sign in sign up
<<
Home , Bastar state, His Exalted Highness, Jaipur State, Kasim Razvi, Kondapalli Seetharamaiah, Malabar rebellion

Colonialism and Patterns of Ethnic Conflict in Contemporary

by Ajay Verghese

B.A. in Political Science and in French, May 2005, Temple University

A Dissertation submitted to

The Faculty of Columbian College of Arts and Sciences of The George Washington University in partial satisfaction of the requirements for the degree of Doctor of Philosophy

January 31st, 2013

Dissertation directed by

Emmanuel Teitelbaum Associate Professor of Political Science and International Affairs

The Columbian College of Arts and Sciences of The George Washington

University certifies that Ajay Verghese has passed the Final Examination for the degree of Doctor of Philosophy as of August 22nd, 2012. This is the final and approved form of the dissertation.

Colonialism and Patterns of Ethnic Conflict in Contemporary India

Ajay Verghese

Dissertation Research Committee:

Emmanuel Teitelbaum, Associate Professor of Political Science and International Affairs, Dissertation Director

Henry E. Hale, Associate Professor of Political Science and International Affairs, Committee Member

Henry J. Farrell, Associate Professor of Political Science and International Affairs, Committee Member

ii © Copyright 2012 by Ajay Verghese All rights reserved

iii Acknowledgements

Completing a Ph.D. and writing a dissertation are rather difficult tasks, and it pleases me to now finally have the opportunity to thank the numerous individuals who have provided support one way or another over the years. There are unfortunately too many people to recognize so I apologize in advance for those I may have forgotten.

Foremost, I benefited immensely from a stellar dissertation committee.

My greatest thanks go to Manny Teitelbaum, my dissertation chair. Most of what

I know about being a scholar I learned from Manny. Most importantly, he was never afraid to spare the criticism, and my work wouldn't be the same without his tough love. Henry Hale was also a great mentor during the researching and writing of this project, including conducting many a Skype conversation with me when I was lost in the field. Henry also wins the award for fastest e-mail turnaround in the history of dissertation advising. Finally, Henry Farrell was somehow always able to cite just the right book or article when I needed some more guidance.

I also had three excellent external readers for this project: Dane Kennedy,

Irfan Nooruddin, and Kanchan Chandra. All of them provided terrific feedback, pressed me on numerous issues, and I have no doubt my work will be better in the long run due to their generosity in sharing criticisms. Irfan and Kanchan also served as readers at the very last minute, so I am extra grateful for that.

There were several other professors at The George Washington University

iv to whom I owe thanks for various things over the years: Robert Adcock, Bruce

Dickson, Eric Lawrence, Harvey Feigenbaum, Jim Goldgeier, Shawn McHale,

Kimberly Morgan, and Chad Rector. I also made many good friends among my cohort of graduate students, without whom school would have been a lot less interesting: Colm Fox, Enze Han, Brian Karlsson, Craig Kauffman, J.J. Mikulec,

Joseph O'Mahoney, Varun Piplani, and Miles Townes.

There were also scholars at a few other universities who were helpful in talking through ideas or sharing feedback during phases of this project, and I'd like to thank James Mahoney, Lloyd Rudolph, Ron Herring, and Thomas Rosin.

Several institutions provided generous funding for both language training and fieldwork in India: The State Department Critical Language Scholarship

Program, The American Institute of Indian Studies, The George Washington

University's Sigur Center for Asian Studies, Konosuke Matsushita Memorial

Foundation, Cosmos Club, and The George Washington University Office of

Graduate Student Assistantships and Fellowships. Without this funding I would not have been able to learn or spend two years researching in the field.

In India I owe thanks to a great many people, too many to recount. My greatest debt of gratitude is to Asha Sarangi of the Centre for Political Studies at

Jawaharlal Nehru University. Without her constant help I would have been completely lost navigating the myriad bureaucracies of the Indian government. I also had the great fortune to meet a number of young scholars like myself who were also conducting research in India, and I would like to single out Adam

Auerbach, Gabi Kruks-Wisner, Julia Kowalski, Patton Burchett, and Lion König.

v Patton assisted me with finding two major treasure troves of data in , and without one of Lion's small tips about the computers in the National Archives of

India I would have literally wasted months of time. Special cheers to them!

In six different cities in India there were a great many individuals who assisted my research efforts. In Jaipur and Ajmer, I would first like to single out all of the teachers from the AIIS Hindi Language Institute, especially Vivekanand

Upadhyay, Anita Tripathi, and Kumar Janyani. In these two cities I also want to thank Shwati Vashishta, Panjak Sharma and staff at the Man Singh Museum, and the staff of the Institute of Development Studies. Anupam Sharma also helped me out a lot and even taught me how to ride a motorcycle. In New , George and

Rekha Varghese provided me with a home away from home, lots of chai and dosas, and always made me feel like a part of their family. I also want to thank

Ronn Mathew, George Varkey, Radhakrishna , Jaya Prabha Ravindran, and the staff of the National Archives of India. In , Thomas Punnose,

Jacob Punnose, and their respective families went out of their way to help me with everything. Premu Philips provided me with a wonderful place to stay in the heart of the city. I would also like to thank Zareena Parveen and V. Rangaraj at the State Archives and Research Institute, the staff of the Nehru

Centenary Museum, and the staff of the Central Library. In Raipur, I would like to thank Lalit Surjan, Prabha Taunk (for help on translations), Sunil Jaishwal, and the staff of the Deshbandhu Press Library. In , Philip and Mary

Cherian welcomed me into their home, and Cecil Thomas George took a lot of time to help me find great accommodations. I would also like to thank the staff of

vi the Directorate of Archives, , especially Reji Kumar,

Prakash B. (for telling me about the Journal of Kerala Studies), Pauvabhy (for tolerating my poor Malayalam), and Jacob Mathew at the Malayala Manorama.

During the last portion of my fieldwork in the United Kingdom and the

United States, I would like to thank the staff of the British Library and the Library of Congress. I'd especially like to single out Allen Thrasher and Ken Robbins.

I have too many personal friends to list here who deserve my gratitude.

Travis Valentine gets a special nod of recognition for years of putting up with me as a roommate. But writing this project was made easier by his true friendship.

Frank Verano provided a place to stay in the UK during my research, as well as lots of support while writing. Jenn Ortegren's sarcasm and help with procrastination was also much appreciated.

Finally, my family also played a key role during my graduate school years.

My ammachi Mary Verghese was always very supportive of my research – reminding me before I went to India in 2008 that was the capital – and I wish she was here to see the final product. My brother Ashwin deserves thanks for many smart historical discussions over the years, and for his friendship as I wrote up the final product. At last, my father Abraham always filled the house with books when I was younger and gave me a great love of history and politics from an early age, and my mother Anila always made sure that I finished my homework. I dedicate this dissertation to the two of them.

vii Abstract of Dissertation

Colonialism and Patterns of Ethnic Conflict in Contemporary India

Why does ethnic violence in multi-ethnic states revolve around one ethnic identity rather than another? Why, for example, do some conflicts revolve around religion whereas others revolve around language? What, in short, explains patterns of ethnic conflict in multi-ethnic societies? This is an important question for understanding ethnic bloodshed in a variety of plural states in Europe, Africa,

Asia, and elsewhere. I examine these questions through an investigation of India, one of the most ethnically diverse countries in the world. Using a mixed-methods research design that combines a quantitative analysis of 589 Indian districts with

15 months of archival work and elite interviews conducted in six case studies, I argue that the legacies of British colonial rule are the major determinant of contemporary patterns of ethnic conflict.

Three-fifths of India was directly-ruled by British administrators (British provinces) while the rest of the country remained under the control of native kings who exercised large amounts of internal autonomy (princely states). In the provinces, British administrators understood caste as the central organizing principle of Indian society, and they created ethnic policies which benefited high castes, discriminated against low castes and tribals, and protected religious minorities. Native kings, on the other hand, did the opposite: they emphasized religion as the bedrock of princely society, and therefore embraced policies that benefited co-religionists, discriminated against non-coreligionists, and protected

viii low castes and tribals. These sharply discordant policies led to the creation of disparate and durable 'fault lines' of ethnic conflict: former British provinces experience more contemporary caste and tribal conflict whereas former princely states experience more religious conflict.

Bifurcated colonial rule in India embedded master narratives of conflict in specific regions, reinforced them through local institutions, and ultimately engendered commonsensical understandings of how ethnic conflict is legitimately organized. Furthermore, because the new Indian government failed to reform many of the vestiges of its colonial past, the transition to independence in the mid-20th century did not disrupt these patterns of violence. Most of the ethnic conflicts which plague the country today are therefore best understood as legacies of colonialism. This system of bifurcated rule in India became a model for later British colonial expansion into parts of Africa and Southeast Asia, and this project therefore has major implications for understanding the historical roots of ethnic conflict in a number of multi-ethnic states around the world.

ix Table of Contents

Acknowledgements iv

Abstract of Dissertation viii

List of Figures xi

List of Tables xii

List of Abbreviations xiv

List of Maps xv

A Note on the Documentation of Archival Material xvi

Glossary of Terms xvii

A Note on Translation and Syntax xix

Introduction 1

PART I: THEORY AND QUANTITATIVE EVIDENCE

Chapter 1 – Colonialism and the Fault Lines of Ethnic Conflict 36

Chapter 2 – The Colonial Roots of Ethnic Violence: Quantitative Analysis 109

PART II: QUALITATIVE CASE STUDIES

Chapter 3 – Ethnic Violence in : Jaipur and Ajmer 132

Chapter 4 – Ethnic Violence in Kerala: Malabar and 196

Chapter 5 – Ethnic Violence in the Red Corridor: Bastar and Hyderabad 277

Conclusion – The Indian Model of Colonialism 334

Bibliography 367

Data Appendix 401

x List of Figures

Figure 1: Case Studies 28

Figure 2: Ethnic Policies in Colonial India 55

xi List of Tables

Table 1.1: Major Princely States in India by Gun Salute, 1931 45

Table 1.2: Princely Land Left by Muslims During Partition 78

Table 1.3: Historical Communal Composition of Mughal Mansabdars 85

Table 1.4: Riots Records, British Library 91

Table 2.1: Colonialism and Caste/Tribal Conflict Regression Results 119

Table 2.2: Colonialism and Caste/Tribal Mechanisms Regression Results 122

Table 2.3: Colonialism and Hindu-Muslim Conflict Regression Results 125

Table 3.1: Jaipur and Ajmer Comparison, 1901 and 2001 134

Table 3.2: Contemporary Religious Violence in Rajasthan (per 100,000) 142

Table 3.3: 2009 SC Violence in Rajasthan (per 100,000) 153

Table 3.4: Civil List, Mahakma Khas, 1929-30 161

Table 3.5: Communal Representation in Ajmer Gov't, 1925 172

Table 3.6: Hindu and Muslim School Enrollment in Ajmer, 1933-4 174

Table 3.7: Composition of Castes in Jaipur and Ajmer, 1901 177

Table 4.1: Malabar and Travancore Comparison, 1931 201

Table 4.2: Malabar and Travancore Comparison, 2001 201

Table 4.3: Communal Riot Casualties in Kerala, 1950-1995 (per 100,000) 215

Table 4.4: Communal Representation in Travancore Gov't, 1930s 238

Table 4.5: Composition of Castes in Malabar and Travancore, 1901 255

Table 4.6: Evictions in Malabar, 1862-80 260

Table 4.7: Land Tenure System in Travancore, 19th Century 265

xii Table 4.8: Agricultural Statistics for Kerala, 1956 273

Table 4.9: Agricultural System in Kerala, 1957-8 274

Table 5.1: Jagirdars and Maktadars Revenue in , 1920s 325

Table 5.2: Mansabdars Revenue in Hyderabad State, 1920s 325

Table 5.3: Hindu and Muslim Literacy, Hyderabad State, 1881-1921 326

Table 5.4: Salaries of Employees of Hyderabad State, 1920s 327

xiii List of Abbreviations

BJP – Bharatiya Janata Party

CDC – Centre for Rights (Jaipur)

INC –

MHA – Ministry of Home Affairs

MP – Member of Parliament

OBC – Other Backward Class

SC, ST and SCST – Scheduled Castes and Scheduled Tribes

RMAD – Rajasthan Minority Affairs Department

RSMC – Rajasthan State Minorities Commission

RSS – Rashtriya Swayamsevak Sangh

xiv List of Maps

Map 1: British Provinces and the Princely States, 1909 43

Map 2: Colonial 134

Map 3: Post-Colonial Rajasthan 134

Map 4: Colonial Kerala 200

Map 5: Post-Colonial Kerala 200

Map 6: Bastar and Hyderabad State, 1903 284

Map 7: The Red Corridor, 2007 285

Map 8: Colonial Burma 359

Map 9: Colonial Malaysia 363

xv A Note on the Documentation of Archival Material

All archival material for this dissertation was collected from the following locations: National Archives of India (NAI), the British Library in London (BL), the Jaipur City Palace Archives (JCP), the Andhra Pradesh State Archives

(APSA), the Deshbandhu Press Library in Raipur (DPL), and the Kerala State

Archives (KSA). I also visited a number of libraries in India, the UK, and the US for the collection of secondary historical material, but those volumes are not listed here.

I utilize the following citation formats for archival material:

1. NAI: Department, Branch, Year, File Number

2. BL: Shelfmark, Year

3. JCP: File Name, Year

4. DPL: Subject Number, Bin Number, Article Number

5. APSA: Installment Number, List Number, Serial Number

6. KSA: Department, File Number, Year and Bundle Number

Sometimes larger files from these archives were numbered so I also include page numbers where applicable.

xvi Glossary of Terms

Adivasi – 'original inhabitant', a member of an indigenous tribe

Dalit – a member of the Scheduled Castes; an Untouchable or Harijan

Darbar – the princely court, used to refer to the government of a princely state

Dargah – a shrine, in this case used to refer to the tomb of a Sufi saint in Ajmer

Diwan – Chief Minister of a princely state, also spelled 'Dewan'

Dora – the landed elite of ; also known as

Istimradars – the landlords of Ajmer

Jagir – a land grant in Rajputana

Jagirdari – an area of land controlled by a nobleman

Jenmi – Hindu landlords in Malabar; also spelled 'Janmi'

Kanamdars – land supervisors in Malabar

Khalsa – an area of land controlled directly by the state (British or princely)

Khilafat – an Islamic movement in India to preserve the Ottoman Empire

Mappillas – Muslims of Malabar

Nawab – a Mughal regional governor

Nizam – title for the Muslim rulers of Hyderabad

Pandaravaka – land controlled by the Travancore princely government

Puravaka – private land in Travancore

Raj – rule, or government; India was known as the

Raja, – king, '', i.e. Hindu ruler of a princely state

Rajputs – prominent warrior caste in Rajasthan

xvii Ryotwari – land in which the tribute was collected directly from the peasant

Verumpattamdars – cultivators in Malabar

Zamindari – a landlord area, usually in British India; a zamindar is a landlord

[Other terms are defined in the text]

xviii A Note on Translation and Syntax

1. There are a few cases where Hindi terms are used in the text. I opt for popular spellings rather than technical spellings. For example, the term Rajput would, using traditional Hindi transliteration (see Rupert Snell's scheme in Beginner's

Hindi) actually be spelled 'Raajpuut' (राजपूत). However, I just use the more popular 'Rajput'. There are also some Malayalam terms utilized in the text for which I employ the same practice.

2. In the mid-1990s the name of many Indian cities were altered: British names were replaced with more traditional names. So, for example, Calcutta became

Kolkata and Madras became . I still use the British names because most interview respondents in India utilized them. For example, I use the term

Trivandrum to refer to the capital of Kerala, although its name is now

Thiruvananthapuram. Most interview respondents in the region still say

Trivandrum. At other points in the dissertation I make reference to other Indian cities and almost always use British names.

xix Introduction

“Men make their own history, but they do not make it just as they please; they do not make it under circumstances chosen by themselves, but under circumstances directly encountered, given and transmitted from the past. The tradition of all the dead generations weighs like a nightmare on the brain of the living.” – Karl Marx1

“They realize at last that change does not mean reform, that change does not mean improvement.” – Frantz Fanon2

Why does ethnic violence in multi-ethnic states revolve around one ethnic identity rather than another? Why, for example, do some conflicts revolve around religion whereas others revolve around language? In short, what explains patterns of ethnic conflict in multi-ethnic societies? This is an important question for understanding ethnic bloodshed in many plural states in Europe, Africa, Asia, and elsewhere. I examine these questions through an investigation of India, one of the most ethnically diverse countries in the world. India is birthplace to three of the world's major religions, home to thousands of distinct castes and indigenous tribes, and the state has recognized 26 official language groups. Of its enormous diversity Lloyd and Susanne Rudolph have written: “India is the state

Europe would have become had the Holy Roman Empire embodied itself in a modern polity” (1987: 64). This striking pluralism has also become implicated in a variety of different conflicts throughout the country.

Different regions of multi-ethnic states may experience different kinds of

1 1852/2008: 15 2 1961/2004: 79

1 ethnic violence.3 India is no exception; scholars have noted that specific parts of the country contain a 'master narrative' which dictates the identity around which ethnic conflicts will revolve.4 In the city of Hyderabad, for example, Hindu-

Muslim riots have become endemic, but in northern Kerala caste conflict has defined the region's politics. However, these conflicts are puzzling because they are not necessarily based on the group sizes of ethnic groups in a particular region.5 Almost every district in India is Hindu-majority6, for instance, yet the

Hindu-Muslim divide only forms the axis of conflict in certain parts of the country. In other areas caste, tribal, or linguistic7 violence constitute the dominant narrative of violence. I posit that the 'fault lines' of ethnic conflict are not created simply by the particular constellation of ethnic groups in a given area, but rather by historical legacies.

I argue in this dissertation that the legacies of British colonial rule played the most important role in generating variations in ethnic conflict in the contemporary Indian state. India was the crown jewel of the , but

3 To be clear, I am not suggesting that multi-ethnic societies are more prone to violence than other kinds of states. As Fearon and Laitin (1996) note, most ethnic groups tend to live in peace, not conflict. But violence is naturally an endemic feature of large, modern human societies – therefore, assuming that ethnic conflict will occur, I am concerned with the specific question of why it revolves around one ethnic identity rather than another. 4 See Varshney 2002: 35 – “Either there are several master narratives in a country, one hegemonic in some regions and others powerful elsewhere, or some other locally based factors intervene between the potential power of narratives to cleave society and generate violence and the actually observed patterns of violence.” 5 This is, of course, contra Posner 2005. I discuss his work in detail later. 6 As of 2001, the main exceptions are located in the state of Jammu and Kashmir. 7 The only kind of ethnic violence which I do not analyze in this dissertation is linguistic violence. There are several reasons for this. First, time constraints limited the scope of the project. Second, conflict around language has been shown to be less violent than other forms of ethnic conflict (Laitin 1999). In fact, Steven Wilkinson (2008) finds that linguistic mobilization in India is statistically negatively correlated with ethnic conflict deaths and injuries. Other fine scholars, however, have written on linguistic violence in India – see Lacina (forthcoming).

2 only three-fourths of its population was directly-ruled by British administrators in territories known as provinces. The rest of the population remained under the control of native kings who exercised large amounts of internal autonomy in territories known as princely states.8 It is the principal argument of this dissertation that this dichotomy in colonial rule conditioned and ultimately structured the way ethnic violence occurs in modern India. This dissertation finds that former British provinces experience more contemporary caste and tribal conflict, whereas former princely states experience more contemporary religious conflict.

Bifurcated colonial rule9 led to the creation of disparate and durable fault lines of ethnic conflict. In the provinces, British administrators understood caste as the central organizing principle of Indian society, and they created ethnic policies which benefited high castes, discriminated against low castes and tribals10, and protected religious minorities. On the other hand, native kings emphasized religion as the bedrock of princely society. They embraced policies that benefited co-religionists, discriminated against non-coreligionists, and protected low castes and tribals. These sharply discordant policies in the

8 The British also referred to these areas as the 'native states' or just the 'states'. I sometimes utilize these two terms at various points throughout the dissertation. 9 Let me be clear about the usage of this term as it is likely to provoke some confusion. Lange et al. (2006) identify four broad types of British colonialism: settler (Canada), indirect (Nigeria), direct (Singapore), and hybrid (India). When I say that British India experienced 'direct rule', I mean that British administrators were in charge of governing these territories, although Indians themselves were also involved. When I say that the princely states experienced 'indirect rule' or 'native rule', I mean that while British rulers were the final authority native kings still had large amounts of internal autonomy. India was a unique colony in this regard, combining two forms of colonial rule together. See the Conclusion for comparative implications – a few other colonies also fit into the hybrid category. 10 I use the term 'low castes' to refer to Scheduled Castes. Tribals in India are often also known as adivasis (original inhabitants); I use both terms interchangeably.

3 provinces and princely states had long-lasting effects on ethnic violence in India.

Bifurcated colonial rule embedded master narratives of conflict in specific regions, reinforced them through local institutions and power structures, and ultimately engendered commonsensical understandings of how ethnic conflict is legitimately organized (Laitin 1986). The transition to independence in the mid-

20th century did not disrupt these patterns of conflict, mainly because the new

Indian government failed to reform many of the vestiges of its colonial past. The socioeconomic condition of minority ethnic groups in different regions has remained static. Most of the ethnic conflicts which plague contemporary India are therefore best understood as legacies of colonial rule.

India is one of many post-colonial, multi-ethnic states around the world.

The era of European colonialism had major, long-term effects on the ethnic politics of a number of polities in South America, Africa, and Asia, among other regions.11 Colonial administrators engineered massive data-gathering projects about native populations, created some ethnic groups out of thin air, altered their pre-existing sizes and spatial distributions, and even promoted animosity and violence between them. Daniel Posner, for example, surveys the immense impact of colonialism in Africa:

“Howard Wolpe shows that the Igbo of Nigeria were a product of colonial boundaries. Terence Ranger shows that the Manyika of Zimbabwe were 'created' by missionaries. Crawford Young traces the origins of the Ngala of present-day Congo to Henry Stanley’s misinformed labeling of the people he encountered on

11 The study of the state's effect on classification is very broad, going all the way back to the work of Weber, Foucault, and Bourdieu. On colonialism's effect on thnic politics, see as a small sampling Horowitz 1985, Laitin 1986, Young 1994, Mamdani 1996, Posner 2005. The comparative implications of this research project do not extend only to post-colonial states; the broader point is that historical legacies, colonial or not, influence patterns of ethnic conflict.

4 his river exploration. Philip Gourevitch shows that the emergence of Hutus and Tutsis as distinct identity groups in Rwanda was a product of Belgian administrative fiat.” (2003: 127)

The British experience in India, however, was unique in two important ways.

First, British administrators did not control the entire country, resulting in the creation of two very different but concurrent colonial histories on the subcontinent. Second, the success of the 'Indian model' – that is, combining direct and indirect rule12 – led to its adoption in several other British colonies.

The 'Indian model' of colonialism was eventually exported to parts of eastern and southern Africa, as well as states within Southeast Asia (Fisher 1991, Metcalf

2007). This project highlights that the idea of 'colonial rule' is not uniform – many states around the world experienced different types or styles of colonialism.13 Understanding these often subtle distinctions may have a major effect on understanding variations in ethnic conflict. India has the potential to be an especially important case because it allows researchers the opportunity to compare colonial rule (British India) with its relative absence (princely India). In short, we can potentially isolate the effects of colonialism on ethnic conflict.14

The rest of this introduction is structured as follows: I begin by defining several key terms utilized throughout this dissertation. The second section then consists of a literature critique, in which I argue that existing work on ethnic violence – both on India specifically, and conflict throughout the world more broadly – suffers from several important shortcomings. In section III I then put 12 Indirect rule is a term often used to describe the system of governance in princely India. British administrators were supreme, but I argue that native kings had a large amount of internal autonomy. I discuss this further in Chapter 1, section 1.3. 13 See for example Blanton et al. 2001, Lange et al. 2006. 14 I want to thank James Mahoney for originally bringing this idea to my attention.

5 forward my theory of ethnic conflict in India, upon which I will expand in subsequent chapters. In section IV I lay out the research design of the project, detailing both the quantitative and qualitative methods employed. Finally, in section V I briefly sketch out the remaining chapters of the dissertation project.

I. Terminology

This dissertation is primarily concerned with explaining patterns of ethnic conflict in contemporary India. Before delving into the argument it is necessary to define certain principal terms. First, I follow Donald Horowitz in defining ethnic as “easily embrac[ing] groups differentiated by color, language, and religion; it covers 'tribes,' 'races,' 'nationalities,' and castes.” (Horowitz 1985: 53).

The three specific kinds of ethnic identities, or ethnic groups, with which I am concerned in this dissertation are those based on caste, tribe, and religion.15 I argue, further, that ethnicity is essentially about uncertainty reduction. Ethnic markers are but one way of reducing the complexity of the social world16, but they are uniquely powerful because ethnic groups share myths of common origin, a sense of a common fate, culture and symbols, physical similarities, and face reduced barriers to communication (Hale 2004, 2008).17 Ethnicity is therefore not simply instrumental; it connotes a way of interpreting and making sense of

15 I lump together castes and tribes – see Chapter 1, section 2.4 for an explanation. 16 Hale (2008: 242) notes that “...ethnicity derives from the inherent human need to reduce uncertainty in the world, a need well established in psychological research and consistent with evolutionary theory.” 17 Chandra (2006a) has criticized the idea that ethnicity is important at all; however, in earlier work (2004: Chapter 2) she shows how ethnic identities may provide informational shortcuts that are useful to individuals. For example, an individual's last name in India provides knowledge about their caste, religion, and linguistic group. Information about other identity categories – an individual's class, for example – is not nearly as accessible. For authors who argue that ethnicity relates to deep, emotional bonds see Connor 1993, Kaufman 2001, and Petersen 2002.

6 the social world. Much of this is discussed further in Chapter 1. Relatedly, I accept the constructivist insight that individuals potentially have a multitude of different ethnic identities18; Kanchan Chandra (2012) carefully delineates between nominal and activated categories. The former is the set of ethnic identities to which a person may belong; the latter is the set of identities to which an individual professes membership or is assigned such by others. In almost every region of India there are multiple potential fault lines of ethnic conflict, but violence occurs around one dominant axis.

Ethnic violence19 refers to violence which occurs largely along ethnic lines.20 However, I am not concerned with ethnic wars, civil wars, or separatism.

On the contrary, I am particularly interested in everyday forms of violence, or what may be termed low-level conflict. Finally, many scholars have made the argument that what seem like ethnicity-based conflicts are often not explicitly ethnic, but rather interpreted or officially stylized that way after the fact (Brass

1997, Mueller 2000, Kalyvas 2003). However, while remaining sensitive to this argument in analyzing the various conflicts documented in this dissertation, I find that the historical construction and treatment of ethnic groups is the main driver of contemporary patterns of conflict.

The actual ethnic violence that occurs can take many different forms, but there are two principal ways of classifying conflict: top-down21 or bottom-up. For 18 See Posner 2005: 11. 19 I use the terms 'ethnic conflict' and 'ethnic violence' interchangeably. 20 Chapter 5 of this dissertation focuses on the caste and tribal conflict which occurs in eastern India. In that chapter, I focus on the way low castes and tribal groups have initiated violence against the state. They are not technically fighting another ethnic group, but rather the Indian government. 21 Top-down does not refer to the state, but rather a higher-positioned group in a social

7 example, top-down caste or tribal violence refers to atrocities committed against

Scheduled Castes and Scheduled Tribes.22 Top-down religious violence refers to riots, as they are generally directed against Muslims23, who form a religious minority in almost every district of the country. This kind of violence is driven primarily by long histories of antagonism between different ethnic groups, and ethnic majority groups initiate violence to 'teach a lesson'24 to minorities. But violence is not only top-down; marginalized ethnic groups often retaliate.

However, because they are weaker these groups employ very different methods of violence – generally what may be termed 'armed conflict'. This includes bombings, assassinations, arson, etc. This kind of violence is driven primarily by the socioeconomic condition of ethnic minorities – where they are poorer, they have a greater likelihood of starting violence. Colonialism structures both of these kinds of conflicts – I argue that it created long-term antagonisms between ethnic groups, as well as the relative power distributions between different groups. All of this is merely a way of disaggregating the variegated ethnic conflicts which may occur in a country, and I am careful to do this in my case study analysis. For the sake of conceptual simplicity, however, I lump the two principal forms of violence together. I am interested in the totality of ethnic conflicts in contemporary India.

hierarchy persecuting a lower-positioned group; i.e. high castes attacking low castes. 22 Scheduled Castes (SCs) and Scheduled Tribes (STs) – sometimes combined together as SCSTs – are low castes and tribal groups who have been judged by the Indian government to have experienced historic discrimination. SCs are also known as , Harijans, Untouchables, or Panchamas. All these terms are used interchangeably, but I mainly use the term 'low castes'. 23 Brass, for example, has noted that the term 'pogrom' may be more fitting than riot, and has investigated why the latter term is not used more throughout India. See his 2003 work. 24 This was a phrase often recounted to me during fieldwork – in Hindi, sabak sikhaana.

8 II. Literature Critique

The study of ethnic conflict in the social sciences consists of an extraordinarily vast literature.25 To deal with its complexity I structure my critique into two overarching categories: India-specific literature, and literature on ethnic conflict more broadly. I opt for this approach because the study of ethnic violence in India has led to a number of impressive books published on this subject over the last decade (Brass 1997; 2003, Varshney 2002, Wilkinson

2004) that merit individual attention.26

The existing literature on ethnic conflict in India has made several important insights into the study of conflict. Most of these works have centered specifically on post-independence Hindu-Muslim riots. This literature has rightly been called “one of the most striking examples in recent years of the development of a cumulative research program in political science” (Chandra 2006b: 207).

Three main scholars of conflict have put forward different theories to account for both the causes of religious violence and variations in religious conflict throughout India. Paul Brass (1997, 2003) has argued that communal27 riots in

India are not simply random acts of violence, but are rather primarily driven by

'riot specialists', many of them belonging to a class of militant Hindu nationalist organizations called the Sangh Parivar (Family of Associations). Two important such groups are the Bharatiya Janata Party (BJP) and the Rashtriya 25 This literature is too broad to adequately document; see Posen 1993, Horowitz 1985; 2001, Huntington 1993, Fearon and Laitin 1996; 2000; 2003, Mueller 2000, Kaufman 2001, Petersen 2002, Kalyvas 2003, and Cederman et al. 2010. 26 These are just some of the major works; also see Ali Engineer 1989; 1991, Krishna 1985, Shani 2007, and Bohlken and Sergenti 2010. For different takes on religious violence, also see Weiner 1978, Ollapally 2008. 27 The term communalism in India refers primarily to religious conflict; I use this term at times.

9 Swayamsevak Sangh (RSS). Brass argues that entire cities – like Aligarh in the state of Uttar Pradesh, which has been the focus of much of his work – develop

'institutionalized riot systems' (2003: 32). Riots are engineered precisely because they serve a clear benefit: “It should be clear enough by now...how valuable

Hindu-Muslim opposition, antagonism, and violence have been to the fortunes of the BJP” (2003: 8).

Ashutosh Varshney (2002) has focused more explicitly on the causes of ethnic peace. He argues, following Robert Putnam (1993), that intercommunal links – those which cross religious divides – can prevent violence. These links may be formal (i.e. organizations, neighborhood peace committees), or informal

(i.e. Hindu and Muslim families eating together), although formal links are posited to be stronger. Where these links exist in India, Hindu-Muslim violence is minimized. In a clear reformulation of Brass' theory, Varshney then argues that cities create 'institutionalized peace systems' which can either stop Hindu-

Muslim riots or prevent small-scale incidents from escalating quickly into them.

Finally, Steven Wilkinson (2004) has posited an electoral theory of Hindu-

Muslim violence. He argues, following Brass, that riots are planned and not spontaneous.28 Wilkinson posits that politicians have the ability to prevent riots through their control of the police force. Local politicians will do this if either of two conditions occur: a.) Muslims form an important voting bloc29, or b.) the

28 Brass posited a link between electoral politics and Hindu-Muslim violence, although he did not expound on it like Wilkinson. See 1997: 9, emphasis added – “When full-fledged riots develop, the local politicians and authorities are often either incompetent or they themselves desire the riots to take place...” 29 Specifically, either Muslims are important to the governing party or a coalition party.

10 party system is so competitive that governing parties will need future minority group coalition partners. If either of these conditions are met, then politicians will order the police force to stop riots from occurring. Therefore, Hindu-Muslim violence is driven by political elites and their electoral incentives.

While these works have illuminated fascinating and undoubtedly important aspects of Hindu-Muslim violence in India, all of these accounts suffer from some similar and serious underlying problems. Specifically, I argue that there are two overarching shortcomings of the existing literature: first, its lack of historical depth; and second, its narrow focus on religious violence. Both of these weaknesses, as I will show, constitute critical challenges to these arguments.

The first underlying problem with existing scholarship is a serious lack of historical depth. This is not to say that the aforementioned scholars have not taken Indian history seriously – rather, they have not taken enough of it seriously. All three of the main theories of ethnic violence in India have generally focused only on the post-1947 period. Yet the problem of communalism is much more deeply entrenched than this, and a wide literature, much of it published by historians, has been devoted to it.30 For instance, the earliest recorded Hindu-

Muslim riot occurred in Ahmedabad in 1714 (Krishna 1985: 61), even before the

British had firmly come into political control of the subcontinent. The causes of communalism cannot be adequately understood if divorced from its long history on the subcontinent.

For example, most current scholarship on Hindu-Muslim riots has focused

30 Just as an introduction see Bayly 1985, Pandey 1990, Keppley Mahmood 1993.

11 on the impact of militant Hindu nationalist organizations. There can be no doubt that these organizations are a critical part of the story behind modern Hindu-

Muslim riots. But the problem with focusing on them is obvious: communal riots predate their existence. Communal riots even predate the existence of older

Hindu nationalist organizations such as Mahasabha (founded in 1915), a precursor of the BJP. So the rise of militant Hindu nationalism may help explain modern riots, but not their longevity. A deeper historical study is necessary to uncover the root causes of communalism. Note that the problem of history also seriously undermines Wilkinson's argument, at least as a general theory of ethnic violence in India: riots again predate the introduction of Indian electoral politics.31 So is Wilkinson's account only meant to explain modern riots, or are there other underlying causes of violence?

Along these lines, there is a very curious lack of attention paid specifically to the colonial period in the existing literature. It is not entirely ignored: for example, Varshney points to the 1920s as the key period during which associational life was constructed (2002: 16-18) and Wilkinson also delves into it at times.32 However, colonial rule as a potential explanatory variable for ethnic violence is not seriously considered. This oversight is puzzling: despite these authors' singular focus on Hindu-Muslim riots, there is a stunning neglect of the argument that British colonialism promoted communal violence via divide-and-

31 This is true even if we take into account that during the British period some elections did occur, although without universal suffrage. For example, the Morley-Minto Reforms of 1909 allowed some Indians to vote for legislative councils. Communal riots still predate these developments. 32 Interestingly, most of Wilkinson's mention of the colonial state is in reference to their policies on caste, not religion. See 2004: Chapter 6.

12 rule policies. This argument, so well-known and central to much of the historical literature on communalism in India33, is not taken seriously as a possible determinant of post-independence riots by any of the authors in question. In

Chapter 1 I address this issue in detail, and I test it statistically in Chapter 2.

To deal with this first of two important shortcomings, this dissertation project considers ethnic violence within a longer historical time frame; it takes care to consider political developments from the pre-colonial, colonial, and post- colonial period. To be clear: existing scholarship has looked mainly at the proximate causes of contemporary violence, but this project is concerned with its underlying causes. In doing so, this project aims to contribute to the 'historic turn in the social sciences'.34 Misunderstanding the long-term roots of ethnic conflict will lead to dubious conclusions about why ethnic groups are in conflict, as well as how best to prevent violence.

The second major shortcoming of the major work written on ethnic violence in India has been its focus on communalism.35 These two terms, though overlapping, are not synonymous. Hindu-Muslim riots, i.e. religious violence, constitute only one type of ethnic violence in India, and there is, furthermore, little evidence that it is even the most important type of conflict. Religious identity forms the basis of one type of ethnic category. But linguistic groups, castes, and tribes are also important ethnic identities. These latter ethnic

33 See Chapter 1, section 2.5. 34 See North 1991, Steinmo et al. 1992, Pierson 2000; 2004, and Mahoney and Rueschemeyer 2003. 35 Brass (1997) does discuss caste violence, but is primarily interested in religious riots. Varshney and Wilkinson also make references to caste violence, but their gaze is focused mostly on Hindu-Muslim riots.

13 categorizations, and the conflicts which revolve around them, have unfortunately been largely ignored in the recent political science literature.36

Perhaps there is some evidence that Hindu-Muslim riots are more important than other kinds of ethnic conflicts. The Varshney-Wilkinson database, which used Times of India newspaper reports to record incidents of communal rioting from 1950-1995, found roughly 7,000 deaths over that 45-year period.

Even considering that this number is likely underestimating the actual number of deaths, this is a small figure taken in comparative perspective with other ethnic conflicts in India. Data on caste and tribal conflict which I utilize (from the

Worldwide Incidents Tracking System) for my statistical analysis in Chapter 2, for example, show that a similar number of deaths occurred from these conflicts just from 2005-9. Caste violence has been a significant problem in India – in

2010 alone a total of 32,712 crimes were committed against members of the

Scheduled Castes.37 It is further estimated that every day in India three Dalit women are raped and two Dalits are killed (Scuto 2008: 5). Similarly, since the

1940s in many parts of eastern India major tribal rebellions have been raging relentlessly against the Indian state. The weight of the evidence thus suggests that fewer people have probably died from communalism in India than caste and tribal conflict – therefore, Hindu-Muslim violence may be the least prevalent type of ethnic violence in India.

36 This is not to say that they are completely unaddressed, though historians have done most of the heavy lifting. On caste conflict see Gough 1974, Guha 1999, Urdal 2008. On tribal violence see Mathur 2004, Gupta 2007. On linguistic violence see Lacina (forthcoming). 37 Data from Crime in India (2010), published by the National Crime Records Bureau, Ministry of Home Affairs.

14 Why then have communal riots achieved such a special focus in contemporary political science research? Part of the interest might be driven by the fact that Hindu-Muslim riots are generally spectacular conflagrations which receive overwhelming media attention. The burning of a train which led to the deaths of 59 Hindu pilgrims in Gohdra and started the is a perfect example of this phenomenon. Or perhaps communal riots receive such attention because they are open to primordialist, essentialist interpretations of an ancient struggle between and Muslims. Whatever the reason, the exclusive study of Hindu-Muslim riots does not deserve its privileged position in contemporary political science research on India.

Religious, caste, and tribal conflicts are all ethnic conflicts, yet the latter two remain severely under-studied. This dissertation aims to also correct this important shortcoming by looking at the three main forms of ethnic violence in contemporary India: caste, tribal, and religious conflict. Naturally, one risk with such a research program is that different types of ethnic violence may have different causes. In fact, this is probably true – for instance, caste and tribal conflict is often driven by land issues. However, all of these conflicts are ethnic conflicts, and variations in conflict are driven by the understanding individuals have about their specific ethnic categorizations.38 Why individuals with multiple identities choose to fight over one ethnic identity rather than another is an important question for the study of ethnic politics more broadly.

38 See Horowitz 1985: 53 – “This concept of ethnicity means that ethnic conflict is one phenomenon and not several. To be sure, that conflict takes different courses...[but] the putatively ascriptive character of ethnic identifications imparts to ethnic conflict its intense and permeative qualities.”

15 Finally, aside from these two major shortcomings which affect the work of all three aforementioned scholars, there are several other smaller though perhaps no less important criticisms to note. As just one example, take the work of

Varshney. There are several reasons to be skeptical of his argument about the importance of intercommunal links. First, it is not clear that 'institutionalized peace systems' can overcome the power of 'institutionalized riot systems' (Brass

2003: 27). Second, there are a wide variety of historical cases in which two communities with a long lineage of interethnic linkages still managed to engage in brutal violence. For example, Robert Bates et al. note on the ethnic cleansing in the Balkans during the 1990s:

“Not only did the 45 years between the end of World War II and the outbreak of the recent war remain remarkably peaceful, but also, during that era, high rates of intermarriage took place among ethnic groups.” (1998: 609)39

Finally, associational life in India is among the weakest in the world (Chhibber

199940), and therefore placing special focus on intercommunal organizations seems unjustified. The Varshney example is only meant to show the smaller problems with each of the specific arguments of the existing literature on ethnic violence in India; in the empirical chapters of this dissertation (Chapters 3-5) I return to these arguments again to explain why they are insufficient to explain violence outcomes in my case studies.

39 See also Van Evera 2001: 22 on this: “Intermarriage among Serbs, Croats and Bosniaks was fairly common in Bosnia before 1992 but this did little to slow the Bosnian slaughter of 1992- 95. The lesson is that ethnic groups and ethnic conflicts can survive a good deal of intermarriage.” 40 Chhibber's claim is buttressed by data on associational membership in India from the World Values Survey.

16 2.2 Broader Arguments on Ethnic Conflict

Beyond the India-specific literature, there is a much broader social science literature on ethnic conflict. For instance, James Fearon and David Laitin (2003), in their well-known broad gauge study of ethnic civil wars, find that various factors – poverty, geography, etc. – are the primary drivers of conflict. But as I noted earlier, my work is not concerned as much with the proximate causes of violence, and this has been the focus of most social scientific literature on ethnic conflict. I am, rather, concerned with the historical legacies that often underpin violence41, and in explaining patterns of conflict.42

This latter question has been investigated, most notably by Daniel Posner

(2005).43 Posner set out to discover why political competition in Zambia focused on either tribal identities or linguistic identities. He posits that colonialism structures the repertoire of available ethnic identities, but that individuals then select a specific identity of their choice based on an essentially rationalist framework. In short, individuals consider ethnic group sizes44 and determine which identity most benefits them. Given institutional changes – in this case, the transition from single-party to multi-party rule – Posner shows that individuals choose to emphasize different ethnic identities. Therefore, the salience of a particular ethnic cleavage is determined by both individual actor preferences and the institutional rules of the game. He writes: 41 Some very interesting work has been done along these lines; see Kapferer 1988 and Prunier 1995. 42 Wilkinson (2008) does examine ethnic mobilizations and violence, which is to some extent an investigation of patterns of conflict. But he is interested in determining whether some kinds of ethnic mobilizations are more violent than others, not geographical patterns of conflict. 43 See also Lipset and Rokkan 1967. 44 A similar dynamic, ethnic head counts, is also posited in Chandra 2004.

17 “...ethnic groups are mobilized or joined not because of the depth of attachment that people feel toward them but because of the usefulness of the political coalitions that they define – a usefulness determined exclusively by their sizes relative to those of other coalitions.” (2005: 12)

My work, while sharing Posner's focus on the importance of the colonial period, nevertheless differs considerably from his account. First, while Posner is specifically focused on explaining the salience of particular ethnic cleavages, this dissertation is concerned with ethnic violence.45 There is little reason to believe that the two phenomena are similar; as Posner himself admits: “...to the extent that the dynamics of ethnic violence are different from the dynamics of non- violent ethnic politics, the account presented here may be applicable only to explaining the latter” (2005: 10). For example, Posner's account places a lot of importance on voting, but the way that actors view themselves in the voting booth does not necessarily correspond to the way in which ethnic violence unfolds. Posner's theory supposes that individuals can choose their identities and discard them at will. In the realm of ethnic violence, however, it is not clear that this is possible.46 There is scant evidence that individuals in India shape or manipulate ethnic identities in the realm of violence in a way that is comparable to voting. Is it possible, for example, that oppressed low castes in a violent area can attempt to create a Hindu-Muslim divide where there is none? Can they strategically position themselves within the Hindu fold if caste conflict in the

45 Posner does write that his book is about “ethnic conflict” (2005: 1). However, he does not deal explicitly with violence; rather, his work is really focused on why certain ethnic cleavages become salient whereas others do not. The type of 'conflict' on which he focuses is not bombs and bullets; rather, he means political competition. 46 There is, for example, a broad literature on how ethnic violence 'hardens' ethnic identities and makes identity shifts nearly impossible. See Kaufmann 1996, Byman 2000, and Van Evera 2001. For a dissenting view, see Kalyvas 2008.

18 region has been the dominant narrative? Finally, Posner's argument is predicated on the idea that ethnicity is 'thin', and used merely for instrumental purposes. In the next chapter I detail a substantially different theory of ethnicity.

My argument veers more closely toward the results of David Laitin's study of ethnic cleavages and conflict in Africa (1986). Laitin set out to explain why politics in Yorubaland revolved around ancestral city membership rather than religion. He found that British colonial administrators in the region 'expunged' religion as a legitimate mode of ethnic classification, therefore creating ancestral city membership as the main axis around which conflict would occur. He writes:

“The point being made is that the political situation in Yorubaland was fluid at the point of British conquest, and that the British resuscitated a declining set of authorities in order to rule through them. In doing so, they created a cultural hegemony of ancestral city membership. And this is why it seems obvious to nearly all Yorubas that religious conflict is bloody and led by fanatics, whereas regional (i.e. tribal) conflict is natural and led by responsible authorities. This commonsense perception did not come from the free marketplace of ideas; it was forged by British colonial strategy.” (1986: 158)

Colonialism therefore created a hegemonic understanding of ethnic politics which continued to persist into the contemporary period. Very similar dynamics apply in India: British rulers in the provinces also expunged religion as a mode of ethnic classification and sought to minimize its importance in political life. In fact, as I detail in the Conclusion, the British experience in India created the model of bifurcated rule which was applied to Yorubaland, as well as other states in Africa and Southeast Asia. Note finally that Laitin's account is not dependent upon the group sizes of ethnic groups or individual actor preferences47; instead, it

47 Posner does discuss Laitin's work; see 2005: 281-82. Interestingly, Posner finds that his model cannot explain the existent ethnic cleavage structure in Yorubaland – i.e. his model would predict political competition revolving around religion, which it does not. Posner does not explicitly cite Laitin as a counterargument to his own work, but I treat it as such here.

19 posits that colonial rule creates commonsensical understandings of how ethnic politics and ethnic violence is legitimately organized. Similarly, the account presented in this dissertation does not depend on group sizes, etc. but rather the different kinds of administrative rule experienced in the provinces and the princely states.

III. The Argument

The primary argument of this dissertation is that contemporary patterns of ethnic violence in India have been conditioned and structured by the bifurcated history of colonial rule on the subcontinent. There are two specific components to the argument. First, I argue that the disparate ethnic policies of British administrators in the provinces and princely rulers in the native states led to the creation of different kinds of 'ethnic fault lines'. In the provinces, castes and tribes became the primary ethnic classification around which conflict revolved; in the princely states, on the other hand, conflict revolved around religion.

The second aspect of the argument deals with the post-colonial period and the persistence of patterns of ethnic conflict. I argue that colonial rule embedded a master narrative of conflict in a region, reinforced it through local institutions and power structures, and this led to commonsensical understandings of how ethnic conflict is legitimately organized. Furthermore, I posit that the new Indian government failed to reform pre-existing ethnic policies during the post-colonial period. In the former provinces, many low caste and tribal groups are still suffering under the weight of historic discrimination, and reforms efforts – although impressive on paper – have ultimately failed to minimize violence. In

20 the former princely states, on the other hand, minority religious groups suffer from numerous discriminatory legacies, and reform efforts have also failed to alter their socioeconomic condition. As I detail extensively in my case study chapters, the lives of ethnic groups in conflict often remain fundamentally unchanged for hundreds of years.

Many scholars will naturally be skeptical about the supposed persistence of ethnic relations or patterns of ethnic conflict from the past.48 Most constructivist scholars of ethnic politics have showcased how ethnic classifications have changed – often radically over very short periods of time

(Chandra 2012). This has led to the conclusion that ethnicity itself may be epiphenomenal and ultimately immaterial for various political outcomes, including violence (Chandra 2006a). While ethnic identities can in principle change, it is incumbent to show the mechanisms by which these changes occur.

In India, as I detail in depth during the next chapter, many of the traditional mechanisms of change (new state policies, political parties) have, instead of reordering society, explicitly built on the inherited edifice of colonialism. To be clear, I do not discount the immense ability of the state to change ethnic identities or reform ethnic politics. Carter Johnson and Diana Dumitru, for example, emphasize this in their work on interethnic cooperation during the

Holocaust:

“Our cases strongly suggest that state citizenship and nationality policies, if they are focused on integration and inclusion and backed by strong political commitment, are able to build cooperative interethnic relationships that are

48 In a broader sense, this is a criticism of 'path dependent' arguments – see Mahoney 2000, Thelen 2004, Streeck and Thelen 2005, Page 2006.

21 internalized and that endure beyond the life of the state itself.” (2011: 2).

But this only highlights two missing variables in post-independence India: a focus on integration and inclusion, and strong political commitment. As I detail further at the end of Chapter 1, despite a vast array of reform efforts the position of minority ethnic groups in many parts of the country has not altered considerably. For instance, the Sachar Committee (2006) conducted for the first time a serious investigation into the position of Muslims in India, revealing the deep socioeconomic discrimination still faced by the community. Similarly, despite a number of official reservations (i.e. quotas) for members of the

Scheduled Castes, their representation in powerful cabinet portfolios is still incredibly weak (Jensenius 2012). About tribal groups Ramachandra Guha notes that “...compared with India's other subaltern groups, such as the Dalits...and the

Muslims, they are well nigh invisible.”49 All of this is not meant to suggest that no real changes in the post-independence period have occurred, or are impossible.

But rather, the key variables which may lessen the impact of the past or alter pre- existing patterns of ethnic conflict are not present in contemporary India.

IV. Research Design

This dissertation project utilizes a mixed-methods research design50, combining together both quantitative and qualitative analysis. I first utilize a quantitative analysis of 589 Indian districts to establish the broad patterns of

49 “A War in the Heart of India”, The Nation, June 27th 2007. 50 The mixed-methods movement in political science has become very important over the past decade. This literature is too large to adequately recount here, but for some important foundational works see Almond 1988, Caporaso 1995, Laitin 2003, Brady and Collier 2004, and Lieberman 2005.

