The political origins of tribal identity

Before the arrival of British, there was no term equivalent to “tribe” in the South Asian lexicon (Pathy Reference Pathy, Sharma and Oommen2000). It was used by the colonial administration to categorize numerous disparate communities that did not fit into their Brahmin-informed understanding of caste-based South Asian society. Thus, as a result of the colonial encounter, communities that were previously considered to be nations, kingdoms or peoples came to be referred to as tribes (ibid.). The process of designating or scheduling tribes acquired a systematic character in the 1931 census (Béteille Reference Béteille1986). Following the Government of India Act (1935) – which established a federal structure, introduced direct elections, and granted Indian Provinces greater self-rule – a list or “schedule” of depressed castes and tribes for whom parliamentary seats were to be reserved was created. The Constitution of India (1950) extended affirmative action to higher education and public sector jobs.

In contemporary India, Scheduled Tribes are recognized as a distinct group and are routinely enumerated in national surveys and censuses. Nevertheless, a definition of what constitutes a tribe has been “neither clearly formulated nor systematically applied” (Xaxa Reference Xaxa1999, p. 3589). Over 84 million people, belonging to 698 separate communities, are identified as Scheduled Tribes (GoI 2001). They encompass a wide range of communities that vary enormously in terms of population size, geographic spread, mode of livelihood, social organization, language and customs (GoI 2006). Traditional theories of postcolonial violence tend to define factional identities – and explain armed conflict – in terms of class or ethnicity. A number of influential scholars, perturbed by the fact that Adivasis cannot be reduced to economic or cultural identities, have criticized the conceptual utility of the term precisely because it is an administrative construct that includes communities with a variety of cultural and socioeconomic characteristics (Cohn Reference Cohn1990; Appadurai Reference Appadurai, Breckenridge and van der Veer1993). Nevertheless, the concept of Scheduled Tribe is what Mamdani (Reference Mamdani2002) refers to as a “political identity” – a legally enforced group that can only be understood with reference to state formation during colonialism and its historical legacy (also see Scott Reference Scott1998). But, while the term was originally imposed from above, it has been appropriated by the communities themselves and forms the basis of political claims made at the national and international level (Xaxa Reference Xaxa1999). In Murray Li’s (2000) formulation, Adivasis claim the tribal or indigenous “slot” (also see Karlsson Reference Karlsson2003).

Tribal communities and Hindu society

Béteille (1986, p. 316) points out that, despite their differences, historically tribal communities had one thing in common: “they all stood more or less outside Hindu civilization”. Or as Xaxa (1999, p. 3593) clarifies, tribes “continued to be distinct because they escaped colonisation and subjugation”. This is not to say that they were isolated from Hindu society. They existed within, but on the geographical and social margins of, dominant polities of South Asia (Xaxa Reference Xaxa1999; Sundar Reference Sundar2007; Padel Reference Padel2009). Thus, Adivasis are similar to Scott’s (Reference Scott2009) concept of people who inhabited “shatter zones” or “zones of refuge” – relatively inaccessible spaces that were, until the past century, more or less ungoverned by the state.

Béteille (1986, p. 311) distinguishes between “heterogeneous, differentiated and stratified” Hindu society and “homogenous, undifferentiated and unstratified” tribal society. These categories should be seen as ideal types – abstract models that exaggerate certain essential characteristics – rather than a description of social reality. Indeed, as Sundar (Reference Sundar2007) points out, there was some stratification within tribal villages, with the headmen, whose power stems from their supposed ability to mediate with the local gods, tending to have more or better land. Acknowledging this does not, however, undermine the crucial point that, in contrast to the rest of India, there was no landowning class distinct from a landless labouring class in Adivasi communities (Béteille Reference Béteille1986; Xaxa Reference Xaxa1999; Padel Reference Padel2009). Thus, in Weberian terms, while there were differences in status within Adivasi communities, there was no class differentiation.

The relationship between Hindu society and tribal communities is a dynamic one. In the past, it was widely argued that over the longue durée of South Asian history tribes have been slowly but continuously absorbed into Hindu society. Kosambi (1975, p. 25) wrote that the “entire course of Indian History shows tribal elements being fused into general society”; Bose (1941) referred to the “Hindu method of tribal absorption”; and Ghuyre (1963) believed that Adivasis were “backward Hindus” who underwent a process of “Hinduisation”. A fundamental aspect of this process is Adivasis being dispossessed of their land and compelled to work as agricultural labourers on other people’s land. In the past, tribes would cease to exist as independent entities when they lost their distinctiveness. However, since the Schedule sealed the boundaries between tribals and non-tribals, this is no longer the case (Xaxa Reference Xaxa1999; Béteille Reference Béteille1986). Thus, the category Adivasi encompasses communities with a variety of different characteristics, ranging from those who have been absorbed into Hindu society as landless labourers to Adivasis who continue to cultivate the same land as their ancestors did.

