about archives

Vvgnli

Ailh

Publications

Research Projects

Research Partners

Archival Sources

Acknowledgements

Events

Contact Us

Home Web Survey Search  Collection Map    

Labouring Histories: Agrarian Labour and Colonialism

Naladri Bhattacharya

(Neeladri Bhattacharya is with Jawaharlal Nehru University, New Delhi)

Preface

The essay, 'Labouring Histories: Agrarian Labour and Colonialism', by Neeladri Bhattacharya, forms part of a specially commissioned series, Writing Labour History, initiated under the Integrated Labour History Research Programme (ILHRP), of the V.VGiri National Labour Institute. This research programme, established in collaboration with the Association of Indian Labour Historians (AILH), aims at establishing and maintaining an apex repository of labour history documents, an Archive of Indian Labour History, with special emphasis on digital storage and retrieval. It is envisaged that all research projects designed and commissioned by the ILHRP will add to the continuous process of Writing Labour History of India, covering neglected areas of labour history research.

The concern towards constructing a rich and complex history of working people in India has emanated due to the recognition of three interrelated propositions. Firstly, there has been a growing consensus among the scholars that for too long the history of labour in India has been fixated on the industrial factory labour that emerged in the late colonial period. This has led to serious neglect of the overwhelming bulk of the working people who labour outside the pale of the factories and formal regulations. Secondly, scholarly research in labour history has focused exclusively in the colonial period, thus in some sense, neglecting the much more important post independence period developments. This may be one reason why labour history research has been seen as lacking contemporary relevance and has been deprived of the invigorating influence of engagement with the enormous social transformations in the labour scenario in the recent past. Thirdly, the history of the labour movement has been too easily equated with its organised and institutionalised aspects. This has led to the neglect and some times erasure of the rich experience of struggles in the unorganised sector by countless anonymous workers.

In the present essay, the author critically reviews some of the seminal research works on agrarian labour in colonial India and discusses the transformations within the frameworks. While focusing on the changing structure of arguments, the author probes into some of the problems in*conceptualising agrarian labour. The author highlights the nded for understanding agricultural labour as a relationship that sustains agrarian societies and cautions the danger in attaching greater importance to its 'personal' element. Further, the author argues that in analysing agrarian labour there is a need to move beyond the standard binary oppositions such as forced labour and free labour; voluntary labour and involuntary labour and so on. It is emphasized in the discussion that though there is a discernable trend of blurring of the boundaries between these neat categories, one could also see a continuity of erstwhile distinctions, forms of coercions and unfreedom.

On the whole, the author invites the attention of scholars and practitioners into certain hitherto unattended aspects of labour historiography. I hope that labour historians and students of labour studies would find this essay thought providing and useful.

(Uday Kumar Varma)

Director

This essay first looks at some of the important studies on agrarian labour in colonial India (Part-I), discussing the shifts within the frameworks, the structure of the arguments, and some of the problems with the framing ideas. Then (Part-II) it explores some of the themes raised in Part I, by drawing on recent researches on agrarian history.

PART I

The question of numbers

The early studies of agrarian labour in India focussed on the question of numbers. In a book that was to become a nationalist Marxist canonical text, SJ Patel argued with force that agricultural labourers had increased dramatically under British rule: from a mere one-seventh of the agricultural population at the end of the nineteenth century to over one-third by 1931. In the society before British rule, argued Patel, there was little room for the existence of an independent and distinct class of agricultural labourers. Within the traditional self-sufficient village communities, cultivators and artisans lived together, exchanging their products with each other; and peasants produced with the assistance of family labour. British rule saw the dissolution of the old order, the collapse of traditional social relations, a process of immeserization, and the emergence of a substantial class of landless agricultural labourers.

Patel's study clearly was not simply about agricultural labourers. Through a calculation of their numbers Patel was reflecting on the dynamic of colonialism, specifying an empirical index through which to unravel this logic. The dramatic increase in the number of agricultural labourers was seen as clear evidence of impoverishment and pauperization under British rule. In the 1950s, the study of agricultural labour in India was not yet framed within narratives of transition from feudalism to capitalism, the growth of agriculture labour was not yet read as a sign of agrarian capitalism.

Patel's text is pervaded by a sense of loss: the loss of the old world, the loss of self-sufficiency, the breakdown of village community. This story of loss was inevitably linked to a story of great rupture, of dramatic change, 'the most thoroughgoing transformation in India's long history'. Such narratives of dramatic break always schematize the contrast between two times, before and after the moment of rupture, between the time of concord and discord, harmony and conflict, prosperity and poverty. And Patel did just that. But Patel was writing at a time when colonial stereotypes about Indian past were not yet opened up for critical scrutiny, when the theories of self-sufficient village community and jajmani system were not yet critiqued. If Patel built his argument through an uncritical reading of colonial sources, and accepted colonial representations without questioning them, if he was using Census data unproblematically, then he was doing what most academics did in the early 1950s.

Patel's text quickly became a classic. For the first time census figures were being used to study the emergence of agricultural labourers and to map a picture of depeasantization, not only at the national level but with all its regional variations. It was a study that gave the old nationalist thesis of impoverishment under colonialism a new empirical basis, concretizing the picture, establishing it with seemingly reliable data. In the 1950s and 60s, numbers had a seductive power.

By the early 1960s the Patel thesis was under attack. In 1962 Dharma Kumar published her classic essay 'Caste and landlessness in South India', and her book on the subject appeared three years later. Kumar argued with polemical vigour that the proportion of agricultural labourers to the total labour force did not change in the nineteenth century: it fluctuated between 17 to 25 percent. She showed that Patel's thesis of dramatic change under colonialism was based on an uncritical reading of colonial sources, and that landlessness and agrestic serfdom were common in pre-British Madras; they were characteristic of a rural landscape structured by the hierarchical caste order. The increase in the number of agricultural labourers was a phenomenon of the twentieth century and not the long nineteenth century. And if indeed the number of agricultural labourers did not substantially increase over a century of British rule then the nationalist thesis that Patel sought to affirm was itself under question. The social existence of agricultural labourers in the nineteenth century could not reveal a history of impoverishment, depeasantization, and increasing landlessness consequent on high levels of land revenue demand. It revealed rather the persistence of inequality structured and reproduced by the caste order. In 1900 as in 1800 most of the agricultural labourers were untouchables.

Within the Marxist nationalist circle in India, Kumar's thesis was received with unease. She was quickly identified as presenting an optimistic picture of British rule, dissociating the link between impoverishment and colonialism, projecting a simple continuity between the pre-colonial and colonial periods. By the mid-sixties Marxists had re-conceptualized the image of pre-colonial Indian economy, rejected the romanticized picture of self-sufficient village societies; but a rethinking about the colonial economy was yet to begin. For a long time every argument on the colonial economy remained firmly locked in the stultifying poverty / prosperity paradigm. Against every questioning of the old nationalist thesis, Marxist nationalists reacted by reasserting the truth of old canons. There was a refusal to see that many of those simple propositions needed to be rethought and problematized. The terms in which Kumar questioned the nationalist thesis were problematic, but the act of questioning was not.

