Communal Violence in India - Perspectives on Causative Factors (Part 1) (Roshni Sengupta)

[The Economic and Political Weekly - May 14, 2005 ]

Communal Violence in India - Perspectives on Causative Factors

This paper examines recent theories advanced by political scientists to explain the occurrence of communal violence in India. Thus blame for incitement could lie in the persistent hate campaign waged by right wing organisations. The degree of capacity and autonomy vested in a state government also determines its capability to resist a riot situation. Further the degree of communal amity also depends on the bonds of civic engagement and associational activity that exist between communities. While causative factors could differ, it is clear that compulsions of electoral politics and the increasing anonymity and disparity of urban life are playing a greater role in communal conflagrations in post-independent India.

Roshni Sengupta

Even as riots raged in Gujarat for most of 2002, myths were circulated, accusations were levelled, all manner of explanations and counter-explanations were advanced about how members of the minority community had ‘started’ the riot in the first place by sending the train up in flames.1 The subsequent murder and mayhem was then categorised by the Gujarat government (and the central government as well as seen by their silence on the pogrom) as an ‘equal and opposite reaction’ to the train-burning incident at Godhra.

As reports suggest, various fears were expressed by the BJP government and its Sangh2  affiliates about the ‘lesson’ that needed to be taught to the Muslims failing which they would have taken over the state, both in numbers as well as power. Further the ‘Pakistani agents’ who thronged the streets of Ahmedabad and Vadodara had to be shown their place. But such politico-religious reasons given by those part of the governmental machinery in the state and the countrywide network of the Sangh parivar failed to convince anyone. After all, they smacked of a lack of logic and compounded obduracy.

The carnage, however, would lead a number of scholars, both young and old, to re-examine the existing explanations of the various causative factors of religious conflagrations. Paul Brass terms these explications as ‘unsatisfactory’ and ‘mystifying’ in his work on Hindu-Muslim violence in contemporary India [Brass 2003]. Naturalising accounts of riots have invariably portrayed communal riots as inevitable eruptions of anger and violence between communities divided by deep, incommensurable and often historical differences.

This paper makes an attempt to take a broad but detailed account of these explanations and the theorists credited with advancing them. A study of the major authoritative works done by Paul Brass, Ashutosh Varshney, Imtiaz Ahmad and Steven Wilkinson leads to a most varied and differentiated mass of thought and scholarship which could go a long way in putting the issue of Hindu-Muslim violence in perspective.

The veracity with which the Gujarat state government pushed for early elections in the aftermath of the worst communal violence since independence startled both scholars as well as media-watchers who had been witness, even though as were distant and mute spectators, to the horrors that unfolded across the state in the aftermath of the Godhra incident.3  The dissolving of the state assembly only confirmed what some had feared the violence was possibly a result of the insecurity of the BJP after weighed down by defeat at the municipal and panchayat elections in 2001.

Dissecting Gujarat

Steven Wilkinson, in the seminal ‘Putting Gujarat in Perspective’ (EPW, April 2002) deconstructs the entire debate on state complicity by delineating a wider shift in Indian politics. The state government’s delay in calling in the army established beyond doubt that the anti-minority pogrom was being carried out with the involvement of not only the politicians but the state machinery as well.

As in most riots in the country before and after independence, minorities suffered disproportionately in Gujarat from February 27, 2002 to the beginning of April 2002.4  Minorities Commission figures on the communal riots that took place from 1985-87 stand testimony to the fact that in almost every bout of communal violence, the minorities suffer most.5 Given the fact that the state governments are responsible for law and order in India, the question why some state governments are successful in controlling riot situations and preventing them from occurring and others are not is not only relevant but also vexing.

Wilkinson argues that ethnic riots6  are far from being spontaneous eruptions of anger. Instead they are often planned by politicians for a clear electoral purpose. Subsequently, it then follows that these very politicians will also prevent riots if and when it is in their interest to do so. These violent conflagrations are, therefore, caused by political elites who play on existing communal tensions to advance a political agenda [Wilkinson 2004].

