. ⁵¹ Moreover, colonial bureaucrats claimed that prostitution was the trade of Berias (who were not proclaimed in 1890 but were often linked to Sansiyas): ‘the majority of the womankind are prostitutes, and the men live on their earnings’. ⁵² Crooke represented Beria women as ‘prostitutes’ but not Sansiya or Kanjar women. ⁵³ Yet NWPO officials portrayed almost all ‘nomadic’ women as sex workers.

This was a magnified example of the colonial stereotype of the sexually immoral ‘criminal tribeswoman’, which had consolidated in the 1860s and 1870s. Sexual and conjugal practices featured heavily in colonial accounts of the ‘criminal tribes’, especially the ethnographic reports of district administrators and published caste and tribe glossaries. The NWPO’s reports to the central government justifying a group’s designation as a ‘criminal tribe’ referred to marriage and sexual ‘customs’ to prove the group were a ‘distinct tribe’ that was ‘criminal as a whole’. ⁵⁴ British administrator scholars criticized criminalized communities’ conjugal practices when they did not adhere to dominant-caste norms. ⁵⁵ Colonial commentators argued that criminalized women’s sexual immorality was hereditary: ‘generation succeeded generation in which the men were professional thieves, and the women openly immoral’. ⁵⁶ Administrators and ethnographers claimed that criminalized women had sex before marriage, lived with men who were not their husbands, or were in the ‘keeping of’ landlords or village headmen. ⁵⁷

Colonial portrayals of ‘criminal tribeswomen’ drew upon dominant-caste Hindu representations of Dalit and Shudra women as sexually immoral, since the British believed that most criminalized groups were of low-caste status. From the mid-1800s, educated, professional men from upper-caste Hindu ‘service gentry’ communities increasingly identified as ‘middle class’ and distinguished themselves from both the old elite of rulers and nobility and the lower orders, through their claims to morality and respectability. ⁵⁸ These men defined the morality of dominant-caste women through contrasts with Dalit and Shudra women, who were characterized as sexually available, obscene, loud and morally corrupting kutnis (vamps or pimps). Dalit women were particularly subject to such stereotypes because they often worked in public spaces alongside men, which was ‘unfeminine’ according to dominant-caste, middle-class norms. ⁵⁹ These elite Indian representations shaped hypersexualized colonial accounts of low-status ‘criminal tribes women’. Yet British officials assumed that criminalized women in peripatetic communities were more ‘immoral’ than primarily sedentary women, who reportedly had extramarital sex but were not explicitly labelled ‘prostitutes’, like ‘nomadic’ women. ⁶⁰ Apparently, peripatetic women were entirely beyond the bounds of ‘respectable’ ‘settled’ society, in their both mobility patterns and sexual practices.

The Sansiya ‘prostitute’ discourse might suggest that colonial commentators roundly denounced prostitution but colonial policies were highly equivocal. Under 1864 and 1868 contagious disease laws, a system of regulated brothels was established to provide (especially British) men with sexual services, while protecting them from venereal disease through the registration and medical examination of women brothel workers. ⁶¹ The colonial state was heavily invested in sexual commerce. Hence, colonial courts often narrowly interpreted the 1860 Penal Code to limit the prosecution of traffickers and employers of child sex workers, ⁶² which the colonial state was itself ‘complicit’ in. ⁶³ At the same time, colonial courts attempted to suppress diverse female performers classified as ‘temple dancing girls’ on the basis that they were ‘prostitutes’, thereby defining ‘a vast residual category’ of ‘sexual activity’ beyond dominant-caste Hindu marriage norms as ‘prostitution’. British judges defined the dedication of girls to temples (including by parents) as ‘disposing’ or ‘obtaining possession’ of a minor for ‘the purpose of prostitution’ under Sections 372 and 373 of the Penal Code. ⁶⁴ In sum, the colonial government variously criminalized, regulated or facilitated practices defined as ‘prostitution’. Nonetheless, nebulous colonial definitions of prostitution leant weight to provincial governments’ applications to notify certain communities as ‘criminal tribes’.

The Sansiya ‘prostitute’ stereotype obfuscated the sexual violence that criminalized women experienced in their interactions with police. In 1888, Hobart, the Inspector-General of Police, warned that when police escorted mobile Sansiya ‘gangs’ from one police jurisdiction to another (a common practice), ‘the solitary policeman … is ordinarily too human to resist the blandishments or the violence of the females; and in this connexion, it may be noted that prostitution, certainly free love, is common among the tribe’. ⁶⁵ What should we make of this account? From the 1860s, some ‘criminal tribeswomen’ had formed long-term relationships with Indian policemen and thereby avoided registration. ⁶⁶ Yet these relationships occurred within a wider context of both sexual commerce and violence at police stations. ⁶⁷ Evidence from the early 1900s suggests that many criminalized women experienced sexual harassment and assault by police. ⁶⁸ Such abuse was partly enabled by official discourses of criminalized women’s sexual immorality. Similarly, Charu Gupta argues that dominant-caste Hindu narratives of Dalit women’s sexual availability erased women’s experiences of sexual violence at the hands of higher-status men. ⁶⁹

Space, Family and Marriage

Familial relationships had been pivotal to the enforcement of the CTA since its passage in 1871. Authorities usually registered entire families, while individuals were recorded on registers under their ‘head of household’. Police kept domestic spaces under surveillance. The NWPO’s aim was to produce industrious families. ⁷⁰ However, the Sansiya project attempted to redraw the relationship between families, social groups and space in much more drastic ways than had been attempted in groups confined to their own villages.