Kumar's exercise in quantitative measures focussed on the nineteenth century: comparing the census estimates of 1900 with rough estimates for1800. She did not fundamentally question the Census figures that Patel was using, or his interpretation of those figures. In fact her contention that agricultural labourers undoubtedly increased in the twentieth century, though not in the nineteenth, was premised on an unproblematized acceptance of the Census data. In an important essay published in 1972, J Krishnamurty returned to the Census figures that Patel had used, subjected them to a close scrutiny and rejected the conclusions that Patel had drawn. By now it was generally accepted that the occupational data in the Census were suspect (Daniel and Alice Thorner1962). But Krishnamurty showed how the problems with the figures affected Patel's conclusions. Not only were the figures non-comparable but they had a consistent bias. First: the 1871-2 estimates related to 'adult males', the 1881 to 'males' and the subsequent one's to 'persons' (males and females) Second: the area covered by the Census expanded over time. Third: in the early censuses the concept of 'worker' is used while in the later decades the concept of 'population supported' (workers plus dependents). So a comparison of the later figures with the early Census estimates inevitably exaggerated the increase of agricultural labour. Krishnamurty's recalculations, after eliminating the biases, reaffirmed the picture of stasis: between 1901 and 1961 the share of agricultural labourers in agricultural workforce in the Indian Union fluctuated around 25 per cent for males though for females it declined from 31 per cent to 26 per cent. (Krishnamurty 1972)

Can this picture of changelessness be sustained? I will return to this question in later sections of the essay. For the moment I wish to make only two points. First: the thesis of continuity that emerges in Kumar's study flowed from her methodological approach. Since she limited her study to the estimation of lower castes, assuming that they constituted the bulk of the labour force, she excluded from her estimates those impoverished peasants who turned to agriculture labour to supplement their living. The evidence in her book itself suggests that this exclusion is problematic. In the 1890s, according to Kumar, 65 per cent of the ryotwari  pattas in Madras Presidency were too small to support a family; in Ramnad, 75 per cent of the ryots worked as agricultural labourers in the 1890s; in  Anantpur, two-thirds of the pattas were for plots which yielded only about three months of food for one man (pp. 171-72). Consider this evidence in relation to the agrarian context described in the book. Kumar shows how the extension of cultivation and increase in productivity failed to keep up with the growth in population, and high levels of revenue demand and repeated scarcities and famines led to emigration. Impoverishment may not have lead to total landlessness, but the pauperized peasants undoubtedly depended on agricultural labour for their subsistence. The composition of the people who laboured on the fields clearly changed over time, even if their numbers are difficult to capture through census statistics. The harvesters who cut the crop were not all lower caste Malas and Madigas.

Second: even if Kumar is right in saying that most agricultural labourers in 1900 as in 1800 were from lower castes, does this imply that there was no profound change in the social form of labour?  In reacting against Patel's simplified picture of dramatic transformation, we need not throw over board the very idea of change itself? Kumar is right in pointing to the resilience of native institutions and the durability of caste as an institution: embedded structures are not easily washed away under the pressure of external forces. But are they not refigured in significant ways? To track such changes we cannot simply look at figures. We need to re-evaluate the evidence of change, relocate the sites at which they occur. And instead of the binary between continuity and change, the long duree and the event, we need to see how they are intimately connected, how they constitute each other. This is a theme I will return to in the second part of the essay.

 Narratives of transition:

By the 1970s reflections on agricultural labour got linked to the prevailing debate on modes of production. Economists and historians began to identify the specific stage of transition from feudalism to capitalism, and searched for indices through which the transition could be measured.  Agricultural labour inevitably became a point of focus, and discussions on it were all framed within meta-narratives of transition.

One site of the debate was around the question of hired labour. What did it signify? For some, the increased use of hired labour and the expansion of commodity economy in the post-Independence period were evidence of the growth of capitalism (Kotovsky, Gupta 1962, Chattopadhyay 1972a, 1972 b). Others saw the relationship between capitalism and hired labour as more complicated. In a set of important essays Utsa Patnaik argued that hired labour was a necessary but not a sufficient condition for the operation of capitalism (Patnaik 1971; 1972). It existed in earlier social forms. Only when integrated to a system oriented towards an expanded production of surplus value, did hired labour embody capitalist relations of production, only then could it be considered as surplus value producing wage labour. Under colonialism, she argued, capital expanded in the sphere of circulation without entering production, landowners employed hired labour without reinvesting the income back into cultivation; technological growth was aborted and agriculture stagnated. While the pressure of revenue, rent and interest pauperized the peasantry, there was no simultaneous rise of modern industry that could absorb the landless, and help create an internal market (Patnaik 1971)

While Marxist economists differed on whether hired labour signified capitalism, there was a seeming unanimity about non-wage labour. Everyone seemed to agree that attached labour  - whether sustained through traditional customary ties or through debt-peonage - was necessarily pre-capitalist, and ought to be described as feudal or semi-feudal. It is an argument premised on a set of analogous binaries: force and freedom, coercion and consent, involuntary and voluntary, personal and contractual, feudalism and capitalism. If force and coercion was by definition pre-capitalist, and free wage labour the signifier of capitalism, the evidence of coercive labour could then be identified without a problem as feudal and semi-feudal. And the shift from coerced to free wage labour could be seen as marking the transition from feudalism to capitalism. Till the seventies, these reductive assumptions shaped the transition debate not only in India, but also in Europe (Hilton 1978, Brenner 1977) Even critics who set out to question these assumptions, operated in many ways within its terms. Consider the essay by Jairus Banaji on the Deccan peasantry (Banaji 1977). Drawing upon Marx's newly published Appendix to Capital, Banaji developed a powerful critique of the thesis of semi-feudalism. Marx had distinguished two stages within the historic process of subordination of small producers to capital. One was the stage formal subsumption when merchant capital invaded the process of production and took it over. It was a stage in which the labour process was still technically fragmented and decentralized, but producers lost control of their means of subsistence, and their capacity to reconstitute the process of production from one cycle to another without the intervention of merchant capital. The stage of real subsumption entailed the production of relative surplus value - a form that was more adequate to the process of specifically capitalist production.  In this stage the process of production was reconstituted, the means of production and labour power were centralized, and capital lost its individual character and became social capital. Building on this conceptual separation, Banaji suggested that systems of advances and share cropping in late-nineteenth century Deccan, the region of his study, were not semi-feudal relations; they were mechanisms through which small producers were subordinated to the sway of capital.

This was an important argument that led to a conceptual rethinking of the idea of capitalist domination of labour. Yet even this argument remained firmly embedded in the teleology of transition. Instead of a shift form forced to free wage labour, the transformation was now visualized through intermediate stages: from pre-capitalist modes of labour, through pre-formal and formal subsumption of labour, to real subsumption of labour. The 'real subsumption' was the telos towards which the history of labour moved, through tortured steps: capital first took over earlier forms of labour, subjecting them to the power of capital, without transforming them, without entering them, but controlling them from the outside. This formal subsumption expressed capitalist relations of production, surplus value production; but it did not presuppose the bourgeois mode of production in its advanced form; it was based on a labour process that was continuous with earlier modes of labour. In this argument, the true bourgeois form was co-terminus with the real subsumption-process, when the relations of production would emerge purged of the inherited pre-capitalist remains alien to the pure motion of capital. Banaji's thesis begins with the argument that coerced labour, in its multifarious forms, is compatible with capitalism - capitalism in fact develops through these forms - but it ends by arguing that these forms of coerced labour are in fact anomalous to the real logic of capital. Formal subsumption can only be a stage within a linear move towards a purer capitalism, untainted by force and coercion, based on the logic of relative surplus value production through technological innovation - the real subsumption of capital.

The transition debate of the 1970s was, at one level, productive. It opened up a range of conceptual issues for debate, and revealed the problems of categorizing labour forms. But after a point the interest in theoretical classification became semantic exercises in labelling. As if to label and classify was to gain visibility, to know. As if to designate a relation as free wage labour or forced labour, semi-feudal or capitalist, was to understnad the relation, its attributes, its characteristics. The project of classification often obscures and homogenizes, flattens and essentializes. Diverse social forms are made intelligible by fitting them into categories, their differences are repressed in order to discover their social essence, an essence that could be captured through the framing category. When these framing categories form a part of narratives of transition, different social forms of labour are allocated different spaces within the linear flow time, as so many evolutionary stages from one pure form to another, forms whose essences are pre-given. Inevitably these evolutionary forms are then seen as embroynic manifestations of supposed later forms, or dissolving remnants/residues of earlier forms.  What such exercises often block are historical explorations of the formation of specific labour forms, the logic of different labour regimes, enquiries into the processes of their production and re-production, legitimation and de-legitimation.

