Doing ‘coolie’ work in a ‘gentlemanly’ way: Gender and caste on the famine public works in colonial North India

Abstract

The second half of the nineteenth century was marked by regular famines and scarcities in India, and famine public works were one of the chief ways for the colonial state to provide relief. Famine public works involved labourers, including a large number of women, working in the construction of railways, roads, canals, and tanks in return for a subsistence wage. The present article contextualizes the practices of famine public works, especially the segregation of famine public works into large departmental and village works, within the intersecting processes of labour, caste, and gender. Drawing on evidence from North Western Provinces and Punjab, the article makes two arguments. First, it shows that segregation in famine works was driven by a shared understanding of the dominant castes and colonial state regarding labour, property, and caste which ensured that village works were reserved for dominant castes. A relational definition of labour was central to the construction of caste respectability on famine works. Second, by comparing the sex ratio of labourers in the two kinds of famine works, the article argues that women's labour was not merely a marker of caste, but constitutive of it.

Introduction

In the late-nineteenth century, shifts in the rural economy and techniques of colonial governance concretized caste identity around property and labour in new ways.¹ This was also a period of recurrent famines and scarcities² and the state's organized attempts to provide relief through a programme of famine public works. Work, mainly earthwork (digging and carrying earth), was given out in the famine-affected areas in different types of construction work—roads, railways, tanks, and canals—in return for a subsistence wage. This article locates the practices on the famine public works within the interconnected trajectories of labour, caste, and gender in the second half of nineteenth- century India. It makes two main arguments. First, it demonstrates that the segregation of famine public works into large departmental and village works was a result of a shared understanding between the colonial state and dominant castes³ in rural India of the relationship between caste, labour, and property. The demand by dominant castes to secure ‘respectable’ labour⁴ for themselves within famine relief drove this principle of segregation and resulted in the reservation of village works for dominant castes. Further, in the case of famine public works, respectable labour was defined in terms of where and with whom one had to work rather than the nature of work itself. The markers of respectable labour—who worked where and with whom—both drew from and reinforced the categories of landed and non-landed, and labouring and non-labouring castes. In the second half of the nineteenth century, the specific alignment of caste, labour, and property contributed to the process of labour control and subordination in rural India, ‘freeing up’ labour for the various factories, mines, mills, and plantations. It also reconstituted caste power over economic resources, including labour, within the changing agrarian order.⁵

Second, the article explores the gendered implications of the segregation of famine public works. It argues that not only were the site and nature of women's work crucial to the claims of being landed or labouring castes, but that women's labour and their mobility for labour was central to the very sustenance of rural stratifications, such as caste. The reservation of village works for the dominant castes directly created the conditions of work for women from non-dominant castes who had to seek work outside their villages on large departmental famine works. Notably, the claim of being respectable castes rested on women's labour being confined to their ‘own fields’. Drawing on the evidence from famine public works, this article argues that the relationship between gender and construction of identity, either through processes of ‘othering’ or self-fashioning, should be located within the power structures of the labour regime. This approach can elucidate the material processes that inform social norms and cultural constructions regarding gender relations and their connections with other identities like caste. In making this argument, the article draws from the existing frameworks of understanding caste, gender, and labour in relation to each other, especially Brahmanical patriarchy in the context of South Asia.⁶ While the significance of gender in the maintenance of caste boundaries, especially through endogamy and hypergamy, has been researched in South Asian scholarship, within the discipline of history the interrelatedness of gender, caste, and labour has not been explored sufficiently.

Famines and scarcities in nineteenth-century India became part of a variety of colonial and nationalist discourses and also impacted on the contours of colonial governance and institutional practices.⁷ Scholars have also shown how infrastructural expansion in the second half of the nineteenth century created uneven spaces under capitalism and colonialism, and was central to the making of colonial governance and imperial power in South Asia.⁸ However, scant attention has been paid to the intertwined histories of famines and the expansion of public works in this period.⁹ This article builds on the historiographical departure marking the more recent scholarship on infrastructures that views them as ‘lived artefacts’ producing their own ‘field of power’.¹⁰ In this article, we analyse famine public works as an interrelated set of institutional arrangements and labour regime in their relation to the history of caste power in rural India. Caste and colonial governance is a rich field of historiographical debate with implications for both historicity of caste and the nature of its modernity. One of the strands of the debate—the extent and the process of the role of the indigenous elites—is particularly useful to trace the economic power of caste.¹¹ For the period under consideration—the second half of the nineteenth century—historical investigations into the process of sedentarization have brought together the questions of economic change, colonial governance, and the reconstitution of caste power. The intensification of agriculture in the wet zones, extension of agriculture in the dry zones, increasing production of cash crops, and decline in the older forms of rural industry and services concomitant with the colonial construction of the ideal agrarian (the ‘village’¹² with secured property rights and a settled itinerant population) had far-reaching consequences for caste in both rural and urban areas. It is not just that caste claims of being ‘customarily’ propertied or increasingly exclusionary definitions of labouring castes were made, but that through legal, bureaucratic, and ethnographic means property and labour became constitutive to caste categorizations. Historians have elaborated the role, agency, and collusion of the rural elites in this process in both maintaining and extending their access over resources and subjugating and controlling labour.¹³ This article traces the making of caste power in the process of segregation of famine public works through claims over respectable labour.