The NWPO’s 1890 plan to control the Sansiya population involved geographically redistributing family units. In the government’s view, the Sansiya community consisted of large, mobile ‘camps’ with fluctuating memberships that were impossible to fix, in both colonial knowledge and place. Thus, the provincial government broke up these larger social formations by forcibly relocating Sansiya families from the western districts in which they were ‘concentrated’ (Muzaffarnagar, Meerut, Aligarh and Mathura) ⁷¹ and scattering them across the eastern parts of the province. ⁷² In 1890, the NWPO Secretary, J. Woodburn, wrote that the geographical dispersion of families offered a ‘hope of reform’: ‘Members of the tribe amenable to influence … may in detached families … take to agriculture, and, under the pressure of isolation, abandon the criminal pursuits of their race’. ⁷³ Authorities aimed to place one or two families in each village because ‘families should be isolated as far as possible, not concentrated’. ⁷⁴ This was referred to as the ‘planting out of families in districts’, ⁷⁵ a horticultural metaphor for ‘scattering’ seeds (that is, families) and rooting them in place, which expressed both the spatial dimensions of the Sansiya project and government aspirations to extend agriculture by turning the ‘criminal tribes’ into cultivators. ⁷⁶ The family was central to colonial agrarian schemes. Colonial officials did not recruit single men to migrate to the Punjab Canal Colonies but rather, recruited entire families. ⁷⁷ When the Andaman Islands authorities attempted to create a self-sustaining penal colony through the agricultural settlement of convict couples, they assumed that ‘a healthy society’ was ‘one that was based on stable families’. ⁷⁸ The ‘planting out’ of Sansiya families echoed these agrarian projects but officials sometimes doubted Sansiyas’ suitability for cultivation and proposed ‘sweeping and scavenging or other more menial forms of labour’. ⁷⁹ In the context of the CTA, ‘planting out’ was a novel project: the NWPO had not previously attempted forced transportation and resettlement on this scale. ⁸⁰

Criminalized people were relocated to the estates of zamindars and taluqdars; these land-holders were responsible for providing employment and keeping families under the surveillance of their chaukidars (watchmen) and agents. ⁸¹ The British had long attempted to co-opt landholders into criminal law enforcement to produce a ‘domesticated’, loyal elite. ⁸² ‘Planted-out’ people could be punished if they refused to do assigned work. To travel beyond their new village, they needed a pass from the landlord or, if their travel was for more than 14 days, from the Magistrate or District Superintendent of Police. ⁸³

The provincial administration planned to ‘plant out’ only the ‘least criminal’ Sansiyas. The relatively more ‘criminal’, both ‘single men and families’, would be ‘treat[ed] with restraint’ in a penal settlement located in an old jail at Sultanpur, 500–800 kilometres from Sansiyas’ reputed ‘chief resort’ in western NWPO. ⁸⁴ Notwithstanding official instructions that ‘jail discipline’ should not be enforced, Sultanpur was ‘in all respects a prison’, in its both architecture—it was a converted jail—and administration. ⁸⁵ Moreover, the provincial government planned to separate some Sultanpur children from their parents and transport them approximately 500 kilometres away to a ‘Reformatory Settlement’ at Fatehgarh. The Government of India ordered that ‘as the Act does not authorize the separation of children from their parents or guardians, it will only be possible to send them to Fatehgarh (a) by consent or (b) for a breach of discipline [by the child or their parents]’. ⁸⁶ Yet official records indicate that the NWPO fulfilled neither of these conditions and child removal was forcible. ⁸⁷ Even Lieutenant Governor Colvin later admitted that ‘the law may have been somewhat strained’. ⁸⁸ The official policy on the spatial relocation of families seemingly pulled in opposite directions: ‘planting out’ reinforced the family unit, whereas the removal of children from their parents fragmented families to ‘reform’ children.

On 13 October 1890, police attempted the forced transportation, by foot and train, of 1,687 people. Around 1,000 people from 190 families were allocated to ninety-three different landlords and Court of Wards estates in thirteen districts. The remaining population was earmarked for Sultanpur. ⁸⁹ Berrill, a high-ranking police official, claimed that ‘no violence was resorted to’ in capturing registered people. But Berrill’s definition of ‘violence’ was extremely narrow: he noted that at least one district, Muzaffarnagar, ‘sent some of the men down in leg irons’. Many criminalized people arrived with little besides the clothes on their backs, having been forced to leave behind their household possessions and animals. ⁹⁰ Some British officials ‘simply deposit[ed] Sansiahs in the villages’ without ensuring that the landlords would provide employment. ⁹¹ One man died when he jumped from a moving train in a bid to escape, which indicates the acute distress of the forcibly transported. ⁹² Scores of people bolted when the police attempted to capture them. By 8 November, only 1,077 out of 1,687 registered Sansiyas had ‘been secured’. ⁹³ Nonetheless, Sansiyas were the second-largest ‘criminal tribe’ population in the NWPO. ⁹⁴

