Steven I. Wilkinson
There is often a tendency to treat caste and communal conflicts and politics as separate. In fact, however, the degree to which one of these identities is salient in politics or conflicts at a particular time is often linked to the institutional and economic incentives supporting mobilization around the other, or to another identity such as language or class.1 André Béteille pointed out long ago for instance that communal politics seemed to take a Hindu– Muslim pattern in the north but have a caste pattern in the south, and he noted that even within the south there was substantial regional variation, with Muslim political mobilization strongest in those areas such as Kerala and parts of Andhra, where the non-Brahman movement had been weakest.2 In the late 1980s and early 1990s, the interconnectedness between the salience of caste and communal identities became even more apparent in the violent political contest between “mandir” and “mandal”: the Bharatiya Janata Party (BJP) and the Sangh Parivar, on the one hand, pushing a policy of Hindutva (“Hinduness”), and parties representing a variety of backward and lower caste interests pushing a policy of caste reservations.
The most important fact to understand about the development of caste and communal politics and conflicts since Independence is that, politically and constitutionally, caste is a privileged category, one that can deliver tangible benefits to communities and voters (e.g., reservations in education, employment, and sometimes in politics) in a way that religious identities cannot. So while attempts have been made, at various times since Independence, to use anti-Muslim, anti-Christian, and (much more rarely) anti-Sikh mobilizations and violence to create a Hindu majority for a particular political party (and that party has not always been the BJP), these attempts have only been successful in the short term, and have typically foundered on the much greater resonance of caste appeals to the state, local, and even the national electorate.3 From 1989 to 1992, for instance, it seemed as if the Ayodhya campaign around the Babri Masjid and other “disputed” sites, which involved large-scale yatras (processions) and demonstrations across India, involving millions of participants, might be capable of generating a permanent Hindu majority for the BJP. On the back of the campaign to build a Ram Mandir on the site of the Babri mosque, the BJP’s representation shot up from two seats in the Lok Sabha to 88 in the December 1989 elections, and then to 120 seats in 1991. But, after the destruction of the Babri Masjid on 6 December, 1992, the national Hindutva agitation temporarily ran out of steam, as the immediate goals seemed to have been achieved, the violent destruction of the mosque and subsequent communal riots turned off many supporters, and the Congress government of P. V. Narasimha Rao imposed emergency rule on four BJP-ruled states. The BJP, contrary to its own expectations, then suffered very severe electoral reversals in 1993 state elections in these states at the hands of parties such as the Bahujan Samaj Party and Samajwadi party in UP, which promised concrete policies to particular lower castes rather than the unclear future benefits of a Hindu Raj.
Turning now to communal conflicts, the fundamental fact about communal polarization and violence in post-Independence India is that whether it happens is the outcome of political decisions. Riots are often, although not always, fomented for political purposes, and they are prevented or stopped by the state police and administration when it is in the interests of those who control the state government to do so.4 To understand the political incentives facing the state politicians who control the 28 state governments which in turn control the police is therefore the most important factor in understanding why communal violence takes place.
Riots pay political dividends: they unify Hindus behind the party that seems best able to defend “Hindu” interests, they help break up the coalitions of other parties, and they temporarily make the Hindu–Muslim cleavage appear more significant than other political issues, such as caste, urban vs. rural cleavages, or development. Christophe Jaffrelot has rightly pointed out that embracing Hindutva and the organizational energies of militant organizations such as the RSS, VHP, and Bajrang Dal seems to be especially attractive to BJP leaders when the party has suffered reverses, and therefore looks unlikely to win on other issues.5 In 2001, most notably, the BJP government in Gujarat, which had performed badly since its election in 1998 in terms of development and rehabilitation after the January 2001 earthquake, turned decisively to a “hard” Hindutva policy to save itself after a succession of defeats in local elections pointed to likely defeat in upcoming Vidhan Sabha elections. These defeats convinced the party leadership that only a sharp turn to the right would help, and the incumbent chief minister was replaced by hardliner Narendra Modi. In March and April 2002 the Modi government reportedly fomented large scale riots and pogroms against the state’s Muslim minority in order to solidify a majority behind the party in upcoming Vidhan Sabha elections. Perhaps a thousand people, mainly Muslims, died in these disturbances and tens of thousands more were forced to flee their jobs and homes.6 The riots paid clear electoral dividends, and, in December 2002, the BJP won a crushing victory in state elections over Congress, doing especially well in riot-affected districts. Polls taken during the elections suggested that the riots were a major issue in helping swing voters decide in favor of the BJP, as well as in increasing turnout among the BJP’s core supporters.7 But, overall the anti-minority mobilization of Hindutva and communal polarization, like most other religious ideologies, makes much more sense as an oppositional ideology, a temporary way of unifying people against a clear target, than it does as a way of governing.This is because in itself it (like secularism) offers no clear roadmap to decide what Harold Lasswell long ago identified as the key questions of politics: who gets what, when and how?8 In the absence of rules bolstering religious identities and favoring one religion over another—rules that would be unconstitutional in the secular framework created in India in 1950—religion is inherently limited as a political ideology, compared with linguistic or caste identities that do benefit from this government and institutional support.