Analysing the labour regime on the famine relief works allows us to explore the ways in which the practices of segregation reiterated caste-based claims of being essentially propertied, or its obverse, that is, essentially non-labouring. The very negotiation of segregation of famine public works into village and large departmental works between the colonial state and the dominant castes was a claims-making process.¹⁴ These claims, we shall see below, were sought to be settled within the framework of rights over land and labouring ‘habits’ and ‘customs’. The ownership of land or rights over land and the definition of agricultural labourer has been principal to the question of free and unfree labour in South Asia.¹⁵ Accounting for caste in the definitions and connotations of agricultural labourers has further complicated the question. For instance, Rupa Vishwanath challenges both Dharma Kumar and Gyan Prakash on whether partial rights over land or ownership of homesteads justified the blurring of distinctions between the Untouchable landless labourers and tenants. She argues that, as far as control over labour of pariahs was concerned, rights were circumscribed by punitive and extra economic measures that were exclusively applied to Dalit labourers.¹⁶ While this argument serves well the purpose of delineating Dalit labourers from others within the universe of agrarian labour and proving the case of agrestic slavery in Madras presidency,¹⁷ it does not fully exhaust the ways in which caste aligned with property and labour claims in colonial India. This article argues that dominant castes’ claims of being essentially property-owning were made not only through subjugation and control over the labour of those below them in the caste hierarchy, particularly Dalits, but also through a relational definition of their own labour. Hence, the self-fashioning by the dominant castes as property owning was contiguous with their demand for respectable labour on the famine public works.

Earthwork on public works construction was strenuous, tedious, and often hazardous. Combined with low wages and low ‘skill’,¹⁸ the shifting nature of work, and the caste background of those who were engaged in digging and carrying on public works¹⁹ contributed to earthwork being perceived as low status work.²⁰ Both the nature of work (usually the wages and conditions) and the identity of the workers (gender, caste, race, ethnicity, migrant status) determine the low status associated with certain kinds of work: more often than not both these markers of low status work coincide. Labour historians have directed our attention to the historical shifts within not just low status work, but also ‘stigmatized’ labour. Hence, in municipal sanitation and leather work associated with Dalit labour, caste was reinscribed in the ‘modern’ work spaces through shop floor division of work, industrial training, labour control measures such as contractual work, and the mediated use of technology.²¹ It has also been possible for upwardly mobile peasant castes to claim that the manual labour of ploughing was respectable work that did not necessarily hinder their ascent in the caste hierarchy.²² This article unravels the question of caste, labour, and respectability to interrogate the different factors that made it possible for the dominant castes to simultaneously disavow famine labour as ‘coolie’ work and reserve the same type of work for themselves in village famine works. In other words, how was the relational definition of labour constructed in the context of famine public works to claim caste respectability?

Respectability and gender have been studied by historians of South Asia in relation to identity formation under colonial rule. Women's sexuality, education, work, and practices of conjugality have all been central to the ways in which different classes, castes, and religious communities have fashioned their identity in relation to others as well as the colonial state.²³ However, the processual link between labour, gender, and caste has been most illuminating in the research on women's work, especially rural work.²⁴ Caste and kinship networks determined the kind of work women did and the site of their work. In the last decades of the nineteenth century, the consequence of mechanization of rural industries and the low returns from labour and craft-based work on women's socio-economic lives was differentiated by caste.²⁵ In urban areas, caste-marked practices of marriage, sexuality, motherhood, and domesticity contributed to particular forms of working class femininity that were instrumental in maintaining class divisions.²⁶ For instance, in the jute mills of Calcutta in the early twentieth century, the livelihood and cohabitation strategies of men and women (particularly single migrant women) from labouring households were constructed as promiscuous and immoral.²⁷ In the industrial context, the narratives that folded caste and class hierarchies into specific constructions of working class families were useful to employers in dealing with economic exigencies, as well as in responding to the various labour legislations that sought to control the working hours and working conditions of women.²⁸ The growing influence of the male breadwinner ideology on the working classes as well as the decline of women in the workforce throughout the twentieth century has been noted in phenomena like effeminization, housewifization, and the domesticity of women, which were aligned to the prevalent wider gendered notions of work that, in turn, conformed to the dominant caste norms.²⁹ Chitra Joshi has shown that the degree of seclusion of women depended on the caste of the labouring households.³⁰