The chaotic process of transporting criminalized people produced considerable correspondence about individuals’ marriage and familial relationships. ⁹⁵ British officials found that relocating family units was complicated because the composition of families changed over time, due to marriages, as well as births and adoptions. A woman’s registration status usually followed that of her father and after marriage, her husband. ⁹⁶ Berrill reported that the ‘planted-out’ population had changed due to the addition of people who ‘married into the proclaimed districts’ and the subtraction of some who ‘married into non-proclaimed districts’. ⁹⁷ Forty-one ‘excess’ people arrived in Saharanpur from Meerut district, due to ‘men having … taken more wives and adopted other children’. ⁹⁸

When the police rounded up people, they occasionally separated married couples. To reunite husbands and wives, the provincial administration tried to pin down conjugal relationships. Woodburn, the NWPO Secretary, issued specific instructions on several couples, including that ‘the wife of Mohan should be sent to her husband’ and ‘in Aligarh, the woman, belonging to a man sent to Sultanpur, should be sent there also’. ⁹⁹ This policy of keeping couples together meant that the government had to negotiate with men and women about their marital status. In March 1891, ‘a few married women [in Sultanpur] stated that their husbands were interned out in villages, and asked to be allowed to join them’. Colvin ordered that ‘[g]reat care should be taken that in every such case the woman is released and sent to her husband’. ¹⁰⁰ In other instances, women refused to join men who claimed to be their ‘husbands’. When the Prison department ordered the ‘wife and child’ of Hardeo to be sent from Meerut to Sultanpur, where Hardeo was detained, the Meerut Superintendent of Police reported, ‘it appeared that the woman was not the wife of Hardeo, nor was the child his, and that she objected to join him’. The NWPO administration ‘declined to interfere in the matter’. ¹⁰¹ In sum, some women escaped Sultanpur by demanding to join their husbands, while conversely, others refused to go to Sultanpur and repudiated purported ‘husbands’.

A central assumption of the population redistribution project was that only families which were ‘complete’ from the colonial perspective—with a resident male ‘head’—could be ‘planted out’. In contrast, some districts ‘sent families without men’, whereas others did not ‘interfere’ with families ‘until the heads of those families could be accounted for’. ¹⁰² Colvin ordered that, ‘[t]he families of all Sánsiahs … who have no adult males that can earn a living for the families in agricultural employment’ should be sent to Sultanpur. ¹⁰³ Since the late 1880s, NWPO officials had expressed concern about ensuring the material support of Sansiya women and children, given mass male incarceration. In 1888, Hobart, the Inspector-General of Police, warned that ‘a scheme for supporting … [Sansiya] women at the expense of Government’ might be necessary because the incarceration of men left ‘the women and children destitute’, as ‘prostitution and beggary seem to be their only resources’. ¹⁰⁴ British officials repeatedly interlaced sexual immorality with economic unproductivity in framing Sansiya women as a ‘problem’, highlighting the tight connections between sexual and economic norms in colonial conceptions of a governable population. ¹⁰⁵ During the ‘planting out’ project, British officials assumed that without the penal discipline of Sultanpur, Sansiya women would not ‘settle down to any work or industrial calling’, ¹⁰⁶ least of all agriculture, which they did not ‘know anything about’. ¹⁰⁷ The NWPO’s position was that in the absence of a male breadwinner, the state needed to directly assert its paternal power over women and children through their interment in Sultanpur.

Meanwhile, in the process of relocating ‘detached families’, ¹⁰⁸ British administrators defined what constituted a family. But people registered as Sansiyas had more expansive notions of family than British (and possibly Indian) colonial officials. Several complained that they were not relocated with their ‘entire family’. ¹⁰⁹ Some criminalized people successfully ‘pleaded’ with officials ‘to accompany their relatives’, instead of going to their assigned district, making the spatial distribution of population uneven, to the administration’s frustration. ¹¹⁰ Registered people adopted numerous strategies—from lobbying authorities to escaping—to remake families fractured by forced transportation, and more generally, to evade the CTA regime. By March 1893, over 400 ‘planted-out’ people had absconded and only 205 registered people remained. ¹¹¹ ‘Planting out’ ostensibly bolstered the family unit but in actuality, colonial authorities broke up and reconstituted families, often unsuccessfully.

Marital status, age and gender structured the spatial organization of the Sultanpur settlement to a degree that the NWPO had not previously attempted in the implementation of the CTA. The government planned to confine ‘single men and families’ who were considered especially ‘criminal’ in Sultanpur. But as we have seen, officials sent substantial numbers of unmarried women to Sultanpur, as well as married women with absent or imprisoned husbands. As such, the 1890 Sultanpur rules dictated that ‘Part of the building will be allotted … for the accommodation of Sánsiah families; a second portion for the accommodation of single men; a third for single women; and a fourth for the old and infirm’. ¹¹² To facilitate spatial segregation, the Sultanpur management compiled lists showing ‘the civil status’ of each inmate, ‘i.e., married, unmarried, widower or widow’. ¹¹³ Officials enforced a division of labour based on gender and marital status. Adult men did construction work, tended the settlement ‘garden’, ¹¹⁴ made baskets and rope, wove cloth, parched gram and made bricks. Only men who had ‘families in the Settlement, whom they might be unwilling to abandon’, cultivated fields outside the compound—evidence of the penal manipulation of marital and familial ties. ¹¹⁵ The Sultanpur management assigned feminine-coded occupations to women, whose work was determined by their age, marital status and the age of their children. Women ‘with infants or very young children’ were ‘not employed on ordinary labour’ but cooked and ground grain for their families. Additionally, there were three categories of ‘labouring females’: ‘first, young unmarried women, who live[d] together’ and spun cotton; second, ‘young married women’ without children, who spun wool; and third, ‘married women of middle age or over, without young children’, who tore ‘wool with their fingers’. ¹¹⁶ Conjugality, familial relationships and gender norms shaped the envisaged organization of space and work in Sultanpur in much more fine-grained ways than in the policing of ‘planted-out Sansiahs’ and other criminalized communities in the late 1800s. More ambitious spatial projects entailed more intensified governance of kinship, marriage and gender—precisely because the family was central to the new social geographies that the NWPO hoped to create through the CTA.