Patronage and exploitation

Jan Breman’s  Patronage and Exploitation (1974) was such an effort to explore the working of a specific labour regime. Building on anthropological insights, he looked at the cultural constitution of the relationship between landlords and labourers in south Gujrat, analyzing the framework of norms that bound them vertically through ties of mutual duties and obligations. In developing his argument, he also reworked the terms within which sociologists in India had conceptualized such systems of relationships. Till Breman, the debate on jajmani system had been largely framed within the binary opposition of reciprocity and exploitation. The argument of reciprocity emphasized the organizing ideals of the system: mutualities, equivalent exchanges of services, co-operation and interdependence, all operating within an insulated autarchic world, insulated from the market and money economy (Wiser 1936, 1946, Neale 1957). It was as if reciprocity excluded exploitation, expression of emotion and affection was incompatible with the pursuit of interest, gift giving was unmediated by power, and exchange of services against grain was devoid of market calculus. The focus on exploitation inverted the terms, looking at the under lying relations as essentially exploitative and oppressive, unequal and conflict ridden, and driven by market rationality (Biedelman 1959, Hjelje 1967; Gough 1960).  Within this frame, the norms of reciprocity appeared as a mask, the language of mutualities was seen as a deception that cloaked the reality of exploitation and oppression, caste was an expression of class. Distancing himself from the exclusivist terms of these oppositions, Breman looked at the inter-articulation of patronage and exploitation. The relations of patronage and gestures of generosity allowed and lubricated a system that was exploitative and unequal. If the system of patronage made exploitation possible, it also defined the limits of oppression. Halipratha was a system of unfree labour mediated by patronage.

Within the halipratha system that Breman discussed, halis were mostly Dubla tribals, and the masters were predominantly Anavil Brahamins. The hali bound not only himself but also his wife and children. He worked on the land, did odd jobs and ran errands. It was a relationship of beck and call, with working hours never well defined. The wife ground the grain, fetched water, swept the floor, washed utensils, cleaned the stable, did housework and field labor. The master was a patron with obligation and duties. He provided food, allotted a site to build a hut, gave clothing, some firewood, toddy, and medicines. Beyond the basic means of subsistence, he helped the hali tidy over periods of subsistence crisis – in the lean seasons every year, and during scarcities and famines. The Anavil’s loans made it possible for the hali to meet the expenses of life cycle rituals, of marriage, death and birth.

This, according to Breman, was not a system of forced labour: 'The initiative to begin the relationship was taken by both parties and it was in the interest of both to maintain it.’ Within an economy of high seasonality of labour demand and possible shortage of supply in the peak season, Anavils found it necessary to ensure a secure supply for the peak season through the halipratha. Halis on their part preferred servitude to free labour since regular employment was scarce, labour demand was seasonal, and the alternative to attachment was starvation. Even if the remuneration of the halis was low, they were entitled to credit, care and protection. Dublas, in fact, desired permanent attachment: termination of the relationship was the last thing they wanted. Servitude within this system was therefore voluntary.

The Dublas accepted dependence in lieu of protection and care. If the Anavil gave the hali security, he expected the hali to be loyal, submissive, timid, and devoted. The Dubla could not openly express dissatisfaction, or show himself to be disobedient and disloyal. Through loyalty he could ensure rewards, special care, and an access to the intimate details of the household, an access that signified trust. The submission of the hali accentuated the power of the master, while at the same time the ideology of patronage defined the limits of tyranny and exploitation.

In the end, halipratha disintegrated primarily because Anavils and halis, for different reasons, saw the system as no longer serving their interests. With the expansion of industries and railroads, Dublas found alternative employment and were unwilling to enter into agreements that bound them. At the same time the interests of Anavils shifted, and the ideology of patronage lost its function. With the decline of sugarcane cultivation the demand for labour contracted, and Anavils increasingly considered urban occupations as the source of status and prestige. Now agriculture was not valued to the same extent and the command over a retinue of halis was no more perceived as the foundation of power and influence.

Breman’s fascinating study thus explored the ideology of patronage that both expressed and sustained the relations of domination and exploitation in South Gujrat.  To engage with the argument we need to locate it within the larger literature on patron-client bonds. In the 1970s, studies of patron-client relations developed in opposition to earlier functionalist frameworks of looking at social and political formations.  In developing the argument, these new studies borrowed from a longer tradition of works on the subject, but reworked the meaning of the key concepts. Earlier studies of patronage, conceptualized the relation in opposition to democratic, bureaucratic and modern structures.  Framed within a narrative of modernization, patron-client relations were seen as a pre-modern phenomenon that would disappear with the growth of democratic or class-consciousness. Departing from these assumptions, the new studies saw patron client relations as integral and central to the growth of modern economic and political structures ( Wolf 1966; Scott, 1972, 1977; Gellner 1977; Hall 1974;  Weingrod 1972; Eisenstadt and Roniger 1980)

These new studies had certain recurrent themes. They explored the way hierarchy, inequality and asymmetry was structured within a system of interpersonal relations, the way power was constituted symbolically through notions of honour and shame, relations of friendship and kinship. They showed that a system of patron-client relations was characterized by a simultaneous exchange of resources: support and loyalty against duties, long-term credit, and obligations.  This exchange was simultaneous, and was sustained by a two-way flow. The system could not work if the flow stopped from any one side. Within the system, solidarities were forged through the language of inter-personal loyalty, attachment, personal honour, and obligation. The relations were not contractual, but were based on informal understanding. Ties were established voluntarily and abandoned voluntarily. Since they forged vertical bonds, these ties undermined horizontal group organizations. Premised on the patron’s access to means of production, markets and resources, these were relations of inequality and power, which appeared as relations sustained through inter-personal sentiments and emotions.

Breman’s study belonged to this new genre of writings that explored the inner working of the patron-client system. He developed, elaborated, and worked out themes that resonated with other writings on Southeast Asia. Reacting to the notion of forced labour, the argument of patronage, emphasized the voluntary nature of the pact. Patrons and clients both appear complicit in establishing a relationship that served the interests of both simultaneously; and hence both patrons and clients wanted its continuance. Breman’s argument, thus, redefined the field of agrarian labour studies, deepening and enriching it, forcing historians and anthropologists to look not only at structures and their inner functioning, but persuading them to explore the perceptions of those involved in the relations, their calculations, their strategies of survival.

While Breman inflects the notions of patronage and exploitation, his use of the concept  ‘voluntary’ remains rather unproblematic.  How does one define what is voluntary and what is forced? Are the two so distinct and opposed? Did not force underlie what appeared as voluntary? In any case, were the attitudes of the Dubla so unambiguous, particularly when choices are made under such severe constraints, when rupture of the relationship is so difficult, when Anavils brought run-away halis back to work?  Can we conceptualize such relationships through the usual binaries: freedom/ unfreedom, voluntary/ forced?

The focus on mutual consent as the necessary premise of the working of the system also leads Breman to under-theorize the role of contestation in the constitution of the system. The halis may not have risen in revolt, but they did question their master, sought redressal of grievances, and held out the threat of flight. The idealization of dubla loyalty by the Anavils was underwritten by the unstated fear of disloyalty; generosity and beneficence was sustained by the threat of flight; the experience of insubordination in the present led to a nostalgic return to a mythic past of harmony and order. The public script of loyalty and deference can, as Scott as argued, coexist with a hidden script of every day resistance. This resistance may not subvert the system, but it does alter its contours.

Breman's text, it appears to me, is pervaded by an inner tension. It oscillates between description of the system as model and the system as practice. He is acutely aware of the difference, frequently reminding the reader of this, and yet he often moves from a discussion of one to the other rather unproblematically - a move that creates an impression that the world of the halipratha in someway corresponded to the model, if not in the present, sometime in the past. This inner tension of the text is related possibly to another, one that has to do with questions of representation. Breman's work is wonderfully evocative, he takes the readers into the world of halipratha, makes them listen to the anavils and dublas, understand their calculations, perceptions, and experiences. In the process however, the reader begins to see the system through the idealized projections of the informants, the past begins to appear as a time when the model worked in its ideal form, the present was a time of breakdown. The distinction between the model and the practice is temporalized: it becomes analogous to the difference between the past and the present.

The discourse of bondage

Gyan Prakash, in his finely crafted study of the kamauti system in South Bihar, emphasizes the need to understand the power of colonial discourse in re-figuring the kamia-malik relationship in Bihar. The discourse of freedom figured the kamia as a bonded slave, as a person who had lost his natural right to liberty. It was in opposition to the notion of free labour that the term kamia was translated  as un-free labour, and seen as bondage. The progress from unfreedom to freedom, from bonded to free labour was seen as a progress towards modernity, a process of restoring the natural rights of liberty. Within this narrative of progress, bondage had to be seen as a legacy of pre-British, pre-modern past, and the movement against slavery a feature of colonial rule. Prakash shows how the conceptions of the British conflicted with the perceptions of the kamia’s in constituting the relationship.