From tribal rebellions and Maoist insurgency

During the nineteenth and early-twentieth centuries, there were a significant number of “tribal rebellions” against the colonial state (Raghavaiah Reference Raghavaiah1971; Guha Reference Guha1983). Many of these occurred in areas that are now affected by insurgent activity. For example the Muria (1876) and Bhumkal (1910) rebellions in what is now southern Chhattisgarh (Sundar Reference Sundar2007); the Kalahandi rebellion (1882) in eastern Orissa (Padel Reference Padel2009); and the Kol uprising (1831-1832) and Santal insurrection (1855) in what is now West Bengal, Bihar and Jharkhand (Duyker Reference Duyker1987).

Since independence, Adivasis have predominately taken part in violent conflict in alliance with the Maoist insurgents (Kennedy and Purushotham Reference Kennedy and Purushotham2012). The insurgent leaders, even at the local level, are non-Adivasis and, most frequently, well-educated, dominant or upper-caste and originating from urban areas of Andhra Pradesh and West Bengal (Harrison Reference Harrison1956; Sundarayya Reference Sundarayya1972; Bhatia Reference Bhatia2005; Shah Reference Shah2006; Guha Reference Guha2007; Banerjee Reference Banerjee2008). The insurgent leaders provide previously discrete rebellions with a unified organizational structure that is capable of sustaining armed conflict against the state, while Adivasis supply the insurgents with various resources needed to fight the insurgency, such as shelter, food, intelligence and recruits (Kennedy and Purushotham Reference Kennedy and Purushotham2012).

The link between Maoist insurgents and Adivasis first manifested in insurgency in Telengana between 1946 and 1951, and then in various parts of the tribal belt from 1967 until the early 1970s (ibid.). After the Emergency (1975-1977), most surviving insurgents were released from jail and some attempted to revive the armed struggle. In the 1980s the Maoist Communist Centre (MCC) and People’s War Group (PWG) were the most important insurgent organizations, most prominent in the plains of Bihar and Andhra Pradesh respectively (Bhatia Reference Bhatia2005; Kunnath Reference Kunnath2006; Balagopal Reference Balagopal2006b). Over the past two decades the insurgents have receded in these areas but have expanded their activities in the tribal belt. In 2004 the MCC and PWG unified to form the Communist Party of India (Maoist) (CPI [Maoist]), which marked the first time in 35 years that the vast majority of Maoist insurgent activity was prosecuted by a single organization.

In order to understand why Adivasis support the Maoist insurgency, we must answer two questions. First, what are the specific characteristics of the Adivasis that provide support for Maoist insurgents? Or, more specifically, is it Adivasis who have been dispossessed of their land and compelled to labour on other’s land, or those who remain in control of their land, who are most likely to be involved in insurgent activity? And second, how and why do Adivasis form lasting and meaningful alliances with non-Adivasis insurgent organizations?

Marxist, Durkheimian and Millian approaches to understanding insurgency

Over the past twenty years political science and economics have come to dominate research into civil war, while sociology has been side-lined (Tarrow Reference Tarrow2007; Hegre and Sambanis Reference Hegre and Sambanis2006). Consequently structural explanations that dominated the field in the 1970s and 1980s are largely ignored, and instead rationalist approaches have become more or less hegemonic (Walder Reference Walder2009).

The assumptions underlying the dominant rationalist studies of insurgency can be traced through Popkin (Reference Popkin1979) and Mancur Olsen to John Stuart Mill (Tilly Reference Tilly1978; Paige Reference Paige1983). These studies concentrate on insurgent organizations and the state, largely overlooking the socioeconomic characteristics of the population in the areas in which they operate. This is because it is assumed that, as long as political entrepreneurs provide suitable “selective incentives”, utility-maximizing actors will take part in the insurgency. The most influential large-N cross-national studies emphasize the importance of greed, opportunity or feasibility (Fearon and Laitin Reference Fearon and Laitin2003; Collier and Hoeffler Reference Collier and Hoeffler2004). But several recent qualitative studies of insurgency that have been undertaken by political scientists have clear rationalist foundations (Sambanis and Collier 2003; Wood Reference Wood2003; Kalyvas Reference Kalyvas2006; Weinstein Reference Weinstein2007). Rationalist theories have heavily influenced qualitative studies of the insurgency undertaken by scholars based outside India: Shah (Reference Shah2006) argues that the insurgents have not generated mass support based on their political programme, but, in the context of a very weak state, they control the “markets of protection”; Suykens (Reference Suykens2010) claims that the success of Maoists is a consequence of their ability to control the market for Tendu patta (leaves used to make traditional beedi cigarettes); and Miklian (Reference Miklian2011) suggests that insurgent activity is largely a function of the various states’ counterinsurgent strategies. Although such studies raise some important points, they ignore a vast amount of scholarship emanating from India that shows how aggrieved local actors play a crucial role in the insurgency.