Criminalized Peoples’ Conjugal Strategies and Government Matchmaking

People classified as—persistently lobbied to carry out their marriage practices under conditions of enforced sedentarization and detention; in some cases, demanding the direct involvement of colonial authorities in marital matters. In early 1891, the question of how Sultanpur inmates would pay for wedding expenses arose. ¹¹⁷ Sultanpur detainees earned only 5 per cent of the estimated produce of their work or 5–10 per cent of the estimated wages. ¹¹⁸ When a man named Kallu agreed to marry his daughter, Musammat Sona, to Bhagu, the groom’s father, Darba, asked for permission to borrow ₹30 from another Sultanpur inmate to spend in the purchase of ‘ghí, gur, rice, and cloth’ for distribution on the occasion. ¹¹⁹ On the advice of the two British and three Indian members of the Sultanpur Committee of Visitors, the provincial government banned inmates from lending or borrowing money but sanctioned ‘the distribution of extra rations’ of food and cloth at weddings if the family had insufficient funds. The government’s aim was to prevent indebtedness and encourage inmates to save their meagre wages. Darba and Kallu had also petitioned for permission to ‘sing and dance for two days when all the Sansiahs might be exempted from work’ and for use of a dholak (drum) and sarangi (string instrument). ¹²⁰ Instead, the NWPO permitted Sultanpur detainees to celebrate weddings on Sundays, their day off, and to use a dholak only. ¹²¹ Hence, officials aimed to minimize the impacts of weddings on labour extraction and discipline.

In other contexts, colonial population management also involved the regulation of marriage expenses, though significantly, officials were usually more concerned with elite marriage celebrations. For example, in the mid-1800s, colonial governments drew up agreements on permissible marriage expenses with dominant-caste Hindu groups, ostensibly in a bid to prevent female infanticide, which the British claimed was caused by excessive dowry demands. ¹²² Colonial governments sought to eliminate ‘frivolous waste’—such as (relatively minimal) expenditure on wedding-service providers ¹²³ —to instruct Indian elites in thrift and productivity. While dominant-caste men accepted limits on various wedding costs, they successfully resisted colonial proposals to cap dowry expenses. ¹²⁴ Radhika Singha argues that anti-infanticide campaigns targeted the ‘fiscally ruinous’ status strategies of indigenous elites and the terms on which their ‘rank and patriarchal right[s]’ could ‘coexist with colonial rule’. ¹²⁵ In contrast, the Sansiya ‘experiment’ aimed to create a sedentary working poor. Nonetheless, both infanticide policies and the management of marriage celebrations in Sultanpur inculcated colonial economic and consumption norms.

Colonial officials also became heavily involved in matchmaking among criminalized people. In March 1891, the NWPO launched a marriage-arranging scheme, the immediate impetus for which was the residence in Sultanpur of both young women and (apparently) disorderly men. Colvin, the Lieutenant Governor, was worried that:

The presence of so large a number of young women in the [Sultanpur] Settlement is obviously undesirable, and is outside the object of the Settlement, which is to intern the older people of either sex and provide them with means of employment; or the more ruffianly and violent of the men, who are thus kept out of crime. ¹²⁶

These women, whom Colvin estimated were between eighteen and twenty-six years old, ‘represent[ed] themselves either as unmarried women’ or said that their husbands were imprisoned or ‘wandering about’. Colvin ordered that these ‘girls’ be ‘made over’ to ‘their husbands or other relations … or if it is found that any of the Sansiahs living in the districts under surveillance are prepared to marry one or other of the girls she can be forwarded to that district’. ¹²⁷ In June 1891, district Magistrates received a list of thirty-five ‘marriageable young girls’ from Sultanpur between the ages of twelve and twenty-five, and a second list of fifty-three unmarried ‘Sansiah males from the age of twelve and upwards’ who were registered in the districts, including men as old as seventy. ¹²⁸ J. Woodburn, the NWPO Secretary, ordered that Magistrates of districts ‘in which there are settled Sansiah families, proclaimed or unproclaimed, with unmarried males, will communicate to the heads of those families the details [of marriageable “girls”] received from the Sultanpur Settlement’. ¹²⁹

In an important article on colonial criminal law, Sandria Freitag briefly analysed this marriage-brokering scheme, viewing it as a product of ‘administrators’ concern … to ensure that women in the penal settlement did not mate with the male Sansiahs who surrounded them’. ¹³⁰ Certainly, officials fretted about unmarried women’s interactions with married and unmarried men in Sultanpur. The NWPO government also hoped to marry women to unregistered male Sansiyas, as well as the ostensibly ‘less criminal’ registered men in the districts, thus removing them from the influence of the ‘ruffianly’ men of Sultanpur. ¹³¹ According to Freitag, the matchmaking project assumed that ‘criminality was genetically transmitted’ and women ‘functioned simply to pass on more or less criminal genes’. ¹³² The strongest evidence for this is the late-1880s suggestion of NWPO official A. Caddell that ‘the imprisonment of the women for lengthened periods (when found guilty of offences) would help to prevent an increase in the members of these tribes’. Yet Caddell’s proposal was to reduce Sansiya births overall, not to encourage Sansiya women to procreate with ‘less criminal’ Sansiya men. ¹³³ Beyond this one example, I have not found any eugenicist discussion of either ‘breeding out’ criminality or reducing the Sansiya population in this period in the NWPO.