22 ethnic conflict in contemporary India. Then, I use qualitative archival and interview evidence collected from case studies of six carefully chosen territories to see whether large-N trends are also visible at the local level, as well as to investigate the mechanisms that have produced violence outcomes.

4.2 Quantitative Analysis

The results of my quantitative analysis, presented in Chapter 2, are based partly on a new dataset of ethnic conflict in contemporary India.51 I also utilize a variety of data collected from, among other sources, the Indian government, the

Worldwide Incidents Tracking System, the Indian Human Development Survey, and private statistical firms. Through a series of regression models, I showcase that British direct rule is positively correlated with contemporary caste and tribal violence, but negatively correlated with contemporary religious violence. I also test several mechanisms that may produce conflict, most of them pertaining specifically to caste and tribal violence.

4.3 Qualitative Analysis

The qualitative portion of this dissertation project, presented in Chapters

3-5, consists of archival research and elite interviews from case studies of six territories, compiled from fieldwork conducted over a period of 15 months in

India, the United Kingdom, and the United States. Archival evidence was collected from a number of different locations: The National Archives of India,

Jaipur City Palace Archives (Man Singh Museum), Andhra Pradesh State

Archives, Deshbandhu Press Library, Kerala State Archives, the British Library,

51 Much of this work was conducted in collaboration with my dissertation advisor Emmanuel Teitelbaum.

23 and the Library of Congress. Aside from a wealth of primary sources from these locations, I also used a number of secondary source materials. Many of these secondary sources were books, reports, and various other publications which are not readily available in the United States. Furthermore, almost all of it was written by Indian experts and scholars whose work is too often ignored by

Western scholars. For instance, in Kerala I made use of the local Journal of

Kerala Studies, published by the University of Kerala. This publication consists of the work of a number of important scholars of Malayali politics and history who are, unfortunately, largely unknown in the West.

Elite interviews were conducted among roughly 75 individuals in the six different cases. Respondents in every city were chosen from a wide variety of backgrounds in order to triangulate responses.52 Typical respondents included government officials (from a number of different agencies and departments),

NGO officials, police officers, journalists, and academics, among others. Because this project was considered somewhat controversial in India, interview respondents are not identified by name. The primary purpose of elite interviews was to provide information about the state of ethnic minority groups in the six case studies, as well as relations between groups. While some quantitative information on this topic was available from various sources, most often it was not. Data on Muslims throughout India, for example, is very difficult to acquire.53

I therefore supplemented my quantitative work with the use of qualitative interview evidence where there was a lack of reliable existing data. I also relied on

52 Information about how respondents were chosen is located in the Data Appendix. 53 I discuss why in Chapter 2.

24 contemporary newspaper reports (from local newspapers) to collect this information.

Six case studies were selected for the qualitative analysis. Because of

India's enormous subnational variation, it was extremely important to select cases from different regions of the country. No single district or state can reasonably be said to be representative of the country at large. My case studies are pitched at three different levels of analysis: the district, state, and regional level. This allowed me to control for different variables that exist at different levels of political administration. The primary purpose of analyzing these cases is twofold: first, to see if the patterns established using quantitative analysis are also present at the small-N level; and second, to investigate the mechanisms that produce specific ethnic violence outcomes.

The first two pairs of case studies are what I refer to as controlled historical comparisons. This terminology is adopted from Alexander George and

Andrew Bennett's idea of a controlled comparison, defined as “the study of two or more instances of a well-specified phenomenon that resemble each other in every respect but one” (2005: 151). I add the term 'historical' because my comparisons are longitudinal, i.e. the two cases are compared over a very long time period. The difficulty with this strategy is finding cases that can meet George and Bennett's strict criteria. In fact, it is probably impossible to find cases that are similar 'in every respect' – however, what controlled comparisons really require are two cases with several important independent variables held constant. In my district-level and state-level controlled historical comparisons from Rajasthan

25 and Kerala, for example, the two territories I selected for comparison have similar geographies, languages, economies, ethnic cleavage structures, etc.

Therefore, a number of potentially confounding factors are controlled for. Then, the main difference between the territories is the key independent variable of interest – the nature of colonial rule. In both controlled historical comparisons the two territories differ only in their colonial experience: one case was a British province whereas the other was a princely state.

The first controlled historical comparison consists of two neighboring districts in the north Indian state of Rajasthan: Jaipur and Ajmer. These two cases are remarkably similar except for the fact that during colonialism Jaipur was a princely state whereas Ajmer became a British province. This created a radical historical divergence between these two territories. Furthermore, I argue in Chapter 3 that Ajmer was selected for annexation due to reasons that would not have had any effect on ethnic violence outcomes (its geographical location in the center of what was then known as Rajputana). Therefore, the comparison between the two areas can be said to be quasi-random.54

The second controlled historical comparison comes from the south Indian state of Kerala. I specifically chose a southern state to account for vast cultural differences between the north and south. I compare the northern Malabar region of Kerala with the southern Travancore region; the former came under the control of the British whereas the latter remained an independent princely state.

54 I use the term 'quasi-random' for two reasons. First, it is not clear that social scientists can actually conduct random or natural experiments (Sekhon and Titiunik 2012). Second, because the decision to annex territories happened hundreds of years ago, it would be hard to prove conclusively that the selection of cases was completely random.

26 British administrators from the early 20th century even remarked that these two territories were exceptionally similar, and that their divided political administration was an historical “accident” (Meyer et al. 1909: 56). Furthermore,

I present historical evidence that the British were also interested in conquering

Travancore but, for various reasons, were unable to do so. Therefore, the selection of Malabar is also quasi-random – even more so than the Ajmer case.

The final two case studies are not a controlled historical comparison – rather, they are deviant cases from eastern India. Bastar and Hyderabad are two former princely states which – unlike almost all other former native states – experience extremely high levels of caste and tribal conflict. I therefore seek to explain this unusual outcome and whether my argument can account for it. These two cases are pitched at the regional level, as they cover two different contemporary states ( and Andhra Pradesh) within the broader 'Red

Corridor' of Naxalite conflict.55

Figure 1 below shows the location of the three pairs of case studies in this dissertation. The red circle indicates the Rajasthan comparison, the green circle indicates the Kerala comparison, and the blue circle straddles the border between the Andhra Pradesh and Chhattisgarh cases.

55 This refers to the region of eastern India where an immense amount of tribal violence occurs. Naxalites, most of whom come from Scheduled Tribes, are Maoist groups rebelling against the Indian state. See Chapter 5.

27 Figure 1: Case Studies

There is, however, an underlying problem with selecting cases for a research project such as this: endogeneity (Geddes 1990, Collier and Mahoney

2006).56 That is to say, are patterns of ethnic violence in the country a product of 56 Iyer (2010) opts for a clever strategy to overcome endogeneity in her comparison of economic outcomes in former provinces and princely states. Her study, which is entirely quantitative, utilizes an instrumental variable analysis. The specific instrument she uses is the (which I discuss briefly in Chapter 1). In short, during one period of Indian colonial history the British implemented a policy, called the Doctrine of Lapse, through which princely states were annexed if kings died without natural heirs. Therefore, the annexation of states

28 colonial rule or were they already established prior to colonialism? The key issue is therefore how the British selected areas of India for annexation. If they selected areas that already had high levels of caste and tribal violence, for instance, then colonial rule is not an important variable in explaining conflict outcomes.57 So how did the British choose which areas became their provinces?

First, it is clear that the pre-existing level of ethnic violence in a region did not matter to British administrators. There is no historical evidence that they had a master plan to annex either violent or peaceful regions, and there are plenty of examples of both kinds of territories coming under their control.58 Rather, the main selection criterion for annexation was the availability of productive land.

Historian Barbara Ramusack explains:

“As the English Company extended its political control throughout India, it annexed the most economically productive areas, both agriculturally and commercially. Thus it began in Bengal, long fabled as one of the richest Indian provinces; it quickly gobbled up most coastal areas to facilitate commercial enterprises and gradually engorged the fertile Gangetic plain to Punjab. When Queen Victoria renounced any further British annexation, the princely states were located mainly in less economically productive areas.” (2004: 186)

Therefore, the British were not selecting cases on the specific dependent variable of interest: ethnic violence. This seems to have played no part in their

during this period was presumably random (i.e. occurring when a king died). On closer inspection, however, this instrument is highly flawed: several princely states in which the king died without a natural heir were not annexed by the British (see pg. 712). The selection of territories to be annexed even via the Doctrine of Lapse, then, was definitely not random. This therefore undermines her use of this policy as an instrument. 57 For a book which makes a kind of similar claim, i.e. negating the supposed impact of colonialism, see Herbst 2000. 58 Arguments could be made that they wanted to control either violent regions or peaceful regions. Violent regions may have been easier to conquer because the British could come in as a third-party against two competing forces. Peaceful regions may also have been desired precisely because there was no conflict and therefore these areas would have required less manpower, arms, and money to annex. However, most of the archival records consulted for this project, many of which included the personal correspondence of British administrators or official minutes from meetings, showcase that these thoughts were not likely on the minds of the men who conquered India.

29 calculations. However, it is still possible that the actual selection factor – the availability of productive land – might itself in some way be related to ethnic violence outcomes. As a hypothetical example, productive land may have been heavily rack-rented, meaning that levels of caste violence in these regions may have already been high prior to British conquest. However, the relationship between the availability of productive land and caste violence outcomes is not obvious – it is difficult, based on the value of land, to infer anything concrete about caste relations in a given area.59 Ultimately then, the British selection of areas for annexation does not appear to be directly related to ethnic violence, nor does the real selection factor (fertile land) influence ethnic violence outcomes.

The reasoning behind the British annexation of a territory was divorced from the causes of ethnic conflict.

Nevertheless, to further counteract the potential problems associated with endogeneity, I carefully selected two princely states among my cases which were considered to be economically productive areas. Ramusack continues:

“Notable exceptions [of princely states] with extensive natural resources included the coastal states of Travancore and Cochin with a small but lush agricultural base in both food and cash crops, the cis-Sutlej Punjab states possessing fertile soils and early access to canal irrigation, and Hyderabad and with diverse economies.” (2004: 187)

Travancore and Hyderabad were therefore areas which do not fit the traditional profile of princely states in India. Also, as I noted previously, in Chapter 4 I argue that Travancore was even more unique in the sense that the British did want to

59 To further illustrate this point, consider that Banerjee and Iyer (2005) find that the most productive land in India today is located in areas that were formerly ryotwari settlements – i.e. where land taxes, etc. were collected directly from peasants without the inference of middlemen like zamindars. Therefore, they essentially show that productive land today was not rack-rented in the past.

30 conquer it during the 19th century but were unable to do so. Finally, in case the religion of the ruler makes a difference I made sure to select both Hindu rulers

(Jaipur, Travancore, Bastar) and Muslims (Hyderabad).

V. Dissertation Outline

The rest of the chapters of the dissertation are laid out here with a brief description of their contents. Part I of the dissertation consists of Chapters 1 and

2, and is devoted to theory and a large-N quantitative analysis of patterns of ethnic conflict across India. Part II of the dissertation consists of Chapters 3, 4, and 5, and forms the bulk of the project. These chapters are devoted to six in- depth qualitative case studies, and make use of extensive archival research and numerous elite interviews. The sixth and final chapter then concludes the analysis and puts forward several comparative implications of the project for other cases around the world.

The first chapter is a broad historical treatment of the colonial period in

India. I argue that during this period, the main ethnic 'fault lines' of conflict were created. I detail the history of the British in India, why certain territories remained under native rule, and the disparate ways in which British administrators in the provinces and native kings in the princely states constructed ethnic policies. While the British pursued policies that led to low caste and tribal discontent, they simultaneously enforced protectionist policies for Muslims, thus limiting religious violence. On the other hand, princely rulers pursued policies that were discriminatory toward non-coreligionists, but protected low castes and tribals. I then conclude the chapter by highlighting that

31 many of these important ethnic policies instituted during colonialism did not dissipate during the post-independence period.

The second chapter presents quantitative evidence from 589 Indian districts which showcase the main pattern of ethnic conflict in contemporary

India. Using a variety of datasets and several regression models, I demonstrate that former British provinces experience more caste and tribal conflict whereas former princely states experience more religious violence, while controlling for a number of alternative explanations. I also test several mechanisms posited in

Chapter 1 that may have produced these outcomes.

The third chapter deals with the first of three qualitative pairs of case studies. In this chapter I examine a controlled historical comparison of two districts from the north Indian state of Rajasthan: Jaipur and Ajmer. I detail the histories of both regions, why Ajmer came under British rule, and the disparate ethnic policies of British administrators there and the Hindu rulers of Jaipur. I find that while the British exacerbated caste conflict in Ajmer, the policies of the of Jaipur led to communal discord. I then trace ethnic violence outcomes in both regions over time, and I find that conflict in Jaipur still revolves around religion, whereas conflict in Ajmer continues to revolve around caste.

The fourth chapter deals with the second pair of case studies, a controlled historical comparison of two areas within the south Indian state of Kerala:

Malabar and Travancore. I begin by describing the histories of both regions, explicating why Malabar was conquered by the British whereas Travancore was not. I then present historical evidence which indicates that British administrators

32 also wanted to take control of Travancore, but for various reasons were unable to do so. I then examine the disparate ethnic policies of the rulers of both regions, and trace violence outcomes over time. While Malabar has experienced a long history of caste and tribal conflict, Travancore became an area beset by serious communal conflict.

The fifth chapter details the final pair of case studies, the deviant cases of

Bastar and Hyderabad, both located in the 'Red Corridor' region of India. This area is known for its immense tribal and caste (i.e. Naxalite) violence. However, two of the main sites of conflict are the former princely states of Bastar and, specifically, the Telangana region of Hyderabad. In this chapter I examine why these cases experience such intense violence. I find that in Bastar, the British essentially came to control the state by the mid-19th century, and that before that time there was very little tribal violence in the region. Second, I find that the conflict in Telangana was not driven by the ruler of the Hyderabad State, but rather by a unique class of local landlords who were given proprietary rights over the land much like the zamindars of British India. I then examine the post- colonial period, and document how the Indian government failed to reform the ethnic politics of both regions.

The sixth and final chapter is the conclusion. I first revisit my theory, and summarize the main arguments from every chapter. I also address some of the major contributions of this project for existing scholarship on ethnicity, ethnic conflict, and colonial legacies. I then seek to explain the comparative implications of this dissertation, examining India's position within the larger British imperial

33 system. The creation of a separate provincial and princely India led to a specific model of colonial rule – the so-called 'Indian model'. I highlight that this model was subsequently exported to other colonies in Africa and Southeast Asia. I then examine whether insights from the Indian case can explain patterns of ethnic violence in these other post-colonial states around the world, and I place special emphasis on the colonies of Burma and Malaysia.

34 PART I: THEORY AND QUANTITATIVE EVIDENCE

35 Chapter 1 Colonialism and the Fault Lines of Ethnic Conflict

“British colonial officers, who were largely responsible for the creation of the two , tended to see them as almost different worlds. William Barton, who after a distinguished career in the political service became Resident in Hyderabad, wrote in 1934 that the achievements of the Indian political genius were to be found at their best in the states. That idea of an ancient but vital political tradition, cultivated also by the British, made many Indians consider the states not only as symbols of past glory, but also harbingers of future independence, islands of freedom washed by the waves of imperialism.” – Dick Kooiman1

“Aspects of the British-princely distinction persisted in the political behavior of a generation of postindependence Indian politicians and their constituents. In certain contemporary Indian states, however, the significance of the British/princely distinction has been even greater: historical-political heterogeneity – briefly, the fact and the consequences of British versus princely rule in historically separate territorial units – has fundamentally influenced the formation of the contemporary state and its evolution as a political community.” – John Wood2

In this chapter I present an historical theory3 of colonialism and ethnic conflict in contemporary India. I offer an explanation of how bifurcated colonial rule led to disparate ethnic violence outcomes which have persisted into the modern period. I argue that during the colonial era, the British understood caste as the central organizing principle of society in the provinces. On the other hand, princely rulers primarily emphasized their religious identities. This overarching dichotomy affected the way these rulers then constructed ethnic policies toward different groups. This ultimately had, as I will detail, long-lasting effects on

1 2002: 19 2 1984: 66 3 Taking seriously Ian Lustick's (1996) warnings about selection bias in historically-oriented political science work, I have attempted to, within the text itself and through the use of footnotes, cite authors who have offered contrasting historical views to my own – what Lustick calls the method of 'explicit triage' (pg. 616). Naturally there is no singular 'correct' interpretation of Indian history. This way, however, other scholars may at least compare my arguments with others and decide for themselves what is more convincing.

36 patterns of ethnic violence in the contemporary Indian republic.

This chapter begins with a brief overview of the formation of the British

Indian Empire, especially detailing how several hundred princely states came to remain semi-independent principalities. I then compare ethnic politics in the two areas, highlighting that the British discriminated against low castes and tribals, but protected religious minorities. Princely rulers, on the other hand, discriminated against non-coreligionists, but protected low castes and tribals. I then conclude by detailing why ethnic conflict outcomes persisted into contemporary India, highlighting that colonialism embedded master narratives of conflict into local institutions and power structures. Furthermore, the post- independent government in India also failed to reform its colonial past, thereby enabling pre-existing patterns of conflict to continue into the modern period.

I. The British Provinces and the Princely States

In 1585 trade representatives from Queen Elizabeth I of England arrived in

Surat, in modern day Gujarat, as part of an expedition to request trading rights from the Mughal Emperor . This was the humble start of the British

Empire in India.4 Within two centuries this trading company would come to control a vast portion of the subcontinent, and the empire that succeeded it at its peak governed over 600 million people. As Francis Hutchins succinctly put it,

“India was conquered for England by merchant adventurers” (1967: 3). India would come to be considered the most important of British colonies – the 'crown jewel' of the empire. As Lord Curzon, Viceroy of India, once stated: “We could

4 There were other European powers who sought a foothold in India – the Dutch and the French, for example. I do not consider those powers in this discussion.

37 lose all our [white settlement] dominions and still survive, but if we lost India, our sun would sink to its setting” (Metcalf and Metcalf 2001: 131).

When the British arrived in India the Mughal Empire5 was firmly consolidated. However, within a century this changed dramatically. Historian

Michael Fisher notes that the “gradually dissolved into its constituent regions during the course of the eighteenth century” (1991: 46).

Regional leaders – often known as nawabs – remained officially tied to the

Mughal Emperor, but in Bengal, Hyderabad, and other areas they in effect created independent regional kingdoms. New powers such as the Marathas from south India also began to challenge Mughal supremacy. By the early 18th century the British had become an emerging power on the subcontinent.

The beachhead of British India was Bengal, and this was the area where the British began their commercial exploits. After the initial granting of trading rights in 1613, the British gradually expanded their clout in the region. First, in 1698 they purchased zamindari rights – the ability to collect taxes – from the villages in their territory around Calcutta. Then, in 1717 they purchased additional zamindari rights in another 38 villages in Bengal. The

Company at this time officially operated as a client of the Mughal Emperor. This continued until the aggressiveness of the British in their commercial expansion resulted in a backlash from the local nawab of Bengal, who managed to briefly expel the Company from Calcutta. The British responded with military force; they

5 The Mughals were an Islamic power from Central Asia. Mughal is the Persian term for Mongol, as the founder of the empire, , was descended from Timur and Gengis , rulers of the Mongol Empire.

38 defeated the nawab’s armies at the decisive Battle of Plassey in 1757 which, for most historians, signaled the beginning of British political control over Bengal.

Plassey was followed soon thereafter with the Company obtaining diwani

– chief land revenue collector – rights in Bengal (1765). Philip Mason recounts that this “was the biggest single step in the transformation of traders into rulers”

(1985: 36). The Battle of Buxar, in 1764, against the nawab of Awadh, also marked an historical turning point because it allowed the British to expand into the Gangetic heartland and prompted the first dispatching of Political Residents6 to the courts of princely rulers (Fisher 1991: 49–50). In short, the post-1757 period resulted in the expansion of Company power over most of the remaining northern portion of India. Among the myriad changes in this ‘post-Plassey’ period was the introduction of the ‘Permanent Settlement’ by Lord Cornwallis in

1793, which was a new system of land ownership and revenue collection that attempted to create a class of landlord rulers (zamindars) in India similar to the large landowners of England (Guha 1981).

The period from 1765-1813 was dominated by a policy of creating a 'ring- fence', or a buffer between the Company's territories and external threats. The

British during this period were but one of many powers in India, and it was not yet clear that they would become dominant. From 1813-1858, however, the situation changed and the Company became more aggressive, creating a policy of

'subordinate isolation', which was intended to establish British supremacy in

India by annexing native states and increasing the power of Residents. The years

6 Residents were British officials sent to princely courts to 'advise' rulers; see Fisher 1991.

39 1848-1856 were especially eventful; it was during this period that Lord Dalhousie established the 'Doctrine of Lapse'. Prior to this policy, the British had allowed the independent princes of India to adopt heirs when they were unable to produce them naturally. Once the Doctrine of Lapse was instituted in 1848 the

British disallowed this practice and annexed territories where princes did not produce natural heirs (i.e. these territories lapsed to the British). The increasing annexations undertaken during this period by the East India Company created widespread disapproval and discontent (Chopra et al. 2003: 137-140).

As British power was increasing over India it is worth considering their strategy for expansion. I dealt with this briefly during the Introduction when I discussed endogeneity and case selection, but I will discuss the issue in greater detail here. Overall, it is important to recognize that British expansion in India did not follow any master plan – rather, it proceeded somewhat haphazardly.

There were a number of reasons for this. First, the British government was not actually annexing territory – it was the East India Company, a trading corporation, which was expanding the boundaries of the empire. Second, British administrators were hardly experts on India and therefore did not understand enough about the country to have developed a coherent blueprint for expansion.

Third, the British also vacillated between strategies depending on specific

Governors-General – Lord Cornwallis was known to be very aggressive, for example, whereas his successor John Shore was not. Finally, the British were also driven always by local developments – the power of the Marathas and the

Kingdom of Mysore in the south, for example, presented stiff challenges for

40 imperial expansion in that region.

It does seem likely, however, that the British, if not for the Rebellion of

1857 (discussed below), would have conquered the entire country. As late as 1856 they were expanding aggressively throughout India and there is little reason to believe that this would have suddenly stopped. Ian Copland notes along these lines: “If events had not intervened, the remaining states would probably have suffered the same fate as befell Satara, Jhansi, , Awadh, and the Punjab between 1848-1856 – absorbed into the expanding Indian Empire” (2002: 15).

However, I am not arguing that the British were simply randomly annexing territory even before 1857. They clearly had some criteria upon which they based policy; the main selection criteria which I highlighted in the Introduction was the availability of fertile land. Driven primarily by commercial interests, the British were looking for agriculturally productive territories, mineral wealth, and coastal areas which could facilitate the export and import of goods.

1.2 The Rebellion of 1857

Company political rule continued to strengthen in India in the 19th century until the Rebellion of 1857.7 This conflict began when sepoys – native Indian soldiers serving in the Company army – refused to use new rifle cartridges coated with either beef or pork fat, which offended both Hindus and Muslims on religious grounds. The British imprisoned several insubordinate sepoys which then resulted in the beginnings of a widespread army mutiny throughout the whole of north India. The rebels eventually marched to Delhi to restore the

7 Also known as the Mutiny, or the First War of Independence.

41 Mughal Empire under the leadership of the aging emperor Bahadur II.

Most historians offer different causal variables for 1857; some point to increased racial antagonism and fears of forced religious conversion to

Christianity (Mukherjee 1990) while others offer economic explanations focusing on declining sepoy wages relative to the cost of living (Streets 2001). The most important causal factor, however, was the policy of subordinate isolation and aggressive British annexation, exemplified by the introduction of the Doctrine of

Lapse by Dalhousie. The capture of two prominent territories in particular –

Punjab in 1848 and Awadh in 18568 – generated widespread discontent among the native population.

Whatever the primary causes of 1857, British power in India fundamentally altered in its wake. The conflict took a full year for the British to suppress, and control of India thereafter shifted from the East India Company to direct rule by the Crown. The Government of India Act 1858 was passed in

Parliament which contained a number of far-reaching reforms; most importantly, it led to the creation of a Secretary of State for India (in London), a Viceroy of

India (considered the Crown's representative in India), and a new bureaucracy named the Indian Civil Service. The British also took direct control of the colony; Queen Victoria was eventually styled as the Empress of India in 1876.

The final change occurred when the Mughal Empire was officially dissolved –

Bahadur Shah II was exiled to Burma after the conflict.

The Rebellion was a critical historical juncture in Indian history,

8 Neither Punjab nor Awadh were annexed using the Doctrine of Lapse; however, numerous other territories did lapse to the British. See Iyer 2010 for a full list.

42 essentially locking in place the boundaries between the British provinces and the princely states. As Queen Victoria stated in her declaration to the princes after the conflict: “…all treaties and engagements made with them [India’s princes] by or under the authority of the Honourable East India Company are by us accepted, and will be scrupulously maintained”. She further assured the princes that the

British sought “no extension of [their] present territorial possessions”. And in fact no further states were annexed after the Rebellion. Map 1 below showcases the divisions between British India and the princely states; the former are pink and the latter are yellow. British India constituted at this time roughly three- fifths of the area of the country and three-fourths of its population.

Map 1: British Provinces and the Princely States, 1909

43 1.3 The Post-1857 Period

After the Rebellion the boundaries between the provinces and the princely states were essentially made permanent. There were at one point over 600 princely states, although the vast majority of them were very small. Dick

Kooiman notes that the 24 states which received a gun salute9 of 17 or higher accounted for roughly 70% of the entire princely state population (2002: 18).

Princely India therefore referred both to massive kingdoms like Hyderabad as well as tiny states like Mandi or Faridkot. Table 1.1 highlights the 24 major states in India in 1931. Because of the immense variation among the native states it is hard to generalize about them. In this dissertation I selected cases from among the larger native states, and I find some important similarities between them which I highlight in the ensuing chapters.

9 Gun salutes were given to certain princely rulers by the British. The highest salute was 21 guns, and generally the higher the salute the more prestigious the state.

44 Table 1.1: Major Princely States in India by Gun Salute, 193110 Princely State Location Size Pop. Gun Salute Religion Hyderabad AP 82,698 14,436,148 21 Muslim MP 26,383 3,523,070 21 Hindu Mysore KN 29,475 6,557,303 21 Hindu Jammu and Kashmir JK 85,885 3,646,243 21 Hindu Baroda GJ 8,135 2,443,007 21 Hindu Travancore KL 7,625 5,095,973 19 Hindu Bhopal MP 6,902 729,955 19 Muslim Kolhapur MH 3,217 957,137 19 Hindu Indore MP 9,570 1,318,217 19 Hindu RJ 12,915 1,566,910 19 Hindu Kalat 73,278 342,101 19 Muslim Cutch GJ 7,616 514,307 17 Hindu Bikaner RJ 23,315 936,218 17 Hindu RJ 35,066 2,125,982 17 Hindu Jaipur RJ 16,682 2,631,775 17 Hindu Tonk RJ 2,553 317,360 17 Muslim Bharatput RJ 1,993 486,565 17 Hindu Bundi RJ 2,220 216,722 17 Hindu Karauli RJ 1,242 140,525 17 Hindu Kotah RJ 5,684 685,804 17 Hindu Patiala RJ 5,932 1,625,520 17 Hindu Bahawalpur Pakistan 15,000 984,612 17 Muslim Cochin KL 1,418 1,205,016 17 Hindu MP 13,000 1,587,445 17 Hindu

The British policy toward princes since 1813 had been to prevent them from controlling their foreign affairs and external communications – both undertaken to prevent them from fighting wars with other each other, or uniting against colonial rule. Many princes were also forced to provide a subsidy to the

10 Princely states in bold are case studies in this dissertation. Location refers to the modern location – some princely states overlap several modern states so I simply cite the largest one. Size is area in square miles. Religion refers only to the religion of the ruler.

45 British in return for military protection, and princes were still threatened with removal in the case of gross misrule. While some princes were deposed even after

1857, it is not clear whether this threat was taken seriously by most princes.11 As I will note in my case study chapters, the British made threatening claims but were often brushed aside by princely rulers. Hutchins posits that overall, due to British fears of another rebellion, colonial administrators:

“now actually set about resurrecting rulers of princely states and classes of landlords. No dereliction, incompetence, enfeeblement, or immorality would henceforth persuade the British to take the step which was taken in Oudh in 1856 when the debasement of the Oudh Court led to its extinction at the hands of Lord Dalhousie...their [the princes'] influence was declared to be sanctioned by the immemorial traditions of Indian society.” (1967: 171)

In the parlance of the times, the princely states were technically under the system of indirect rule, also known as paramountcy or . It was not always clear what this system meant; in 1928, in response to princely inquiries, the British defined the relationship between themselves and the native rajas rather obliquely: “Paramountcy must remain paramount” (Rudolph and Rudolph

1966: 139). However, the British were undoubtedly the dominant power in India.

In response to complaints from the , Lord Reading, Viceroy of India, famously responded: “The Sovereignty of the British Crown is supreme in India, and therefore no Ruler of an Indian state can justifiably claim to negotiate with the British Government on an equal footing” (Leonard 2003: 369).

11 Iyer notes that 21 princes were deposed (out of her dataset of 71 princely states) from 1858- 1947. But taken against the total number of princely states, 21 depositions for over 600 states over an 89 year period is a rather low figure (2010: 697). Furthermore, deposed princes were replaced with a family member, not a British administrator, so the princely line generally remained intact. Finally, during the period of 'subordinate isolation' a number of princely rulers – even when they governed well – were deposed by the British. Consider Fisher's discussion of Awadh in 1856 (1991: 385). So were princes forced to change their practices to better please the British? It seems doubtful.

46 All of this prompts the major underlying question: were the British, even after 1857, through various mechanisms able to exercise substantial control over the princely states? For example, the British instituted the aforementioned

Residency system in which advisors were sent to princely courts to offer advice to rajas and curb their excesses. During the early history of British power in India

Governor-General Warren Hastings even openly noted that “[i]n our treaties with

[princes], we recognize them as independent sovereigns. Then we send a

Resident to their courts. Instead of acting in the character of ambassador, he assumes the functions of a dictator” (Majumdar 1951: 15). Is it therefore not possible that even after 1857 the British remained firmly in control of the states?

While there is no universal agreement between historians of India on this point, I argue in this dissertation that princely rulers after 1857 – and especially in the early 20th century – generally had internal autonomy over their territories. First, it is worth discussing the progression of historical scholarship on this question. There was a period in which most historians of South Asia did not take princes seriously as independent rulers. At worst, rajas were even considered collaborators with British rule. Waltraud Ernst and Biswamoy Pati document “earlier, hegemonic [historical] accounts that depicted princes as at best mere decorative stooges of British imperial power” (2007: 3).12 Edward S.

Haynes perhaps best summarizes this position in writing that “[u]nder the

12 They make a very interesting point in Chapter 1 – they generally argue that princes were not independent, and that historians who argue that princes were autonomous “become, in an endeavour to rescue the princes from hegemonic accounts of history, victims of colonial ideology themselves” (4). As I noted, there is no real consensus on this issue, so I endeavor in the subsequent chapters to show the specific ways in which the princes negotiated, altered, or outright resisted colonial interference.

47 British, the rajas became princes fit only for the garden party...” (1990: 492).

On the whole, however, most historical scholarship on the princely states has cast much skepticism on the belief that princes were merely under the thumb of the British. There are a number of scholars who have, on the other hand, emphasized the importance of the princes as semi-independent rulers.13 Barbara

Ramusack (2004) elaborates this position, writing that while princes did not control their defense or external affairs, they did control tax collection, the allocation of state revenues, criminal and judicial power, internal law and order, and cultural activities and institutions. John Hurd II argues: “Nevertheless, in spite of British influence and control, whether potential or actual, the

Government of India rarely tried to force princes to alter their political systems to any significant degree” (1975: 171). John Wood writes about the Residency system specifically: “Ordinarily, however, the prince possessed preeminent substantive and personal power. It often appeared, in fact, that the Resident was more zealous of the prerogatives of 'his' princes than they themselves were”

(1984: 71). Furthermore, this position is not solely the view of modern historians interpreting Indian history – there is evidence even contemporary audiences understood princes to be semi-independent sovereigns. In 1911 The Times collected a series of articles for 'Empire Day' and in their reporting on the princely states they wrote:

“[Princes] are not independent sovereigns, for they cannot wage war, or form alliances, or maintain foreign relations, but they direct the internal government

13 Many historians working on the princely states have taken them seriously as semi- autonomous kingdoms; for example, Manu Bhagavan, Ian Copland, Michal Fisher, Dick Kooiman, Robin Jeffrey, Mridu Rai, Barbara Ramusack, etc.

48 of their States, and some have the power of life and death over their own subjects.” (Times 1911: 21)

Why did the princes retain this independence? There are four main reasons. First – and most importantly – the British after 1857 needed the princes.

As Lord Canning, Governor-General of India, put it using a phrase that would be oft-quoted, princely states were “breakwaters in the storm which would have swept over us in one great wave” (Kooiman 2002: 15). This is to say, most of the princely states had supported the British during the Rebellion, and in the post- conflict period they expected to be repaid for their loyalty. The Rebellion of 1857 had shocked the British, especially some of its more brutal episodes such as the infamous massacre of British civilians at Cawnpore (Kanpur). Fears of another rebellion were never far from the minds of British administrators in post-1857

India. Furthermore, in the early 20th century the British were also involved in two world wars, and they needed princely support during both of them.

Second, the British began to realize that princes could be an effective ally not only in preventing rebellion, but also perpetuating colonial rule in India – provided that they were kept happy. As Sir John Malcolm, Governor of Bombay, said in 1825, “the...security of our vast oriental possessions is involved in the preservation of native principalities which are dependent on us for protection...their co-existence with our rule is of itself a source of political strength” (Kooiman 2002: 15). This led to a policy of non-intervention in the princely states. By 1867 Secretary of State for India Stafford Northcote was suggesting to a British administrator that he should avoid lecturing a prince “...as

49 if he were a minor of whom we are the guardians” (Metcalf 1965: 227). In a 1905 speech in Udaipur Lord Minto stated that “[o]ur policy is, with rare exceptions, one of non-interference in the affairs of the States” (Majumdar 1951: 803).

The third reason for princely semi-independence was that the model of

'indirectly' ruling certain territories was recognized as being less expensive for the

British – monetarily, and in terms of sheer manpower. India was one of the largest colonies in the British system, and one of its most populous. Furthermore, its enormous cultural diversity made it a difficult colony to understand, and therefore govern. Along these lines, India presented very difficult natural terrain for the British – an inhospitable climate and a host of native diseases generated high mortality rates for settlers (Acemoglu et al. 2001).

The fourth and final reason for princely autonomy was that during the early 20th century, especially in the 1920s and 1930s, the British began the process of devolving authority to Indians in greater expansions of self-rule. While this was a much more complicated endeavor in the provinces, princely rulers were directly given more power in their territories. As Hydari, Prime

Minister of the Hyderabad State, summarized the situation in 1930:

“Luckily, for the States, the present situation is favourable. On the one hand, the pressure exerted by British India for devolution of power, all along the line, has awakened the British Government to their obligations to the States; on the other, the abandonment of the policy of isolating the States has enabled them to discover that their grievances are common and, on that common ground, to unite in pressing their demand for the removal of those grievances.”14

The reference to abandoning the policy of isolating the princely states refers to the fact that, prior to the 1930s, the princes were forbidden to form alliances.

14 APSA, Installment #47, List #10, File IV, 82

50 However, the 'Chamber of Princes' was formed in 1920 in order to appease the princes and to allow them to communicate on issues of mutual interest. British administrators on the whole recognized that they could not devolve power in the provinces without also doing the same in the princely states. As newspaper editor

Lovat Fraser wrote in 1911: “We cannot...on the one hand announce our intention of giving greater liberty to the people of British India, and on the other, turn the screw upon the Indian princes” (Copland 1973: 213). By the 20th century the situation had changed considerably; this period “witnessed a marked and notable revival in the authority, prestige and efficiency of State administrations”

(Majumdar 1951: 967).

All four of these factors created a situation wherein, post-1857 and especially in the 20th century, princely rulers had considerable internal autonomy in their territories. To be clear, I am not arguing that the British exercised no control over the princely states after 1857. In fact, in every princely state discussed in this dissertation the British did involve themselves in local affairs to varying degrees. But it is not true that princely territories were simply puppet states in which the British had a free hand to interfere as they pleased. As I will detail in my empirical chapters, the British were often surprisingly limited in what they could do in princely states, especially as time progressed. And at the absolute least, compared to provinces the princely states were undoubtedly areas where the colonial footprint was less pronounced.

Finally, it is worth noting that taking the princely states seriously as semi- independent kingdoms has several other merits that pertain specifically to

51 scholarship. As Hira Singh smartly notes:

“The notion of indirect rule is...a tool of the colonial mode of historiography. According to the notion of indirect rule, Indian States were ‘preserved’ by the colonial state for their instrumental value to the functional needs of metropolitan capital and the colonial state. However, as I have argued elsewhere (Singh 1998), the historic alliance between the colonial state and landed aristocracy (the ruling class in the princely states) in India was a compromise which the colonial state, given its vulnerabilities, was forced to accept in the face of resistance by the latter. The notion of ‘indirect rule’ eschews the element of resistance. While privileging the metropolis, it denies the colonized subjects (here the landed aristocracy) any agency, excepting as the instruments or ‘puppets’ of the colonial state. Repudiation of the notion of ‘indirect rule’ is therefore a necessary step towards de-colonizing the historiography of colonial India.” (203: 224)

Similarly, Karen Leonard writes:

“Historians of South Asia seem still obsessed with the British, with colonialism and postcolonialism, overlooking ways of interpreting the past and present that do not make colonial rule responsible for all that has happened. Indirect rule and the so-called princely or native states of India provide opportunities for comparative investigation of modern South Asian cultural configurations.” (2003: 364)

The princely states still remain, although now to a much lesser extent, on the margins of South Asian history. Yet, understanding the states is pivotal to understanding colonialism in India, and, more specifically for this dissertation, understanding patterns of ethnic violence in contemporary India.

II. Ethnic Policies in Colonial India

Bifurcated colonial rule in India led to two sets of authorities reigning in their respective areas; in the provinces British administrators governed whereas in the princely states native kings remained in control of local politics. Most importantly, this dichotomy led to disparate sets of ethnic policies. Before delving into the colonial construction of ethnic politics, a basic working definition of ethnicity needs to be established. I began a cursory address of this issue in the

52 Introduction, but I return to it here. This is an important exercise – if ethnicity is merely epiphenomenal or instrumental, as many constructivists have argued, then ethnic identities (and patterns of ethnic conflict) are likely to change easily.

In the next section I present a definition of ethnicity different from most existing literature, and I highlight several reasons why ethnic affiliations persisted long after their initial construction during colonial rule.

2.2 A Theory of Ethnicity

I follow Henry Hale in arguing that ethnicity is primarily about uncertainty reduction (Hale 2004, 2008). Individuals use identity categories to reduce the complexity of the social world, but ethnic identities are uniquely useful for a number of reasons. Ethnic groups band together over myths of common origin, a sense of a common fate, symbols and culture, physical similarities, and they face reduced barriers to communication. Ethnicity therefore connotes a way of thinking about and interpreting the social world. As I will detail in the next section, both British administrators and princely rulers conceived of ethnicity and ethnic politics in different ways; while the British emphasized caste, princely rulers emphasized religion. They then both tried to construct ethnic policies in ways which reinforced this key distinction.

Therefore, I deny that ethnicity is epiphenomenal or merely instrumental.15 Rather, once ethnic identities are constructed they often become internalized by individuals. What I am putting forward, then, is a 'thick' account of ethnicity. For example, it was widely understood in British provinces that caste

15 In fact, Hale (2008: 50) argues that ethnicity is “pre-rational”.

53 was the central mode of ethnic classification, and that religious distinctions were secondary and of less importance. This ethnic category became embedded in local institutions and power structures, and came to eventually define British India; in short, caste became the commonsensical mode of ethnic classification in the provinces. This naturally extended to ethnic violence – caste conflict in the provinces was deemed to be a legitimate kind of conflict whereas religious conflict was, on the other hand, considered illegitimate. In the next section I delve into the disparate ethnic policies of British administrators and princely rulers, and how they institutionalized the salience of these different kinds of ethnic categories.

2.3 Ethnic Policies in the Provinces and Princely States

After the Rebellion there were two distinct Indias: British India and the

India of the princely states. This was a pivotal period in the politics of the subcontinent, a period in which the main ethnic cleavages which still dominate

Indian politics were constructed. The fact that British India and princely India remained separated for almost a century has had long-lasting effects on many aspects of contemporary Indian politics. Most important for this dissertation, however, is the construction of ethnic politics in both areas. British administrators and princely rulers had very different ideas on the issue of ethnicity, and how ethnic politics should be organized. I argue here that the

British in the provinces emphasized the importance of caste, and discriminated against low castes and tribal groups while protecting religious minorities. The situation in the princely states was the exact opposite: princely rulers emphasized

54 religion, and discriminated against non-coreligionists while protecting low castes and tribal groups. This dichotomy still forms the basis of the fault lines of ethnic conflict in modern India, and is diagrammed in Figure 2 below.

Figure 2: Ethnic Policies in Colonial India

BRITISH PROVINCES PRINCELY STATES

High Castes Dominant Religious Groups

Low Castes / Tribals Minority Religious Groups

Favorable Policies Discriminatory Policies Protectionist Policies

What led to these specific conceptualizations of ethnic politics? To some extent, this is simply the way British rulers and princely rulers comparatively thought of society – their understandings of ethnic relations, even if based on nothing concrete or firm, were decisive in the construction of their policies. In the ensuing sections I delve into this further, and discuss the specific configurations of ethnic politics in the provinces and the princely states.

2.4 The Provinces: Policies on Low Castes and Tribals

After the Rebellion the British were confronted with the fact that their knowledge of India was perilously lacking. After the conflict ended in 1858, colonial administrators undertook a large project to gain information and knowledge about India, mainly done through massive data gathering projects

55 (Cohn 1996). In their quest to better understand their prized colony, the British reached several important conclusions. Foremost among them, the British came to understand caste as the central organizing principle of Indian society. As

Nicholas Dirks writes, “...under the British...'caste' became a single term capable of expressing, organizing, and above all 'systematizing' India's diverse forms of social identity, community, and organization” (2001: 5). In short, caste became the primary means of classifying an overwhelmingly complex society. Herbert

Risley, director of the census and himself extremely important in reifying caste identity, wrote in 1915 that “...caste...is regarded as the natural law governing human society” (Saumarez Smith 2008: 70).16 Caste was also a convenient analogy for many British officials who were used to the hierarchical class structure of Victorian society. Along these lines, many of the British administrators who came to control India were inclined to believe that caste was not only a mode of classification for society, but also a method of social control.

Caste was, unlike religion, a hierarchical or ranked system. Therefore, through the use of the caste system the population of India could be divided between, at two extremes, elites and those who were considered to be barely human. Caste was a means through which low castes, for example, provided a massive supply of free labor as the servants of British administrators. Hutchins notes that this appealed to many of the officials who came to India:

“...it was the autocrat who sought in India a field to fulfill ambitions stymied in England. India attracted a person who was disturbed by the growing democratization of English life…Once the target of reformers, India had now become the hope of reactionaries.” (1967: 10-11)

16 I discuss how tribes fit into this discussion a bit later.

56 Caste actually refers to two specific systems: varna and jati. The varna system contains four broad categories: Brahmins, Kshatriyas, Vaishyas, and

Shudras. Brahmins are priests; Kshatriyas are kings or warriors; Vaishyas are agriculturalists or merchants; and Shudras are laborers. Untouchables are below these groups and technically outside of the structure of the system – they do work that is considered ritually unclean (such as removing animal carcasses). Jati, on the other hand, refers to a birth group which is an endogamous community – it is utilized much more than varna throughout contemporary India (Bayly 1999).

Furthermore, unlike the broad typology associated with varna, there are thousands upon thousands of jati groups all over the country.

The British certainly did not conjure caste out of thin air. It is an ancient system of classification that is partly derived from Hindu religious texts, although these same texts have also been used to justify opposition to the caste system.17

The word itself is Portuguese in origin, casta, a term that refers to animal husbandry. The centrality of caste in Indian society prior to British arrival is a question that still divides historians. However, most scholars accept that the

British at the very least placed a new and distinct emphasis on caste and caste relations.18 Partly this was due to the fact that the British centralized power in

17 There are a wealth of opinions on this point. Some scholars argue that the caste system is supported by ancient Hindu texts such as the Bhagavad Gita and the Laws of Manu, whereas others argue that the caste system is not justified by Hinduism, or was never intended to be a central facet of Hinduism. See Bayly 1999: Chapter 1 for an historical discussion on this. 18 This view is widespread among scholars of colonial India; just as a small sampling see Cohn 1996, Bayly 1999, Dirks 2001, Banerjee-Dube 2008. For example, for a long time Brahmins were considered lower on the caste hierarchy than Kshatriyas, but once the British arrived in India they relied on the former group for interpreting Hinduism, its ancient texts, and the caste system. Brahmins naturally argued that they were at the top of the system, and this view still remains widespread today (Bayly 1999).

57 India in a way no previous ruler had been able to accomplish.

To be clear, I am not arguing that religion was not an important ethnic category in the provinces. The British definitely placed an importance on religion; they believed, for example, that Hindus and Muslims were two ancient and distinct groups who could not be reconciled. However, the British were a secular power – they pursued a policy, as I will detail later, of religious neutrality, and administrators always tried to subvert the sacred to the secular. As Mridu Rai writes, “...in British India religion was, even if only theoretically, relegated to a space subordinate to the ‘secular’ and the ‘public’” (2004: 93). Similarly, she notes that Queen Victoria's proclamation after the Rebellion “was implemented through the construction of a colonial state claiming to arch loftily above religion...” (82). This was not a British policy that was specific to India either –

David Laitin (1986) notes that British administrators also sought to 'expunge' religion as a legitimate mode of ethnic classification in Yorubaland in Africa

(1986: 150-158). Furthermore, colonial administrators prioritized caste over religion; Risley himself seemed to believe that Hinduism was a component of the caste system. He was quoted in The Times stating that Hinduism has an

“...intimate connextion with, or rather dependence upon, the social institution of caste” (Times 1911: 198, emphasis added). Max Müller, noted scholar of India and well-known among the British public, similarly noted that “Modern Hinduism rests on the system of caste as on a rock which no arguments can shake” (Ibid).

Religion was considered to be an ethereal category; caste, on the other hand, was scientific, concrete, and specific.

58 Beginning from the assumption that caste defined India, British administrators then emphasized, reified, and rigidified the caste system through various policies. The central mechanism through which this occurred was the census, first carried out in 1871. The entire process of classifying castes had a significant effect on, as G.S. Ghurye put it, “a livening up of the caste-spirit”

(Cohn 2008: 30). The census was seen as an important tool in creating the foundations for a new social order, one which was, Dirks notes, removed from its religious and supernatural origins (2001: 51). There was a high level of experimentation with the census – the 1871 round used varna as the main component of caste, whereas by 1891 census administrators switched to focusing on occupations. W.C. Plowden, a census commissioner, during the 1881 round focused only on castes larger than 100,000 individuals. One main result of the census was that caste began to mirror the whims of colonial administrators more than it did the system which had actually existed prior to their arrival.

Another important aspect of British rule in heightening the salience of caste was the attempt by the colonizers to link with dominant castes. As British power grew in India, so did its relationship with dominant castes throughout the subcontinent. Partly this was a need on the behalf of British administrators to link with the pre-existing elite of the country. As Charles Canning wrote to

Charles Wood in 1860 after the Rebellion, “If we are wise, we shall lose no time in binding to ourselves the Chiefs, the Landholders, great and small, and the wealthy classes...by stronger and more substantial ties. We must impress them all with this plain conviction that, come what may, the fall of our power will be no

59 gain to them” (Metcalf 1965: 237). Many of the dominant castes linked with the

East India Company, buying official titles and ranks from the company in order to strengthen their control over a country in flux. This led subsequently to increasing stratification among caste groups.

As the power of dominant castes grew, however, the position of low castes deteriorated. This was not only a function of increasing political, economic, and social disparities between high and low castes – the colonial state also actively enforced policies that discriminated against them. I consider these groups in the subsequent analysis. I focus on two main mechanisms which led to their declining power in colonial India: the census and the subsequent intensification of caste antagonisms, and new landholding and tax policies.

As the census was instituted in colonial India, the position of low castes deteriorated even from their decidedly poor starting point. The census officially designated certain groups as low castes. Prior to this, these groups had been relatively fluid and there was even some mobility among the lower rung of the caste system (Cohn 2008, Mayaram 2008). But the census hardened caste identities and this naturally led to increased antagonism as low castes fought to achieve higher status, or as some caste groups tried to designate rivals as inferior.

For example, Commissioner Plowden remarked that low castes would be easy to categorize as those groups “which the popular voice designates as inferior” (Cohn

2008: 33). The 1901 census was also pivotal in that it began to rank caste groups, a practice which was later disavowed although its reverberations were still felt long afterward. Because of the political nature of rankings, which included

60 jockeying for position, the mid-19th century saw an explosion in the number of narrow caste organizations and associations (Rudolph and Rudolph 1960).

British administrators themselves were cognizant of the overall effect which their census policies had on the caste system, and low castes in particular.

Mr. Middleton, a census commissioner in Punjab, wrote in 1921 that:

“Our land records and official documents have added iron bonds to the old rigidity of caste. Caste in itself was rigid among the higher castes, but malleable amongst the lower. We pigeon-holed every one by caste, and if we could not find a true caste for them, labelled [sic] them with the name of an hereditary occupation. We deplore the caste system and its effects on social and economic problems, but we are largely responsible for the system we deplore. Left to themselves such castes as Sunar and Lohar would rapidly disappear and no one would suffer...Government's passion for labels and pigeon-holes has led to a crystallization of the caste system, which, except amongst the aristocratic castes, was really very fluid under indigenous rule...” (Ghurye 2008: 43, emphasis added)

The second major change for low castes engendered by British rule dealt with landholding. Prior to British rule, the Mughal system of landholding was, in the phrasing of Barrington Moore Jr., essentially an Asian form of royal absolutism (1966: Chapter 6). It is important, therefore, to delineate the specific ways in which the British changed – and worsened – the existing agricultural system of the Mughals. I highlight three main changes. First, the British transformed the zamindars who had been revenue agents of the Mughals into private property holders. Second, the British introduced a whole onslaught of new taxes and revenue burdens on rural India, increasing inequality and promoting violence. Finally, the new court system provided a legal basis to uphold these discriminatory policies.