In their own oral epics, the Bhuinyas located the origin of their historical status as a ritually polluted group engaged in agricultural labour, through a narrative return to a world inhabited by birs and Gods, a world in which the Gods appear mean and deceitful, and the lowly appear divine. The stories develop through a series of reversals. Tulsi bir, the original ancestor, is represented as a warrior, as one with an elevated status, someone who employs labour and confronts Gods. Yet he has at the same time marks of dependence: his father had worked for Bhabanas. Tulsi’s loss of power is explained through a series of pollution themes. Through divine trickery he comes into contact with and consumes some polluting substance, beef in some tales and placenta in others. This story of pollution reveals the bir’s inner purity and the meanness of the Brahmanical gods.

Prakash suggests that the oral tales show how important the question of power and caste was within the imagination of the kamias. On the one hand, the ancestors are seen as birs, as heroes who could confront Gods, as raja like figures. On the other, the kamia category existed only in so far as it related to and was signified by the caste order. The origin of their lowly status was in an originary polluting act, a fact that acquires meaning only within the terms of caste norms. From these oral traditions and other written texts, concludes Prakash, the Kamias in the pre-colonial period, emerge not as unfree labourers, but as dependent servants of dominant landlords. It is in the colonial texts, and through colonial practices, that dependence was re-conceptualized as unfreedom.

As maliks consolidated their power through their command over money and produce,  kamias were subjected to a variety of practices that defined and documented them as persons who had lost their natural rights to freedom because of the power of loans - of things, money, land and grain. Identified as the essence of humanity, freedom was projected as a natural object of desire, an inspiration behind the anti-slavery movement. The kamias who had lost this freedom through debt and loans, lived in a world of suspended rights. Debt bondage freed the kamia from association with caste that earlier characterized his relation with the malik, and tied him anew with loans.

Building on Bourdieu's ideas, Prakash re-conceptualizes the relationship of patronage and exploitation as an economy of gentleness and violence. The life of the kamia was crucially dependent on this economy of gentleness. Since their wages sustained them only for ten months in a year, their needs in other months could only be met through the master’s assistance. The master’s help and loans in times of crisis appeared as acts of generosity and paternal care, binding the kamin in structures of dependence. But if the kamia refused his labour seeking to work elsewhere, then the malik ensured submission through a variety of forms of physical violence. If the kamia fled to other villages, contesting the limits of malik oppression, then the malik fell back on the policy of violence and gentleness. When Keso Bhuinya disappeared to his wife’s village for four days during the time of transplantation, Birji Babu beat him unconscious on his return. Reacting to the violence, Keso Bhuinya disappeared again to his wife’s village for three and a half years. Birji Babu could now re-establish the relation only through a display of gentleness and symbolic power. He arrived on his elephant and showered Keso with gifts.

Unlike Breman’s framework, contestation has an important space in Prakash’s narrative. Kamia-malik relations were defined by a process in which kamias both  contested the relations and were complicit in their elaboration. They continued with every day acts of resistance – working for others, stealing time, and pretending to be sick when the malik expected him to work. Such daily acts of resistance, carried through under pretence of acquiescence, limited the malik’s exercise of power. ‘But in so far as resistance was disguised as acquiescence, such acts also acknowledged the malik’s dominance.’

Prakash demonstrates with great effect the importance of the discourse of freedom in both defining the kamia-malik relationship and narrativizing its history. But, I think, there are a few problems with the argument. First: the use of the concept of freedom, and the notion of bondage does not necessarily imply that we operate within the liberal teleology of progress, and see capitalism and modernity as inevitably liberating. Debates on bondage, forced labour and slavery have shown how bondage not only co-exists with but also is consolidated within the framework of capitalism and an expansive market economy. (see Steinfeld 1991, 2001; Brass and Linden 1977; Banaji 2003) These studies do not necessarily see the move towards free wage labour as a logical corollary to the expansion of market, or the unfolding of legal constitutional system (Steinfeld1991; 2001). The move is defined by the nature of struggles, structured by co-relation of forces and the politics of a time.

Second: the use of the notion of freedom is not as problematic as Prakash suggests. If liberals have discovered in freedom, the essence of human nature, and seen freedom as a natural right, which needs to be recovered whenever and wherever it is lost, we need not operate within the terms set by them, giving the term over to liberal essentialists. Concepts can be deconstructed, as Prakash does, but also refigured, emptied of sedimented meanings and refilled with new meanings.  It is possible to see freedom as specifically defined by societies, in particular locations and times; its meaning constituted through social practices and negotiation. Steinfeld for instance shows how the idea of 'free labour' and 'free contract' had continuously shifting meanings in nineteenth century Britain. Agricultural workers imagine their own idea of freedom, imaginings that sustain their lives and inspire their struggles. What we need to see is not whether 'freedom' and 'coercion' existed, in some hypostatized form, but how particular ideas and structures of coercion/freedom are defined and historically shaped. We need to differentiate determinate notions of freedom and coercion.

Third:  Prakash’s shift from the idea of 'bondage' and 'un-freedom' to the notion of 'dependence' has the same conceptual problems that he seeks to criticize. Dependence can only be defined in opposition to the idea of independence, and the notion of the free individual subject. It is a notion firmly rooted within the liberal discourse of self: the Enlightenment notion that independence from pre-given societal constraints is crucial to the making of an autonomous, self-constituting, meaning attributing individual, someone who can exercise his free will, free agency, private volition. How then does the use of this category help us transcend the limits of Enlightenment discourse? To me it appears that the need is not so much to invent new terms (this too is undoubtedly important), untainted by the burden of the past; but to rework each term, by exploring its assumed essence and by unpacking the frameworks of thought that saturate it with specific meanings. And this Prakash has done with great effect. If elements are plucked out of a paradigm or a structure of discourse and reintegrated to another, then their original essence is ruptured, their meanings are re-figured. We need not be so anxious, so stricken with the guilt of complicity, in using terms seemingly tainted by Enlightenment thought.

Fourth: Prakash tends to overstate the codifying power of discourse.  Clearly we need to understand the nature of the framing discourses of labour, but we must equally recognize the limits of their shaping power, and look at the ways in which framing categories themselves were questioned and reworked. In any case, was there only one source of this framing discourse? Did not colonial conceptions of village society and official descriptions of patron-client relations within rural areas, draw upon anti-Enlightenment ideas of community and  'folk', affective ties and emotional links, as much as Enlightenment ideas of freedom and bondage? The finest accounts of these relations often were penned by conservative paternalists officials who had a complicated and tortured relationship with Enlightenment ideas.

To recapitulate the discussion in Part  I: in the last four decades the field of agrarian labour studies in India has continuously broadened and mutated. The framing questions have changed, the operating assumptions have been rethought, and the conceptual terms of discussions have shifted. The earliest studies focused on the question of numbers. The nationalist assumption that agricultural labourers were absent in pre-colonial period, and that they dramatically increased under British rule, was effectively and powerfully questioned by Dharma Kumar. While scholars debated the picture of stasis that Kumar presented, and Marxists reflected on the problem of conceptualizing labour forms, Breman began to explore the world which labourers and employers inhabited. Historians now began looking at the cultural categories through which social relations were ordered, the perceptual frames through which the labourers and employers looked at each other, and the duties and obligations that tied them together. Recent studies of Gyan Prakash and others have shown how our notions of labour are also structured by the discourses of freedom and bondage that dominate the post-enlightenment world.

But to assess our knowledge about agricultural labour in the colonial period, we need to return to many of the old issues, synthesizing the existing evidence and re-thinking our answers. New historiographical insights do not condemn old questions into oblivion; they force us to re-look at them in new ways.  We need to reassess the question of numbers, discuss the forms of labour force, and reflect on the nature of the labour market. This I seek to do in the next part of the essay.

Part II

Persistences: numbers

First we return to the question of numbers. A look at the calculations may not tell us the exact number of labourers, but they do inform us of the limits of codification – what could and could not be classified, what the census makes visible and what it erases. Behind the numbers we need to look at the politics of the record that shapes our notion of the labour force.

