Conjugated oppression within contemporary capitalism: class, caste, tribe and agrarian change in India

ABSTRACT

Neoliberal globalisation has resulted in the bypassing of agrarian transition-led industrialisation and classic proletarianisation, and class-for-itself class struggles are rare. Drawing on analyses of class relations, racism and other forms of social oppression, this contribution explores how processes of ‘conjugated oppression’ are central to the spread of contemporary capitalism. The focus is on India and on how the co-constitution of class relations and social oppression based on caste, tribe, gender and region is entrenching Dalits and Adivasis at the bottom of social and economic hierarchies. The analysis has deep-seated consequences for how we think about political struggles, in this case ones that foreground caste and tribe and focus on both labour and land.

KEYWORDS:

• class
• caste
• tribe
• India
• racism
• conjugated oppression
• capitalism

Ever since the Communist Manifesto declared, ‘the history of all hitherto existing society is the history of class struggles’ (Marx and Engels 1969 [1848], 14), class analysis has formed a central part of understanding capitalist development and of shaping political action within and against capitalism. There are, however, different views on what should be at the heart of a class analysis. Some of the core assumptions of radical political economy traditions have been questioned. For scholars, political parties and activists of the Global South, this has included a challenge to the assumption that the conditions of the capitalist development of northern Europe and the US – generally seen as the classic model of development – are being reproduced across the world. With that challenge comes an implicit undermining of the theory of agrarian-led transition to industrialisation, a transition that is assumed to lead to the formation of an oppressed workforce who self-identify as a proletariat. Another critique has come from feminist Marxists and Marxist scholars of racism and ethnicity who have drawn attention to other axes of oppression which are not accounted for in the classic perspective.

This contribution advances these critiques of classic traditions of political economy by drawing on the case of India. It begins with a discussion of recent debates on agrarian transition to show how the analysis of class relations changes in the absence of full-scale proletarianisation. It argues that a class analysis must consider how the spread of capitalism has been marked by class relations that have given rise to informalised, insecure work across the agrarian/non-agrarian divide, and it must acknowledge that these class relations are co-constituted by relations of race, ethnicity, caste, tribe, region and gender.¹ In the Indian context, we focus on the contemporary situation of Dalits (the ex-untouchable castes) and the Adivasis (tribal groups) who together make up one quarter of the Indian population. We explore how processes of what Philippe Bourgois has called ‘conjugated oppression’ – the co-constitution of class-based relations and oppression along the lines of race, ethnicity, gender and, in India, caste and tribe – is entrenching Dalits and Adivasis at the bottom of social and economic hierarchies in India. We also show how these processes that are central to an understanding of the spread of capitalism and contemporary agrarian change have deep-seated consequences for how we think about political struggles that foreground caste and tribe in particular, and focus on land/forests and labour.²

Agrarian transition and the absence of global proletarianisation

An agrarian-led transition to industrialisation, and the related formation of a dominant proletarian condition with workers solely reproducing through capitalist labour relations, is not taking place in most countries of the Global South. This departure from the classic model poses a challenge to the activists and scholars who expected to develop a political strategy based on workers recognising themselves as a social class and therefore becoming and acting as a ‘class-for-itself’. It indicates the necessity to re-assess how we think about class.

Lenin’s paradigmatic definition of class emphasises ‘class’ as a relational concept, placing at its core the ways in which labour is exploited and surplus appropriated from it:

Classes are large groups of people differing from each other by the place they occupy in a historically determined system of social production, by their relation (in most cases fixed and formulated in law) to the means of production, by their role in the social organisation of labour, and, consequently, by the dimensions of the share of social wealth of which they dispose and the mode of acquiring it. Classes are groups of people one of which can appropriate the labour of another owing to the different places they occupy in a definite system of social economy. (Lenin 1919)

This relational concept of class is, of course, a very different approach to class than that provided by a Weberian, economic, stratification-based analysis, or more ‘culturalist’ analyses which only define groups through their cultural markers and shared identity. The relational notion of class is the starting point for most Marxist class analyses, but there are some problems with it. One pitfall comes from the determinist streak in classic Marxism in which capitalist development is seen as inevitable, and with it, also inevitable, is the emergence of a proletariat as the only class that could lead the revolutionary struggle, due to its ‘economic role … in large-scale production’ (Lenin n.d. [1917]). Such abstract models of predetermined capitalist development are still prevalent today (see, for example, Warren 1980; Sender 2016).

Criticisms of such predetermined trajectories of change (which have been characterised as ‘post facto determinism’ Thompson 1963; McNally 1993), and the argument that abstract theory must be historicised (as expressed in the book title Theory as History – Banaji 2010), go back to Marx himself. As shown by, among others, Kevin Anderson, Marx’s thinking was guided by concrete historical analysis, and in the later years of his life this overrode any deterministic focus on the ‘proletarian’ condition (Anderson 2010). Indeed, in the 1870s and early 1880s, he was exploring whether a revolution in Russia could be based on its existing agrarian communal forms, as opposed to having to await the development of a fully industrialised society (Shanin 1983; Anderson 2010). Such an approach did in fact form the basis of the strategy and practice of many revolutionary movements of the Global South, starting with the Chinese Communist Party in the 1920s. Their historically grounded analysis deviated from the industrialisation–proletarianisation model and instead de facto identified the poor peasantry as the main revolutionary class.

More recently, from the late 1990s onwards, Henry Bernstein has questioned this determinism from an agrarian political economy perspective. He argues that under neo-liberal globalisation and global finance capitalism, agrarian transition to capitalism is no longer a main engine for capitalist growth in the Global South (see, for example, Bernstein 2016). The national ‘agrarian question of capital’ has been bypassed in the sense that the development of capitalism in the non-agrarian economy does not require inputs from capitalist development in agriculture: global finance has taken over from agriculture as the main source of capital; markets for non-agrarian produce are now global and agrarian production and profits in many parts of the world are global too.³ This also means that the formation of class alliances involving major sections of the bourgeoisie around a goal of national capitalist development based on agrarian capitalism is no longer likely or possible. Furthermore, the relentless drive by capitalism towards labour-saving production processes, combined with the absence of ‘Fordist’ social compromises – as capital absolves itself from the need to create its own domestic market by turning labour into consumers – means that in the Global South the transition of the exploited agrarian classes into an industrial proletariat is now the exception, not the rule.

Today, in most parts of the Global South, the bulk of the working population are in informal, temporary jobs, or in petty commodity production in the agrarian and non-agrarian sectors, with their work changing throughout the year, throughout their lifetimes and across the household. As pointed out by Bernstein, the appropriation of labour and surplus value occurs both through direct labour exploitation in the production process and through exploitation of those petty commodity producers who depend indirectly on the sale of their labour power. Bernstein conceptualises these class relations as relations between capital and what he labels ‘classes of labour’. Classes of labour encompass ‘various and complex combinations of employment and self-employment’, i.e. all those who ‘have to pursue their reproduction through insecure and oppressive … wage employment and/or a range of likewise precarious small-scale and insecure, “informal sector” (“survival”) activity, including farming’ (Bernstein 2007, 6; see also Breman 2003, 13). As Bernstein indicates (2007, 9), this absence of a working-class-in-itself also challenges simplistic notions of transformations to class-for-itself consciousness and related class-based action by the classes of labour.

