REVIEW SYMPOSIUM

Copyright © 2006 SAGE Publications (London, Thousand Oaks, CA and New Delhi) 1468-7968
Vol 6(1): 102–117;062238
DOI:10.1177/1468796806062238
http://etn.sagepub.com

Review Symposium: The Production of

Hindu–Muslim Violence in Contemporary
India by Paul R. Brass (published May 2005,
pbk)

Politics by other means

THOMAS BLOM HANSEN
Yale University

Paul Brass’ latest book The Production of Hindu–Muslim Violence in

Contemporary India (2003) is an extraordinary work that sums up almost
40 years of research on politics, religious identities and violence in northern
India. Focusing on the politics of Hindu–Muslim relations in one city in
Uttar Pradesh – Aligarh – over the entire post-Independence era, Brass
argues that riots are permanent features of Indian politics, produced and
staged by ‘institutionalized riot systems’. Condemning and bemoaning riots
and casualties have become part of India’s modern political culture – as
much as the riots themselves.
Brass’s introductory chapter takes aim at what he sees as the unsatisfac-
tory and, ultimately, mystifying explanations that have been advanced in
explaining riots. ‘Naturalizing’ accounts portray riots as inevitable eruptions
of anger between communities divided by deep and incommensurable
differences. Others view riots as pathologies of Indian political life, result-
ing from the cynical manipulation of religious passions by criminal business
people and ill-intentioned politicians focused on short-term electoral gains.
These explanations, Brass argues, not only obscure the processes at work;
they are complicit in the very regime of interpretation that perpetuates
riots. Portraying them as either ‘justified’ or as short lived ‘aberrations’,
these explanations fail to recognize the integral, normalized roles riots play
in political competition and communal organization in large parts of India.
 REVIEW SYMPOSIUM 103

With its prosperous Hindu bania (trader) communities and substantial

Muslim artisan population, Aligarh is a typical north Indian city. At the
same time, the presence of India’s premier Muslim institution, Aligarh
Muslim University, the deep and enduring communal polarization in the
city, and the early alliances between Muslims and ex-untouchables make the
city special. Unlike Ashutosh Varshney in his recent work (Varshney, 2002),
Brass rejects the official classification of certain cities as ‘riot-prone’. The
Production of Hindu–Muslim Violence in Contemporary India emerges as
a welcome rejoinder to Varshney’s influential and overly schematic analysis,
where communal riots result from the absence of civic ties across com-
munities. In explaining specific riots, Brass obviously finds analytical distinc-
tions between ‘politics’, ‘civil society’, and ‘the state’ less plausible than the
detailed stories of individual careers and sociopolitical networks in the
city’s neighbourhoods.
Riots are best understood as the results of actions by identifiable
specialists and networks of specialists in ‘riot-production’: the systematic
rehearsal, staging and interpretation of collective violence as spontaneous
acts of self-defence or retaliation against what is perceived as unjust and
murderous threats from the other community. Understood in this way,
Brass sensibly concludes, riots can be policed and prevented more effec-
tively. How does Brass reach his conclusion? How does his conception of
riots as complex localized productions compare to other explanations of
riots in South Asia?
The evidence presented in the book is comprehensive and represents,
thus far, the most systematic exploration of ‘riot production’ in India (and
possibly anywhere).
Brass presents the context and development of a sequence of riots in
Aligarh since 1925 and explores the changing roles of Aligarh Muslim
University, national political campaigns, local rivalries over space and
livelihoods, and policing strategies. He presents data on the changing
spatial and demographic features of Aligarh, testing the popular thesis of
Hindu–Muslim economic competition as the source of rioting. In subse-
quent sections Brass analyzes the nature of political competition and local
electoral arithmetic with cogency and precision, while the role of the police
and the media are treated in separate chapters.
The sheer volume and complexity of this unique longitudinal study
comprising interviews, official reports, statistical evidence, and biographies
of key figures prevents strong conclusions. Brass reminds us that riots do
not happen in most places most of the time, not even in times of general-
ized tanav (tension) between communities. His material convincingly
demonstrates that over the decades, riots have repeatedly occurred in only
four or five specific localities in Aligarh. These localities are all character-
ized by the presence of seasoned riot specialists, men whose activities
span business, politics, and cultural-religious organizations; men who are
104 ETHNICITIES 6(1)

willing and able to translate rumours and general discourses into local
mobilization.

