Response to Ashutosh Varshney
By Paul R. Brass
Ashutosh Varshney has written a hostile and unprofessional review of my new book on The Production of Hindu-Muslim Violence in Contemporary India, originally published by University of Washington Press in February 2003 and issued by Oxford University Press-India in September-October of this year. The review, published in the10 November issue of India Today, is so inaccurate and dishonest that it is difficult for me to know where to begin to rebut it. Varshney does not even take the trouble to summarize the book, but merely picks out and misrepresents at random aspects of my arguments.
Varshney begins by insinuating that I have spent 40 years of my life studying one city, Aligarh, and suggests that I have produced nothing of value from my labors. While it would be unseemly of me to write about my own professional accomplishments in my work on India, I believe it is well enough known among scholars, journalists, and politically knowledgeable people in India that I have written widely on many aspects of the politics of India over these years, and some may know that I have published some 14 books on those subjects as well as rather numerous articles. My works have ranged from detailed studies at the local level to works that cover politics in all India, including my text on The Politics of India Since Independence, the second edition of which is still available. I have personally carried out field work, during approximately 25 visits to India, in the states of Uttar Pradesh, Bihar, Punjab, Tamil Nadu, Gujarat, and Assam. I have also interviewed many politicians from all other parts of India during my visits to Delhi.
It is true, however, that I have labored hard and long, including for a good part of the six years between 1997 and 2003 in Seattle, poring over my interviews, documents, and other data collected over four decades in Aligarh, to ensure that I minimize the possibility of mistake on so serious a matter as Hindu-Muslim violence. Nor, indeed, despite Varshney’s sneering remarks, has he been able to point to a single error.
Having found no inaccuracies, Varshney seeks to undermine my arguments in a personally insulting way. He claims that I have simply “recycled” my “old arguments” from two books of mine that are well-known in India and elsewhere, Theft of an Idol and Riots and Pogroms. Varshney himself wrote an extremely laudatory review of the latter book (for the Journal of Asian Studies, published in February, 1999), in which he expressed his “admiration for the superb contribution by Brass” and praised “the great merit and compelling brilliance of his reasoning” (p. 133). In the same review, he made laudatory comments on Theft of an Idol. Evidently, something has changed in Varshney’s attitudes, on which I will comment below, but it has nothing to do with the quality of my work or its arguments. It cannot be so since Varshney has also made considerable (mis)use in his own writings of my central argument that the best explanation for the persistence of riots in sites where they appear to be endemic—such as Aligarh, many other cities and towns in India, and many other places around the world at different times, including the twentieth century U.S. and nineteenth century Russia—is the existence of what I have labelled Institutionalized Riot Systems. Varshney has completely misread my description of such systems in his own work, as well as in the India Today review, imagining that all that is meant by the concept is that politicians and criminals protected by them, “especially the Hindu nationalists,” are involved in riots and “keep the communal pot boiling.” He again strikes a derisory note by calling his misunderstanding of my construct “a boiling-pot theory.” This is quite a travesty of my conception, which is that Institutionalized Riot Systems are composed of networks of specialists who play varied and multifarious roles in the instigation and perpetuation of communal animosities, in the enactment of riots, and in the interpretation of riots after they occur.
The metaphor I have used is, as far as I know, quite different from anything anyone else has used in the study of collective violence, namely, the conceptualization of riot production as comparable to that of a grisly theatrical drama, in which there are three phases: preparation/rehearsal, performance/enactment, and interpretation/explanation. This is not a trivial one-off comment on riots, a “boiling-pot theory,” but an elaborate analogy of a type that should be familiar to anthropologists and others who know the work of the great anthropologist, Victor Turner, particularly his Dramas, Fields, and Metaphors.
In his own work on peaceful cities and towns in India, Varshney copies my argument by inversion as it were, claiming that they have “institutionalized peace systems.” However, his use of both terms, mine and his inversion of it, lacks logic, precision, and a basis in worthwhile empirical data. But, not content to invert my argument, he has been reported in the India International Center Diary (Janauary-February 1999) to have presented, at a talk at the Centre, my original argument (incorrectly as usual) as if it were his own invention. Perhaps the Centre journal has misunderstood him, but no contradiction of his use of my concept as his own has yet appeared.
At one point in the review, he makes a tortuous move from his misreading of my argument to a statement that it is “historically inaccurate” because “Hindu nationalists were not prominent in Aligarh before 1967.” Here he is trying to insinuate that I am misleadingly emphasizing the important role played by what he calls “Hindu nationalists” in producing riots in Aligarh. He then cites various electoral statistics to say that this cannot be accurate because “the Hindu nationalists were not prominent in Aligarh before 1967.” These electoral statistics are quite beside the point here. The plain facts are that, though many communal riots in Aligarh and elsewhere in India have involved persons and parties not part of the Sangh parivar, militant Hindus have played a central role in every single large-scale riot in Aligarh at least since 1961, however electorally strong or weak they were, and my book demonstrates it very clearly.