Instrumental political explanations for violence have been labelled ‘unsatisfactory’ not only by Wilkinson, but also by other theorists. Wilkinson lends credence to his criticism of instrumental explanations by invoking the actions of certain state governments. The first point of refutation that he makes is that scholars who look at political elites and their reasons for inciting violence offer little insight into why some politicians tend to do exactly the opposite and use their political capital and control of the state to prevent ethnic conflict (Wilkinson ibid, Chapter 1). The second major problem identified by Wilkinson, many political explanations for ethnic violence often fail to account for the variation in patterns of violence within states.

In order to clear any confusion arising out of the criticism made against existing theories, Wilkinson posits three ‘possible’ explanations [Wilkinson 2002] for the differences in state performance, which will be elaborated upon in the course of this paper. First, decades of corruption, criminalisation, politicisation and a general lack of state capacity have left Indian state governments too weak to prevent riots. Second, Indian state governments are unable or unwilling to protect minorities because they systematically under-represent them within their governments, police forces and local administrations. Lastly and most importantly, the degree of party competition affects the value, governments place on attracting ‘Muslim swing voters’, which in turn affects whether or not the government will order the respective administrations to protect the minorities.

Since Wilkinson’s scholarship primarily deals with state action in fomenting or preventing ethnic violence between Hindus and Muslims in accordance with incentives offered by both communities electorally, it is imperative to take a detailed account of the arguments that he posits in this regard. He focuses on state and town level electoral incentives which remain important even if we assume various other factors, socio-economic and otherwise, to be constant and controlled.

At the local level, politicians would try to ensure that the identity that favours their party is the one that is most salient in the minds of a majority of voters [Wilkinson 2004:2]. Those parties that represent elites within ethnic groups will invariably use polarising anti-minority events in order to encourage the members of their ethnic category to identify with their party and the ‘majority’ identity.

The most effective method, according to Wilkinson, for elite dominated ethnic parties to mobilise those target voters who are at risk of voting for the main rival parties will be to use ‘ethnic wedge issues’. This is summarily accomplished by highlighting the ‘Muslim threat’ especially in urban areas7  where the party hopes to win over the pivotal Hindu voters.

The form of anti-minority mobilisation taken recourse to depends on both the identity which the party wants to make salient and also the fact that the Indian state, similar to other states, privileges some form of public ritual or procession which culminates in immense mobilisation.8 In fact, C A Bayly shows that riots occurred because of, or in the wake of, religious processions in the late 18th and early 19th centuries. Scholars like Stanley Tambiah have also noted that processions often degenerate into rioting as a result of the manipulation by leaders who wish to bring about some form of mobilisation in their favour [in Christophe Jaffrelot 2001].

Defensive counter-mobilisation by the minorities is portrayed as ‘anti-national’ and the myth of the ‘foreign hand’ is often invoked. When large crowds face each other, the threat of the situation deteriorating increases which is then construed to be the handiwork of the Muslims alone.
Elections and Riots

Conditions such as these produces ‘community consciousness’ that translates into a wave for or against the party which further leads to a swing that either reaps rewards or brickbats. Wilkinson (2004) argues that to win an election it is not necessary to appeal to each and every voter but to the pivotal swing voters who generally are those undecided voters who fear the consequences of not taking a defensive stand against the members of the other community.

Statistical evidence suggests that proximity to an election sharply increases the likelihood of a riot [Wilkinson 2004]. Factors such as economic competition, Muslim population and percentage of refugees from Pakistan are not numerically significant in explaining the occurrence of a Hindu-Muslim riot even though computation shows that as the Hindu-Muslim balance of a town reaches 50-50, the possibility of a riot goes up a few notches.