The Indian Constitution of 1950 provides certain benefits for scheduled castes and tribes, and also in a 1951 amendment specifies that benefits may also be provided for other backward classes (OBCs)—meaning, in effect, castes—a category that might potentially incorporate most of the population, since backwardness is in the eye of the political beholder. In sharp contrast, however—and understandably given that the Constitution was drafted after Congress’ long struggle with the Muslim League during the campaign for India’s independence, as well as the communal violence of Partition that followed—the constitution is unambiguous in its opposition to religious preferences, such as job reservations, educational reservations, or the separate constituencies that existed before 1950 for Muslim, Sikh, and Christian minorities in various provinces.The future of these religious reservations was extensively debated by the Constituent Assembly from 1947–49, and the assembly decided to abolish them, the substantial support they still enjoyed at the time from many in the Muslim and Sikh communities notwithstanding.9
Nehru would have liked to abolish caste reservations as well, and move away entirely from a society in which caste or religious labels were important. His ultimate goal, as he wrote to Charan Singh in 1954, was to end the caste system, which he saw as “the biggest weakening factor in our society.”10 That Nehru could not achieve this goal, however, was largely because caste was already entrenched in politics in two different ways. First, in the previous four decades there had been a very substantial “non-Brahman” movement in the south, especially in the province of Madras and in the princely states of Mysore and Travancore-Cochin, against upper caste dominance in government employment, education, and politics. This upper caste dominance had been overwhelming in the early twentieth century, with Brahmans, for instance, accounting for around 50–80 percent of government employees in many branches of the subordinate civil service in Madras despite accounting for only 3.5 percent of the population, a percentage that reflected their much higher levels of wealth and education.11 Further, 68 percent of the graduates of Madras University in 1918 were Brahmans.12
After the Second World War non-Brahman movements used their access to sympathetic policymakers in Madras and the princely administrations of Mysore and Travancore and Cochin as well as their control of the new elected provincial government in Madras (1920) to institute widespread government reservations for backward classes.13 These reservations created large numbers of politicians, employees and voters who invested in backward caste identities—the number of castes formally recognized as backward in Madras shot up as a result from 45 to 245 by the mid-1920s—and formed a well-entrenched interest group that was able to resist legal and political challenges to the system of employment reservations after Independence.14 In 1950 and 1951, for example, there were large and violent protests in Madras, led by E.V. Ramaswamy Naicker’s Dravida Kazhagam Party, after court rulings that placed caste reservations in the state in jeopardy. These protests led the Madras government to pass a motion defending the reservations system in 1951, and soon after the Indian National Congress backed down over the issue, passing Amendment 15(4) to the constitution, which permitted reservations for “backward classes.”15
The second way in which caste was entrenched was through a historic compromise that Congress itself had to make with Dr Ambedkar in the mid-1930s over the question of political reservations for what came to be called the Scheduled Castes. (The “Schedule” refers to a list appended to the 1935 Government of India act, specifying castes that were treated as untouchable by caste Hindus.) Congress was generally opposed to such reservations, and Gandhi in particular opposed them, seeing them as an insidious part of a more general British divide-and-rule policy. But Congress was forced to compromise over the issue in the 1932 Poona Pact. The compromise involved accepting British proposals for reserved SC seats but not their proposals for separate caste electorates for these seats as the price of enlisting lower caste support in the campaign for independence. The tangible electoral impact of such support at the time may not have been great, given that lower castes constituted a small share of the electorate because of the property-based franchise (around 14 percent could vote after the 1935 act), but the symbolic value was high and securing Ambedkar’s support also gave the British one argument less to use when they claimed that general devolution of power had to wait until Indians were united in their demands and that granting independence would not unduly disadvantage any particular important minority.