In order to understand the specific characteristics of Adivasis that take part in insurgency, we must revisit the structural theories. Paige (1976, Reference Paige1983) argues that a combination of proletarianized rural inhabitants – landless labourers or sharecroppers – and “backward capitalists” who have nothing but their ownership of the land as the basis of their domination, is most likely to lead to insurgency. Further, in the context of the Maya of Guatemala, Paige (1983) claims that indigenous peoples supported left wing insurgents because they were exploited landless labourers rather than because of any distinctive indigenous identity. The dominant accounts of Maoist insurgency from India, which argue that Adivasis support the insurgents because they have been dispossessed of their land and forced to become agricultural labourers, fit in with this neo-Marxist class conflict analysis. Guha (2007, p. 3305) relates Adivasi involvement in the insurgency with the process of “dispossession and exploitation” and Banerjee (2008, p. 27) argues that “Adivasis” who “were ousted from their lands by colonizers” form the Maoist’s support base. The Government of India (2008, p. 9) also sees “land alienation, forced evictions from land and displacement” as causes of Adivasi unrest, and the CPI (Maoist) (2004, p. 21, p. 14) argue that the “vast majority of the Adivasis have long been deprived of their land” and see the solution to this problem is “land to the tiller”.

The moral economy approach, which shares with Durkheim a concern with the anomie wrought by industrialization and capitalism and its affect on the collective consciousness, has historically been the foil to class analyses (Wolf Reference Wolf1969; Scott Reference Scott1976). It argues that sharecroppers and landless labourers are unlikely initiators of rebellion because they are vulnerable to repression from a distinct landowning class. Landholding cultivators living in communal villages outside direct landlord control are more likely to be involved in insurgent activity for two reasons. First, their relative “organizational autonomy” or “tactical leverage”: they own and cultivate their own land, directly control the immediate processes of economic production and are free from the tight control of landlords. Second, rural inhabitants with strong communal traditions and few sharp internal class divisions are more likely to revolt than those that are structurally and hence socially divided. Wolf (Reference Wolf1969), again referring to the Maya of Guatemala, argues that the distinctive “closed corporate” nature of indigenous communities provides organizational and material advantages for collective resistance against outside oppressors. Few academics in India advocate this approach, although the CPI (Maoist) (2004) and the Government of India (2008) do acknowledge that Adivasis are distinctive from other sections of Indian society.

An alternative Weberian approach

Prima facie, none of these theoretical approaches appear to provide an adequate framework for understanding the Adivasis’ involvement in the Maoist insurgency. The Millian approach concentrates on insurgent organizations but overlooks the role played by local level supporters, whereas Marxist and Durkheimian frameworks stress the importance of certain sociostructural characteristics at the local level but do not take into account the role of insurgent organizations. In addition, none of these frameworks acknowledge the complex and dynamic manner in which social structure and political context interact in the development of social movements (Tilly Reference Tilly1964; Kalyvas Reference Kalyvas2006; Walder Reference Walder2006, Reference Walder2009). In this section we sketch an alternative Weberian approach, which takes into account the role of both the insurgent organization and its support base, as well as the inherent dynamism and complexity of insurgency. Following Kalyvas (Reference Kalyvas2006), we conceptualize insurgency as a transaction between insurgents and their supporters. Insurgents provide disparate groups of supporters with a unified organizational structure, while the supporters endow the insurgents with access to resources needed to sustain the insurgency, such as food, shelter, intelligence and recruits. Thus, insurgencies are processes that provide a medium through which a variety of grievances can be addressed within the space of a greater conflict, rather than binary conflicts neatly organized around a single master cleavage (ibid.). Although actors might be motivated by fundamentally different objectives, their respective interests are served by pursuing guerrilla warfare and seizing local power.