The zamindars of India were originally intermediaries between Mughal

61 rulers and the peasantry19. They collected taxes and supervised the cultivation of agricultural properties. With the rise of the British, however, the zamindars were transformed into a powerful class of landlords. The Permanent Settlement in

Bengal formed the foundations of this experiment. The thinking behind this policy lay in English liberal ideas about the importance of proprietary rights and the need for the creation of a local landed elite (Metcalf 1965, Guha 1981). The

British were actually interested in improving the land – they reasoned that by granting proprietary rights to landlords, they would, in turn, reinvest in the land and ultimately improve it. As Eric Stokes writes, “...the British believed they were innovating...[they] proclaimed their purpose as being not to overturn existing rights but to give amplification and legal certainty to rights that had hitherto remained vague and inchoate” (1978: 2).20 But it was not clear that they had properly understood existing rights – in Chapter 4, for example, I detail how

British administrators completely misunderstood the nature of agrarian relations in Malabar with devastating results. The rule of zamindars – and I highlight this in all of the subsequent case study chapters – intensified peasant unrest. The granting of the proprietary right led to land concentration in the hands of a smaller number of individuals and increasing levels of displacement among

19 I use the term peasantry often in this chapter, mainly to refer to low castes and tribals. This is because, as Debal K. SinghaRoy has argued, “Peasants in India broadly represent a vast mass of landless agricultural labourers, sharecroppers, tenants, poor artisans and small and marginal cultivators having a close social interface with the socially deprived, such as the scheduled tribes, scheduled castes, other backward classes and women. The so-called 'outcastes' of the Varna hierarchy in the real sense of the term form the core of the peasantry in rural India” (2005: 5505). Of course, other groups – like Muslims for example – could also be peasants (see Chapter 4), but I use the term peasant mainly to refer to low castes and tribals. 20 Stokes also notes that by the mid-19th century, most British administrators believed that prior to colonialism land in India was jointly managed by village leaders (pg. 4).

62 peasants. By 1963, for example, some 2 million individuals out of a total population of 63 million controlled all of the land in the United Provinces (Ibid.:

205). Peasants became tenants-at-will, and land evictions increased throughout

India during the colonial period (Gough 1974).21

The zamindari model was not implemented everywhere in India, however.

Other models also existed – for example, throughout the south the British implemented the ryotwari system in which revenues were collected directly from peasants. In other areas the British gave control over the land to village communities (mahalwari). But there was, throughout India, a high degree of experimentation with land policy. Karl Marx even briefly commented on the subject in Das Kapital:

“...the history of the English in India is a string of futile and really absurd (in practice infamous) economic experiments. In Bengal they created a caricature of large-scale English landed estates; in south-eastern India a caricature of small parcelled properties; in the north-west they did all they could to transform the Indian economic community with common ownership of the soil into a caricature of itself.” (Chandra 1999: 5-6)

The zamindari system was the most detrimental for peasants, but the rapid implementation of different tenurial models also led to increased hardship. As I will note in Chapter 3 about Ajmer Province in Rajputana, the British changed the tenurial model several times during the 19th century, a policy which bred resentment among the agricultural population.

Tied to these new landholding policies were heavily increased British revenue demands. Colonial administrators placed enormous new burdens on

21 For some specific figures on eviction rates, see Chapter 4, section 5.2 on the peasants of Malabar.

63 rural economies, specifically in terms of the taxes which were collected. This was even a sharp change from the 'royal absolutism' of the Mughal period; for example, during the reign of Emperor Aurangzeb crop failures often led to remissions and extra produce was left with the cultivator in these instances

(Majumdar 1951: 533). Under the British, however, the rural population was not spared – colonial administrators established fixed land tributes. As a result, famines on the subcontinent became almost endemic in the colonial period, such as the Bengal Famine of 1943 which led to several million deaths. Further, while

Indians perished from starvation the British government continued massive crop exports from the country.22 As I note in my discussion of Ajmer and Malabar, two

British provinces, colonial administrators enforced new revenue demands that increased famines. But increased revenue was only one demand – colonial administrators also stipulated that provinces begin producing cash crops such as indigo, sugar, and rice for export, as well as demanding the opening of domestic markets for English goods (sans import taxes) which subsequently destroyed a number of native industries.

Finally, the new judicial system implemented by the British was used by dominant castes to provide legal justification for their rule and its various attendant policies. Courts became, for instance, the main tool utilized by landlords to displace peasants from their land.23 British land reforms had often empowered zamindars not only as landlords, but also as the final judicial

22 Davis 2001, quoted in Lange 2009: 1-2. 23 A number of scholars have made this point – see for example Dhanagare 1977: 119, Hardgrave Jr. 1977: 63, and Prakash 1988: 56.

64 authority in their lands. In Chapter 4, I demonstrate how the new judicial system in Malabar was used to massively increase the number of evictions of Muslim peasants during the 19th century. Because these new court systems were captured by high castes, they were used as an instrument to undermine the position of the peasantry.

On the whole, then, British rule caused a severe deterioration in the position of low castes. Furthermore, this occurred in a way which was different from previous rulers of the subcontinent. New technologies of colonial rule – designed to derive knowledge, or modernize the agricultural sector – had several deleterious effects for low castes. Land alienation increased, as did rural inequality; tensions blew up between landlords and peasants. In short, it was the low castes, paraphrasing Barrington Moore Jr., who paid the price of progress.

While the British understood caste as forming the bedrock of Indian society, what precisely was the role of tribal identities? The relationship between between caste and tribe is complex – some scholars contend that adivasis are not actually outside of the caste system. For example, Alfred Gell in his study of the adivasis of Bastar notes that “India's 'tribals' (or Adivasis) are not really tribes...at all, but are numerically dominant agricultural castes which hold, or used to hold, land in clan-based village communities in the more remote, forested and hilly parts of the subcontinent” (1997: 433). Other scholars argue that tribes have always existed outside the caste system and are functionally distinct from castes. V.C. Simhadra, for example, notes that tribes – unlike castes – do not have links to any specific occupation (1979: 2).

65 Whatever the case, the British did at the very least believe in a clear linkage between the two ethnic categories; District Gazetteers, for example, always recorded “Castes and Tribes” together, suggesting that British administrators perceived some kind of kinship between the two. Other important government reports and periodicals did the same – for example, Edgar

Thurston's 7-volume Castes and Tribes of Southern India and H.H. Risley's The

Tribes and Castes of Bengal. Even if tribes were considered distinct from the caste system, the British understood that the two concepts were related and lumped them together schematically. Along these lines, discriminatory policies toward low castes were also matched by discriminatory policies toward tribals.

The position of adivasis in India deteriorated considerably after the inception of colonial rule. Prior to British arrival, it is unclear the extent or nature of the relationship between adivasis and pre-colonial governments. Christoph von Fürer-Haimendorf, noted tribal anthropologist, has written about the relationship between tribals and the Mughal administration:

“Now and then the campaign of a Mughal Army extending for a short spell into the wilds of tribal country would bring the inhabitants briefly to the notice of princes and chroniclers, but for long periods the hillmen and forest dwellers were left undisturbed. Under British rule, however, a new situation arose. The extension of a centralized administration over areas which had previously lain outside the effective control of princely rulers deprived many of the aboriginal tribes of their autonomy...” (1982: 34)

The adivasis may have benefited from the fact that Mughal authority was not highly centralized, but this changed during the British period. The adivasis, many of whom were congregated in forests and jungles, attracted the interest

(and ire) of a colonial government intent on expanding control over land and

66 extracting mineral wealth. In this realm I highlight three specific policies which harmed India's tribal population: the British taking direct control over forests, heavy immigration into adivasi lands that led to tribal land displacement, and the establishment of Criminal Tribes legislation.

The British colonial government became the conservator of India's vast forests in 1878 through the Indian Forest Act. This was a singular act in the history of the country – never before had any native power claimed ownership of forests. Adivasis were therefore transformed into tenants-at-will on land which they had occupied relatively unfettered for centuries. As Ramachandra Guha and

Madhav Gadgil note, “...by one stroke of the executive pen [this bill] attempted to obliterate centuries of customary use of the forest by rural populations all over

India” (1989: 145). The bill allowed the government to designate certain areas as

'reserved' forests. These areas were cut off from local populations, and were opened to commercial interests (for example, the timber trade). Later reforms to the system of forestry rights made minimal changes – take, for example, the

Indian Forest Act, 1927, which “contained all the major provisions of the earlier

Act” (Kulkarni 1983: 192).

Second, tribal land displacement skyrocketed under the colonial state. An influx of new groups led to the displacement of adivasis from their land – immigrants, moneylenders (with usurious rates), and British officials all contributed to a steady encroachment on tribal lands. As K.S. Singh has noted,

“the self-contained subsistence economy based on agriculture and forest produce was weakened and destroyed” (Hasnain 2007: 277). Illiterate adivasis were often

67 also bilked out of their land, and colonial administrators were often complicit in the process (Ibid.: 224). In Chapter 5, I detail many of these processes in the tribal kingdom of Bastar.

Finally, the British implemented Criminal Tribes legislation, laws which governed the actions of tribes who had a history of criminal behavior (Simhadra

1979, Radhakrishna 2001, Hasnain 2007). But this legislation did little to blunt the problem of criminality among tribes – for example, it criminalized every member of a Criminal Tribe, including those who had never personally committed a crime. These tribes were also heavily stigmatized, prevented from free movement, and were often jailed for not reporting their movements to colonial authorities. In 1949 A.V. Thakkar, a member of the Bombay Depressed

Classes and Aboriginal Tribes Committee, noted that “it may be said that a much larger percentage of these tribes were led into criminality by giving them the stigma of criminal tribes” (Hasnain 2007: 237). Barbara and Thomas Metcalf also noted the highly imprecise nature of designating a tribe as criminal:

“The first major compilation of such photographs was The Peoples of India, published by the Government of India in 1868 in eight volumes. The picture of the Banjaras, for example, nomadic herdsmen and traders...was accompanied by a description of them as having ‘a reputation for perfect honesty’, but they were later relegated to the status of ‘criminal tribe’, a reminder of the fantasy that passed for exactitude.” (2001: 112)

Aside from the three central mechanisms I have highlighted here, there were also a number of smaller but no less significant changes to the lives of adivasis under British rule: the colonial state replaced the customary tribal barter system with a monetary economy, tribal law with state decrees, and local panchayats (village councils) with bureaucracy.

68 All of these new discriminatory policies against low castes and adivasis led to a sharp increase in the number of peasant revolts in colonial India.24 There is a paucity of data on pre-colonial peasant revolts, but Kathleen Gough notes that

“...British rule brought a degree of disruption and suffering among the peasantry which was, it seems likely, more prolonged and widespread than had occurred in

Mogul times” (1974: 1392). Peasant revolts became a routine occurrence during the British period: in 1855 the Santhals rebelled; in 1868 the Naikdas; in 1873 the Kolis; in 1875 the massive Deccan Riots occurred; and in 1895 the Birsas revolted. This is only a very small smattering of the total number of conflicts.

Ranajit Guha (1999) has documented over 110 different colonial-era peasant revolts. Gough (1974) records at least 77 since the advent of British rule.

Ultimately then, the colonial state created caste as the central organizing principle of the provinces, and tribal identity was also lumped together with it.

This emphasis on castes and tribes had a long-lasting effect: the fault lines for future ethnic conflict would revolve around these specific identities. Colonial administrators enforced policies which benefited high castes, but discriminated against low castes and tribals. The result was an increase in caste violence in the

British provinces, and this is a legacy which has persisted into modern India.

2.5 The Provinces: Religious Neutrality

As the British colonial state pursued policies that increased caste antagonisms in the provinces, a very different set of policies were simultaneously being implemented in the realm of religious policy. The first term that

24 For excellent overviews see Gough 1974, Stokes 1978, Guha 1999; specifically on tribal revolts see Simhadra 1979, Mathur 2004, Hasnain 2007, Desai 2007.

69 presumably comes to mind with respect to British religious policy in India is

'divide-and-rule'. According to this theory, the British divided Hindus and

Muslims for political reasons, thereby promoting antagonism between them. The rise of communalism throughout India is therefore taken as an enduring legacy of

British colonialism. It is true to some extent to suggest that the British pursued an agenda which heightened divisions between Hindus and Muslims, though mainly because they perceived the two to be ancient, irreconcilable communities.

The British established notions of Hindu law versus Muslim law, for example, and did allow for the creation of separate Hindu and Muslim electorates.

However, on the whole there a number of reasons to be skeptical about the divide-and-rule argument.

Foremost, as Cynthia Keppley Mahmood notes, this hypothesis rests on a shaky foundation:

“The political myth of communal conflict today is that it is a modern phenomenon out of sync with the rest of Indian history. The cultural reality is that, insofar as its roots are in fact historically deep, communalism unfortunately represents a far more intractable problem than most current dialogue admits. An image of India based on the notion of consensus, deviations from which require explanation, must be replaced by a paradigm according full weight to the long- term dialectic of communalism that is, unhappily, showing no signs of abating.” (1993: 373)

This is an important point – the implication of the divide-and-rule hypothesis is that communal conflict was not a major feature of Indian politics prior to British arrival. This seems highly doubtful, as religious conflicts – even in the form of riots – occurred throughout the pre-British period (Bayly 1985). As Gopal

Krishna has noted in his detailed study of communal conflict throughout India,

Hindu-Muslim riots have existed at least as far back as the 18th century, and

70 probably even earlier (1985: 61). Furthermore, there is also the question of interpreting conflicts – do wars between the Marathas (Hindus) and Mughals

(Muslims) constitute communal conflicts? If so, then Indian history prior to colonial rule has been littered with religious bloodshed.

Nevertheless, is it right to say that the British pursued a divide-and-rule strategy? There are several problems with this argument. First, Hindus and

Muslims were already significantly divided prior to the arrival of the British. The divide-and-rule theory does not take seriously the very real differences that exist between the two communities. As historian Francis Robinson astutely notes:

“...the religious differences which separated [Muslims] from the Hindus were fundamental. The Hindus worshipped idols, the Muslims abhorred them. The Hindus had many gods, the Muslims had one. The Hindus revered the cow, the Muslims ate it.” (1974: 13)

Similarly, Mra Baig highlights innate and important differences in the two communities:

“What should be understood is that the Muslim belongs to a transnational community; the Hindu to a sub-national caste. To the Muslim, the Hindus next door is a fellow-citizen, but the Algerian, Indonesian and, even Pakistani, is a brother. To the Hindu, the Muslim next door is a fellow-citizen, but the caste- fellow on the other side of India is a brother. Psychologically, the epicentre of Islam is abroad, Mecca; the epicentre of Hinduism is in India, Benares.” (1974: 27)

Gopal Krishna Gokhale, a leader of the Indian National Congress, bemoaned these innate divisions, writing that “...over the greater part of India, the two communities [Hindus and Muslims] had inherited a tradition of antagonism which, though it might ordinarily lie dormant, broke forth into activity at the smallest provocation. It was this tradition which had to be overcome.”

(Majumdar 1951: 158, emphasis added). Both Hinduism and Islam at this time

71 were also going through revivalist periods.25 Hindu revivalism attempted to create a more explicitly Hindu state in India whereas Islamic revivalism led to increased calls for Muslim separatism. Both sides created a number of new associations, institutions, and began to focus more on conversion than they had previously. But the larger point is that Hindus and Muslims in India in the 19th and 20th centuries were already sufficiently divided, and needed no assistance from the British to exacerbate tensions between them. Copland writes that new historical scholarship has therefore “...effectively demolished the romantic notion that, prior to the coming of the British, the prevalence of tolerant attitudes towards variant beliefs ensured completely harmony between Hindus and

Muslims” (2005: 11). Ultimately, Robinson concludes about the policy of divide- and-rule: “Does this mean that the British divided and ruled? In the crude sense, no. There was no deliberate attempt to foster communal hostility; indeed the aim was to avoid it” (1974: 348-9). As an Indian delegate to the 1931 Round Table

Conference summarized it to a British administrator, “We divide and you rule”

(Kooiman 2002: 26, emphasis added).

Furthermore, both Hindus and Muslims through their political organizations played a large part in dividing the two communities, irrespective of

British policy. For example, the Indian National Congress contained a specific contingent of Hindu nationalists, which worked to alienate Muslims from the organization. The song Vande Mataram, for example, was often played at INC gatherings even though Muslims opposed it. On the other side, the Aligarh

25 For more on this see Copland 2005: 81-85, and Jaffrelot 2007: 7-10.

72 movement led by Sir Syed – who, as Rafiq Zakaria notes is, more so than

Muhammad Ali Jinnah, the real father of Muslim nationhood (1970: xxii) – did much to increase Muslim separatism. Sir Syed created the Aligarh movement partly to reestablish strong relations between the British government and

Muslims after the Rebellion. He pushed for Muslims to embrace Western education and argued that only through closer ties to the British government would they be able to survive in the modern era. Sir Syed “most of all widened the psychological gulf between Hindus and Muslims” (Baig 1974: 51). Zakaria notes:

“In this they [Muslims] were motivated not so much by any hatred of the Hindus as by an instinct of protection against the possibility of eventual domination by the majority community. They looked upon the British as arbiters of inter- communal disputes. As outsiders, the British, they believed, had no reason to destroy Muslim religious and cultural identity; but the Hindus, being the major community, having suffered oppression through many generations, were bound – so thought the Muslim leaders – to impose their own supremacy on others.” (1970: 355)

Furthermore, as Hindus and Muslims continued to grow apart during the 20th century the Muslim League eventually adopted a policy of demanding a separate state. Therefore, both sides, Hindu and Muslim, made several tactical mistakes during the decades preceding independence that exacerbated tensions and undermined the possibility of reconciliation.

Contrary to the idea of divide-and-rule, the religious policy of the colonial state was one of religious neutrality – of a government which would, in the famous words of Lord Curzon, hold even the scales between Hindus and

Muslims. The colonial state would govern as a secular authority, and would work to actively limit the domination of any one religious group in the public sphere

(Smith 1963). As Donald Smith notes, “As Britain's Indian empire expanded the

73 conviction deepened that religious neutrality was sound imperial policy as well as good business” (1963: 68). Even Queen Victoria's 1858 proclamation made special note of this policy:

“We declare it to be our royal will and pleasure that none be in anywise favoured, none molested or disquieted, by reason of their religious faith or observances, but that all alike shall enjoy the equal and impartial protection of the law; and we do strictly charge and enjoin all those who may be in authority under us that they abstain from all interference with the religious belief or worship of any of our subjects on pain of our highest displeasure.”

Dick Kooiman further elaborates on this point by looking at the specific religious policy of the :

“We may conclude that the British Government in Madras had arrived at a reasonably firm policy of not involving itself in matters of religion and of upholding a stance of imperial arbiter in secular matters. The state was firmly in British hands and there was little ground for any community to see the ruler- subject relationship in a proprietorial manner, even though some groups were strategically better placed than others. The British were convinced that this neutral bureaucracy was a major factor in the discouragement of communal strife.” (Kooiman et al. 2002: 56)

The policy of religious neutrality had one important corollary, however – protectionist policies for religious minorities, which in most cases meant

Muslims. What endeared Muslims to the British rulers of India? There were a number of factors. In one sense, the British may have sought to promote the cause of Muslims as a counterweight to the power of Hindus, and Hindu nationalism. Though the British linked with high caste groups throughout India, they were not uncritical toward the idea of a 'Brahmin' state (Bayly 1999: Ch. 2).

But, more than this, the British were driven by their recognition of the importance of Muslims as the previous rulers of the country, as the natural governing elite of India. For example, when meeting with Lord Minto before the

74 Morley-Minto Reforms of 1909, the Muslim delegation demanded that:

“...the position accorded to the Mohammedan community in any kind of representation, direct or indirect, and in ally other ways affecting their status and influence, should be commensurate not merely with their numerical strength but also with their political importance and the value of the contribution which they make to the defence of the Empire” (Majumdar 1951: 128)

Minto was also requested to consider “the position [Muslims] occupied in India a little more than a hundred years ago” (Ibid). The British were sympathetic to this argument. As Peter Hardy observes:

“British servants of the East India Company were intensely conscious that they were the successors of Muslims as rulers in India...In British India many Muslims' belief that they were the natural ruling aristocracy of the country gave them a confidence in politics all the greater because they knew the British shared their belief.” (1972: 11)

Francis Robinson, whose work focused on the United Provinces, has pointed to the strength of Muslim representation in the region – in 1882, despite being only

14% of the population they held 45% of the uncovenanted service posts (1974: 22-

3). He writes that “[i]t is not remarkable that so many Muslims held government posts. Government service was the traditional occupation of many Muslims; they had dominated the province for a long period...” (23). I further show in my case studies that in every British province under investigation, Muslims were greatly overrepresented in government service and were generally afforded a privileged position in society.

Second, there was also the fact that – according to many colonial administrators themselves – the British were simply “mortally afraid of the

Muslims” (Ibid.: 348). As British administrator W.S. Blunt once noted: “In

England we were perpetually scared at the idea of a Mohammedan rising in

75 India, and any word uttered by a Muslim was paid more attention to than that of twenty Hindus” (Majumdar 1951: 300). Secretary of State for India Lord Olivier also similarly noted that “No one with a close acquaintance of Indian affairs will be prepared to deny that on the whole there is a predominant bias in British officialdom in favor of the Moslem community” (Ibid: 270-1). Part of this fear was rooted in the fact that the British believed that Islam, more so than Hinduism, lent itself to 'fanaticism', and that Muslims therefore needed to be kept happy.

British support for Muslims, however, was not uniformly steadfast throughout the duration of colonial rule. There was a wide perception among

British administrators after the Rebellion of 1857 that it had essentially been an

Islamic uprising – they were certainly aided in this perception by the rebels choosing to reinstate Bahadur Shah II, the Mughal Emperor. After the Rebellion the British distanced themselves from the Muslim population. The – and its bloody finale in Kerala, which I discuss in Chapter 4 – also led to a disruption among British-Muslim relations. In response to this, Sir Syed formed the Aligarh movement in order to ensure that Muslims repaired their relationship with the British government, and that the linkage between the two groups would remain strong going forward. Moreover, as independence loomed

Muslim leaders began to fear that they would be left, in the words of Muhammad

Ali, “at the mercy of an angelic majority” (Majumdar 1951: 152). As Nawab

Waqar-ul-Mulk noted in a public speech in Aligarh:

“God forbid, if the British rule disappears from India, Hindus will lord over it; and we will be in constant danger of our life, property and honour. The only way for Muslims to escape this danger is to help in the continuance of the British rule.

76 If the Muslims are heartily with the British, then that rule is bound to endure.” (Ibid.: 151)

Due in part to the Aligarh movement as well as other various circumstances, in the 20th century British administrators and Muslim leaders patched up their differences. Many of the same circumstances that revived the position of princes in the early 20th century also applied to Muslims – they came, once again, to be viewed as pillars of support for colonial rule (Metcalf and Metcalf 2001: 106). As the Islam Prachak remarked about the situation between the two groups in 1905,

“...under British rule we are undeniably dwelling in great peace and happiness”

(Kaura 1977: 202).

In 1947 the country was partitioned into India and Pakistan as British colonial rule came to an end. But did Partition serve as a major disruption between the colonial and post-colonial period? It is true that a large population of

Muslims left the country for newly-formed Pakistan. But it is not clear that

Partition affected the provinces more than the princely states, and this is the pivotal point. Partition ripped apart the whole of north India, and its reverberations were even felt in the south. If the argument is that the provinces have less religious violence today because the large-scale emigration of Muslims

'solved' the problem of Hindu-Muslim violence, then there are a number of reasons to be wary of this hypothesis. First, India still has a population of almost

140 million Muslims, one of the largest Muslim populations in the world. Second, thousands of Muslims also left their land and property in the princely states during Partition. Table 1.226 below highlights several princely territories which

26 Source is Hasan 1997: 181.

77 experienced large-scale emigration during this period.

Table 1.2: Princely Land Left by Muslims During Partition Agricultural Land Princely Left by Muslims State (Acres) Patiala 338,996 Nabha 53,045 Jind 36,060 Kapurthala 194,000 Faridkot 40,978 419,000 Bharatpur 117,000 Bikaner 1,052,801

Therefore, if Partition was a solution to the problem of communal violence in

British India, then what explains why the princely states would not have also benefited in the same way?

Along these lines, while the British have received a large amount of blame for Partition – specifically for the reckless speed with which they exited the country – it would be incorrect to solely blame them for the catastrophe. As

Copland notes, “[t]here is no evidence in the records that the viceroy

[Mountbatten] or his his close advisers, or for that matter any of the Congress politicians in his cabinet, had even the faintest apprehension of the scale of the calamity that was about to occur” (2002: 658). And, as I stated previously, both

Hindus and Muslims themselves had done much to ensure that Partition would become a reality.27 By the 1930s, for example, the Muslim League took a hard line 27 It is also worth noting that the view that the British were solely responsible for Partition has been challenged from interesting quarters. BJP leader Jaswant Singh, for example, in a 2009 book argued that the policies of were largely responsible for Partition. He was briefly expelled from the party after the book's publication.

78 against any outcome except the creation of a separate Pakistan state, formalized in the Lahore Resolution of 1940. Congress, which initially rejected this idea, eventually also relented.

Finally, it is important to recognize that the British policy of religious neutrality also affected the princely states. While Queen Victoria's proclamation after the Rebellion highlighted that the princes would remain independent sovereigns, she also intimated that the British would intervene in native administrations in order to “secure religious toleration for the subjects of the protected states” (Lee-Warner 1910: 307). In many of their treaties with the native states, the British made sure that laws establishing religious toleration were “written in large characters on the Statutes of the realm” (Ibid.: 199). As I will detail in later chapters, it was often the case that Muslims from the provinces protested the treatment of their religious counterparts in the princely states.

Therefore, the British rulers of the provinces advanced a set of ethnic policies which benefited high castes, discriminated against low castes and tribals, and protected religious minorities. Caste formed the bedrock of British Indian society, but a firm policy of religious neutrality maintained the position of

Muslims. In the next section I detail the very different ethnic policies instituted by native kings in the princely states.

2.6 The Princely States: Syncretism in Princely India?

Princely rulers had a very different conception of how ethnic politics should be organized in contrast to British administrators in the provinces. While caste defined the provinces, religion formed the foundation of princely society.

79 Again, I am not suggesting that princely rulers did not care about caste – in fact, most of them also made sure to emphasize their caste ties. But their religious identities were prioritized – as Copland notes, “princes, by definition, were practising Hindus or Muslims. Their rule had religious sanction” (1998: 211). He notes that the religiosity of princely culture was arguably “the antithesis of British

Indian secularism” (Ibid.). All of the princely states under examination in this dissertation match this criteria – the rulers of Jaipur, Travancore, Bastar, and

Hyderabad all emphasized their religious identities to a high degree, and more so than their caste identities. They relied, like most native rulers, on religion as the basis of their power (Ramusack 2004: 207). Furthermore, as Susan Bayly notes, religion was a firm and concrete ethnic categorization in the princely states, but caste was not:

“...pre-colonial kings and their subjects did not treat caste norms as one- dimensional absolutes, but as reference points to be negotiated, challenged or reshaped to fit changing circumstances.” (1999: 30)

The princely states were best described as theocracies, and princely rulers emphasized religion in their polities in various ways. The culture of the darbar revolved around ornate religious displays and elaborate religious functions. The chaplain of the court was, in Hindu kingdoms, a Brahmin, and they served as the main advisor to a . Young princes were also educated by Brahmins, giving these priests immense power throughout the reign of a king. Princely rulers made large donations to religious institutions, rescinded or heavily minimized taxes for them, and often granted religious groups large tracts of land for temples, mosques, and shrines. Many princely territories were governed explicitly by

80 religious law. Many states also enacted anti-conversion laws in order to maintain the strength of the dominant religious group in society. Raigarh created a law to this effect in 1936; Patna in 1942; Sarguja in 1945; Udaipur in 1946. Many of the states which allowed conversion still stripped the convert of many rights, even going so far as to remove them from their children. In the provinces, on the other hand, there were no restrictions upon religious conversion.

Princely rulers created religious polities, but in the 19th century the British also undertook the project of reifying the princely states in this fashion.28 Their efforts only further deepened the conception of the princely states as theocracies.

As Pamela Price argues, “Both imperial administrators and Western-educated elites came to assume that authoritarianism in government and community identification by religion were indigenous to India” (1993: 495). Rai notes that in preparation for Queen Victoria's Imperial Assemblage in 1877 the British created an official chronicle of India prior to colonialism, and in it they divided the country into different ages which were derived from the religion of rulers (2004:

89). This was a widely held view – for instance, administrator James Grant described to Warren Hastings in 1786 that the Muslim invasion of India “effected the total subversion of the ancient inferior empire of the Hindoos...” (Grant 1786:

107). Partly this also served the purpose of creating a traditional India to compare

(unfavorably) with the modern, secular British-ruled provinces. Ramusack notes, however, that “[in classifying states by religion, the British]…served to reify the princely states as political, religious and ethnic entities, with serious implications

28 This was, as Rai notes, “entirely in keeping with the practice of 'traditional' Indian rulers” (2004: 87).

81 for popular political activity in the twentieth century” (2004: 53). As I will detail, this emphasis on religion in the princely states had long-lasting and far-reaching effects.

Despite the obvious importance of religion in the princely states – especially the religion of the king – most scholars have argued that princely rulers embraced a policy of religious tolerance termed 'syncretism' (Bayly 1985, Bayly

1989, Burman 1996, Copland 2005, Ollapally 2008). There are at least two main definitions of this term. The first comes from Peter van der Veer, who defines it as “the borrowing, affirmation, or integration of concepts, symbols, or practices of one religious tradition into another by a process of selection and reconciliation” (1994: 200). This definition is rather non-controversial: there is little doubt that over centuries of co-existence the various religions of India have blended and melded practices together. Second, syncretism has also been used to refer to a general culture of tolerance that pervaded the history of the subcontinent (Bayly 1985, Burman 1996). This is often then connected to the tolerant religious policies of the rulers of the Indian states. Syncretism in this vein then means that princely rulers did not discriminate against non- coreligionists. In this dissertation I focus specifically on this latter definition.

Scholars argue that princely rulers were on the whole tolerant toward rival religious groups. However, I argue the opposite: while the British enforced religious neutrality, the princely states – though publicly embracing tolerant policies – were generally discriminatory to non-coreligionists.

First, how did syncretism supposedly come to exist in India? There are a

82 number of possible causal factors. Some have argued that syncretism is simply a cultural attribute of a religiously diverse and plural nation. India is the birthplace of Hinduism, Buddhism, Sikhism, and Jainism. The country also has sizable

Muslim, Christian, and Jewish populations – it is, in short, one of the most religiously diverse states in the world. The existence of so many disparate religions co-existing for such a long time has been one of the hallmarks of Indian history. The desire to blend them together is perhaps best exemplified by the

Mughal Emperor Akbar's religious doctrine Din-i-Ilahi, which combined together what he considered to be the best aspects of Islam and Hinduism. This situation also persisted into modern India. During the independence movement Congress leaders, especially Gandhi and Nehru, constantly emphasized the need to embrace all of India's various faiths. Gandhi embraced and lauded the philosophy of sarva dharma sambhava, or equality between religions. On many occasions he even referred to himself as an adherent of different religions. The same is true for contemporary leaders. In 2007 Prime Minister Manmohan Singh noted that:

“This universal worldview is the foundation of our composite culture. India has been home for centuries to all great religions of the world...But we have not merely tolerated each other. We have not merely learnt to live and let live. We have in fact learnt to live together, grow together, learn together. Even as each one of us remains devoted to our own individual faith, we have learnt to respect the faith of another. This has been the basis of our nationhood.”29

A more specific argument along these lines is that syncretism was an outgrowth of the processes of state formation. As Christopher Bayly writes:

“The imperatives of regional and local state building during the eighteenth century and the flexible, all-India warrior society to which they gave rise both tended to confirm or newly to create relationships between the religious

29 Full speech available at: http://pmindia.nic.in/lspeech.asp?id=524

83 practices of the major faiths.” (1985: 181)

Syncretism in this sense appears to have been a by-product of the intermingling of the various powers who came to control India over the centuries. Hindu kings, for example, often governed states with sizable Muslim populations and needed to devise policies for these religious minorities. Elite incorporation of non- coreligionists was the simplest way to achieve this end.

But it remains to be seen whether or not syncretism was in fact the policy of pre-British rulers. Consider at the very least one specific era: the Mughal period. There are data available on the integration of Hindus into the Mughal imperial system. Table 1.3 below details the strength of Hindus and Muslims who were Mughal mansabdars (rank-holders) during the reign of three emperors:

Akbar, Jahangir, and Shahjahan. Mansabdars were non-hereditary leaders in the

Mughal Empire, and they often commanded armies and governed territory.

Mansabdars were given different sets of rankings (the term mansab itself means rank), and it was presumed that Hindus were an important part of the system.

84 Table 1.3: Historical Communal Composition of Mughal Mansabdars30 Relative Position of Hindus and Muslims among Year Mansabdars Highest High Medium Percentage 1595 (M) 17 80 135 83% 1595 (H) 4 18 25 17% 1621 (M) 17 120 60 83% 1621 (H) 3 28 10 17% 1656-7 (M) 13 171 220 79% 1656-7 (H) 4 52 50 21%

These data cast doubt on the argument that the Mughals were very good at incorporating Hindus into the structure of the mansabdari system – in fact,

Hindus were relegated to a subordinate position despite being an overwhelming percentage of the population. As the data also show, they very rarely achieved the highest rankings within the imperial system. This discussion is not meant to suggest that Hindus were oppressed by Muslim rulers; rather, it simply highlights that accounts of the syncretic nature of pre-colonial rulers may be overstated.

As I noted previously, many scholars have argued that syncretism became the main policy of princely rulers. The main evidence for this is the way in which kings were known to patronize rival religious traditions. Copland writes:

“Every year in Gwalior the Hindu maharaja rode at the head of the tazia procession which marked the end of the Muslim festival of Mohurrum. The Jaipur government patronized Mohurrum as well as the tomb of the Sufi saint Mu'in al-din Chishti at Ajmer. In 1930, the Sikh maharaja of Kapurthala spent £30,000 on the construction of a mosque in his capital. In 1939, the nizam of Hyderabad endowed a chair at the Benares Hindu University.” (1998: 212)

Princely rulers also sometimes utilized the language of previous rulers (i.e. many

Hindu kings continued to use Persian as the language of their kingdoms even

30 All data from Char 1991: 159; M = Muslim, H = Hindu

85 after the collapse of the Mughal Empire) and also employed non-coreligionists in their governments. Sir , for instance, was a famous Muslim who served as the Diwan of the Hindu kingdoms of Mysore and Jaipur.31 Princely rulers are also said to have guaranteed freedom of worship to religious minorities, access to mosques and temples, and the ability to carry out religious festivals without interference. As I will note in the ensuing empirical chapters, princely states like Jaipur and Travancore have been assumed to be syncretic by a number of scholars working on these territories.

But, as Dick Kooiman wryly notes, “[t]his picture of social harmony in the states is too good to be true” (Kooiman et al. 2002: 49).32 The main problem is that scholars have been much more likely to look at official pronouncements or laws on the books rather than the actual religious policies pursued by princely governments. Naturally, most princely rulers were quick to adopt the language of tolerance and acceptance, often as a means to keep British intervention in their territories to a minimum. But were they actually syncretic in practice? Jaipur and

Travancore are good examples; they are widely considered progressive states on issues of religious freedom, but, as I will detail in Chapters 3 and 4, their real communal records differ, often considerably, from the conventional wisdom. In sum, there is a substantial difference between state policies and state practices that has often been overlooked by scholars of the princely states.

31 Sometimes princely rulers would name a non-coreligionist as diwan as a kind of token appointment. As Kooiman notes, even when this happened the bureaucracy was overwhelming staffed by coreligionists (Kooiman et al. 2002: 51). 32 However, to be clear: Kooiman does believe that on the whole the princely states were better for intercommunal relations.

86 Princes throughout India were much less tolerant of religious minorities than typically assumed. In fact, I argue that many of them practiced outright discrimination toward non-coreligionists. Aside from the cases analyzed in this dissertation, other examples throughout India emphasize this point. The rulers of

Kashmir – Hindu kings presiding over an overwhelmingly Muslim population – were brutally repressive. A contemporary account from 1941 notes: “The poverty of the Muslim masses is appalling. Dressed in rags and barefoot, a Muslim peasant presents the appearance of a starving beggar...Most are landless laborers, working as serfs for (absentee) Hindu landlords...Almost the whole brunt of official corruption is borne by the Muslim masses...” (Bose 2005: 16) Even the raja's Foreign and Political Advisor resigned in 1929, making a public speech stating:

“Jammu and Kashmir State is labouring under many disadvantages, with a large Mahommedan population absolutely illiterate, labouring under poverty and very low economic conditions of living in the villages and practically governed like dumb driven cattle.” (Bhattacharjea 1994: 63)

Aside from Jaipur, other states in Rajputana were also backward on religious policies. James Tod, a famed British expert on the region, noted that Zalim Sing, the raja of Kotah, was a “zealot in all the observances of religion” (Tod 1829:

420), and Tod also noted with incredulity about these princely states, “I have seen charters which, it has been gravely asserted, were granted by upwards of three thousand years ago!” (Ibid.: 470). Jai Singh, the ruler of Alwar, established Rama as the state deity in the 1930s, banned and Persian from all state schools, and even prohibited the wearing of beards (Mayaram 1997). The

87 states of Alwar and Bharatpur also became sites of massive communal violence during Partition. Kooiman notes overall that in most states “...in spite of all darbari rhetoric, institutionalised [communal] discrimination was rampant...”

(Kooiman et al. 2002: 52). As Donald Smith summarizes:

“The religious policies of the Indian states were a part of the heritage which India received with her independence in 1947. With few exceptions these policies were not nearly as conducive to the realization of a secular democracy as those evolved in British India.” (1963: 98)

The account being constructed, therefore, is one in which the provinces embraced a policy of religious neutrality whereas the princely states enacted discriminatory policies toward non-coreligionists. But there is one major problem with this conclusion: by most accounts the provinces experienced more religious violence – some scholars say much more – than the princely states. This position has not been challenged, and is often offered with little in the way of evidence.

For example, Asghar Ali Engineer notes that “[generally] the princely states were less prone to communal conflict” (1989: 2467). Even the Simon Commission

(Indian Statutory Commission), sent to India in 1927 to oversee government reforms, commented that the princely states were relatively free from violence.

How can this be reconciled with the argument made here that the policies of the princely states were worse for religious minorities?

I argue that, contrary to popular conception, it is not clear that the provinces actually experienced more religious violence than the princely states.

This fact has been assumed, but few scholars have actually undertaken a systematic comparison. One of the notable exceptions is the work of historian Ian

88 Copland. In a comparison of data on violence collected from the Simon

Commission, press reports, and various other sources, Copland finds that during the period 1920-40 the provinces “suffered between fifteen and eighteen times more [communal violence than the princely states]...” (1988: 213). Copland, furthermore, attributes this relative lack of communal conflict in the native states to the syncretic nature of princely rule.

But there are several major problems with Copland's comparison. First, it is worth considering whether provinces and princely states constitute a clean, apples-to-apples comparison. The greatest impediment to this comparison is that the princely states were autocracies and the provinces were proto-democratic.

Naturally the regime type will account for a huge portion of the disparity. In the princely states, riots could be minimized with a level of force and repression not possible in the provinces. Hutchins notes, for example, that British administrators understood this point well:

“Law enforcement as practiced in the princely states also drew substantial praise, which increased with the emergence of the Indian nationalist movement. Many Englishmen pointed to the way in which the princes dealt with political troublemakers in their own states and deployed the chaos which English administrators seemed incapable of checking in British India where legal process played into the hands of the agitators.” (1967: 162).

Therefore, even if there was a clear absence of religious violence in the princely states, this does not necessarily mean that religious groups did not want to instigate violence, merely that they were unable to do so. As Kooiman notes, “the autocratic milieu [of the princely states] did not allow for protest and gave short shrift to political troublemakers” (Kooiman et al. 2002: 51).

Second, Copland uses only one control variable in his analysis –

89 population. Yet, there are many more issues that might conceivably matter: the relative strength of the Muslim population, economic factors, infrastructural factors, etc. The need to control for a variety of additional variables necessitates the use of statistical analysis, which I utilize in Chapter 2. The results of my analysis cast significant doubt on the belief that the princely states experience less religious violence than provinces (at least in the contemporary period).

Third, it is not clear that reports of religious violence in the princely states were well-covered by the press. The states were 'backwaters', and the strength of the press was weaker there as compared with the provinces. Also, princes were known to ban newspapers which recorded unflattering portraits of life in the states, and riots would fit under this category. Finally, it is worth noting that riots were not uncommon in princely India. Table 1.4 below, for example, lists a very basic search of princely state riots which I conducted in the British Library. This is only a small sample of the total number of conflicts.

90 Table 1.4: Princely State Riots Records, British Library Princely State Year of Riot British Library Record Baran (Kotan State) 1947 IOR/R/1/1/4589 Baroda 1946 IOR/R/1/1/4487 Bikaner 1946 IOR/R/1/1/4470 State 1946 IOR/R/1/1/4450 1945 IOR/R/1/1/4282 Poonch 1945 IOR/R/1/1/4242 Hyderabad 1938 IOR/R/1/1/3089 Jind 1934 IOR/R/1/1/2624 Kashmir State 1933 IOR/R/1/1/2424 Alwar 1933 IOR/R/1/1/2395 Indore 1927 IOR/L/PS/11/276, P 802/1927 Porbandar 1895 IOR/R/2/574/7

Finally, even if Copland's analysis is correct, he notes that by the 1930s the princely states, even just accounting for population and nothing else, became as violent as the provinces (1988: 214). He quotes C.S. Ranga Iyer, a member of the

Indian independence movement, as stating that “There was a time when our politicians like Gokhale rightly used to take pride in Indian States being free from communalism which was a vice in British India...But the table appears to have been turned” (Ibid.). Furthermore, during Partition, princely states also accounted for an immense amount of communal violence – states in the Punjab, and Alwar and Bharatpur in Rajasthan saw wholesale ethnic cleansing (Mayaram

1997, Copland 2002).

This point has significant implications – as the former princely states today experience more violence than the former provinces, the rise of communalism may have been stunted in princely India until the 1930s. At that

91 time, as I detail in the ensuing chapters, princely rulers were given more internal discretion over their territories and began to harden in their discriminatory practices. Subsequently, communal violence increased, and this may have continued into the post-independence period. In the provinces, however, communal relations were also heightened, but then gradually receded over time.

As I will highlight in the provincial cases of Ajmer and Malabar, relations between Hindus and Muslims today are quite peaceful; in the latter case, which saw major Hindu-Muslim conflict in the early 20th century, communalism today is almost non-existent.

Ultimately, for all of these reasons it is not clear that the British provinces did in fact experience more religious violence during colonialism than the princely states. As Peter van der Weer observes:

“So much for contemporary syncretism and tolerance. There is no reason to argue that India was a syncretistic and tolerant society in the precolonial period but has become the opposite as a result of the colonial construction of communalism.” (1994: 201)

2.7 The Princely States: Policies on Low Castes and Tribals

There is a perception that the princes of India could be even more brutal toward the peasantry – low castes and adivasis – than British administrators.

For example, Ramusack notes, “Many observers claimed that princes extracted more from their peasants than did the British Indian Government, but that the peasants in princely states were ‘happier’ than those under colonial rule” (2004:

187). However, I will detail that princely rulers on the whole were better in their policies toward peasants than British administrators, though this certainly

92 depended upon the princely state in question. Most of the case studies in this dissertation confirm this argument. Take, for example, 'feudal' Rajputana, in which I highlight (Chapter 3) that the rulers of the Jaipur State fought against reactionary nobles and also enforced policies much more beneficial policies toward the peasantry than British administrators in neighboring Ajmer.

Similarly, in the repressive and backward regime of the nizam of Hyderabad, I highlight (Chapter 5) that despite embracing policies that were brutal toward non-coreligionists (Hindus), the nizam at the same time enforced some of the most progressive policies toward adivasis in all of India, and even some beneficial policies for low castes.

Princely policies toward low castes were generally protectionist. First, the princely states initiated the process of low caste uplift, and it was the British who followed suit. There are a number of reasons why the princes pursued these policies. First, peasants in the princely areas also often had a closer link to princes than zamindars or British administrators, presumably because the princes were the native rulers of India. Lord Lytton, Viceroy of India, remarked along these lines that “...the Indian peasantry is an inert mass. If it ever moves at all it will move in obedience, not to its British benefactors, but to its native chiefs and princes...” (Copland 1973: 210).

Consider also the changes which occurred and deteriorated the position of low castes in the provinces. In princely India, many of the same mechanisms were missing. The census which was conducted in British India was also implemented in princely India, although the effects of caste classification were

93 dampened. As I note in Chapter 4 on Travancore, the Hindu rulers of the state used the census to attack the position of non-coreligionists like Christians rather than as a means of increasing caste antagonisms. In terms of landholding, the princely states employed a variety of tenurial models, but princes generally treated cultivators on their lands as hereditary and did not displace them arbitrarily (Singh 2003). In the 20th century many princes began to establish ryotwari settlements in which land tributes were collected directly from cultivators, thus curtailing the intermediary groups that were a source of endless strife in British India. In many princely states there were hereditary nobles who controlled large portions of the land, and princes often had little control over these areas. The position of cultivators was generally worse in these noble- controlled lands, but princes eventually also mounted efforts to reform these territories as well. For example, in the 1930s Jaipur's raja Man Singh initiated a strict set of reforms for the nobles of a large principality within his kingdom called Sikar where peasants were constantly in rebellion (Rubin 1983). Baroda

State was also progressive on rural issues; it implemented several reforms geared toward the peasantry, lowering agricultural taxes and incorporating cultivators into a closer and more direct relationship with the government (Hardiman 1978).

During the same time Mysore, Bikaner, and Cochin all began to spend more state revenues on agricultural improvements such as irrigation networks (Ramusack

2004: 191).

The princely states also certainly had new revenue burdens placed on them by the British, but their taxes on the land did not approach the excessive amounts

94 dictated by British administrators. I detail this in Chapters 3 and 4, as the British- controlled areas in Rajputana and Kerala had higher levels of rackrenting and, subsequently, more peasant unrest than that of their princely counterparts.

Similarly, princes also maintained control over their judicial systems, and courts were not utilized as a means by which to expropriate peasants as was the case in the British provinces.

Aside from the lack of the same mechanisms which created caste conflict in the provinces, several princely state rulers were noteworthy for taking a stand against caste discrimination, especially the rulers of Baroda, Kolhapur, Mysore, and Travancore. The Gaekwar (ruler) of Baroda, for instance, personally paid for the education of the Dalit reformer B.R. Ambedkar. The ruler of Kolhapur responded to low caste agitation by stipulating that fifty percent of all government jobs were to be reserved for non-Brahmin castes (Mendelsohn and

Vicziany 1998: 129). In Mysore, all non-Brahmins were designated as backward, and were also entitled to ameliorative government policies (Bayly 1999: 279). In

Travancore, maharaja Shri Chithira Thirunal Balarama Varma took the most liberal measure of all by allowing untouchables to enter Hindu temples. In the subsequent empirical chapters I highlight other various examples of protectionist policies toward low castes initiated by princely rulers.

Princes all over India were similarly protectionist toward adivasi groups.

In Rajputana, both the Bhils and Minas were incorporated into the structure of the princely government.33 These tribes were also charged with ceremonially

33 The believed that the area had been controlled by adivasis, Minas, in the twelfth century.

95 placing the rajatilaka on the brow of the newly crowned king. In Jaipur, the

Minas were the guardians of the royal treasury (Ramusack 2004: 210). In

Travancore and Cochin, tribes were given ownership of their land, government subsidies to improve it, and special policies that limited the imposition of the outsiders who were a major problem for tribes throughout British India

(Mohanty 2006: 178). In Jammu and Kashmir many members of the Bakkarwal tribe were employed as tax collectors (Zaildaars) and became an important part of the government (Tandor 2005: 24). In 1942, most of the rulers of the Eastern

Feudatory States approved a draft policy (although it was not implemented) declaring that tribal groups ought to be the first claimants to forest lands and should also have the right to be governed by independent village panchayats

(Edmunds and Wollenberg 2003: 184). In short, princes displayed much more tolerance for tribal groups and adivasis fared better under their rule than that of

British administrators in the provinces. For Verrier Elwin, noted anthropologist of adivasis, it was “most refreshing to go to Bastar from the reform-stricken and barren districts of the ” (Guha 1996: 2379).

Therefore, the rulers of the princely states ultimately advanced a set of ethnic policies which benefited coreligionists, discriminated against non- coreligionists, and protected low castes and tribals. These territories, unlike

British India, were dominated by religion, but princely rulers also enacted policies that maintained the position of low castes and adivasis. The ethnic policies of British India and princely India were therefore mirror images of one another. The fault lines of ethnic conflict in the former revolved around castes

96 and tribes; in the latter, around religion. In the next section, I explain why these particular axes of conflict have persisted in the post-independence state.

III. Post-Colonial India: The Persistence of the Old Regime

The salience of ethnic identities and patterns of ethnic violence which were initially constructed during colonialism did not simply dissipate in the post- independence era. In this section I lay out a number of reasons why the post- colonial period has experienced key continuities with the past. There are two components to this explanation. First, it is necessary to delineate the specific legacies of colonial rule which have persisted into the modern period. On this point, I argue that the post-colonial Indian government has failed (or refused) to implement effective reforms of the past, especially with respect to policies governing ethnic minorities. The second component of explaining post-colonial continuities is addressing why patterns of conflict have been actively reproduced.