These steps towards abandoning a determinist class analysis, though, need to be further developed. In this paper we advance perspectives which suggest that the analysis of the development of capitalism and class must be explored in relation to social oppression based on race and ethnicity – in our case caste and tribe – as well as gender (and sexuality, etc.), and that such relations are, in fact, co-constituted.

Racism, ethnicity, gender and class: conjugated oppression

Marxist-feminist scholars have highlighted the need to analyse gender relations together with class, work and capitalism. In particular, they have pointed out that the totality of capital–labour relations includes the social reproduction of labour. Silvia Federici, for instance, has shown that, in Europe, the historic subordination of men under capital during early capitalism (i.e. primitive accumulation) was linked to and made possible by the simultaneous subordination of women both under capital and under men. Women, through a sustained orgy of violence (constructed as a witch hunt), were beaten back from positions in production to reproduction within the household (Federici 2004).⁴ Today, increasing numbers of social reproduction theory scholars analyse capital accumulation, class, gender, race and sexuality as co-constituted through the process of capital accumulation (Bhattacharya 2017). The analyses have moved from the reproduction of labour power within the working-class family (Vogel 1995) to, among others areas, analyses of migrant women performing care work in the most developed capitalist countries, which facilitates the work of local women from the developed countries in other sectors (see, for example, Farris 2015; Ferguson and McNally 2014).⁵

Literatures analysing the relationship between race and political economy go back to Marx, and have been developed further in the North American and European contexts (see Reed 2002; McNally 2015; Camfield 2016). Focussing on capitalism and ethnicity, Anderson detailed how Marx was concerned with the political implications of ethnic divides within the working class. For instance, Marx highlighted that a major obstacle to the progression of the class struggle in England was the fact that Irish workers occupied the lowest rungs of jobs in England and that there was no working-class solidarity across this ethnic divide (Anderson 2010).

Although Marx himself did not provide a theoretical underpinning for these concerns (Camfield 2016), methodologically he eschewed an essentialist approach. David McNally argues that for Marx, abstract essences of any phenomena could not stand alone. Specific forms of appearance and essential forms have to be co-theorised as ‘concrete universals’ (emphasis in original); analysing wage-labour as a class means understanding ‘all the concrete determinations and particulars that constitute living social groups’ such as social relations of race, gender and sexuality (McNally 2015, 131, 133, 135). But while race etc. may be a part of the ‘concrete universals’ of class relations, to characterise them simply as that is, in our view, also reductionist. David Camfield (2016) rightly argues that class relations and race (as well as relations based on gender and sexuality) are co-constitutive. In our view, this also means that, conversely, while class is part of gender, race, etc. relations, such relations cannot be reduced to class either.⁶

Borrowing from Philippe Bourgois (1988, 1989, 1995), we call the multiple forms of oppression simultaneously at work ‘conjugated oppression’. We do so because of his insight that ethnic discrimination and class exploitation ‘interact explosively’ to ‘produce an overwhelming experience of oppression that is more than the sum of the parts’ (Bourgois 1995, 72). Bourgois first used the concept in relation to two ethnic groups of workers on banana plantations spanning the borders of Costa Rica and Panama and the different kinds of oppression they faced, both at work and outside work (Bourgois 1988, 1989). While Bourgois distinguishes between ethnicity as an ideological relation and class as an economic relation, we expand the term ‘conjugated oppression’ to express how multiple axes of oppression based on social relations such as race/caste/tribe/ethnicity/region or gender and sexuality, etc. are co-constitutive of and shape class relations, potentially producing extreme relations of oppression, inseparable from each other in capitalist accumulation.⁷

Several authors have sought to historicise such a relationship between race, class and capitalism. In the late 1940s Oliver Cromwell Cox, in his Caste, class and race (1970 [1948], 331–43, 344) tracing European colonialism in the Americas after 1492, linked colonisation and capitalism to argue that race prejudice was a ruling class instrument justifying the exploitation of a group or its resources.⁸ Cox’s insights were developed in different ways by various authors to show that race and class were not fundamentally distinguishable – not dichotomous, but overlapping in a singular system of social power and stratification rooted in political economy.⁹

The relationship between capitalism, racism and ethnicity, and with that the racial fragmentation of workers, is, perhaps, nowhere better depicted than by Etienne Balibar (1991a, 1991b). Like Cox, Balibar argues that racism was brought into being together with capitalism and colonialism. ‘Class racism’ was created in Spain from the late fifteenth century onwards, as the ruling classes constructed themselves as a pure race, linking them to the ‘Spanish people’ and allowing them to conquer, commit genocide and enslave the people of their colonial empire (Balibar 1991a, 208). The Industrial Revolution and the expansion of capitalism gave rise to ‘the new racism of the bourgeois era which has as its target the proletariat in its dual status as the exploited  …  and politically threatening population’ (Balibar 1991a, 209). Drawing on Louis Chevalier (1981 [1973]), Balibar proposes that the ‘labouring classes’ were discursively divided into two: the ‘dangerous classes’, who were ‘an object of fear to society’ and therefore ‘dangerous’, and those who were made to be no longer dangerous. The privileges of the latter group might include citizenship, voting rights and so on, while the former were excluded from such positions and markers (Balibar 1991a, 208–11, 1991b). Balibar argues that in France, today, this divide runs between immigrant workers and those considered proper ‘French’ workers, and that this corresponds closely to hierarchies of work, with immigrants relegated to menial, unskilled work. Working-class racism – that is, racism between those groups who perceive themselves as linked to the dominant classes, and the ‘dangerous’ parts of the labouring classes – plays a major role in the workings of class racism (Balibar 1991b, 224–26).

Exclusionary practices linked to racial oppression by the dominant classes in conjunction with parts of the working class are widespread. Many studies of such practices are set in the US. Du Bois’ path-breaking study on the agency of black people in the American civil war and the reconstruction period that followed details how racism squashed the threat to capital that was posed by working-class action. Northern capital and southern propertied classes allowed the white working class and peasants in the south to benefit just enough for them to become active agents of a brutal racism, partaking in the dispossession and disenfranchising of the black population in an orgy of violence (Du Bois 1998 [1935], esp. 670–710). David Roediger (1991) likewise documents how in the nineteenth century, US capitalism created a racialised working class, with the skilled–unskilled division between workers reinforced by race, while racism flourished among unskilled workers as well, fuelled by the competition for work. This was underpinned by stigmatising black labour as an ‘inferior race’, as ‘ignorant’, ‘smelly’ and ‘black rats’ (178).