So, even before the Constituent Assembly was elected and independence attained—and at a time when Congress leaders spoke out forcefully against the system of separate electorates for religious minorities—caste reservations were well entrenched in the south, and reservations for “depressed classes” (SCs) had been accepted in principle by Congress leaders in the Poona Pact.The practical political effect of so many people already being nominally included within the reservation system, at least once southern politicians successfully blocked legal efforts to end reservations in 1950–51, was that politicians representing lower and backward castes had no incentive whatsoever to end reservations, and in fact if they wanted to add supporters it was much more effective to simply extend the principle of reservation to more and more castes. This political logic played itself out very quickly in the south after Independence, as politicians recognized more and more castes as “backward” and eligible for reservations throughout the 1950s and 1960s: by the mid-1950s over 40 percent of positions in employment and education in Tamil Nadu were reserved for members of the backward and most backward castes, a proportion that eventually rose to 69 percent. In the north, as a result of backward caste mobilization, large-scale reservations were gradually extended to the OBCs in the same way in the 1970s and 1980s, a process systematically explored in Jaffrelot”s 2003 book India’s Silent Revolution.16 By 1980, according to the Mandal Commission, the number of castes officially recognized as “Backward” in India had risen to 3,743, compared with 2,394 in 1955, at the time of the first backward caste commission headed by Kaka Kalelkar.17 In some states, the rise was even more dramatic, with the number of OBCs in Tamil Nadu that qualified for reservations reportedly rising from 150 before 1970 to 310 castes in 1994.18
Much of this increase has come about as the result of explicit quid pro quos, as politicians have used promises of reservations to peel off supporters from larger groups allied with another party or leader, and caste leaders themselves have indicated that the support of their community can be obtained in return for reserved status. In Rajasthan, for instance, the Meena community was reportedly recognized as a Scheduled Tribe in return for the support of 13 MLAs for the chief minister during a party leadership contest in 1957.19 In 1994 the Vokkaligas and Lingayats in Karnataka were recognized as OBCs in return for their support of Mr Veerappa Moily in the state elections.20
The extension of reservation to more and more jobs and positions and of reserved status to a greater share of the population has, of course, been resisted by upper castes, as well as others (such as those on the Left, at least until recently) who think that entrenching caste identities in jobs and education might not be the way to get beyond caste identities and end caste inequalities. In 1989–90, most notably, there were violent upper caste protests in large cities and on university campuses against the V. P. Singh government’s proposed implementation of the Mandal Committee’s nearly decade-old recommendations to extend the scope of OBC reservations in central government employment. The commission had controversially estimated the OBCs at 52 percent of the population, and the proposals involved increasing the number of reserved places in government employment and universities by 27 percent, while decreasing the number of “merit” places by a similar amount. One student at Delhi University, Rajiv Goswami, set himself on fire, and several other students across the country followed suit in angry demonstrations. And in 2006, in what became known as the “Mandal II” protests, there were renewed demonstrations by upper castes against plans to extend OBC reservations to one of the few areas not yet affected, namely, higher level graduate education, by Human Resources and Development Minister Arjun Singh. But the fact that opposition in both cases took the form of public demonstrations by groups of students without substantial political direction was, paradoxically, a sign of the anti-reservation movement’s fundamental political weakness. The fact is that, in both cases, but especially in 2006, the political arithmetic in favor of reservations is simply so overwhelming that no major politician will come out openly against them. In a national survey done in 2004, 61 percent of the Indian population supported reservations and only 22 percent opposed them, proportions very close to the percentages of dalits and Backward Castes in the population, on the one hand, and forward castes on the other.21 Given this overwhelming political support, violent demonstrations in urban areas where upper castes are a larger share of the population are one of the only ways, together with court cases, in which opponents can try to slow their growth.
The political currency of reservations has, however, become devalued through overuse, with the number of groups being made eligible for reservations increasing much more rapidly than the supply of government positions or other benefits. So, in response, powerful caste groups have asked for and politicians have promised more valuable forms of reservation: such as inclusion within “Most Backward Caste” classifications that offer more benefits than are available to general backward castes, or specific “quotas within quotas” that guarantee groups that they will receive a specific share of benefits, rather than simply being included within a larger category in which better educated and wealthier jatis might secure most of the benefits.22
These efforts have, not unexpectedly, led to fierce conflicts and even some quite substantial caste violence because some castes that benefit from the current classifications seek to prevent any changes that would disadvantage their groups. As one backward caste minister who opposed such changes in UP put it in 2001: “Come what may, we will not allow anybody to take away from our share. If separate reservation is required for the most backward castes, let there be an increase in the [percentage of] reservation.”23
Politicians and caste leaders can use various methods to block changes to the reservation system that they do not like. Politically influential backward castes in Kerala, for instance, blocked a caste census proposed in 1995 that would have increased pressure for reform of the existing reservation system by demonstrating that their own “backward” groups were in fact doing better than some “forward” groups.24 Three years later, in September 1998, the census was finally dropped.25 The Yadavs and other relatively well-off OBCs in Uttar Pradesh successfully blocked Rajnath Singh’s proposals to create a southern-style MBC category in UP, a measure Singh hoped would split the political coalitions created by the BJP’s rivals, the SP and BSP. Further, in Andhra Pradesh, Madigas and Malas have frequently come to blows since 1994 as the latter have tried—through direct action as well as their support for particular parties—to block efforts by the worse-off Madigas to reform the scheduled caste reservation system in a way that will disadvantage the Malas. In 1998 the Madiga Reservation Porata Samithi (MRPS) launched a statewide agitation in favor of the subdivision of the SC category that led, over the course of a week, to ten attempts at self-immolation (one ending in death), 1,100 arrests, several large-scale strikes, and the burning or partial burning of 86 buses.26 This violence was largely repeated by the MRPS in 2004. Similar conflicts have also arisen, so far less violent,over subdividing reservations for OBCs and SCs, for example between the Meenas and Gujjars in Rajasthan, between Jatavs and Pasis in Uttar Pradesh, and between the Mahars and Mangs in Maharashtra.We can expect these distributional conflicts within ethnic categories to rise in number and intensity in the future, unless the private sector should quickly create large numbers of good jobs for middle and lower castes outside the reservation systems, which seems unlikely given the very poor level of state primary, secondary, and higher education to which many of them have access.