For insurgents motivated by ideology or the pursuit of state power, the creation of a base area is a step towards the seizure of state power. In seizing local power the insurgents undertake a project of “competitive state building” (Kalyvas Reference Kalyvas2006, p. 218). That is, they challenge the state’s monopoly of violence resulting in a situation that Tilly (1978, p. 191) refers to as “dual sovereignty”. The crucial issue here is how the insurgent organization establishes its authority – we could use Weber’s term, Herrschaft – over the population. All the theories that have been considered so far overlook this issue, even though, as Weber (1968, p. 941) points out, it is “one of the most important elements of social action”. It is possible that the insurgents establish control by generating support through the provision of incentives, which might be either selective, as envisaged by the rationalist theories, or “collective”, in the sense that they are offered to certain sections of the population (see Skocpol Reference Skocpol1982; Goodwin and Skocpol Reference Goodwin and Skocpol1989). But the insurgents might also use collective and selective sanctions, most notably violence, to discourage collaboration with the state (Kalyvas Reference Kalyvas2006). As Weber (1991 [1948], p. 79) points out: “obedience is determined by highly robust motives of fear and hope”.

Weber’s concept of legitimacy provides us with a means of understanding some of the dynamism inherent to insurgency. Legitimacy refers to validity claims of rulers in the eyes of the ruled. It is not an internalised constant but an emotional feeling that arises from assessing a state – or state building organization – at any given moment and can vary from being non-existent to very high (Weber 1968). Individuals must interpret the world around them – they do not react mechanically to their structural position or to selective or collective incentives and disincentives that they are offered. In situations of dual sovereignty two political organizations within the same territory compete to claim and establish legitimacy. There are tendencies regarding which organization is viewed as legitimate or illegitimate that are related to people’s position in the structure of power and privilege. But, because actors’ assessments of legitimacy are subjective, they can be either reinforced or altered by short-term processes, such as the dynamics of violence (ibid.; also see Snow et al. Reference Snow, Rochford, Worden and Benford1986; Wood Reference Wood2003; Kalyvas Reference Kalyvas2006; Walder Reference Walder2006, Reference Walder2009; Jasper Reference Jasper2011).

Variables

In this section we analyse quantitative data to investigate whether insurgent activity occurs in areas where Adivasis remain in control of their land or where they have been dispossessed. The spatial unit is the district, the administrative level below the state and the smallest division at which all the variables are available.

The dependent variable is insurgent activity. We documented all fatalities in incidents involving Maoist insurgents that were reported in The Times of India (Mumbai edition), India’s most widely read English newspaper, between 1982 and 2011. There were 1,105 fatal incidents in this period, resulting in 4,698 deaths. We then created a dummy variable using the mean number of deaths per decade (four) as the cut off point.

The main independent variables are the percentage of Scheduled Tribes that are either landless “agricultural labourers” and landholding “cultivators”. A landless labourer “has no risk in the cultivation, but merely works on another person’s land for wages […] in money or kind or share” (GoI undated c). Landholding cultivators, on the other hand, take on the risk of cultivation on land that they either own or rent. These two categories represent, respectively, Adivasis who have been dispossessed of their land and those who remain in control their land. We control for the total population values in order to ensure that it is not simply the general level of landless labourers or landholding cultivators that explains the presence or absence of insurgent activity. We also control for the percentage of Scheduled Tribes in the population to ensure that it is not simply the large Adivasi population that accounts for the increased probability of insurgent activity.

We control for several other variables that are enumerated by the Census of India. First, we include the log transformed total district population because, due to the way that insurgent activity is operationalized, more populous districts are more likely to be affected by insurgent activity (Hegre and Sambanis Reference Hegre and Sambanis2006). Second, we control for the level of development because less developed areas are more likely to be affected by insurgent activity (ibid.). Cross-national studies would tend to use per capita GDP, but, because this data is not available at the district-level, we use literacy rates. Third, because state power tends to be weaker in rural areas, insurgents are able to build up base areas in the countryside without interference from the state (ibid.). We therefore control for the proportion of inhabitants that live in rural areas. We include a dummy variable to control for the occurrence of insurgent activity in the district in the previous decade. Finally, we add a calendar year variable to capture possible changes in the geopolitical climate over time. Appendix A shows descriptive statistics and appendix B is a correlation matrix of the main variables.

Sample

We run our regressions with two different samples. The first, sample A, includes all districts in India, except those in Jammu and Kashmir, and the North East, where there have been non-Maoist insurgencies in the period under analysis. The states of Haryana and Punjab, as well as five districts, drop out of the analysis when we include tribal specific variables because they do not have Scheduled Tribe populations. In 2001 our sample covered 916 million people and 72 million Scheduled Tribes – that is, 89 % of the total Indian population and 86 % of the India’s Scheduled Tribe population.

We do not believe that there is only one route – in this case the tribal route – to Maoist insurgency – an assumption that is implicit in rationalist studies of internal conflict. Indeed, ethnographic data indicates that lower-caste landless labourers supported Maoist insurgents on the plains of Bihar and Andhra Pradesh (Kunnath Reference Kunnath2006; Bhatia Reference Bhatia2005; Balagopal 2006b). Sample B is limited to the tribal belt in order to partially isolate the “tribal route” to insurgency. The total population of the area covered by the Fifth Schedule was 166 million and the Adivasi population was 45.5 million in 2001– respectively 18 % and 63 % of sample A. We predict that the effect size will be greater in this sample because it partially excludes other mechanisms that lead to insurgent activity.