In line with these arguments, Camfield (2016, 56–63) suggests that racism relates to three dimensions of capitalism. The first is imperialism – that is, the ongoing globalised history of capitalism. The second is the profitability of racism (in Hall’s words, ‘the law of value  …  operates through and because of the culturally specific character of labour power’ Hall 1986, 24, emphasis in original). The third is the effort of dominant groups to preserve their advantages relative to the racially oppressed (that is, the racism within the working classes directed against the ‘dangerous classes’). Such axes of oppression of the labouring classes are central not only in shaping the struggles from below but also to the class struggle from above, waged by capital for its dominance and towards further accumulation.

In sum, we have argued that in most of the Global South capitalism has not led, and is not about to lead, to full-scale industrialisation or to a proletarian working class. Instead, groups of ‘more-or-less’ free labourers (Banaji 2003) have been constituted, often with one foot in agriculture and the other restricted to informalised and insecure work and precarious petty commodity production outside of agriculture. These labouring classes are constituted through the conjugated oppression of class, race, ethnicity – in India, caste and tribe – gender, etc. As Stuart Hall (1986, 2016) and others such as Anna Tsing (2009, 148) have argued, diversity in the form of gender, race, national status and other forms is structurally central to global capitalism, and not ‘decoration on a common core’. This is rarely dealt with head-on by class-based analyses, despite the fact that it has serious implications for class consciousness and struggles.

Conjugated oppression in India

The social relations of conjugated oppression will vary from context to context. For India we argue that the spread of capitalism has been marked by class relations that are mutually constituted with caste, tribe, gender and region-based oppression. We trace these processes by focussing on Dalits (the ex-untouchable castes) and Adivasis (tribal groups)¹⁰ who together constitute one-quarter of the Indian population, and to a lesser extent on gender and region.

Historically, Indian class relations were enmeshed with and inseparable from caste. Dalits were historically seen as India’s ‘untouchables’: an ‘impure’ and ‘filthy’ class of slave-like, landless agricultural labourers (see e.g. Kumar 1965; Breman 1974, 2007; Viswanath 2014), at the very bottom of a caste hierarchy based on religiously sanctioned ritual purity and pollution, below the ‘touchable’ or ‘clean’ castes (Dumont 1970; Habib 1995, 166). Their touch and even their shadows were seen as polluting; they were left to do only the hardest and most demeaning jobs and were treated as the higher castes saw fit. This extreme domination also included widespread sexual harassment of Dalit women, and abuse, boycotts, beatings and killings would enforce the relationship (Ambedkar 1989). Although there was a tribe–caste continuum, as pointed out by Frederick Bailey (1961), the Adivasis of the hills and forests, in comparison to the Dalits, lived in relatively independent or autonomous communities, with much more direct access to land and forest resources, and without the same domination by higher caste groups that Dalits faced on a day-to-day basis. But they were stereotyped as ‘wild’, ‘savage’ and ‘childlike’ (Skaria 1997). The closer their interaction with Hindu societies, the more their domination and exploitation might become more like that of the ‘Untouchables’, earning the label ‘tribal castes’ by some scholars (e.g. Breman 1974).

We argue that the social relations of oppression of Dalits and Adivasis continue to be pervasive, but that these relations have changed over time and have become integral and systemic to capitalism. We aim to show that this social oppression is an inseparable aspect of the constitution of exploitative capitalist class relations.¹¹ As part of this, tribe and caste also shape the specific ways in which the dominant Indian classes and castes ally with parts of the labouring classes, reinforcing conjugated oppression and defending the (for many, quite meagre) privileges and resources of these labouring groups from those below them in the social hierarchy, through the stigma they attach to them.¹²

The mutual constitution of class relations and other social relations of oppression in India is not much studied by Left scholars. Class-based analyses highlight important aspects of what unites and divides particular groups of labourers (e.g. Parry 2009) and how informalised labour struggles to survive under adverse conditions (e.g. Breman 2003).¹³ But Dalit scholars have advanced our analysis and pointed out the inseparability of Leninist-type class relations and other axes of oppression (Guru and Chakravarty 2005; Teltumbde 2010). Critiques and self-critiques of radical Marxist movements have made similar points (see Ghandy 2011; Ismail and Shah 2015; Srinivasulu 2017 on the Marxist Leninists/Maoists; and Lerche 1999 on the Communist Party of India [Marxist]). Such political critiques are supplemented by historical analyses (e.g. Habib 1963; Chandavarkar 1999) and the rare agrarian class analysis along such lines, such as that of Arvind Das (1984). There are also studies showing how capital segments and fragments labour through the manipulation of caste, gender and other social identities (Harriss-White and Gooptu 2000), and how contemporary processes of accumulation accentuate dynamics of caste and ethnicity, gender and class (e.g. Kapadia 1995; Da Corta and Venkateswarlu 1999; Whitehead 2016; Pattenden 2016). These are important contributions but they are exceptions to a general unease about exploring the mutual constitution of class and other axes of social oppression in relation to capitalism in India.

Such an exploration should include the well-established point that the historical forms of social oppression of Adivasis and Dalits have lost some of their power. With Independence, affirmative action for Dalits and Adivasis was introduced, and in the following decades extreme oppressive historical forms of exploitation, such as hereditary debt slavery, became much less common, and practices of untouchability became less extreme – although they most certainly do still exist (see, for example, Shah et al. 2006; Desai and Dubey 2011; Deshpande 2011). These changes were rooted in economic processes in agriculture and other sectors which loosened the historical, occupational, village-based ties. Linked to this, for Dalits, was the – albeit limited – widening of their formal political space, something Adivasis experienced less of (Guha 2007).

However, the transformative effects of economic development since Independence had severe limitations. There has been no classic agrarian transition: agriculture’s share in gross domestic product (GDP) and in employment has declined, but this has not been offset by industrial growth.¹⁴ Instead, people from rural areas mainly get work in the fast-growing construction sector and in services.¹⁵ The expansion of capitalism has also not led to growth in secure employment. The biggest divide within the labouring classes is between the eight percent of the workforce in formal-sector regular jobs and the rest – 92 percent – who are engaged in precarious work, either in the small-scale informal sector or in informalised employment for formal sector enterprises, or in informal agriculture.¹⁶ Most of these are workers trapped in low-wage jobs or vulnerable petty commodity production self-employment, with poor work conditions, no job security, no health insurance, no pensions, no ability to unionise and so on.¹⁷ They are forced to survive in households where some members work in far-away places and others keep hold of what few assets they have in the predominantly rural areas that are their home. Unsurprising, poverty levels are high.¹⁸