Writing in October 1991, in the aftermath of the violent street conflicts over the Mandal Commission, the eminent sociologist André Béteille rightly predicted that the economic liberalization then beginning in India would, sooner or later, be bound to collide with the system of caste reservations. The economic reforms would reduce the relative share of jobs controlled by the center and the various state governments, and the principles and mechanisms of the market would conflict with those of government planning and reservation.27 Since 1991, despite a growing population and growing demand for jobs, the number of positions in central and many state governments has remained stable, and the massive retrenchment of many public sector units (PSUs) has also meant that the overall number of reserved places in industrial enterprises under state or central government control has also been stagnant. Overall, central government employment, in fact, dropped by 2.66 percent between 1995 and 2001, to 3,876,000.28 So, with the number of options for expanding reservations within the state sector diminishing, the political focus of demands for reservations has, since the late 1990s, begun to shift to the private sector, which has long been a bastion of upper caste dominance.This extension of caste reservations to the private sector is not, of course, a new idea.As far back as 1990, the then Union Social Welfare Minister, Ram Vilas Paswan, floated the idea of job reservations in the private sector as a way of filling the gap between the aspirations of the backward classes and SCs and the available supply of government-controlled jobs.29 But the cries for the extension of reservation have grown more insistent since then as the number of government jobs, their status, salaries and perquisites have failed to keep up with the obviously booming private sector, especially given the very large number of backward class parties and politicians on which coalition governments in India now survive. The privatization of public sector units (PSUs) has been a particularly big flashpoint, prompting BSP leader, Mayawati for instance to make several forceful speeches in parliament in 2001 claiming that the privatization policy and reforms were “nothing but an attempt to deprive us from getting jobs.”30
The Congress government elected in 2004 appointed a committee (staffed with known supporters of reservations, such as Laloo Prasad Yadav and Ram Vilas Paswan) in August that year to look at the issue of whether and how private sector reservations might be implemented. But in the short term not much seems likely to happen, partly because of very substantial business resistance to reservations (the two main business federations, CBI and FICCI, both came out against formal private sector reservations) and partly because of larger questions about how such reservations would be implemented. In the near term, the most likely outcome would seem to be some voluntary affirmative action programs similar to those in some US companies, with requirements that companies doing substantial business with the government demonstrate that they employ significant numbers of SCs and OBCs. In the longer term, however, the issue looks sure to return, and has the potential to cause massive conflict between the largely upper caste-controlled business world and the increasingly OBC and SC-controlled world of politics.
One aspect of caste politics and conflicts that has not been extensively explored by sociologists is the extent to which claiming “backwardness” in the narrow context of reservations over a substantial period of time, and formally claiming kinship with other castes for instrumental political purposes, might have long-term effects on the way in which caste is practiced in other spheres, such as in social or market interactions. Srinivas, writing in the mid-1990s about the Vanniyar community in Tamil Nadu, implied that the effect of successfully claiming backwardness in a political context had only a minimal impact on other spheres, and that the community was “Janus-faced … claiming high caste status in a traditional context and a low one in the fierce struggle for access to scarce resources.”31 It seems to be the case that some communities which cooperate in caste politics are still at loggerheads in local disputes over land and local political power. But studies of intracaste conflicts in different spheres have been few and far between, so there is little firm information on whether cooperation in one sphere will ultimately reduce conflicts more generally in areas such as disputes over intercaste marriages, caste practices, or land.