The presence of women in Sultanpur without husbands or male relatives was, rather, a wider problem of population management. In the dominant colonial vision of a governable population, women were subject to patriarchal authority. ¹³⁴ For some time, NWPO officials had worried that Sansiya women with absent husbands posed both moral problems (‘flagrant immorality’) and economic problems (destitution and begging), which might necessitate ‘a certain amount of State support’. ¹³⁵ From 1890, Sultanpur’s female population had grown precisely because officials thought it necessary to assert intensified state control over women and girls in the absence of their adult male relatives. Having assumed this responsibility, the NWPO administration sought to reduce the burden on the state by placing women and girls with either male relatives or husbands, suggesting that sexual reproduction was not the primary concern. Hence, in March 1891, Colvin ordered that young women should be ‘pass[ed] … to the relations or to be married’. ¹³⁶ Freitag argues that the matchmaking project required ‘the government … to ignore the characterization that “the women were almost as bad as the men”’. ¹³⁷ In contrast, the problem of Sansiya femininity was at the centre of official decision-making. The immorality and criminality of Sansiya women required their subjection to the control of the colonial state, or, given the actual limits of its reach, of relatively ‘less criminal’ Sansiya men. ¹³⁸ Thus, the NWPO aimed to marry Sultanpur women to men outside the settlement, not due to a fixation with distinguishing ‘more or less criminal genes’, ¹³⁹ but rather in an effort to distinguish between men who were more or less capable of exercising patriarchal power.

Another factor that propelled the government matchmaking scheme was the demands of criminalized people that the government assist them in marrying their children. In 1889, Spedding, Banda’s Magistrate, correctly predicted that criminalized people would ‘hardly submit to be deprived of the right to intermarry’ if their ‘communication with each other for arranging marriages’ was cut off through forced relocation. ¹⁴⁰ The transportation of prospective brides and grooms to different places interrupted a number of planned matches, including four previously arranged engagements noted on an 1891 list of marriageable ‘girls’ in Sultanpur. ¹⁴¹ By June 1891, ‘Complaints [had] been made from settled Sansiahs that families, in which marriage would have been arranged, have now been interned in Sultanpur’. ¹⁴²

Criminalized peoples immediately availed themselves of the government matchmaking programme. Within the first year, twenty-three of the thirty-five ‘young unmarried women’ at Sultanpur had married either registered or unregistered men in the districts. ¹⁴³ In 1895 alone, ‘53 young girls were allowed to leave the [Sultanpur] settlement on their marriage’. ¹⁴⁴ The annual reports also mention several marriages between ‘planted-out’ people. In 1895 in Dehradun district, a registered woman from Raynorpur estate married a registered man from Doiwala estate. Another registered woman in Budaun district relocated to Bareilly upon her marriage. ¹⁴⁵ As the ‘planted-out’ population dwindled due to widespread absconding, it is likely that Sultanpur women and girls increasingly married unregistered men and boys. Moreover, fewer officially mediated marriages probably occurred between ‘planted-out’ people because illegal mobility facilitated marriage arranging—and absconding was much easier from villages than Sultanpur.

When arranging marriages, the authorities did not communicate with prospective brides but rather with their families and with possible grooms. A ‘petition presented by the Sansiahs located at Chunár’, southwest of Benares, ‘represent[ed] their inability to secure, at Chunár, husbands for their marriageable daughters’. ¹⁴⁶ The NWPO government accommodated the Chunar petitioners’ demands and ordered the Magistrate of Mirzapur to:

…communicate with the Magistrates of the districts in which the desired husbands lived with a view, if they were free Sansiahs, to their coming to Chunár to arrange about the marriage. But if the desired husbands were planted-out Sansiahs, the Magistrate of [their] district … might be asked to ascertain whether they agreed to the proposed marriage, and if they did, to arrange for their being sent to Chunár under proper custody with a view to the marriage being effected, after which they should be returned to their holdings with their wives. ¹⁴⁷

Magistrates were expected to ascertain the consent of the ‘desired husbands’ and the prospective brides’ relatives. The government’s order to consult directly with possible grooms may have curtailed the role of male relatives as marriage brokers for both prospective brides and grooms, which Crooke noted among Sansiyas and Kanjars. ¹⁴⁸

The extent to which women’s and girls’ choice of partner was a consideration for their relatives is difficult to determine from official sources. But the authorities envisaged a very limited role in the matchmaking process for the ‘marriageable’ women and girls themselves. Similarly, Clare Anderson suggests that marriages between convict men and women in the Andamans represented a male/state ‘extension of power’, rather than ‘the choices of convict women’. ¹⁴⁹ In the anti-Sansiya campaign, British administrators sometimes actively circumscribed women’s capacity to choose their partners. When Colvin initiated the matchmaking scheme, he insisted that ‘Care will … be taken at present that no intimation is given to these [Sultanpur] women of what is contemplated’. ¹⁵⁰ When a woman’s wishes differed from those of her relatives, officials followed the family’s view of whom the woman should marry, as the case of a widow named Musammat Husaini demonstrates:

Musammat Husaini, widow of Nathan, desired in August to marry a Nat of [Bijnor] district; but the Governor of the [Sultanpur] Settlement objected that her family wished her to marry Sabira, the very man whom the Magistrate had long been asking to have sent to Bijnor to help the widow. Husaini expressed her acquiescence in the match [with Sabira], but the bridegroom never appeared, and Musammat Husaini absconded in March with her children. ¹⁵¹

Husaini communicated her ‘desire’ to marry a Nat man to the authorities, resulting in an archival trace of her marital plans. But the Governor of Sultanpur clearly prioritized the ‘wishes’ of Husaini’s family and the Bijnor Magistrate’s opinion, over Husaini’s ‘desires’. We see a complex picture of Husaini’s agency: though she acquiesced to her relatives’ and the authorities’ choice of a husband, she eventually escaped the CTA regime by running away. ¹⁵² At the same time, some female detainees in Sultanpur were able to leverage marriage to leave the settlement and, if they married unregistered men, to escape registration under the CTA. ¹⁵³ Yet the colonial state allowed women and girls very limited scope to choose who their future partner would be, with such power being vested in their relatives, prospective grooms and officials.

The colonial state’s disregard of criminalized girls’ and women’s marital consent may seem to depart from a colonial concern with consent evident in various legal reforms, ¹⁵⁴ especially the 1891 Age of Consent Act. However, women’s choice in sexual and conjugal matters was not the primary concern that propelled the 1891 law, which raised the age of consent from ten to twelve. To be sure, Indian women reformers argued that early marriages structured by ‘clan, caste and kin’ undermined women’s ‘liberty’, and they sometimes promoted adult marriages by choice, rather than arranged marriages. ¹⁵⁵ But Pande argues that Indian and British men who argued for consent reform deployed eugenicist arguments that prioritized the ‘question of the health of the race and nation’, over women’s ‘well-being’. ¹⁵⁶ The autonomy of women and their ‘happiness’ in marriage rarely, if ever, concerned pro-reform British and Indian men. ¹⁵⁷ It is thus unsurprising that in the Sansiya matchmaking project, the NWPO did not attempt to raise the age of marriage above the new age of consent and, thus, considered twelve the age of male and female marriageability. ¹⁵⁸

When the NWPO attempted to marry off youths detained in Fatehgarh, the young women, young men and their families had little say in the matches. In 1894, the provincial administration proposed to marry Fatehgarh inmates upon their discharge (at 16–8 years old) and send the newly married couples to the Caribbean as indentured labourers. The NWPO hoped indentured migration would ‘rid this country’ of the Sansiya ‘terror’, while reforming Sansiyas through labour. ¹⁵⁹ The government also aimed to prevent Fatehgarh youths from returning to their relatives upon their release by permanently severing their familial ties. ¹⁶⁰ This family-separation aspect was quite different from the transportation to the Andamans of ‘criminal tribe’ Bhatus/Bhantus serving prison sentences along with their families in the 1920s. ¹⁶¹ Conversely, the Fatehgarh emigrants would create new families. Thus, the NWPO administration saw marriage as a prerequisite for the overseas migration of young people from Fatehgarh.

In early 1894, eight Fatehgarh ‘boys’ and eight ‘girls’ were married and entered into contracts of indenture, ¹⁶² with two couples going to British Guiana and the rest to Suriname. The records suggest their families were not involved in arranging the matches. In addition, thirty-one Sultanpur detainees emigrated to British Guiana, though it was the emigration of Fatehgarh ‘young couples’ that the NWPO publicized. ¹⁶³ The Times of India reported on this scheme in an article titled ‘Citizenising the Criminal’: the long-distance migration of young people joined ‘in the bonds of matrimony’ and the extraction of their labour, would make ‘citizens’ of the ‘criminal tribes’. ¹⁶⁴ The emigration project soon floundered due to opposition from sugar colonies worried about importing ‘criminals’, and from criminalized people, among whom emigration was ‘unpopular’. But it is further evidence that the colonial governance of the ‘criminal tribes’ involved the intertwined management of conjugality, labour and mobility. Hence, the NWPO continued to encourage the marriage of young people on their release from Fatehgarh. For example, in 1895, ‘Two boys and two girls were married and sent to work as warders in district jails, one couple to Lalitpur and the other to Jaunpur’. ¹⁶⁵

In their role as marriage brokers, district Magistrates had become intimately involved in conjugal and familial relationships. Yet, authorities were often vague on the finer points of marital practices. For instance, official correspondence does not show a concern with either endogamy or exogamy. ¹⁶⁶ The NWPO ordered that ‘Sansiahs’ should be married to ‘Sansiahs’, though this was an enormously broad and internally diverse official category. ¹⁶⁷ William Crooke reported that Kanjars and Sansiyas practiced forms of exogamy within ‘sub-castes’ or ‘sections’. ¹⁶⁸ But such considerations did not factor into British administrators’ descriptions of marriage-brokering. Two annual reports even noted marriages between registered ‘Sansiyas’ and ‘Nats’, without any comment on the implications of this for the official Sansiya category. ¹⁶⁹ Notwithstanding officials’ vagueness on criminalized peoples’ marital practices, matchmaking had become part of the routine policing of Sansiyas.