RIOTS AS ROUTINE POLITICS

This book is the work of a mature mind and does not discount the broader
cultural and psychological explanations of how the history of Hindu–
Muslim enmity has, over time, produced a rich archive of mythical knowl-
edge of ‘the other’, an archive which defies logic and reasoned argument.
Brass is, however, more interested in when, where and how this archive is
activated and transformed into arguments for action and violence. His
insistence on ‘demystification’ is refreshing and this book once again shows
the immense value of sustained and localized field research.
The most suggestive conclusion to emerge from Brass’s book is that riots
are integral and routinized aspects of India’s modern political culture, and
that condemning and bemoaning riots and their casualties have become as
much a part of this political culture as the riots themselves. Recent studies
of lower caste movements and other forms of political mobilization in
India suggest that activists are groomed to regard politics as a permanent
state of warfare. Violence no longer represents the breakdown of political
communication, but lies, rather, at the heart of contemporary Indian
political practice.
For all its merits, the book leaves a range of questions unanswered. We
hear a great deal about the ‘riot systems’ constructed over decades by
various Hindu nationalist figures in Aligarh. The riot systems on the other
side, among Muslims, appear less documented – almost non-existent –
despite stories of links between Muslim criminals and academics at Aligarh
Muslim University. Is this due to the difficulties involved in gathering
information from marginalized communities? Or is it because their
networks are differently organized? Or absent? Or is the whole idea of
symmetry, of equally apportioning blame and culpability to Hindus and
Muslims – really a myth in itself and a part of an official interpretative
regime that absolves Hindu nationalists of their prime responsibility for
what are, increasingly, anti-Muslim pogroms?
Another question left open is why riots occur in localities without
established ‘riot systems’. Brass’s answer would undoubtedly be that
‘new’ riots signify initial and necessary steps by local operators in organiz-
ing more permanent ‘systems’ that will ensure both their influence over
a constituency and the political effectiveness of future riots. Yet, the
question is whether this comes close to a tautology. Can one, for instance,
assume that a riot always represents more of the same logic? The
evidence on the effects of the Babri Masjid controversy in Aligarh
 REVIEW SYMPOSIUM 105

indicates that the national scale and systematicity of Hindu nationalist

campaigns in recent years constantly transform new areas into loci of
communal conflict and violence and thus seem to reduce the significance
of local factors.

THE ARITHMETIC OF HATRED

Although Brass has qualified his earlier, more hard-headed ‘instrumental-

ist’ position on how and why ethnic-communal identities are created and
maintained, assumptions of underlying political rationalities reverberate
through the book. Riots are ultimately rational mechanisms organized and
orchestrated in order to consolidate political constituencies and to repro-
duce paranoia and mistrust. The problem with this ‘switch on/off’ theory of
riot production is that it assumes that political operators working behind
the scenes always have their eyes on the larger, supra-local picture. Brass’s
own evidence, however, provides several examples of how this was not
always the case. His interviews with key figures also make it plain that they
are deeply immersed in what he brands irrational and ‘fantastical’ ideas
about the threat posed to the Hindu majority in India. Maybe we should
see riots as truly political actions, i.e. tentative, chaotic, and complex occur-
rences, immersed in dominant social and political ideological formations
and always unpredictable in their effects. We cannot extrapolate causes
from effects. We can, Brass reminds us, always be sure that riots are inten-
tional and organized with certain objectives in mind. Yet, the ‘riot systems’
do not always manage to produce the desired effects.
The postscript on the pogroms in Gujarat in 2002 provides additional
support for Brass’s thesis of riot systems being systemic features of India’s
political culture. Aligarh has experienced almost a decade without violence:
during this time political alignments have shifted and the Muslim population
has grown in strength, while the Bharatiya Janata Party (BJP) in Uttar
Pradesh remains paralysed. As a consequence, a non-communal candidate
was elected by both Muslim and Hindu voters in 2002. Simultaneously, in the
neighbouring state of Gujarat, the ‘riots systems’ painstakingly constructed
by the Hindu nationalist movement organized a gruesome pogrom against
Muslims, in complicity with the police and parts of the government. While
the riot systems were dormant in Aligarh, they flourished in Gujarat because
the BJP and aligned forces seized the opportunity to use public violence to
consolidate ‘Hindu sentiments’ and their political constituency.
If Brass is right, the same can happen again in Aligarh when the
combination of national political discourse, electoral arithmetic and local
grievances make it possible and expedient for the seasoned riot specialists
of that city to resume their deadly game.
106 ETHNICITIES 6(1)

References
Brass, Paul R. (2003) The Production of Hindu–Muslim Violence in Contemporary
India. New Delhi: Oxford University Press.
Varshney, Ashutosh (2002) Ethnic Conflict and Civic Life. New Haven, CT: Yale
University Press.