On this point, I argue that the salience of particular ethnic categories which were constructed during the colonial period continued to form the foundations of post- independence politics, and that ethnic conflict itself can further harden ethnic ties. Finally, I posit that these ethnic identities were embedded in a variety of local institutions and power structures.

What I present in the ensuing two sections is meant to be a broad, national-level story about how the legacies of the past have continued to influence post-colonial India, as well as a delineation of the mechanisms by which ethnic conflicts have been actively reproduced. However, in Chapters 3-5 I analyze these same two issues at the district, state, and regional levels. Using in-

97 depth qualitative analysis, I address the legacies of the past and the active reproduction of ethnic conflicts in each pair of case studies.

3.2 The Failure of Reform

At the dawn of Indian independence in 1947 one of the fundamental problems facing the new republic was the integration of hundreds of princely states (Furber 1951, Menon 1956, Wood 1984). Most of these states were incorporated into the union through what was called a 'merger agreement', which allowed princes to retain their wealth and titles in return for giving up jurisdiction over their territories. Once these princely states were integrated into the republic they adopted the same democratic institutional framework as that of the provinces. The decades after independence have seen state boundaries change numerous times: in 1956, according to linguistic lines, and most recently with the creation of several new states (Jharkhand, Uttarkhand, etc.) in 2000.

Yet, ethnic identification still plays an important role in post-independence India, primarily because an overarching 'Indian' national identity has not firmly developed. For example, a famed quote from a Member of Parliament (MP) in

Orissa highlights the hold local ethnic identities still have on the contemporary population:

“My first ambition is the glory of Mother India. I know in my heart of hearts that I am an Indian first and an Indian last. But when you say you are a Bihari, I say I am an Oriya. When you say you are a Bengali, I say I am an Oriya. Otherwise, I am an Indian.” (Tambiah 1990: 742)

For various other reasons upon which I elaborate below, the colonial past is still singularly important for understanding post-independence ethnic politics.

98 First, in the broadest sense, India did not experience what Barrington

Moore Jr. terms a 'revolutionary break from the past' (1966); in fact, there was a large sense of continuity between the colonial and post-colonial periods

(Chatterjee 1986, Brass 1994). For instance, the primary organization spearheading the independence movement – the Indian National Congress (INC)

– morphed from a broad-based political movement into a broad-based political party, one which still dominates the politics of the country. Most of the machinery of colonial rule also survived relatively unscathed into the post- independence setting: the bureaucracy, police, and courts all exist in more or less the same format.

Second, attempts at reform in post-colonial India have met with very limited success. I addressed this briefly during the Introduction, but here I offer several specific examples. First, consider policies on landholding. On the surface, the post-independence period has featured an impressive array of reforms aimed at British landholding policies. However, as Barbara and Thomas Metcalf note:

“...the extent of the transformation can easily be exaggerated. Despite protests from India’s socialists, the landlords were guaranteed compensation for all property taken. Further, under the constitution land reform as a subject was allocated to the states, not the centre, with the result that the well-to-do peasant castes who dominated the state Congress parties saw to it that ceilings were set high enough so that they would not be adversely affected. In addition, the abolition legislation was itself riddled with loopholes. By adroit measures, such as division of their estates among family members before ceiling legislation came into effect, or placing land in groves or under personal cultivation, which exempted it from confiscation, many landlords were able to preserve substantial properties, and with them a footing in the new political order. Landless peasants gained little from zamindari abolition, while the cultivating tenants, to gain full title to their land, had to make payments to the government over a series of years.” (2001: 244)

Furthermore, the INC even prevented real reforms from occurring in many

99 different parts of India. In fact, in the words of Kapil Kumar, “...[the landlords'] relationship with the INC was aimed at 'disciplining' the peasantry and crushing any challenge to their own authority” (1988: xi). In Chapter 5 I extensively document this fact – I show how the post-independence Indian government fought with great force to prevent land reforms in Bastar and Telangana in eastern India.34 Along these lines, consider also the way the post-independence government has approached colonial forest policies, the main driver of tribal conflict. The post-British governments of India have not significantly changed earlier policies – the state still controls forest lands, still reserves large tracts of lands, especially for development projects, and still has done little to prevent adivasi land displacement (Guha 1983, Haeuber 1993).

Policies toward religious minority groups have also largely remained the same from the colonial period. The Muslims of India, the primary religious minority in the country, still remain severely behind Hindus in their socioeconomic development.35 By 1980, for example, Muslims constituted only roughly 3% of the Indian Administrative Services, one of the central governing institutions in the country. The Indian Social Institute's report on the position of

Muslims has noted:

“…discrimination, social stagnation and educational marginalisation have cumulatively resulted in the growing economic backwardness of the Muslims in large parts of the country. A noteworthy feature is that post-partition, the share of Muslims in government services and other sectors has dropped drastically.” (2006: 8)

A related point here is that – and this may seem intuitive, but is

34 See also Gough 1974: 1395. 35 The best treatment of this is the Sachar Committee Report of 2006.

100 nevertheless under-addressed – political reform is quite difficult to achieve. It is certainly not easy for any government to enact changes to long-held policies, especially a country as populous and diverse as India. Even well-intentioned reforms tend to be limited in their scope due to inefficiency, excessive bureaucratization, or a lack of adequate funding.

The larger point is that the reform of the conditions of India's ethnic minority groups has not been successful. Therefore, the position of these groups in different parts of the country exist largely as they did during the colonial period. And key distinctions between the condition of ethnic groups in British and princely India persist. The position of low castes and tribals is still poorer in former British areas; the position of Muslims is still poorer in former princely states – the power distribution between groups has not been altered. This point is something which I am careful to showcase in the ensuing chapters: I provide evidence that the socioeconomic position of minority ethnic groups in Rajasthan,

Kerala, Chhattisgarh, and Andhra Pradesh has not fundamentally changed from the colonial era. The divide between the provinces and princely states still resonates in many ways, as “particularistic values and segmented political thinking have continued to affect political behavior in a number of, perhaps all regions of India” (Price 1993: 495).

3.3 Reproducing Ethnic Conflicts

The failures of the Indian government to effectively reform its past have enabled the legacies of colonial rule to persist into the modern period. However, there are a variety of factors which have also actively reproduced patterns of

101 conflict in the contemporary state. In this section I delineate several of these factors. In sum, I argue that once patterns of violence were established in the colonial period, master narratives of conflict were embedded in specific regions and ultimately led to commonsensical notions among local population about how ethnic conflict should be organized. Cities like Jaipur developed a reputation as being 'communally sensitive'; regions like Malabar, on the other hand, became well-known as 'caste-ridden'. People in these regions to this day understand the kinds of conflict which are legitimate and those which are not. These patterns of ethnic conflict are self-reinforcing; once the Hindu-Muslim axis of conflict comes to dominate a region, for example, violence itself hardens Hindu-Muslim ethnic identification. Finally, I argue that violence patterns were also reproduced via local institutions and power structures.

Ethnic identities constructed and institutionalized during the colonial period became embedded in local contexts – in the provinces, caste became the primary mode of ethnic classification, whereas in the princely states religion formed the bedrock of society. The colonial period also therefore created histories of violence that eventually came to define specific regions. So, for example, the long legacy of communal violence in Hyderabad has led to the city becoming known throughout India as a 'communally-sensitive' place. Since the moment communal riots became a feature of the city's politics in the 1920s, Hyderabad has become a major site of periodic Hindu-Muslim violence. Despite all of the changes to the city – economic or demographic changes, for example – it is still known throughout India as one of the most riot-prone cities in the country.

102 Members of ethnic groups in conflict are well aware of these histories of conflict, conveyed either through written or oral traditions. As Roger Petersen succinctly notes, “Violence becomes memory” (2012: 17)36. As I will detail in my three empirical chapters, respondents in given areas often make reference to previous historical events or episodes in explaining modern conflicts.

These patterns of violence then created commonsensical notions of how conflict is organized. Certain kinds of conflict are deemed to be legitimate, whereas other kinds are not. As Laitin demonstrates about Yorubaland, the way individuals are conditioned to think about ethnicity and ethnic relations matters greatly in the production of violence outcomes. The situation is similar in India. A common refrain from respondents in Ajmer, a former British province, was: “But everybody knows that Hindus and Muslims do not fight here”.37 Recall that I have defined ethnicity as a useful heuristic for reducing the uncertainty of the social world. Once ethnic categories were developed during the colonial period, and it became commonplace to order societal relations in one way or another, these rules of thumb were still useful in the post-colonial context.

Along these lines, several scholars have noted that conflict itself hardens ethnic identities (Van Evera 2001, Kaufmann 1996, Bates et al. 1998, Byman

2000). As Chaim Kaufmann writes: “...in ethnic wars both hypernationalist mobilization rhetoric and real atrocities harden ethnic identities to the point that cross-ethnic political appeals are unlikely to be made and even less likely to be

36 This book is currently in press; the page number refers to the existing manuscript. Petersen's work will be Chapter 12 of the book. 37 This was specifically a comment made to me by a top Ajmeri police official during fieldwork.

103 heard” (1996: 137).38 As Ornit Shani has noted in her work on Hindu-Muslim riots in Gujarat:

“An act of large-scale violence that is widely labelled or perceived to be taking place along a Hindu–Muslim divide clearly defines the boundaries between two such groups. Violence that seemed to be communal is often revealed to be, on closer examination, a manifestation of disparate social conflicts and the settling of various scores between social groups, and between them and the state. Yet the general representation of violence as communal, in official and popular discourses, simplified this complexity and hardened its communal characterisation.” (2007: 158)

Violence therefore serves the purpose of rigidifying ethnic divides. Violence between ethnic groups can maintain the salience of particular ethnic categories; if a region is dominated by upper caste-low caste violence, for instance, then these ethnic affiliations themselves are hardened because of conflict.

Second, these patterns of ethnic conflict were reinforced by local institutions. Take, for example, party systems, which have been a focus of special emphasis by constructivist scholars of ethnicity as a key engine of identity change

(Chandra 2004, Posner 2005). However, there is an underlying problem of causal directionality with the constructivist argument. Party systems did not appear out of thin air; these systems in India are probably based on ethnicity and not the other way around. For example, as I show in Chapter 4, the reason why caste matters in Malabar is not because of the Communist Party; rather, the

Communist Party matters in Malabar because of caste.39

This argument about parties is also important because of the way parties

38 Kaufmann's work, like most of these studies, focuses on large-scale violence such as wars or separatist movements. I argue here, however, that the same logic applies to low-level conflict, a question which Kaufmann does not analyze. 39 Wilkinson (2004) also starts with parties, ignoring that they may themselves have been formed by ethnic cleavage structures. See also his chapter in Chandra 2012.

104 in India operate. The dominant Congress Party, for example, has defined itself as a 'catchall' party – the party of all classes, religions, castes, etc. However, as

Pradeep Chhibber and John R. Petrocik note:

“Our image of the Congress is of a political party that is a coalition of State (and ultimately local) groups, which rests upon the divisions and conflicts of the State and community in question...To describe the Congress as a heterogeneous, national party of consensus, representing all groups, is formally accurate but substantively misleading. A more accurate description comes from Sisson, who finds that the national Congress is 'organized into mutually exclusive factional coalitions which tend to nucleate around a dominant leader or faction which has its own regional base of political support.'” (1989: 196-7)

Even during the early days of Congress dominance, state-level parties were built on localized ethnic cleavage structures. For example, when Hyderabad was integrated into the Indian Union in 1948 the Congress Party was, according to officials at the time, “...prepared to recognize the historical position of the

Muslim community by agreeing to special weightage for it.”40 In this way, local politics in Hyderabad became dominated by religious identification, and this has persisted into the present.41 Along this vein, it is important to note that princely rulers continued to play a large role in the critical early post-colonial period when they served as governors (Rajpramukhs) of their former states and still controlled the selection of civil servants. Many of them ran for elected office in the early post-colonial period, and they had an overwhelming success rate

(Richter 1971). Even today princely families exercise influence: Jitendra Singh, the current Minister of State for Home Affairs, and a former MP representing

Alwar (a princely state), is the grandson of the last maharaja of the kingdom. The

40 Collected from the APSA; the name of the article was missing, but it was published in The Hindu, December 6th 1947. 41 See Varshney 2002: 180-181.

105 creation of durable ethnic cleavage structures during the colonial period therefore affected the creation of parties more than parties themselves created ethnic cleavages. This is but one example of how local institutions continued to reinforce colonial-era ethnic divides, and in my qualitative chapters I discuss other examples.

This section has presented a broad, national-level story about why the legacies of the past still matter in post-independence India. Several colonial legacies were not reformed by the new Indian government, but ethnic classifications from the past were also actively reinforced via a variety of factors.

In Chapters 3-5 I offer more detailed analyses of both of these issues at the local level.

IV. Conclusion

This chapter presented an historical theory of colonialism and ethnic conflict in contemporary India. I began with a history of the British in India, arguing that the East India Company had initially intended to bring the entire country under their control. However, the Rebellion of 1857 locked in place a set of accidental boundaries between British India and princely India. Informed by a new wave of South Asian historical work, I argued that the native rulers of the princely states were not simply puppets of the British; rather, for the most part they exercised internal autonomy in their territories.

The divide between British and princely India was the single most important factor in producing distinct patterns of ethnic conflict. Starting with the argument that ethnicity is a heuristic for reducing the uncertainty of the

106 social world, and connotes a way of thinking about social relations, I argued that

British administrators and native kings had distinct views on how society should be ordered ethnically. In the provinces, British administrators placed a new focus on caste and tribal identities, and pursued policies which benefited high castes and discriminated against low caste and tribals. This created castes and tribes as the main ethnic identities around which violence would revolve. In the princely states, however, native rulers emphasized religious identities, and pursued policies which benefited coreligionists and discriminated against non- coreligionists. This then created religion as the main axis of ethnic violence.

Furthermore, the British enforced protectionist policies for religious minorities

(Muslims), whereas princely rulers enforced protectionist policies for low castes and tribals. These disparate sets of ethnic policies led to more caste violence in the provinces, but more religious conflict in the princely states.

Finally, in the third section I addressed why ethnic identities initially constructed during colonial rule continued to remain salient in the post- independence period, and why patterns of conflict between ethnic groups also persisted. I differentiated between legacies of the past and factors which actively reproduced ethnic conflicts. First, I argued the post-independence government in

India either failed to reform, or did not attempt to reform, the socioeconomic conditions of minority ethnic groups. Despite a number of reforms geared toward

SCSTs, or Muslims, the position of these groups has not improved considerably in the decades after 1947. Second, patterns of conflict were actively reproduced by actors because colonialism embedded master narratives of conflict in specific

107 regions, thereby creating commonsensical notions of how conflict is to be legitimately organized. Violence itself can also harden identities; Hindu-Muslim riots, for instance, only further reinforce religious divides. Finally, I showed that local institutions in post-colonial India also reinforced patterns of ethnic conflict; using the example of party systems, I showed the regional parties in India were often based on ethnic divides and not the other way around. Therefore, the colonial past still continues to play a major role in producing patterns of ethnic violence in contemporary India.

In the next chapter, I test this historical theory of colonialism and ethnic conflict using a quantitative study of 589 Indian districts. Utilizing a variety of data sources and a number of different regression models, I provide strong empirical evidence for my theory. I show that British rule is correlated with contemporary caste and tribal conflict, but negatively correlated with contemporary religious conflict, while controlling for a number of alternative explanations. I also test some plausible mechanisms of conflict, all of which shows that bifurcated colonial rule is a key factor in producing variations in contemporary ethnic violence.

108 Chapter 2 The Colonial Roots of Ethnic Violence: Quantitative Analysis

“...all we want is our hills.” – Orissa adivasi, to anthropologist Verrier Elwin1

“[The British government in India] has two wives, Hindu and Moslem, and the Moslem wife was the Government's favorite.” – Sir Bampfylde Fuller, Lieutenant Governor of Bengal2

The previous chapter laid out an historical theory of ethnic conflict in contemporary India. I hypothesized that bifurcated colonial rule created distinctive patterns of ethnic violence. While three-fifths of India was under the direct control of British administrators, the remainder of the country consisted of princely states governed by native kings who exercised large amounts of internal autonomy. British and princely rulers had disparate views on ethnicity, and therefore constructed ethnic policies in vastly different ways. Caste formed the foundations of British India, and colonial rulers enforced favorable policies for high castes, discriminatory policies for low castes and tribals, and protectionist policies for religious minorities. The princely states experienced the opposite: princely rulers favored coreligionists, discriminated against non-coreligionists, and protected low castes and tribals. I argued that the long-term effects of these policies produced different fault lines of ethnic conflict, and due to the failure of post-independence reforms and various other factors these patterns have persisted into the modern period.

1 Guha 1996: 2379, date of the quote unknown. 2 Sunderland 1924: 270; the date of the quote is unknown, but he served from 1902-1905.

109 In this chapter I test this historical argument through the use of statistical analysis. I present the results of a large-N quantitative study of 589 modern

Indian districts which investigates the effects of bifurcated colonial rule on contemporary patterns of ethnic violence.3 Using a variety of data and a number of different regression models, I show that British rule is positively correlated with contemporary caste and tribal violence but negatively correlated with contemporary religious violence, even when controlling for a number of alternative explanations. In short, this study provides strong empirical evidence for the theory advanced in Chapter 1: British rule generated higher levels of caste and tribal conflict, but princely rule generated higher levels of religious conflict.

This chapter is organized into four sections. In section I describe the statistical models which I employ in the analysis, as well as how I code the key independent and dependent variables. In section II I then examine the effect of

British rule on contemporary caste and tribal conflict. I also lay out several alternative explanations which were analyzed, as well as a test of plausible mechanisms that may be linked to conflict. Finally, in section III I analyze the effect of British rule on religious conflict, controlling again for various alternative explanations.

3 Much of this data collection and analysis was done in collaboration with my dissertation advisor, Emmanuel Teitelbaum. Part of this work will be the subject of a future co-authored publication on British colonialism and Naxalism.

110 I. The Regression Models

I utilize multivariate regression analysis in order to investigate the effects of bifurcated colonial rule on ethnic conflict in modern India. The dependent variables of interest in the main series of regression analyses (Tables 2.1 and 2.3) were count (or event) variables – i.e. the number of attacks, deaths, or injuries.

Therefore, I selected a negative binomial regression model.4 There are two main reasons for using this particular model over other alternatives. First, the distribution of an events model is not linear, so traditional linear models are inappropriate. Second, the dependent variables in the analysis are highly overdispersed – for example, some districts in India have hundreds of deaths from caste or religious violence whereas others have none at all.5 Violence outcomes across a country in general tend to be very volatile.

In order to conduct a cursory test of some of the mechanisms that may be linked to caste and tribal violence outcomes, I also utilize an Ordinary Least

Squares (OLS) regression model (Table 2.2). I examine the effect of British rule on certain factors that may lead to conflict, specifically the level of trust between castes, and land ownership. However, as I will detail later, I am only able to investigate mechanisms for caste and tribal conflict outcomes because there is a paucity of comparable data pertaining to the Muslims of India.

4 See Cameron and Trivedi 1998 and Hilbe 2007 for more on negative binomial regression analysis. This model has been used very often in analyses of conflict; see for example Urdal 2008, Kalyvas 2008, Bohlken and Sergenti 2010. However, other work on violence uses logistic regression to account for conflict 'onset'; see for example Cederman et al. 2010. 5 Many scholars use the Poisson model for counts, but the negative binomial model is more appropriate here because it includes a dispersion parameter. See Wilkinson 2004: 42 for a discussion of the advantages of this model over others.

111 1.2 The Independent Variable

The key independent variable of interest is colonial rule – specifically the type of colonial rule. I depart here from most existing studies of Indian politics in choosing to examine this question at the district level of analysis. I did this for a number of reasons. First, a state-level analysis would be problematic because present-day state boundaries have changed, often considerably, from the colonial period. While some contemporary and erstwhile state boundaries match up almost perfectly, others do not. The colonial territory of Rajputana, for example, is roughly the same as the present-day state of Rajasthan, but the princely state of

Hyderabad on the other hand is now strewn across the modern states of Andhra

Pradesh, , and . Therefore, using states as the level of analysis would not accurately capture the differences between British and princely India.

British rulers, however, placed an importance on district level administration, and this has persisted into modern times. As Paul Brass has noted:

“…the entire system of British district administration, which...concentrated authority at the district level…was retained virtually intact, particularly the central importance of the District Magistrate and the district courts and the police.” (1994: 138)

Princely rulers also often divided their states into districts, adopting the system from British India. Therefore, using districts as the level of analysis for the quantitative study is more appropriate for considering the impact of the colonial period. Finally, in terms of coding it was also relatively easy to match colonial and

112 post-colonial districts – even many of the names have not changed over time.

The district strategy, however, was not without its own complications. The number of districts in India has expanded considerably since independence. By

1971 there were 356, whereas as of this writing there are 640. This creates difficulties in adding control variables because data for modern districts have to be incorporated into older districts. Because many of the controls are taken from the 2001 census, I chose to use the list of 2001 districts. A total of 589 districts were ultimately included in the analysis, but several small districts were excluded.6 This strategy therefore includes almost every area of the country instead of simply selecting a number of states, or a random or representative subset of districts from each state. This analysis can ultimately provide a more complete picture of ethnic violence throughout India.

I coded all of the 2001 Indian districts as having belonged to either a

British province or a princely state. The historical data that informed this coding process came from a variety of sources: The Imperial Gazetteer of India, various

District Gazetteers, and district websites maintained by the Indian government.7

In her study of bifurcated colonialism and economic development in India,

Lakshmi Iyer (2010) also utilizes a district-level coding scheme. However, my coding process improves upon her study in two ways. First, I use districts from

2001, whereas she utilized data from 1991; therefore, the results of my study are

6 For example, Andaman and Nicobar Islands, Chandigarh, Dadra and Haveli, etc. The Delhi districts however, which are substantial in terms of population, are included in the analysis. The amount of observations in the various regression analyses varies due to missing data. However, the NB regressions never drop below 450 observations, and the OLS regressions do not drop below 300. 7 For precise coding procedures, see the Data Appendix.

113 more recent. Second, Iyer compares colonial and post-colonial maps8 to code districts, whereas I relied on historical data pertaining to each individual district in the dataset in order to ensure the highest level of accuracy.

I created two measures of the independent variable. I first constructed a dummy variable (province) for each district (coded as 1 if it was a British territory and 0 if it was not). This is the primary measure of bifurcated colonial rule.9 Then

I constructed a secondary measure of the variable, one which coded the number of years each territory was under the control of the British (all princely state values were zero). As I show later, the results of the analysis do not vary considerably with one variable or the other, lending credence to the argument that colonial rule does have an impact on contemporary ethnic violence outcomes.

1.3 The Dependent Variables

There are several dependent variables which are utilized in the analysis; they are broken down into two broad categories: those pertaining to caste and tribal conflict, and those pertaining to religious conflict (in the OLS regressions I use two different non-conflict dependent variables; they are discussed at the end of section II). The two conflict dependent variables come from two different sources: the former from the Worldwide Incidents Tracking System, and the latter from the Varshney-Wilkinson dataset on Hindu-Muslim riots.10

8 See pgs. 698 and 711. 9 I am mindful of the fact that this variable should ideally be continuous. Not all princely states were the same, and while the provinces were generally governed under similar frameworks, even they had important differences among them. However, due to the sheer number of different princely states and provinces, I simplified and coded a dichotomous variable. 10 More information on these datasets is also provided in the Data Appendix.

114 First, I use data on the Naxalite movement in eastern India as my main proxy for caste and tribal conflict. The Naxalites are Maoist rebels drawn mainly from the SCST population that are rebelling against the Indian state.11 This is a highly useful proxy for a few different reasons. First, the traditional measure of caste and tribal conflict comes from the Indian government – a metric called

'SCST Crimes'. These data record the number of atrocities committed against

SCSTs.12 However, there are serious problems with using these data. Both reporting bias (i.e. the failure of the police to investigate and register crimes) and enforcement bias (i.e. variation among states and their enforcement of SCST legal protections) seriously undermine the value of this metric. Ultimately, these data are an unreliable measure of contemporary caste and tribal conflict.

Due to these problems I utilized data on Naxalite violence instead – it is perhaps the best example of armed caste and tribal conflict in modern India. I investigate whether British rule correlates with casualties from the Naxalite conflict. These data cover incidents from 2005-2009. Using Naxal violence as the main proxy also has one upshot: because I use Hindu-Muslim riots as the main dependent variable in my analysis of religious conflict, I can therefore present regression results on the effects of colonialism on both 'top-down' and 'bottom- up' ethnic conflicts. As I noted in the Introduction, Hindu-Muslim riots are top- down incidents of violence whereas the Naxalite conflict is driven from the bottom-up, i.e. by SCSTs themselves.

11 See Chapter 5 for more on the Naxalite movement. 12 See the Ministry of Social Justice and Empowerment's definition of atrocities, available at: http://socialjustice.nic.in/faqs1.php#q3

115 The second series of dependent variables pertain to contemporary religious conflict. Data on Hindu-Muslim riots are taken from the Varshney-

Wilkinson dataset; they were collected for the time period 1950-1995 from newspaper reports. I constructed pooled district-level figures for the total number of deaths and injuries from riots. The dataset listed the number of Hindu

Muslim riots, deaths, injuries, etc. with corresponding information on year and location. I then compiled the number of total casualties from Hindi-Muslim riots per district, pooling the figures over the time period covered by the dataset. So, for example, for the district of Jaipur in Rajasthan I recorded 155 total casualties from 1950-1995.

Both of the conflict dependent variables in my analyses – for caste and tribal conflict, as well as religious conflict – record the total number of casualties, defined as deaths and injuries combined. I chose to look at this specific measure rather than the number of attacks, for example, for a few reasons. First, I found that for some districts a large number of incidents occur in which there is actually very little or no violence whatsoever. These incidents might officially be recorded as a riot or an attack, but are actually rather benign. Second, I am more interested in the total number of casualties from conflict, as the number of individuals who have been killed or hurt in episodes of ethnic violence more accurately reflects the human devastation which occurs when conflict breaks out.

116 II. British Rule and Contemporary Caste and Tribal Conflict

The first series of regression analyses pertained to contemporary caste and tribal conflict in India. There were several control variables included in the analysis that can also account for alternative explanations. Because the Naxalite conflict is technically an ethnic insurgency13, I incorporated insights from the well-known Fearon and Laitin study of ethnic insurgencies and civil wars (2003).

In their article, the authors examined the causes of civil wars from 1945-1999, and they posited a variety of factors that promote insurgency. There are, as I detail below, several variables I include in my analysis of the Naxalite conflict in order to account for their arguments.14

First, I included a log of the population of each district, compiled from the

Indian census. It is presumed that districts with larger populations will have higher levels of conflict, or that rebels will be more difficult to apprehend and stop in more populous districts. Similarly, I also include a variable that measures the percentage of each district's SCST population, also compiled from the census.

These are the two main groups who are engaged in Naxalite conflict, so districts with larger SCST populations may be more prone to conflict. Third, I included a log of the area of the district, compiled also from the census. Again, larger districts may be more prone to violence. Fourth, I included a variable that measured the percentage of the population which lives in designated rural areas, compiled from the census. Most caste and tribal conflict in contemporary India

13 Fearon and Laitin (2003: 75) define an insurgency as “...a technology of military conflict characterized by small, lightly armed bands practicing guerrilla warfare from rural base areas.” 14 All control variables are from 2001 otherwise unless noted.

117 occurs in the countryside. Fifth, I included a variable to account for geography, as rough terrain is often a factor presumed to aid rebels in fighting against the state.15 This variable measured the percentage of each district covered by forest, compiled from IndiaStat.16 This is a key variable for the Naxalite conflict as battles often occur in heavily-forested areas through eastern India. Sixth, I included a variable to account for the argument that the Naxalite movement is driven by the natural resources of a region (Miklian and Carney 2010). This variable codes whether or not a given district contains any kind of mining operations, compiled also from IndiaStat. Finally, I include two additional variables that are measures of socioeconomic development. The seventh variable is a log of the number of health centers per-capita for each district, compiled from IndiaStat. The eight and final variable is the literacy rate (percentage) per district, compiled from IndiaStat. Since many insurgencies are driven by poverty or a lack of socioeconomic development, these variables may be correlated with violence.

The results of the first regression analysis are displayed in Table 2.1 below.

I ran the model17 with the use of the provincial dummy as the main independent variable first, and then in the second model I swapped in the number of years of

British rule. The dependent variable in both models is Naxalite casualties – the number of individuals killed and injured in Naxal attacks from 2005-9. Also, all of the regression models utilized in this chapter were run using robust standard

15 On this, see also Scott 2009. 16 An Indian statistical firm; see http://www.indiastat.com 17 All of the regressions were run using Stata/SE 11.1

118 errors to account for potential heteroscedasticity due to large variations in district population.

Table 2.1: Colonialism and Caste/Tribal Conflict Regression Results ------(Model 1) (Model 2) ------

Province 1.189*** (0.36) Years British 0.014*** (0.00) Population (Log) 2.717*** 2.264*** (0.47) (0.50) Area (Log) -0.682 -0.136 (0.41) (0.45) % Rural 5.106** 5.422** (1.68) (1.87) % Forested 0.052*** 0.062*** (0.01) (0.02) % SCST 0.035 0.038* (0.02) (0.02) Mining 1.026* 0.863 (0.41) (0.45) Health Centers 0.473 0.181 (0.79) (0.82) Literacy Rate -0.060*** -0.052** (0.02) (0.02)

_cons -31.878** -34.296** (9.69) (10.74) ------lnalpha _cons 2.799*** 2.722*** (0.13) (0.14) ------

Observations 547 503 Log likelihood -573.275 -507.125 Wald Chi2 146.07 172.35

Negative Binomial Regressions Robust Standard Errors in Parentheses *** p<0.01, ** p<0.05, * p<0.1

119 The results indicate that British rule in both models correlates positively with the number of casualties from Naxalite attacks, and is significant at the 1% level. Determining the substantive effects of variables (i.e. interpreting the coefficient) is not as straightforward in a negative binomial model as it is in an

OLS model. I therefore additionally calculated the 'percent change coefficient' of all the independent variables.18 Put in substantive terms, the first model shows that a former province experiences roughly 228% more Naxalite casualties than a former princely state, holding all other variables constant.19 A number of these other variables also matter according to the analysis: the results show that forested districts which have larger, more rural populations and lower levels of social development experience higher levels of Naxalite casualties. This adds to the likelihood that the model is correct: most of the variables point in the predicted direction. The Naxalite movement in India is widely understood as a rural insurgency driven by socioeconomic factors, and the rebels are believed to be aided by thick forest cover.20 But colonial rule also plays an important role in driving the conflict, and the results of the second regression analysis highlight two specific mechanisms which are linked to violence.

I utilized an OLS regression model to test two plausible mechanisms of caste and tribal conflict which I posited in Chapter 1: I investigate the effect of colonial rule on caste antagonisms and land inequality, respectively. There are two dependent variables in this regression analysis: the first is a measure of trust

18 See Long and Freese 2006: 247-248. 19 Note: in order to calculate this figure a special Stata package (spost) must be installed. 20 See Chapter 5 for more on this.

120 between castes, taken from the Indian Human Development Survey.21 This is meant to test the argument that British rule led to long-term antagonism between disparate caste groups. As I noted in Chapter 1, through mechanisms such as the census, British administrators reified and hardened what were once fluid caste categories, and the result was wide-scale competition between castes for access to government patronage. The process of conducting the census was highly politicized, as many caste groups sought to avoid the 'Untouchable' label, and a variety of new, narrow caste associations were formed in order to achieve better positioning. The census was not used in the princely states in the same way; rather, princely rulers prioritized religious identification. Therefore, the level of trust between castes should be lower in former British provinces.

The second dependent variable in the OLS regression records the percentage of full-time workers engaged in agricultural labor. This is a proxy for landlessness, as these individuals work the land but do not own it; they essentially constitute landless laborers. This is meant to test the argument that

British rule, through the granting of proprietary rights to zamindars, increased landlessness among peasants. Prior to the institution of the Permanent

Settlement in Bengal, Indian peasants were generally hereditary tenants, and could not easily be displaced from the land. Once private property rights were granted, however, land dispossession among the peasantry increased. Therefore, the percentage of landless laborers should be higher in former provinces.

The results of the OLS regression analysis are displayed in Table 2.2 below.

21 For more on this dataset see the Data Appendix.

121

Table 2.2: Colonialism and Caste/Tribal Mechanisms Regression Results ------Jati Trust % Ag. Worker Jati Trust % Ag. Worker ------Province -0.109*** 4.619*** (0.03) (0.92) Years British -0.001*** 0.050*** (0.00) (0.01) Literacy Rate 0.007*** -0.037 0.006*** -0.010 (0.00) (0.04) (0.00) (0.05) % No Assets 0.284** 27.300*** 0.250* 27.411*** (0.12) (3.40) (0.12) (3.46) % SCST -0.000 -0.208*** -0.001 -0.148*** (0.00) (0.02) (0.00) (0.02) _cons 2.132*** 12.260*** 2.220*** 7.418* (0.12) (3.31) (0.13) (3.50) ------

Observations 345 534 317 491 R Squared .0973 0.2153 .0900 .2375 Root MSE .2803 10.41 .2814 10.291

Ordinary Least Squares Regressions Robust Standard Errors in Parentheses *** p<0.01, ** p<0.05, * p<0.1

122 These results indicate that British rule is negatively related to trust between castes, and positively related to higher levels of landlessness. Again, the two codings of the independent variable are significant at the 1% level across all models. The mechanisms that may be linked to conflict which I laid out previously therefore do seem to be important. British rule led to long-term caste antagonisms as well as increased landlessness, but princely rule did not lead to either of these outcomes. Both of these factors help explain higher levels of caste and tribal conflict in the former British states of contemporary India.

III. British Rule and Contemporary Religious Conflict

The second major component of the statistical analysis deals with the effects of British colonial rule on contemporary Hindu-Muslim riots. By including a study of riots, as I noted earlier, I am able to account for both bottom-up and now top-down ethnic conflict. Since this analysis deals with conflict outcomes, I return again to a negative binomial regression model.

The dependent variable of the model is riot casualties, taken from the

Varshney-Wilkinson dataset. The number of casualties per district were pooled over the 45-year period covered by the dataset. However, data on important control variables were not available at the district level from previous decades – for example, controls on poverty and infrastructure from the 1950s and 1960s are not readily accessible to researchers. Instead, I continue to use control variables only from the 2001 census. In some cases these controls may not have changed considerably – the Varshney-Wilkinson dataset begins in 1950, so the percentage of the population that is Muslim recorded in 2001 can still account for major

123 post-Partition population shifts, for example.

For the regressions on religious violence I use several controls, many of which were described previously. First, I included a log of the population of each district. Just like caste and tribal conflict, larger districts may also be prone to experiencing riots. The second variable is a log of the area of a district; again, larger districts may experience a higher number of riots. The third variable records the percentage of the population which is Muslim. There are a number of districts where the Muslim population is minuscule, and we would not expect riots to occur in these areas. All three of these control variables came from the

2001 census. The last two variables are proxies for socioeconomic development.

First, I include a variable on poverty: the percentage of households per district who claim no economic assets, compiled from IndiaStat. Examples of economic assets include televisions, radios, cars, and access to some form of banking.

Poverty may be a main driver of religious conflict; Anjali Thomas Bohlken and

Ernest Sergenti (2010), for example, find that a 1% increase in the growth rate of a state reduces the likelihood of a Hindu-Muslim riot by over 5%. Finally, the last variable is the literacy rate percentage per district, also compiled from IndiaStat.

Again, lower levels of socioeconomic development may breed religious conflict.

The results of this regression analysis is displayed in Table 2.3 below.

124 Table 2.3: Colonialism and Hindu-Muslim Conflict Regression Results ------(Model 1) (Model 2) ------

Province -0.737** (0.29) Years British -0.006*** (0.00) Population (Log) 2.816*** 2.817*** (0.24) (0.25) Area (Log) -1.142*** -1.175*** (0.21) (0.21) % Muslim 3.256*** 3.070*** (0.92) (0.88) % No Assets 1.087 1.277 (0.93) (0.95) Literacy Rate 0.055*** 0.053*** (0.01) (0.01)

_cons -31.686*** -31.296*** (3.09) (3.20) ------lnalpha _cons 2.339*** 2.293*** (0.09) (0.09) ------

Observations 526 481 Log Likelihood -1232.3427 -1202.1811 Wald Chi2 190.64 172.93

------

Negative Binomial Regressions Robust Standard Errors in Parentheses *** p<0.01, ** p<0.05, * p<0.1

The results indicate that British rule is now negatively correlated to violence outcomes, meaning that while colonialism led to more caste and tribal violence it simultaneously generated less Hindu-Muslim conflict. The variable coded for British rule is also significant at the 5% level in the first model and at the 1% level in the second. In substantive terms, the first model shows that a former princely state experiences roughly 52% more Hindi-Muslim riot

125 casualties than a former province, holding all other variables constant.

A number of other variables also matter in this analysis, although not always as predicted. The population of a district is important, as is the percentage of the population that is Muslim. Both of these results are unsurprising, and mirror the first regression analysis displayed in Table 2.1. However, there are some unusual outcomes of this analysis: the area of a district is negatively correlated to Hindu-Muslim riot casualties and the higher the literacy rate of the district the more likely violence is to occur. There are some plausible reasons for these outcomes. The first finding might suggest that very small, densely populated areas will be more likely to experience riots. The second finding might suggest that overall levels of social development do matter, but note that this variable does not account for specific levels of Muslim social development.

Measures of Muslim social development are difficult to obtain. Areas with high levels of social development might still experience more violence due to higher disparities in development between Hindus and Muslims. If these districts, for example, have Hindu populations with high levels of social development but

Muslim populations that significantly lag behind, violence between the two communities may be more likely.

Unlike the regressions pertaining to British rule and caste/tribal conflict mechanisms (Table 2.2), I am unable to investigate the mechanisms that may lead to religious conflict because finding additional control variables pertaining to India's Muslim population is difficult.22 For example, IndiaStat had several

22 Even the Indian Human Development Survey, for example, featured a survey question about intercaste trust, but surprisingly nothing on intercommunal trust.

126 datasets available on socioeconomic indicators for SCSTs, but nothing other than simple district level population figures for Muslims. There are two reasons why this is the case. First, the Indian government after Partition has been careful to de-legitimize claims based on religion, and has resisted data collection efforts on religious groups.23 Second, investigations into the position of Muslims in India have been resisted by Hindu nationalist organizations.24 In order to counteract this important shortcoming I investigate mechanisms that might produce Hindu-

Muslim violence using in-depth qualitative analysis conducted in my case studies, presented in Chapters 3-5.

Finally, it is important to note that several arguments from the existing political science literature on Hindu-Muslim riots which I detailed in the

Introduction also unfortunately could not be tested quantitatively due to data limitations. All three of the major arguments from existing work on religious violence are difficult to test statistically. First, Brass' argument about the importance of the BJP cannot easily be quantified.25 If, for example, there were data on the number of BJP or RSS shakhas (branches) at the district level, then it would be possible to test the strength of Hindu nationalism as a plausible independent variable which promotes riots. The same applies to Wilkinson's electoral incentives argument – data on election results for every district in India over time would need to be collected. Finally, Varshney's argument about the 23 A story from fieldwork will illustrate this. I had a conversation with a civil servant at the Rajasthan Department of Education about collecting data on minority groups. He was more than willing to give me information about SCSTs, but when I asked about Muslims he simply replied, “We 't bifurcate the data.” 24 The term often thrown about by the BJP, for example, is “Muslim appeasement”. 25 Brass himself (1997, 2003) at times voices a very critical appraisal of statistical methods and even causal analysis.

127 strength of intercommunal links was not tested quantitatively at the national level in his work, again due to a lack of reliable data. These three arguments (the limitations of which I already discussed in the Introduction) are therefore also addressed in Chapters 3-5 in the particular context of each of my case studies.

IV. Conclusion

This chapter has presented quantitative evidence that links the colonial period in India to contemporary ethnic violence outcomes. I tested the impact of

British rule on modern ethnic conflicts using a large-N quantitative study of 589

Indian districts, utilizing a variety of data sources and a number of different regression models. The results of this quantitative study provided strong empirical evidence for the primary argument of this dissertation, which was laid out in detail in Chapter 1.

First, I found that British colonialism promoted higher levels of contemporary caste and tribal conflict. Using the Naxalite insurgency as the main proxy for this kind of violence, I found that being a former British provincial district increased Naxalite casualties by 228%. These results were highly significant, and I controlled for a number of alternative explanations. Then, I found that British rule increased caste antagonisms and landlessness, two mechanisms which I laid out as possible drivers of conflict in Chapter 1.

Regression results showcased that British rule was negatively related to trust between castes, and positively related to the percentage of the work force who are landless laborers, and these results were also strongly significant.

Furthermore, as my theory predicated, another set of regression results

128 highlighted that princely rule promoted contemporary religious conflict. While controlling for a number of alternative explanations, I found that being a former princely state district increased Hindu-Muslim riot casualties by 52%. Therefore, while British rule led to higher levels of caste and tribal conflict, princely rule generated higher levels of religious violence.

This clear dichotomy in violence outcomes highlights that the legacies of colonial rule play a key and under-addressed role in explicating why ethnic conflicts in India revolve around one identity rather than another. Bifurcated colonial rule had an important long-term effect in conditioning and structuring violence outcomes. British policies which emphasized caste – benefiting high castes while discriminating against low castes and tribals – led ultimately to more insurgency in the countryside. However, British religious neutrality and protectionist policies for Muslims seem to have decreased communal conflict over the long term. On the other hand, princely policies which emphasized religion – benefiting coreligionists while discriminating against non- coreligionists – increased communalism in the native states. But the protectionist policies of princely rulers toward low castes and adivasis minimized the Naxalite movement in these territories.

In the ensuing three chapters I turn to qualitative evidence for my theory, and this forms the bulk of the empirical work in this dissertation. The regression results displayed here provide evidence of a broad and discernible national-level trend in ethnic conflict outcomes. In Chapters 3-5, however, I address the microfoundations of violence. I utilize historical and localized data on ethnic

129 violence outcomes in all of my case studies to examine whether the same trend showcased here is existent at the small-N level of analysis. I also more thoroughly investigate the mechanisms that produce these ethnic conflicts. Evidence from

Rajasthan, Kerala, Chhattisgarh, and Andhra Pradesh all provide further support for the quantitative results detailed in this chapter. The legacies of bifurcated colonial rule still play a major role in structuring patterns of ethnic conflict in the contemporary Indian state.

130 PART II: QUALITATIVE CASE STUDIES

131 Chapter 3 Ethnic Violence in Rajasthan: Jaipur and Ajmer

“Muslims have a good situation here. Go into their neighborhoods. They eat chicken biryani all the time.” – Police Inspector, Interview A in Ajmer, January 24th 2011

“When you head out that way, you'll see it's about caste. Conflict in Ajmer revolves around caste.” – Journalist, Interview in Jaipur, September 9th 2010

The previous chapter presented quantitative evidence from a study of 589 districts showing that former British provinces in India experience higher levels of caste and tribal violence today while former princely states experience more contemporary religious violence. This chapter examines ethnic violence outcomes in the first of three pairs of case studies: Jaipur and Ajmer districts in the north

Indian state of Rajasthan. In this chapter I seek to first determine whether the quantitative results from Chapter 2 are evident at the micro level; and second, to assess the causal mechanisms that produce variation in conflict outcomes.

In the first section I discuss the history of both Jaipur and Ajmer, and why they form a controlled historical comparison. Then, in the second section I present figures on post-colonial religious violence in both districts, highlighting that Jaipur experiences more Hindu-Muslim conflict. I also provide a qualitative account of the socioeconomic condition of Muslims in both districts, as well as information on intercommunal relations, both of which I collected from a number of interviews conducted in the two districts. In the second section I also consider some existing explanations for communal conflict, drawing on existing

132 work in the political science literature. In the third section I then examine contemporary caste and tribal conflict, and I find that in this regard Ajmer suffers more violence than Jaipur. In the fourth and fifth sections, I turn to historical analysis: I showcase that bifurcated colonial rule led to different kinds of ethnic policies in both districts. In Jaipur, the Muslim community was badly discriminated against, whereas low castes and tribals were protected. British administrators in Ajmer, however, protected the Muslim minority but instituted a host of policies that decimated the position of low castes and adivasis. I argue that these disparate sets of policies embedded different narratives of conflict in both districts, and post-colonial reform efforts have failed to change them. The origins of ethnic violence patterns in contemporary Rajasthan, therefore, lie in the colonial period.

I. The Case Studies

Jaipur and Ajmer are two neighboring districts located in the heart of the north Indian state of Rajasthan. During the colonial period this entire area was a conglomerate of princely states1 known as Rajputana (the land of the Rajputs).

Rajputana was a massive landlocked territory extending over 330,000 square kilometers. Both Ajmer and Jaipur were located along the eastern part of the state, slightly north of the Aravalli mountain range. Maps 2 and 3 below indicate the geographical positions of Jaipur and Ajmer in the colonial and contemporary period, and Table 3.1 shows basic comparative figures for both districts over time.

Though Jaipur has always been a more populous territory than Ajmer, both areas

1 Aside from Jaipur, others major princely states in the region were Udaipur, Jodhpur, Bikaner, and Jaisalmer.

133 are otherwise comparable: they share a similar geography, economy, culture, and language. But there was a key difference between them – during colonialism

Jaipur was a princely state whereas Ajmer was a British province.

Maps 2 and 3: Colonial Rajputana; Post-Colonial Rajasthan2

Table 3.1: Jaipur and Ajmer Comparison, 1901 and 20013 Jaipur Ajmer Population (1901) 2,658,666 476,912 Population (2001) 5,251,071 2,181,670 Percentage Hindu (1901/2001) 90% / 88% 80% / 86% Percentage Muslim (1901/2001) 7 / 10% 15 / 11% Percentage SCST (2001) 22.67% 20.11%

The princely states of Rajputana were, according to most historical sources, patrimonial and backward.4 British administrators often liked to compare them to the kingdoms of feudal Europe. They were considered 'antique states' (Ramusack 2004: 43) in which tradition and lineage played a paramount role. Most of the Rajput territories were initially formed by clans who used the language of Hindu mythology to trace and legitimize their kingdoms. One of

2 The map on the left is colonial, and the pink area denotes Ajmer-Merwara province. The map on the right is post-colonial. 3 The first set of population figures refers to Jaipur State and Ajmer-Merwara province; the latter figure refers to the present-day districts. 4 For a great overview on the Rajput states see Rudolph and Rudolph 1984.

134 these clans – the Kachhawas, who claimed to be descended from the god Ram – established a kingdom in Rajputana which was initially called Amber. In 1727 the maharaja of Amber Jai Singh II began to construct the city of Jaipur as the new capital of his expanded kingdom.

Amber, later renamed the Jaipur State by Jai Singh II, rose during the period of Mughal supremacy in India. It quickly became an imperial client state and was incorporated into the Mughal political system. One of Amber's earliest rulers, Bahar Mal, was given a mansab by the Mughals, and the house of

Amber intermarried with the Islamic rulers of India. A later king of Amber, Jai

Singh I, was recruited to serve the Mughals during their Deccan military campaigns in the late 17th century. Jai Singh II also served the Mughals and earned the Sawai5 from the Emperor Aurangzeb for his military valor.

Jaipur's relationship with the British began at the start of the 19th century, by which time Mughal power in India had declined sharply. By this period

Rajputana, trapped in the resultant power vacuum in north India, had become a war-torn region. New emergent powers such as the Marathas and Pindaris led vicious raids against the Rajput states which effectively decimated the area. In the midst of this mayhem in 1803 the British, under Lord Gerard Lake, signed a treaty with Jaipur's raja Jagat Singh. The Jaipur State had actually initiated contact with the British, seeking protection against raids. As was the case with most of these treaties, Jaipur agreed to pay an annual tribute and provide troops to the British in return for military protection. The relationship would, however,

5 Sawai meant 'one and a quarter' and was intended to mean that Jai Singh II was more capable than one mere man.

135 be short-lived. In 1806 the British abruptly terminated the agreement because officials in London were deeply worried about imperial overextension in India

(Stern 1988: 13). For the next decade Jaipur was left on its own and the raids resumed in force. Then, in another about-face, in 1818 the British, with recently renewed ambitions in India, demanded that Jaipur once again sign a treaty of subsidiary alliance. Jaipur eventually relented, and once again became a protected state.6 In 1820 the British for the first time then sent a Resident to the

Jaipur court. There were many succeeding political reforms that occurred in the state over the next several decades, but after the Rebellion of 1857 – as happened throughout India – the British began to take a hands-off approach to governance, and devolved more internal authority to Jaipur's rulers. By 1896 Jaipur's raja enjoyed a 17 gun salute, one of the highest that could be bestowed on a prince.

Robert W. Stern writes that the state eventually became, in relation to the British,

“a client tail wagging a patron dog” and that Sawai Madho Singh, the second-to- last ruler of Jaipur, had essentially become an “absolutist monarch” (1988: 18-19) unencumbered by British constraints.

Neighboring Ajmer was known as Ajmer-Merwara province during the colonial period. This territory was also once a princely state whose pre-colonial history was very similar to that of Jaipur. Ajmer also became an imperial client state once the Mughals invaded and conquered India. It was designated a subah or province of the empire and it was even briefly the headquarters of the Emperor

Aurangzeb. Like Jaipur, once the Mughal Empire began to decay regional

6 NAI, Jaipur Residency Records, Short History

136 warlords attempted to establish their own autonomous kingdoms in Ajmer. The territory came under domination in the mid-18th century and was later placed under the control of a Maratha regional governor.

By 1816 the British had determined to crush Maratha power once and for all in north India, and enlisted the support of Daulat Rao Sindhia, the maharaja of the nearby . After Maratha power had successfully been broken

Ajmer was ceded in 1818 to the British for over 500,000 rupees by Sindhia

(Baden-Powell 1892: 321). British administrators then set out over the next several decades negotiating and writing treaties with the surrounding princely states to pacify the borders of their new outpost.