At independence in August 1947, few would have predicted that India’s first few decades would be relatively free from communal conflict. Under the British, India had a system of communal reservations in politics and administration, and a system of “class recruitment” in the army whose effect (and its intention) was to accentuate communal divides and preserve their own rule.32 The partition itself led to the killing of perhaps 200,000 people and the mass migration of 13 million more, in a process that was to continue well into the 1950s.33 Congress itself was also vulnerable to pressure from the Hindu right, especially in the north and west, with both the powerful Hindu Mahasabha and Congress right wingers such as Purushottam Das Tandon pushing for a more supremacist policy towards members of the Muslim minority.34
But, under Nehru, the communal temperature was significantly lowered. The Muslim League was clearly a spent force, with much of the Muslim political elite having left for Pakistan.The Muslim social and economic elite left behind was largely broken by the zamindari reform of the 1950s and the loss of economic opportunities and discrimination so poignantly displayed in M. S. Sathyu’s 1973 film Garam Hava. Anti-cow slaughter legislation passed in most major states in the late 1940s and 1950s, as did legislation enshrining the status of Hindi written in devanagari script as an official language of India, taking both these important symbolic issues off the political agenda after decades of conflict. Further, pressure from the right wing diminished after the 1948 assassination of Mahatma Gandhi, which allowed the temporary ban of many organizations and, more importantly, cast a very negative light over assertive support for a Hindu right agenda within the Congress party.The death of Sardar Patel in 1950 also allowed Nehru to take stronger action against hardliners within congress, most notably in his 1951 power struggle with Congress President Purushottam Das Tandon over the exclusion of a prominent Congress Muslim, Rafi Ahmad Kidwai, from the Congress Working Committee. This standoff, in which Nehru threatened to resign, ultimately led to Tandon’s own resignation as Congress president in September 1951.35
That is not to suggest that everything was perfect with communal relations under Nehru. Periodic episodes of violence against Hindus in Pakistan led to refugee flows into India, sparking tit-for-tat violence in West Bengal on more than one occasion in the late 1940s and 1950s. In early 1950, for instance, more than 50 people were killed and 256 injured in West Bengal in strikes and riots that broke out in that state in protest at the death of perhaps 600 Hindus in Dacca, violence that was to be repeated again in 1964 in similar circumstances.36 Although Nehru was personally secular and demanded that Muslims be treated as full citizens, he could do little about day-to-day employment discrimination against Muslims in the states, despite urging chief ministers to address the issue in 1959 and again in 1961.37 The Muslim proportion in state police forces and administrations declined rapidly in the decade and a half after Partition, and much of the legislation and ordinances passed to protect Muslim educational interests, such as regulations on the provision of Urdu schools, or requirements that government servants communicate in Urdu with citizens under certain circumstances, were not enforced because of political opposition.38
Moreover, throughout the long period of Congress dominance in the post-Independence period, there were periodic attempts by politicians (sometimes Congress ones) to whip up communal issues and instigate violence for political or electoral advantage on one pretext or another: the 1956 riots over a book with an offensive biography of the Prophet Muhammad, which Nehru thought had been engineered to help Hindu parties in the upcoming elections;39 the anti-cow slaughter agitation in 1966 in New Delhi and elsewhere, designed to help the Jana Sangh and other communal parties in the run-up to the 1967 elections;40 the 1967 Ranchi-Hatia riots over Urdu, designed to destabilize the coalition government in Bihar; and the horrific 1969 riots in Ahmedabad, apparently instigated by the RSS and Jana Sangh.
In the short term, the decline of Congress from the mid-1960s seemed to many to be directly related to the rise in communal violence in India, which they blamed on the absence of the steady Nehruvian hand at the center, the decline of Congress as a party organization, the growth of caste and communal parties, and the increasing marginalization of congress in state politics in some areas. But ultimately, as I have suggested elsewhere, the decline of Congress and the growth of caste politics in the states was not, as if often viewed, a bad thing for Hindu–Muslim relations. It has, in fact, been helpful for communal relations in the long term in several important ways.41 It was in Kerala, lest we forget, that Congress first lost power (in 1957), and in which communalization of politics into definable caste and religious parties has been most advanced, and yet Kerala has had one of the best records, compared to other states, in preventing communal violence.
The growth of OBC parties has been a good thing for Hindu–Muslim relations in two ways. First, parties such as the Dravida Kazhagam in Tamil Nadu and the Communists in Kerala specifically included Muslims within their broad concepts of “Dravidian” or “back-ward” in order to build coalitions capable of challenging Congress. Second, and more importantly, the growth in each state of a larger number of effective political parties created a much more competitive environment for Muslim votes, in return for which Muslims could demand that states provide them better security. Even if a coalition did not, at the moment, need Muslim votes and Muslim-supported parties, a competitive environment in which there were five or six effective parties in a state made it very likely that it would need such support in the future, which gave it an incentive to protect Muslims. Why should Muslims, rather than militant Hindus, benefit from such increased competition along caste lines, and become pivotal swing voters in many states? First, because Muslim demands tended not to conflict with those of caste supporters of backward caste parties because Muslims are constitutionally banned from making effective claims for reservations on the grounds of religion equivalent to those made on the basis of caste by the OBCs and SCs;42 second, because Muslims placed a very high premium on the state providing physical security, a demand that was relatively cheap for Hindu politicians to supply, as long as they were not seen to be intervening too aggressively on behalf of Muslims.
The rise in political competition in the states as a consequence of the rise in OBC and SC parties—there are now an average of 4.4 effective parties competing in large states and the average level of electoral volatility (the seats changing hand at each election) has gone up from 20 percent in the 1957 election to 40 percent in the mid-1990s—has created a powerful incentive for state politicians to cultivate the large Muslim community, 13 percent of India’s total population but concentrated in particular states and especially in urban areas.