The number of agricultural labourers is difficult to calculate with any degree of precision. But in general census figures show an amazing stability in the percentage of agricultural population to total work force, in most provinces of India. Apart from Punjab there was no substantial increase in any other region. Krishnamurty’s recalculation of the all India Census figures shows that that the proportion of male agricultural labourers to the total workforce fluctuated around 25 per cent (Table I). This also meant that the spatial distribution of agricultural labourers did not change substantially. The inter-regional variation in the incidence of labour force did not narrow. In 1881, as in 1951, Bengal, Bihar, Central Provinces, Madras, Kerala, had a larger incidence of agricultural labourers than provinces like Punjab, UP, Rajasthan.

Table 1

THE SHARE OF AGRICULTURAL LABOURERS IN THE

AGRICULTURAL WORKFORCE IN INDIA: 1881-1931

Source: Krishnamurty (1972)

What does this seeming persistence tell us? To what extent can calculations of numbers confirm the picture of changelessness within the agricultural labour force?  I would suggest that if SJ Patel was wrong in using Census data to suggest that there was a dramatic change in the structure of labour force, it is equally problematic to use the Census data to demonstrate the picture of stability. We need to read the figures to see what it cannot tell us, not simply to calculate what it does. What I seek to do is not to question Krishnamurty’s calculations, but to point to its limits. This might give us a different kind of a socially embedded picture of the agricultural labour force.

 If we read the Census data in relation to the specific social and ecological specificities of each region, then an interesting pattern can be seen. Let me summarize a few features of this pattern:

1. The weight of agricultural labourers within the agricultural population was most significant in wet zones where labour intensive crops could be grown and an intensive cropping pattern could be adopted. Here labour demand was high and a large supply of labour could be absorbed within agriculture.

2.  Intensive cultivation could sustain a higher density of population in the wet zones. The increased pressure of population and the consequent fragmentation of holdings created a large reserve of potential labour.

3. The proportion of agricultural labour was likely to be higher in regions where cultivated land was controlled by landowners, sometimes of higher castes (Brahmans, Rajputs, Vellalas), who abstained from manual work, compared to regions where cultivating peasant castes (Jats, Kunbis) held land, and family labour was widely utilized. This pattern reinforced the dry and wet region contrasts since productive soils were more likely to sustain "unproductive" elites. Historians have usually found caste ..000``hierarchies more deeply entrenched in fertile regions producing a substantive surplus. (Stokes 1978; Stein 1982; Ludden 1985).

The contrast between wet and dry zones can be seen within each province. In Maharashtra, the proportion of agricultural labourers was higher in the coastal Konkan belt with high rain-fall, intensive rice cultivation (primarily in north Konkan), high density population, small sized holdings and Brahman domination (Charlesworth 1985). The drier eastern upland Maharashtra, the millet zone with large holdings, had a lower incidence of labourers.

The Madras wet zones, usually the valleys, were again characterized by intensive paddy cultivation, high-density population and smallholdings. These were the regions of Brahman-Vellala power and widespread agrestic servitude (Kumar 1968, Baker 1984, Ludden 1985). In the drier uplands, agrarian hierarchy was weakly developed and agrestic servitude less prevalent. In North Arcot, Salem, Madura, Coimbatore the collectors and judges reported that slavery was non-existent. Census figures show that in these districts the proportion of agricultural workers to total work force was 15 to16 per cent, compared to 29.5 per cent in Malabar, a wet coastal district.

In Punjab, the proportion of agricultural labourers to total work- force was the highest in the canal irrigated western cotton tracts, followed by the intensively cultivated central wheat zone; and the lowest in the southeastern dry millet zone. In 1931, agricultural labourers formed 29 per cent of the work force in Lyallpur, a canal district, 22 per cent in Jullunder in the central wheat zone, and 9.3 per cent in the arid south- eastern district of Hissar (Bhattacharya 1985).

Such general explanations, ecological or sociological, suggest valid broad contrasts. The way determinants inter-relate at the local level need a more concrete analysis. Sociological determinations worked in complex, and sometimes apparently conflicting, ways. Caste, for instance, did not always define work ethic. The Brahman mirasidars of Tambraparni valley, a fertile wet tract in Tirunelveli district of Madras, refused to soil their hands, preferring to live on the labour of Pallas and Pariahs. In the neighbouring drier tracts, north and south of the valley, the ritual taboos against polluting manual work tended to break down. Here, the Brahmans were hardy self-cultivating peasants (Ludden 1985). So the conditions of cultivation in different eco-zones could be more important than caste in defining labour relations. The same caste in two different eco-zones could have opposing work ethic.

The ecological argument similarly needs to be extended with caution. Wet and dry zones were not characterized necessarily by uniform labour relations and invariant work ethics. Bengal, for instance, was a wet zone, but with a range of distinct regional social structures (Chatterji 1982 ; Bose 1986). The east, with its fertile new alluvium soil, intensively cultivated with aman rice and jute, was predominantly a region of small holding self-cultivators. In the stagnant old alluvial delta of West Bengal, susceptible to unpredictable monsoons and distant from the fertilizing action of annual floods, zamindars cultivated their khas-khamar with the help of a large class of agricultural labourers. In north Bengal, a newly reclaimed area, jotedari and share cropping were widely prevalent (Bose 1986).

4. Regional variations in numbers could be produced by statistical biases.  Census data, I have argued elsewhere, seriously understate the number of agricultural labourers from peasant households. Rajputs, Jats, Kunbis, or Brahmans rarely stated agricultural labour to be their occupation. Persuaded by norms of purity/pollution and a sense of honour, peasant proletarians considered themselves to be cultivators, not agricultural labourers. Only lower castes admitted to being labourers.

Thus, in regions where lower castes (or tribals), and permanent farm servants are significant constituents of the labour force ( as in Bihar, Madras, Central India, or South Gujrat) the census would show high figures of agricultural labourers. In regions with a large preponderance of casual harvest labour, a low incidence of lower caste labour, and a high percentage of peasant labourers, the census would show a low proportion of agricultural labourers.

Historical studies based on the census tend to reproduce the bias of their source and underestimate the number of agricultural labourers from peasant background. This distorts comparisons not only across regions of India but even within each province.

Consider Bengal. Census figures show that agricultural labourers to total agricultural work force was low in the eastern districts, around 19 per cent in 1931. The proportion was much higher in west and central Bengal, around 43 per cent. Underlying this contrast was a difference in the composition of the labour force. In west Bengal, agricultural labourers  formed a class socially distinct from cultivators. The former were mostly lower castes or tribals (Bagdi, Bauri, Santhal); whereas the latter were from relatively higher castes (Mahashya, Sadgop, Aguri). In east Bengal, on the other hand, cultivators and agricultural labourers were not two identifiable social groups. Most labourers were pauper peasants who turned to wage work during harvest months. Jack tells us that landless labourers were unknown in Faridpur;  but 30 to 37 per cent of the peasants with land were working as agricultural labourers. The recorded proportion of agricultural labourers was invariably low in such regions where the identity of a cultivating peasant was difficult to separate from that of a labourer.

Thus, a focus on numbers produces a picture of stability. Census figures would suggest that the relative weight of agricultural labour force did not significantly increase over time, nor did its spatial distribution change very much. Commercialization did not necessarily lead to an increase in the agricultural labour force. In fact, Punjab, one of the most commercialized zones, had the lowest incidence of agricultural labourers. Once we probe the categories we realize that they invisiblize certain forms of labour and repress changes within the labouring population.

Transformations: Forms of labour

Quantitative measures are important. But they have their limits. They tell us little about the social forms of existence of labour. What we need to consider are the changes within forms of labour, the basis of their production and reproduction. These cannot be always quantified. Pointing to the predominance of lower castes within the agricultural labour force, Kumar had proposed her thesis of continuity, arguing that the resilience of caste structure absorbed the transformative impact of British rule (Kumar 1962). But apparent similarities in social forms of labour very often conceal changes in their essential characteristics.

Consider the Kamiyauti relations in south Bihar discussed by Prakash. Within this system, as we have seen, long term ties bound the kamiya, mostly Bhuiyas in the pre-colonial period, to the upper caste malik (Prakash 1990). He worked all his life for the same landlord. His son inherited the bondage.