Adivasis and Dalits continue to find themselves at the bottom of India’s social and economic hierarchies. There has only been minimal social differentiation among them, mainly related to affirmative action programmes. With 82 percent of Adivasis and Dalits below the international poverty line in 2009–2010 (Kannan 2012), Dalits (and Adivasis even more so) are still worse off than all other groups across India.¹⁹ Workwise, it is not just that the formal/informal divide cuts across the labouring classes; the divide follows the historical lines of social oppression of castes and tribes: the Adivasis and Dalits are overwhelmingly working in the informal economy while formal-sector regular employment is dominated by the higher castes.²⁰

We argue that social relations of oppression of Dalits and Adivasis are pervasive in the way capitalism has expanded across the country. We propose that these social relations of oppression in the expansion of capitalism in India are entrenched through at least three interrelated processes.²¹ The first is the historical inherited inequalities of power, which enable dominant groups and the state to control the adverse incorporation of Adivasis and Dalits in the capitalist economy. The second is the super-exploitation of casual migrant labour, where local labour power is undercut by a more vulnerable workforce, enabling capital to fragment the overall labour force and therefore better control and cheapen it. Third, conjugated oppression of class relations and multiple oppressions based on caste, tribe, class, gender and region is a constitutive part of these processes and includes all-India stigmatisation and other oppressive relations furthering the division of the working classes, also between and within Adivasi and Dalit groups (Shah & Lerche et al. 2018).

These propositions are based on our collaborative research into the inequalities and poverty of Adivasis and Dalits in the context of economic growth in India. As part of this research, five scholars each undertook one year of ethnographic field research, in separate sites. In each site, a survey of occupations, assets and education of all Dalits and Adivasis and a selection of other classes and castes was undertaken, together with recording of genealogies. The sites were: the tea belts of the Western Ghats in Kerala (where Jayaseelan Raj worked); the chemical industrial estate in Cuddalore district of Tamil Nadu (Brendan Donegan’s field site); the Bhadrachalam Scheduled Area in Telangana (where Dalel Benbabaali conducted research); the Chamba Valley in the mountains of Himachal Pradesh (the field site of Richard Axelby), and the Narmada Valley and adjoining plains in Maharashtra (the research sites of Vikramaditya Thakur).²²

Inherited inequalities of power

The historical powerlessness of Dalits and Adivasis in relation to dominant social groups and institutions has continued to mark their livelihoods, even as they have moved beyond the bounds of the village and into the wider informal economy. Not only do historical inequalities legitimise that Adivasis and Dalits have less land, less capital, less education and less political clout than other social groups; they also sanction that they are consigned to low-end work. Across all our research sites, the global processes of the expansion of capitalism work through locally dominant caste groups. Many of the upper caste and dominant-caste Hindus, who used to command hegemonic power at the village level through caste-based hierarchical land, labour, social and political relations, have transformed themselves into powerful players in the new non-agricultural economy.

For instance, in the expanding industrial belt of Tamil Nadu in South India, an export-oriented gelatine factory is managed by caste brethren of the local dominant Nadar caste landlords. Here, the Dalits find their access to work in the factory controlled by the very landlords they used to work for as bonded labourers (see the chapter by Donegan in Shah et al. 2018 our book). In addition, they are relegated to the worst jobs in the most polluted circumstances, as the dominant castes have monopolised the managerial jobs while middle castes have been able to monopolise semiskilled and skilled work. Similarly, in a Telangana village, it is also the old landlord (of the dominant Kamma caste) who mediates access for Dalits and Adivasis to informalised jobs and patronage in the nearby large paper factory – while, at the same time, doing his best to continue the everyday dominance of the lives of the Dalits, Adivasis and also other villagers, and of the local political scene (see chapter by Benbabaali in Shah et al. 2018 our book).

In some of our sites the historical disadvantage and powerlessness is extreme. This is the case for Irulas in Tamil Nadu, who had little recourse as government officials persistently refuse to certify that they are Adivasis. This means that they have no access to affirmative action programmes (see Donegan’s chapter in Shah et al. 2018 our book). It is also the case for the Bhil Adivasi petty farmers in northern Maharashtra who were forced to watch their papaya crop rot in the fields as traders, unpunished, reneged on purchasing deals (see Thakur’s chapter in Shah et al. 2018 our book).²³ Elsewhere the persistent powerlessness is more mundane. For example, in Kerala, the Tamil Dalit plantation workers could only watch as, from the 1990s onwards, the plantation owners broke the Plantation Labour Act and ceased providing a modicum of welfare measures and social security benefits while the state government – of the most progressive state in India – turned a blind eye (see Raj’s chapter in Shah et al. 2018 our book). At the other end of the country, in the Saal Valley in the Himalayas, the Gaddi and Gujjar Adivasis are the only social groups who still cannot access any kind of decent work in government or in the private sector. Historically they have lived at the social and geographical extremes of society, and now, all that is left for them is to branch out into road construction work and local petty business, or to migrate to rear cattle in Punjab (see Axelby’s chapter in Shah et al. 2018 our book).

The processes we are describing here are more than simple outcomes of differences in initial material conditions such as access to land or wealth. The old exploiters have become part of the new capitalist classes and through this have reinvigorated their power to dominate, divide and rule. More than any other groups, Adivasis and Dalits are at the receiving end of this, and this has had a direct impact on their position in the hierarchies of work and income.

The super-exploitation of migrant workers

Exploitation and oppression of international migrant workers in Europe, North America, etc. is a central part of global labour exploitation. Balibar argues that migrants are racially constructed as the ‘dangerous classes’, bereft of the rights and positions of the rest of the working class, and others have focussed on how gender and social reproduction as well as race are central to their oppression and exploitation (Farris 2015; Ferguson and McNally 2014). In contrast to this, the oppression of migrant labour in India is not a product of external immigration but ‘internal alien-ness’ (see also Roberts 2016) based on low-caste, tribal and region-related status. There are millions of circular migrant labourers, or ‘wage hunters and gatherers’ (Breman 1996, 222), who hold onto whatever little land or housing they may have in rural areas while migrating seasonally, transcending both the agricultural–industrial and the rural–urban divides. They are paid less than local low-caste workers and render the overall labour force resolutely insecure and super-exploitable.

Claude Meillassoux, in the context of southern Africa, pointed out the division between local workers reproducing themselves entirely within the capitalist industries, and a rural circular migrant labour force which only partly reproduces itself in this way (Meillassoux 1981 [1975], 120–25). We use the term ‘super-exploitation’ to describe how capital cheapens labour from the countryside to far below the cost of such local workers, through lower pay and worse terms and conditions of work, as well as through relying on the cost of their reproduction to be dealt with back at home. Their wages are below what is required for their social reproduction, and in many sectors this is becoming the ‘new normal’.²⁴ This is reinforced by discrimination against the migrant labour, including racism, which makes it difficult for them to put forward demands, demonstrate publicly, or ally with local workers (Meillassoux 1981 [1975], 120).