The changing political incentives in the states are absolutely critical because the Indian Constitution clearly makes local law and order the responsibility of the 28 state governments, not local or central governments.43 If a riot breaks out in a town or district, the army or central paramilitary forces may intervene only at the explicit invitation of the district magistrate or state government, even if there is a barracks just a few miles from the area in which the riot is taking place, as was the case for instance at Ranchi-Hatia in 1967. In theory, the central government can threaten to use its emergency constitutional powers and get rid of a state government that allows communal riots to take place. In practice, however, central governments only used this power five times between 1950 and 1996 over the issue of communal riots, despite the many large riots that took place over this period (e.g., Moradabad 1980,Ahmedabad 1969) and even then only in cases where the party in power in a state was not their own party, and where the center therefore had a clear electoral motivation for dismissing the state government. So, while the perceived threat of the imposition of President’s Rule can occasionally be useful— as Congress threats seem to have been in persuading the Modi government to quickly call in the army when riots broke out in Vadodara in May 2006, for instance—in practical terms the security of Muslims is largely dependent on state politics and the actions of the state government.44 This is even more true after the Supreme Court’s March 1994 Bommai judgment, which severely limits the freedom of the central government to impose President’s Rule in cases where there is not clear proof—subject to judicial review by the court—of the breakdown of the constitution. The effect of Bommai in restraining the center has also been magnified because the growth of central coalition governments in recent years in which many of the parties that were victims of the misuse of President’s Rule in the past are important participants, make them loath to sign off on any use of Article 356 outside Kashmir.
The crucial importance of the minority support base of the party in power together with the overall level of party competition in a state, in determining whether communal violence will be controlled or not, was tragically demonstrated during the massive riots that afflicted Gujarat in 2002. In Gujarat itself, the incumbent BJP government had no Muslim support, according to 1998 exit polls done by CSDS, and Gujarat also had very low levels of party competition, in what was basically a straight fight between the BJP and Congress. The Modi government, unconcerned about losing Muslim support and standing to gain all the support that fell away from Congress as a result of the riots, acted in a biased and partisan way throughout, even going so far as to transfer 27 officials for taking too aggressive a stance towards Hindu rioters.45
Outside Gujarat, though, the state political environments in 2002 were all favorable to controlling communal violence. Every state government in 2002 either relied heavily on Muslim voters directly, as for instance was the case in Madhya Pradesh and Rajasthan, or else was in a state that was very competitive in terms of overall levels of party competition.46 Bihar, for instance had 7.7 effective parties, Maharashtra had 5.64, Uttar Pradesh 4.99, and Tamil Nadu 4.84 (compared with 2.97 in Gujarat).47 These governments, therefore, had an enormous incentive to act strongly to prevent violence when the RSS, VHP and Bajrang Dal organized massive demonstrations, protests and strikes in the aftermath of the Godhra massacre of 57 Hindus in Gujarat on 27 February, 2002. In Gujarat, these demonstrations were a prelude to the pogroms of March and April. Outside Gujarat, however, owing to very decisive police action, including preventive arrests of thousands and, in some cases, deadly firing on rioters, large-scale anti-Muslim pogroms were completely avoided. In Rajasthan, for instance, a state adjacent to the violence in Gujarat, the state police force used deadly force to prevent riots from breaking out in Gangapur and Silwara.48 In Andhra Pradesh, the state police force successfully prevented riots from breaking out in the highly sensitive capital of Hyderabad, that city’s long history of communal riots notwithstanding, which one might have thought would predispose it to violence.49
None of this is meant to imply that communal relations in India are satisfactory. There has been a creeping communalization in many state administrations, with the growing display of Hindu symbols and fraternization of state servants with members of Hindu nationalist organizations such as the RSS and VHP.There have also been periodic attempts to rewrite school textbooks to accentuate the conflictual and anti-Muslim strands of Indian history rather than its more hopeful aspects: more Aurangzeb and temple destruction, in other words, and much less about Akbar and other rulers who employed many non-Muslims in their administrations and endowed temples across the land. And social and physical segregation are still realities in many places across India, with anti-Muslim prejudice and fears preventing many upwardly mobile members of the Muslim minority from obtaining housing outside of recognizable “ghetto” areas in the major cities.