Three kinds of situations may develop which prove that as electoral competition increases, the level of riots goes down. First, the existence of three or more parties provides a security blanket to the minorities as the importance of swing votes increases provided the majoritarian party is not trying to attract the fringe Hindu votes. A bipolar state party system creates a potentially dangerous situation for the minorities, especially when the majoritarian party, which owns the anti-minority issues, tries to foment violence.

The third situation can be exemplified by Gujarat and the events that unfolded in that state in 2002. Disaster awaits the minority community, when the anti-minority, majoritarian party is in power. The anti-minority government will prevent violence only when the destruction threatens to wean away the loyal voters. Any potential advantage to the government will not be sacrificed in order to protect the minorities.

The number of elections that have been preceded by communal killings has taken an upswing over the years, especially since the Hindu nationalist BJP gained political prominence. If the L K Advani9 led rath yatra (chariot procession) that traversed most parts of the country in 1989 left a trail of blood in its wake, the mammoth saffron wave that swept India in election after election in the 1990s proves beyond doubt the fact that Hindu-Muslim riots increases the likelihood of the BJP improving its electoral performance.

The above argument refutes the claim made by Ashutosh Varshney in his work on ethnic violence (2002) in which he almost absolves politicians, the strategic roles played by them in fomenting violence and the Sangh parivar that has been associated with most post-independence Hindu-Muslim conflict by stressing instead on the existence or absence of civic ties between members of the two communities as the major factor that either leads to violence or to peace.

Electoral incentives are the prime movers of an ethnic riot. Wilkinson, however, further investigates the reaction of a state government and its administrative machinery to communal violence, whether in controlling the conflagration or letting the fires burn. The two major indices that are used for this purpose are – state autonomy10 and state capacity.11 

Wilkinson’s central argument is that state weakness does not account for interstate differences in the level of Hindu-Muslim violence. He finds that the relationship between state autonomy (the lack of political interference or otherwise) and state capacity is inversely proportional to variations in occurrence of Hindu-Muslim riots. States that are said to be the lowest on the autonomy and capacity barometer have done remarkably well to prevent riots. Even the weakest state governments, like Bihar and Uttar Pradesh, still seem to possess the minimal state capacity necessary to prevent Hindu-Muslim riots if this is prioritised by the state’s political leaders.

The mere fact that there are sharp state level variations in the occurrence as well as the prevention and control of riots suggest that the problem is not so much state capacity as the instructions given by politicians to state officials telling them whether to protect or not to protect minorities. Certain specific aspects however are linked to state capacity and their poor performance in preventing riots.

The financial weakness of some state governments can be a cause for concern. Interestingly enough, the Gujarat state government was hardly cash-strapped when the pogrom took place in the state. Secondly, the police and judicial systems in many states are understaffed and overloaded, which reduces the perceived risk rioters face of arrest, prosecution and conviction.

Punitive transfers also have an independent negative effect on riot preparedness, because frequent transfers reduce officers’ knowledge about their districts, the potential trouble spots and the one best way to prevent a riot. Political interference with state autonomy is alleged to help increase Hindu-Muslim violence in different ways. Instructions from the political bosses to either drop or go slow in investigating cases of mob brutality and murder by certain influential members of the citizenry or threatening political retribution, unless they delay in taking stern action especially against groups that enjoy state protection are the common ways in which state autonomy is eroded and in some cases demolished.

Obviously enough, the central thrust of the work done by Steven Wilkinson traces the linkages between party competition and ethnic violence but at the same time he finds that high levels of electoral competition can reduce as well as precipitate ethnic violence. Both situations can be reproduced in equation form in the following manner:

Party competition ( ? ) + Muslim swing votes ( ?) = (Ø ) Levels of violence
Party competition (Ø) + Muslim swing votes (Ø) = ( ?) Levels of violence

Forms of Engagement

That the existence or otherwise of civic ties between Hindus and Muslims at the town level is the primary cause of conflict or the absence of this between the two is the central assertion made by Ashutosh Varshney in his work on the subject (2002).12 Where such networks of civic engagement exist, tensions and conflicts are regulated and managed; where they are missing, communal identities lead to endemic and ghastly violence.