In both the rural and urban context, the space of women's work (for example, working on family owned or other's fields) and visibility (the different adoptions and adaptations of purdah and seclusion as well as withdrawal of women from the workforce) was fundamental to the self-fashioning of respectability. However, these very markers of respectability were subject to change, especially under the economic pressures that were exerted on labouring households. Famines and scarcities of the late nineteenth century represented such moments of economic crisis for large sections of the rural labouring population. This article analyses the ways in which the deployment of women's labour as a response to economic hardships was incorporated into the formations of caste. While women comprised nearly 50 per cent of the total labouring population on the famine public works by the end of the nineteenth century,³¹ there was considerable difference in the sex ratio of labourers in large departmental and village works. There were close to equal numbers of women and men working on the large departmental works, while in the case of village works, the number of women was very low in comparison to men. This article presents this reversal of the sex ratio on large departmental works and village works as a consequence of the reservation of village works for the dominant castes which not only separated the work done by women belonging to different castes, but also reiterated caste order in rural India.

This article mainly relies on famine records from North Western Provinces [henceforth, NWP] and Punjab. The official archive on famine public works can be found in the Public Works Department [henceforth, PWD] and the Revenue and Agriculture Department.³² While the accounts of the famines of the late nineteenth century are also present in non-official sources like newspapers and missionary records, the substantial amount of information on famine public works comes from official and semi-official sources, including reports. The Famine Commission reports [henceforth, FCR] of 1880, 1898, and 1901, along with their province-specific evidence volumes, are of immense value when read critically. The article has also used material from colonial ethnographies and vernacular and English-language newspapers and journals.

Classification of famine works: Large departmental works and village works

The emphasis on famine public works within the colonial relief system in the second half of the nineteenth century³³ reflected the dominant understanding of the causes of famines as primarily due to a shortage of agricultural work.³⁴ Further, famine public works were congruent with the extant and emerging ideals of the ‘useful poor’ and colonial charity.³⁵ The organization of famine public works responded to two imperatives of the colonial state—economy of governance and efficiency of governance.³⁶ Consequently, famine works represented a particular kind of labour regime that sought to bring together the principles of laissez-faire with the aim of mitigating the effects of famines and scarcities and creating the ideal labouring relief subject. This labour regime was not unmediated. The interests of rural elites, the private employers of rural labour in agriculture as well as public works, and the labourers themselves shaped the practices of famine public works.³⁷ For the purpose of our analysis of caste, labour, and gender in famine public works, the present section looks at one particular aspect of this labour regime—the spatial organization of famine works, especially their division into large departmental works and village works.

A number of factors influenced the spatial distribution of famine public works. One of the major concerns was to ascertain the ‘genuine’ relief seekers. Famine public works were chiefly aimed at ameliorating the conditions of out-of-work agricultural labourers who were ‘able-bodied’ but weakened due to the lack of sustenance, both because they could not find employment during famines and due to the high prices of food grains.³⁸ Thus rules of work and wages on the famine public works (more on this below) were contrived to allow this category of relief seekers to earn an ‘honest’ living. However, famine officials feared that ‘professional’ labourers (that is, labourers who were accustomed to the type of work given out on construction sites), labourers who were not weakened by famines, and those who could easily find work in the non-famine public works or as field labourers would all be attracted to famine public works because of what was perceived as the laxer conditions of the nature of the work required on the famine public works.³⁹ Consequently, different types of ‘tests’ were devised to ensure the authenticity of the relief seekers on the famine works. One of these tests was the distance test.⁴⁰

The distance test was based on the belief that only those people who were in actual need during times of famine and scarcity would travel a certain distance from their home to access the relief provided by famine public works. Those who did not pass the distance test were deemed to be not in real need of relief. However, all the famine commissions were of the view that distance between the residence of the labourers and the famine public works was not a good test of distress as it did not account for the general enfeebling effect of famines and scarcities on the labourers. While famine officials agreed that works need not be provided very close to the labourers’ homes, it was also apparent that the distance tests could not be applied practically. In fact, it was found by the FCR, 1898 that the distribution of famine public works varied from one province to another and was affected by a different set of factors in each of the provinces.⁴¹

The higher density of population in the Bengal province as opposed to NWP and Punjab meant that the famine public works had to be closer to each other in Bengal. The type of terrain also determined the nature of the works and subsequently its density. The feasibility of opening large famine public works at any given distance from the affected villages was also a consideration. For instance, in 1897, the works opened in NWP usually served villages within a diameter of 16 km; the figure for Central Provinces was 24 km and for Bengal it was 10 km. The complaints of private employers also affected the selection of projects for famine relief. In response to allegations by the East India Railway Company and the Bengal North Western Railways that famine public works drew labourers away from nearby railway line construction, the chief engineer of the Bengal PWD recommended that the government refrain from opening relief works within eight or ten miles of any private works.⁴²

In 1898 famine public works were divided into village and large departmental works. This was a shift from the earlier classification of large and small works, depending on the number of people working on the construction site.⁴³ Famine public works were also classified according to the agency which carried out the works—PWD or the civil administration.