Hence, the Sansiya marriage-brokering scheme differed from the governance of conjugality in other ‘criminal tribes’ between 1871 and 1911. To be sure, the degree to which everyday life was policed under the CTA meant that district officials inevitably managed marriage practices through the pass system. In 1885, Special Officer Munshi Gurdyal reported that for arranging marriages and attending weddings, ‘the Barwars have to obtain passes to go to other villages with great but at the same time limited liberality’. ¹⁷⁰ Gurdyal had also instructed the: ‘head-constable [to] make a report with respect to such a woman who … obtains a pass and goes to another village where she is married: so that the register may be corrected, and she may be accounted for in the village where she goes to’. ¹⁷¹

Indian and British officials may have coerced criminalized people by occasionally withholding passes for marriage purposes. However, this quotidian management of marriage practices through the pass system differed substantially from the Sansiya matchmaking scheme, in which colonial officials, especially district Magistrates, effectively became marriage brokers. Why did the NWPO make more extensive and intimate interventions into conjugality in the case of Sansiyas, compared with other ‘criminal tribes’? The NWPO attempted more radical transformations in the social geography of the Sansiya population, which rested on the spatial redistribution of family units, meaning that marital relationships assumed a greater importance. Meanwhile, the social networks of criminalized people were severed to such a degree that they persistently demanded colonial interventions in matchmaking—even as women, girls and Fatehgarh detainees had little say in whom they married.

The Limits of Intimate Colonial Governance

In April 1896, a new Lieutenant Governor, Anthony MacDonnell, ended the programme of child removal and abruptly closed both Sultanpur and Fatehgarh. MacDonnell ordered that the 403 Sultanpur inmates and seventy-one Fatehgarh children be relocated to an ‘agricultural colony’ at Sahibganj, a tract of government-owned, uncultivated land in Kheri district. The 131 ‘planted-out’ people who had not absconded would remain in their various districts. ¹⁷² Perplexingly, in 1894–5, the province had pressed for the legalization of forced child removal under the CTA, ¹⁷³ but by 1896, the NWPO was reuniting parents with their children. ¹⁷⁴ J. P. Hewett, the India Home Department Secretary, rebuked MacDonnell for disbanding Fatehgarh, since the girls would be ‘removed from a life of probably virtue to one of shame and vice’. ¹⁷⁵

MacDonnell closed Sultanpur and Fatehgarh due to economic considerations, as well as a recognition of the limits of intimate colonial governance. First, Sultanpur and Fatehgarh were apparently expensive to run, whereas a ‘self-supporting’ settlement could bring a ‘financial saving to government’ of ₹538 per month. ¹⁷⁶ Second, ‘experiments’ with governing familial relationships had seriously stretched the provincial government’s capacity to transform criminalized peoples’ lives. ‘It was difficult to find employment for the boys [released from Fatehgarh], and it was still more difficult to find husbands for the girls’. ¹⁷⁷ The NWPO administration had taken responsibility for certain criminalized children but it could not achieve (or would not attempt) the level of economic and intimate interventions required to employ all young men and marry all young women. Third, registered people’s strategies of resistance and persuasion undermined the colonial governance of familial and marital ties. Criminalized people were demanding a retraction of government interference in some areas and greater intervention in others. They wanted the return of their children, along with further government assistance in arranging marriages. Parents ‘felt strongly their enforced separation from their children’ and persistently petitioned the government. ¹⁷⁸ NWPO officials had intended child removal to discipline parents as well as children, but breaking up families had undermined the government’s control over parents, who ‘would not settle down with contentment in an agricultural colony unless their children were restored to them’. ¹⁷⁹ Moreover, the matchmaking scheme had failed to counter registered people’s demands for government assistance. The Sultanpur detainees had ‘forcefully’ and ‘truthfully’ complained that their internment had interfered with the ‘marriage of their children’. ¹⁸⁰ In short, the various dimensions of intimate governance involved in the anti-Sansiya campaign had either prompted resistance, undermined political control, or proved to be beyond the power of the colonial state.

There was another, more uncomfortable and less-cited reason for the closure of Sultanpur. Sadly, 10 per cent of Sultanpur inmates had died in a twelve-month period spanning 1895–6, which may explain the urgency with which the new Lieutenant Governor closed the settlement. ¹⁸¹ The Prison Department would not run the new ‘colony’ at Sahibganj lest it has the physical appearance or discipline of a prison. When Sahibganj opened in April 1896, the authorities’ first task was to return children to their parents. ¹⁸² Police were ordered not to enter the Sahibganj settlement. Nonetheless, the settlement manager held regular roll calls; there was always at least one chaukidar on duty; and a district Magistrate visited weekly, to hear ‘misconduct’ cases and order whippings of offenders. ¹⁸³

MacDonnell opened Sahibganj, in part, to provide greater scope for detained people’s matchmaking. ¹⁸⁴ Yet the new ‘agricultural colony’ neither ended the authorities’ concern with conjugality, nor stopped criminalized people from lobbying for help with marriage arrangements. Two months after Sahibganj opened, the Commissioner of Lucknow, J. S. Mackintosh, reported,

There is a large number of young men in the settlement, and also a considerable number of marriageable girls; but the parents will not allow their daughters to marry unless the would-be husbands pay considerable sums. It is said that before their incarceration, ₹300 or ₹400 was commonly paid for a Sánsiah girl. No one in the settlement can afford to pay this money now, and the consequence is that no one can get married. ¹⁸⁵