This pre-colonial form of labour relation did not disappear with the spread of market forces. Yet, as Prakash shows, its basis was transformed in significant ways. In the beginning of the nineteenth century, an advance of money was not a defining feature of the relationship. It "rested on unequal social hierarchy and was secured through reciprocal action." By the late nineteenth century, advances acquired a new significance. Now Kamiyas were being employed by Koeris who were low caste peasants. The Koeries had acquired money and could advance credit, but they had no power within the ritual hierarchy. Bhuiya oral tradition defined the collective position of Bhuiyas in terms of their dependence on upper class maliks.  When labour relations could not be secured through ritual power and prescribed status, money advances inevitably gained in significance. Gradually non-Bhuiyas were also bonded through kamiyauti loans. Bondage, earlier a ritual relation, was now sustained by the power of money.

In Punjab, the sepidari system was crucial to the functioning of the agrarian society. The allocation of labour within the sepidari system was a function of caste prescriptions. The system regulated the supply of diverse forms of labour, essential not only for agrarian production but for sustaining everyday life. Apart from their specialized crafts, each sepidar had specified ritual and social roles to perform: on the occasion of marriage and death, during religious ceremonies and festivals, at harvest and sowing times.

If the system specified sepidar functions, it also assured their rights and dues: their customary share after harvest, food and clothes during social occasions, help during situations of crisis. These rights of subsistence and work were the basis of the system's viability, its acceptability. They were an integral, not a dispensable, part of the relationships.

Within the system, power and domination were secured through specific norms. Acquiescence to caste specific roles were not based on coercion alone. Essential to the relationship were acts, almost ritually enacted, which projected the zamindar as benevolent and humane. Gestures of generosity initiate and sustain social relations. Indebted, the recipient of 'kindness', is persuaded to reciprocate with what he can offer - labour. Hidden behind relationships of reciprocity, bondage did not appear as such.

By the late nineteenth century, the norms sustaining the sepidari system were breaking down. The landowning cultivators attempted to maximize their gains and the sepidars distrusted the 'beneficence' of the former. With increased commercialization of agriculture and general rise of prices on the one hand, and the tendency of the kamins to assert themselves on the other, the cultivators were reluctant to accept customary rights. The sepidars were increasingly paid by the job. This deprived them of all claims to customary rights and dues: the right to pick greens for their daily meals or draw water from the wells, the right to take the carcasses of dead animals. the right to collect dry wood for fuel, the right to a share of the crop at harvest time, and the right to glean in the fields after that. Within the earlier system the kamin sepidars were not just labourers: they performed a multiplicity of functions, and were involved in diverse forms of exchanges. By the turn of the twentieth century the sepidars were seen as labourers.

Stability in the caste composition of the labour force, and a seeming persistence of traditional structures of subordination of labour, do not imply a lack of change within the existing social forms. Change need not be made visible and intelligible only through liberal teleologies that look for the transition from forced to free wage labour.

  Integration or segmentation?

Did commercialization of agriculture create an integrated labour market? A discussion of this question is important for our purpose. Within liberal teleologies it is often assumed that market expansion leads to a process of integration that equalizes opportunities, enables labour mobility, dissolves primordial forms of labour, and creates a free labour force. There is a powerful reductive logic implicit in this teleology of the market that simply does not work. It is a logic that can neither be conceptually sustained not historically demonstrated - a point we have already discussed in the earlier section. To develop the argument further we need to explore the meaning of integration - its structure and implication.

Labour mobility was not entirely a new phenomenon in the nineteenth century. And this mobility increased over time. In general, the movement was from dry insecure tracts where labour demand was marginal to wet regions or irrigated tracts where intensity of cropping was higher and labour was much in demand; from single to double cropped areas; from millet tracts to paddy fields.

In Andhra, during transplantation and harvesting of rice, local labour supply was not sufficient to meet the demand: by the end of the nineteenth century in the peak season over 50 per cent of the labourers were migrants.. From the 1860s we hear of poor peasants from the drier upland taluks of Vizagapatnam coming in annually to work as day labourers and returning home after the operations were over (Reddy  1979;1985). In Tamilnad thousands moved every year from the poorer districts of Vizagapatnam, and from the uplands of Godavari, Kistna and Guntur to the riverine lands of Kistna and Godavari (Baker 1984; Bates and Carter 1992). In Punjab, labourers moved, sometimes with their families, from the dry southeastern tracts to central Punjab, and from central Punjab to the Canal Colonies. (Bhattacharya 1985)  Migration flows out of the arid tracts of southeast Punjab were regulated by the nature of harvest. In a bad year the numbers swelled. Pauper migrants wandered from village to village in search of work. In Maharashtra, labourers from the dry Deccan tracts moved to the northern cotton belt during the cotton picking season. This flow was also regulated by the conditions of harvest in the dry zone (Charlesworth 1985).

In eastern India there was again a complex network of migration flows (Van Schendel  and Faraizi 1984). (1) Labourers from south-western Bengal, south Bihar, Orissa, and Central Provinces went to north Bengal to reclaim jungles. This was a tribal stream of Santhals, Oraons, and  Mundas.(2) 'Upcountry' labourers from Orissa, Uttar Pradesh and north Bihar  moved to west and central Bengal during winter, and returned home during spring and early summer. These 'upcountry' labourers could be seen harvesting, transplanting, and husking rice, and taking to earthwork. (3) Within Bengal, local labourers were on the move in winter months. Generally the movement was from west to east; from the stagnant old alluvial delta adversely affected by the shifting course of river, to the fertile new alluvial delta producing aman rice and jute. Labour demand and wages increased as one moved east. Some migrants could hope to cultivate, as seasonal cultivators, the chars (fresh alluvial lands) thrown up regularly by the rivers, or the marshes of Sunderbans (bhati). But most seasonal migrants worked as harvesters. Jack describes how the poor ryots from Nadia moved to Faridpur, where they were lodged and fed by Faridpur ryots while the harvesting continued (Jack 1916, also Sen 1916 quoted in Van Schendel and Faraizi 1984). In the meantime the Faridpur ryot "takes boat and goes off with a party of his neighbours to the fertile land of Bakarganj and Mymensingh...".  The luxuriant paddy crop in Bakarganj, we are told, was regularly harvested by the "hordes of labourers (who) come from Dacca and Faridpur during the harvest season and do all the work connected with harvest" (Sen 1916 quoted in Van Schendel and Faraizi 1984). Rural labourers from Chittagong went further east into Burma and worked on the fertile coastal plains of Arakan as harvesters. It was reported that without Chittagong labour "the Aracan crop could not be gathered, or prepared for export."

So: agricultural labour in the nineteenth century was mobile. Yet integration of the labour market was limited. Wage variations between regions continued to be marked: only after the 1960s does the co-efficient of variation in wages show any definite fall. Between 1900 and 1950 the variation across regions was around 35 to 40 per cent . After the 1950s it declined to about 25 per cent (Reddy).

Village surveys of the early twentieth century show that wage rates in neighbouring villages very often varied sharply, and this variation persisted over time. The movement of labour from one village to another did not equalize wages. This variation was, in fact, a structural feature of a labour market that was becoming integrated at one level and fragmented at another. Differences in labour demand and wages were bound to persist as long as agricultural growth in different zones continued to be uneven; and millet and paddy fields, dry and wet agriculture, single and double crop areas continued to co-exist. In such a context, mobility of labour between agro-zones would co-exist with wage disparities between them.

Fragmentation was sustained by other characteristics of the labour market. Labour mobility was not determined by wage differential alone, nor did labour invariably flow into regions with the highest wage. The poor, always on the look out for work, stopped at any place where work was available. And so long as there was work in one place, higher wages in a neighbouring area did not attract labourers there. In Bengal, Nadia labourers worked in Faridpur, but did not go upto nearby Bakarganj where wage rates were higher.

But why did Faridpur peasants labour in Bakarganj and employ outside labour in their home village? Wage differentials provide a partial explanation: wages in Bakarganj were higher. But social considerations were also at work. Agricultural work away from home was more acceptable to peasants: it was less degrading. Decisions about the nature of work continued to be determined by such social considerations, and not by wage differential alone.