Robustness Tests

We performed a series of robustness tests on our findings. First, we removed potential outliers according to a liberal definition of standard deviations in the residuals greater than |2|. Second, we tested whether our findings were sensitive to our operationalization of insurgent activity by halving (to 2) and doubling (to 8) the threshold of deaths needed for a district to be affected by insurgent activity. Third, we introduced state dummies to test if our variables were merely a proxy for a state’s shared traits. These tests did not materially alter the significance of the main explanatory variables or the models’ pseudo R².

Case selection

The statistical analysis reveals several interesting points but leaves crucial questions unanswered. If dispossession and exploitation cannot explain the Adivasis’ involvement in the Maoist insurgency, what does? If Adivasis do not share a common consciousness with Maoist insurgents, how and why do they form a lasting alliance? This section addresses these questions by undertaking a case study of one district – Dantewara, the southernmost part of the former “Jungle Kingdom” of Bastar that now forms part of Chhattisgarh state – which is informed by fieldwork carried out in 2008, archival data such as government publications, insurgent documents and newspaper articles, and secondary ethnographic data.

Dantewara is a crucial case in the context of this study for several reasons. First, it is widely seen as the epicentre of the Maoist insurgency. It has been a haven for insurgents since the early 1980s and their presence has undermined the sovereignty of the Indian state, as there are areas of the district where neither the police nor other state employees dare to venture. An Indian Administrative Service report from 1990 stated that the insurgents were running a “parallel government” in the area (Sharma Reference Sharma1990). In 2011, Jairam Ramesh, Minister for Rural Development, admitted that there were “liberated zones” in Southern Chhattisgarh, where “the state’s writ doesn’t run” and “Our officers […] don’t venture in these places” (Ojha Reference Ojha2011). What is more, since 2005 Dantewara has been the scene of a brutal war between the insurgents and a counterinsurgent militia, Salwa Judum, and this provides a fascinating opportunity to better understand the relationship between Adivasis and Non-Adivasi insurgents.

Second, for several centuries Dantewara has been an important battleground for understanding India’s tribal population and its relationship with Hindu society. In the mid-nineteenth century, the Deputy Commissioner of Bastar described the tribal population of the region as “the most perfect specimens of the aboriginal race” (Lyall 1868, p. 25). British “anthropologist administrators”, such as Elwin (Reference Elwin1936) and Grigson [1991 (1949)], lived with and studied the Adivasis of Bastar, and this has been continued by, inter alios, Sundar (Reference Sundar2007) and Gell (Reference Gell and Appadurai1986, Reference Gell1997). In 2001 78.5 % of the population were Scheduled Tribes, the third highest proportion in India (Census of India 2001). The socioeconomic characteristics of Scheduled Tribes in Dantewara are representative of those in insurgent-affected districts, as identified by our statistical analysis: 89 % of the Adivasi workers are landholders, whereas only 5 % are landless labourers (ibid.).

What is more, it has been noted that the tribal population in the region was relatively well off in the mid-1980s, when the insurgents were expanding their influence. This fits in with our statistical analysis, which challenges the dominant narrative that associates the insurgency with landless and immiserated Adivasis. BD Sharma, Commissioner for Scheduled Castes and Scheduled Tribes (1987, p. 68), described how “The luxuriant forest on the hills and the mighty river Indravati below provide plentiful sustenance to all and even the landless, the old and the infirm”. Gell (1986, p. 124) notes that “Despite being a notoriously ‘backward’ area, supposedly occupied by miserable, poverty-stricken tribals […] the local economy is in a flourishing condition, prosperous in good years and well able to withstand the rigors of bad ones”. Gell also points out that landholdings held by Adivasis in the district are “enormous” by Indian standards (ibid.). Even now, the size of the average landholding of an Adivasi in Dantewara is 4.8 hectares compared to the national mean of 1.6 hectares for the sample (GoI undated b).