But ultimately India’s strong caste, regional and linguistic cleavages, and above all the institutionalized nature of caste identities through the Constitution, reservations, and political parties have sharply undercut the likelihood of massive polarization along religious lines, despite the occasional terrible episodes such as Gujarat. The way in which strong lower and backward caste identities crosscut Hindu identities, and provide massive support for an overall “secular majority” in India is quite nicely demonstrated by recent survey data on support for majoritarian versus pluralist policies among Indian voters. In a large 2004–05 survey of the State of Democracy in South Asia, 5,389 Indians were asked if they agreed with the statement that “minorities should adopt the ways of life of the majority community.” Overall the good news is that there is a substantial pro-diversity majority among the Indian population, defined by the pollsters as the ratio of those who strongly disagreed with the statement compared to those who agreed with it.50 In India this “pro-diversity ratio” was 3.56. The corresponding ratio in Bangladesh, just for comparison, was 2.78 and in Pakistan a very depressing 0.60, indicating considerably more supporters of majoritarianism in that country than those who supported a more pluralist policy. Among Indian Hindus, though, the poll found considerable variation in terms of support for majoritarianism. Support is highest for a pro-majority policy among upper caste Hindus (prodiversity ratio of 1.79) and lowest among OBCs (2.77) and dalits (4.90).51 Thus, whether because of their own lower and backward caste ideologies, their association of Hindutva with upper castes, or the lack of tangible benefits that Hindutva supplies to them, the middle and lower castes seem to be strongly resisting majoritarian ideologies.
1 In Madras, for instance, the Hindu–Muslim divide before the late 1930s was crosscut by the divide between Brahmans and non-Brahmans; see Eugene Irschick, Tamil Nationalism in the 1930s (Madras: Cre-A, 1986). In Bihar, the Hindu–Muslim divide was crosscut in the same period by the conflict between Biharis and Bengalis over the latter community’s disproportionate share of government employment. See, e.g., the questions asked about Bengali overrepresentation in the police in Bihar: Legislative Assembly Debates Official Report, Vol. 4, No. 37 (26 April, 1937), pp. 2443–44.
2 André Béteille, Castes: Old and New (Bombay: Asia Publishing House, 1969), p. 50.
3 Some have argued that Hinduism does not work as a unifying identity in Indian politics because it is cross cut by other identities such as caste and language. It is certainly correct that religious identity in India is both diverse and is crosscut by multiple other identities, but this seems to me less important than the institutional benefits available to people who mobilize on the basis of caste rather than religion. In fact, all potential identities in India are crosscut in some way or the other, so intragroup differentiation is not a sufficiently good argument to explain why religion does not work but caste or language (both of which can offer tangible benefits) do, and, further, religion was, of course, highly salient as a unifying identity in Indian politics before 1947, when it was institutionalized by the British through separate elections and employment reservations.
4 This fundamentally political view of riots, in opposition to the more local and sociological view taken in Varshney (2002), is put forward in both Brass, The Production of Hindu–Muslim Violence in Contemporary India (Seattle: University of Washington Press, 2003) and in Wilkinson, Votes and Violence: Electoral Competition and Ethnic Riots in India (Cambridge: University Press, 2004).
5 Jaffrelot presentation on Hindu nationalism, Michigan State University, April 2003.
6 See the three volumes of the Concerned Citizens Tribunal—2002 for the most complete account of the violence, available at http://www.sabrang.com/tribunal/.
7 Sanjay Kumar, “Gujarat Assembly Elections 2002: Analyzing the Verdict,” Economic and Political Weekly,Vol. 38, No. 4 (25 January, 2003), pp. 270–75.
8 Harold Lasswell, Politics:Who Gets What,When, How (New York: McGraw-Hill, 1936).
9 Members of religious minorities within Congress generally supported reservations for their communities in employment and politics before Independence, although they opposed the separate electorates insisted on by the Muslim League. Opinion surveys conducted among Muslim legislators in the 1960s suggested that many were still in favor of separate electorates and employment reservations; Theodore P. Wright, “The Effectiveness of Muslim Representation in India,” in Donald E. Smith (ed.), South Asian Politics and Religion (Princeton, NJ: University Press, 1966), pp. 102–37.
10 Sarvepalli Gopal, Jawaharlal Nehru: An Anthology (Delhi: Oxford University Press, 1983), pp. 324–25.
11 S. Saraswathi, Minorities in Madras State: Group Interests in Modern Politics (Delhi: Impex India, 1974), p. 48.
12 Saraswathi, p. 49.
13 See Irschick, Tamil Nationalism in the 1930s; David Washbrook, The Emergence of Provincial Politics, The Madras Presidency 1870–1920 (Cambridge: University Press, 1976), pp. 271–87.
14 Irschick, pp. 36–37.
15 P. Radhakrishnan,“Backward Class Movements in Tamil Nadu,” in M. N. Srinivas, Caste: Its Twentieth Century Avatar (New Delhi: Viking, 1996), pp. 110–34, 120–22.
16 Jaffrelot, India’s Silent Revolution:The Rise of the Lower Castes in North India (New York: Columbia University Press, 2003).
17 Srinivas, p. xxviii.
18 “Racketeering in Quotas,” India Today, 15 November, 1994, pp. 36–42.
19 “Racketeering,” p. 37.