The Indian government estimates that there are about 140 million internal seasonal labour migrants (Government of India 2017 and see e.g. UNESCO 2015 for earlier estimates). But seasonal migrants are parts of households that tend to have multiple means of livelihoods, and household members often take turns to migrate. That means that several hundred million people – in fact, most rural households – are directly affected by seasonal casual labour migration.

Adivasis and Dalits are overwhelmingly represented among the circular migrant workers (see also Breman 1996; Mosse et al. 2002; Rogaly et al. 2002) across nearly all sectors, and they dominate numerically among migrant workers in the low-skilled, back-breaking brick kiln and construction sectors (Ministry of Housing and Urban Poverty Alleviation 2017, 14–17).²⁵ This was also shown to be the case in our research. For example, the Adivasis of Maharashtra migrate seasonally mainly to brick kilns and for sugarcane harvesting, while at the Kerala plantations the young Dalits migrate to the towns for work as informalised contract labour, as do some from the Irula Adivasi households in Tamil Nadu. The Adivasis of the Himalayas combine this kind of migration with less distant, short-term migrant work.

Region also overlaps with caste/tribe to divide the workforce. Typically, the most exploited seasonal migrant labour comes from central and eastern India (Jharkhand, parts of Bihar, Chhattisgarh, Odisha, parts of Madhya Pradesh and parts of Andhra Pradesh and Telangana), and they are predominantly Adivasis.²⁶ Our research shows that as seasonal migrant labourers they are treated even worse than the local Dalits are in the rest of India. This is the case in the tea estates and in the gelatine factory in South India, where the Adivasi and Odisha in-migrants are the most exploited workers. This is helped by their intense isolation caused, at least in part, by not speaking the local language. Similarly, in the Himalayas, local labour contractors consider the seasonal labour migrants from Jharkhand, Bihar and elsewhere to be more malleable than local labourers.

The vulnerability of migrant labour is also central to keeping labour in general cheap and docile in India. The Odisha seasonal migrant labourers at the gelatine factory were originally brought in to crush a strike in which the local Dalits took part, while in the Kerala plantations a strike by the local Dalit labourers was weakened by the presence of Adivasi workers from Jharkhand who continued to work during the strike.

Adivasis are worse off than Dalits on most indicators but, importantly, they are much more likely than Dalits to have access to some land and forests which gives them some autonomy over their lives, especially in the central and eastern regions which have historically been less developed. But this autonomy is under threat (see Shah forthcoming) as land and forest alienation is rife, not least due to encroachment by dams (as in the case of the Narmada dam), factories (such as the Telangana paper factory), and large-scale coal and mineral mining by major Indian and transnational mining conglomerates. This is likely to continue as an engine of labour migration.

The new, seasonal migrant labour relations involve new forms of oppression and exploitation. While some labour contractors are little more than foremen arising from the midst of migrant labourers, the long-distance migration is often organised by chains of labour contractors and many labourers at the bottom of the labour hierarchy are tied to their contractor through advance payments. They may also be charged exorbitant rates for ‘services’ provided, often for travel costs or accommodation as is the case for the gelatine factory migrant labourers.

The seasonal migrant workers are also prone to oppression outside the production process. We have met them housed in sheds and tents, or in crowded rooms where they had to sleep in shifts, or simply in smaller spaces than local workers. They often do not speak the local language and rely on their labour contractor to negotiate and bargain for them. Unlike the local workers, they are excluded from the Public Distribution System which provides subsidised rations of rice, sugar and kerosene, and they have no access to local government services such as schools.²⁷ Like immigrants in many other countries of the world, they are not seen as citizens in the Indian states where they work. In that sense there are clear parallels to the situation in China where circular migrants working in the big cities do not have the same rights as the local workers with urban hokou.

Conjugated oppression

Conjugated oppression of class, caste and tribe started long before the spread of capitalism.²⁸ As mentioned earlier, caste-based oppression and stigmatisation, especially of ‘untouchable’ castes, was part and parcel of maintaining a large rural proletariat of agrestic slaves in pre-colonial times. For the Indian tribes the issues are different, as their forest autonomy enabled them to be somewhat outside the Hindu caste system – except, of course, when their areas were conquered or encroached upon by outsiders (Shah forthcoming). But they are still discursively constructed as part of the ‘dangerous classes’, as wild, savage and barbaric.

Today, Dalits and Adivasis continue to be excluded from certain jobs by explicit discrimination and stigmatisation. Our research documents several clear-cut instances. The textile and garment sector factories in Tamil Nadu would not hire semi-skilled Dalits and there were cases of entire communities of Dalits hiding their name and caste background for years, including from their non-Dalit co-workers, in order to get work (see chapter by Raj in Shah et al. 2018 our book).²⁹ For others, for example a Dalit plumber in a Kerala town, everyday discriminatory and stigmatising behaviour of work colleagues and bosses made it impossible for him to stay in the job – had he stayed on, it would have been at the cost of his sanity and self-respect.³⁰ In the Kerala tea plantations, the Dalits are stigmatised not only as Dalits but also because of their history as enslaved bonded labour, because they were Tamil speakers in Malayali Kerala, and because they were from the ‘wild’ highlands rather than from the settled and ‘civilised’ lowland valley.

At the gelatine factory in Tamil Nadu only Dalits, Adivasis and other migrant labourers handle the animal bones in the smelly, unhygienic heart of the factory (see Donegan’s chapter in Shah et al. 2018 our book). In Telangana stigma is used to divide and rule as the Kamma landlord sees the Adivasi Koyas as ‘rebellious’ whereas the ‘obedient’ Adivasi Lambadas get preferential employment in his fields and as informal labour in the paper factory (see chapter by Benbabaali in Shah et al. 2018 our book). In Maharashtra, Gujar landowners use abusive language and taunt their Adivasi Bhil workers (see chapter by Thakur in our book); while in Chamba government officials and wealthy urban upper castes and Muslims view the Adivasi Gaddis as ignorant and lazy, and the Adivasi (Muslim) Gujjars as dishonest, dirty, and ill educated (see chapter by Axelby in our book).

While Adivasis and Dalits (as well as Muslims) are overwhelmingly informalised, so are major sections of other relatively low-ranking castes. The process of conjugated oppression maintains and strengthens the divides between these groups. The non-Dalit/Adivasi castes may kick downwards in relation to excluded and stigmatised Dalit workers (see Raj’s chapter in our book). Similarly, Donegan reports from Tamil Nadu that the Vanniyar middle-ranking caste were allowed to monopolise semi-skilled work at the bone factory by the higher caste factory management, keeping the Dalits and the seasonal migrants confined to the more hazardous and poorly paid factory floor work. In the same locality, Adivasi Irulas had their fishing nets destroyed and the engines of their new fishing dinghies stolen, probably by people from the Natter fishing caste who, as could be expected, got away scot free. Underwritten by the broad acceptance of those classes and castes which are not at the bottom of the hierarchy, this conjugated oppression is essential to the social relations of inequality on the ground.