THOMAS BLOM HANSEN is Professor of Anthropology in the Depart-

ment of Anthropology, Yale University. Address: Department of Anthro-
pology, Yale University, PO Box 208277, New Haven, CT 06520–8277.
[email: Blomhansent@hotmail.com; thomasblom.hansen@yale.edu]

Review of Paul Brass

A.R. MOMIN
University of Mumbai

It can hardly be disputed that communalism, particularly in its most horri-

fying manifestation in Hindu–Muslim riots, poses the gravest threat to the
cohesiveness and stability of society and state in India. The social, economic
and human costs of communal violence are enormous and in fact incal-
culable. There is an unfortunate dearth of systematic, intensive and empiri-
cally based studies on communal violence; most of the studies on the subject
tend to be rather descriptive and lack analytical rigour and theoretical
refinement. This book (Brass, 2003), based on an intensive case study of
communal violence in the riot-prone town of Aligarh in northern India
carried over a period of nearly 38 years, makes a valuable contribution to
the systematic study of communal violence in contemporary India.
Brass argues that persistent communal violence in Aligarh, as elsewhere
in the country, is embedded in a discourse of communalism which is
premised on a deliberate and calculated accentuation of religious and
cultural differences between Hindus and Muslims, hostility towards
Muslims, and militant Hindu nationalism. Riots, according to him, are not
spontaneous occurrences which can be facilely explained in terms of mob
fury. They are essentially a planned, orchestrated and institutionalized
phenomenon. Brass argues that there exists at sites of endemic communal
violence an ‘institutionalized riot system’ which works as a central factor
in the genesis and persistence of communal riots. This system, which is
 REVIEW SYMPOSIUM 107

nourished and sustained by the discourse of communalism, involves a multi-

plicity of roles and a network of relations between specific individuals,
economic interests, organizations, criminals, politicians, and the police.
Brass maintains that there is a close linkage between communal violence
and the political process, especially electoral competition and political
mobilization.
In seeking to explain the persistence of communal violence Brass draws
on a functionalist theory and argues that communal riots have functional
utility and benefits for a wide range of groups and organizations in society,
particularly the Bharatiya Janata Party (BJP) and other political organiz-
ations affiliated to the Rashtriya Swayamsevak Sangh (RSS). He points out
that the Hindu–Muslim divide and polarization, which is rooted in the
discourse of communalism and militant Hindu nationalism, has been
extremely valuable to the political fortunes of the BJP.
In keeping with the view that riots are a product of planning and strategy,
Brass speaks of three stages of riot production: preparation, actual enact-
ment, and post-hoc explanation or interpretation. He provides, within the
conceptual framework sketched out in the foregoing, a detailed explication
and analysis of the pattern and dynamics of communal violence in Aligarh
since Independence. The focus of the study, he emphasizes, is on how riots
are produced, rather than on why they happen. The book contains some
highly perceptive observations and penetrating insights into the dynamics
of communal violence. It is written in a lucid, straightforward and candid
style.
While Brass’s systematically collected and enormous data on Aligarh
and his analysis of the varied dimensions of communal violence are
immensely valuable, his conceptual and methodological approach leave
something to be desired. Though he is critical of causal analysis and empha-
sizes the central role of human agency in social processes, one can never-
theless notice a lurking positivistic propensity in his generalizations. Thus,
he frequently talks about causal linkages, such as the ‘direct causal link
between riots and electoral/political competition’ (p. 33). This nomothetic
predilection is also betrayed in his hope that the findings of the book can
be generalized to ‘other parts of India and to other times and places in the
world’ (Preface, emphasis added). Apparently, Brass assumes or implies
that all forms of violence in the world are fundamentally alike, and that they
have basically the same dynamics, which can be explained in terms of a few
parameters (or laws) having universal applicability. This positivistic assump-
tion is a matter of belief rather than an empirically verifiable proposition.
It disregards the enormous variations in the patterns and manifestations of
violence in different social, cultural and political contexts and in their
aetiology and motivations.
Drawing on the functionalist perspective, Brass argues that large-scale
riots serve the interests of particular individuals, groups, organizations and
108 ETHNICITIES 6(1)