Sahibganj inmates were clamouring for official help: ‘they think the Government will pay the money required for marriages according to their notions’. ¹⁸⁶ But Mackintosh made it clear that ‘the Government is not in the least likely to give them ₹300 or ₹400 for each girl they give in marriage’. ¹⁸⁷ Nor did the authorities attempt another matchmaking scheme to marry Sahibganj ‘boys’ and ‘girls’ to people living outside the settlement. The authorities aimed to marry Sahibganj detainees to each other ¹⁸⁸ due to the aim of producing a ‘self-supporting’ colony based on the institution of the family. ¹⁸⁹

To solve the bride price problem, Mackintosh proposed (and the NWPO approved) passes for one year of leave, to allow Sahibganj men with ‘lawful employment’ to work outside the settlement. ¹⁹⁰ Mackintosh reasoned that young men could thereby save up for bride price. Moreover, he wanted to threaten bride-givers with the removal of marriageable young men from the settlement, and the possibility that they might go ‘elsewhere in search of wives’, in order to manipulate parents to ‘give their daughters for what their suitors can afford to give’. ¹⁹¹ In the same year, 1896, Crooke wrote that Sansiyas paid bride price but Kanjars practiced dowry. ¹⁹² Mackintosh’s plan to manipulate bride price may not have grasped the Sahibganj inmates’ marriage practices, which probably varied due to the classification of numerous groups as Sansiyas. In fact, in early 1897, the administration concluded that the majority of Sahibganj detainees were not actually Sansiyas. Numbers varied but one tally suggested that out of 474 inmates, 106 were Beriahs, 236 Bhatus, and 33 Nats—all of whom were not proclaimed as ‘criminal tribes’—and only 99 were Sansiyas, an official designation which included Kanjars, Radhua Kanjars and Gidhias. ¹⁹³ In 1890, the police had ‘illegally detained’ mobile and marginalized people with diverse affiliations—and likely, diverse conjugal practices. ¹⁹⁴

Officials failed to resolve contestations over marriage through pressure upon the parents of daughters to lower their bride price demands. In 1897, five girls from Sahibganj were married. ¹⁹⁵ Yet poverty prevented many inmates from marrying their children due to not only bride price (or dowry) but also other marriage costs. The detainees continued to insist that the government should contribute to marriage expenses. Richardson, a Kheri police official, reported in 1897 that one of the ‘principal reasons given by the tribe for their being unable to settle down’ was that ‘[t]hey are unable to obtain money for their marriage expenses’. ¹⁹⁶ In 1900, Brereton, the Inspector-General of Police, wrote that the Sahibganj inmates ‘find difficulty in marrying their sons’. ¹⁹⁷ Kheri officials repeatedly refused to provide marriage costs, which they viewed as a threat to the order and productivity of Sahibganj. The Deputy Commissioner complained that the Sahibganj detainees ‘asked the most absurd sums, and it was generally only an excuse to eat and drink in excess’. ¹⁹⁸ In the official view, weddings were a cause of disorder but marriage was imperative, since cementing the institution of the family was necessary to ‘give some hope of stability to the colony’. ¹⁹⁹ Yet the impoverishment of Sahibganj detainees—itself a consequence of the CTA—hindered marriage.

Conclusion

The existing literature on the CTA is especially concerned with the reasons why certain groups were classified as ‘criminal tribes’. While drawing attention to the role of gender, sexuality and conjugality in colonial knowledge, I have focused on how the CTA worked. With this shift in focus, we see that the CTA involved the colonization of subaltern families. Since 1871, the family had been crucial to the policing of the ‘criminal tribes’. But the NWPO attempted closer governance of criminalized peoples’ familial relationships, including marriage, in populations that it understood as ‘wandering criminal tribes’. In the government view, controlling such groups necessitated far-reaching changes to their geographical location and distribution. According to the logic of colonial population management, families would be central to the spatial reorganization of populations. The NWPO took a deeper interest in conjugal matters in the Sansiya population due to the unprecedented extent of forced transportation across land which the Sansiya project entailed. In 1896, the NWPO closed Sultanpur and Fatehgarh, in large part, because it could not sustain the highly interventionist governance of intimate relationships—between parents and children, relatives, and prospective brides and grooms—that it had attempted. The NWPO repeatedly deemed the Sansiya project a failure, but as colonial policy was reformulated, conjugality continued to emerge as an official concern.

Moreover, people criminalized as Sansiyas were unable to marry their children due to the fracturing of their familial and social networks and their impoverishment under the CTA. This fuelled criminalized peoples’ resistance to the CTA regime, as well as their efforts to make the colonial state work for them, by enlisting officials to assist them in celebrating and arranging marriages. ²⁰⁰ Those government policies on conjugality that were more ‘successful’, were those that criminalized people demanded or appropriated, notably the 1891–6 matchmaking project. In contrast, the indentured emigration of ‘young couples’ (and others), as well as efforts to marry Sahibganj detainees to each other, failed due to a lack of cooperation from criminalized people. We should not underestimate the asymmetrical power relations between ‘criminal tribes’ and the state—or the ways that gender patterned these power dynamics. Indeed, criminalized women and girls could rarely choose their own marriage partners under the official matchmaking scheme. But in a number of instances, criminalized people leveraged colonial interventions into conjugality to their own ends.