The large departmental works, which mostly included construction of roads, railways, and canals, were under the supervision of the PWD and were governed according to the rules set out in the various famine codes and by the Famine Departments of the respective provinces. Labourers were usually employed on earthworks, which was considered to be familiar work for those who sought relief by working on the famine works. According to the famine codes, the work had to be set at a lesser quantum than the usual work given out in the non-famine public works. Labourers were organized into gangs, with mates from among them and a set of famine officials⁴⁴ to supervise the work and payment of wages. There was extensive discussion on setting the wages for famine public works, with the diet of prisoners, labourers, and emigrant labourers to plantations used as the terms of reference. Wages were calculated on the basis of the minimum required diet for every person working in the gang and was differentiated according to the age and sex of the labourer. The careful calculation of the wages, which were subject to further cuts by the famine officials on the construction site as well as to the vagaries of the market price of grains, arose out of the concern that famine works should not be too ‘attractive’ to labourers not affected by the vicissitudes of famines and scarcities.

While non-famine construction work relied on a complex system of contractors and sub-contractors,⁴⁵ the famine codes warned against the use of contractors in famine public works as it was suspected that their presence would depress the wages that ultimately reached the labourer.⁴⁶ Both piecework and wage work was given out on the famine works, though a minimum wage was usually stipulated in the case of piecework so that the labourers could earn a subsistence wage as directed in the famine codes. It should be noted that these code prescriptions were applied to varying degrees in the provinces. Nonetheless, all these features of large departmental works—individual calculation of diet, absence of contractors, task work, and minimum wage—were criticized by both private employers and government famine officials. The former argued that the wages given on the famine works raised the overall wages of rural labour and amounted to labour market ‘interference’ by the state.

Village works included the construction of wells and tanks within a village or group of villages and were carried out by the Revenue Department, private individuals, or both. The emphasis in village works was on the utilization of local resources and agencies, which included work being carried out through middlemen like zamindars⁴⁷ and headmen. For instance, the FCR, 1898 recommended construction and repair of wells, tanks, and reservoirs for storage of drinking water and irrigation purposes, irrigation channels, embankment of fields, the weeding of noxious weeds from fields, and improvement of village roads as tasks that could be carried out as village works.⁴⁸

The role of the PWD was limited in the case of village works. To illustrate, in 1906 earthen embankments or bandhs had to be constructed in Banda district of the United Provinces⁴⁹ as part of village works. They were carried out under the Ken River canal division and the sub-engineer was ordered by the Irrigation Department to give advances to zamindars and lambardars ⁵⁰ identified by the Collector and the naib tehsildars ⁵¹ of the district for the construction of embankments and bandhs. The sub-engineer had to carry out a topographical survey of the area and identify suitable projects. It was recommended by the Irrigation Department that ‘unpretentious projects’ of simple earthwork should be chosen for famine relief. The rate to be paid to the lambardars had to be officially communicated from time to time. The sub-engineer was instructed not to concern himself with what the lambardar paid his labourers.⁵² Hence, the PWD's contribution to the village works was mainly in the form of the technical expertise of its engineers. In this scenario, the question of labour mobilization became heavily dependent on the local elites.⁵³

Divergent opinions were expressed by famine officials over the relative usefulness of large departmental works and village works. It was the opinion of the FCR, 1898 that large departmental works should be the ‘backbone’ of famine relief and village works should be opened either during times of lesser or decreasing distress or when the public works were threatened by epidemics. While agreeing with the view expressed in the FCR, 1898, the FCR, 1901 reiterated the advantages of village works (more economical and useful, less exposed to dangers of cholera and other epidemics, and easier to supervise) drawing special attention to two advantages—first, that village works were less open to accusations of labour market ‘interference’ and neglect of agricultural dwellings and stock, and were a hindrance to the early resumption of agricultural activities; and second, that village works were less likely to loosen moral and domestic ties. These two points, it was reported in FCR, 1901, were much appreciated in non-official circles. Despite this, the FCR, 1901 upheld large departmental works ‘not only as an alternative, but also as a safeguard’ [emphasis added].⁵⁴ The greater dependence on large departmental works meant that labourers had to travel some distance outside their villages to access relief work. Who travelled outside the village to work on large departmental works managed by the PWD and what factors impelled the higher reliance on large departmental works?