Regular migration flows, once established, acquired a certain rigidity. Labourers usually moved in established circuits, often in groups with a group leader. Community and regional solidarities, which could provide a sense of security in alien milieus, defined community patterns of labour flows. People from one community, or one region, all looked for work in the same place. Once links were forged, employers preferred labourers from known catchment areas, and jobbers linked specific labour catchment areas with particular points of supply. So in each region we find some homogeneity in the community composition of immigrant labourers - a fact that again fragmented the labour market.

Commercialization of agriculture, thus, created a labour market that was at once integrated and fragmented, linked at one level and segmented at another. This continued fragmentation was no reflection of weak penetration of market forces within the rural economy. Framed within the terms of liberal economics, such an argument sees the market as a powerful dissolving agent which recreates the world in its own image, freeing labour from primordial ties, creating a free labour market. Such liberal teleologies that associate the idea of the market with the idea of freedom and progress, and progress with the dissolution of the old order, have now been widely questioned. Existing social institutions and bonds do not evaporate with the spread of the market: they in fact define the specific form in which the market operates; they shape the peculiarities of the market. Occurring within institutional and social contexts, labour market integration and segmentation were complementary processes, not opposed.

Circulation of labour did not weaken the structures of segmentation; they deepened them, reproducing them on an expanded scale. In the post Independence period, peasant capitalists have utilized this segmentation to restructure the labour market, resist the demands of local labourers, employ migrant labour, ensure conditions of surplus labour supply, and check wage rise. In Bardoli, Kanbi Patidars employ Khandeshi migrants on their farms, in preference to the assertive local halapatis (Breman 1985). The tribal migrants are seen as vulnerable and naïve: they accept low wages, work for long hours, and are submissive. In Punjab similarly cheap Bihari migrant labour is employed to displace the unionized local labour, cut wage costs, and establish control over labour force. Segmentation fractures the working community, pits one section against another, and intensifies the conflict between the locals and migrants. Accumulation does not proceed through an uncomplicated process of unification of the labour market: this unification not only comes up against barriers and resistance, it in fact occurs through processes of segmentation, fragmentation and differentiation of the labour force.  Conversely, we could say, segmentation happens through circulation of labour, employers welcome the flow of pauper migrants that allows them to restructure the labour market and intensify forms of coercive control over both outside and local labour (Breman 1985, Tom Brass 1986, 1997). By intensifying the internal conflict within the labouring groups, it constrains the possibilities of struggle and negotiation through which conditions of labour usually change.

Concluding comments

I wish to end by re-emphasizing two points. One: the activity of agriculture labour is easier to identify than the person 'agricultural labourer'.  Agriculture labour is a relationship that sustains agrarian production, but is not always an identity stamped on the body of a person. The same person could be a poor peasant clinging to his land, a migrant looking for a job, an unemployed tramping around, and an agricultural labourer in the peak season. The poor peasant who worked as a harvesters or thresher, most often preferred to see himself as a peasant rather than a labourer. His relationship to land defined his identity more than his relationship to labour.

Second: in looking at agricultural labour we do need to move beyond the usual opposition between forced labour and free labour, voluntary labour and involuntary labour. Since the nineteenth century, the language of law has sanitised coercion, the fiction of freedom has masked the pressures of compulsion (Steinfeld 1991, 2001;Banaji 2003) But we need to be wary of the counter move. A blurring of the boundaries between these neat categories does not mean that we discover sameness where we earlier saw difference, that beneath the appearance of freedom and the language of contract, we see only a continuity of earlier forms of coercion and unfreedom. When new languages become necessary to sustain old forms, we need to explore the new language and the history of its necessity, and we need to see how the old forms are modified in new contexts. This requires that instead of sameness we look for differences, that we explore the variety of ways in which labour regimes work, and the ways in which these are reproduced and legitimated within specific historical situations.

References:

 Amin, Shahid, Sugarcane and Sugar in Gorakhpur: An Inquiry into Peasant production for Capitalist Enterprise in Colonial India, Delhi 1884.

Anon., 1979, 'Plight of migrant labour', E.P.W., XIV, 1979.

Baker, Christopher John, An Indian Rural Economy, 1880 - 1995: The Tamilnadu Countryside, Oxford and Delhi1984.

Baker, CJ, An Indian Rural Economy: The Tamilnadu Countryside, 1885-1955, Oxford, 1984

Banaji, Jairus, 'Capitalist domination of and the small peasantry: Deccan districts in the late nineteenth century', Economic and Political Weekly, 7, no. 33-34 Special Number, August 1977, pp. 1375 1404.

Banaji,-----, 'The fiction  of free labour: contract, coercion and so-called free labour', Historical Materialism, 11, 3, 2003, pp.69-96.

Bandyopadhyaya, N., 'Causes of sharp increase in agricultural labourers 1961-71: a case study of social-existence forms of labour in North Bengal', E.P.W., XII, Dec. 1977,  Review of Agriculture, A111 - 26.

Banerjee, Himadri, 'Agricultural labourers of the Punjab during the second half o the nineteenth century', The Punjab Past and Present, 11, 1977.

Bardhan, Kalpana, 'Rural employment, wages and labour markets in India: a survey of research ', E.P.W., XII, Dec. 1977, Review of Agriculture, A 34 - 64.

Bardhan, Kalpna, 'Rural Employment, Wages and Labour Markets in India: A Survey of Research', E.P.W., July 23 1983, pp.1319 - 1334.

Bardhan, Pranab and Ashok Rudra, 'Labour mobility and the boundaries of the village moral economy, The Journal of Peasant Studies, 13. 3, April 1986, pp. 90-115.

Bates, Crispin, and Marina Carter, 'Tribal migration in India and beyond', in Gyan Prakash ed., The World of the Rural Labourer.

Behal, Rana Pratap, and Prabhu Mohapatra, ' "Tea and Money versus human life": The rise and fall of the indenture system in the Assam tea plantations 1840-1908', The Journal of Peasant Studies, 1992.

Berreman, G.D., Caste and Other Inequalities, Meerut 1979.

Beteille, Caste, Class and Power, Berkley 1965.

Bhatia, B.M., 'Famine and agricultural labourers in India: a historical perspective', Indian Journal of Industrial Relations, 10: 4, 1975.

Bhattacharya Neeladri, ‘Agricultural labour and production: central and south-east Punjab 1880-1950’, in K.N. Raj , et.al.ed., Essays on the Commercialization of Indian Agriculture, OUP, Delhi, 1985.

Biedelman, T.O., A Comparative Analysis of the Jajamani System, Monographs of the Association of Asian Studies, VIII, New York 1959.

Bose, Sugata, Agrarian Bengal: Economy, Social Structure and Politics, 1919-47, Cambridge, 1986.

Bose, Sugata, Peasant Labour and Colonial Capital, CUP, 1993.

_________, Credit, Markets and the Economy of Colonial India, Delhi 1994.

Bougle Celestin, Essays on the Caste System, Cambridge 1971.

Brass, Tom, 'Unfree labour and capitalist re-structuring in the agrarian sector: Peru India', Journal of Peasant Studies, 14, 1986a

-----, 'Some observations on unfree labour, capitalist restructuring and deproletarianization', in Tom Brass and Marcel van der Linden, eds., Free and Unfree Labour: The Debate Continues, Berne 1997

Breman, Jan, Patronage and Expolitation, Manohar, New Delhi, 1979.

_____, Of Peasants, Migrants and Paupers: Rural Labour Circulation and Capitalist Production in West India, OUP, Delhi, 1985.

Brenner, Robert, 'The origins of capitalist development: a critique of neo-Smithian Marxism', New Left Review, 104, July-Aug, 1977, pp. 25 -92.

Charlesworth, N., Peasants and Imperial Rule: Agriculture and agrarian society in the Bombay Presidency, 1850 - 1935, Cambridge 1985.

Chatterjee, Partha, 'Agrarian structure in pre-partition Bengal', in Asok Sen, Partha Chattrjee, Saugata Mukherji, Perspectives in Social Sciences 2, Calcutta 1982.

Chatterji, B., 'Agricultural labour, enterprise and land reforms in India', Enquiry, 2 & 3.

Chattopadhyay, ----, 'Mode of production in Indian agriculture: an anti-kritik', E.P.W., 1972.

Chattopadhyay, Paresh, 'On the mode of production in Indian agriculture: a preliminary note', E.P.W, March 25, 1972.