Varshney here is acting out his own role in the communal discourse in India, namely, that of the BJP/RSS apologist who, though he is not himself a member of the Sangh parivar, chooses to ignore their undeniably central role in rehearsing, enacting, and interpreting communal riots after the fact. His statement that he agrees with me—in his words not mine—“that Hindu nationalism is a dangerous project and if it succeeds it will destroy India” is nothing but a pious, throw-away line for a person whose work virtually frees the BJP and the RSS from responsibility for the production of riots. For example, in his own book, Ethnic Conflict and Civic Life, the index contains only three references to the RSS and the VHP, of which only one includes a very brief description of these organizations, from which one learns that their “activities … include running ideological camps for the youth, schools and dispensaries for the tribals and scheduled castes, and organizations for women.” We learn that their activists also do “relief work”at times of heavy floods. The Bajrang Dal, the principal organization for violence in the Sangh parivar, receives no mention at all.
Varshney’s review also mixes together incomprehensibly some questions and answers that are unrelated to each other, as if they undermined my arguments when they have nothing to do with them. He asks, “Why was the Congress government, in its days of nearly uncontested hegemony, unable to prevent riots?” He then answers sarcastically: “Presumably, the DMs and the SPs who, according to Brass, had been instrumental in Aligarh’s intermittent stretches of peace, were not so compromised at that time.” This is all totally misleading. Most governments in India, including those of the Congress and the BJP, for that matter, have been able to prevent, contain, and control riots when they chose to do so. Nor have I said that Aligarh has had “intermittent stretches of peace.” Rather, there have been intermittent periods when large-scale riots did not occur, during which the riot network was only partially inactivated, but kept in readiness.
But then, somehow, Varshney has a different answer to his questions, totally irrelevant to them, but important to understanding the malicious character of his review. He says: “Commenting on the Aligarh of the 1950s, Nehru was forced to call attention to the rebirth of Muslim communalism at the AMU. …Brass neglects the role of Muslim communalism in the city.” This is dishonest, mendacious, and motivated. In fact, I have not neglected the role of Muslim communalism in the city. I have also drawn attention to the contribution of elements in the Muslim community, including politicians, University personnel, and AMU students in maintaining communal attitudes and in participation in riots as well. However, there is simply no doubt in my mind, amply demonstrated in my work, that the BJP/RSS has been far more deeply implicated—perhaps because it is far better organized than the Muslim network—most especially during the decade of the 1980s up to, and including especially, the great Aligarh riots of 1990-91. Varshney is here simply avoiding my main conceptual arguments concerning the process of riot production, throwing up a false statement against me and pitting me against Nehru himself in the process. Moreover, Varshney is here doing what the BJP/RSS people do: blaming the AMU, which has rarely been at the physical center of Hindu-Muslim violence, though it has often been targeted by militant Hindus and has been generally used as a justification for violence against Muslims.
Varshney is here also showing his ignorance of the political geography of Aligarh, though he has a chapter on Aligarh in his own book. I have emphasized, in my book, the very sharp separation between Muslims and Muslim politics at the AMU and Muslims and Muslim politics in the old city. There have been some forms of border-crossing, as it were, but, historically, riots have been produced in the old city where there is a juxtaposition of Hindu and Muslim mohallas, not at the AMU. In contrast, in 1990-91, the militant Hindu riot system extended its range dramatically across the boundary of the Grand Trunk road and the railway line and all around the outskirts of the city in a pattern that has been revealed by me and others elsewhere, in Kanpur (by me), in Bombay by many other scholars, in Gujarat by most commentators, and so forth.
The second argument Varshey criticizes, headlined in the India Today review as “Aligarh is not India,” concerns the generalizability of my arguments. He quotes me correctly as follows: “The findings herein can be generalized to other parts of India and other times and places in the world.” (This quotation comes from the Preface, however, not from the heart of the book where the arguments are presented in full.) He then asserts falsely that I have ignored places in India where riots have not occurred. My book indeed centers on Aligarh, though my work on riots has extended throughout north India and the Punjab in interviews, and throughout the rest of the country in my reading of both primary and secondary sources. My argument here is not that Aligarh stands for or represents all of India, which is nonsense, but that the pattern that I have described for Aligarh applies to other cities and towns in India that I know well from my own personal research. Moreover, I have presented my argument as a social science hypothesis for other scholars to test in other places in other parts of the world. Far from being an old argument recycled, my argument needs testing elsewhere. Such testing would not prove or disprove what I have described and discovered in Aligarh. But, insofar as its generalizability is concerned, this is an important question that begs for further research. For, if I am right, then most research on, and explanations for, riots, pogroms, massacres, and some genocides as well, have been not only wrong, not only false, but misleading and contributory to the perpetuation of systems of violence.