Caste respectability and the labour regime on the famine works

The spatial organization of famine public works and their division into departmental and village works was driven by a number of factors that addressed one of the core concerns of famine public works—determining the legitimate relief subject. One of the principal ways to ascertain the legitimacy of the relief subject was their ability and willingness to labour on the public works construction within the parameters set by the colonial famine relief programme. This section will show that caste was an important qualifier to this ‘labour test’ and moulded the practices on famine relief—more specifically the rationale and application of the segregation of relief works into village and large departmental works. The voices of dominant castes from NWP and Punjab arguing for the segregation of famine works by caste are strongly registered in the famine records. The rationale for famine public works segregation can be found in the claims-making process by the dominant caste over the relief provided through village works and the colonial state's consideration or refashioning of these claims. The main terms of reference for both the dominant castes and the colonial state were labour and property. We see below that the demand made by the dominant castes for providing avenues for respectable labour was based on their claims of being essentially non-labouring. These claims were relational in nature, that is, they served the function of separating the dominant castes from others. Notably, the arguments of being essentially non-labouring were made on the grounds of both the nature and site of their labour as well as claims of being property owing.⁵⁵

The following discussion first provides the rationale found in the famine records for why the dominant castes required a different kind of famine relief works. The main arguments the dominant castes used for the segregation of famine works included their cultivator status; their affinity to ‘home’ and aversion to migrate for work; their non-association with the non-dominant castes in the normal course of their work; and, finally, the nature and site of women's work belonging to these castes. Second, it shows, through evidence from the NWP, that this understanding of caste, property, and labour led to a conscious policy of ‘selection’ by the colonial state and village works were reserved for the dominant castes.

Rationale for segregation

The nature of the practices that embodied the village works resonates in the uncharacteristically ambivalent tone of the FCR, 1898 regarding the utility and place of village works in the relief system. Even as it introduced the classification of the relief system into village works and large departmental works, the FCR, 1898 cautioned,

When it is necessary to open village works on an extensive scale in consequence of widespread or acute distress, they (village works) should, as far as possible be managed professionally on the same principles, as near as may be, as public relief works. Detached village works, or works carried out by means of advances from Government, will be carried out by the Collector with the aid of his assistants and the relief circle staff under the rules to be framed for such works. When such works are to be numerous, they should if possible be supervised by a Civil Works inspector, appointed by the Public Works Department under the orders of Government, but working entirely under the orders of the Collector.⁵⁶

The wariness stemmed from the ground reports on the composition of labourers on large departmental and village works. In 1896, in Banda district of NWP, the engineer made enquiries regarding the caste of labourers working on the large departmental works. It was found that the bulk of those present belonged to castes such as chamars and kols.⁵⁷ The report noted the absence of brahmins and thakurs on these works.⁵⁸ In Etawah in NWP, 50 village works had been sanctioned. It was stated that while these works were sufficient for the employment of ‘high caste labourers’, it was necessary to provide large departmental works for chamars and other ‘low caste men’.⁵⁹ Similarly, in Bara (Allahabad Division, NWP), the Collector commented on the problems in making the village relief system work in this district. He reported that the zamindars did not give employment to those ‘most in need of it and excluded the low castes and non residents’.⁶⁰

What, then, was the vision behind village works? Did the colonial relief system simply buckle under the pressures exerted by rural elites or was it a result of a shared understanding of social stratification and its bases in colonial rural India? It was not as if the colonial administration did not see the merits of village works. Hence, in 1896, in Ambala district of Punjab province, one of the stated advantages of digging tanks was that it could provide relief to the ‘land owning classes’. This was in response to the reluctance of certain castes to work on large departmental works. Relief officials identified three castes that attached the notion of shame to digging work in this region—rajputs, jats, and gujjars. The experience of famines in Karnal in 1869 was cited to argue that these castes had refused to work on roads even when famines had been severe.⁶¹ Further, it was stated that their refusal to work was linked to these castes’ strong bond with their homes. This was contrasted with the non-dominant castes’ tendency to emigrate for work: ‘in famine time kamin and kangal has no particular home’. These caste distinctions were also used to justify the division of work provided as relief. Hence, the idea of digging tanks as being appropriate for village works was mooted.

The tank system finds great favour in my mind, and in order to render it acceptable to zamindars, and in order to prevent this class of labour from quickly giving out, I am inclined to let it be known—that ‘tanks’ are for the zamindars and for kangals the ‘roads’. There is no great hardship to the labourer in this. He is accustomed to seek work and can take his family with him and has no home of value to leave behind. On the other hand, zamindars will shy away from work if they are to be mixed up with ‘kangals'.⁶²

Further evidence for this attitude was provided from Lukhi, a village in this district. In this village, rajputs agreed to work on tanks if they were assured that the work would be reserved for the zamindars. It was reported that the rajputs were agreeable to the idea that the ‘kangals would go and work like coolies on roads, while they would be asked to dig together at home in a gentlemanly way’ [emphasis added]. Moreover, it was argued that in the absence of such an exclusive arrangement of relief for these castes, they would resort to borrowing which would lead to further indebtedness in Punjab province.⁶³