‘even society as a whole in concrete, useful ways that are beneficial to them’
(pp. 23–4, emphasis added). One can readily understand and agree that
communal riots have functional utility for a wide range of groups and
organizations that stand to benefit from them. However, one is at a loss to
understand how riots could have functional utility or benefits for society as
a whole. Evidently, riots bring about devastation, disruption and dis-
harmony in the larger society. In other words, they are dysfunctional to
society. Robert Merton, whose functionalist paradigm undergirds Brass’s
conceptualization, has clearly stated that certain forms of culture or social
processes may have functional as well as dysfunctional consequences.
Undoubtedly, the functionalist perspective has some heuristic value but, as
the eminent British anthropologist Sir Edward Evans-Pritchard cautioned
many years ago, in its extreme form functionalism often leads to absolute
relativism, a crude teleology, and a naïve determinism.
In tune with the functionalist perspective, Brass focuses on persistence
and continuity and ignores the role of changing social and political contexts.
The ‘institutionalized riot system’ may undergo radical changes over some
period of time or may become ineffective or may even break up and disinte-
grate under the pressure of an efficient, impartial and forceful local
administration, a vigilant local leadership, and concerted civic action. Brass
refers to a key player in communal riots, described as a ‘conversion
specialist’, who has a knack of turning a mere local, trivial incident into a
potent tool for instigating a riot. But he fails to consider the possibility that
the machinations of the ‘conversion specialist’ may well be thwarted by a
watchful and responsible local administration with the cooperation of a
vigilant local leadership. This has in fact happened in some places – in
Bhiwandi (Maharashtra) in 1992 and in Surat (Gujarat) in 2002. Unfortu-
nately, Brass does not take sufficient cognizance of the potentially effective
role of the local administration and civic action in averting or preventing a
riot, though he makes passing references to it (pp. 142, 169–70).
Brass’s study raises, inadvertently, some methodological problems. The
case study method employed in the study undoubtedly has the advantage
of providing an in-depth, micro-view of the complexity and dynamics of
specific social situations. However, it is not free from drawbacks. One tends
to take one’s field or area of study as a kind of yardstick of comparison and,
on the basis of it, makes rather sweeping generalizations and extrapolations.
Thus, Brass maintains that he is seeking an explanation for the persistence
of communal violence in Aligarh which ‘in turn can be generalized as an
explanation for recurring riots elsewhere as well’ (p. 38). Similarly, he
asserts that the political context of communal violence is decisive, that it is
riots that produce communal solidarity, not electoral politics that produces
riots (pp. 367–71). Looking at the Gujarat carnage of 2002, one is inclined
to believe that electoral politics, among other factors, may well lead to a riot.
It may be pointed out that generalizations based on a single case study
 REVIEW SYMPOSIUM 109

ignore the enormous diversity and complexity of Indian society in general

and of riot-prone sites and riot situations in particular. Furthermore, such
generalizations are usually made within a static framework and therefore
do not take due cognizance of changing social and political contexts. In
many respects, Aligarh has a rather unique character as a town and as a
site of endemic communal violence, different from other riot-prone towns
and cities. If one focuses on Aligarh, where the Hindu–Muslim divide and
communal riots have been conspicuous, one would invariably be led to
ask the question that Brass addresses: how riots persist and recur with
such frequency and intensity. However, if one were to focus on a differ-
ent riot-prone town or city – such as Surat or Bhiwandi, which have
experienced violent Hindu–Muslim clashes as well as long periods of
peaceful coexistence – one may be led to ask an altogether different, and
far more important, question: how can riots be effectively averted or
prevented?
Brass emphasizes, exaggeratedly, that riots are wholly a product of
rational planning and preparation. To buttress his argument, he draws on
the dramaturgical perspective and maintains that riots are ‘dramatic
productions’ (p. 369). He elaborates:
The people of Aligarh, like devotees of theatre, are kept in a state of readiness
for the next production (riot) through advertising of all kinds of trivial incidents
that hold the promise of a great drama to follow . . . each production involves
audience participation. (p. 358)
The contrasts between a riot and drama are too stark and glaring. While
watching theatre brings great joy and amusement, riots bring nothing but
fear, anxiety and despair for the people. Furthermore, the involvement of
the audience in theatre is wilful, whereas large numbers of people are
caught up in the riot situation against their will. The characterization of
Hindu–Muslim riots in the metaphor of drama or theatre is not only an
over-simplification but a grotesque distortion of communal violence.

Reference
Brass, Paul R. (2003) The Production of Hindu–Muslim Violence in Contemporary
India. New Delhi: Oxford University Press.