Conjugated oppression works through gender relations. Across India, women bear the greater burden of the social reproduction of the household. This is never considered ‘work’ nor is it remunerated, even though it enables men to work for wages and employers to have workers. Sylvia Federici, among others, has shown that the rise of capitalism was interlinked with the imposition of a ‘patriarchy of the wage’ (2004, 97–100) – a war on women turning them into housewives. Within the ‘classic patriarchy’ of South Asia (Kandiyoti 1988), women of higher castes tend to be confined to work within the household, whereas at the bottom of the social and economic hierarchy both men and women do paid work outside the house. Even though women do not get the most remunerative of work and are usually paid less than men for the same work, Adivasi and Dalit women’s paid work usually enables their greater autonomy within the household and hence more egalitarian gender relations than otherwise. Upward mobility among Dalits and Adivasis may have intra-household costs, as women can be confined to domestic work, leading to less-equal gender relations (see Kapadia 1995; Still 2014 for Dalits, and Higham and Shah 2013 for Adivasis). But there is no overall trend of upward mobility, nor of Adivasi and Dalit women withdrawing from the paid workforce in the sites of our work. In fact, in some of our sites, such as the Kerala tea plantations, neoliberal cutbacks have meant a feminisation of the workforce, with predominantly women doing the paid work in the plantations and thereby also maintaining their right to a plantation worker’s home, which also means that the Dalit men taking up migrant work have a house to return to (Shah & Lerche et al. 2018).

The flipside of this is that working outside the household involves conjugated oppression of Dalit and Adivasi women by higher caste employers, including sexual exploitation as, for example, in Thakur’s site where Bhil Adivasi women report that this happens, particularly during work migrations. Dalit and Adivasi women are potentially oppressed both by the men of their households and by their employer, while the men, to differing degrees, rely on women taking care of the reproductive sphere. That said, gender relations among Dalits are less extreme than among high castes. It is also the case that Adivasi gender relations tend to be more egalitarian than those among Dalits, a comparison that needs to be studied in greater detail (Shah forthcoming).

Conjugated oppression includes other processes beyond the workplace and intra-household relations. Although the ideology of caste based on purity and pollution has been diluted since Independence, Adivasis and Dalits continue to be ‘othered’. They can be treated as the non-humans of India with no rights, who can be discriminated against, and against whom atrocities can be committed with near impunity. Adivasis in central and eastern India have been stigmatised to an extreme degree as a ‘dangerous class’, supposedly all actively supporting the Naxalite revolutionary struggle. Labelled terrorist supporters, they have been subjected to brutal vigilante and police action. Whole villages have been burnt to the ground and their populations displaced and women raped, and thousands have been arrested, tortured and killed under the guise of the ‘civilising’ mission of development (Shah 2011, 2018; Sundar 2016). In recent years, Dalits and their organic intellectuals across the country have been increasingly targeted as ‘anti-nationals’ (that is, disloyal to the nation) with what this entails in terms of threats of prison or of extra-juridical killings. At the same time, extreme religious motifs are revived to entice mobs to target Muslims, Dalits or Adivasis under the pretext of punishing beef eating, while everyday violence in the most brutal forms persists, including murder and rape. The dominant classes in society still view the lives, lifestyles and customs of Adivasis and Dalits (and Muslims) as dirty or uncivilised, decry emancipatory politics from their midst as anti-national, and consider them in need of civilisatory education before they can join the ‘nation’. What we have as a result of the construction of Adivasis and Dalits as ‘dangerous classes’, based on their conjugated oppression, is a super-exploitable workforce, the control of which is enforced and supported by an oppressive ‘civilising’ mission that is increasingly being meted out by the police and other state forces in collusion with corporate capital.

This potent cocktail of oppression and exploitation of Adivasis and Dalits in India cannot be captured by analyses focusing solely on degrees of classic untouchability or other pre-capitalist ideological structures. It also cannot be explained by mainstream theories of discrimination, as they do not focus on the relationship between capital accumulation, class and social oppression. Here we have argued for an analytical approach to class relations that eschews assumptions of full-scale proletarianisation, and that understands conjugated oppression along the lines of tribe, caste, region and gender as part of the class struggle by capital and by the state. We have sketched out how this involves analysing the ways in which Indian capitalism works through inherited inequalities of power, how it makes use of super-exploited domestic seasonal migrant labour, and how this is underpinned by relations of conjugated oppression which not only cheapen the labour force but also divide it politically. This kind of grounded class analysis provides, we think, a strong basis from which to think about implications in relation to struggles and class-based political strategies, and this is what we turn to now.

Struggles and strategies

E.P. Thompson (1963) argued that class consciousness is the way in which class experiences (determined by the productive relations) are ‘handled in cultural terms: embedded in traditions, value-systems, ideas, and institutional forms’ (9–10). Class struggle, for him, is people’s struggles around antagonistic points of interest when they find themselves in structurally different positions in relation to their experience of exploitation, and – crucially – may be expressed in a range of different idioms (including, as we have shown in the Indian case, caste or tribe/indigeneity). Class struggle takes many forms other than struggles by ‘the proletariat’ for-itself, and a class analysis which puts conjugated oppression at its centre, as we have argued for here, will help us in understanding various struggles and their potentials and pitfalls. In a similar vein, Jeffery Webber (2011) has called the interweaving of politics/class struggle and indigenous liberation ‘combined oppositional consciousness’, to capture the interpenetration of class and indigenous identity and how working-class identity and organising was constituted in and through rich traditions of indigenous resistance.³¹ As David McNally puts it, ‘racialized workers, women workers, queer workers, disabled workers must all be able to determine, in and through their own activities and aspirations, the very meanings of working-class identity and struggle’(2015, 144).

Centring an analysis of conjugated oppression means embracing many of the issues focused on by agrarian populists, subalterns and post-structuralists as co-constitutive of class politics and class mobilisation. The point, then, is to analyse relations of social oppression (race, caste, ethnicity for example) from a perspective that does not reify them into abstract essences but acknowledges them as central to class struggle. Ethnic movements, caste-based movements, peasant movements, indigenous movements, feminist movements, etc. can thus be important drivers of progressive social transformation, not as identity-based movements per se, but as co-constituted by and through class relations. Class-based political movements have often ignored these other axes of oppression, as is evident from, for example, the tensions that have long existed between working-class and indigenous liberation movements in Latin America (Becker 2008), or Maoist movements and indigenous movements in South Asia (Ismail and Shah 2015) or from the critiques of the Cuban revolution for failing to take into account racial inequality within Cuba (Dominguez 2013).

In the preceding sections we showed how, in India, class struggle from below is not being led by the proletariat understood as a class fully reproduced inside a capital–labour relation. The proletariat in India is small, it is not expanding, and the main sections of it are relatively privileged when compared to the majority who are informal workers (Parry 2013) with many being seasonal migrants. Instead, the starting point of an analysis of class struggle from below should be the positions and strategic interests of the mainly informalised labourers transcending the rural–urban and the agrarian–non-agrarian divides and how they can overcome their conjugated oppression along lines of class, caste, tribe and gender.