These networks can be broken down into: associational forms13 of engagement and everyday forms14  of engagement. Both forms of civic engagement, if intercommunal, promote peace. The capacity, however, of the associational forms to withstand national level exogenous shocks is substantially higher. Varshney argues that associational forms turn out to be sturdier in the face of politicians trying to foment communal trouble. Vigorous associational life acts as a serious constraint on the polarising strategies of political elites. The mechanisms that connect civil society (non-state domain) and ethnic conflict can be broadly classified into two categories [Varshney 2002:9].

By promoting communication between members of different religious communities, civic networks often make neighbourhood peace possible. People come together routinely to form temporary organisations in the face of tensions. These can be highly effective, and are known as ‘peace committees’. Such organised bodies are difficult in those urban concentrations where civic engagement between Hindus and Muslims does not exist. The second mechanism describes why associational forms of engagement are sturdier than everyday forms in dealing with ethnic and communal tensions. Vibrant organisations serving the economic, cultural and social needs of the two communities can promote communal peace, which can be solidly expressed. Varshney, however, presents a profound paradox. Everyday engagement is so complete in the village level in India that associational forms of engagement are few and far between. Yet rural India has experienced fewer riots since independence. In contrast, even though associational life flourishes in the cities, even petty rumours can cause deadly bouts of communal violence.

The argument fails to stand its ground once the ‘anonymity’ argument is put forward. Cities tend to be less interconnected and more anonymous. Size can reduce the extent and effectiveness of everyday interaction. This explanation leads us to the two other major findings that makes its appearance in Varshney’s work. First, the share of villages in communal rioting is remarkably small. According to Varshney, Hindu-Muslim violence is primarily an urban phenomenon.

The violence in Gujarat can be termed unique in several ways including the occurrence of large-scale communal violence in rural areas. The adivasi dominated Dangs region and Sabarkantha district saw large-scale violence in the rural belt; this was, however, much less that what was seen in urban centres like Ahmedabad and Vadodara.
Riot-prone Cities

The second argument that Varshney makes as far as town-level variation is concerned, is somewhat more problematic than the first one. He differentiates between ‘riot-prone cities’ and others, (he identifies eight cities) – Ahmedabad, Mumbai, Aligarh, Hyderabad, Meerut, Vadodara, Kolkata and Delhi – to be particularly high on the riot scale.15 Eighty two per cent of the urban population therefore, is not riot-prone.

India’s Hindu-Muslim violence therefore, is city specific. State and national level politics provides the context within which the local mechanisms linked with violence are activated. Cities are also the sites for large-scale civic engagement, which constrains local politicians in their strategic behaviour. Riot-prone cities lack such forms of engagement that leads to the political elites taking advantage of the volatility of the situation.

In ‘peaceful’ cities, an institutionalised peace system exists, where organisations are communally integrated. These civic organisations, for all practical purposes become the ears and arms of the local administration. It then follows that if the civic edifice is interethnic and associational, it can take on ‘ethnic earthquakes’ such as a partition and desecration of a holy place. Contrarily, if the form of civic engagement is intraethnic in everyday life, earthquakes of smaller intensity can bring the edifice down [Varshney 2002:11].

Varshney, thus states that a multiethnic society with few interconnections across ethnic boundaries is very vulnerable to ethnic disorders and violence. In Hyderabad city, for instance, most Hindus and Muslims do not meet in a civic setting where mutual relations can be formed. Lacking these networks even competent police officers and administrators watch a riot unfolding helplessly.

The emphasis on civic engagement as the mechanism that either foments or controls communal violence tends to displace focus from the role played by the Hindu nationalist Sangh parivar in engineering riots and pogroms since independence. Paul Brass16  is critical of Varshney’s ‘blame displacing’ theorisation and the fact that he identifies the Gujarat violence as the first of its kind in India.

[... continued in subsequent post]