This understanding of the objectives of the village works was reiterated by the FCR, 1898 in a ‘Note on Village Works For the Relief of Distressed Agriculturists’.⁶⁴ The FCR, 1898 recognized that large departmental works served a bigger population of labourers who were not only accustomed to the type of work provided in the large departmental works, but were also used to going a certain distance from their homes to earn their livelihood. Hence, there was ‘no special hardship involved in this so far as the majority of the workers are concerned’.⁶⁵ However, there was another class of people in the villages who would not avail themselves of the work provided in large departmental works. This class was identified as ‘self-cultivating owners and tenants living on the profits of their cultivation’. Further, this class, ‘though poor and leading a simple village life and accustomed to labour in their own fields, belong to a much higher stratum [emphasis original] of society than the ordinary labourers. Their women in most places work in association with their male relations, but they do not hire themselves out to other cultivators, far less go out of their villages [emphasis added] in quest of employment.’ It was pointed out that this class of people could not be expected to work in large departmental works with its ‘privations and inconveniences, its enforced companionship of strangers and exposure of their womankind to the familiarities of all manner of men gathered together at a relief camp’.⁶⁶

Clearly, in the view of the British administrators, a uniform ‘labour test’ failed to ascertain the relief requirements of those affected by famines and scarcities. It was suggested that the landowning cultivators, ‘whom perhaps it is much more important to help than the majority of those who crowd the relief works’,⁶⁷ be relieved through a different class of works that required labour, but removed the necessity of association with the non-dominant castes. Village works were considered ideal for this purpose.

Policy of ‘selection’ on village works

The understanding that village works were meant for the relief of what were considered to be non-labouring castes in the villages led to a conscious policy of ‘selection’. This policy meant that the dominant castes (as we will see below this meant brahmins and thakurs in the case of NWP/United Provinces) were the preferred subjects of relief through village work. In 1906, it was reported from Karwi and Banda in United Provinces that efforts were being made to induce zamindars and lambardars to take petty contracts for small amounts of work near their villages.⁶⁸ If the zamindars and lambardars did not agree, contractors would have to be employed, with the condition that they would engage only local labour. This was the last resort, as employing contractors on the famine public works was not seen as suitable as a matter of policy. Fifty thousand rupees had been allotted in this district for carrying out agricultural works, which primarily included constructing bandhs. There were 23 such works in this district in the month of January and 3,241 persons were employed on it. Labourers coming from any village within five miles of the work were given relief with this type of work. It was noted that ‘special arrangements’ had to be made for thakurs and brahmins who would not work on the ‘ordinary terms on large works’.⁶⁹ It was concluded that small works close to their villages was a suitable form of relief for these castes.⁷⁰

A similar arrangement was made for the thakurs in Bhoginipur, in Kanpur district in 1906. Advances were paid for building eight tanks in this area to thakurs who ‘would not come on to work on ordinary relief works’.⁷¹ By 24 February 1906, the number of such tanks had increased to 19 and instructions had been issued to ‘not allow low caste people to work on them’.⁷² The number of tanks increased to 20 the following month and it was reported that ‘thakurs and brahmins are generally employed on them’. By April, the overall numbers on all types of works went down with the beginning of the rabi harvest season. In Bhoginipur, thakurs declined to work on the 20 tanks after the revision of famine rates by the commissioner. He remarked that, ‘if the thakurs of Bhoginipur will not work at the tanks below the maximum allowed by the Code, they cannot be in very great trouble’. In May 1906, work of this kind stopped completely in Bhoginipur.

In a neighbouring region—Ghatampur in the same district—there was work being done on six tanks, with the same condition that ‘low caste people are not allowed on this work’. In Unao too there was special selection in village works expressly for the thakurs and brahmins ‘who will not resort to regular works. Efforts are being made to induce them to work on these themselves and send the lower classes of residents of their villages to the public works.’⁷³ In Hamirpur too, the village works were restricted to ‘higher castes’.⁷⁴ In Jhansi a large number of village works (between 44 and 46) were ‘opened for the employment of better classes’.⁷⁵ There is similar evidence from United Provinces in succeeding years.⁷⁶ Hence, we see that the policy of selection implied that the village works were to be reserved for the dominant castes and excluded others. In reserving village works for the dominant castes, the colonial state had not only secured relief for them during famines, but also reproduced caste through hierarchies of famine public works.

Who worked where and why: Gender, caste, and labour

Dominant castes used two main features of women's work to buttress their claim of being respectable. First, that women belonging to these castes worked in their own fields. Consequently, they did not hire out their labour and worked alongside their ‘own’ men, that is, men to whom they were related through ties of kinship and family. Second, that they did not go out of the village to seek work. Labouring on large departmental works would violate both of these conditions associated with respectable labour. On the other hand, it was precisely such conditions of work that were created for the women from non-dominant castes due to the segregation of famine works by caste. Hence, the labour of women was constitutive of caste and not merely its signifier. The following pages first delineate some of the ways in which gender and caste characterized famine relief and then present the gendered effects of the segregation of relief works on village and large departmental works by comparing the sex ratio on these two kinds of famine relief works in five districts of United Provinces.