A.R. MOMIN is Professor of Cultural Anthropology and Head of the

Department of Sociology, University of Mumbai. Address: Department
of Sociology, University of Mumbai, Kalina, Mumbai 400 098, India.
[email: armomin@mu.ac.in]
110 ETHNICITIES 6(1)

Review of Paul Brass

ROGER PETERSEN
Massachusetts Institute of Technology, USA

Paul Brass’s The Production of Hindu–Muslim Violence in Contemporary

India (2003) is an empirically rich and theoretically ambitious work. In this
brief review, I wish to address only one issue, but an issue that is central to
the book – agency. By agency, I mean two things: actors possess a range of
options and they actively and consciously choose among this range of
options. If actors have only one realistic option, if they follow one option
unconsciously or if the options have been obfuscated or manipulated, then
the actors do not possess agency.
Brass’s work is centered on the agents behind communal violence, not
on the larger structural, economic, and demographic factors behind them.
His essential question is ‘on how riots are produced rather than on why they
happen’ (p. 16). Claiming to eschew causal analysis, Brass states, ‘All that
the social scientist can reasonably achieve in such situations is to expose to
full view, as far as possible, the agents involved in the production of
collective violence and the interests that are served by those who seek to
capture its meaning’ (p. 387).
With these goals in mind, Brass focuses his study on an ‘intermediate
space’ between the general public and the state. In this space, Brass locates
the agents who create the conditions for riots. Brass repeatedly turns to a
metaphor in which riots are likened to ‘dramatic productions, street theater
performances that are meant to appear spontaneous, but involve many
people in a variety of roles and actions that include inciting the people of
the audience, the dramatization and enlargement of incidents into a fit
subject for performance, and, finally, the production of events’ (p. 358).
Collective violence in India is the result of institutionalized riot systems
which include a plethora of roles – informants, propagandists, vernacular
journalists, rumor mongers, recruiters, thugs, and crucially ‘fire tenders’ who
bring incidents to notice and ‘conversion specialists’ who give the signal for
violence. For Brass, these are the agents of riots and the social scientist
should expose them. The riot system is embedded within a broader communal
discourse and persists partly due to a weak state, but to focus attention there
is to abdicate the search for agency – the specific who and how – in order
to fruitlessly chase after explanations based on broad causal forces and
psychological states which only serve to obfuscate and displace blame.
The metaphor of riot as dramatic production or street theater perform-
ance is a vivid one, but it raises some questions about agency. In particular,
what about the agency of the audience? Why, particularly in a democracy,
 REVIEW SYMPOSIUM 111

does the audience keep showing up for this violent dramatic production?
Brass argues that ‘The people of Aligarh, like devotees of theater, are kept
in a state of readiness for the next production through advertising of all
kinds of trivial incidents that hold the promise of a great drama to follow’
(p. 358). This passage, and many others, suggests very little agency for
members of the audience. They are ‘kept’, and apparently easily ‘kept’. In
this work, riots lead to increased communal solidarity that in turn affects
political competition, especially elections, in ways that benefit a small elite.
If Brass really wishes to concentrate on agency, he needs to better explain
to the reader (especially one not familiar with Indian society and politics),
how the audience appears to be ‘kept’ in a desired state by what appear to
be the same tactics time and time again and why the audience accepts the
same post-riot framing devices over and over.
Other scholars writing on Indian politics do not provide such a Pavlovian
portrait of Indian mass society. In Kanchan Chandra’s recent Why Ethnic
Parties Succeed: Patronage and Ethnic Head Counts in India (Chandra,
2004), Indian voters are rational and strategic actors who cannot always be
swayed by emotional ethnic appeals. More directly on the issue of riots,
Ashutosh Varshney’s Ethnic Conflict and Civic Life: Hindus and Muslims
in India (2002) argues that members of Indian society have the resources
and the power to prevent collective violence through local networks of civic
engagement. Brass quickly brushes aside Varshney’s work. He states that
Varshney’s method leads him to ‘a doubtful conclusion that civic engage-
ment between Hindus and Muslims can prevent violence, when it is more
likely that the creation of institutionalized riot systems overrides and
displaces whatever forms of civic engagement and interethnic cooperation
exist at specific sites’ (p. 27). In Brass’s treatment, societal resistance against
a well-oiled riot system seems almost futile. Indeed, Indian citizens, locked
into continual blame displacement are not even able to see communal
violence as a fundamental violation of their most basic democratic values
(p. 24).
If the audience does not possess agency, who does? The bit players in the
dramatic production of riots do not seem likely candidates. As Brass
mentions, they are ‘always ready to come out for a fracas’ (p. 358). Their
response is even more predictable and automatic than the audience’s. The
actors who do seem to have a clear ability to define their own interests and
act on them are local political elites, especially those wishing to affect elec-
tions, and the police. Brass’s treatment of the police in the Chapter 14 shows
a sophisticated and capable actor. Brass summarizes: ‘It was a matter of
some surprise to me to find that local police inspectors and station officers
in the mohallas of Aligarh in the riot-prone areas expressed their views on
riots very coherently, fully, articulately, and sensitively, with many nuances’
(p. 333). The police are not easily fooled by statements of political elites.
They could also prevent riots if they possessed the will to do so. But this is
112 ETHNICITIES 6(1)