Ajmer-Merwara would become the only British province in the whole of

Rajputana, and colonial administrators intended it to be “a model to the surrounding States”7. It was different from most other British provinces because it was not located close to any major centers of colonial power; in fact, it was 220 miles away from any British territory. Yet, despite its remote position it was not an inconsequential area. Ajmer became a British province before many other well-known cities – before Delhi, Agra, Lahore, Lucknow, Allahabad, Karachi, and Nagpur.8 Also, because of its isolated position it was governed through a

Chief Commissioner who was responsible directly to the Foreign Department of the Government of India.9 Robert Holland, Chief Commissioner of Ajmer-

Merwara, noted:

7 NAI, Rajputana Agency, Ajmer-Merwara Files, 1924, #31 8 NAI, Foreign & Political Department, Internal Branch, 1925, #100-I of 1924 9 Imperial Gazetteer of India (Vol. 5): 413

137 “The province resembled in some way a spoilt child because it is equipped with loving grand parents namely the Government of India, devoted and fussy parents, in the Chief Commissioner and the Commissioner, and a host of other fond relations in the shape of officials who would not be found in an area corresponding in size in British India.”10

After Indian independence in 1947 Rajputana was designated a 'Part B' state – i.e. a former princely state or conglomerate of princely states. In the next year it was incorporated into the Indian Union. The maharaja of Jaipur was named the of the area, and Jaipur city became the capital of the new state. Finally, in 1956 under the States Reorganisation Act, Ajmer-Merwara province (considered a 'Part C' state as a former Chief Commissioner's Province) was integrated into the rest of Rajasthan. The boundaries of present-day

Rajasthan have remained largely the same as those of the colonial period.

1.2 The Controlled Historical Comparison

Why was it that Ajmer came under the control of the British but Jaipur remained independent? This is the central concern underlying the choice of these districts as a controlled historical comparison. As I noted in the Introduction, endogeneity questions could potentially undermine the selection of cases. I argue here that the selection of Ajmer instead of Jaipur as a British province was a quasi-random historical outcome. As I also noted in the Introduction, if the

British were selecting parts of India to control along any basis it was fertile land that most interested them. In this sense, Rajputana as a whole was not an area of colonial interest because the massive Thar Desert stretched across most of the region. In fact, records from the time noted that “agricultural conditions in

10 BL, IOR/V/27/249/1, 1925, pp.6

138 Ajmer-Merwara are precarious...[t]he soil is generally shallow, and the rocky strata are near the surface...rainfall...is uncertain, and its frequent failure makes the Province peculiarly liable to scarcity and famine”.11 The British also wrote that adjoining Merwara was “ a wild and hilly tract, giving little opportunity for productive agriculture”12. Ajmer-Merwara, however, did have one specific benefit for the British: it was well-placed in the center of Rajputana. Chief Commissioner

Holland noted that “[Emperor] Akbar realised the great importance of Ajmer as a

'point d'appui' in the midst of Rajputana, commanding the route to Gujarat on the one side and to Malwa on the other.”13 He added: “...every power who became paramount in Upper India has striven to possess Ajmer as one of the symbols of sovereignty”.14 In 1818 the British took control of Ajmer for precisely this reason, using it as their lone station in Rajputana.

In the same year the British also concluded a treaty of 'subordinate cooperation' with the ruler of Jaipur. Why didn't they take control of this state as well? Stern notes:

“The British knew what they wanted in Jaipur, but not yet how exactly to go about getting it. They wanted, of Rajputana in general, a pacified border area bordering on their empire proper; and from Jaipur, in particular, they wanted money." (1988: 66)

Jaipur was a considerably more prosperous princely state than Ajmer, and the

British ultimately needed to pay for the wars in which they engaged to pacify

Rajputana. In 1818 they therefore set the revenue tribute for Jaipur at a

11 Imperial Gazetteer of India (Vol. 5): 149 12 NAI, Rajputana Agency, Ajmer-Merwara Files, 1924, #31 13 Ibid. 14 BL, IOR/V/27/249/1, 1925, pp. 5-6

139 staggering 8 lakhs15. Even colonial officials residing in Jaipur at the time considered it excessive. In 1842 British authorities finally relented and reduced the size of the tribute to only 4 lakhs annually, and they also remitted 46 lakhs to the Jaipur government.16 Stern notes that the British, instead of conquering

Jaipur, intended to “maintain, more-or-less, Jaipur's own political system and try to establish over it controlling British influence” (1988: 1-2). I argue in this chapter that they were not ultimately successful in this plan, but they preferred this option over fighting yet another costly war in the region.

The important point here is that the British were not selecting Ajmer based on the dependent variable of interest – ethnic violence. There is no evidence that this factor played any role in the process. Furthermore, the real selection criterion – geographical location – is not related to nor does it seem to influence the dependent variable. Geographical location would have no effect on ethnic violence: Ajmer being placed in the center of Rajputana would not have made it more or less likely to experience any particular kind of ethnic conflict. Plus, its geographical attributes were incredibly similar to that of Jaipur. Therefore the paired selection of Jaipur and Ajmer allows me to effectively overcome the problem of endogeneity.

15 A lakh is equivalent to 100,000 rupees. 16 NAI, Jaipur Residency, Jaipur Agency, 1947, #6

140 II. Religious Violence in Contemporary Rajasthan

How is ethnic conflict organized in contemporary Rajasthan? Are the patterns of violence established in Chapter 2 evident when examining this particular pair of cases? I begin this district-level analysis by examining communal conflict in both Jaipur and Ajmer, and in section III I turn to addressing caste and tribal conflict. For data on modern religious violence I utilize three main sources. First, I use the Worldwide Incidents Tracking System database (2005-2009), and the Varshney-Wilkinson dataset (1950-1995) on

Hindu-Muslim riots. More information about both of these datasets is included in the Appendix. Finally, I also include additional riot data from the Ministry of

Home Affairs which were published by Gopal Krishna in 1985. These latter data are unique because Home Affairs data are rarely disaggregated to the district and town level (one of the reasons why Varshney and Wilkinson used newspaper reports), but Krishna had special access to this information. The main drawback is that the data only cover a truncated time period (1961-70). Nonetheless, they offer a rare highly detailed glimpse into riot metrics for the entire country.

Finally, considering three unique sets of data increases the validity of the comparison. The results are laid out in Table 3.2 below.17

17 These are calculated per-capita; the denominator was the Hindu and Muslim population of each district. Casualties means deaths and injuries.

141 Table 3.2: Contemporary Religious Violence in Rajasthan (per 100,000)18 Hindu-Muslim WITS Casualties Riot Casualties Jaipur 5.13 3.01 Ajmer 1.04 0

Since independence Jaipur has experienced a much more violent communal history than neighboring Ajmer, according to the first two primary sources of data. The WITS data record one major instance of violence in Jaipur: the June 2008 bombings that killed 64 people and injured 200. The Indian

Mujahideen took responsibility for the attack. Ajmer, on the other hand, had only

3 deaths and 19 injuries in a 2007 bomb attack by Muslim assailants from the group Harakat ul-Jihad-I-Islami.

The Varshney-Wilkinson dataset uses Times of India newspaper reports to record Hindu-Muslim riots over a 45-year period. According to the dataset,

Jaipur is listed as one of India's 'riot-prone' cities.19 Jaipur experienced 32 deaths and 123 injuries, although because the data are recorded from one source it is likely that the overall toll is much higher. Ajmer, on the other hand, experienced no casualties. The number of events is also noteworthy – Jaipur experienced riots in 1952, 1969, 1989, 1992, 1993 and 1995.20 In contrast, Ajmer did not have a single recorded riot from 1950-1995.

18 Copland (2005) has compiled a list of communal incidents in many different parts of North India, including Ajmer and Jaipur. It is not completely clear from what source Copland has constructed his list. He cites, for example, only three riots in Jaipur: 1926, 1932, and 1940. However, the number of incidents I found mentioned in the National Archives of India and the Fortnightly Reports in the British Library were much higher and I have included as much documentation as possible to this effect. 19 See Varshney 2001: 372. 20 This database probably excludes one other riot in Jaipur in October of 1990. I found information on this episode from a 2007 report by the “Working Group of National Integration Council to Study Reports of the Commissions of Inquiry on Communal Riots”, which was issued by the Ministry of Home Affairs.

142 Finally, Krishna's dataset also details Hindu-Muslim riots but unfortunately does not present figures on the total number of casualties.

However, it does record that Jaipur was the most violent city in Rajasthan during the 1960s (1985: 69). It experienced riots during four of the ten years of the decade, enough to be labeled an “urban centre...prone to periodic rioting”. Ajmer, however, did not appear on Krishna's list.

When all three sources are taken together, the pattern is clear: Jaipur is a city of communal strife and Ajmer is a city of communal peace. This provides micro-level confirmation of statistical results highlighted in Chapter 2: the former princely territory of Jaipur suffers higher levels of communal conflict than the former British province of Ajmer. At first glance, however, this result is not completely intuitive. Both Jaipur and Ajmer have a similar Hindu-Muslim divide

(in terms of percentage of the population), and both regions therefore seem equally susceptible to communal conflict. What, then, accounts for the divergence in violence outcomes? Why has Hindu-Muslim violence constituted such a major problem in contemporary Jaipur but not Ajmer? In the next section I detail part of my argument: that the socioeconomic position of Muslims in Jaipur is severely stunted, and that relations between Hindus and Muslims in the district are extremely poor. To be clear, the socioeconomic state of Muslims matters more for bottom-up violence; where Muslims are poorer they will be more prone to initiating violence against dominant religious groups. Relations between Hindus and Muslims relates more to top-down violence; where there is a high level of antagonism between these groups, riots are more likely to occur. In Ajmer,

143 however, the situation is completely reversed: Muslims have a comparatively better socioeconomic condition, and relations with Hindus are generally marked by comity.

2.2 Communalism in Contemporary Jaipur and Ajmer: Qualitative Evidence

There is a considerable lack of credible and rigorous information on the

Muslims of Jaipur and Ajmer. This mirrors a broader problem in India, which I discussed in Chapter 2. To overcome data insufficiencies I relied on qualitative interviews with a wide array of respondents in both cases to piece together as much information as possible on the socioeconomic condition of Muslims, and the state of Hindu-Muslim relations. I conducted interviews with government officials, members of the police administration, NGO groups, religious leaders, academics, and journalists in both districts.21

In almost 300 years the Muslim population in Jaipur has not moved from the earliest neighborhoods which it inhabited – the ghettos of the 'Old City'. Nor have they changed their major professions or sources of income. For Jaipuri

Muslims time has stood still: they were born a marginalized group and have remained one. The Muslim religious leaders in the city have been a continuously important presence in their lives and are best informed about their overall situation. They were among the most important interview sources for this analysis. One Muslim religious leader noted about the Muslim community's history:

21 Information on how interview respondents were chosen, what questions they were asked, etc. are all included in the Appendix.

144 “Sawai Jai Singh [founder of Jaipur] brought labor class Muslims here when Jaipur was constructed. Because most of them were poor, they were never integrated into society and did not achieve quality education. They live in the same areas even today – mainly in the eastern part of the Old City.”22

While it is hard to come by systematic figures on the state of these Old City

Muslim neighborhoods, Shail Mayaram (1993) found a Jaipur Development

Authority report which included some basic and useful information. On the availability of water Mayaram recounts:

“It is significant to note that all those four wards, namely, chowkri Gangapol, chowkri Topkhana Huzuri, chowkri Topkhana Desh and chowkri Ramchanderji, where proportion of households per public tap is considerably high, have quite a high percentage of Muslim population residing there. And these are mostly the poor Muslim households, living under slum-like conditions, who depend on public taps for their water requirements.” (1993: 2525)

She continues more broadly:

“The story of draining, sanitation and public utility services, such as police chowkis, post offices, medical facilities, reading rooms is very much the same.” (Ibid.)

Aside from the slums in which they reside, most Muslims work in the same kinds of occupations as they did during the princely period (but now without even the opportunity for princely patronage). A Muslim NGO activist put it like this:

“Muslims here today work as artisans, in handicrafts, polishing gemstones, textiles, printing, handmade paperwork. They've done the same thing for a long time.”23

Work in the gemstone trade has remained perhaps the important source of economic stability for Muslims in Jaipur. But very few of the sellers of gemstones are Muslims; rather, they are Hindus, as was also the case during princely rule.

Exact figures are not known, but it is estimated that around 25% of the Muslim

22 Interview in Jaipur, November 30th 2010 23 Interview in Jaipur, October 27th 2010

145 workforce in Jaipur are children involved in the gemstone trade. Hindu children, in contrast, only constitute about 7% of their respective workforce (Wal 2006:

Chapter 19).

For information on communal violence in the district I relied on interviews with numerous members of the police administration – from high-level inspectors down to street-level constables. One police officer noted:

“There are periodic Hindu-Muslim riots here in Jaipur. There is not a lot of understanding between the two communities. They are segregated. And all it takes is one miscreant to start a large conflict.”24

Another stated:

“The violence erupts from time to time, but it's especially bad during 'tense' moments. Think of Ayodhya or other problems like that. Whenever things like that happen, Jaipur will become violent.”25

Riots are, of course, only one form of religious violence. There are many other types of small, every-day conflicts that may go unnoticed. A member of the

Rajasthan State Minorities Commission (RSMC) noted that during 2010, for example, there were over 500 state-wide complaints by Muslims in Rajasthan, and Jaipur was a major problem area. Of these complaints, approximately 50% dealt with violence. The main cause of the conflict was the demand (almost always from Hindus) for Muslims to leave their neighborhoods.26

Overall, the socioeconomic conditions of Muslims in Jaipur is very backward; they live largely in ghettos and work menial jobs. Furthermore, relations between Hindus and Muslims in the district are best characterized as 24 Interview in Jaipur, October 12th 2010 25 Interview in Jaipur, October 21st 2010 – Note: Ayodhya refers to a city in Uttar Pradesh which was the location of the famous Babri Mosque. This mosque was destroyed by Hindu nationalists in 1992, setting off massive communal riots throughout the country. 26 Interview in Jaipur, November 11th 2010

146 highly contentious. The communal divide in the city has provided the main axis of conflict since independence, and because of this the district has experienced periodic religious bloodshed since that time.

The state of Muslims in neighboring Ajmer is quite different. While the position of Ajmeri Muslims is not great in absolute terms, a series of interviews conducted in the district revealed that their situation is better than that of

Muslims across the border. As in Jaipur, religious leaders were the most important interviewees. One summarized the difference between Ajmer and

Jaipur:

“Things for Muslims are not so bad here. The main key here in Ajmer is that the Hindu-Muslim situation is roughly equal. There is not a lot of enmity between the two groups due to one being ahead of the other.”27

On the historical situation the respondent recalled:

“During British rule the situation was good. In the princely states everything was done according to religion. But British India was different. There was not this kind of favoritism for the dominant group. Muslims received government jobs. And most importantly, they were protected from Hindus. The British and the Muslims both feared Hindu India.”

Another religious leader echoed these sentiments:

“I wouldn't say that the situation of Muslims is great here. But there is a sense that there is some equality between Hindus and Muslims; that there is not a huge gap here. There is some discrimination against Muslims, but no real hatred between the two groups.”28

This respondent added about the colonial period:

“The British enforced order. They did not let Hindus get away with discrimination or violence against Muslims. Before Partition things here were very good for Muslims.”

Many respondents associated with Muslim NGOs spoke about the

27 Interview in Ajmer, December 22nd 2010 28 Interview in Ajmer, December 25th 2010

147 economic situation of the community. Most pointed out that Muslims worked in a variety of industries, unlike in Jaipur where they were concentrated working as artisans or in the gemstone trade. One respondent put it like this:

“In the rural areas Muslims work in agriculture, but in the urban areas Muslims work in lots of different jobs: they own their own businesses, work in construction, work for the government. Of course, a lot of them work for the dargah.”29

About the dargah the respondent continued:

“Go to the dargah area. You'll see, it's not like the situation of most Muslims. They [those that run the Dargah] are powerful people. They are well-off. They are at the top.”

Several police officers were interviewed about Hindu-Muslim violence. An overwhelming number of them indicated that riots were not a major problem in

Ajmer district. One Inspector noted generally:

“The communal condition here in Ajmer is cool. There is not a lot of Hindu- Muslim violence at all – at least not a lot of riots. There are small things here and there, like, one Hindu shopkeeper fighting with a Muslim shopkeeper. Or maybe a Muslim hits a Hindu with his scooter. But not riots.”30

One noted about 'tense' moments:

“When Babri Masjid happened I was in the police. I remember it. A lot of people were very worried, but there weren't any major riots. Definitely we have to be much more vigilant at those times.”31

The Babri Masjid episode is an important and telling one – after the destruction of the mosque, violence erupted throughout a number of cities all over India. In

Rajasthan, Jaipur experienced rioting, but neighboring Ajmer largely remained peaceful (Ghosh 1997: 106).

29 Interview in Ajmer, December 9th 2011; the dargah is a shrine of the Sufi saint Mu'in al-din Chishti. 30 Interview A in Ajmer, January 24th 2011 31 Interview B in Ajmer, January 24th 2011

148 The overall situation for Muslims in Ajmer, then, is the opposite as that of

Jaipur. Muslims, although still constituting a backward group, work in a variety of industries and interview respondents were more apt to display optimism about their community. Furthermore, Hindu-Muslim relations are not a major source of tension in the district, and religious violence has therefore been minimized.

The position of Muslims in Jaipur and Ajmer are mirror images of one another, and religious interaction in both districts differs greatly. The overall result is that

Jaipur has periodically exploded into communal violence, but Ajmer has remained peaceful.

2.3 Existing Explanations for Religious Violence in Rajasthan

It is undoubtedly the case that religious violence, especially Hindu-Muslim riots, began a disturbing increase during the final decades of the 20th century. The rise of a group of militant Hindu nationalist organizations – the Sangh Parivar

(Family of Associations), which includes the BJP and the RSS – is often identified as a key causal factor in the increase of violence. All three of the main political scientists working on Hindu-Muslim riots in contemporary India have pointed to the Sangh Parivar as a central promoter of violence (Brass 1997, 2003; Varshney

2002; Wilkinson 2004). All of their individual analyses highlight important features driving contemporary ethnic violence, but I argue that none of their arguments can explain the particulars of the Rajasthan case. I considered these arguments in the Introduction, but it is worth considering them again in light of the specifics of the Jaipur-Ajmer comparison.

Brass summarizes his main argument as follows:

149 “Hindu-Muslim opposition, tensions, and violence have provided the principal justification and the primary source of strength for the political existence of some local political organizations in many cities and towns in north India linked to a family of militant Hindu nationalist organizations whose core is an organization founded in 1925, known as the Rashtriya Swayamsevak Sangh (RSS).” (2003: 6)

As I have stated previously, Brass is correct to note that Hindu nationalist organizations have played a major role in promoting religious violence. But as I have also noted previously, riots pre-date the rise of these organizations. Even if we take into account that the BJP traces its roots to colonial organizations such as the Hindu Mahasabha, riots still pre-date these organizations. Second, if these organizations are the main promoters of violence then we might expect Ajmer to have more Hindu-Muslim conflict than Jaipur. After all, Hindu nationalist organizations on the whole throughout India laid down their roots in directly- controlled British areas, and this was also true for Ajmer-Merwara. In fact, in the

1930s Ajmer was one of the only places in Rajputana where the RSS was active

(Mayaram 1993: 2538). Only in the final years of colonial rule were Hindu nationalist organizations able to begin organizing in princely areas like Jaipur. In modern times, the BJP as a local party has been successful in both districts – the party dominated Lok Sabha elections in both constituencies from 1989-2009, so this cannot be the source of the variation which exists in religious violence outcomes.

Next, recall Wilkinson's argument. He summarizes it as follows:

“My argument…is that ethnic riots, far from being relatively spontaneous eruptions of anger, are often planned by politicians for a clear electoral purpose. They are best thought of as a solution to the problem of how to change the salience of ethnic issues and identities among the electorate in order to build a winning political coalition.” (2004: 1)

150 The main problem with Wilkinson's hypothesis is obvious: political parties and competitive elections did not exist during most of the colonial period yet riots still occurred. There can be little doubt that contemporary politicians manipulate ethnic issues and offer protection to minority ethnic groups who are politically valuable to a coalition. However, riots occurred well before the introduction of any kind of electoral system in India. Therefore, while Wilkinson may be correct to note the role played by political parties in promoting contemporary ethnic violence, his work leaves the deeper causes of conflict unaddressed. Furthermore, the particular constellation of ethnic groups in both Jaipur and Ajmer is similar –

Muslims form roughly the same percentage of the population in both areas, so they are not more valuable as political actors in one district.

Finally, Varshney has argued that civil society plays the key role in determining where outbreaks of communal violence occur. Specifically, he argues that intercommunal networks (those that include both Hindus and Muslims) can prevent violence. These networks can exist in two forms: associational or quotidian. Associational organizations are formal; examples might include business associations, and trade unions. Quotidian organizations are informal; they might include acts as simple as Hindu and Muslim families routinely eating together. Varshney finds survey evidence that where intercommunal networks exist violence is minimized.

There are a number of major problems with this argument, some of which

I highlighted in the Introduction. Recall, India is not a country with a rich

151 associational life. Only a very small percentage of Indians belong to any voluntary association (Chhibber 1999). Second, in regards specifically to Jaipur and Ajmer, it is not clear that civil society groups matter for violence outcomes. For example, perhaps it is the case that Ajmer has more intercommunal organizations than

Jaipur and this led to the minimization of violence. To test this argument, all interviewees in Jaipur and Ajmer were asked about the existence of Hindu-

Muslim organizations, but no one was able to name even one such formal association which existed.32 Therefore, it seems unlikely that these organizations play a role in reducing ethnic conflict.

III. Caste and Tribal Violence in Contemporary Rajasthan

In this section I turn to the other major type of ethnic conflict in contemporary India: caste and tribal violence.33 The main source of data for violence against SCSTs usually comes from the Ministry of Home Affairs (MHA).

However, MHA data are generally considered by researchers to be very problematic for a number of reasons which I discussed in Chapter 2. Because of these shortcomings I used figures on caste atrocities compiled by a local organization, the Centre for Dalit Rights (CDC) in Jaipur. These data unfortunately only pertain to violence against SCs but they were available for several districts in Rajasthan, including Ajmer.

CDC data on caste violence are superior to MHA data for a few reasons.

First, SC crime victims are often reluctant to file reports of violence against them

32 Some interviewees, for example, would claim that they existed but could not provide the names of any actual organizations. 33 WITS data are not used here because Jaipur and Ajmer do not have militarized conflicts; rather, most of the caste and tribal conflict is low-level atrocities.

152 because they believe that the police will not be helpful, or often harass them and make the situation worse. The CDC, on the other hand, collect data through three avenues. First, victims who are aware of the organization will come directly to its office and report crimes. Usually these victims will not trust the police to properly register incidents. Second, volunteers (such as witnesses) submit information about crimes. CDC has a network of over 1,000 lawyers and activists. Third, CDC employees will use recorded instances of caste crimes in print and electronic media, offering their services to victims by conducting fact-finding missions and even submitting their reports to government officials if no action is taken. For all of these reasons CDC data serves as a better source of information for violence outcomes. The results of a comparison between Jaipur and Ajmer district for the year 2009 using CDC data are detailed in Table 3.3 below:

Table 3.3: 2009 SC Violence in Rajasthan (per 100,000)34 Jaipur Ajmer Crimes Against SCs 9.0 26.0

The results showcase that now Ajmer is the violent district: it experiences more caste and tribal conflict than Jaipur. To understand the contemporary socioeconomic position of SCSTs in both districts, as well as relations between caste groups in these areas, I conducted a number of interviews there, detailed in the next section.

34 These figures are calculated as the number of crimes against SCs divided by the SC population (2001). This only pertains to the year 2009.

153 3.2 Caste and Tribal Violence in Contemporary Ajmer and Jaipur: Qualitative Evidence

A variety of respondents were interviewed about the state of SCSTs in

Ajmer and Jaipur. One Jaipuri journalist summed up the comparative state of caste and tribal conflict in these areas:

“Ajmer struggles with caste and tribal issues in a way Jaipur does not. You can read the papers and you'll see the violence that happens there on a daily basis. The Dalits are usually the victims, but STs suffer too. The high castes will do anything to ensure that the position of Dalits does not change. They will prevent them from getting an education, from getting political representation – anything.”35

In fact, a Times of India article from 2009 summed up the political history of

Ajmer by noting that “[it] has been dominated by caste politics.”36 Caste still plays the central role in society there today; it is the particular ethnic identity which reverberates the strongest. This has always been the case; in fact, Jawaharlal

Nehru wrote in the 1950s that the local Congress Party was split badly into caste factions, and that development efforts there were among the least successful in

Rajasthan (Naipaul 1971).

Another respondent from an NGO noted that the position of low castes has not changed considerably over time:

“The government in Rajasthan has made a lot of promises to SCs, but no one trusts the government to fulfill its promises. The law states, for example, that a certain percentage of the state budget must be put toward programs for Dalits. We monitor this. But it never happens.”37

Respondents were also asked about the implementation of caste reservations

35 Interview in Jaipur, September 9th 2010 36 “Ajmer Sees a Change in Caste”, by Kshitiz Gaur, Times of India, May 7th 2009; see also the article “Pilot Bridges Caste Differences in Ajmer” by the same author in the same paper, June 1st 2009. He notes there that “Ajmer was dictated by caste politics.” 37 Interview in Jaipur, September 16th 2010

154 after the Mandal Commission.38 A journalist summed up the result of the changes brought on by the law:

“Mandal does not help. The reservation system only lends more credence to the caste system and encourages discrimination and anger against Dalits.”39

Along these lines, some interesting qualitative work has been done on the viewpoints of SCs in Rajasthan. Narayan Mishra's work (2001: 240) on the perceptions of SC students toward education reforms in the state found that of a number of districts surveyed in Rajasthan, Ajmer was the only district where less than 50% of SC students agreed that the reservation system set up by the Mandal

Commission has been 'fairly useful'.

On the specific causes of violence, the aforementioned journalist respondent stated:

“What you will notice is that the situation in the villages around Ajmer has remained the same. The Shudras (OBCs) have become immensely powerful. They are the ones who commit violence against SCs, not high caste Hindus. When a Dalit is killed, a Shudra is usually behind it. They will not allow any of their gains to go to the Dalits.”40

Christophe Jaffrelot (2000) has noted that most OBCs in North India were land holders during the British period and have continued to maintain their dominance in the countryside. This is the case in rural Ajmer today. The SCSTs who bear the brunt of violence in contemporary Ajmer find themselves targeted by powerful OBCs. An NGO respondent noted:

“The main problem, the main cause of the violence is casteism...This means the strength of the caste system. In some areas it is very rigid and in other areas it is

38 The Mandal Commission created institutionalized quotas in education and government representation for groups who have faced historic discrimination: Other Backward Classes (OBCs), and SCSTs. 39 Telephone interview in Jaipur, September 14th 2010 40 Telephone interview in Jaipur, September 14th 2010

155 not. Ajmer has a very rigid caste system, and the Dalits there suffer for it. Any attempts to reorganize the caste system result in violence. A Dalit bridegroom cannot even mount a horse for his wedding because this is an upper caste custom. If he does, there will be a backlash.”41

In Jaipur, on the other hand, most respondents indicated that caste violence was not the major problem in the district, citing instead the salience of religious identities. One journalist noted:

“The violence in Jaipur is communal. We all wait for the next Hindu-Muslim riot. There are some caste issues in the rural areas outside the city – in the villages – but most of the conflict here is communal.”42

When asked to compare Jaipur and Ajmer the respondent added:

“The last major conflagration here in Jaipur was in 2008 when bombs were detonated in the Old City. If you ask people about violence, this is what they think of – Hindus and Muslims fighting. But caste violence – you won't find strong opinions about that. But if you go to Ajmer – you'll find a lot of caste conflict. The issues are very complicated there.”

Furthermore, Mishra's work on Scheduled Caste students revealed that in Jaipur the highest percentage of students (74.1%) agreed that educational facilities provided by the government for them were 'quite useful' (2001: 236), indicating a major discrepancy in the way Jaipuri and Ajmeri SCs perceive the usefulness of government reservations and special programs. Ajmeri students were much less sanguine than their counterparts about their educational opportunities.

The socioeconomic position of low castes and tribals in Rajasthan differs greatly between Jaipur and Ajmer. In the former district, religion has formed the main axis of conflict, and caste violence has not been a major problem. In Ajmer, however, SCSTs remain stuck in a rigid caste system, and violence in rural regions is especially bad. As I will detail later, higher levels of caste and tribal 41 Interview in Jaipur, September 16th 2010 42 Interview in Jaipur, September 29th 2010

156 conflict in Ajmer were the result of historic discrimination, rooted in the treatment of low caste and tribal groups during the colonial period.

IV. The Colonial Roots of Religious Violence in Rajasthan

In the ensuing two sections I turn to an examination of the historical roots of religious, caste, and tribal conflict in Rajasthan. As I have noted previously, most existing political science work has focused squarely on the contemporary period; however, in the next two sections I investigate the deeper, underlying causes of the variations in ethnic conflict experienced in Jaipur and Ajmer today.

I begin by examining religious violence – I argue that the poor socioeconomic condition of Muslims in Jaipur, as well as the higher levels of religious tension in the district, are due to a legacy of historic discrimination dating back to the period of princely rule. In Ajmer, on the other hand, Muslims were a protected minority group during British rule, and therefore Hindu-Muslim conflict has been minimized in the district today. In section V I then turn to the colonial roots of caste and tribal conflict.

Jaipur has often described throughout its history as a city of communal peace. Shail Mayaram, for example, writes that “Rajasthan...has hardly had a history of communal violence” (1993: 2524), and that princely states like Jaipur

“...by and large, maintained their record of communal harmony” (2525). Stern notes that Jaipur “enjoyed a well-earned reputation for...communal peace”

(1988: 276). Ian Copland comments that the “Jaipur government patronized

Mohurrum43 as well as the the tomb of the Sufi saint Mu'in al-din Chishti at

43 Mohurrum is a Muslim festival.

157 Ajmer” (1998: 212). Contrary to these standard accounts, I argue here that Jaipur perfectly describes the false claim of syncretism in the princely states which was discussed and critiqued in Chapter 1 (section 2.6). In reality, Jaipur was a communally-sensitive city and remained so for most of its pre-independence history. Additionally, the cause of this long history of religious violence began with the discriminatory policies of the Jaipur princely state toward the Muslim minority. In contrast, I show how communal violence has been historically much less of a problem in Ajmer, where British administrators pursued policies that greatly benefited Muslims and reduced the long-term opportunities for violence.

4.2 Communal Policies of the Jaipur State

The rulers of the Jaipur State were orthodox Hindus and they established and ruled Jaipur as a Hindu state. The religiosity of the founder of the city Jai

Singh II was well-known:

“[Singh] was personally very religious and well versed in theology. About the time that he began to construct Jaipur, he received from the Mughal emperor the titles of Maharaja, Mirza Raja and Sawai. Monika Horstmann points out that the title of Mirza Raja, which denotes a kinship relation with the Mughals, was not used in his own documents since it did not fit his image as a ruler...by 1727, when Jaipur was founded, a new seal was adopted that proclaimed allegiance to the deity Sita Rama, not to any Mughal overlord.” (Datta 2008: 236)

Singh embraced Hindu customs and traditions, even designing the city of Jaipur according to specific Hindu precepts. He also famously performed the Vedic

Asvamedha sacrifice44 in Jaipur, which to that point had not been performed by any Hindu ruler in some nine centuries (Ibid.: 232). Much of this was typical of the Rajput states. James Tod, famed British chronicler of the area, noted that

44 Horse sacrifice.

158 “[t]here is scarcely a state in Rajpootana in which one-fifth of the soil is not assigned for the support of the temples, their ministers, the secular Brahmins, bands and genealogists.”45 He also observed the powerful role that Brahmins played in the raja's daily life:

“The [king] is often surrounded by lay-Brahmins as confidential servants, in the capacities of butler, keeper of the wardrobe, or seneschal [majordomo], besides the Guru or domestic chaplain, who to the duty of ghostly comforter sometimes joins that of astrologer and physician, in which case God help the prince!”46

Like many other princely states in Rajputana, a large portion of the land in Jaipur was also given to priests tax-free as a gift (inam) and used for the erection of temples. Rulers throughout Rajputana generally dispensed large amounts of patronage to religious institutions. As I will show in further detail later, this religiosity also extended to later rulers of Jaipur. For instance, in 1911 The Times published a collection of articles to celebrate 'Empire Day' and singled out for their readership the Maharajah of Jaipur as a “prominent...ardent supporter of orthodox Hinduism” (Times 1911: 26).

When Jaipur was initially being constructed Jai Singh II brought a large contingent of Muslim laborers to his new capital city. Jaipur is famous as one of the earliest planned cities in India and Muslim labor played a large role in constructing it. Later, once the city had been constructed, Jaipur's Muslims worked in primarily two occupations: either as artisans or in the large international gemstone trade. In the latter, they mainly worked on the manufacture of gems, i.e. purchasing stones and polishing them. Most of the

45 Tod 1829: 403 46 Tod 1829: 406-7

159 actual sellers of gemstones were well-to-do Hindus. As I noted earlier, the main occupations for Muslims have not changed in almost three hundred years.

It has been argued that Muslims were an integral part of the fabric of

Jaipuri society, and further that the Jaipur State had a general record of tolerance and inclusion toward them. Mayaram notes that “Although [Jaipur] was theocratic, its rulers were known...for their financial support to the Ajmer dargah and to tazia processions. Persian was the language of the state administration and

Muslims were an integral part of the ruling elite and revenue administration...”

(1993: 2525). Rajat Datta notes that “Jai Singh saw himself as a dharmic

[religious] king...but his state did not promote policies that discriminated against non-Hindus” (2008: 236). It is also frequently noted that the last maharaja of

Jaipur, Sawai Man Singh, personally intervened during Partition to prevent violence against Muslims and implored them not to emigrate from the state.

A deeper look into the record of the Jaipur State, however, reveals a starkly different picture. In contrast to official pronouncements or public displays of tolerance and inclusion, the princes of Jaipur in reality enforced many discriminatory policies against the Muslim minority. These policies were constant across several different . While official state policies were rarely openly discriminatory, underneath the surface Muslims suffered under the control of a Hindu raj. They were stunted economically, socially, and politically.

First, consider Mayaram's argument about the inclusion of Muslims in the

Jaipur government. Using data collected from the City Palace Archives in Jaipur,

Table 3.4 below displays their representation in the Jaipur State during the 1920s

160 and 1930s (the only period for which data were available). As I show, Hindus in reality dominated the upper echelons of the princely government. The Council of

State, known as the Mahakma Khas, was the central governing institution of the kingdom and consisted almost entirely of Hindus. For example, during the year

1929-30 Hindus were in charge of the Vice-Presidency, Foreign & Home

Ministry, Finance Ministry, Military Ministry, Education Ministry, and a Hindu was the Chief Justice of the Court. Only the Public Works Ministry was headed by a Muslim.47

Table 3.4: Jaipur State Civil List, Mahakma Khas, 1929-30 Position Office-Holder Vice-Presidency Mr. B.J. Glancy Foreign and Home Rai Bahadur Pandit Sri Gopi Nathji Minister Finance Minister Rai Bahadur Pandit Amar Nathji Atal Revenue Minister C.L. Alexander Khan Bahadur Maulvi Mahomed Ashfaq Public Works Minister Hasan Khanji Military Minister Thakur Devi Singhji of Chomu Rao Bahadur Thakur Narendra Singhji of Education Minister Jobner Chief Justice Raj Bahadur Pandit Seetla Prasadji Bajpai Inspector General Mr. L.C.B. Glascock

Only one Muslim headed a ministry in 1934-5 and again in 1937-8.48 Even the posts headed by Muslims were not always secure. One British administrator noted in 1940:

“I spoke to Resident [sic] and also to the Prime Minister at Jaipur a few days ago about the question of a successor to Khan Bahadur Abdul Aziz as Revenue Member, Jaipur. Apparently the Maharaja has now decided that he would definitely prefer to have a Hindu Revenue Member and intends to appoint a Muslim to hold some other portfolio.”49

47 JCP, Report on the Administration of the Jaipur State, 1929-30 48 JCP, Report on the Administration of the Jaipur State, 1934-5 and 1937-8 49 BL, IOR/R/1/1/3346, 1939

161 Moreover, as Table 3.4 also shows, the Jaipur State was much more likely to appoint the British to important positions rather than Muslims. In 1937-8, for example, six of Jaipur State's major bureaucratic posts were held by British men

(Rudolph and Rudolph 1966: 140).

In 1942 Jaipur's raja Sawai Man Singh did appoint a Muslim as his

Dewan (Chief Minister): the acclaimed Mirza Ismail. Prior to coming to Jaipur,

Ismail had been well-known throughout India for his efforts to rapidly modernize the south Indian . Despite having a Muslim in the highest office the benefit to the Muslim population was minimal. For example, when

Mirza instituted a constitutional reform process Muslims were left on the outside looking in. Stern writes:

“The Muslim members of the constitutional reform committee wanted for their constituency in Jaipur the same pattern of positively disproportionate representation, 'weightage', and separate electorates that had been enjoyed by Muslims in British India since the Morley-Minto Reform of 1909. They didn't get it...Most of Jaipur's Muslims were among the state's poorest and least educated people, and they were for the most part unenfranchised by Mirza's constitutional reform.” (1988: 317)

As I will detail in the succeeding chapters, this was a common pattern – Muslims in the princely states often held as their standard for treatment the way their counterparts lived in British India. The idea of disproportionate representation in the government was especially coveted. Later I detail how the Muslims of Ajmer fared in this regard.

In the early 20th century, all of the princely states were devolved greater levels of autonomy in their internal affairs. During this time Jaipur became more hardened in its discrimination of Muslims. For its treatment of them the princely

162 government periodically became the target of complaints by the Muslim League during the 1930s and 1940s. For example, the League wrote about the lack of responsiveness by the Jaipur government toward the condition of the Muslim community:

“[We have] every sympathy with the Muslim Muhajirin from Jaipur who have been undergoing great hardships and suffering...The Committee urges upon the Jaipur Durbar to meet the just demands of the Jaipur Muslims…” (Jinnah 1996: 378)

They commented again on the same topic (this time including a mention of other princely states):

“…in the various Indian States namely Jaipur, Jodhpur, Bhavanagar, Cutch and Patiala...The Committee most earnestly urges upon the rulers of the States concerned to redress the grievances and protect the lives, honour, properties and liberties of their Muslim subjects thereby creating a sense of security amongst them and preventing the situation from taking a serious turn.” (Ibid: 642)

The British, after spying on his personal mail, believed that the last prince of Jaipur Sawai Man Singh (1922-1949) was biased against Muslims and had connections to the Hindu Mahasabha (Stern 1988: 304). For Muslims, the situation under Man Singh grew worse. Even mundane affairs resulted in discrimination. One of the things taken for granted by authors who have argued that the princely states were syncretic is the right of religious minorities to worship without interference or harassment. But even this was not always guaranteed in Jaipur. In January 1939 the Muslims of Jama Masjid (the largest mosque in the city) submitted an application to Man Singh to expand the entrance of their mosque, as the walkway was too small for Muslims to enter and exit at the same time. The request was denied. Frustrated Muslim leaders ordered that construction begin anyway. In response, the prince sent his police to halt the

163 project. A small disturbance ensued and six Muslims were shot and killed on the spot.50 Syed Ghulam Bhik Nairang, a member of the Legislative Assembly

(Central) created a report on the incident for the Muslim League Party of the

Central Legislative Assembly entitled “The Jaipur Juma Mosque Tragedy”.51 Even the Muslims of British India, including those from Ajmer, protested against the actions of the Jaipur darbar52 – that is, British Muslims had to come to the defense of their princely coreligionists.

At the same time, instances of communal violence in Jaipur State increased. One communal riot broke out in 1926 on the outskirts of Jaipur city.

Another larger riot occurred in 1927 as the entire city was shut down (hartal) for five days.53 In 1931 Mohammed Ishaq, editor of the newspaper Muballigh wrote to the editors of the Muslim Outlook newspaper about another riot that year:

“Hindu Muslim tension at [Jaipur] has assumed [a] dangerous aspect.”54 He continued:

“Hindus have completely boycotted Muslims who threatened with starvation, scarcity of water and intolerable aggressiveness of Hindus have left home and migrated to neighbouring places. Cemetary fences have been burnt by Hindus and no legal steps are taken against perpetrators while Muslims are being involved in false cases. Tahsildars and Police Officers are doing their worst to oppress Muslims.”55

Muslims wrote to the British Resident at this time:

50 1939 Jama Masjid Financial Report (Urdu) 51 BL, IOR/R/1/1/3306, 1939 52 Ibid. 53 These data are from the Report on the Administration of the Jaipur State for the aforementioned years and also The Jaipur Album. See Bibliography for further details. 54 NAI, Jaipur Residency, Part II, 1933, #466 55 Ibid. – Note: Many of the documents that will be quoted hereafter contain spelling or grammatical errors, some of which I have corrected. This may be because the documents were often telegrams or short notes in which the author was trying to conserve space. Also, other messages were written very hastily during the middle of a disturbance or riot.

164 “...[Hindus] have combined together to reduce us to abject misery and distress and to compel us to leave our homes...They made undue claims upon our religious rights and places of worship, they stopped business and dealings with us and they have banned their schools and dispensaries upon us. We are poor and helpless and depend upon our Hindu brethren. Our plight can therefore be well imagined...”56

For their part, Hindus blamed the 1931 conflict on Muslims, noting in their own letter to the Resident:

“...the direct and the immediate cause of the present disturbance is that the Mohammedans forcibly constructed a wall in night hours on Mohurram night with beat of drums and 500 armed persons outside the mosque on Khalsa (Thikana owned) land and took a Pipal tree inside it.”57

A.C. Lothian, the British Resident in Jaipur, seemed to side with the Muslims. He noted in 1933 that the Jaipur government played a role in promoting the violence:

“Relations between Hindus and Muslims of the Jaipur State are very strained. One of the main grievances of the Muslims is that they are being denied services in the Household and other departments under the verbal orders of the Maharaja. The posts which were formerly occupied by Muslims are now being given to Hindus in general and to Rajputs in particular.”58

Another major incident occurred in 1936 over the Muslim festival of

Mohurrum. From contemporaneous accounts and documents it is impossible to know with certainty what happened. However a press communique from 1936 noted:

“The Moharram processions on the Ashra or 10th day of Moharram were marred in Jhunjhunu town, the Head-Quarters of the Shekhawati District of the Jaipur State, by an untoward incident which occurred when the Chejaras or Mohammedan masons completely demolished a wall of a Hindu house, which they alleged interfered [sic] with the passage of their Tazia. The demolition was carried out by a large mob of Mohammedans in defiance of the orders of the Nazim or District Magistrate, who was present on the scene...[n]ext day the

56 NAI, Jaipur Residency, Part II, 1933, #466 57 NAI, Jaipur Residency, Part II, 1933, #466 58 NAI, Jaipur Resident Part III, Confidential, 1933, #1134

165 Hindus observed a complete hartal...”59

Another incident occurred in 1946 in Jaipur and this time the Muslims appealed to their new Muslim Dewan Mirza Ismail. They described the incident:

“In the Jaipur State...the Musalmans are of late facing circumstance [sic] which are far from happy...On the 18th March, 1946, certain local Hindu goondas [thugs] decided to take out their holi60 procession with music being played before the mosque, with threats of social boycott and violence to the Musalmans in case they placed obstructions in their way...When the Musalmans were engaged in their Zor (afternoon) prayers, the goondas raised a huge noise and pounced yelling upon the defence-less Namazis inside the mosque. The Police Constables were also helpless. The Mussalmans were taken unaware and received bullet wounds, lathi blows and beatings in quick succession.”61

The sharp increase in discrimination and resultant episodes of communal violence prompted The Sarosh, an Urdu weekly in Bombay, to write tellingly on the eve of independence in 1946 that: “...so far as repression and excesses against the Musalmans are concerned this state [Jaipur] has surpassed even the states of

Kashmir and Gwalior.”62 The Jaipur Police Superintendent noted the next year that “...[this all] goes to show that the feelings amongst the Hindus and

Mohammedans in Jaipur are somewhat strained, and might burst forth any day.”63 And in fact further riots occurred again in 1947 and 1948 as the country was partitioned.

Overall, the situation of Muslims in Jaipur was – in sharp contrast to traditional accounts – distressed. They were subjected to economic, social, and political discrimination by the Hindu-dominated Jaipur princely government.

Periodic explosions of violence were the result of this discrimination – Jaipur's

59 NAI, Jaipur Residency, Part II, 1937, #278 60 Holi is the Hindu festival of colors. 61 NAI, Jaipur Residency, Part II, 1944, #214 62 Ibid., emphasis added 63 NAI, Jaipur Residency, Confidential, 1948, #1-P/48

166 Muslims agitated against their maltreatment while the Jaipur government and the Hindu majority of the princely state used force to maintain their strength.

Tensions between both communities consistently ran high. The syncretism that existed on face in the Jaipur State was merely a facade; the government badly mistreated Muslims for most of its history. Even after Partition, several Muslim members of the Jaipur government who had emigrated to Pakistan complained that their back wages had never been delivered to them.64 Post-independence and contemporary communal violence in Jaipur are therefore only the latest episodes in a long history of conflict which first began during the princely period. Both the socioeconomic stagnation of the Jaipuri Muslim community today, as well as the sharp Hindu-Muslim divide which dominates the district, are direct relics of the princely period.

4.3 Communal Policies of Ajmer-Merwara Province

The comparative situation in Ajmer during the colonial period was very different from Jaipur. The city was relatively peaceful during British rule. There were only two riots which occurred prior to Partition. The first was in 1923.65

Authorities at the time noted:

“As a result of Hindu processions playing music near mosques in Ajmer two collisions have occurred between Muhammadans and Hindus. Bitter feeling prevails between the two communities.”66

Another contemporaneous telegram noted:

64 NAI, Ministry of States, G(R) Branch, 1948, #11(1)-G(R)/48 65 NAI, Rajputana Agency, Chief Commissioner's Branch, 1923-4, #1195; Ian Copland had argued that Ajmer was a riot-prone city, and that it “was the first town in Rajputana to blow up, communally speaking...” (2005: 203). The second statement is true, but the first is not. Ajmer had two minor riots prior to Partition, and communal violence in neighboring Jaipur was considerably worse during the same period. 66 Ibid.

167 “The first signs of trouble occurred on the night of the 15th instant when a Hindu procession was set upon in the Durgah Bazar by Mohamedans. This was a small affair and was adequately dealt with by the Police in half an hour at the cost of a dozen arrests and the disappearance of a gold bangle from the idol.”67

During the riot five people died while there were over one hundred injuries.

Over the next decade – when Jaipur plunged into communal chaos – there was only one minor riot in Ajmer (in which no individuals were killed):

“On May 22nd, at 11.10 a.m. Constable Lahri Singh of the local C.I.D. Reported to the Kotwali on telephone that a riot of communal nature was taking place at the gate of the Carriage shops and that stones were being freely pelted. The S.O. Kotwali accompanied by some Sub-Inspectors, Head Constables and constables hastened to the scene. The rioting was then going on but it was brought under control... It was ascertained that 19 workmen, including 11 Hindus and 8 Muslims had received injuries during the course of fracas [sic].”68

The Fortnightly Reports69 located in the British Library in London contain detailed reports of the communal situation in Ajmer during a wide swath of the early 20th century. The reports paint a picture of a relatively peaceful province in communal terms. Most communal incidents recorded were minor in nature. For example, a dispatch from February 1931 stated:

“In consequence of the fact that the month of Ramzan coincides this year with the Hindu marriage season, there has been some slight friction over the playing of music by Hindu processionists when passing Mosques at the time of Tarahvi prayers. The attitude of the leaders on both sides appears to be reasonable and no serious developments are apprehended.”70

Or consider this dispatch from May 1931:

“The Muharram celebrations passed off peacefully, although there were one or two contretemps. In one case it was alleged that water had been thrown from a house occupied by Hindus on the party accompanying a Tazia. The police pacified the Taziaars and induced them to pass on.”71

67 NAI, Rajputana Agency, Chief Commissioner's Branch, 1923-4, #1195, Confidential Letter #36C 68 BL, IOR/L/PJ/7/1079, 1936 69 See Appendix for more information on this and other sources. 70 BL, IOR/R/1/1/2068, 1931 71 Ibid.

168 At other times the British noted no disturbances of any kind whatsoever. In

January 1922 the dispatch stated: “There is a distinct lull in agitation in the City of Ajmer. The Commissioner states that the City is quieter than it has been for 13 months.”72 In 1943 when Jaipur experienced rioting, the December dispatch for

Ajmer-Merwara noted under the heading of communal conflicts: “Nothing of interest”.73 The reports also indicate that most of the demonstrations and rallies in Ajmer were not ethnic but rather related to the independence movement. Civil disobedience and non-cooperation were much more active in Ajmer than Jaipur as the INC was much less active in general in the princely states. The last major incident of communal violence in Ajmer was during Partition, although almost the whole of north India experienced some violence during this time.

There are two possible existing explanations for the lack of communal violence in Ajmer during the colonial period. First, is it possible that Jaipur was less able to prevent violence than Ajmer-Merwara? To answer this question I looked at the size of the police force in both Jaipur and Ajmer during the colonial period. In Ajmer in 1903 there were a total of 704 policemen in the regular police force. There were also 2,505 combined men in the Imperial and Native armies stationed in Ajmer-Merwara.74 Jaipur, on the other hand, had 877 policemen; the army consisted of 5,000 infantrymen and another 5,782 cavalry were under the control of the nobles who served the darbar.75 If we consider population figures

72 BL, IOR/R/1/1/1362, 1923 73 BL, IOR/L/PJ/5/283, 1943 74 Imperial Gazetteer of India (Vol. 13): 165 75 Ibid.: 397

169 from the 1901 census, this means that Ajmer had 1 police officer for every 148 citizens while Jaipur had one police offer for every 228 persons. While there is a clear disparity, it does not seem likely that Jaipur lacked the capacity to prevent violence, especially considering princely states could be much more brutal in the application of force than the proto-democratic provinces.