As mentioned earlier, most of the women working on the public works belonged to castes that either specialized in construction work or were generally involved in field labour. We do have evidence of women from dominant castes like brahmins and thakurs working on the famine public works. For instance, in March 1874 the Government of India set up an inquiry to verify the claims made in the London Daily News that ‘hundred[s] of high caste women are to be seen labouring on the government relief works with common coolies’.⁷⁷ This news had been reported from Tirhut and Champaran districts from the Bengal province. The subsequent inquiry concluded that while the presence of certain ‘high’ caste women on the famine public works was not out of ordinary in this region, the conditions of scarcity were not severe enough to ‘drive the better classes to great extremities’.⁷⁸ Similarly, in Punjab, the plight of ‘well connected women who had never been even to the doors of their houses’ in Hissar is presented in a particularly emotive register in the newspaper Wakil. These women are described as labouring on the famine relief works with ‘eyes fill [ed] with angry tears’, ‘repress [ing] their feelings’.⁷⁹ In Etawah (United Provinces) the case of one thakur widow was mentioned separately in 1906. She was reported as the only woman on that particular site who was of a ‘good caste’ and special care was taken to transfer her to a nearby canal works.⁸⁰ A caste census in the same year on a canal works in the same province showed that out of the 428 women working on the construction of the canal, two women were brahmins, five were thakurs, 102 women were from kori caste, and 171 were reported to belong to chamar caste.⁸¹ Women from dominant castes present on the famine public works in small numbers were notably described as exceptions in the famine records. In fact, in case of the example from Bengal province cited above, we see that the very presence of women from these castes is taken as a proof of the severity of famines.

The distinction in famine relief based on the caste of women relief seekers also becomes apparent from the type of work given out to pardanashin women.⁸² These women were allotted work that they could do within their houses in order to maintain their ‘respectability’. They were either given doles or put to tasks such as spinning cotton and grinding grains. Missionaries also gave work to women at home as part of their own famine relief initiatives and differentiated such type of work (spinning, weaving, lace making) from that done on the famine public works, describing the latter as ‘rough work’.⁸³ The division of women relief recipients between ‘gratuitous’ relief and famine public works was for all purposes based on caste.⁸⁴ In 1906, the commissioner of Kanpur observed the following about the people listed on gratuitous relief: ‘there are not a large number on the village gratuitous relief list, but they are nearly all pardanashins, and I am extremely doubtful if they should remain. I directed further inquiry. There is a tendency to pile on these pardanashin, and omit the miserable chamars, dhanuks and similar castes.’⁸⁵

What did the policy of segregation of famine relief works according to caste mean for women? A closer look at the relief statistics on different types of works suggests some patterns. The number of women and men on the famine public works was almost equal in most parts of the country, but in some cases, the number of women was higher than men. Further, if we take the example of one famine circle⁸⁶ (Famine Relief Works, Second Circle) from NWP, in the two famines years of 1896–1897 and 1897–1898, seven out of 11 districts returned more women on the famine public works than men.⁸⁷ However, in order to map the gendered consequences of segregation of famine works, we have to examine the gender segregated data on different types of works. We have this type of data from five districts in United Provinces—Etawah, Agra, Kanpur, Muttra, and Hamirpur—for the year of 1906. These data have been collated from the fortnightly reports from these districts that were submitted as part of routine famine reporting. While the exact numbers were sometimes subject to revisions, the data are interesting in the pattern of sex ratio that they reveal. The numbers of men and women relieved by working on large departmental and village works are provided.⁸⁸ We have calculated the sex ratio—by which we mean number of women working per hundred men—in both types of relief work (see Table 1).⁸⁹

Table 1: Sex ratios on large departmental works and village works, United Provinces, 1906

In Etawah, in the days from 14 April to 7 July 1906, the number of women was higher than men on large departmental works, with the exception of one fortnight. The sex ratio in large departmental works ranged from 79 to 119. The numbers for village works in this district were provided for two fortnightly periods and the sex ratio was unfavourable to women in both cases. In Agra, from 31 March to 26 May, the sex ratio in large departmental works ranged from 198 to 286. Hence, there was no unfavourable sex ratio in large departmental works. The data on village works in this district are available from 31 March to 7 July and the range for the sex ratio was 26 to 118, with only two out of 14 entries showing a sex ratio favouring women. In Kanpur, the data are available from 27 January to 18 August and the sex ratio on large departmental works ranged from 115 to 79. The range of the sex ratio on village works from 10 February to 18 August was 15 to 92. In this district, the sex ratio was unfavourable in large departmental works in six out of ten entries. However, in all cases bar one, the sex ratio was above 90. In contrast, on village works, where the sex ratio was unfavourable in all cases, it was usually below 50. In Muttra, the data for large departmental works are available for the period from 3 February to 7 July and the range for the sex ratio in these works in this district was 94 to 121, with two out of 12 figures not favouring women. The village works numbers are available for the period between 17 March and 23 June in this district and the sex ratio range was from four to 33. In Hamirpur, the figures are available for the period between 24 February and 22 September with the female to male ratio ranging from 35 to 289 in large departmental works. Hence, here the sex ratio is unfavourable in two out of 16 cases. The ratio was more than 200 in two cases. The sex ratio in village works ranged from 23 to 97.