just the point. The police have agency, but, as a Hindu-dominated organiz-
ation, appear to have little will to use it. This brings us back to the more
general point about agency behind communal violence in India.
Brass invokes Merton’s functionalism to describe the persistence of riot
systems. Like political machines seen previously in the United States, the
system possesses momentum of its own. The system can maintain itself so
long as no powerful party has a reason to destroy it. No powerful agent need
be behind the perpetuation of the machine or system. This reference is apt.
Although Brass wishes to discuss agency, there is not much agency in this
book. Institutionalized riot systems destroy the agency of the mass and its
voters, provide steady employment for marginal bit players and fail to
provoke a reaction from the police. Riots are a routinized form of ‘politics
by other means’ that elites fall back on as a default option instead of using
their intelligence and imagination.
However, Brass’s book suggests another possibility about agency.
Perhaps there are forces that affect actors in similar ways and create
common preferences. Then it may not be that a small political elite is
manipulating and obfuscating action as much as all actors are consciously
and willingly deciding to participate or support communal violence. The
power of the overarching Hindu–Muslim discourse may be such a force.
While the presence of institutionalized riot systems may explain much of
the variation in the time and place of riots, the high number of riots and
their persistence would seem better explained by the larger forces that
Brass would wish to avoid. Brass himself recognizes the power of a Hindu
ideology that seethes at Hindu displacement from a rightful position of
dominance. In fact, he concludes that, ‘In this world, the riots that
comprise my theme and subject are but specks in a much bigger frame’
(p. 327). If this is the case, then it may be that the audience, the bit
players, the politicians, the police are not driven by a largely concealed
riot system that operates on deception. Rather, the power and pervasive-
ness of this communal discourse aligns the preferences of all these actors
in the same direction. Many actors in north India and in riot-prone areas
such as Aligarh – elite and mass, criminal and police – may be pursuing
interests in full knowledge of their consequences. They know that they
have a range of options, they understand the consequences of their
actions, they are not being manipulated. They participate in the violence
because they want to, because larger cultural and structural forces compel
them to participate or tolerate riots. In discussing the infamous events
involved with the destruction of the Mosque at Ayodhya, Brass writes,
‘the evidence of preplanning was overwhelming and the personal satis-
faction of all those I saw and interviewed after the fact can only be
described as joyous’ (p. 355). But Brass does not wish to pursue the
puzzle of how a wide strata of the population can come to experience
such common joy about collective violence.
 REVIEW SYMPOSIUM 113

While Brass wishes to avoid assigning causal properties to vague

historical processes and large structural changes, other social scientists
place this widespread joy within such contours. For instance, Susanne and
Lloyd Rudolph (1993) saw the Ayodha events corresponding to a rise and
challenge to traditional Hindu dominance. They wrote of these new
forces:
Such mobility is unhinging a severely hierarchical society, creating social stress
bred of envy and resentment. Old established Hindu middle classes, mostly
from the upper literate and landed castes, suddenly see a whole range of
Johnny-come-latelies at their side who only yesterday were their inferiors in
status and income, both low-caste folks and Muslims. (Rudolph and Rudolph,
1993)
While Brass does give us a sense of the power of India’s Hindu–Muslim
discourse, we get little sense in his book of how the structural shifts and
their political effects discussed by the Rudolphs might motivate action in
combination with this powerful discourse. Although Brass recognizes
widespread joy at violence, he holds an absolute aversion to discussing
psychology. Brass writes of his approach: (A)ll forms of deep psychologi-
cal interpretation of individual motivations for participation in riotous
activity will be eschewed as essentially futile; instead the focus will be on
the hidden face itself, what can be made visible behind the mask of
discourse’ (p. 17). However, the forces described by the Rudolph’s do not
seem so ‘deep’ or unfathomable. One also wonders whether the discourse
is so hidden, whether the mass of individuals and voters are so easily
fooled, whether the bit players don’t really have something else to do.
When one looks at the overwhelming power of the communal discourse
since the 1947 partition and considers the visible structural and demo-
graphic forces at work in Indian society that can combine and shape the
direction of this force, one might also wonder why we have to avoid causal
analysis, why we need to spend our time ‘looking behind a mask’. Perhaps
we should look for individuals and organizations pursuing more obvious
interests and we should look for institutional solutions to the problem of
communal violence.