One other explanation for the lack of violence in Ajmer comes from the nature of Islam in the district; specifically, the dominance of Sufism. Mu'in al-din

Chishti, the founder of the dargah in Ajmer, was a Sufi saint. A few interviewees in Ajmer made reference to the special impact of Sufism as a tolerant form of

Islam. Yet, the existence of a strong Sufi identity in Ajmer is a poor explanation for the lack of communal violence. It is not clear that militant Hindu groups make any distinction between the different sects of Islam; rather, they uniformly oppose the entire religion. Similarly, the dominance of Sufism does not dissuade

Muslim groups that initiate violence (as was the case in 2007). Along these lines, the dargah in Ajmer does bring together the Hindu and Muslim communities, thus potentially creating social and economic ties that could prevent violence.

But, Saumitra Jha has found some statistical evidence that Muslim pilgrimage sites have increased chances of experiencing religious violence because rival religious communities frequently practice their rituals in close proximity to one another (2008: 20). After all, Babri Masjid in Ayodhya was once an important relic for both Hindus and Muslims – and this was a major part of the reason it was the site of such communal tension and violence.

Overall, Ajmer experienced much less communal conflict during the

170 colonial period than neighboring Jaipur. Before Partition there were only two small riots and much less total bloodshed in Ajmer, even taking into account the population disparity between the two territories. Jaipur, compared to Ajmer, was a city of deep communal discord. This fact must also be considered in context. In many ways, Ajmer is a most likely case for Hindu-Muslim conflict. We might expect Ajmer to have experienced a large amount of communal conflict during colonialism for a number of reasons. First, it was the last capital of the Hindu

Empire in India, and therefore an important historical site for Hindu nationalists.76 David Ludden has noted that the local fort of Prithvi Raj , the last Hindu emperor, “has today become an icon of militant Hindu nationalism” (1999: 112). Second, Ajmer was home to several powerful Hindu organizations who used it as an important meeting point. It was the headquarters of the Hindu reform movement , founded by Dayananda Saraswati

(Sikand and Katju 1994). In 1925 a massive gathering of Hindu Jats occurred right outside Ajmer, overseen by the ruler of the princely state Bharatpur, during which the crowd passed a resolution demanding the forced conversion of Muslim

Jats. Further, as I noted earlier, Ajmer is an important pilgrimage site for both

Hindus and Muslims. For the former, nearby Pushkar is a devout site for the worship of Brahma; for the latter, Ajmer is home to the dargah of the famed Sufi saint Mu'in al-din Chishti. Yet, despite all these factors which might produce high levels of conflict, Ajmer was a relatively peaceful place and experienced less communal violence than neighboring Jaipur. How did Ajmer keep the peace?

76 NAI, Foreign & Political Branch, Internal Branch, 1925, #100-I of 1924

171 One reason for the comparative lack of violence in Ajmer was the fact that

Muslims lived in better conditions and the gulf between Hindus and Muslims was not as wide as that of Jaipur. Unlike their backward state in Jaipur, Muslims were an integral part of the ruling elite in British-controlled Ajmer-Merwara.

Table 3.5 below displays the number of government posts Muslims occupied in

Ajmer in 1925.77 The data show that Muslims were actually greatly over- represented in Ajmer.

Table 3.5: Communal Representation in Ajmer Gov't, 1925 Total Jobs Hindus Muslims Percentage Muslim 2340 1241 1051 44.91%

Muslims held almost 45% of all jobs in the province in 1925. These jobs were also distributed across various departments of the government, unlike in Jaipur where a Muslim generally governed only one department. Muslims also held several of the top government posts – Additional Sessions Judge, Public

Prosecutor, and Inspector of Police, for example. In the early 1940s the Chief

Commissioner of Ajmer-Merwara was also a Muslim, Khurshid Ahmad Khan.

The British noted in 1925 that overall “out of 73 posts of a superior nature 27 are held by Mohammedans and 46 by Hindus”.78 This means that Muslims controlled almost 37% of the top posts in Ajmer-Merwara. They even surprisingly comprised a majority of the Ajmer police force. This, of course, may have played a crucial role in limiting communal violence. A majority Muslim police force may have worked hard to prevent spasms of violence in which Muslims generally suffered disproportionately. Moreover, Muslims dominated the key posts in the police – of

77 NAI, Foreign & Political Branch, Establishment Branch, 1925, #509-E 78 NAI, Foreign Department, Establishment Branch, 1926, #1-4

172 87 Head Constables, 59 (67.8%) were Muslims and only 28 were Hindus.79

Information about Muslims in government posts was released by the

British in response to complaints from newspapers that Ajmer constituted, as they put it, a 'Hindu Raj'.80 The British responded by conducting an internal audit of government posts and publicly releasing the information. This whole process was carried out in stark contrast to the autocracy of the princely states. Religious minorities there had fewer outlets to complain or organize for a redress of grievances. In 1939, for instance, Jaipur banned several major newspapers – The

Hindustan, Vir Arjun, Jaiya Praja, Sainik and Qaumi Gazette – due to

“scurrilous editorials intending to excite disaffection both towards the Ruler of the State and his Government”.81 This was a common practice in princely states throughout India. Critiques of the government, especially on the grounds of religious intolerance, were not prohibited.

The Chief Commissioner of Ajmer S.B.A. Patterson noted to his superiors about the number of Muslims employed in government service in Ajmer in 1925:

“...[the data] will show that taking into account the compete total of employees a larger number of Mohammadans is employed than their proportion to the population warrants.”82

Likewise, the Commissioner noted that Muslims “enjoy a position of very considerable weight”.83 Another Ajmer-Merwara official, Denys Bray, responded to the complaints from Muslim newspapers by noting:

“In view of the fact that Mahommedans number about one third of the urban and 79 Ibid. 80 For example: “Hindu Raj at Ajmer”, Muslim Outlook, August 16th 1925 81 NAI, Jaipur Residency, Part II, 1939, #315 82 NAI, Foreign & Political Department, Establishment Branch, 1925, #509-E 83 Ibid.

173 one fifth of the rural population of Ajmer-Merwara it would appear from these statistics that the need for special measures to safeguard their interests in this respect has not been clearly established. The local Administration are not unmindful of the matter...”84

At the same time, however, Bray also admonished those within the Ajmer-

Merwara government to understand the “undesireability (sic) of any one community [holding] an unduly disproportionate share of posts in Govt. service.”85 This was, as I highlighted in Chapter 1, the British policy of religious neutrality – holding the scales even between both communities.

The British also made other efforts to push for the betterment of Ajmeri

Muslims. One of these initiatives dealt with education policy. Table 3.6 below details the percentage of Hindu and Muslim children (males and females) enrolled in schools in 1933-4 in Ajmer-Merwara.86 It shows that a higher percentage of Muslim boys were enrolled in school than Hindu boys and that the percentage of girls from both religions was comparable. This latter statistic is surprising because of the general cultural reluctance at the time toward enrolling

Muslim girls in school.

Table 3.6: Hindu and Muslim School Enrollment in Ajmer, 1933-4 Hindu Hindu Muslim Muslim Males Females Males Females Primary School 5.02% 1.10% 6.87% 1.05% Attendance

Overall then, the British in Ajmer-Merwara pursued generally the same beneficial policies for Muslims as they did throughout the rest of India.

84 NAI, Foreign & Political Department, Establishment Branch, 1925, #509-E 85 NAI, Foreign Department, Establishment Branch, 1926, #1-4 86 The denominator in the official government report was the total population so the percentages were in fact probably much higher than recorded. From: NAI, Home Department, Public Branch, 1935, #31/38/35

174 Government documents from the time period, highlighted by the remarks of

Denys Bray, show that officers were concerned about protecting the Muslim minority against the Hindu majority. Muslims, even well into the 20th century, provided the backbone of the Ajmer-Merwara government and were afforded a favorable position in society that was incommensurate with their population strength. Their situation in Ajmer-Merwara was on the whole superior to their treatment in the Jaipur State.

4.4 Post-Colonial Religious Reforms

In the post-colonial period, a variety of government organizations have undertaken to reform the position of Muslims in Rajasthan. At the national level, these reforms aimed at the Muslim community have largely been a failure – I noted in Chapter 1 the results of the Sachar Committee inquiry of 2006. I argued that the failure of reform has not eliminated the British-princely distinction in

India; therefore, reforms to the position of Muslims in former princely states have not improved their status. The Jaipur-Ajmer comparison only lends credence to this argument.

First, it is important to note that the main administrative organizations which are responsible for improving the situation of Muslims in Jaipur are poorly-staffed, poorly-funded, and largely ineffective. One interviewee at the

RSMC noted that there was currently no Chairman and no members on the administrative board as they were still waiting for (stalled) political appointments.87 Or consider the Rajasthan Minority Affairs Department

87 Interview in Jaipur, November 11th 2010

175 (RMAD), which oversees the distribution of scholarships to minority religious groups. From 2009-10 they were stripped of over 10,000 scholarships from their fixed allotment.88 Second, as I have documented, Muslims in Jaipur have not seen much change to their fortunes from the princely period; they work the same jobs and live in the same neighborhoods. On the other hand, Muslims in Ajmer have – despite the exodus of some Muslims to Pakistan after Partition – remained in a better position than their Jaipuri counterparts. The distinction between princely and British rule still therefore resonates in both areas.

V. The Colonial Roots of Caste and Tribal Violence in Rajasthan

In this section I consider the underlying causes for variations in caste and tribal conflict in both Jaipur and Ajmer. I highlight that while the Jaipur State pursued protectionist policies for low castes and tribals, British administrators in

Ajmer pursued a number of discriminatory policies that severely undermined the position of these groups. I argue that this is the key explanatory factor for why

Ajmer today suffers more caste violence than Jaipur.

Unfortunately there is no comprehensive account or historical dataset of caste and tribal conflict in India during the colonial period. This has been a major oversight of the existing literature on ethnic violence in India, as more weight has been placed by scholars on communalism. While I cannot present figures on the number of caste riots or tribal revolts in Jaipur and Ajmer during the colonial period, I can discuss and expound upon the disparate policies enacted in both territories and the comparative effects they had on low caste and tribal groups.

88 Interview in Jaipur, November 3rd 2010

176 The basic comparative caste compositions of Jaipur and Ajmer in the early

20th century is displayed in Table 3.7 below. Brahmins were the numerically superior caste in Jaipur and over 53% of the population made their living from the land. In Ajmer-Merwara, however, the Mahajans (a mercantile caste) were the largest group; approximately 55% of the people there made their living from the land. The dominant castes – Brahmins and Rajputs – were not involved in agriculture. It was the low castes and, to a lesser extent, tribal groups that dominated the countryside.

Table 3.7 – Composition of Castes in Jaipur and Ajmer, 1901 Jaipur State Ajmer-Merwara Brahmins – 349,000 Mahajans – 37,027 Jats – 265,000 Gujars – 36,278 Chamars – 218,000 Rawats – 32,362 Mahajans – 212,000 Shaiks – 31,972 Gujars – 184,000 Jats – 27,952 Rajputs – 124,000 Brahmins – 25,896 Malis – 116,000 Rajputs – 15,430 Source: Imperial Gazetteer of India, (Vol. 5): 145 (Ajmer-Merwara) and (Vol. 13): 389 (Jaipur)

As for the adivasis, the main tribes in each district were Bhils and .

There were a number of smaller tribes as well. Neither Jaipur nor Ajmer were heavily forested areas – according to the 2001 census only 4.53% and 4.19% of the respective land was covered by forest. Colonial figures were not much higher.

Most of the tribes lived either in these small forested tracts or in hill regions.

Before considering the roots of caste and tribal violence it is imperative to understand the nature of landholding in Rajputana.89 Land has always been at the heart of caste and tribal conflicts throughout India. When analyzing religious

89 For greater detail see Stern 1988: 2-3, and Sisson 1971: 24-28.

177 violence, the communal representation of minority groups in government is paramount because Muslims were the previous administrators of India; most of them lived in urban areas, many worked in government service, and the strength of the community at large flowed from these factors. Caste and tribal conflicts, on the other hand, are largely rural, and agricultural policies therefore determined the position and condition of different caste groups.

There were two main types of land in Jaipur and Ajmer: khalsa land and jagirdari land. Khalsa land (also known as crown land) was directly-controlled by the government. In princely areas, for example, the maharaja owned these lands and received revenue directly from the cultivators who worked on it. On khalsa land cultivators had no proprietary rights, but they were still “...as a rule, treated as hereditary, being seldom interfered with so long as they pa[id] the revenue”.90

Jagirdari land, on the other hand, was controlled by a hereditary noble and the darbar did not exercise strict rights over these areas, although they did collect revenue from them. The term jagir refers to a land grant. Jagir land in

Rajputana had been granted historically for some kind of service to the raja

(often military service), and the rights to the land typically expired upon the death of the grantee. However, over time jagirs evolved into hereditary grants as new claimants (such as the son of a nobleman) simply paid a fee to the darbar

(nazarana) and thereby retained control of the territory. The term jagirdar refers to Rajput noblemen; in their estates they were “assignees of the rights of

90 The Rajputana Gazetteer (1879): 141

178 the State, [but had] the same rights as the State itself”.91 Jagirdari lands could be extremely small parcels or massive estates – the larger ones were often known as thikanas, and they could have their own courts, tax collectors, and police forces.

Powerful jagirdars could even assign smaller jagirs within their kingdoms. The jagirdars were known by a wide variety of names: sardar, thakur, thikanedar, and rao raja.

Finally, in Rajputana there were also small parcels of land known as bhums, which were free land allotments. They did not require any kind of tax payment and were granted for life. They were given both by the state and noblemen. While this discussion outlines the general contours of the Rajput landholding model, in reality every state in Rajputana was different.92

5.2 British Policies on Low Castes and Tribals in Ajmer-Merwara

It is difficult to know for certain what exactly the system of landholding was in Ajmer-Merwara prior to British arrival. The British administrators who governed the region and surveyed existing land records were of the belief that khalsa lands were governed under a ryotwari system in which the tribute was not collected by landlords, but directly from the cultivators. Recall that this is the tenurial model which produces the best agricultural yields today and also produces the least amount of violence (Banerjee and Iyer 2005). Yet once Ajmer came under the direct control of British administrators in 1818 they began to disrupt the existing landholding system in a number of different ways (Gold and

91 Ibid.: 23 92 For example, in Jaipur there were also a small number of zamindari or biswedari areas.

179 Gujar 1997).93 The results were disastrous to low caste groups and produced a high level of rural unrest and antagonism between castes. It is important to note that the British were often, as Marx's quote in Chapter 1 indicated, experimenting with different land policies. If certain policies did not work they were abandoned and replaced. For example, Ajmer-Merwara was originally opened up to the unfettered cultivation and sale of opium, but once the district became rife with smuggling this policy was quickly abandoned (IDS Part 3: 41).

In 1818 one of the most important tasks of the first Chief Commissioner of

Ajmer was calculating and instituting a new land revenue policy. Commissioner

Wilder determined the land revenue as one-half of the value of all crops produced in Ajmer-Merwara. This amount was calculated after Wilder had commented that the Ajmer population before British rule had been “sadly thinned by oppression” (Gold and Gujar 2002: 67). This new figure mirrored the crushing revenue demands made in direct-rule districts throughout the rest of

British India. Later British administrators noted about Wilder's policy:

“[He] had made the mistake of over-estimating the resources of the District, and the baneful effects of this error extended many years. This, added to several years of distress...reduced the District to a state of abject poverty.”94

The last tribute collected by the Marathas in 1818 was totaled at 128,978 rupees;

Wilder increased it within two years to 179,457 rupees, although it was subsequently lowered after widespread protest.95 Later administrators noted the results on low caste agriculturalists: “cultivators and landless labourers...are

93 Once the British came to control Ajmer, khalsa lands became British property. I detail later how they dealt with jagirdari areas. 94 The Rajputana Gazetteer (1879): 161 95 Ibid.: 90

180 generally in debt, and the latter live hand to mouth.”96

Aside from increased revenue demands, the British did not otherwise alter the system of khalsa lands until 1849 when a mouzawar system was established in which proprietary rights were given over to village bodies. In essence this new policy:

“...effected a radical change in the [khalsa] tenure. It transformed the cultivating communities of the khalsa – each member of which possessed certain rights in improved land, but who as a community possessed no rights at all – into...proprietary bodies.”97

Yet this new system was, as administrators bemoaned, “hardly understood, and

[was] not appreciated by the people.”98 Anna Grodzins Gold and Bhoju Ram

Gujar note that during the early days of colonial rule in Ajmer, the British were focused on “above all tax assessment and revenue collection” (2002: 67). This was the chief concern of British administrators, and they disrupted the existing landholding system in attempting to maximize revenues. Along this line, another important change with the onset of colonialism was an increase in the number of khalsa villages. They almost doubled during the early years of British rule in the district, increasing the number of revenue-bearing villages under the control of the government. While many of the early caste conflicts over land in India began in Bengal Presidency, the British in Ajmer-Merwara noted that they actually had comparatively stronger proprietary rights over the land in this district.99

The result of these policies was the complete economic deterioration of the

96 Imperial Gazetteer of India (Vol. 5): 153 97 Imperial Gazetteer of India (Vol. 5): 153 98 Ibid. Baden-Powell notes that “This course [introduction of the mouzawar system] was afterwards much regretted” (1892: 332). 99 The Rajputana Gazetteer (1879): 31

181 peasantry. In 1926 Commissioner of Ajmer-Merwara Lieutenant Colonel R.J.W.

Heale traveled to an 'average village' in the district to conduct an inquiry on the state of cultivators. He noted:

“The village is typical to the extent that it suffers from the difficulties and presents the problems that are common to the whole District, but it is more backward than very many villages...The difficulties may be summed up in three words; ignorance, poverty and lack of water.”100

Mukat Behari Lal Bhargava, member of the INC and future Lok Sabha member from Ajmer constituency, also summarized the state of peasants in Ajmer in

1938:

“The condition of the peasantry in the province in general and in the Istamrari101 area in particular is indeed miserable. In the latter the system of Begar102 is in vogue and the tenants have to pay a number of 'lags' (cesses) besides the usual share of the produce, and are liable to be ejected at the whim and caprice of the Istamradar though they may have been cultivating the same plots of land for generations...The lot of the five and half lacs of people of this important but unfortunate province is no better than the hewers of wood and drawers of water and they drag on their miserable existence as mere dumb driven cattle...”103

This particular quote highlights another important facet of the colonial land system in Ajmer-Merwara: the increasing power of zamindars. By 1901, for example, 1,693,728 acres of land were cultivated under the zamindari system whereas no areas were under ryotwari tenure (Keltie 1903: 152). Ajmer-Merwara was thus similar to other British areas such as Bengal, Oudh, the Northwest

Provinces, and Punjab. In Ajmer the chiefs of non-khalsa land became independent property holders and were called istimradars. The istimradars claimed that they had always been private property holders – a dubious claim –

100 BL, Q/12/9, 1926 101 This term refers to a kind of zamindari tenure. 102 Begar refers to bonded (forced) labor. 103 Quoted in Gold and Gujar 2002: 77.

182 but the British accepted it nonetheless. Colonial administrators in the district generally practiced non-intervention in zamindari areas. Cultivators in these areas were therefore subjected to the caprices of the istimradars – peasants were mere tenants-at-will and could be thrown off their land for any reason.104 British rule in Ajmer then reinforced and strengthened the power of landlords.

Finally, Ajmer-Merwara under British rule also became prone to famine.

For example, in the nineteenth century there were episodes of food scarcity (of varying degrees) in 1819, 1824, 1832-3, 1848, 1868-9, 1891-2 and 1899.105 Part of the problem was that the British failed to make long-term investments in the land

– for example, irrigation – in Ajmer, and famines therefore hit the district especially hard.

Outside the realm of landholding additional aspects of British rule in

Ajmer-Merwara also harmed the position of low caste groups. Most importantly,

R.C. Bramley, a British Superintendent, was sent to the district in preparation for the 1901 Census. This was the first census in British India which tried to rank caste groups. Bramley, along with Captain A.D. Bannerman, may have been the first Englishman to use the term 'untouchable' (Charsley 1996: 2). In his report he painstakingly noted all the castes in the district which (according to his estimation) constituted the untouchables, such as Balais, Chamars, and Regars among others (Ibid.). The very process of determining untouchability promoted long-term caste antagonisms, and the stigma associated with this new untouchable designation was difficult to erase. While census operators were also

104 The Rajputana Gazetteer (1879): 28 105 Imperial Gazetteer of India (Vol. 5): 393

183 sent into princely states like Jaipur, as I laid out in the Introduction, the rulers of the native states emphasized religious identities and subverted caste to a secondary position of importance in their territories.

British policies on adivasis were no better. The tribes of Ajmer-Merwara resided primarily in hilly regions or small forested tracts in the district. The

British, like they did throughout provincial India, commandeered these areas as property of the state when they came into power in 1818, arguing that they had the right of “management of any tract of waste or hilly land, the proprietary right...being invested absolutely in Government as long as the land is required for forest purposes.”106 The result was the alienation of lands which tribal groups had used unfettered by any state interference for hundreds of years. Although there is no information about Maratha forest policy, as I noted in Chapter 1, it is generally established that the British were the first rulers of India to interfere significantly with tribal forest rights. They were led by their belief that without conservation policies the forests would be destroyed by villagers, who, British administrators wrote, “wanted the free run of the forests to help themselves as they pleased with no thought to the future” (Gold and Gujar 2002: 73). However, there is little evidence that Ajmer prior to British arrival suffered from excessive deforestation.

The British also established Criminal Tribes legislation in Ajmer, and members of the large tribe were designated as criminal. British administrators noted that there were actually two 'classes' of Meenas: “...the zamindari or agriculturalists, and the chaukidari or watchmen; the former are

106 The Rajputana Gazetteer (1879): 32

184 industrious and well-behaved, while the latter were, and to some extent still are, famous as marauders. As noticed later on, it is not always easy to distinguish one class from the other...”107 As often happened due to the Criminal Tribes legislation, members of a tribe who were not actually criminals were stigmatized and suffered discrimination. In 1933, for example, hundreds of Meenas, Sansies,

Baories and Chakras were convicted under the Criminal Tribes legislation in the province.108 Policies against the Mers tribe were especially brutal. British official

B.H. Baden-Powell noted that the Mers were “rude tribesmen...[who] were reclaimed under British rule” (1892: 319109) – that is, they were designated as a

'predatory tribe' and in 1824 the British initiated military operations against them.110

5.3 Jaipur State Policies on Low Castes and Tribals

In the Jaipur State it could hardly be argued that the situation of low caste groups was good – their condition was not good anywhere in India. Yet the comparison here is with British-controlled Ajmer-Merwara, and – far from its reputation as a 'feudal' state – Jaipur pursued a variety of beneficial policies for low caste groups. First, consider how the state of landholding in Jaipur differed from Ajmer. The darbar probably controlled only about two-fifths of the land

(khalsa) while the rest was “alienated in grants to nobles, ministers, priests or courtiers”.111 As I will demonstrate, especially in the next chapter, this percentage

107 NAI, Jaipur Residency, Jaipur Agency, 1947, #6, pp. 9-10 108 NAI, Home Department, Public Branch, 1935, #31/38/35 109 Emphasis added. 110 BL, IOR/F/4/810/21725, 1824 111 NAI, Jaipur Residency, Jaipur Agency, 1947, #6

185 of land controlled by the darbar in Jaipur was extremely low in comparison to most other princely states.

In the khalsa areas, however, the Jaipur government had a markedly different set of policies than those which the British instituted in Ajmer's khalsa lands. Beginning in the early 20th century, Jaipur State “followed a strategy similar to that in Baroda to achieve peasant support and political stability. [In the

1920s and 1930s]...a ryotwari settlement began to bring peasants into direct relations with the central administration” (Ramusack 2004: 176). Under this system the land revenue was taken directly from the ryot – the cultivator of the land. Jaipur's ryotwari system cut out middlemen (an endless source of corruption in British India) and produced more security in land tenures. In 1939 the Jaipur State went further and abolished a large number of land taxes which had previously been levied only on low castes.112 In contrast to the excessive land tributes demanded by the British in Ajmer, the darbar in Jaipur generally requested less; tributes could be as low as one-sixth of the crop, although the autumn demand could be much higher.113

The benevolent nature of the darbar was also reflected in the difference between land tenure rights in khalsa as opposed to jagirdari land. On khalsa land, the economic condition of the peasantry was superior. Richard Sisson notes on Jaipur:

“...primarily in khalsa areas, the position of tenants was neither so tenuous nor their economic obligations so severe. Frequently tenants enjoyed katedari and pattedari rights which provided permanent rights of occupancy and

112 NAI, Jaipur Residency – Part 3, Confidential, 1939, 1442 113 Ibid.:142

186 inheritance if certain rules of tenure were met and in some cases provided for the right of transfer and the assumption of mortgage.” (1971: 27)

Lloyd and Susanne Rudolph add:

“In the judgment of an inquiry committee of high civil servants from British India that reported on tenant rights at the time of independence, 'the tenant [of khalsa] has rights defined and secured, and there is, on the whole, little cause for him to complain.'” (1984: 29)

On jagirdari areas over which the darbar had limited control, however, the position of tenants was much less secure and more prone to strife. The

Rudolphs note that even by 1947 only 32% of all jagir villages in Rajasthan

(16,780 total) had settled revenue obligations, meaning the rules which governed taxes and revenue were not codified and therefore tended in practice to be capricious. Jaipur State tried to push for settlement in jagir lands in the late

1930s and 1940s, although with mixed success. The main point of contention was that nobles in Jaipur did not believe themselves to be subservient to the darbar, but rather considered themselves independent rulers of hereditary estates.114 This enabled them to govern their territories in any fashion they chose, which usually resulted in feudal tenures. In fact, most associations or organizations that emerged to fight for peasant rights in the princely states targeted not princes but jagirdars. Ramusack puts it well in writing: “Although these princes were hardly utopian reformers and peasants on khalsa lands still had significant grievances, the conditions of peasants in jagirdari areas could be much worse” (Ramusack

114 On this see Ramusack 2004: 171 – “Views on the origins of the jagirs differed. Rulers asserted that they granted jagirs, retained control over succession to them, and had their superior position acknowledged by the payment of tribute from jagirdars. Jagirdars often challenged this interpretation. In the case of in Rajasthan, Denis Vidal has documented how nobles argued that their ancestors had obtained their lands when the kingdom was in the process of formation and thus had autonomous rights in the land.”

187 2004: 228). This situation in Jaipur, I will argue, mirrored the conditions of the

Telangana region of Hyderabad (See Chapter 5). In that area, where the ruler of the state (nizam) had limited control, land tenure rights for peasants were extremely backward and repressive due to an independent class of landlords which dominated and brutalized the region. But the princes of India generally enforced more benevolent policies in their lands than their nobles.

The differences between peasants on khalsa and non-khalsa land is well illustrated by the political history of Sikar, the largest thikana in the Jaipur State.

Sikar was the site of peasant (Jat) uprisings in 1935 when hundreds of cultivators refused to pay excessive revenue demands and protested their feudal conditions, which then resulted in a violent backlash by Sikar officials (Rudolph and Rudolph

1966: 158). At this time thousands of Jats traveled from Sikar to Jaipur City in order to speak to the maharaja, and Stern notes that “[O]ne of the sabha's

[group's] leaders, Harlal Singh, remembers a courteous reception at the palace and a sympathetic hearing of the Jat's complaints” (1988: 282). Yet the raja of

Jaipur had limited ability to interfere in the Sikar thikana. In fact, in response to attempts to meddle in 1938 the rao raja of Sikar decided to launch an armed rebellion against the Jaipur maharaja (Rubin 1983). During this struggle, the

Jats sided with the maharaja against the rao raja. The Jats understood that their situation in Sikar – uncertain land tenures, excessive revenue demands, harassment by petty officials – was appreciably worse than that of the peasants of the khalsa lands of Jaipur. Sisson notes that on the whole Jat movements in the princely states were “most intense in areas ruled by jagirdars...rather than in

188 those under the direct rule of the royal house” (1969: 947). The Sikar Rebellion was eventually put down by the Jaipur raja with the assistance of the British, and the thikana was finally subjected to a set of serious land reforms.

The peasants of Jaipur fortuitously benefited from this endless infighting between the raja and his nobles. It was in fact partly because of an attempt to weaken the nobles that the rajas of Jaipur in the 20th century sought improvements in the lives of the peasantry. For example, when Mirza initiated constitutional reforms in the 1940s:

“The Praja Mandalists on the constitutional reform committee were so successful in reducing the general franchise qualifications of wealth and/or education for Jaipur's untouchable castes as to enfranchise a positively disproportionate number of untouchables.” (Stern 1988: 317)

These specific reforms were ultimately blocked by high castes, but the darbar succeeded in passing legislation in the late 1930s which tried to eliminate the more feudal aspects of jagirdari tenure (Kumar 2001: 265). By aligning themselves with the peasantry the rulers of Jaipur were able to stunt the autonomy of the large landholding jagirdars in the state.

Finally, Jaipur State was, in contrast to the British in Ajmer, also very proactive in response to famines.115 During the 1868-9 famine the darbar distributed large amounts of cooked food from the capital and started a massive relief works program. Low caste groups were also specifically targeted; they were allowed to graze freely in the forests and acquire enough produce to prevent starvation.

In short, then, the position of the low castes in Jaipur was superior to that

115 Imperial Gazetteer of India (Vol. 13): 393

189 of Ajmer-Merwara. The darbar, even if it was slow to act and limited in its reforms, pursued policies that were more progressive than those which occurred in British-controlled Ajmer. There was more security in land tenure, better government responses to famines, and, over time, the darbar connected more closely with the peasantry in land settlements. Furthermore, the darbar did its best to curb the most feudal aspects of jagirdar rule, pushing for peasant land rights in those areas as well.

Jaipur State, like a surprising number of princely territories, also implemented protective policies toward tribal groups. I have argued that the main reason for this was because princely rulers considered the tribes of the area to have been the original inhabitants of the land. For example, in the states of

Dungarpur, Banswara, and Pratapgarh, the Bhil tribe were considered the earliest occupants of the land (Varma 1978). The rulers of Jaipur were no different. One account of the history of the state argues that “Meenas...held a good deal of this part of the country in the twelfth century”116 before the Rajputs came to control it.

Because of this belief, many tribes were treated with a large amount of respect.

When a new raja was placed on the throne in Jaipur State it was a member of the

Meena tribe who placed the ceremonial rajtilaka117 upon the king. Meenas were also reserved an official allotment of government jobs: specifically, they guarded the royal treasury and were hunters employed by the state (Ramusack 2004:

210). In 1939 Jaipur created a new job training program for members of the Dais

116 Imperial Gazetteer of India (Vol. 13): 393 117 A ceremonial mark, usually made of a red powder, placed on the forehead.

190 tribe.118 Their situation, and that of other tribes in Jaipur, was better than that of neighboring Ajmer.

The British, for their part, were not happy about Jaipur's protective treatment of its tribal population. They tried to enforce the same Criminal Tribes legislation which existed in Ajmer-Merwara in Jaipur also, though they met with little success. For instance, in 1930 the British Resident wrote to complain to the

Jaipur government that “If there are really members of Criminal Tribes – Minas and Baories – [there] they should be dealt with according to the Criminal Tribes

Rules...”. The Jaipur government brushed them back:

“...we have the honor to state that these Minas & Baories are living here since the [state] was founded & are only cultivators. There is also no complaint against them. We, therefore, recommend that these Minas & Baories...may not be declared as Criminal Tribes.”119

When the British discussed the Criminal Tribes legislation at a conference in

Ajmer in 1923 Inspector of Police D.R. Wright complained that “I see no officer from Jaipur was present at this conference but the Darbar had been informed of it and had intimated that one would attend.”120 Attempts by the British to limit the movement of the Meenas in Jaipur also failed – the Jaipur State refused to implement these policies, much to the chagrin of British administrators.121

Therefore, there were clear differences in the treatment of low caste and tribal groups in Jaipur and Ajmer. British policies implemented in Ajmer-

Merwara were devastating to these groups; low castes and tribals were subjected

118 NAI, Jaipur Residency – Part 3, Confidential, 1939, #1442 119 NAI, Jaipur Agency Part II, Jaipur Residency, 1930, #253 120 NAI, Jaipur Agency Part II, Jaipur Residency, 1930, #349 121 Ibid.

191 to crushing new revenue demands, alienated from their land, and suffered an increase in the power of local landlords. New census operations also produced antagonism between castes, who were officially ranked for the first time. For their part, tribal groups were cut off from their forests and many were criminalized. In

Jaipur, on the other hand, the darbar pushed for several beneficial policies for low caste groups, even trying to improve their condition in jagirdari areas.

Adivasis were also protected; they formed an important part of the princely administration, were allowed to remain in their forests, and were shielded from

British attempts to criminalize them.

5.4 Post-Colonial Caste and Tribal Reforms

In the post-independence period, Rajasthan like most states tried different policies to improve the condition of marginalized castes and tribes. The caste system was technically 'abolished'. Rajasthan also embarked upon an ambitious set of land reforms. In 1949 the central government appointed a Jagir Enquiry

Commission which recommended the abolition of jagirs. In 1952 the major piece of land reform legislation was passed: The Rajasthan Resumption of Jagirs Act.122

This act “resumed” (commandeered) all jagirs, provided compensation for jagirdars, created khudkasht (personal cultivation) allotments, created a system of land revenue for all holdings, and – most importantly – created a fixed system of tenurial rights for cultivators. The legality of the legislation was immediately challenged by the jagirdars, and although they tied up the courts for years they were ultimately unsuccessful. Initially, small Rajput landholders (bhumiyas)

122 For more on this see Sharma 1993: 13-16

192 were excluded from the legislation, but they were eventually also stripped of their land by the Congress Party. The Rajasthan Resumption of Jagirs Act, as well as later legislation that banned Zamindari and Biswedari tenures, were all implemented at a state-wide level. Ajmer came under the ambit of these reforms when the Government of Rajasthan passed the Ajmer Abolition of Intermediaries and Land Reforms Act in 1955.

However, the implementation of land reforms has been variable across

Rajasthan and not every district has had the same level of success (Rosin 1978).

In Jaipur, post-independence changes to the land tenure system were a continuation of pre-existing reforms which had been instituted by the darbar in the early 20th century. Low caste cultivators and tribal groups had already been the focus of efforts to reform their land ownership and economic situation. In

Ajmer, on the other hand, land reforms were more difficult to implement because

British policies had already strengthened the position of landlords, and had also stigmatized tribal groups. In Ajmer, as Thomas Rosin writes:

“Putting a ceiling upon the size of land holdings to shift land to the landless was easily subverted and not applicable to the myriad of smaller holdings. The small land titleholders here seem to have little sympathy for the folks of lower castes whose labor they need to augment their own.” (2010: 7-8)

Land reform efforts in the district, therefore, have largely been stalled.

The backward position of low castes and tribals in Ajmer today, then, is a legacy of discriminatory policies first instituted during the colonial period. These policies increased land inequality, generated tension among caste groups, and ultimately led to increased violence. In Jaipur, however, the position of low castes

193 and tribals was protected, and this has led to a minimization of caste and tribal violence in that district today.

VI. Conclusion

This chapter has examined patterns of ethnic conflict in contemporary

Rajasthan through a controlled historical comparison of Jaipur, a former princely state, and Ajmer, a former province of British India. First, I highlighted that the pattern of ethnic conflict which I detailed in Chapter 2 through the use of statistical analysis was clearly evident in these two case studies: Jaipur has higher levels of contemporary religious conflict, but Ajmer has higher levels of caste and tribal violence. Using qualitative interviews in both regions, I showcased that in contemporary Jaipur the socioeconomic position of Muslims lags significantly behind that of Hindus, and that Hindu-Muslim antagonism in the district is high.

In Ajmer, on the other hand, Muslims were on the whole more sanguine about their situation and spoke to a larger sense of equality between Hindus and

Muslims. The result was the opposite for low castes and tribals: in Jaipur, caste did not form the main axis of conflict, and low castes and adivasis were much further ahead of their Ajmeri counterparts. Within Ajmer, however, caste tensions and antagonisms formed the major problem in local politics.

I then turned to the historical roots of this variation in conflict outcomes.

In Jaipur I argued that, despite claims by numerous historians of princely syncretism, the darbar enforced a number of policies that discriminated against the Muslims of the state. In Ajmer-Merwara province, on the other hand,

Muslims enjoyed a vastly disproportionate amount of power in the state

194 apparatus and in society. I subsequently examined the roots of caste and tribal conflict. In this realm the situation was reversed. New British policies instituted in Ajmer-Merwara – such as increased land taxation and Criminal Tribes legislation – led to the degradation of low caste and tribal groups. In Jaipur, on the other hand, the darbar pursued beneficial policies for low caste groups and were protective of the rights of tribals in the area. Overall, I found strong support for the historical theory which I laid out in Chapter 1: British administrators and the native rulers of Jaipur pursued vastly different ethnic policies, and the legacies of these policies – largely unreformed in the post-colonial context – created the patterns of ethnic violence in Rajasthan today. Bifurcated colonial rule ultimately embedded different narratives of violence in Jaipur and Ajmer.

The next chapter investigates the second pair of case studies in this dissertation. I examine another controlled historical comparison from the southern state of Kerala: I compare the British-controlled north (Malabar) and the princely south (Travancore). I examine these two cases to determine if the same pattern of ethnic violence existent in Rajasthan is also present in Kerala, and whether bifurcated colonial rule produced variations in contemporary violence outcomes.

195 Chapter 4 Ethnic Violence in Kerala: Malabar and Travancore

“[The Travancore population is] unhappily stricken with the plague of communal animosities...” – Donald Field, British Resident to Travancore, 19351

“...the Moplahs of E. Malabar...are most unfortunate subjects for fanaticism; and more especially so when the land tenures are such as if arranged specially for the purpose of making people discontented.” – F. Fawcett, Superintendent of Police, Malabar District, 18972

The previous chapter examined the first of three pairs of case studies which form the empirical bulk of this dissertation. In Chapter 3 I analyzed patterns of ethnic conflict in Jaipur and Ajmer districts in Rajasthan. I found that while Jaipur experiences a large amount of contemporary religious conflict, ethnic violence in Ajmer revolves around castes and tribes. I then located the source of this divergence in the colonial period. Jaipur was a princely state, and its Hindu rulers discriminated against the Muslim minority but were progressive in their policies toward low castes and adivasis. In contrast, Ajmer was a British province, and colonial administrators there enforced discriminatory policies toward low castes and tribal groups but were protective of the Muslim minority.

The different historical experiences of these neighboring districts therefore accounted for a striking variation in contemporary ethnic violence patterns.

This chapter considers a second pair of case studies, this time from the south Indian state of Kerala. While the previous chapter compared two districts,

1 Quoted in Kooiman 1995: 2130 2 Fawcett 1897: 296

196 this chapter analyzes two regions of the same state. I compare the British-ruled

Malabar region of northern Kerala and the former Travancore princely state of southern Kerala.3 Because this is a state-level comparison, a number of state level factors can be taken into consideration in the analysis. Furthermore, by choosing a state from south India, which has a completely different history and culture from the north, I can potentially increase the validity of my main hypothesis if patterns of ethnic violence across Kerala match violence outcomes in Rajasthan.

On face, Kerala may seem an unlikely case for the study of ethnic violence patterns because the state is known throughout India for its caste politics, stemming from its historically elaborate and oppressive caste system. For example, the 1931 census recorded over 500 unique castes, and there were strict, ritual restrictions governing intercaste contact. The Imperial Gazetteer of India

(1909) noted for instance that in Malabar “a high-caste man returning from his bath shouts out to warn others of his approach, so that they may step aside into the fields and not pollute him”.4 Yet, as this chapter will demonstrate, Kerala has also experienced its share of communal conflict, and – most importantly – there is a clear distinction in the pattern of violence between the north and south.

I first discuss the history of Malabar and Travancore, and why it constitutes a controlled historical comparison. I highlight, along these lines, that the British also wanted to annex Travancore, but for various contingent reasons

3 This chapter is a partial attempt to correct the problem noted by Margret Frenz, who has written that a “comparison between Tiruvitamkur [Travancore] and Malabar in the early modern period is still a desideratum” (2003: 5). 4 Imperial Gazetteer of India (Vol. 70): 60

197 were not able to do so. In sections II and III I then present evidence on post- colonial ethnic violence patterns. I highlight that the Malabar region experiences higher levels of caste and tribal violence today while Travancore's main contemporary ethnic conflict revolves around religious identities. As I did in

Rajasthan, I also conducted a number of interviews in both regions to examine the socioeconomic state of ethnic minority groups, as well as relations between them. Then, beginning in section IV, I show that this contemporary pattern of ethnic conflict is rooted in the varying colonial experiences of both territories.

British administrators in Malabar enforced a number of discriminatory policies toward low castes and adivasis. Their position on religious minorities was more complicated, which I discuss in sections 4.3 and 4.4. The Hindu rulers of

Travancore, on the other hand, discriminated against non-coreligionists (both

Muslims and Christians), but protected low castes and adivasis. These disparate legacies, I argue, have not been reformed in the post-colonial context. Finally, I conclude by taking stock of the similarities evidenced in the Malabar-Travancore comparison with the previous Jaipur-Ajmer comparison from Rajasthan.

I. The Case Studies

Kerala is a small coastal state in southwestern India, bordered on the north by Kannada and separated on the east from Tamil Nadu by the Western

Ghats mountain range. The state lays on the extreme southern portion of what is known as the . It is, along with Karnataka, Andhra Pradesh, and

Tamil Nadu, one of the four states that comprise south India. The local language

198 is Malayalam and the inhabitants of the state are known as Malayalis. Maps 4 and

5 show the location of Malabar and Travancore in the colonial period, and the present state and its districts, respectively. Table 4.1 provides basic comparative figures on both regions from the colonial period, and Table 4.2 provides similar data for the contemporary period. Both regions have similar overall caste population dynamics, but the figures on religious groups is a bit more complicated. Both regions have Hindu majorities, but Muslims are much larger of a minority in Malabar whereas Christians are much larger of a minority in

Travancore. Overall, however, the constellation of ethnic groups in both areas is very similar. The Malabar region today consists of the districts of Kannur,

Kasargod, , Malapurram, , and Wayanad. Travancore is comprised of Alappuzha, Idukki, Kollam, Kottayam, Pathanamthitta, and

Thiruvananthapuram districts.

199 Maps 4 and 5: Colonial Kerala; Post-Colonial Kerala5

5 On the west coast of Map 4, the province of Malabar is pink and the princely state of Travancore (to the south) is yellow. The other large yellow area (to the north of Malabar) is the princely state of Mysore.

200 Table 4.1: Malabar and Travancore Comparison, 19316 Malabar Travancore Area (sq. miles) 5,800 7,660 Total Population 3,534,000 5,096,000 Caste Hindus 20% 21% 'Polluting' Hindus 48% 40% Muslims 30% 7% Christians 2% 32%

Table 4.2: Malabar and Travancore Comparison, 20017 Malabar Travancore Total Population 13,515,737 14,222,184 Hindus 52.58% 53.22% Muslims 41.10% 9.61% Christians 6.20% 23.16% SCSTs 10.25% 10.33%

In the pre-colonial period Kerala was a major international center of commerce, especially in the highly lucrative . Even before Vasco De

Gama famously arrived in Kozhikode (Calicut) in 1498 the area had been frequented by Greek, Roman, Jewish, and Arab traders for centuries. Their immigration to the area and intermingling with the local population led to enormous cultural diversity in the region. There were many ancient which ruled over Kerala in the pre-colonial period: the Cheras, Cholas, and

Pandyas all controlled parts of the region at various times. Kerala also always had a close relationship with neighboring Tamil Nadu, and Malayalam itself 6 Adapted from Jeffrey 1978: 30 7 All figures taken from the 2001 Census.

201 probably emerged some time in the 6th century as a variant of the Tamil language.

The Portuguese were the first European colonial power to arrive in Kerala, establishing trading outposts there during the 16th century. After them, the Dutch became the dominant power, and the Dutch East India Company invested a lot of effort to establish trading supremacy in the region. However, the rising

Travancore State upended these plans, defeating the Dutch during the

Travancore-Dutch War (1739-41). By the end of the 18th century the Dutch were eventually completely eclipsed by the British East India Company. Much of this

Western competition for influence in India was intertwined with the Napoleonic

Wars simultaneously raging in Europe.8

The northern portion of Kerala became known as Malabar. The of Calicut was the major local power prior to British rule, but there were a number of other small rajas and petty chiefs. In the 1760s Malabar came under the control of Hyder Ali, the Muslim ruler of the neighboring Kingdom of Mysore, who invaded the region and crushed the Zamorin. Many of the Hindu rajas in

Malabar fled into the surrounding jungles and others took asylum in Travancore.

Mysore then subsequently engaged in four wars with the British (Anglo-Mysore

Wars), and, after the final war in 1799 the region came under the direct control of the British. It was eventually named the Malabar District and incorporated into the Madras Presidency.

8 There were also some French settlements in south India, notably at Mahé on the Malabar Coast and Pondicherry. The French retained some of these territories as colonies even after the British established their supremacy over the subcontinent. In fact, some French possessions did not officially merge with the Indian Union until the 1950s.

202 The southern portion of Kerala constituted the princely state of

Travancore. Prior to the 18th century the territory that would become Travancore was, like the rest of Kerala, a series of small, warring principalities. In 1729, however, the young and ambitious raja Martanda Varma came to the throne and, through a series of military conquests over the ensuing decades, he expanded the borders of his kingdom considerably. His successor, Rama Varma, also continued this expansion until the kingdom reached its greatest territorial extent.

Travancore covered 7,900 square miles and was the third-largest princely state in

India by population. It had a 19 gun salute, and was generally considered to be a prestigious territory and model for other princely states due to its progressive maharajas. The last ruler of the state, Chithira Thirunal, was even designated a

Knight Grand Commander of the Order of the Star of India. Travancore was a

Hindu kingdom, and its rulers were Kshatriyas.

Travancore did not suffer large amounts of British interference. Robin

Jeffrey notes that from 1800 to 1860 “little attempt was made to tamper with the state's society or local politics” (1976: 5), although the British did try to force the maharajas to modernize the state in the mid-19th century. Still, Travancore State was generally regarded favorably by the British, and it was often subject to special treatment. For example, the royal house in 1907 had no male heir but the British did not interfere in princely succession, instead allowing the rulers to adopt two girls.9 Most of the power resided with the maharaja and his diwan. A British

9 NAI, Madras Government – Travancore and Cochin Residency, Political Dept. (Confidential), 1907, #323; see also the Imperial Gazetteer (Vol. 24): 8 – “The laws which govern succession

203 report from 1914 noted that:

“The machinery of the administration is of the old world model. The government is conducted in the name, and under the direct control, of His the Maha Raja. The Dewan is His Highness' Chief Minister, with whom all departments correspond in official matters. It is a recognised maxim that no order issues in any official matter from His Highness except through the Dewan. There is no Executive Council; the machinery centres in the Dewan to an extent for which there is probably no parallel just now in any other Native State.”10

By the start of the 19th century the continuity which had defined the region encompassing Malabar and Travancore changed forever as the British came to control the north while the south remained independent. The main patterns of ethnic violence which are prevalent in contemporary Kerala were forged during this critical historical period.

After the end of colonial rule the princely states of Travancore and Cochin11 were formed together into the state of Travancore-Cochin. Malabar, on the other hand, was merged together with the new Madras State. Upon linguistic reorganization in 1956, most of Travancore and Cochin, along with a small part of the state of Karnataka, was merged together with Malabar, and the new state of

Kerala was formed.

are peculiar, descent being traced in the female line, according to the prevalent usage on the West Coast...[o]wing to failure of natural heirs, two girls have recently been adopted, with the sanction of the British government.” 10 NAI, Madras Government, Political Department (Non-Confidential), 1914, #282-3, pp. 3 11 Cochin was a former princely state located in the center of Kerala. However, I do not include it in the analysis of this chapter.

204 1.2 The Controlled Historical Comparison

Chapter 3 detailed why the Jaipur and Ajmer case studies formed a controlled historical comparison. I argued that the British took control of Ajmer not for any reason that might have been related to the dependent variable of this dissertation (ethnic violence), but rather for purely geographical reasons. The

Malabar-Travancore comparison, I will argue, is an even stronger pair of case studies for counteracting the problem of endogeneity. That is because there is significant historical evidence that the British wanted to conquer all of Kerala, but, for various reasons, were unable to annex Travancore. I highlight these reasons in the ensuing section; most of all, if not for the threat posed by the

Kingdom of Mysore the British might have conquered Kerala completely.

Travancore additionally benefited from an isolated location, as well as strong rulers who, though they could not easily be dislodged, were still willing to trade with the British. Therefore, I posit that the fact that Malabar came under the control of the British whereas Travancore did not is also a quasi-random historical outcome.