Tribal history of Bastar

In the pre-colonial period, Adivasis in Bastar, as well as the tribal belt more generally, had strong symbolic and material links to the forests (Elwin Reference Elwin1936; Guha and Gadgil Reference Guha and Gadgil1989; Sundar Reference Sundar2007; Padel Reference Padel2009). They practised shifting cultivation, hunted and collected forest produce for food, fuel, medicines, buildings materials and alcohol, as well as to exchange with traders for salt, cloth and cash (GoI 1861). Surplus extraction from the tribal areas of Bastar did not occur through taxation on land or the exploitation of a class of landless labourers (Gell Reference Gell1997; Sundar Reference Sundar2007). The “1861 Report on Bustar and Kharonde dependencies” stated that “savage hill tribes […] pay no tax whatsoever” (GoI 1861, p. 44). Instead, the abundant natural resources – timber, non-timber forest produce and minerals – were the chief source of state income, although commercial exploitation of the forest did not occur in a systematic manner (Guha and Gadgil Reference Guha and Gadgil1989; Sundar Reference Sundar2007). The situation changed with the arrival of the East India Company, which set about acquiring control over India’s natural resources, using bureaucratic and legal means to recast patterns of ownership and rights to common resources in a manner that was detrimental to the indigenous population.

This is illustrated by two pieces of legislation, which the British encouraged the Rajah and Diwan (Prime Minister) to introduce. First, the Forest Act of 1878 gave the state the right to “reserve” – expropriate, administer and keep Adivasis out of – all forests in colonial India. It was not until the last decade of the nineteenth century that the Forest Act was applied to Bastar, but between 1891 and 1910 one-third of all the forest was reserved (Sundar Reference Sundar2007). Second, from 1889 onwards, the state began to regulate the collection of forest produce by granting monopsonies to non-Adivasi traders and imposing duties on Adivasis collectors (ibid.). This legislation led to an influx of Non-Adivasis into tribal areas in the form of forest guards who kept the Adivasis out of the forests and charged them to collect forest produce. It also led to the emergence of traders who dominated the market for forest produce, and officials and police to administer and enforce the system. The manner in which these changes negatively affected the lives of Adivasis is neatly summed up by a tribal who described to Elwin (1936, p. 22) his vision of heaven as “miles and miles of forest without any forest guards”.

The changing terms of interaction between Adivasi communities and Hindu society resulted in several “tribal rebellions” in Bastar (Gell Reference Gell1997; Sundar Reference Sundar2007). The 1910 rebellion, which was the most violent and widespread, was a response to the state’s attempt to increase surplus extraction from, and control over, the forests. Five years previously the state had revealed plans to reserve two-thirds of Bastar’s forest and the rebellion drew its supporters from precisely those areas in which reservation had taken place. It is interesting to note that the 1910 rebellion was popularly referred to as the Bhumkal, a reference to the social solidarity that binds members of a clan to each other and to their specific bhum (earth), as well as the political authority of the council of elders. Sundar (2007, p. 133) suggests that “one might read the Bhumkal as a mode of protest juridically sanctioned by local authority, that of the elders in the name of the earth, a pitting of indigenous law against colonial law”.

From British colonialism to internal colonialism

On paper, the provisions of the Fifth Schedule of the Indian Constitution and later legislation protect the Adivasis’ rights to their land and permit them a certain degree of political autonomy. But, in practice, the post-colonial Indian state exercises neo-colonial forms of control over Adivasi areas (Guha Reference Guha2007; Sundar Reference Sundar2007; Padel Reference Padel2009; Scott Reference Scott2009; Kennedy and King Reference Kennedy and King2011). The Forest Act (1927) remains the foundation of forest administration in India and the Forest Department continues to own and administer the vast majority of forests. In Dantewara 52 % of the forest is reserved and another 31 % is “protected” (Government of Chhattisgarh 2005). The state, currently in the form of the Chhattisgarh State Minor Forest Produce, attempts to control the market for non-timber forest produce and aims to keep prices low, passing on the benefits to Non-Adivasi traders and industrialists (Saxena Reference Saxena2003; Sundar Reference Sundar2007). Additionally, large mining and steel companies with strong links to the Indian political elite – most notably ESSAR and Tata – have increasingly coveted Dantewara’s sizable reserves of iron ore (Guha et al. 2006; Sundar Reference Sundar2007). The state uses the Land Acquisition Act (1894) to expropriate Adivasis’ land for “development” projects. The Panchayats (Extension to the Scheduled Areas) Act (1996) should, in theory, allow the local community to veto such projects, but in practice villagers are threatened by “goons” who coerce voters into agreeing to their demands, the minutes of Gram Sabha (village meetings) are rewritten, or fake Gram Sabha are used to approve a project (ibid.).