References
Brass, Paul R. (2003) The Production of Hindu–Muslim Violence in Contemporary
India. New Delhi: Oxford University Press.
Chandra, Kanchan (2004) Why Ethnic Parties Succeed: Patronage and Ethnic Head
Counts in India. Cambridge: Cambridge University Press.
Rudolph, Susanne Hoeber and Lloyd I. Rudolph (1993) ‘Modern Hate’, The New
Republic, 22 March: 24–9.
Varshney, Ashutosh (2002) Ethnic Conflict and Civic Life. New Haven, CT: Yale
University Press.
114 ETHNICITIES 6(1)

ROGER D. PETERSEN is Associate Professor of Political Science in the

Department of Political Science, Massachusetts Institute of Technology.
Address: Department of Political Science, Massachusetts Institute of Tech-
nology, 77 Massachusetts Avenue, Room E53–470, Cambridge, MA
02139–4307, USA. [email: rpeters@mit.edu]

Response to reviews by Thomas Hansen,

A.R. Momin, and Roger Petersen
PAUL R. BRASS
University of Washington, Seattle, USA

RESPONSE TO THOMAS HANSEN

The review by Thomas Hansen is of the kind that every serious scholar
should hope for, namely, one written by someone who has himself done
first-rate research on the subject, that gives a fair treatment of the work as
a whole, understands the arguments fully and raises at the same time diffi-
cult questions that deserve a response, further discussion and additional
research. First, to the question of the absence in my work of a discussion of
riot systems among Muslims in Aligarh. In many, if not most riots, violence
certainly occurs between the local populations of Hindus and Muslims in
adjacent mohallas. However, the core element of riot participation on the
Muslim side was virtually obliterated with the extra-judicial execution of
more than 40 alleged Muslim criminals, who were taken outside the city
limits and shot in 1979 during the incumbency of the Senior Superintendent
of Police, B.P. Singh. Even allowing for Muslim involvement in riot produc-
tion, however, it is the case that the systematic production of Hindu–Muslim
riots for the last quarter century or more, with differences from city to city,
has been a specialty of the militant Hindu organizations, who have the
largest share of responsibility for most of them that I know of.
How do riots get started without a pre-existing system of riot production?
I have shown how riots have become both more organized, more deadly and
more spatially extended, especially during the three decades from 1961 to
1991 when the system was at its most active. The spatial extension from the
center of the city to outlying areas previously untouched by such riots has
required the recruitment of riot specialists and criminal groups from particu-
lar mohallas in the city to attack Muslims in these outlying areas.
 REVIEW SYMPOSIUM 115

With regard to the effects of national political movements upon local

communal conflict and violence, it depends upon local factors, upon how
well organized the local system is, and upon the political context, that is, the
local electoral advantage to be gained from fomenting riots during a
broader political mobilization.
Finally, on the question of political rationality, passion, and ideology in
the production of riots. I accept that a considerable part of the Hindu popu-
lation, especially of the upper castes, are imbued with intense nationalist
feelings that include a dislike for, even a passionate hatred of, the Muslim
population of the country that is blamed for the partition of India in 1947.
Nevertheless, political rationality on the part of the principal riot producers
is critical: the ends in view are clear, the prospects of success reasonable
enough to make the effort worthwhile, and the results often enough
precisely what are desired.

RESPONSE TO A.R. MOMIN

Concerning my use and critique of causal analysis. I do not reject

completely the attempt to find causal linkages and connections. But I do
reject the often mindless, fruitless, and blame displacing misuses of causal
analysis in political science and sociology.
I do not assume or imply that ‘all forms of violence in the world are
fundamentally alike’. But, it also seems to me, and has seemed to others
who have read my work, that my arguments concerning organization, prep-
aration, planning, and deliberation are relevant to several forms of collec-
tive violence, and that my concept of ‘institutionalized riot systems’ may
have some bearing even in the extreme cases of genocide, as for example,
in Rwanda and Burundi.
It is true that I have found Mertonian functionalism a more useful form
of explanation of one kind of question that is not addressed satisfactorily
by any form of causal analysis, namely, how does one explain the persistence
of forms of social or political organization or practices that are universally
condemned in society. I concede that the implication that Indian society as
a whole may benefit from the persistence of riots was misleadingly phrased,
but my elaboration of the argument in the rest of the book is what should
matter, namely, that the existence of so-called Hindu–Muslim conflict has
been used as a justification for a strong state in India to retain the unity of
the country in the face of its internal conflicts.
I have made it as clear as can be in the book that any local administration
with a competent and effective district magistrate and senior superinten-
dent of police, with the backing of the state government, can prevent and/or
stop quickly any of the riots of the type that have been common in northern
116 ETHNICITIES 6(1)