In the pre-colonial period Malabar and Travancore were very similar territories. British administrators themselves noted the unintended or incidental state of Kerala's political divisions:

“The early history of Malabar is inseparable from that of the adjoining state of Travancore. Identical in people, language, laws, customs and climate, the whole of ancient Kerala is homogenous [sic] in every respect, except in the accident of a divided political administration.” (Meyer et al. 1909: 56, emphasis added)

205 British administrators were initially interested in controlling the entire Malabar coast, which includes Travancore. S.M. Mohamed Koya notes:

“At this time the thinking of the Company's servants was in terms of reaping commercial benefits by acquiring ports like Karwar, Honore and Mangalore. It was hoped that if the Malabar coast were to come under the British as a result of the Mysore war it could be restored to cultivation and its products could be exported to Europe...” (1975: 514)

In fact, the first British trading settlement in all of Kerala was not in Malabar but in the city of Anjengo in Travancore (Logan 1887: 350). As I noted in the

Introduction of this dissertation, the British in India were interested primarily in annexing agriculturally productive territories. Like Malabar, Travancore was also an extremely fertile and productive area. Lord Curzon once noted rather lyrically about the region:

“Since I have been in India I have had a great desire to visit Travancore. I have for many years heard so much of its exuberant natural beauties, its old-world simplicity, and its arcadian charm. Who would not be fascinated by such a spectacle? Here Nature has spent upon the land her richest bounties; the sun fails not by day, the rain falls in due season, drought is practically unknown, and an eternal summer gilds the scene.” (Menon 1956: 274)

The Madras Government also documented in a 1907 report:

“[Travancore] is one of the most picaresque portions of South India, containing an extensive hill region, numerous rivers, and a succession of backwaters and vast forests.”12

Finally, Dick Kooiman adds about the uniqueness of Travancore:

“[most] princely states were located inland (Kashmir, Rajputana, Hyderabad, Mysore) rather than in coastal areas. In that sense, Travancore with its coastal location and sea ports was rather exceptional.” (2002: 16, emphasis added)

12 NAI, Madras Government – Travancore and Cochin Residency, Political Dept. (Confidential), 1907, #323

206 What made Travancore exceptional? That is, what factors prevented the British from taking control of this princely state as they did many others?

The major cause was rooted in the military campaigns of the neighboring

Kingdom of Mysore. In the mid-18th century the de facto ruler of the state, Hyder

Ali, began a series of military invasions throughout southern India. He was succeeded in these efforts by his son, . These campaigns brought

Mysore into conflict with the British East India Company, but Mysore was uniquely successful against the British. In fact, arguably no native power in

Indian history blunted the strength of the British like Mysore. Because of the threat posed by this kingdom, the British needed to make an alliance with

Travancore. This ultimately spared the state from British annexation and thereby sharply changed its historical trajectory. There were also a number of other additional factors which assisted Travancore in avoiding annexation, all of which

I discuss at length below.

1.3 British Expansion and Travancore

In the early 1760s the commander (dalwai) of the Mysorean army, Hyder

Ali, effectively took control of the state, overthrowing the ruling Hindu Wodeyar . In 1761 he established a Muslim sultanate and, with his son Tipu Sultan, began a series of aggressive military campaigns throughout south India.13 Four

13 These wars were, of course, highly complex historical events. For example, the first Anglo- Mysore War involved the British, the Kingdom of Mysore, the Nawab of the Carnatic, Hyderabad State and the Maratha Empire. I offer only brief details on these conflicts here. I mainly intend to show how Travancore became involved in these wars, and why the British needed their support to defeat Mysore.

207 subsequent wars involved the British East India Company, and Kerala was often an important site of conflict.

The First Anglo-Mysore War began in 1767 and was a complicated affair – the Kingdom of Mysore united with Hyderabad State against the British (who later formed an alliance with the Maratha Empire) in order to gain control over large portions of south India. During the Second Anglo-Mysore War (1779-1784),

Hyder Ali, who had by this time allied with the French, declared war against the

British East India Company after they took over the French port of Mahé. In the midst of this war Ali died suddenly and Tipu Sultan took control of the Mysorean forces. The war ended in 1784 with Mysore in control of southern Malabar.

The important point to note here is that many historians have argued that both of these wars were – at worst – stalemates for the Kingdom of Mysore. In reality, Mysore was one of the few native powers to have ever fought the British to a standstill. Many events from the conflict became known throughout the rest of

India. The first war had seen the armies of Hyder Ali reach and prepare to invade the British stronghold of Madras, and the second conflict witnessed Tipu Sultan besiege the British-controlled city of Mangalore and wrest it back under his rule.

The treaty which ended the Second Anglo-Mysore War, the Treaty of Mangalore, was also considered a major embarrassment for the British. Officers in India were

“indignant at the peace”, Warren Hastings especially, whereas Tipu Sultan

“believed himself to have won the war and behaved accordingly” (Dodwell 1934:

592). The British had been humbled by their conflicts against Mysore, and after

208 the Second Anglo-Mysore War they had come to regard this kingdom, as well as

Tipu Sultan, as a formidable enemy.

Buoyed by his successes, in 1788 Mysore began to threaten Travancore's borders, ostensibly because the state had provided refuge to Hindu rajas who had escaped from Malabar. In 1789 Sultan finally invaded Travancore, triggering the Third-Anglo Mysore War. However, this conflict was different from the preceding two because “[f]or the first time during his reign Tipu Sultan had to fight against the British without any allies, whilst they could rely on the support of several allies, including the Nisam of Haiderabad” (Frenz 2003: 82-3). By 1792 the conflict had ended and Sultan had been delivered a crushing defeat. In the subsequent Treaty of Seringapatam Sultan was forced to cede half of his kingdom to the British, including large swaths of the Malabar region.

During the start of the Third Anglo-Mysore War the various Malabar rajas and chiefs accepted the terms of British alliance; that is, they would unite to defeat Sultan and afterwards they would retain their territory. However, once

Sultan was defeated in 1792 the British reneged on their promise and took direct control of the region. The official policy of the British was to “regard [the Malabar rajas] as in the sense of Bengali zamindars, i.e. to strip them of their sovereignty and to seal with them appropriate treaties on tax” (Ibid.: 100). Malabar then became a province of British India. Had Tipu Sultan also controlled Travancore, the British most likely would have annexed this area too – however, he had not been able to bring Travancore under his control.

209 During the Fourth Anglo-Mysore War, the final conflict, the British completely destroyed the Mysorean armies by 1799. They were assisted by the nizam of Hyderabad, who had previously been an ally of Mysore. Tipu Sultan himself was killed during the siege of the capital city of Seringapatam in 1799.

After the war, the British took control of the Wayanad region of Kerala, the last portion of Malabar to come under their control, and subsequently re-established the old Wodeyar dynasty in Mysore.

Before the Third Anglo-Mysore War, the British had an interest in the entire coast of Malabar. Yet they signed agreements with Travancore which would ensure its continued political independence. The British needed the support of Travancore in order to fight against a dangerous rival in Mysore.

Barbara Ramusack confirms that John Shore, Governor-General of India after

Cornwallis, made this particular agreement with Travancore because of “the need to protect the districts on the Malabar coast recently acquired from Mysore, and the desire to counter the French” (2004: 67).

Correspondence files between British administrators located in the

National Archives of India also underscore this point. In 1789 administrators at

Fort St. George, the headquarters of the East India Company in Madras, noted as

Tipu Sultan prepared for war the “advisability of taking the Raja of Travancore under the Company's protection”.14 The Imperial Gazetteer also notes that:

“In 1778 the Raja [of Travancore] granted a free passage through his territories to

14 NAI, Descriptive List of Secret Department Vol. VIII, 1789-90, File dated September 9th, 1789, No. 1

210 the British troops sent to attack the French settlement of Mahé, then much valued by Haider as the base of his military supplies. He had already treated with contempt Haider's proposals to become his vassal, and had openly avowed himself the friend of the British, whom he actively joined in their campaign against the Muhammadan ruler.” (v. 24: 6)

Overall, if not for the serious threat posed by Mysore Travancore's colonial history may have been entirely different.

However, there were other reasons beyond the threat of Mysore which explain why the British might have made an agreement with Travancore. For example, the British may have made the calculation that defeating Travancore in military combat was no certainty. Travancore had already demonstrated its unwillingness to acquiesce to foreign control. William Logan, Collector of

Malabar, noted the following story involving Travancore's relationship with the

Dutch (whom they later defeated in battle) that highlights this point:

“In 1739 Mr. Van Imhoff became Governor. He was a most intolerant man, and directly after he arrived he saw the necessity of curbing the rising power of Travancore if the Dutch were to retain their hold of the trade of the country and not allow it to pass into the hands of the English, who were backing up the Travancore Raja. Van Imhoff, it is said, carried to the Travancore Raja his own protest against the Raja's occupation of the territory acquired in 1734. His protest failed, and Van Imhoff nettled at this result spoke of invading Travancore. 'The Raja replied that doubtless he might do so, but there were forests into which he could retire in safety.' Imhoff retorted that 'where Travancoreans could go, Dutch could follow.' The Raja then broke up the conference by sneeringly observing, he had 'been thinking some day of invading Europe!' Unfortunately for Van Imhoff he had no sufficient force at hand to command respect and obedience to his wishes.” (1887: 393)

Travancore also benefited from its geographical boundaries. P.

Shungoonny Menon noted that the state was protected by the Indian ocean on the west, the Western Ghats on the east, and backwaters and rivers on the north,

211 therefore “sav[ing] it from the incursions of foreign invaders” (1878: 7).

There is one final reason why the British did not support annexation in

Travancore: the state's control of the pepper trade. Kooiman notes that the

British alliance with Travancore might have been “built on the exchange of

'cannon for pepper'” (1988: 22). Logan recounted the words of another British administrator who stated that when the British first arrived in Kerala, Travancore was:

“...in a state which did not favour trade...What the English Company would have liked would have been a despotic monarch who could assign to them monopolies of the produce they came seeking and could enforce the same with a strong arm.” (1887: 350)

However, over the next several decades Martanda Varma emerged as that despotic monarch, and the British signed what became known as the 'pepper contract' with him in 1788 (Assary 2009). This agreement stipulated that

Travancore would supply the British with pepper in return for military protection. Historian Leena More notes along these lines that the British had long encouraged Martanda Varma's consolidation of power in the hopes that

“Travancore's supremacy in the region would create the necessary conditions for them in trade” (2003: 236). When the British took control of Malabar there was no comparable ruler of the region – instead, they encountered a series of warring rajas and petty chieftains. Malabar was therefore ripe for annexation whereas

Travancore presented a better opportunity for collaboration.15

15 Could the British have annexed Travancore even after these agreements? This is possible. However, reneging on the agreements, especially after doing so in Malabar, was certainly a risky proposition. But the British did have opportunities. In 1809, for example, the raja's

212 Ultimately, due to various contingent historical factors the British found themselves in a position which prevented them from annexing Travancore. The princely state subsequently became an important pillar of support in southern

India in the ensuing decades. Therefore, the selection of the Malabar-Travancore comparison mutes the problem of endogeneity – the British initially wanted to conquer the entire Malabar coast, but because of the Anglo-Mysore Wars and various other factors, they were unable to do so. This critical historical juncture would alter the long-term politics of both regions.

The next section considers the pattern of ethnic violence in contemporary

Kerala. I argue here that there is a discernible difference between ethnic conflict outcomes across north and south Kerala. The Malabar region in the contemporary period has been a hotbed for caste and tribal conflict, whereas in the former Travancore region religious violence has been the larger recurrent problem. In the ensuing sections I present evidence that highlights this divergence in ethnic violence outcomes.

diwan Velu Thampi attempted to mount an insurrection against the British in the region. However, the raja never supported the rebellion and later even aided the British in fighting against it. T.K. Velu Pillai notes that the British were, in putting down the rebellion, attempting to “rescue the Raja” from his diwan. They did not hold the prince responsible for the revolt, and “[t]hey still treated the Maharaja as an independent ruler and showed considerable respect to the officers of his government” (1940: 24-25). The British had formed a strong relationship with the Travancore raja and did not wish to disturb the internal workings of the government, which provided them with favorable trade policies. In the 1850s the British also threatened Travancore over declining state revenues, after which the princely administration initiated reforms in education and public services. After the Rebellion of 1857 annexation was no longer possible.

213 II. Religious Violence in Contemporary Kerala

How does ethnic violence unfold in contemporary Kerala? Are the patterns of violence established in Chapters 2 and 3 also evident when examining Malabar and Travancore? I begin this state-level analysis by examining communal conflict in both areas, and in section III I turn to contemporary caste and tribal conflict.

Kerala is often lauded as a state which is relatively free from communalism. To some degree, this is true – the state definitely experiences less religious violence than many northern areas, and for this reason it has a sterling reputation throughout India. However, as an article from a state newspaper about the 1992 post-Babri Masjid riots in Trivandrum noted, “...contrary to the claim by the state's politicians, Kerala has never been immune to communal clashes.”16 In this section I highlight that contemporary Kerala suffers from periodic episodes of communal violence, but that most incidents take place in the

Travancore region.

16 “Tremors Down South: Ayodhya sparks off trouble in Kerala”, The Week, August 2nd 1992

214 As I did in the previous chapter, I will present figures from different sources which indicate variations in communal conflict in contemporary Kerala.

However, the WITS database which I previously used does not record any religious violence in the area, mainly because none of the episodes which occur in

Kerala are bottom-up 'armed conflicts' – i.e. bombings, assassinations, etc.

Rather, almost all of the communal violence takes the form of riots. I first present figures from the Varshney-Wilkinson database to compare Malabar and

Travancore longitudinally in Table 4.3.

Table 4.3: Communal Riot Casualties in Kerala, 1950-1995 (per 100,000)17 Malabar Travancore Riot Casualties18 1.58 1.74

Gopal Krishna's fine-grained dataset on Hindu-Muslim riots in the 1960s also highlights that Trivandrum, the capital of the Travancore region and the state, is the nerve center of communalism in Kerala. He notes: “Throughout the decade Kerala experienced communal rioting in four places, and each had only one riot. Of the six incidents 3 occurred in Trivandrum, and of the 5 days of rioting Trivandrum accounted for two” (1985: 67). Krishna does note, however, that Kozhikode in Malabar has also experienced periodic riots, although entirely 17 These figures are based off of the information provided by Varshney-Wilkinson. However, there were some serious problems with the data. Most importantly, the dataset suggested that there were only 16 injuries in Trivandrum from 1950-1995. This is, frankly, impossible – Trivandrum is the most riot-prone city in Kerala. The dataset does note three riots in Trivandrum – in 1961, 1982/3 (Dec-Jan), and 1986. However, there are almost no figures provided on deaths and injuries for these conflicts. I have supplemented their dataset by using information from other books, interviews, and newspaper clippings. I document all of this additional material in the Data Appendix. 18 This figure is calculated as total deaths and injuries divided by the combined Hindu, Muslim, and Christian population (2001) of each region.

215 in rural areas of the district.

Though religious violence in Kerala is much less of a problem than

Rajasthan, these per-capita figures confirm the general pattern of communal conflict thus far established in this dissertation: higher levels of contemporary religious violence in former princely states. Travancore today has become a hotbed of communalism, but Malabar – despite Muslims constituting a much larger percentage of the population – has not experienced comparable religious violence. This is an important contextual point: areas in India where the Hindu-

Muslim divide is roughly equal are especially prone to riots (Wilkinson 2004: 32-

33), and therefore Malabar would be the perfect locale for communal conflict.

Also, the size of the Muslim population in Malabar has increased by 10 percentage points from 1931 to 2001 while the population of Muslims in

Travancore has remained essentially the same. We would therefore expect communal riot casualties in Kerala to be significantly higher in the Malabar region; yet, despite all of these precipitating factors, Malabar today is relatively peaceful. In the next two sections I use qualitative evidence collected from fieldwork in Kerala to further investigate this dichotomy in contemporary communal conflict outcomes across the state.

2.2 Communal Violence in Contemporary Travancore and Malabar: Qualitative Evidence

Since independence, the Travancore region of Kerala has experienced periodic episodes of communal violence. Many parts of Trivandrum have

216 specifically seen immense violence. This area is the center of communal rioting in the state; data from the Varshney-Wilkinson dataset show communal riots in

Trivandrum in 1961, 1982-3, and 1985. These figures actually greatly understate the amount of violence in the city – there were also severe riots in 1968, 1992, an entirely different riot in 1983, and 1995. Furthermore, the communal violence in this region is not merely isolated to Hindus and Muslims. The larger Christian minority in the region has also been involved in various communal incidents. As one example, in 2005 Muslims and Catholics engaged in a high profile riot in which two people were injured and over 1,000 more lost their homes to arson.19

To better understand the dynamics of violence in the area and why the

Travancore region has become plagued by communal conflict, I relied on a number of interviews in the region. A local university professor in Trivandrum summarized the communal situation there:

“There were riots in Trivandrum in the 1990s. Communal outfits are always ready to act. The coastal regions down here are especially dangerous – and it's not just about Hindus and Muslims, but Christians are also involved in the rioting. They are a major presence here in Travancore.”20

He also noted:

“Up north there have been some problems, like in Marad. But not that much violence. The thing you will notice when comparing Muslims in Malabar and Travancore is that while the Muslims down here [in Travancore] are educated, they are extremely religious. Their religion is very important to them. Muslims in the north, though, have become very secularized.”

As I noted in Chapter 1, the British made great efforts in the provinces to expunge religion as a legitimate mode of ethnic classification. They sought, instead, to 19 “150 houses gutted, several injured in clash”, The Hindu, Feb. 28th 2000 20 Interview in Trivandrum, July 22nd 2011

217 establish secular rule. In Kerala, the Mappillas (the name for Muslims in

Malabar) over time became secularized, and the Hindu-Muslim divide subsequently became less salient in that region. In Travancore, on the other hand, religion is the dominant ethnic category.

Another respondent, a local newspaper editor, also commented on the history of riots in Trivandrum:

“I recall major riots here in the early 1980s, around 1982. Very often these occurrences happen during religious festivals. At that time the army was even called in. There is periodic violence in this area [Travancore], and a lot of tension between local religious groups, between Hindus and Christians, and Muslims and Christians. But up north, the major issues are caste issues, issues with depressed castes, those kinds of problems.”21

To be clear, however, the socioeconomic position of the Muslim community in Travancore is not bad – in fact, compared to Muslims in most other parts of India, Travancorean Muslims are quite advanced. This is partly a function of the incredibly high levels of social development throughout Kerala more generally (Heller 1999). The socioeconomic condition of Muslims in the region is not the major driver of violence; rather, religious tensions and antagonisms in the area run very high. That is why conflict does not occur in a bottom-up fashion; the socioeconomic condition of Muslims in Travancore has not led them to initiate violence against other groups. Rather, because tensions are high between religious communities, there are periodic (top-down) riots in which Muslims are largely the victims of attacks by Hindus and Christians. This violence has become deeply embedded in the region. Take, for instance, the latter

21 Telephone interview in Trivandrum, May 27th 2012

218 community: clashes between Muslims and Christians in the coastal areas outside

Trivandrum constitute an “almost...hundred-year conflict” (Singh 1985: 29, emphasis added). Interviews in Travancore suggested that the varied religious communities in the region are deeply suspicious of each other, and violence between them has consequently become endemic.

It is also clear that religion forms the basis of local politics in the

Travancore region, especially the areas around Trivandrum. A local study on the growth of communalism in Travancore noted the recent success of Hindu nationalism:

“Emboldened by the patronage of the NDP and the Congress, the RSS took the first big step in fortifying its base in Kerala by organising early in 1982 a Vishal Hindu Sammelan. It was a mammoth show. The Sammelan gave out the clarion call for the unity of the faithfuls to save Hinduism. The conference of the Nair Service Society followed months later which gave a call for saving not only Hinduism but Hindus' jobs as well.” (O.J. 1983: 210)

Yet the growth of communalism has not affected the Malabar region to the same extent. As I detail later, the Communist Party, which derives most of its support from low castes, has been the larger presence in the north.

While Travancore has been plagued by communal animosities, Malabar has experienced less violence. To be sure, Malabar has had some notable instances of communal conflict in the post-independence period. December 1971 marked the beginnings of a major riot in Tellicherry which, though it produced no deaths, led to a significant number of injuries. More recently, in 2002 the infamous 'Marad Massacre' occurred in which a group of Muslims attacked and

219 killed nine Hindus in Marad, near Calicut. However, despite the occasional outburst of communalism, most of the ethnic violence which has occurred in

Malabar has been, as I will detail later, focused around castes and tribes.

Therefore, the Hindu-Muslim divide – despite the fact that the population of both groups is roughly similar – does not constitute the main axis of conflict in contemporary Malabar. As I detail later (Section 4.4), this is especially surprising considering that this area did experience some communal conflict in the colonial period. Over time, however, tensions decreased and Hindus and Muslims today throughout Malabar exist in relative peace.

2.3 Existing Explanations for Religious Violence in Kerala

In this section I consider the current political science literature and the three dominant arguments for communal violence. I argue that, as was also the case in Rajasthan, none of these arguments can sufficiently explain why

Travancore is today more of a communal riot-prone region than Malabar. In the fourth section I then present my explanation: that the cause of communalism in

Travancore is rooted in the historical policies regarding religious minorities instituted by the princely government. These policies led to severe religious antagonisms that still reverberate in the Travancore region today.

First, Brass' work has focused on the rise of Hindu nationalist organizations in producing communal conflict. As Brass' theory would predict, the Travancore region of Kerala is in fact the major part of the state in which the

Sangh Parivar operates. I detailed earlier the rising power of Hindu nationalist

220 outfits in the area. Dick Kooiman, for example, writes that:

“Over the past few years Kerala has become a fresh recruiting ground for the Rashtriya Sevak Sangh (RSS), a communalist party which has established powerful strongholds especially in Trivandrum, the former capital of Travancore.” (1995: 2130)

However, in Travancore Hindu nationalist groups are merely a symptom of the disease, not the disease itself. As Kooiman importantly adds:

“These manifestations...may be seen as a continuation of the tensions and conflicts that have plagued Kerala, particularly the former Travancore part of it, for a much longer time.” (Ibid., emphasis added)

In short, the rise of the Sangh Parivar in Travancore is recent while communal clashes in the region are not. Therefore, communal conflict bred Hindu nationalism in Travancore and not the other way around.

Next, Wilkinson has placed a lot of importance on the Kerala case in proving his argument about electoral competition and Hindu-Muslim riots. He has argued that Kerala as a whole has seen a smaller number of riots compared to many northern states because of the importance of the Muslim voting bloc in coalition governments. He writes:

“The reason why the level of Hindu-Muslim violence has been so low, despite the existence in Kerala of antiminority mobilizations similar to those that have led to violence elsewhere in the country, is that high levels of party fractionalization have forced successive governments to order the Kerala police force to prevent attacks on minorities in the state at all costs.” (2004: 186).

However, this argument is pitched at the state level and therefore cannot explain the important variations which exist between north and south Kerala. While

Muslims do form a larger portion of the population in northern Kerala, they are also an important minority in the south. The coalition governments which have

221 controlled the state, led either by the Communists or the Congress, cannot afford to ignore Travancorean Muslims. Furthermore, even if this were the case, it is highly unlikely that northern Muslims – or organizations such as the Muslim

League, which are centered in the north – would not be concerned with the plight of Muslims in the south. Wilkinson's argument may explain why Kerala as a whole experiences less communal violence than states in the north, but it cannot explain crucial sub-state variations in rioting.

Finally, Varshney's argument focuses on the importance of intercommunal networks, and he, like Wilkinson, also places special emphasis on Kerala in his book, specifically the city of Calicut (Kozhikode).22 Varshney notes that Calicut has not had a single Hindu-Muslim riot in a century23, and that the city maintains a “deep intercommunal civic engagement” (2002: 120). With specific regards to the Kerala case, interviewees were asked about formal intercommunal associations in the region, and, mirroring fieldwork results from Ajmer, few could provide examples of these kinds of organizations or the effect they might have on religious conflicts.

Therefore, none of the main existing explanations for communalism can explain why Travancore experiences more religious violence than Malabar. In a more general sense, they are not historical explanations, but are rather focused

22 However, Varshney did not compare north and south Kerala; rather, he compared Calicut with Aligarh, Uttar Pradesh. 23 He makes this claim on pg. 121 of Ethnic Conflict and Civic Life. However, his dataset indicates that Calicut has, in fact, had riots, or at least communal disturbances in which people lost their lives or were injured. The dataset notes incidents in 1954 and 1965.

222 squarely on the post-independence period. However, communalism is a deeper problem and, as I will detail later, the best explanation for its rise in Travancore is the legacies of princely rule, especially discriminatory policies that were instituted against non-coreligionist Muslims and Christians in the state.

III. Caste and Tribal Violence in Contemporary Kerala

Despite its reputation as a bastion of caste politics, contemporary Kerala evidences a clear distinction between the salience of caste and tribal identities in the north and south, and levels of caste and tribal conflict also vary significantly in both regions. While Malabar has today become synonymous with simmering caste and tribal tensions, similar violence in the Travancore region has never materialized. In this section I present primarily evidence about the levels of caste and tribal violence in Malabar and Travancore. Unfortunately, figures on both caste and tribal violence were not available for Kerala; the same NGO, or any similar NGO, which I used for data in Rajasthan does not operate in Kerala.

However, there are still a number of other ways of highlighting the higher levels of caste and tribal violence in Malabar as compared with Travancore. I first consider caste violence in the region, then tribal conflict.

Caste is much more salient of an ethnic category in contemporary Malabar than Travancore. First, the Communist Party of Kerala has always had its strongest base in this region (Menon 2007). While Communist Party-led coalition movements have governed the state repeatedly since independence, most of their support comes from Malabar. Donald Zagoria, for instance, writes:

223 “It is, for example, well known that Communist electoral strength is considerable in three Indian states – Kerala, West Bengal and Andhra. But it is less well known that communism is much stronger in certain identifiable parts of these three states – the Malabar portions of Kerala...” (1971: 145)

The success of the Communist Party in Malabar is a strong indicator of the salience of caste and tribal identifications because the party derives most of its strength from low castes and tribal groups (Jeffrey 1978, Menon 2007).24 At the same time, the Communist Party is ostensibly atheist, and does not make religious entreaties to voters.25 Because of the religiosity of Travancore, the communists have been unable to derive the same level of support from this region. It has also failed to achieve the same level of success in Travancore because, as I will show, the princely government preempted the message of low caste equality by passing several progressive policies for these groups in the 19th and 20th centuries. Finally, even outside the organizational capacity of the communists, the Indian peasant movement, specifically the Kisan Sabha

(Peasants Union), has a regional stronghold in Malabar (Karat 1973: 24). The

24 Jeffrey writes about the comparative weakness of the communists in Travancore: “The big landlord – such a useful focus for class consciousness in Malabar – was largely absent [in Travancore]. These Syrian Christian small holders were more prosperous than the Mappilla tenants and laborers of , yet they shared a similarly undisturbed family system and an acute community consciousness reinforced by hundreds of active clergy. They too provided infertile soil for the young leftists of the 1930s” (1978: 83). 25 T.J. Nossiter's study of notes that both the Muslim and Christian communities have provided minimal support to the movement in Kerala. He writes, for example, that Christians have so closely aligned with the Congress Party that it has sneeringly been referred to as 'Christian Congress' (1982: 22). He also notes that: “Such limited inroads as the communist movement has made in Muslim areas have been among the depressed landless Moplahs of the interior” (pg. 24) – i.e. low caste convert groups. Similarly, Varshney writes that: “In areas dominated by Christians and Muslims, the Communists have historically done badly...Communism has made the greatest sense to caste-based Hindu society. Among the Hindus, support for the Communists has been centered among the and other low or scheduled castes” (2002: 164-5).

224 strength of these groups is indicative because both of them are extremely vigilant about caste violence and have made 'social justice' for depressed castes a primary focus of their work.

Tribal violence has also been much larger of a problem in contemporary

Malabar than Travancore. The Malabar region is in fact the only area of Kerala in which Naxalites ever operated. I discussed the Naxalites in Chapters 1 and 2, and why they are a good proxy for caste and tribal conflict. In the late 1960s, a

Naxalite faction of the -Marxist in Kerala organized coordinated attacks against police stations and wireless stations in Malabar, specifically in Tellicherry and Pulpally. Wayanad District has been the very heart of tribal conflict in the state. While the Naxalites have not been an extremely deadly force in the region, they have continued to carry out sporadic attacks in the decades since their founding. In contemporary India, the Wayanad area in

Kerala is now considered to be 'Naxalite-targeted' by the Indian government. In contrast, there has been almost no Naxalite activity at all in Travancore.

Even in the past several years tribal conflict has been a problem in

Malabar. The most famous recent incident was the government-initiated attack in

2003 against protesting adivasis in Muthanga (Bijoy and Raman 2003), which is located in Wayanad. In response to numerous failures by the state government to distribute land to adivasis, a small contingent of tribals entered the Wayanad

Wildlife Sanctuary and took up residence in the area as a protest of the government's failures. A brutal crackdown by the government soon followed. It is

225 no mistake that this tribal disturbance occurred in Malabar, as this region has the highest percentage of tribal landlessness in the state (Ibid.: 1976).

3.2 Caste and Tribal Violence in Contemporary Malabar and Travancore: Qualitative Evidence

In this section I use qualitative evidence to examine the socioeconomic condition of low castes and tribals today in Malabar and Travancore, as well as the state of caste and tribal antagonism in both regions. A local university professor summarized the overall state of conflict in Kerala:

“Caste conflict is concentrated in the north, but it happens all over the state. Caste politics are big in Kerala. But the adivasis, the Naxalites, they were only operating in the north. This was before 1977, and they lost some support over time.”26

Caste antagonisms run high in modern Malabar. Even the Planning

Commission in their Kerala Development Report in 2007 noted that the caste system in Malabar is much more rigid than that of Travancore, and they make a reference to the past in explaining this fact:

“[In Travancore] the institution of attached labour began to wane and that of wage labour, expand. These developments led to the decline of the rigidities of the caste system in Travancore. But in Malabar, the caste system remained entrenched in the prevailing agrarian system.” (2007: 56).

Similarly, the intensive work of P. Radhakrishnan in selected villages in Malabar in the 1970s also highlighted the continued salience of caste, as well as the existence of widespread caste antagonisms:

“While the pattern of tenancy...clearly reflected the virtual dependence of certain social groups on the Nambutiris because of the latter's dominance in the village land system, certain other social groups were reduced to even greater dependence as agricultural labourers, homestead tenants, household servants, and so on.

26 Interview in Trivandrum, July 22nd 2011

226 Such dependence in turn had led to the, social subservience of all the dependent groups which was perpetrated by the Nambutiris by insisting on strict adherence to traditional caste-practices and and caste-based socioeconomic relations.” (1989: 126)

Overall, the socioeconomic position of low castes and tribals in Malabar is significantly worse than that of Travancore. This is why the Naxalites have only operated in this region – the backwardness of SCSTs has led to 'bottom-up' ethnic violence. Antagonisms between different castes also runs high in Malabar, and caste politics in the region are deeply divisive.

In the next section, I trace the roots of Kerala's ethnic violence patterns through a study of both regions during the colonial period. I argue that caste and tribal conflict is such a major problem in contemporary Malabar because this region came under the direct control of the British, who then implemented a series of discriminatory policies which decimated the condition of low castes and tribal groups. Travancore, on the other hand, remained an independent princely state, and the administration, while protective of low castes and adivasis, was primarily discriminatory toward non-coreligionist Muslims and Christians.

Prema Kurien summarizes the main argument which I formulate in this next section, noting that in Kerala “ethnic formation was the consequence of the economic and political compulsions of colonial rule” (1994: 385).27

27 Kurien's argument is mainly about Mappillas, Ezhavas, and Syrian Christians, whereas my argument is about a caste/tribal – religious divide between north and south Kerala.

227 IV. The Colonial Roots of Religious Violence in Kerala

Why is religious violence in Kerala concentrated in Travancore? Why, despite a much larger Muslim population, has Malabar somehow remained relatively immune from similar conflict? The central argument that I advance in this section is that the basic pattern of religious violence seen in the contemporary state of Kerala is an historical legacy of the colonial period. During this time the Malabar region came under the direct control of the British whereas the princely state of Travancore retained its independence. Princely rule in

Travancore created the region's strong religious identity and rendered it an area rife with communal tension. British religious policies in Malabar, however, were more complicated, and I discuss them at length in sections 4.3 and 4.4.

4.2. Communal Policies of Travancore

Travancore has been cited often by scholars as one of the most progressive and tolerant of the Indian princely states. Susan Bayly, who analyzed Kerala as part of her important research on religion in south India, notes that “[t]he Hindu rulers who fought their way to power in the kingdoms of Travancore and

Cochin...were radical and ruthless; at the same time their religious policies were fluid and syncretic...” (1984: 178). She notes that south India as a region is unique because “powerful and dynamic variants of the three major 'world religions' –

Hinduism, Islam and Christianity – have grown, developed and overlapped, all within the comparatively recent historical past, and all within a setting of remarkably rapid social and political change” (1989: 9). J.J. Roy Burman also

228 points out the syncretic practices which are prevalent in Kerala, noting that devotees of the famous Sabrimala Hindu pilgrimage site in Travancore also “visit the temple of Ayyappa to first pay their obeisance to the shrine of a Muslim saint named 'Vavara'” (1996: 1214).

However, I argue here that the situation for religious minorities in

Travancore does not resemble the rosy picture painted by existing scholarship.

On the contrary, Travancore was a strict Hindu state which both enacted discriminatory policies toward non-coreligionists like many other princely kingdoms, and also experienced periodic conflagrations of communal violence, especially in the 20th century. In the ensuing section I focus not only on

Travancore's policies toward Muslims – who were a small minority – but also their policies toward the larger Christian minority. The policies enacted by

Travancore toward these groups form the basis of the communal conflict which plagues the region today.

From its earliest days as a consolidated power Travancore was an orthodox

Hindu state. In 1750 Martanda Varma famously dedicated the entire kingdom to the deity Padmanabha, a name for Vishnu. The maharajas of the state were thereafter technically servants (dasas) of Vishnu, and considered themselves protectors of the Hindu faith and its customs (Smith 1963: 94). G.B. Malleson of the Indian Civil Service in his An Historical Sketch of the Native States of India singled out Travancore as an “example of a territory which, from the earliest tradition, has always been under Hindu rule and governed by Hindu law” (1875:

229 323). Notably, Travancore never came under the control of Muslim invaders from the north, meaning that its Hindu character remained unbroken over time. Dick

Kooiman notes that in Travancore there was an “explicit connection between

Hindu orthodoxy and state government” which gave “local politics...a distinctly communal complexion” (1993: 288). Robin Jeffrey similarly calls it a “staunchly theocratic Hindu state” (1976: 3).

Travancore was also deeply dedicated to the promulgation of Hinduism.

The Travancore Administration Report of 1879-80 notes that the 4th largest expenditure of the state (not including the British tribute) was “The Devasom or religious institutions”28, which cost the state 572,940 rupees a year.29 To put this in perspective, this expenditure was more than the state spent on the judiciary, police, and education combined. This focus on religion was also consistent across many rulers of the state, just as was the case in Jaipur. For instance, when

Travancore acceded to the Indian Union after independence one of its central demands was that the new Government of India continue to subsidize temples at the same level as the princely administration (Majumdar 1951: 274).

Though the British generally viewed and held up the princely states as traditionalistic religious polities, administrators were nevertheless from time to time wary about what they perceived as an excessive focus on religion, and this was often the case in Travancore (Kawashima 1998). A Madras Presidency administrator bemoaned in the late 19th century that “[t]he Rajas were as a rule

28 A “devasom” is a trust used to manage temples. 29 NAI, Madras Government, Political Dept. (Non-Confidential), 1881, #429

230 mere puppets in the hands of certain Brahmans of the Trivandrum temple30...”

(Logan 1887: 350). The British were especially concerned (correctly) that large amounts of spending on religious institutions was hampering the state's ability to provide other services. To counteract this, British Resident Colonel Munro, who briefly served as the diwan of the state in the early 19th century, brought the land of 378 Hindu temples which had previously operated independently under the control of the central administration.31

The Muslims of Travancore were a small but important minority in the state, approximately constituting seven percent of the population during the colonial period. They represented a distinct Muslim community in India – they were not descended from Mughal invaders, but rather from Arab traders and merchants who probably first visited Kerala some time during the 7th century.

Unlike most northern Muslims they did not speak Urdu or Arabic, but

Malayalam. They were also a separate community from the Muslims of Malabar, who were known as Mappillas.

Throughout the early 20th century, contrary to the perception of

Travancore as a syncretic state, there were several instances of Hindu-Muslim violence which occurred. In 1908, for example, there was a major riot in Chalai.

T.K. Velu Pillai notes that this was only the latest explosion in an increasing trend of violence:

30 This refers to the Sri Padmanabhaswamy Temple in Trivandrum, one of the most famous Hindu temples in India. 31 NAI, Madras Government, Political Department (Non-Confidential), 1914, #282-3

231 “There had been other riots before...One of the earliest occurred at Alleppey. The so-called Pulaya riots in Neyyatinkara taluk and adjacent places, the riot at Talayolapparambu in Vaikam and the riot at Cape Comorin were among the more serious. They had to be put down by the strong hand. It may be mentioned that these riots had their origin in local disputes or communal tension...” (1940: 707)

The situation in Travancore began to grow worse in the 1930s as the

British decentralized power and the native states became increasingly independent. Incidents of communal violence in the state began to increase, much as they had in Jaipur during the same time period. In 1930 there was a small Hindu-Muslim riot in Manakad.32 An article from The Muslim of that year noted about the incident:

“Taking the side of those parties whose purses are heavy the Police began to disbelieve the statements of the Mohammedans and rudely dealt with them. Inspector of Police, Mr. Karunakara Kartha and his immediate superior were the persons who were responsible for the investigation of this case. In either eyes all the fault lay on the Muslims. As a result, they refused to hear the just complaints of the Muslims.”33

In 1939 at a meeting of the Travancore Muslim Young Men's Association, an advocate named A.K. Bijili noted that while the state had done much to improve the lot of low castes, they lagged behind in their treatment of Muslims, thus leading to increased communal conflict:

“When the present Dewan of Travancore Sir C.P. Ramaswamy Iyer came into power, we thought that his term of office will mark a new era in the history of the State of Travancore. His bold and ready response to the need of the untouchables in being instrumental to throw open the gates of the temples of the State for them and making these temples accessible to them is indeed an act of which any man can justly be proud. The Temple Entry Proclamation of His Highness the Maharaja of Travancore was an eye-opener to many. But the recent [communal] unrest in Travancore is really an unexpected development.”34

32 KSA, Confidential (1913-1944), 779, 1930, #21 33 Ibid. 34 KSA, Confidential 1944-49, 3629, 1944, #174

232 As I will detail later, Bijili's remarks highlight an important point: there is a clear distinction between how Travancore State responded to low caste agitation and

Muslim agitation. While the state had a progressive streak toward the former group, the latter made few advancements as a community before independence.

In the 1940s Muslim agitation continued to rise in Travancore. Muslim advocate Mohamed Ismail wrote to the diwan Ramaswami Iyer in July 1940:

“Inspite [sic] of the fact that Travancore stands foremost among the Indian States the position of the Muslims in this country leaves many things to be desired. They have been labouring under various disabilities. Attempts were made in the past to improve their lot and to bring them under the influence and guidance of a centralised organisation; but all of them have met with complete failures owing either to lack of response on the part of the members or want of selfless and efficient leadership.”35

In 1946 a large Muslim conference was planned in Balaramapuram, but the

Travancore government banned it. The president of a local Muslim agency noted:

“It happened that the conference was banned under orders of Government. The Police authorities of Neyyattinkara and Nemom, who for special reasons have their axe to grind against us were emboldened by this and thought that this was a fit occasion to harass and persecute us in different ways.”36

There were also other smaller agitations which continued to occur during this period. In 1944 another riot broke out in Trivandrum.37 In 1946 both Muslims and Christians complained about new rules which made it more difficult to construct churches and even cemeteries.38

The situation of the Muslims of Travancore then does not seem to match the account of most scholars who have emphasized the syncretic nature of the

35 KSA, Confidential 1913-44, 77, 1940, #52 36 KSA, Confidential 1944-49, 590, 1946, #187 37 KSA, Confidential 1913-44, 1825, 1944, #80 38 KSA, General Section 1945-46, 1598, 1946, #409

233 state. In reality, because the maharajas of Travancore emphasized their Hindu religious identities so firmly, they did what they could to limit and hold back the

Muslim population. But what about the largest religious minority in the state –

Christians? They formed a much larger percentage of the population than

Muslims did in the region – almost a third of the total population. The Christians of Travancore are an important case because they highlight that religious violence in the princely states, though mainly between Hindus and Muslims, also existed among other minority religious groups. My principal argument has been that princely states were discriminatory in general toward all non-coreligionists, and the Christians of Travancore were no exception.

Christianity had such a strong presence in Kerala because of its long lineage – according to local tradition, Saint Thomas arrived on the shores of

Kerala in 52 A.D. and established Christianity in the region. Over time the number of denominations multiplied in the state: Catholics, Protestants,

Anglicans, Mar Thoma Syrians, etc. The state as a whole has a Christian population of roughly 20%, although Christians are much more concentrated in the Travancore region.

Throughout the 20th century Hindu groups in Travancore worried about the power of the sizable Christian minority in the state, and they periodically

“accused [them] of a hidden scheme to take over their kingdom” (Kooiman 1993:

288). The princely government later linked with the Hindu Mahasabha in an effort to counter the strength of the Christian minority (Kooiman et al. 2002: 61).

234 The state had also been active in trying to prevent conversions to Christianity, and I will detail later their attempts to counter-act the work of Christian missionaries.

For much of its history Travancore's Christian community existed in relative peace and prosperity. However, in the late 19th century this situation began to change dramatically. Susan Bayly has documented the increasing violence between Hindus and Christians during this period. Many Christians, who previously had not held any kind of subservient position in Travancore society, had suddenly come to be considered low caste, or even avarna (without caste). She recounts:

“By the mid-1880s the breakdown in ties between Syrians and Malayali caste Hindus had reached its final stage. Syrians were now routinely excluded from Hindu festival rites, and in many temple centres Hindu office-holders levied fines on Christians who tried to uphold their long-standing temple honours...During the 1880s and 1890s there were innumerable riots between Hindus and Syrians. These outbreaks closely resembled the Hindu-Muslim clashes of north India.” (1989: 294, emphasis added)

What accounted for this sudden change in the perception of Christians in the state? I posit that one of its principal causes was that in 1875 the Travancore

State conducted its first census. As I wrote in Chapter 1, while the British use of the census led to an increase in caste antagonisms, it was not obvious how princely rulers made use of these data (Ramusack 2004: 89). Once Travancore conducted its first census, however, Christians were officially resigned to a lower societal position. It seems that the princely administration used the census, then, not to promote caste antagonisms, but to diminish religious rivals.

235 The Syrian Christians39 in particular saw a major decline in their fortunes.

They had previously been a relatively affluent group with a strong standing in the commercial sector. By the start of the 20th century, however, their situation had completely reversed. As the princely administration in Travancore grew more hardened in its religious orthodoxy, Syrian Christians lost ground and grew increasingly frustrated by state discrimination. They were even prevented from investing in industry (Isaac and Tharakan 1986). Kooiman uses this situation to point out an important distinction between British India and the princely states on the treatment of religious minorities:

“Whereas in British India electoral safeguards to Muslims might help to maintain the empire, in Travancore political concessions to Syrians, the most prominent Christian community, were viewed as threatening the kingdom that became increasingly connected with Hindu orthodoxy.” (1995: 2130)

Both the Muslims and the Christians also pressed for separate electorates in Travancore. Just like in Jaipur, these communities made reference to the separate electorates which existed in British India – in this case, Malabar. The

Travancore princely administration did not acquiesce to these demands, and

Muslims and Christians were not given separate electoral arrangements. The principal argument made by the darbar in response to this demand was that the increased representation of religious minorities in the government was not possible in an explicitly Hindu kingdom. As T.J. Nossiter summarizes this debate: “Did citizenship extend to low-caste helots or to non-Hindus in a state

39 They are known as Syrian Christians because they claim ancestry from that region and their religious ceremonies are conducted in Syriac.

236 dedicated to a Hindu God?” (1982: 77).

This is an important and overlooked point because many authors have suggested that the introduction of separate electorates in British India was a central factor contributing to Hindu-Muslim violence, or even the bloody partition of the country.40 Yet in Jaipur and Travancore the princely administrations did not create separate electorates and both of these states still experienced more communal violence than their neighboring British counterparts. This creates a puzzle; as Kooiman has smartly noted: “...how can we explain the emergence of communalism in Travancore, if there is no government to be blamed for dividing the people by separate electoral arrangements?” (1995: 2130). In fact, the emergence of communalism owes more to the discriminatory treatment of religious minorities by the state.

Muslims and Christians also pressed for increased representation in the

Travancore government. This demand was couched in a much broader movement for democratization within the state. In 1891 over ten thousand signatories from marginalized communities presented the government with the 'Malayali

Memorial', a demand for increased representation in government services. This petition also objected to the Travancore government's practice of importing

Tamilian Brahmins for higher posts in the princely administration. Low caste communities as well as Christians and some Muslims were involved in this petition.

40 Paul Brass and Francis Robinson have engaged in a lively debate on this issue. See Robinson 1977, and Brass 1977.

237 In response, Travancore State established the Sri Mulam Popular

Assembly, which allowed semi-direct elections beginning in 1905. However, by the 1930s Travancore had made little headway in increasing the representation of either Muslims or Christians in the government. Table 4.441 below highlights the inclusion of these religious minority groups in the government apparatus.

Muslims were the worst represented group in the entire administration, and

Christians were also represented far lower than their proportion of the population. Low castes also lagged behind their population percentage, but the state had done more to integrate them into the government than it had for

Muslims and Christians.

Table 4.4 – Communal Representation in Travancore Gov't, 1930s Community Number of Posts Held Percentage Pop. % Hindus 16405 80.5% 61% Muslims 436 2.1% 7% Christians 3537 17.35% 32%

Ultimately, Travancore by independence had become a communally- sensitive state, especially after the dawn of the 20th century. The state had actively limited the position of both Muslims and Christians, and there were a number of communal riots which occurred in response to various discriminatory policies implemented by the state. The roots of communal animosity in the region today are best understood as the long, historical outcome of forces first unleashed during the colonial period.

41 Information from the Travancore State Manual (Volume II): 754-755. Unfortunately the data only refer to the “1930s”. The fourth column indicates the percentage of jobs each community held; the fifth column indicates their overall population percentage in 1931.

238 4.3 Communal Policies of Malabar Province

One of the central arguments in this dissertation is that communalism is a larger problem today in former princely states than former British provinces. The

Malabar case, however, poses a potential problem for this analysis. The various

Mappilla uprisings42 which occurred there in the late 19th and early 20th century, culminating in the massive Mappilla Rebellion (or ) of 1921, are often taken as evidence that British colonial rule promoted religious antagonism between Hindus and Muslims. Many accounts of the Mappilla uprisings, from both contemporaries and modern scholars, emphasize the religious nature of the incidents. Special Commissioner T.L. Strange was convinced that the uprisings were due to religious fanaticism, and historian

Stephen F. Dale argues that “religious leaders were the single, critical determinant of the Mappilla outbreaks” (1975: 87). Steven Wilkinson writes that the uprising in 1921 “was one of the worst outbreaks of communal violence ever in British India” (2004: 185). But were the Mappilla uprisings really religious conflicts?

4.4 Re-Interpreting the Mappilla Uprisings

Mappilla is the term used to denote the Muslim community residing in

Malabar. Muslims formed almost a third of the total population in the region during the colonial period. Among the Mappillas there were some critical

42 I use the term 'Mappilla uprisings' to denote all of the Mappilla conflicts which occurred in the late 19th and early 20th centuries. However, the major insurrection which occurred in 1921 I call the 'Mappilla Rebellion'.

239 differences: those in northern Malabar were a wealthier group, and many of them were involved in business. Those in the south, however, were poorer and worked primarily in agricultural occupations. Most of the agitators during the Mappilla uprisings were specifically from the southern Malabar area – the British derisively referred to this group as the 'Jungle Moplahs'.43

During the 19th century the were involved in a number of violent incidents – colonial administrators often referred to them as 'outbreaks' or

'outrages' – which generally followed the same pattern: Mappillas would kill oppressive Hindu landlords and then take up residence in a public place in order to confront the police in a gunfight. Many Mappillas then committed suicide, and these conflicts were considered to be a kind of jihad.

Then, in 1921, a major uprising occurred, known as the Mappilla

Rebellion.44 In the years preceding this insurrection, the Indian National

Congress had been trying to make inroads among the Muslim population for support in the independence movement. When the British government, after

World War I, disbanded the Ottoman Empire the position of the Caliphate, the leader of worldwide Islam, was suddenly in jeopardy. The Khilafat movement in

India formed in response as an Islamic campaign to support the Ottoman

Empire. After the war, Gandhi and the Congress Party linked with the Khilafat movement, hoping to unite Hindus and Muslims together in their opposition to

43 Mappilla was spelled various ways by colonial administrators: Moplah, , etc. 44 Much has been written about the history of the rebellion. For more information, see Wood 1987, Panikkar 1989, and other sources cited in this text.

240 British rule. Many members of the Khilafat movement subsequently became active in the non-cooperation campaign.

During the first few months Hindus and Muslims in Kerala worked together on the issue of agrarian reform. As the Mappillas had grown more aggressive, the British authorities in Malabar decided to upend the movement by arresting several of its leaders. They also raided a mosque in . This resulted in an enormous backlash as the Mappillas organized and attacked a number of police stations, British camps, courts, and other government buildings. Martial law was declared. For a time the Mappillas controlled large portions of the region – Khilafat kingdoms were officially established in Ernad and Valluvanad.

Many thousands of Mappillas were involved in the ensuing insurrection.

The Congress tried to stem the violence but was unsuccessful, and they eventually abandoned the movement. Press reports indicated that the Mappillas attacked the British, landlords, and then Hindus whom they believed to be collaborating with the colonial government. E.M.S. Namboodiripad, famed Communist leader in Kerala and the first Chief Minister of the State, noted about this:

“The Moplah found that his Hindus compatriots both the leaders and the rank and file deserted him; the military arrived to hunt him out of his abode; his Hindu neighbours helped the military...against him. He naturally got enraged at them...No wonder then if anti-Hindu actions took place. The wonder is rather that they were so few in number and proportion.” (Kusuman 2003: 160-61)

The British eventually had to call in reinforcements from elsewhere in

India to defeat the rebellion, and after six months the insurrection was put down.

241 According to official estimates, almost 2,400 rebels were killed while some 1,700 had been injured. Almost 6,000 more were taken prisoner. The Mappilla community was brutally punished for the rebellion.45 Many Mappilla prisoners were also sent abroad in a penal colonization scheme.46

There are several reasons, however, to question the traditional characterization of the Mappilla uprisings as communal incidents. I argue here that the uprisings were, in fact, agrarian and anti-colonial in nature.