On the ground, forest guards, traders and policemen continue to have a massive and overwhelmingly negative influence over the lives of Adivasis. In the early 1980s a Madhya Pradesh Legislative Assembly minister who had recently completed a padayatra (journey by foot) through Bastar argued that “An unresponsive administration and exploitation of the tribals by a section of officials of the revenue and forest departments and contractors has prompted the people in Bastar to seek help from Naxalite elements” (Times of India 1983b, p. 6). What is more, this worsened as increasing numbers of paramilitary police were stationed in Bastar in response to the insurgency. A report by the People’s Union for Civil Liberties from 1989 stated that “the presence of the large contingent of armed police in the interior Adivasi villagers has changed the environment completely” (Times of India 1989, p. 13). The report documents police raiding villages, stealing property and grain, sexually abusing women, brutally killing some Adivasis, and detaining and torturing others. It is apparent that, in Dantewara, as well as elsewhere in the tribal belt, such incidents continue up to the present day (Guha et al. 2006; Nagarajan Reference Nagarajan2010).

The role of the insurgent organization

Maoist insurgents entered Bastar from Andhra Pradesh in 1980 (Times of India 1980; CPI [Maoist] 2005). The first insurgents were non-Adivasi students from Warangal University in Andhra Pradesh and the leaders of dalams (guerilla squads) and sangham (committees) continue to be, on the whole, Non-Adivasis from Telangana (Times of India 1983b, 1987; Balakrishnan Reference Balakrishnan1993). Nevertheless, the insurgents are supported by Adivasis, who provide them with the resources needed to undertake insurgency, such as food, shelter, intelligence and foot soldiers. The insurgent leaders frame the insurgency in terms that are meaningful to the Adivasis. They support the creation of a separate tribal kingdom or Dandakaranya state in Dantewara and contiguous areas, which was a demand of tribal rebellions before independence (Sundarayya Reference Sundarayya1972; Times of India 1981, 1992; Surendran Reference Surendran1992, Reference Surendran1993; CPI [Maoist] 2005; Kennedy and Purushotham Reference Kennedy and Purushotham2012). The insurgents understand the symbolic importance of the Bhumkal rebellion among Adivasi communities in Dantewara. The CPI (Maoist) (2005, p. i) refers to it as an earlier step on “the path of liberation” and sees “the present generations [of Adivasi insurgents] as their heirs”. In 2009, the insurgents issued a pamphlet warning of widespread violence to mark the centenary of the Bhumkal rebellion (Bose Reference Bose2009) and the insurgents claimed that the ambush that killed 76 paramilitary policemen in April 2010 was timed with this in mind (Times of India 2010b).

The insurgents provide sections of the tribal society with collective incentives. In the early 1980s insurgents would visit Adivasi villages at night and enquire about which forest officials, traders, police and government officials were intimidating them, demanding bribes or paying low wages. The accused would be brought before a Jan Adalat (People’s Court), where the insurgents would “settle the accounts” (Times of India 1984a, p. 12). Often they would be humiliated and warned to behave themselves in future. An Indian Administrative Service Report describes how “an errant constable or forest guard is made to crawl before a village assembly and seek pardon. A forester is forced to hold his ears, and in schoolboy fashion asked to do 100 situps” (Sharma Reference Sharma1990, p. 9). At other times, the insurgents would threaten to cut off the limbs of contractors or to burn their trucks – and sometimes they would carry out these threats (Times of India 1984b, 1987, 1988a; Kher 1991).

It is apparent that the insurgents provided Adivasis with a powerful ally and this brought concrete benefits. They chased away forest guards, improving the Adivasis’ access to the forests. The insurgents secured a fifty-fold increase in the price of tendu patta over the period that they have been active in the area and they claim that prices are 60 % higher in areas where their organization is strong (GoI 2008; Guha et al. 2006). In addition, the insurgents claim to have engaged in considerable development work over the last twenty years, constructing schools, clinics and ponds [CPI (Maoist) 2005]. A Congress MP pointed out in 1988 that “The local population had links with Naxalites because the latter provided them with better protection and help than the state administration” (Times of India 1988b, p. 17). This analysis is corroborated by several recent government reports, which admit that Adivasis see the insurgency as “basically a fight for social justice, equality, protection and local development” (GoI 2008, p. 60; also see GoI 2006, 2010).

The insurgents also provide selective incentives to those people that take an active role in the insurgency. It is estimated that throughout India the insurgents generate between 15 and 20 billion rupees ($300 and $400 million) per year, primarily by extorting money from economic actors ranging from forest produce traders to large mining companies (Times of India 2009; Satapathy, Ohja and Mandal Reference Satapathy, Ojha and Mandal2010; also see Shah Reference Shah2006 and Suykens Reference Suykens2010). The Ministry of Home Affairs recently claimed that insurgents paid their foot soldiers a salary of 3,000 rupees per month together with a cut of extortion money (Times of India 2010a). While these reports might or might not be true, they raise an interesting point. For young Adivasis – both male and female – joining the insurgents is an opportunity to enter an organization that commands respect and fear, provides them with access to food, shelter, and perhaps the opportunity to make money. On the other hand, the insurgents also use targeted violence – selective disincentives – against perceived and real opponents within Adivasi communities, such as suspected informers. This is often done in front of the whole village in Jan Adalats, and the punishments can be brutal; suspected police informers’ limbs are amputated, their eyes are gouged out, and they are frequently killed (for example Times of India 1982, 1983a, 1998; Sharma Reference Sharma2011). Such displays serve a double role: to punish suspected opponents and to send out a warning to other members of the community.