20 “Racketeering,” pp. 36–42.
21 A poll of 17,885 voters conducted 26 July–5 August, 2004.“Mood of the Nation-Poll,” India Today, international edn (New Delhi), August 30, 2004, pp. 16–23.
22 The category of “most backward class” was created in Madras in 1954 by Congress Chief Minister Kamaraj as a compromise way of addressing demands by backward class Dhobis and barbers that they be listed as scheduled castes, which would have entitled them to more benefits; Report of the Backward Classes Commission Tamil Nadu Volume I–1970 (Madras: Government of Tamil Nadu, 1974) (Chairman S. Sattanathan), p. 54.
23 “Uttar Pradesh: The Reservation Plank,” Frontline, 1–14 September, 2001.
24 M. Vijayanunni, “Caste and the Census of India,” unpublished paper, 2003, pp. 10–11.
25 Vijayanunni, p. 11
26 S. Ramakrishna, “Reservation Wars,” Indian Express, 20 June, 1998.
27 André Béteille, The Backward Classes in Contemporary India (New Delhi: Oxford University Press, 1992), pp. 108–10.
28 Ministry of Labor, Government of India, Census of Central Government Employees, 2001 (New Delhi: Ministry of Labour, Directorate General of Employment and Training, 2003).
29 “Reservation in Private Sector Likely: Paswan,” Times of India, 22 June, 1990.
30 “Mayawati Presses for Quota in Pvt. Sector,” The Hindu, 4 March, 2001
31 M. N. Srinivas, “Introduction,” in Srinivas, pp. vii–viii, ix–xxxviii.
32 See, for instance, the clear support for divide and rule in the quotes from Secretaries of State Wood (1862) and Hamilton (1897) in Sumit Sarkar, Modern India 1885–1947 (New Delhi: Macmillan, 1983) pp. 16, 21.
33 I follow the casualty estimate in Moon, Divide and Quit (London: Chatto & Windus, 1961), p. 269, and the refugee estimate in Keller, Uprooting and Social Change:The Role of Refugees in Development (Delhi: Manohar Book Service, 1975), p. 17.
34 For details of this pressure from the right, both inside and outside Congress, see William Gould, Hindu Nationalism in Late Colonial India (Cambridge: University Press, 2005); and Mukul Kesavan, “Invoking a Majority: the Congress and the Muslims of the United Provinces, 1945–47,” Islam and the Modern Age, Vol. 24, No. 2 (1993), pp. 109–30.
35 Jaffrelot, p. 101.
36 Times of India, 24 February, 1950; 15 January, 1964.
37 G. Parthasarathi (ed.), Jawaharlal Nehru: Letters to Chief Ministers 1947–1964, vol. 5, 1958–1964 (New Delhi: Oxford University Press, 1989), pp. 233–46, 427–32, 446–59.
38 See Wilkinson, Votes and Violence, ch. 4.
39 Jaffrelot, The Hindu Nationalist Movement in India (New York: Columbia University Press, 1996), p. 107, citing N. L. Gupta, p. 249.
40 Jaffrelot, pp. 205–10.
41 Wilkinson, Votes and Violence.
42 Politicians in a few states, such as Tamil Nadu, Andhra Pradesh and Kerala, have tried to get around this ban for at least some Muslims by reserving jobs and places in education for certain Muslim jatis on the basis of caste rather than religion.
43 In India, with only a few exceptions, all local law and order is controlled by the state police.
44 “Army Deployed in Vadodara,” The Statesman Weekly, 98, 18 (6 May, 2006).
45 “Modi ties hands of cops who put their foot down,” Indian Express, New Delhi, March 26, 2002, p. 1.
46 The argument here follows that in Wilkinson, Votes and Violence (2004) and Wilkinson, “Putting Gujarat in Perspective,” Economic and Political Weekly, 27 April, 2002, pp. 1,579–83.
47 The effective number of parties measure (ENPV), widely used in political science, allows us to get a measure of the overall level of party competition in a state using a measure that underweights the vote share of the many small parties that secure little support in an election. In 2002 the effective number of parties in major states ranged from 2.78 in Andhra Pradesh (where the two largest parties had 85 percent of the vote) to 7.7 in Bihar (where the two largest parties had only 43 percent of the vote). For more details see Wilkinson, Votes and Violence, ch. 5.
48 Rajasthan Patrika, 27 March, 2002; “Police Firing in Rajasthan: Two Killed,” Indian Express, 26 March, 2002.
49 “Andhra Police on High Alert,” Indian Express, 26 March, 2002.
50 Centre for the Study of Developing Societies, State of Democracy in South Asia: A Report (Oxford: University Press, 2008), Table 5.5, pp. 261–62.
51 The minority communities, not surprisingly, strongly oppose majoritarianism, with prodiversity ratios of Sikhs (5.70), Muslims (16.25), and Christians (11.30).