A clear pattern emerges from the above description. Women worked in large numbers and often in larger proportions than men on the PWD projects. Their presence in village works was very low. This meant that women moved out of villages far more than men to seek employment on large departmental works and, further, that these women belonged to the non-dominant castes. The higher numbers of women than men working on large departmental works can be explained by the fact that there were fewer opportunities for these women to find work than men from the same castes even during times of famine. Overall, women were not the preferred workforce by the private employers, whether inside or outside the villages during famine or non-famine times. However, what is interesting is the reversal of the sex ratio between village works and large departmental works. Village works were clearly reserved for dominant caste men.

Despite the growing scholarship on women's long-distance migration and work,⁹⁰ the overwhelming picture of migration and circulation in nineteenth- and early twentieth-century India is gendered—the mobile, urban, labouring man and his counterpart, the rural woman with limited mobility and a decreasing role in wage labour.⁹¹ An examination of women's labour on the famine public works challenges the view that women were immobile and non-labouring.⁹² Not only did women from labouring castes travel to work on the famine public works, their labour signified a number of other movements which can be mapped on a matrix of season, space, and activity. In agriculture's lean months women moved out of the fields to work on roads, tanks, railways, and canals. The nature of their work also changed in these new spaces from a number of agricultural and allied activities (including food and fuel preparation) to digging and carrying earth. This was, however, a circular movement and women returned to working on the agricultural fields once work was again available for them. The labour of women from non-dominant castes on large departmental works can be located in a continuum of the rather ubiquitous, though not well researched, short-distance migration of women for agricultural work. As Samita Sen has pointed out, the latter did not challenge authoritarian familial structures in the ways that long-distance and overseas migration of women did in colonial India. Such migrations were part of the household's livelihood strategies.⁹³ This article has shown that the deployment of women's labour as part of such strategies was crucial in defining and reproducing caste boundaries.

Conclusion

‘Coolie’ is a historically layered term with meanings and implications that are best understood within specific labour regimes and practices.⁹⁴ While this term was used, along with others,⁹⁵ on the famine and non-famine public works construction to describe the labouring population, this article has unravelled the ways in which the term was employed by the dominant castes and the colonial state to give specific meanings to labour and caste in rural India. Two seemingly opposing principles governed the practices on the famine public works that sought to reinforce caste power within the existing labour regime—the principle of segregation and the principle of relational labour. The article shows that the demand by dominant castes to reserve village works for themselves was made on the basis of particular connotations of their own labour, which needed to be contrasted with ‘coolie’ labour. Resonating with other contexts of ‘coolie’ labour, where their mobility and immobility was central to production and labour processes,⁹⁶ we see that on the famine public works, too, the mobility of non-dominant castes was used to reify their identity. Hence, labour performed on the famine public works by different castes was constructed differently—while the dominant castes were attributed a strong sense of ‘home’ and their labour was considered ‘respectable’, the mobility of the non-dominant castes for work was characterized as a dislocation. This article argues that this relational definition of labour can be seen as contiguous with other kinds of claims that were being made in this period to reinforce different caste-based claims of being propertied or labouring. One implication of this process for the labouring castes was the very definition as ‘labouring’: while caste-based extraction of labour (including forced labour) intensified, they were divested of their customary ownership and usage claims over land. In the case of Dalits, this contributed to the erasure of their agrarian history.⁹⁷

This article has focused on processes and material practices that reveal the ways in which labour, gender, and caste intersected on the famine works. Even as the nature of women's labour was presented by dominant castes as markers of caste respectability, their demand to reserve village works for themselves, which had a specific impact on the labour of women from non-dominant castes, demonstrates that women's labour was not simply a product of social norms governing their seclusion or mobility.

Famine records are silent on the counter claims by non-dominant castes on the question of segregation of famine relief and its implications for caste order. This is in contrast to the assertions and contestations by labourers that we find in other contexts of famine works, which include strikes, desertions, as well as the refashioning of rules on the famine works.⁹⁸ The fact that we do not find counter claims by labouring castes on the question of respectability reflects not only the power inherent in official archives to silence and obscure, but also indicates the caste power of dominant castes who could negotiate with the colonial state.⁹⁹ Further, it strengthens the argument that the understanding of caste, labour, and property was a shared one between the colonial state and dominant castes.