Given the social divisions amongst the workforce, it is unsurprising that they have undertaken little successful united action. The divisions created through inequalities of power, seasonal migration and conjugated oppression have not been overcome by left parties or by mass organisations. Unions and left parties in India rarely raise any specific concerns of Adivasis or Dalits; instead they focus on the universal idea of the proletariat. In practice this has meant that they address the concerns of labour in secure, formal-sector employment, usually leaving most informalised workers high and dry (see also Parry 2009).³² Unions may even be acting against the interest of informalised workers, especially if they are from low castes or tribes. In the Kerala tea plantations Tamil Dalit women tea workers went on strike not only for labour rights and against corrupt trade unions, but also against the discrimination, infantilisation and vilification of Dalits, Tamils and Dalit women in particular (see chapter by Raj in Shah et al. 2018 our book). Seasonal migrant workers are also rarely considered, either in the fight for welfare or the fight for better terms and conditions from employers. This is as true of the left unions as it is of the Naxalites (Navlakha 2010; Shah 2013).

Above all, unions and left parties ignore the fact that major sections of the labouring classes take part in oppressing Dalits and Adivasis, women and other ‘dangerous classes’. Our analysis bears out the reality that overcoming divisions created by conjugated oppression is as important an issue in India today as it was when Marx first identified ethnic divisions within the working class in England as being the major stumbling block for progress of the class struggle. But there is very little understanding of this in analyses which have focused on class relations in India, and the left takes even less action in seeking to deal with it. Strategies based on the idea of a naturally emerging united industrial proletariat which will lead the class struggle seem very hard to let go of.

On the other hand, it is well documented that Adivasis and Dalits have been involved in social struggles throughout history. Such struggles along other axes of oppression than straightforward class lines may be seen as ‘self-racialisation’ of the working class (Balibar 1991a, 214). But they will also often be struggles against their conjugated oppression. Our research revealed several such instances. Decades of struggle against the Sardar Sarovar Project dam shapes the backdrop to the lives of the Adivasi Bhils studied by Thakur. Attempts to stop expropriation of Adivasi land is a major issue in central and eastern India, the main sourcing ground of seasonal migrant labour. In the Tamil Nadu village studies by Donegan, images of figureheads of Dravidian, Dalit and class-based struggles for liberation line the walls of the Dalit-run voluntary educational centre for Dalit youth. And in Telangana, Benbabaali reports on a mainly low-caste and Adivasi protest alliance that succeeded in extracting signed promises from the paper factory to provide drinking water to the village and build roads for the Dalit and Adivasi colonies.³³

The form of class struggle also differs between Adivasis and Dalits. For Adivasis, the struggle to maintain historic land and forest rights is as important as the struggle for labour rights, especially under the new regimes of dispossession to open up for mining and industry and special economic zones, accompanied by land acquisition acts. These rights give Adivasi seasonal migrant workers a small degree of autonomy over their lives away from their exploitation and oppression.³⁴ Dalits, on the other hand, have been without significant land rights for generations, and while land reform demands are still relevant to them, our research indicates that young Dalits focus much more on finding ways to access informalised work in the non-agricultural sectors.

Our analysis suggests that not only are such struggles against conjugated oppression progressive, but also that a central part of a class-based political strategy which acknowledges the importance of conjugated oppression would need to consider how to unite and work with these struggles. This is of course not easy, given the different manifestations of the oppression and exploitation of Dalits and Adivasis, gender-based oppression and differences based on region. The complicity of other labourers in conjugated oppression would also need to be tackled head-on. All this may seem an uphill struggle – but the recognition alone that the ongoing struggles by Dalits and Adivasis are progressive is also an acknowledgment of the fact that people are more active in resisting the oppressive and exploitative moves of capital than can be gleaned from a focus simply on labour relations in the classic workplaces.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Jens Lerche http://orcid.org/0000-0003-3817-7058

Jens Lerche is a reader in labour and agrarian studies in the Department of Development Studies at SOAS, University of London. His research interests include class and caste relations in agrarian transformation; and the political economy of labour relations, unfree labour and rural labour migration. He is the editor of the Journal of Agrarian Change. His publications include Ground down by growth: tribe, caste, class and inequality in 21st century India (co-authored, 2018) and ‘The agrarian question in neoliberal India: agrarian transition bypassed?’ (Journal of Agrarian Change, 2013). Email: [email protected]

Alpa Shah is an associate professor (reader) in anthropology at the London School of Economics. She is the author of Nightmarch: among the guerrillas of India’s revolutionary movement, 2018 and In the shadows of the state: indigenous politics, environmentalism and insurgency in India, 2010; and co-author of Ground down by growth: tribe, caste, class and inequality in 21st century India, 2018. She has co-edited several volumes, on agrarian change, emancipatory politics, Maoists in South Asia, affirmative action, the Indian boom, and Adivasi and Dalit political pathways. She serves on various editorial boards including those of the Journal of Agrarian Change, Dialectical Anthropology and Focaal. Email: [email protected]

Additional information

Funding

This work was supported by Economic and Social Research Council: [Grant Number ES/K002341/1]; European Research Council: [Grant Number ERC-2012-StG_20111124].

Notes

1 We restrict our discussion and analysis to relations of race, caste, tribe, region and gender, although relations based on sexuality, disability, etc. are also of importance.

2 The contribution is based on the work of the Programme of Research on Inequality and Poverty in the Department of Anthropology, London School of Economics, led by Alpa Shah and Jens Lerche. The research programme involved, among others, five postdoctoral researchers (Richard Axelby, Dalel Benbabaali, Brendan Donegan, Jayaseelan Raj and Vikramaditya Thakur). The paper draws extensively on our co-authored book Shah, Lerche, Axelby, Benbabaali, Donegan, Raj and Thakur (2018) Ground Down By Growth. The research was funded by an EU European Research Council (ERC) Starting Grant and a UK Economic and Social Research Council (ESRC) grant awarded to Alpa Shah as Principal Investigator (ERC-2012-StG_20111124 ‘The Underbelly of the Indian Boom: Adivasis and Dalits’ and ESRC ES/K002341/1 ‘An Ethnographic Investigation into the Persistence of Poverty among Adivasis and Dalits in India’).

3 See Terence Byres (1996) for the paradigmatic, opposing view.

4 Jane Humphries’ work (1990) on gender and the proletarianisation of families during the English enclosures in the late eighteenth and early nineteenth centuries is another important contribution to the study of gender, class and accumulation in early capitalism.