Cohn, Bernard S., 'The changing status of a depressed caste', and 'Changing traditions of a low caste', in Cohn, Anthropologists among Historians and Other Essays, Delhi 1987.

Commander, Simon, 'The jajmani system in north India: an examination of is logic and status across two centuries', Modern Asian Studies, 17, 2 , 1983, pp. 283 -311.

Eisenstadt S. N. and Louis Roniger, 'Patron-client relations as a model of structuring social exchange', Comparative Studies in Society and History, 1980, pp. 43-77.

Epstein, T.S., 'Productive efficiency and customary system of rewards in rural south India', R. Firth, ed., Themes in Economic Anthropology, London 1967.

Gellner, E., 'Patrons and clients', in E Gellner and J. Waterbury eds., Patrons and Clients in Mediterranean Societies, London 1977, pp. 1-6.

Ghose, Kamal Kumar, Agricultural Labourers in India: A Study on the History of their Growth and Economic Condition, Calcutta 1969.

Gough, K., 'The social structure of a Tanjore village, in M.N. Srinivas ed., India's Villages, Bombay 1960.

Gould, H.A.,'The Hindu jajmani system: a case of economic particularism', Southwestern Journal of Anthropology, XIV, 1958, pp. 428 -37.

----- 'Priest and contrapriest: a structural analysis of jajmani relationships, Contribution to Indian Sociology, N.S., I, 1967, pp. 26-55.

Gould, ------,'A jajmani system in north India: its structure, magnitude and meaning', ETHN, Vol.III, pp. 12 - 41.

Guha, Sumit, The Agrarian Economy of the Bombay Deccan, 1818 -1941, Delhi 1985.

Gupta, S.C., 'New trends of growth', Seminar, 38, 1962.

-------., 'Some aspects of Indian agriculture', Enquiry, 6.

Hall, A, 'Concepts and terms: patron-client relationship', Journal of Peasant Studies, 1, 1974, pp. 506 -9.

Harper, E.B., 'Two systems of economic exchange in village India', American Anthropologist, 61, 1959, pp. 760 ff.

Hilton, Rodney, ed., Transition from  Feudalism to Capitalism, London 1978.

Hjejle, 'Slavery and Agricultural Bondage in South India in the nineteenth century', Scandinavian Economic History Review, 15:1-2, 1967, pp. 71 -126.

Ishwaran, K., Change and continuity in India's Villages, New York 1970.

Jack, J.C., The Economic Life of a Bengal District: A Study ( 1916; Delhi 1975)

Kessinger, T.G., Vilyatpur, 1848 -1968. Berkeley 1974.

Kolenda, P.M., 'Towards a model of the Hindu Jajmani System', Human Organization, 22, No. I, 1963, pp.11-39. Kotovsky, G.G., Agrarian Reforms in India.

Krishnamurty, J., ‘The growth of agricultural labour in India – a note’, The Indian Economic and Social History Review, 1972.

Kumar, Dharma, 'Caste and Landlessness in South India', Comparative Studies in Society and History, 4: 3, 1962, pp. 337-63.

-------, Land and Caste in South India, Cambridge, 1965.

Kumar, Ravinder, Western India in the Nineteenth Century: A Study of the Social History of Maharashtra, London 1968.

Lewis Oscar, ' Caste and the jajmani system in a north Indian village', in Anthropological Essays, New York 1970.

Ludden, David, Peasant History in South India, Princeton 1985.

Mahar, Michael., ed., The Untouchables in Contemporary India, Tuscon, Arizona 1972.

Marla, S., Bonded Labour in India: National Survey on the Evidence of Bonded Labour, New Delhi, 1981.

Mihir Shah, 'The Kaniatchi form of labour', Economic and Political Weekly 20:30, 1985.

Mohapatra, Prabhu, 'Coolies and colliers: a study of the agrarian context of labour migration from Chota

 Nagpur, 1880 -1920', Studies in History, New Series, 1:2, 1985.

Mukherjee, Ramakrishna, Six Villages of Bengal, Bombay 1971.

Mukhtyar, G.C., Life and Labour in a South Gujrat Village, Calcutta 1930.

Mundle, Sudipto, Backwardness and Bondage: Agrarian Relations in a South Bihar District, Delhi 1979.

Nand, Brahma, 'Agricultural labourers in western India', Studies in History, NS, 1,2, pp. 221 -46.

Neale, Walter, 'Reciprocity and redistribution', in K. Polanyi et. al., eds., Trade and Market in the Early

Empires, Glencoe 1957.

Oberoi, A.S. and K.K. Manmohan Singh, 'Migration flows in Punjab's agriculture'. E.P.W., XV, 1980, A 2-12.

Pandian, M.S.S., The Political Economy of Agrarian Change: Nancilnadu, 1880 -1939, New Delhi 1990

Parry, J.P., Caste and Kinship in Kangra, London 1979.

Patel, Surendra, Agricultural Labourers in Modern India and Pakistan, Bombay 1952

Patnaik, Utsa, 'Capitalist development in agriculture', E.P.W., Sept 25, 1971.

-------, 'On the mode of production in Indian agriculture: a reply', E.P.W.,Sept 30, 1972.

Pocock, D.F.' Notes on Jajmani Relationships', Contribution to Indian Sociology, 6, 1962, pp.78 - 79.

---------, Kanbi and Patidar, Oxford 1972)

.Prakash, Gyan, Bonded Histories: Geneologies of Labour Servitude in Colonial India, CUP, Cambridge, 1990.

--------, The World of the Rural Labourer in Colonial India, New Delhi 1994

Rao, V.K.R.V., ed., Agricultural Labour in India, Bombay 1962.

Reddy, M. Atchi, 'Wage data from private agricultural accounts, Nellore district, 1893 - 1974', I.E.S.H.R., XVI, 3, 1979.

------, 'Female agricultural labourers of Nellore, 1881 - 1981', IESHER, 20: 1, 1983.

-- -----, 'The commercialisation of agriculture in Nellore district, 1850 - 1916: effects on wages, employment and tenancy', in K.N. Raj et al eds., Essays on the Commercialsation of Indian Agriculture, pp.163 - 83.

Rothermund, Deitmar, 'A survey of rural migration and land reclamation in India, 1885', J.P.S., 4, 1977.

Rudra, A. and M.M. Mukhopadhyay, 'Hiring of labour by poor peasants', E.P.W., 1976, Jan 10, pp 33-36.

Sarkar, Tanika, 'Bondage in the colonial context', in Utsa Patnaik ed., Chains of Servitude: Bondage and Slavery in India, Madras 1985.

Scott, J., 'The erosion of patron-client bonds and social change in rural southeast Asia', Journal of Asian Studies, 32, 1, Nov. 1972, pp. 5-37.

Scott, J., 'Patronage and exploitation', in Gellner and Waterbury, eds., Patrons and Clients, pp.21 -40.

Srinivas, M.N. and A.M. Shah, 'The myth of the self-sufficiency of the Indian village', Economic Weekly, 12: 37, 1960

Stein, Burton, 'South India: some general considerations of the region and its early history', in Tapan Raychaudhuri and Irfan Habib eds., Cambridge Economic History of India, Vol. I, Cambridge 1982.

Steinfeld, Robert J., The Invention of Free Labour: The Employment Relations in English and American Law and Culture 1350 - 1870, Chapel Hill, NC, 1991.

------, Coercion, Contract, and Free Labour in the Nineteenth Century, Cambridge 2001.

Stokes, Eric, The Peasant and the Raj: Studies in Agrarian Society and Peasant Rebellion in Colonial India, Cambridge 1978.

Stokes, Eric, The Peasants and the Raj, Cambridge, 1978.

Thorner, Daniel and Alice Thorner, Land and Labour in India, Delhi 1962.

Van Schendel, Willem and Aminul Haque Faraizi, Rural Labourers in Bengal 1880 to 1980, Rotterdam, 1984.

Weingrod, A., 'Patronage and power', in Gellner and Waterbury, eds., Patrons and Clients, pp.41 -52.

Wiser, William H., The Hindu Jajmani System, Lucknow 1936.

------ Behind the Mud Walls, New York 1946.

Wolf, E., 'Friendship and patron-client relationships in complex societies', in M. Banton ed., The Social Anthropology of Complex Societies, London 1966, pp. 1-22.