Insurgents versus counterinsurgents

From 2005 onwards, Dantewara’s population have been able to choose between two competing political organisations, the insurgents and Salwa Judum, a government-sponsored militia (Guha et al. 2006). Salwa Judum aimed to clear the forests of inhabitants in order to deprive the insurgents of their support base. With the help of regular security forces, they forcibly removed Adivasis from their villages and herded them into fortified roadside camps, which acted as “strategic hamlets”. The Government of Chhattisgarh appointed 5,000 Salwa Judum members as Special Police Officers (SPOs). They were given a gun and paid 1,500 Rupees a month to protect the camps and hunt insurgents (Balagopal Reference Balagopal2006a).

The presence of two competing political organizations polarized society. Before the formation of Salwa Judum, the options available to inhabitants of insurgent controlled areas were to either support the insurgents or remain indifferent. After the introduction of a second political organization they were compelled to actively support and seek the protection of either the insurgents or the counterinsurgents. As the Chief Minister of Chhattisgarh declared, “those in the camps are with the government and those in the forests are with the Maoists” (quoted in Balagopal Reference Balagopal2006a, p. 2184). 644 out of 1,220 villages in Dantewara were forcibly evacuated, and between 40,000 and 50,000 people lived in roadside camps at the height of the counterinsurgency (Balagopal Reference Balagopal2006a; Guha et al. 2006). The total district population was almost 720,000 in 2001 (Census of India 2001), indicating that approximately a quarter of a million people chose to remain in insurgent controlled territory or flee to neighbouring areas. In such a situation it is enlightening to ask why some people decided to seek refuge in Salwa Judum camps while others refused.

Many Salwa Judum leaders are Non-Adivasi traders and contractors from the Hindi heartland of northern India. Some were adversely affected by the insurgents’ attempts to provide collective benefits to Adivasis, while others previously profited from trading with the insurgents but came to realize that there was money and influence to be gained from supporting the counterinsurgency. Indeed, Guha et al. (2006) point out that there was little or no oversight or accountability for the huge funds that poured into Dantewara during the counterinsurgency. Notwithstanding, the counterinsurgency cannot simply be presented as a Non-Adivasi movement. Significant numbers of Adivasis were also involved with Salwa Judum at various levels and there were large numbers of Adivasis living in their camps. Why was this so?

It was apparent from talking to Adivasis in Salwa Judum camps that they all had reason to resist or fear the insurgents. In many cases they previously held a privileged status as a result of descending from a particular lineage or collaborating with dominant society. But, in the process of establishing control, the insurgents challenged and undermined their authority. They include kotwars, the traditional administrators in the village who have a reputation for demanding bribes for services such as registering births, and sarpanches, the elected community leaders who are often said to embezzle a proportion of money from government projects. Adivasi teachers and forest guards, who are seen by the insurgents as agents of the state and are often involved in local level corruption, also form a sizeable proportion of Salwa Judum sympathizers.

It is, however, not possible to depict all Adivasis opposing the insurgents as having vested interests that are harmed by an otherwise popular insurgency. There are cases where factional identities result from short-term processes operating during the course of the insurgency – endogenous mechanisms – rather than Adivasis’ relationship to the structure of power and privilege. First, a large proportion of Special Police Officers (SPOs) were Adivasi youth motivated by the promise of a government job, meagre salary and the status of carrying a gun. They did not realize that they were taking sides in the conflict and, in fact, became SPOs for the same reason that others had become insurgents before 2005. In both cases they were “seduced by their new-found – and essentially unearned – authority” (Guha Reference Guha2007, p. 3311). Once they had made this choice, however, they were unable to leave Salwa Judum for fear of reprisals by the insurgents. Second, some Scheduled Tribes who joined the counterinsurgency were previously sangham members who were captured and forced under pain of death to join Salwa Judum. The fear of reprisals for their treachery, as well as shame for what they had done, kept them on the side of the counterinsurgents and prevented them from returning to the insurgents. Third, because the decisions as to who to target are taken in stressful situations and based on imperfect information, insurgent violence is sometimes misplaced. Consequently, some victims and their family and friends do not resent the insurgents and seek the protection of Salwa Judum because of their position in Adivasi society, but due to fear or the desire for revenge as a consequence of being the unintended victims of insurgent violence.