Now, let me answer specificallyVarshney’s question. Anyone, however, who cares to read my book can find the argument laid out carefully there in 476 pages. “Given … variations [from place to place in India in riot incidence], how can Aligarh’s experience be generalised to Uttar Pradesh, let alone the rest of India?” The answer is simple: By testing my hypotheses. First, by the method of confirmation/disconfirmation, that is, by examining sites of endemic rioting to see if institutionalized riot systems can or cannot be discovered. Second, by examining the relationship between party/electoral competition and the incidence of riots in those sites. Third, by examining the consequences of different state policies toward communal riots, my argument being that where the policy of a state government is decisively opposed to communal riots and makes its opposition clear, and where interparty and/or intraparty divisions do not compromise its clarity, riots will be either prevented or contained rapidly. The recent work of my young colleague, Steven Wilkinson, confirms several of my arguments. Wilkinson has also previously questioned parts of my argument, but in an honest, forthright manner, concentrating on the issues at stake. Our discussions have, I think, influenced each other’s work. Moreover, in discussions with him, I believe our mutual work is coming close to a coherent explanation of riot production, though we may still disagree on some aspects of the process. Such, however, is not the case with Varshney’s work on civic engagement, which is a derivative argument from the American social science literature that has very little to do with India. It is a false and artificial transplant, which I have criticized in my book and need not repeat here.
As for the alleged contradictions in my criticism of newspaper reporting on riots in India while also making use of such reports, his disparagement is also totally misleading. My accounts of riots are based heavily upon my own interviews and other primary sources. Where that has been lacking or inadequate, I have used newspaper reports in a careful and critical manner, pointing out where they appear reliable, where not, where biased. I have also criticized sharply Varshney’s uncritical use of newspaper accounts of the precipitants and alleged causes of riots. Moreover, I have noted that his highly touted dataset, based solely on Times of India reporting, is inherently flawed. Furthermore, errors were introduced in coding this flawed data. An huge error was introduced, for example, into the Aligarh data, to which I alerted him through Wilkinson, which Varshney then corrected in his World Politics article with no acknowledgment to me. In short, his own data on Aligarh, on which he claims to have done research, was false.
Then there is the charge concerning my so-called “intellectual schizophrenia.” I suffer from no such ailment. I laid bare my own reasoning concerning riot production in India in this book and in my other recent works on India, and expressed my profound doubts about the enterprise of causal reasoning and analysis as it is conducted in contemporary social science. In his comments on my previous book, Theft of an Idol, Varshney wrote as follows: “Whether or not one can agree with Brass about causality, the great merit and compelling brilliance of his reasoning lies in showing so effectively why the battle over meanings matters, why such battles are as much about knowledge as about power and resources. In doing so, Brass, in this essay [in Riots and Pogroms] as well as in his recent book Theft of an Idol, forces us to re-evaluate the easy certitudes of mainstream social science, if not abandon social science altogether.” Evidently, Varshney has changed his mind about my reasoning.
As for my use of “correlation coefficients,” which he says implicates my work in “mainstream social science,” this is hardly an advanced social science method of causal analysis. It is one of the simplest numerical methods available for establishing associations between variables, from which causal analyses may or may not legitimately be inferred. I have always tended to use such elementary statistical techniques mainly to demonstrate such relationships and suggest the direction of a causal chain where it seems reasonable to say so, but I have mostly used such techniques as supplements to my own type of processual analysis. I have been described by friendlier colleagues as a “closet positivist.” I accept such a friendly statement. But intellectual doubts about the relative merits and utilities of positivist/empiricist vs. other types of social science, historical, and anthropological research hardly constitute “intellectual schizophrenia.”
The most degrading half-sentence in Varshney’s review is his reference to my dedication of the book to Myron Weiner, implying that my work is not consistent with Weiner’s and that the dedication, therefore, is misplaced. I have noted there and elsewhere my debts to Myron, my respect and affection for him, as well as my divergence from his approach. I worked with Varshney on a festschrift for Myron, held at Notre Dame in 1999. It is from that failed collaboration with Varshney that a personal hostility has embittered and has terminated our relationship. I had ultimately to withdraw from editorial collaboration with Varshney on the publication of the conference papers because of his ugly misuse of the occasion to aggrandize himself, advance his own career, prevent other worthy former students of Myron from attending or presenting papers at the conference while ingratiating himself with senior colleagues whom he had previously antagonized, badgering Myron during the last days of his life into allowing him to invite to the conference a person whom Myron strongly disliked, and ultimately disregarding scholarly standards in his attempt to publish the papers from that conference. The volume has not yet been produced.
November 30, 2003
Paul R. Brass
Professor Emeritus of Political Science
and South Asian Studies
University of Washington, Seattle