and western India. As for civic action, it was hardly present in Aligarh
during the 30-year period under investigation.
Case studies are most valuable for the purposes for which I have used
them in this book and in my other writings on this subject, namely, the
uncovering of the dynamic processes involved in riot production rather than
demonstrating universal causal relationships. It is wrong to suggest that
Aligarh is a unique, exceptional case of riot production in India. On the
contrary, I have demonstrated in my own work before and since this book
the existence of systems of institutionalized riot production in Kanpur City
and Meerut. It is the Bhiwandi case that is exceptional. As for Surat, Dr
Arup Singh, at the University of Delhi has, in fact, already gone to that city,
has written about it in his PhD dissertation, and has found that my argu-
ments work quite well there. In any case, if one wants to know ‘how riots
can be effectively averted or prevented’, one has first to know how they are
produced. Moreover, I have presented adequate evidence on the subject of
riot control and prevention in this book itself, as stated above.
Contrary to Momin’s statement, the Gujarat case illustrates the correct-
ness of my argument as perfectly as one could imagine. One may read my
analysis of it in the epilogue to the book as well as in an article published
in Items (Brass, 2002–03).
Momin finds my metaphor of a drama to describe the rehearsal, produc-
tion, and interpretation phases in the process of riot production ‘grotesque’.
My abhorrence of the whole process of riot production is clear enough in
everything I have written. My metaphor of drama in no way underplays it.

RESPONSE TO ROGER PETERSEN

Petersen skews his entire review to a discussion of the notion of agency. I

argue very strongly in this and all my other work that deliberate, purposive,
directed actions by self-conscious persons and organizations (agents) are at
the root of all forms of collective violence. My focus in this book, as Petersen
notes correctly, is on the intermediate space of action between popular
sentiments and state action. Yet, he faults me for not focusing attention on the
space occupied by popular sentiments. In other words, rather than discuss
my book and its focus, he criticizes me for not writing the book of his choice.
In the process, moreover, he distorts my meaning, which is most certainly not
to argue that the general public lacks voice, agency, and purpose.
Petersen’s distortion of my arguments is most apparent where he makes
much of my use of the word ‘kept’, where, drawing upon my metaphor of a
staged drama, I say that the people of Aligarh are kept in a state of readiness
for the next production. Among the 32 meanings of the words ‘keep’ and
‘kept’ in the New Shorter Oxford English Dictionary, the meaning to be
 REVIEW SYMPOSIUM 117

attached to my use of the term here is that the general public in such cities
awaits, that is, anticipates, and often breathlessly, the next riot event, not as
passive persons, but as individuals who are not in fact in control of the events,
but must prepare their own ways of coping with, or participating in them.
Petersen also imagines me to have written things that I have not, that I
say that the purpose of keeping the audience in a state of readiness is to
‘benefit a small elite’. Neither this phrase nor anything conveying such a
meaning is to be found in my book.
Petersen goes on with his extraordinary portrayal of my book by alleging
that I have provided ‘a Pavlovian portrait’ to be compared unfavorably with
the work of other India specialists. Kanchan Chandra’s cited work
(Chandra, 2004) has nothing to do with collective violence. With regard to
Varshney’s (2002) work, I have noted its deficiencies in my book, and at
greater length elsewhere. The Rudolphs (1993) have done no work on
collective violence, but he finds a sentence of theirs in an article written for
a news magazine on a different subject that supports his own point of view.
Finally, Petersen presents my broader argument concerning the critical
role of the discourse of Hindu–Muslim enmity, the militant Hindu nationalist
ideology, the pervasiveness of feelings of ressentiment, and the ways in which
individuals, groups and organizations pursue their political aims and interests
within, or in relation to, these discursive frameworks and feelings, as if they
were his own, criticizing me for not doing so! This is warped criticism.
What is agent Petersen’s purpose in this review? Who knows? But
perhaps it is to justify his own work on interethnic violence, in which he
relies on emotional explanations, based on envy and resentment, presented
in a mechanical, rational choice, oversimplified, over-determined way as the
causal wheels upon which everything else turns.

References
Brass, P. (2002–03) ‘The Gujarat Pogrom of 2002’, Items 4(1): 1, 5–8 [also available
on www.paulbrass.com].
Chandra, Kanchan (2004) Why Ethnic Parties Succeed: Patronage and Ethnic Head
Counts in India. Cambridge: Cambridge University Press.
Rudolph, Susanne Hoeber and Lloyd I. Rudolph (1993) ‘Modern Hate’, The New
Republic, 22 March: 24–9.
Varshney, Ashutosh (2002) Ethnic Conflict and Civic Life. New Haven, CT: Yale
University Press.

PAUL R. BRASS is Professor Emeritus of Political Science and South

Asian Studies, University of Washington. Address: 807 Bowman Road,
Acme, WA 98220, USA. [email: brass@u.washington.edu]