5 For an overview, see Ferguson, Hennessy, and Nagel (2016).

6 While we agree with Camfield’s definition of oppression as ‘systemic harm, arising from social practices’ (2016, 46), we differ when he takes the standard view that oppression is constituted in relations of race, gender, sexuality, etc. but never grounded in just class. Oppression is certainly different from ‘exploitation’ in that the latter specifically involves the extraction of surplus labour and surplus value. However, exploitation often requires direct oppression (for example, the beating of a worker by an employer simply because he is a worker) that cannot be reduced to or understood by a focus on other co-constituted relations of social oppression. To acknowledge the importance of multiple forms of oppression should not lead to ignoring class-based oppression (in addition to exploitation) as a central element of capitalism, co-constituted with oppression based on race/caste/tribe/ethnicity/region/gender and sexuality. This is not clear from McNally either.

7 This perspective differs from that of ‘intersectionality’; a label covering a broad strand of studies concerned with how injustice and social inequality occur on a multidimensional level involving several social identities that ‘interact with each other’ (such as gender, race, social class, ethnicity; Crenshaw 1989 is the classic study). Such studies usefully put multiple identities on the agenda but do not focus on the inextricability of the ways class relations and gender/race/ethnicity, etc. shape each other as they tend to treat these categories as independent ‘variables’ that may or may not ‘interact’ or ‘intersect’ or ‘correlate’. Moreover, class is treated on a par with the other social categories rather than as central in constituting those social relations, and the link to how the co-constitution of class and other social relations are shaped by processes of capitalist accumulation is most often weak.

8 Aspects of Cox’s work were and are disputed, including the extent to which hierarchical racial differences were essential or cultural or material, and his treatment of ‘caste’. For a recent discussion of the usage of ‘caste’ in the American race-related literature, and of US–South Asian comparisons, see Chris Fuller (2011).

9 See the reviews by Reed (2002) and Camfield (2016).

10 There are differences between Dalits and Adivasis and between different castes and tribes included in these categories. We will return to some of these differences below, but will refer to Adivasis and Dalits jointly whenever possible. In 2011, Dalits made up 16.6 percent of the population (201 million people) and Adivasis 8.6 percent (104 million). See Charsley (1996) for a history of the use of the ‘Dalit’ category, and Shah (2010) for the ‘Adivasi’ category.

11 This differs from mainstream positions that see discrimination of caste and tribe as, for example, a primordial relation; as relating to market imperfections in the capitalist economy; or simply as elite capture. See Deshpande (2011) and Desai and Dubey (2011) for discussions of these positions.

12 See Economic & Political Weekly (2016) for a brief outline with several empirical similarities.

13 Many studies of class in India remain Weberian or culturalist and do not view labour exploitation and surplus appropriation as central to class relations, and some omit analyses of major sections of labourers who are seasonal migrants (e.g. Agarwala 2013).

14 The move away from agriculture to a non-farming economy is well documented; see for example the National Sample Survey data in Ranganathan (n.d.) and the collection of village re-studies in Himanshu, Jha, and Rodgers (2016). The sector-wide employment distribution in 2011–2012 was: agriculture 49 percent; industry 14 percent; construction 11 percent and services 27 percent. The share of employment within industry has been near stagnant for the last 30 years (1983 to 2011–2012: 11.5 percent to 13.7 percent) while employment in agriculture has declined by 20 percentage points (Institute for Human Development 2014, 208). For a review of overall trends, see Shah and Harriss-White (2011) and Lerche (2013).

15 From 2000 to 2012 employment grew at 2.2 percent per year. This small growth came especially from the construction sector, which grew in terms of jobs by 17 percent per year from 1999 to 2012 (Varma 2014). From 2013–2014 to 2015–2016 overall employment even declined slightly (Abraham 2017).

16 In the non-agricultural sectors, 86 percent of the workforce is informal (Mehrotra et al. 2014, 52).

17 In the non-agricultural economy, 38 percent of the working population were self-employed, i.e. working in household enterprises, in 2011/2012 (Institute of Human Development 2016, 126). This includes industrial outworkers/homeworkers (see Srivastava 2012, 80) and also disguised wage labour.

18 In 2009–10, 69 percent of the population was below the international poverty line of USD2 PPP (purchasing power parity) a day (Kannan 2012).

19 In many parts of India, the position of Muslims is not much better than that of Adivasis and Dalits. An analysis of this is outside the scope of this contribution but factual information about the position of Muslims is included where relevant.

20 Compared to other groups, the proportion of Adivasis and Dalits who are self-employed outside agriculture is also significantly smaller, and these self-employed Dalits and Adivasis are also significantly poorer than the other groups (Institute for Human Development 2014, 79–80, 83).

21 We draw extensively on the arguments presented in the first and last chapter of our book (see ‘Tribe, caste and class – new mechanisms of exploitation and oppression’ and ‘The struggles ahead’ in Shah & Lerche et al. 2018).

22 See also the respective chapters of our book (Shah & Lerche et al. 2018).

23 The dominant Gujar farmers, on the other hand, got the government to intervene, something which incidentally also benefitted the Bhil farmers.

24 Ben Selwyn argues along similar lines that huge swathes of low-paid workers in export industries in the Global South are super-exploited (2017, 50–72), and Balibar likewise characterises the ‘dangerous classes’ as super-exploited. It should be noted that John Smith uses the term super-exploitation differently and not related to the cost of social reproduction, as to him (following Samir Amin) it denotes that workers in the Global South are paid below the globally determined value of their labour power (Smith 2016, 205).

25 The figures are for male migrants in non-primary sectors only.

26 The adverse conditions of groups of Adivasi workers have been documented by some studies (e.g. Mosse, Gupta, and Shah 2005; Roesch, Venkatasubramaniam, and Guerin 2009) but comparative studies of Adivasis and Dalits are rare.

27 While local construction workers have won rights to Welfare Board Schemes, seasonal migrants are overwhelmingly excluded from these (Srivastava and Jha 2016).

28 Contrary to the dominant theories of racism and capitalism, we view conjugated oppression as potentially being part of the fabric not only of capitalism but of all class-based societies.

29 Carswell and De Neve (2014, 118) report similar instances in Tiruppur.

30 The research on labour-market caste discrimination is growing. This includes, for example, studies of discrimination against Dalits and Muslims in job selection (Thorat and Newman 2010), wage differentials (Madheswaran and Attewell 2007), exclusion of Dalits from wage employment in agriculture and the existence of discriminatory pay rates (Shah et al. 2006).

31 This is in the context of the Gas Wars of El Alto in Bolivia.

32 An exception to this is the new campaigning organisations such as the New Trade Union Initiative (Lerche 2010).

33 We did not come across struggles against gendered work relations, or inter-household gender relations. Overall, such struggles are rare – or, at least, rarely reported – although of course they do exist (see, for example, Anandhi 2017; Kagal 2017; Kapadia 2017).

34 Prathama Banerjee (2016) argues that Dalits articulate themselves as political subjects through representation, whereas Adivasis have done so through autonomy.