Furthermore, the Mappilla uprisings have an under-analyzed caste dimension to them. In that sense, as I will show, the Malabar uprisings confirm one of the major arguments of this dissertation: that British rule, through new land policies, created agrarian unrest and peasant revolt, even in the contemporary period. It just so happened that most of the peasants in Malabar happened to be Muslims.

First, the Mappilla uprisings do not fit into the rubric for most other communal conflagrations. For example, the uprisings were rural, occurring mainly in southern Malabar, especially the area around Ernad. Varshney and

Wilkinson, for example, find that 96 percent of Hindu-Muslim riots are urban.

The Malabar uprisings would be uncharacteristic if religion were at the core of the unrest. Furthermore, most Hindu-Muslim conflict – and religious conflicts in general – are not fought over land. They involve other issues – access to religious sites, competition over jobs, etc. In Jaipur, for example, one main cause of

45 The most famous incident was the 'Mappilla train tragedy', in which 70 Mappilla prisoners being transported in a sealed wagon died of suffocation and heat exhaustion. 46 See NAI, Home Department, Jails Branch, 1927, 293/1927

242 communal strife has been the socioeconomic backwardness of the Muslim population, but their grievances have little to do with land. But one of the central causes of the Mappilla uprisings were agrarian. In fact, the Mappilla Rebellion was, at first, strictly about agrarian issues – securing tenancy rights, the eviction of landlords, etc. Finally, the conflict involved the colonial government in a way which does not match most other colonial-era communal conflicts. In Ajmer, for example, when a Hindu-Muslim incident erupted the colonial government was only involved in stopping it. They were rarely, if ever, the targets of violence, and they tried to avoid siding with either party. Yet all of these things happened during the Mappilla uprisings – the British administration in Malabar was a main target of the rebels, and the government in turn very firmly cracked down on the

Mappillas. Therefore, the Mappilla uprisings did not unfold like most other communal incidents.

Second, as I noted above, at its core the unrest was agrarian in nature. This view is held by a number of scholars (Dhanagare 1977, Hardgrave Jr. 1977, Wood

1976, Panikkar 1989, Herring 2008). K.N. Panikkar summarizes the state of existing scholarship on the Mappilla uprisings by noting that they were “not intrinsically communal” (1989: 190). D.N. Dhanagare adds:

“...the communal sentiment or 'fanaticism' of the Moplahs was only the symptom and not the disease. The 1921 rebellion, like the uprisings that occurred throughout the nineteenth century, was in essence an expression of long-standing agrarian discontent, which was only intensified by the religious and ethnic identity of the Moplahs and by their political alienation...” (1977: 141)

Most of the Mappillas, especially those in south Malabar, were poor cultivators,

243 occupying the lowest rung on the rural hierarchy. Investigating the specific discontent of the Mappillas requires an analysis of the changing land tenure situation in Malabar with the onset of British rule at the end of the 18th century. I detail this at length in section 5.2, but I will offer a brief explanation here. When

Mysore invaded Kerala in the mid-18th century most of the high caste Hindu landholders in the region exiled themselves into the surrounding areas, including

Travancore. In the interim (i.e. in the period before the British took control of most of Malabar in 1792), Mysore's ruler Tipu Sultan forged linkages with the native Muslim population, who in turn increased their control of agriculture in the absence of the traditional Hindu landholders. The major problem arose after

1792 when the British came into possession of Malabar. At this time, the exiled

Hindus returned home and demanded back their old holdings. The British government sided with the Hindu landholders, and the Mappillas were dispossessed in the early 19th century. This act was the basis of most of the agrarian discontent which plagued the Mappilla community throughout the rest of the century. After the Hindu landholders were given private property rights,

British administrator M. Moberly noted that Malabar became “the most rack- rented place on the face of the earth” (Varghese 1970: 78). The British Resident in

Travancore, Atholl MacGregor, summarized it best when he concluded that: “As to the essential nature of the Malabar Mappilla outrages, I am perfectly satisfied that they are agrarian. Fanaticism is merely the instrument through which the

244 terrorism of the landed classes is aimed at”.47

Third, there was an underlying caste dimension to the conflict which has been unaddressed by most scholars. While almost all the proponents of the

Mappillas uprisings were Muslims, they were actually relatively recent low caste converts, mainly from the Tiyya, Cheruman (or Cherumar), and

Mukkuvan castes. Many of the new converts were also previously agrestic slaves.

For instance, in 1856 there were 187,758 Cherumar slaves but by 1881 there were only 64,725 (Dhanagare 1977: 123), a decline which indicates a massive number of conversions among this group to Islam. Or, as British administrator C.A. Innes put it bluntly, Mappillas came from “the dregs of the Hindu population”

(Hardgrave Jr. 1977: 62). Conrad Wood notes that:

“[by] becoming a Moplah a Cheruman labourer or Thiyyan verumpattomdar48 was joining a body that functioned as a self-defence organization of the rural subordinate whose ultimate weapon in this period was the outbreak. Conversion certainly curbed the field of power of the Namboodiri and Nair landlords...” (1974: 23).

Precisely because of the strict caste hierarchies which existed in British Malabar, low caste groups embraced Islam in the hopes of shedding their ritually low status. T.L. Strange noted the effects of mass conversion, writing that the

Mappillas in the southern Malabar region suffered from the “absence of any men of learning to instruct them in their new religion” (Hardgrave Jr. 1977: 60).

British administrators further noted in 1917 that Mappillas “have been cut off from the rest of the Muhammadan world by difference of language, inasmuch as

47 BL, IOR/V/26/312/12 : 1928 48 A cultivator – see section 5.2 for a discussion of the various agricultural groups in Malabar.

245 they know practically no Arabic and have no acquaintance whatever with Urdu and Persian. Their language is a dialect of Malayalam, the speech of the Hindu people of Malabar, with whom they have much in common.”49 Susan Bayly likewise describes the Mappillas as having become “an exclusive and militantly separatist Indian Muslim Society” (1989: 11). While low caste converts made up most of the Mappillas in southern Malabar, the Muslims of northern Malabar – who were not central to the various uprisings – had been converted much earlier and were originally from high castes.

While most of the rebels were low caste converts, most of the landlords in

Malabar were upper castes – Nambuduris and . Surveying the notes on various Mappilla outbreaks between 1836 and 1919, Wood notes that of 80 Hindu victims, 62 were members of high castes. He writes in addition:

“Malabar, and especially that part where outbreaks occurred, was throughout this period pre-eminently a land of the big Nambudri and the Nair official. Nearly all the big were in fact high-caste Hindus. In 1915 Collector Innes gave figures showing that the eighty-six biggest landlord families, owning many hundreds of thousands of acres and paying about a fifth of the total land revenue, were all high-caste Hindus except for two Moplahs, one Tiyyan, and one Goundan.” (1978: 129)

Therefore, the Mappilla uprisings can be re-conceptualized as a movement of low castes against high castes. The antagonism between castes was, like elsewhere in

British India, driven directly by new British land tenure policies which increased rural inequality and economic stratification. Prakash Karat also notes that during the period when Mappilla outrages became endemic, religion was not the only

49 NAI, Legislative Department, Legislative Branch Part A, October 1917, #197-200

246 prism through which rebels acted; rather, “[r]umblings of lower castes against upper caste domination became more and more evident” (1973: 41). These uprisings were unique in India, however, because most of the low castes had embraced Islam in an attempt to escape their downtrodden existence. The

Servants of India Society endorsed a similar view, writing in a pamphlet after the rebellion: “What right have these lovers of Hindoo nation and protectors of Vedic religion to cry over the Mopla rebellion when they did not try beforehand to...remove the cause of distance pollution which prepared the ground for

Mopla atrocities?” (Hardgrave Jr. 1977: 94, emphasis added).50

The obvious question here in re-framing the Mappilla uprisings as caste- based conflicts is whether or not this unique situation might apply to other communal incidents. For example, are Hindu-Muslim riots in Jaipur also actually an expression of underlying caste tensions? The reason this seems unlikely is because the Muslims and Hindus who have engaged frequently in conflicts in

Jaipur are fragmented by caste. The conflicts there, for instance, do not pit high caste Hindus against low caste Muslims. Rather, both the Hindu and Muslim populations come from a range of different castes. Furthermore, there was no large scale conversion of low castes to Islam in Jaipur – the Muslim population there were the descendants of Mughal invaders. Therefore, the situation among the Mappillas and their previously low caste background seems unique.

50 Hardgrave Jr. also, at the end of his piece, investigates why only a certain portion of the Mappilla population supported the rebellion. He notes that the main rebels were poor tenants who had previously been low castes (pg. 97).

247 Fourth, during the initial uprising against the British, Muslims were not the only rebels – low caste Hindus were also initially involved in the rebellion.

This is another important point which is often misunderstood or overlooked. For example, E.M.S. Namboodiripad was very cautious about diminishing the impact of religion. As K.K. Kusuman puts it:

“...Nambudiripad wanted 'the Marxists' to answer certain simple but relevant questions before asserting that the rebellion was only a 'revolt of the peasants of Malabar'. The questions are mainly about the involvement of Mappilas alone in the rebellion. [Nambudiripad wrote] 'The oppression and exploitation of the Jenmi and the officials are as bad for the Hindus peasants as for their Moplah comrades.'” (2003: 160-61)

But the argument that low caste Hindus were not involved in the rebellion is incorrect. While the majority of the rebels were Muslims, Hardgrave Jr. shows that “reports on incidents occurring during the first weeks of the rebellion frequently implicated Hindus in the rioting, and arrests often included

Hindus...” (1977: 82, emphasis added). In fact, the Nayars had been instrumental in trying to get the support of Mappillas against the high caste Nambudhris in the early stages of the rebellion. Namboodiripad was specifically concerned with the later period of the rebellion, when Hindus did abandon the movement, and after which the rebellion gained an explicitly communal complexion.

Furthermore, while most scholarship on this question has focused on agrarian conflict initiated by Muslims, low caste Hindus were not exempt from generating agrarian violence in Malabar. But because they lacked “...means of collective action, [they] in the nineteenth century largely turned to more individualistic

248 'weapons of the weak', largely theft” (Herring 2008: 287). In fact, British administrators who had read the damning reports of William Logan on Malabar noted that “[h]e described the cultivating classes 'as rapidly degenerating into a state of insolvent cottierism' and observed that crime was consequently on the increase”.51 His research on this area showed that between 1865-1880 – a period when most of the Malabar outrages occurred – Hindus were actually the largest group convicted of gang robbery (Dhanagare 1977: 126). This further suggests the prominence of intercaste struggles in Malabar, not communal tension.

Finally, several scholars have rightly emphasized the nationalistic aspect of the Mappilla Rebellion. The rebellion is often understood primarily as a large anti-colonial struggle; in fact, most British administrators themselves interpreted the rebellion this way, as a spontaneous uprising against the foundations of colonial power. The Congress movement had linked with the Khilafat movement for the specific purpose of providing a united front against colonialism. In the initial stages of the rebellion, one of the prime targets of the rebels were British administrators throughout Malabar. Many of the rebels raised Khilafat flags and openly heralded the end of British rule. In comparison to colonial-era communal violence in Jaipur and Travancore – which was almost always fought between two distinct ethnic communities – the Malabar Rebellion featured the British state firmly on one side of the conflict. This complicates the interpretation of the

Mappilla Rebellion as a communal conflict because the Hindu-Muslim divide was

51 British Library, IOR/V/26/312/12 : 1928

249 not the main divide of the conflict.

If the Malabar uprisings were not communal, however, what role did religion play exactly? In other words, why did agrarian unrest in Malabar take on a communal complexion? First, as I noted earlier, most of the low castes who converted to Islam did so in order to escape the ritual injustices of an uniquely oppressive caste system. By converting to Islam, they were able to protest the power of high caste landlords in ways which had previously been inaccessible to them. Second, becoming Muslim meant adhering to a larger ethnic identity, creating, as Ronald Herring puts it, an “advantage in collective action” (2008:

286). For example, mosques served as a type of coordination mechanism against the landlords, and the number of mosques in Malabar nearly doubled from 1831 to 1851 (Dhanagare 1977: 121). Another important and related question concerns why there were forced conversions of Hindus at the hands of the Mappillas.

However, it is unclear how many forced conversions actually occurred and how many were lurid fictional accounts sensationalized by the press. Dhanagare notes, for example, that “the total number of conversion cases...appears significantly small in relation to the size of the Hindu population in the disturbed areas” (1977: 140). Finally, attacks on Hindu temples were also frequent, but powerful landlords were generally the trustees of these temples, so attacks against them were also attacks against the visible symbols of landed elite power

(Guha 1999: 74).

While many traditional accounts have emphasized that British policies in

250 Malabar led to communal conflict, evidenced by the Mappilla uprisings in the 19th and 20th centuries, I argued here that this interpretation is lacking in several key respects. But there is another related point here: I have argued throughout this dissertation that the British pursued protectionist policies toward Muslims.

Therefore, it is important to note that initially, British policies in Malabar had in fact been broadly similar to those instituted throughout the rest of the provinces.

For instance, the colonial government had initially recruited a large number of

Muslims into the new local government after they came to control Malabar. In

1792-3 in Ernad (the main site of the future rebellion), for example, fully two- thirds of all revenue collectors were Mappillas (Wood 1976: 547). But this policy was reversed shortly after the return of the high caste Hindus, and the British government was forced to decide between placating either the Mappillas or the former landlords. The debate within the government raged for some time, and there were even those who suggested that the Mappillas retain control of the land

(Ibid.: 545). However, when confronted between supporting high caste Hindus or

Muslim cultivators, the government eventually sided with the former.

Finally, there is one last question which must be addressed: since many scholars have argued that Malabar was a communally-sensitive area during the colonial period because of these Mappilla uprisings, what explains why this area has become peaceful now?52 This would buck a very notable historical trend in

52 Varshney's survey in Calicut, which is admittedly only one city in the Malabar region, revealed that 83% of Hindus and Muslims eat together in social settings, 90% of them have their children play together, and 84% of them visit each other regularly (2002: 126-7).

251 India – areas that have been violent do not easily become peaceful. What then accounts for this unique change in Malabar?

Ashutosh Varshney has tackled this 'transformation' (2002: 143-148) in his work. He argues:

“The principal reasons have to do with the pre-existing social hierarchy of Hinduism, its grave caste injustices, and an emerging lower-caste mobilization. India’s freedom movement did not introduce mass politics to Malabar. Mass politics there had already taken the form of intra-Hindu struggles over caste, led by lower-caste organizations. Civil rights of the lower castes, especially with respect to pollution, temple entry, and greater access to educational institutions and government employment, were the key issues. Hindu-Muslim issues simply could not match the passions aroused by caste inequalities and injustice.” (2002: 144)

According to this perspective, the Mappilla outbreaks in the late 19th and early

20th centuries seem a brief (but bloody) communal interlude in a region otherwise plagued with a long history of caste struggle.

Furthermore, in the 1920s and 30s the Muslim community in Malabar underwent a major transformation. After the various Mappilla uprisings, Muslim leaders in the region (much like the Aligarh movement after the Rebellion of

1857) pushed for the modernization of the community. The Muslim Aikya

Sanghom, in the words of U. Mohammed, led a “Muslim renaissance” in Malabar

(2007: 152), especially through the opening of new educational institutions. By the dawn of independence, the Mappillas were quite different from the community which had taken part in rebellion only a few decades prior, and their socioeconomic position had improved over time.

On the whole, the Mappilla uprisings were a temporary attempt by low

252 castes to use religious conversion as another means by which to fight uniquely high levels of caste oppression. As in the rest of British India, then, the main axis of ethnic conflict in Malabar has always been caste.

4.5 British Policies on the Malabar Christians

The Mappillas were not the only religious minority group in Malabar – there was a small but important Christian minority in the region as well, and they merit brief consideration here. The Christians presented a difficult case for the

British because, on the one hand, their policy of religious neutrality applied as much to this community as it did to Hindus, Muslims, or any other religious group. On the other hand, the British were a Christian power, and many British administrators openly professed a desire to establish a firm base for Christianity on the subcontinent. On the whole, however, the Christians in Malabar were not given any kind of special treatment by the British government, and this only further emphasizes the strength of the British policy of religious neutrality in the provinces.

Overall then, British religious policies in Malabar cannot be said to have led to higher levels of communal tension. The Mappilla uprisings were a brief communal interlude in the region, but even they were a kind of representation of deeper caste antagonisms. By the dawn of independence, Hindu-Muslim tension in Malabar had largely been minimized, and communal conflict in the region today still mirrors this trend.

253 4.6 Post-Colonial Religious Reforms

In the post-colonial period Kerala was subjected to a number of reforms, but the position of minority religious groups has remained largely the same. First, it is important to recognize that Partition played little role in Kerala. The state was too far removed from north India to have seen major Muslim emigration to

Pakistan. In fact, as I noted previously, the percentage of Muslims in Malabar increased considerably from 1931-2001.

Kerala has the highest rates of social development in India, figures which are comparable to many Western nations. Religious minority groups like

Muslims and Christians have benefited from the rising tide of development that has occurred in the state since the 1950s. In fact, Kerala is one of only a small number of states that has implemented reservations for Muslims. However, as I noted earlier, the main cause of violence in Kerala is not socioeconomic – it is rooted instead in antagonism between different religious communities. Conflict in Travancore is understood to revolve around religion, a fact derived from the long history of communal tension in the region. While caste would eventually come to form the main axis of conflict in Malabar, the legacy of discriminatory religious policies instituted by the Travancore princely state persisted into the modern period. The riots which occur in that region today are merely the latest incarnation of pre-existing religious violence.

254 V. The Colonial Roots of Caste and Tribal Violence in Kerala

In this section I examine caste and tribal violence during the colonial period in north and south Kerala. I have already highlighted that in the northern

Malabar region the amount of peasant violence (Mappilla uprisings) was immense, although I delve into that topic in more breadth here. However, I show, conversely, that the princely state of Travancore enforced a number of protective policies for low castes and tribal groups, which served to limit the amount of contemporary violence in these areas.

Table 4.5 below highlights the major caste groups in both regions.

Table 4.5: Composition of Castes in Malabar and Travancore, 1901 Malabar Travancore Tiyans (Iluvans) – Nayars – 520,941 661,000 Nayars – 391,000 Tiyans – 491,774 Cherumans – 246,000 Pulayans – 206,503 Kammalans – Shanans – 155,846 > 100,000 Paraiyans, Kuravans, and Asaris – 50 to 100,000 Nambudhris – < 20,000 Vellalas, Brahmins, Marans, and Kollans – 20 to 50,000

In Travancore approximately 60% of the population made their living from the land, whereas in Malabar the figure was smaller due to a large proportion of the population working as fishermen in coastal areas. As Table 4.1 at the beginning of this chapter indicated, the percentage of high castes and 'polluting' castes was similar in both Malabar and Travancore. The potential for caste conflict was therefore available in both regions.

255 The two major adivasi groups in Kerala are Paniyas and Maratis. Most of the tribals are located in the Malabar region. Wayanad district has the highest proportion of tribals – in 1991, 35% of the total tribal population of the state lived there. The only district in the Travancore region with a significant percentage of tribals is Idukki (15% of the state total in 1991).

5.2 British Policies on Low Castes and Tribals in Malabar Province

Earlier in this chapter I highlighted that Malabar was the site of an immense amount of peasant unrest, primarily driven by low caste Muslim converts. The agrarian-based Mappilla uprisings were the main form of ethnic violence experienced by Malabar in the 19th and 20th centuries. While I have already focused on the specifics of these uprisings, in this section I delve further into the nature of the Malabar land tenure system and its changes over time, especially after the arrival of the British.

Malabar has come to be considered a notable area of agrarian unrest in contemporary India, but this was not always the case. In fact, T.K. Ravindran notes that “Malabar was...remarkably free from agrarian unrest in the pre-British period” (1978: 126). This is an extremely important point because, by all accounts, an extremely rigid pre-colonial caste system existed in Malabar, yet agrarian unrest was minimal in this period. It was only after the implementation of a series of policies by the British that caste antagonisms began to increase significantly in the region.

What kind of land tenure system existed in Malabar before the arrival of

256 the British? As was the case in the previous chapter, this is a question which cannot be answered definitively and is still a source of debate among historians.

However, a preponderance of the existing evidence suggests that one notable feature of the pre-colonial land tenure system in Malabar was the absence of private property. On the contrary, Dhanagare notes that “[t]he fundamental idea in the Malabar land system...seems to have been 'protection and supervision' rather then 'ownership' in the modern sense” (1977: 116).

There were three major groups in the agrarian system. First, there were the Jenmis, or landholders. They were almost always high caste Hindus, i.e.

Nambuduris, and they had hereditary control over the land. However, like

Rajputs in Rajputana, they did not actually cultivate the land themselves (as this was considered a degrading occupation), but instead turned it over to others for this purpose. Below the Jenmis were Kanamdars (also known as Kanakkaran), or land supervisors. They were mostly Nairs, were from the middle classes, and were responsible for seeking out cultivators for the land. The actual cultivators themselves were known as Verumpattamdars, and most of these were Mappillas or other low caste groups. The uniqueness of this model, known as Kana-Janma-

Maryada, lay in the fact that, as previously noted, the land was not held privately

– rather, the customary relationship between these three groups was collaborative. In fact, the net produce of the land was shared equally between all three (Dhanagare 1977: 116-7). William Logan summarized that “a kanakkaran was as much the proprietor of the soil as the janmi himself was in former days.

257 They were, in short, proprietors bound together in interest by admirable laws of custom” (Panikkar 1984: 174). He further stated:

“In order to understand the Malayali land tenures right...it is...first of all necessary to realise the fundamental idea that certain castes or classes were told off to the work of cultivation, and the land was made over to them in trust for that purpose, and in trust that the shares of produce due to the persons in authority should be faithfully surrendered.” (Somarajan 1983: 132)

The rise of the British in Malabar led to major disruptions in the existing tenurial model. The critical change occurred in 1801 when, after the fourth Anglo-

Mysore War and the return of exiled high caste Hindu landlords from Travancore and elsewhere, the government recognized Jenmis as having a proprietary right over the land. The Kanamdars and Verumpattamdars were subsequently relegated to mere tenants of the land, and therefore able to be displaced from it.

What had led the British to the conclusion that the Jenmis had proprietary rights? On the one hand, they may have actually believed (incorrectly) that this is how the Malayali land tenure model had always worked. Administrator William

Thackeray endorsed this view, noting in an 1807 report that “almost the whole of the land of Malabar, cultivated and uncultivated, is private property and held by

Janmam right which conveys full absolute property in the soil” (Panikkar 1984:

175). British administrators were of course also interpreting the Malabar system through the lens of the land tenure models which existed in Britain. Logan, arguing vehemently that the British had badly misunderstood and destroyed the existing landholding system in Malabar, noted along these lines:

“[t]he idea that the janmi is absolute proprietor of the soil, and as such entitled to

258 take as big a share of the produce as he can command...is founded on error, and has been brought about by engrafting on the customary Malayali law usages which spring from European ideas of property having a totally distinct origin and history.” (Ibid.: 176-7)

He went further:

“...[before the arrival of the British] the Janmi was simply a man exercising authority within a certain defined area, and entitled as to a well-defined share of the produce...of the land lying within that area. But in view of the new rights he secured, the Janmi became the lord of the soil.” (Somarajan 1983: 133)

On the other hand, it is possible that the British did not necessarily misunderstand Malabar's land tenure system – but rather, they had worked out a deal with the high caste Hindus who had returned from exile and demanded back their land. K.K.N. Kurup writes:

“The analysis of the revenue policy adopted by the Company in Malabar at the close of the eighteenth century reveals that it was a policy more or less borrowed from its previous experience under the 'Diwani' in Bengal. Even the administrative structure initiated by the Company consisting of a supervisor and the native revenue collectors who were the traditional Rajas in Malabar, was a part of the administrative plan worked out in Bengal.” (1976: 430)

In either case, the results of this new British policy were devastating for the peasantry. Aside from tearing apart the traditional agrarian model, the power of Jenmis increased steadily at the expense of other groups, thereby creating widespread caste antagonisms. Logan, for example, collected figures on the rise in evictions in the late 19th century, detailed in Table 4.6 below.

259 Table 4.6: Evictions in Malabar, 1862-8053 Average Annual Number Average Annual Number of Persons Period of Eviction Suits Against Whom Eviction Decreed 1862-66 2039 1891 1867-71 2547 3483 1872-76 3974 6286 1877-80 4983 8355 Source: Panikkar 1984: 180

Interestingly, despite the clear evidence of a rise in evictions, the Malabar

Tenancy Committee came to the surprising conclusion in 1928 that:

“As to the extent of unjustifiable evictions, the materials at the disposal of the Committee do not prove that evictions are either so numerous or so unjustifiable as has been represented to us by the advocates of the tenants' interest.”54

A second major disruption to the existing system came in the introduction of a whole bevy of new taxes. Pre-colonial Malabar was a highly unique territory in that taxes were, according to the writings of Dutch Commander Hendrik van

Reede in 1677, “unknown among the people of Malabar...the Nayars are unacquainted with taxes on house, head, land or cattle” (Kooiman 1988: 10).

Because power in pre-colonial Malabar was highly dispersed, kings and chiefs over time became unable to collect revenues from areas other than their own private lands. C.K. Kareem writes:

“Unlike in any other part of the country the rulers or Kings of Kerala had no right over the landed property held by the individuals. As a result of this, land revenue was unknown to the people of Kerala...From the very early days of the history of this country the land was owned by those who actually cultivated the land...There

53 Panikkar notes that the figures are probably too low throughout because many evictions occurred without the knowledge of colonial authorities, or landlords were able to evict tenants without invoking any laws. 54 BL, IOR/V/26/312/12 : 1928, pp. 76

260 was no authority to check the ryots from ploughing and cultivating the land they held. Unrestricted, they enjoyed the fruits of their labour.” (1977: 583)

However, the British were always in the pursuit of new sources of revenue, and therefore taxation came to exist for probably the first time in Malabar. These taxes, as usual, were heavily regressive. The Assistant Collector of Malabar noted in 1829:

“The fact is that the gardens of all Adhikaris and Menons and their relations are shamefully under-taxed. The same remark is applicable to those belonging to opulent individuals who were able to purchase the favour of the public servants. Whilst, on the contrary, the gardens of all the poorer classes are assessed so high that these poor creatures cannot pay the public revenue out of the pattom or proprietor's net rent without the sale of some of the household furniture.” (Panikkar 1984: 162)

At the risk of idealizing pre-colonial Malabar, it seems that it was the rare example of a territory in which land revenues did not exist and different groups in society equally shared the land. Only with the arrival of the British and a new series of policies did the situation change considerably.

As was the case in Ajmer-Merwara, later British administrators in Malabar tried their best to overcome initial mistakes and reverse wrong-headed policies.

Logan himself undertook this task, although with extremely limited success. His

'Malabar Compensation for Tenants' Improvement Act' of 1887, which tried to provide compensation for evicted tenants, proved ultimately to be ineffective.

Many landlords shuffled tenants for higher rents, a practice known as melcharths, and high caste landlords managed to prevent the implementation of a bill that banned this practice by the British government. The Madras

Government was also not convinced that those who tilled the soil needed any

261 legislation to secure their tenure (Hardgrave Jr. 1977: 64). This was in keeping with the general indifference of the Madras government to the cultivating classes.

D.A. Washbrook notes:

“Repeated attempts by Fort St. George to raise the position of untouchables...were rendered nugatory for its predominantly high-caste executive refused to enforce them: land set aside for panchama [Dalit] resettlement disappeared from the records and a High Court ruling of 1917 revealed that slavery still existed in Malabar, eighty-four years after it had been outlawed.” (1976: 40)

Alongside these changes to the existing land tenure system in Malabar was the increasing importance of the courts in providing a legal justification for new policies. The Jenmis were assisted time and time again in their actions against the peasants by court decrees, and this was a situation which had not previously existed. The courts were a new feature of British rule in the region, and their rulings tilted the power balance even further in the direction of the Jenmis.

Finally, as agrarian discontent increased in Malabar, low caste tenants and cultivators were not unaware of the comparative situation in the princely state of

Travancore. This kingdom had been much more progressive than its British neighbor in pursuing agrarian reforms (this is discussed below in section 5.3).

Prior to the Mappilla Rebellion in 1921, for example, when the Congress government had linked with the Khilafat movement and tried to pursue tenancy reform, many protesters noted that the Travancore princely government had already pursued reforms which had left Malabar behind.

Policies toward adivasis in Malabar were also, on the whole, very similar

262 to the policies instituted by the British in Ajmer-Merwara, and the same as those instituted generally throughout India. Most of these changes occurred through the machinery of the Madras Presidency which was in control of Malabar. But the administration in Madras was more liberal about tribal policy than many other governments throughout the rest of India – in fact, it fought several what it perceived to be injurious policies regarding adivasis that were handed down from the British central government. First, forests in Malabar, as in Ajmer, were considered to be officially the property of the state. One administrator, the

Deputy Collector of Nellore, wrote to the government in response that “...there were originally no Government forests in this country. Forests have always been of natural growth here; and so they have been enjoyed by the people” (Guha

2001: 215). However, the British government disregarded this and other similar objections. The result to adivasis was devastating; P.K. Mohanty notes the effects in Malabar:

“The Adivasis were dispossessed in Malabar....The colonization of the forests, now by the Indian government, intensified the displacement of Adivasis.” (2006: 178)

Moreover, the British also pursued Criminal Tribes legislation in Malabar.

Again, several voices within the Madras Presidency objected. Before the Criminal

Tribes Act of 1871 was passed, the Inspector General of Police in Madras wrote to the British government, arguing that there was no problem with criminality among tribes in the region (Radhakrishna 2001: 158). However, the central government again ignored these complaints and passed the legislation.

263 The result of these two policies was a gradual erosion in the state of adivasis in the region. In a post-colonial report from 1955, a Government of India administrator noted the many problems with structuring policies for Malabar's forests:

“There were some agrarian disturbances in the Koothali estate in Malabar in 1954 and the Madras Govt. asked Sri W.R.S Sathianathan at that time Member, Board of Revenue, Madras, to enquire [sic] into the the matter and submit a report. In his report on the situation...Sathianathan expressed the view that the agitation in that estate was only a symptom of a disease which was spreading rapidly over the whole of Malabar Dist. and added that it was necessary for the Govt. to adopt a bold and comprehensive policy for the treatment of private forests and unoccupied private and Govt. lands...”55

Scarcity of land is more of a problem for tribals in Travancore than

Malabar. As I noted previously, landlessness is highest in this latter region.

Various British policies which dispossessed the adivasis and opened up forests to a host of new groups had all combined to place tribal groups in the region in a severely depressed state.

Overall then, British policies in Malabar undermined the position of low castes and adivasis. New landholding policies completely dispossessed peasants throughout the region, led to skyrocketing eviction rates, and sharply increased caste antagonisms. Tribals were also cut off from forests and dispossessed. The roots of caste and tribal conflict in contemporary Malabar can therefore be traced back directly to the arrival of British administrators in the late 18th century.

55 KSA, Revenue Department, 16716/F11/56/RD, 09-05-1955, B-1175

264 5.3 Travancore Policies on Low Castes and Tribals

The land tenure system in Travancore differed considerably from that of

Malabar. In fact, it was very different from even the previous princely state case study of Jaipur. In Jaipur, and in many of the Rajput states in general, princes were limited in their control of the land by jagirdars, or hereditary nobles. In

Jaipur State specifically, the raja actually controlled less than half the land in the kingdom. In Travancore, however, the maharajas controlled a much, much larger portion of the land – by most estimates, almost 75% of the cultivable land

(excluding waste lands and forest areas) was controlled by the Travancore royal family (Somarajan 1983: 131).

Land in Travancore was basically broken up into two overarching categories: private land was called puravaka whereas darbari land was called pandaravaka. As Table 4.7 details, the princely administration had a large amount of power and discretion over land policy in Travancore because it was the supreme landholder in the state.

Table 4.7: Land Tenure System in Travancore, 19th Century Percentage of Total Land Landholder Held State (Pandaravaka) 75.13 Temples 10.22 Jenmis 8.10 Tax free landholders (Karamozhivu) 4.64 Non-Brahmin landlords (Madambimar vaka) 1.05 Maharaja's private lands 0.81 Rani of Attingal's private lands 0.05 Source: Assary 2009: 12

265 In the mid-19th century, as the first spasms of agrarian revolt began to take hold in Malabar, Travancore embarked upon a series of important land reforms.

First, in 1818 the government instituted a ryotwari settlement, although by this point the ryots were still tenants-at-will. Despite this designation, their existence was much better than the peasants of Malabar. Scottish chronicler of India

Francis Buchanan-Hamilton even described Travancore's peasantry as

“contented and prosperous” (Assary 2009: 10). In 1826 and 1829 laws were enacted which prevented the arbitrary eviction of tenants from land. A few decades later the princely administration enacted the Pattam Proclamation of

1865, which C.N. Somarajan calls the “Magna Carta of Travancore ryots” (1983:

133). This created land tenure rights for those who inhabited pandaravaka lands, and a proclamation two years later established the same system for puravaka lands.

Furthermore, the land in Travancore was not concentrated in the hands of large landlords like in Malabar. Rather, land ownership was much more diffuse.

George Kristoffel Lieten notes:

“A notable feature of the size of distribution were the regional differences. The Northern Malabar districts had very huge estates while the Travancore region was mostly characterised by its small peasant proprietors. This is evidenced by the Census of Land Holdings and Cultivation, 1956: while the number of owned holdings of over five acres constituted in Travancore-Cochin only 5% of the total, in the three Malabar districts this percentage went up to more than 30% comprising almost 85% of the land.” (1975: 455)

A British administrator in a 1928 report further noted that:

“The Indian States of Travancore and Cochin, where conditions are similar to those in Malabar, have solved the tenancy problem, so far as it relates to

266 kanamdars.”56

Aside from its progressive land policies, Travancore State also instituted other policies which assisted the improvement of low caste groups. The state was, as I noted at the beginning of this chapter, considered to be one of the most advanced princely states in the union. There were indeed few states who were more progressive on issues dealing with low castes and tribals in India. Most notably, the rulers of Travancore in the 19th and 20th centuries enacted several pathbreaking policies which attempted to improve the condition of Dalits.

In the late 1850s there was a major issue regarding women of the Shanar caste. The tradition in Kerala had been that Shanar women were not allowed to cover their upper bodies in the presence of Brahmins. In 1859 the raja of

Travancore reversed this policy. The diwan, Madhava Rao, wrote to the British

Resident W. Cullen:

“His Highness now proposes to abolish all rules prohibiting the covering of the upper parts of the persons of Shanar women and to grant them perfect liberty to meet the requirements of decency any way they may deem proper with the simple restriction that they do not imitate the same mode of dress that appertains to the higher castes.”57

To be clear: it was not the British who pushed for this policy. On the contrary,

F.N. Maltby, a later Resident at Travancore and Cochin, wrote on this issue:

“On the whole, the opinion which I have formed, and I trust the Government will consider it a just one, is that the question is one in which it is not desirable, or in any way necessary, that the British Government should put any pressure upon the Native State...”58

56 BL, IOR/V/26/312/12 : 1928 57 NAI, Madras Government, Political Department, 1859, #347 58 NAI, Madras Government, Political Department (Non-Confidential), 1862, #284

267 Furthermore, the British did not seem inclined to enforce progressive caste issues in Travancore. Only a few years earlier, in 1853, they recommended to the princely administration that they abolish a festival held specifically for low castes, arguing that it was “endangering the Public Peace”59

Beside the well-known Shanar cloth issue there were other smaller changes which the rajas of Travancore also instituted in order to improve the condition of Dalits. Maltby noted some of the progress in an April 1862 report sent to the Madras government:

“Two important measures...have been commenced, tending to break down caste distinctions. His Highness the Rajah has signified to me his determination to build suitable Courthouses for his Courts instead of their being held, as is now often the case, within religious precincts to which the lower classes are not admissable [sic], and also of directing the preparation of a form of address to the Courts, proportioned to the rank of the Court addressed and not to the grade of the addresser. At present the Christian Converts have just cause to complain that they are compelled to use a form of address expressive of low origin and degraded position.”60

However, the major step taken by the Travancore state for the advancement of low caste groups was the Temple Entry Proclamation of 1936.

This became a famous and seminal event throughout the rest of India. It was issued in November of that year by maharaja Chithira Thirunal, although the effort was pushed aggressively by his diwan C.P. Ramaswami Iyer. The text of the famous proclamation read:

“...We have decided and thereby declare, ordain and command that, subject to such rules and conditions as may be laid down and imposed by Us for preserving their proper atmosphere and maintaining their rituals and observances, there should henceforth be no restriction placed on any Hindu by birth or religion

59 KSA, Cover Files (1728-1903), 15540, 1853, #25 60 NAI, Madras Government, Political Department (Non-Confidential), 1862, #284

268 on entering or worshipping at the temples controlled by Us and Our Government.”61

This act, in the words of A. Sreedhara Menon, “ushered in a bloodless social revolution” (2001: 27). There was no area of British India, for example, which had done anything similar – in fact, only after the proclamation in Travancore did some other British territories follow suit and pursue similar legislation.

Malabar did not completely open up its temples to untouchables until after independence. In fact, many low caste reforms which occurred in Travancore were opposed by the British (Singh 2010: 285). As Robin Jeffrey has noted, for

British administrators the ideal prince was “progressive but not too progressive”62, and the Travancore government was often accused of being too progressive on low caste issues.

In one way the Temple Entry Proclamation was in keeping with

Travancore's reputation as a theocratic state. Ramaswami Iyer and the maharaja understood that pressure from Christian groups and the possibility of mass low caste conversions both threatened the character of the Hindu state. Rather than lose these groups, Travancore struck first and brought them into the Hindu fold.

In fact, upper caste groups were instrumental in urging the government to open schools for low castes, afraid that missionaries would use their schools for religious conversions (Desai 2005: 472). Scholars have rightly noted that the

Hindus of Travancore were not a unified community, but growing attempts to

61 Quoted in Menon 2001: 27 62 Quoted in Sundar 1997: 191

269 proselytize among low castes began to create the emergence of a more unified

Hindu bloc. It is also important to recognize the importance of Christian missionaries in pushing for ameliorative policies for low castes, albeit somewhat inadvertently. Missionaries in Travancore were active in converting low caste groups to Christianity, and, in the words of Manali Desai, “even stimulated caste insurgency” (2005: 463). The Travancore princely administration, partly in response to this practice, issued the Temple Proclamation. However, I do not mean to diminish the importance of the Travancore State, in whose hands laid the ultimate authority to decide how to respond to low caste mobilization. A.

Sreedhara Menon notes that Travancore “produced in the 19th century a galaxy of social reformers who championed the cause of the eradication of untouchability...

[such as] Chattampi Swamikal...Sri Narayana Guru...Kumaran Asan...T.K.

Madhavan...and Ayyankali” (2001: 28). The decision to push for the betterment of low caste groups came ultimately from the princely state itself.

In Travancore State, as was the case in Jaipur, adivasis also occupied a completely different position in society than their counterparts in British

Malabar. Most importantly, many of the seminal land reforms initiated by the

Travancore princely administration in the mid-19th century had adivasis as a primary target. For example, adivasis were given rights over specified tribal lands, and these areas were not open to non-tribal immigration. Moneylenders were also banned from these regions (Mohanty 2006: 178).

Christoph von Fürer-Haimendorf also cited Travancore as an example of a

270 princely state with progressive policies on adivasis:

“In the former princely state of Travancore the rights of such tribes to cultivate free of tax in forest areas was clearly laid down and compact blocks of land of an acreage seven times of that required in any one cultivating season enabled them to continue cultivation permanently in a cycle of rotation.” (1985: 27)

Overall, the distinctions between Malabar and Travancore in terms of their policies on low castes and adivasis is stark. In Malabar, British administrators, much like they did in Ajmer, pursued a host of policies which led to the deterioration of the peasantry. Perhaps the most important new policy was the recognition of Jenmis as lords of the soil. This singular act dislocated tenants and cultivators from the traditional landholding system which was based on land supervision rather than ownership. The result was a sharp rise in the number of evictions, as well as high levels of caste antagonism. British policies on adivasis were no better – the forests of the Malabar region were taken over by the state while tribal groups were dispossessed and some even criminalized. On the other hand, Travancore State pursued roughly the opposite set of policies. A number of reforms in the late 19th and early 20th centuries radically altered the position of low castes, who hitherto had languished in a backward state. Foremost among these new policies was the Temple Entry Proclamation, which opened up Hindu temples to untouchables. Adivasis also benefited from progressive policies, and

Travancore prevented them from losing access to forests or cultivable land. The contemporary pattern of caste and tribal violence in Kerala, therefore, is a direct result of the comparative colonial-ear treatment of low castes and adivasis in

271 British Malabar and princely Travancore.

5.4 Post-Colonial Caste and Tribal Reforms

Kerala has been perhaps one of the most successful states in India in terms of establishing post-colonial land reforms (Herring 1983). The 'land to the tiller' model (or a close approximation) implemented in Kerala has been much more progressive than the reforms pursued by other states in post-independence India.

However, there is still a clear divergence today in the socioeconomic position of low castes and adivasis in Malabar and Travancore. These groups in Malabar lag significantly behind in their development, as I highlighted using qualitative interview evidence earlier, and this has been the cause of a long history of caste and tribal violence in the region.

By the time India achieved independence and by the time they were preparing to linguistically reorganize states, the gulf between agrarian conditions in Malabar and Travancore were immense. In 1956, once the two areas were joined together, the new Kerala state's Planning Department undertook a comparative analysis (which also included the former princely state of Cochin).

The major findings are summarized in Table 4.8 below.63

63 KSA, Confidential (1954-6), 36394, 1956, #960

272 Table 4.8: Agricultural Statistics for Kerala, 1956 Malabar Travancore-Cochin Proportion of cultivated area 45.20% 48.20% to total area Proportion of irrigated area 2.60% 29.20% to total area Average paddy yield per acre 1092 lbs. of dry paddy 1754 lbs. of dry paddy Proportion of villages electrified 6.50% 26.00% Power generation – Nil 810 kilowatts per lakh of population Mileage of roads .38 Mile .70 Mile per square mile of area

The data highlight the deleterious effects of British rule on the Malabar region. By

1956 Malabar lagged behind Travancore-Cochin on almost every single agricultural indicator. The British, through their various policies, had seriously retarded the development of the region. This is important because, as I have noted in reference to the post-independence period, implementing reforms is difficult. It takes an enormous amount of cohesion and political will to initiate reforms in general, or in this case, reforms of the agricultural sector. When post- independence governments in Kerala undertook reforms, Malabar was in a much less privileged position than neighboring Travancore. This created a natural impediment to success.

Similarly, Table 4.9 highlights that in Malabar, reforms would require much more work to empower peasants and turn them from cultivators into owners.

273 Table 4.9: Agricultural System in Kerala, 1957-8 Tenant Status Malabar Travancore Landless laborer 12% 13% Owner-cultivator 10% 56% Tenant-cultivator 75% 25% Rentier 3% 1% Other N/A 5% Source: Nossiter 1982: 16

The entrenched power of landlords was stronger in Malabar, which resulted in this area facing much steeper hurdles in the implementation of land reforms. In Travancore, on the other hand, the position of low castes and tribals underwent a sea change in the 19th and early 20th centuries, and they were already significantly ahead of their Malabar counterparts by independence. Even as late as 1988, B.A. Prakash sadly surveyed the situation of Malabar thusly:

“Though Malabar had been a major exporter of a wide variety of agricultural products to Europe for more than two thousand years and consequently exposed to influences from abroad, it still remains an underdeveloped region with a backward agricultural sector.” (1988: 51)

VI. Conclusion

This chapter examined a second pair of case studies from the south Indian state of Kerala. I compared the Malabar region of north Kerala, which came under the direct rule of British administrators, with the Travancore region of south Kerala, which remained under the control of a Hindu dynasty. I highlighted that today, Malabar experiences high levels of caste and tribal conflict whereas

Travancore experiences heightened religious violence.

I then argued that this pattern of ethnic violence is a product of historical

274 forces stemming from the colonial period. The reason why contemporary

Travancore experiences periodic explosions of communal violence is because of the long legacy of religious marginalization suffered by the Muslim and Christian communities of the region. The northern Malabar area, on the other hand, has become a hotbed of caste and tribal conflict because British policies during the colonial period decimated the condition of low castes and adivasis. Colonial rule embedded different narratives of violence in these two regions, and politics in the north today revolves around caste while religion dominates in the south.

The pattern of ethnic violence which Kerala evidences is therefore similar to that of Rajasthan. Contemporary Jaipur experiences higher levels of communalism whereas contemporary Ajmer experiences higher levels of caste conflict. I showed that this was also due to the legacies of colonial rule: the princely state of Jaipur discriminated against the Muslim minority but protected low castes and adivasis. On the other hand, the British administration in Ajmer targeted their discrimination toward low castes and adivasis while protecting

Muslims. It is worth noting that despite the sizable historical, cultural, and linguistic differences between Jaipur and Travancore, and between Malabar and

Ajmer, these controlled comparisons showcase a strong similarity in their pattern of ethnic violence. Princely rulers and British administrators in both regions implemented extraordinarily similar ethnic policies.

The next chapter turns to two deviant cases from eastern India: the princely states of Bastar and Hyderabad. These cases are located in the 'Red

275 Corridor' – an area rife with Naxalite (low caste, tribal) violence. While Chapter 2 highlighted that British rule correlates very strongly with casualties from the contemporary Naxalite insurgency, Bastar and Hyderabad are two princely state outliers which experience high levels of contemporary conflict – in fact, Bastar experiences the highest levels of Naxalite violence in all of India. The next chapter examines the root causes of why these two cases uniquely experience such high levels of caste and tribal bloodshed.

276 Chapter 5 Ethnic Violence in the Red Corridor: Bastar and Hyderabad

“The Pax Britannica is so firmly established that the idea of overt rebellion is always distant from our minds, even in a remote State like Bastar.” – B.P. Standen, Chief Secretary to the Chief Commissioner, Central Provinces, 19101

“The feudal elements are taking shelter and surreptitiously managing their affairs. They are waiting for an opportunity. They feel that once the Communists are crushed, they would go back and continue as before. They have not changed a bit. They are the same hard nuts.” – Swami Ramanand Tirtha, Hyderabad State Congressman, during a tour of Telangana villages, 19512

“In the same manner our Executive Council should also consist [proportionally] of 85% Hindus. If competent Hindus are not available, incompetent ones should be appointed. The next best alternative would be to appoint competent Muslims.” – Letter from Hindu protesters to the nizam of Hyderabad, October 24th 19313

In Chapter 2 I presented quantitative evidence that former British provinces experience higher levels of contemporary caste and tribal conflict than former princely states. I used the Naxalite insurgency as a proxy for this kind of conflict because the Maoist movement is driven primarily by Scheduled Tribes and, to a lesser extent, Scheduled Castes. This chapter examines two deviant cases4: two former princely states which experience immense levels of

1 NAI, Foreign Department, Secret-1, 1911, #34-40, pp. 4 2 NAI, Ministry of States, Hyderabad Branch, 1951, #1(16)-H/51, pp. 7 3 APSA, Installment #47, List #10, File #82 4 It may reasonably be asked why a deviant case of the other variety (i.e. a British province with high levels of religious violence) is not being studied. Certainly there are major British areas where communal conflict occur. However, the reason why I chose to study Bastar and Hyderabad is because the Naxalite insurgency is so strongly correlated to British rule, and these states, especially the former, are big outliers. For example, Bastar accounts for 1/3rd of all Naxalite casualties from 2005-9. It is the single deadliest area in India for Naxal violence despite the fact that it has a population of just over 2 million. The deadliest city in India for communal violence is a British city, Bombay, but it has a population of 16 million (it is the

277 contemporary caste and tribal conflict.5 Bastar (located in present-day

Chhattisgarh) and Hyderabad (located mostly in present-day Andhra Pradesh) are today centers of the Naxalite insurgency. In fact, Bastar is the single deadliest

Naxalite battleground in contemporary India, while Hyderabad was the site of the famed Telangana6 Rebellion of 1946-51, one of the largest peasant revolts in

Indian history. This chapter attempts to explain why these two cases uniquely experience so much violence. While Chapter 3 looked at the district level and

Chapter 4 the state level, this chapter looks at the regional level by examining two different states within the 'Red Corridor' of eastern India.

I argue in this chapter that Bastar is the 'exception that proves the rule' – it was a princely state which, by the mid-19th century, had come almost completely under the control of the British. I will demonstrate that this level of colonial interference in Bastar was uncommonly aggressive compared to most other princely states, and I will highlight that tribal revolt intensified in the kingdom only after the arrival of British administrators. In Hyderabad, I show that the revolt began in the Telangana region against the local landed elite, called the dora, who had been given proprietary rights over the land much like the zamindars of British India. Furthermore, I highlight that both the rulers of

Bastar and the nizam of Hyderabad, much like the rajas of Jaipur and

biggest city in India). In per-capita terms, Bastar is an unparalleled outlier. Hyderabad must also be studied because the peasant revolt there was one of the largest in the . 5 The conflict in both of these states is primarily tribal conflict, so I focus primarily on the treatment of adivasis during the colonial period. Low castes are/were also involved in these conflicts, but to a lesser extent. 6 Sometimes also spelled Telengana or, less often, Telingana.

278 Travancore, pursued protectionist policies for low castes and, especially, adivasis.

In the ensuing section I first introduce and situate the two case studies.

Using WITS figures and qualitative interview evidence I then showcase in section

II the extremely high levels of contemporary caste and tribal violence in these regions. In section III I then use archival evidence to investigate the colonial roots of violence. I highlight that the British were the cause of the violence in

Bastar, and that the dora were the original focus of the . I also argue that more than any other pair of case studies in this dissertation,

Bastar and Hyderabad highlight the failure of the Indian government to reform its colonial past, and I demonstrate this at length. In section IV I then briefly discuss the state of religious violence in these two regions and its historical causes, attempting to determine whether these two cases also fit into the pattern of religious violence established throughout this dissertation.

I. The Case Studies

Bastar was a princely kingdom located in what is now the state of

Chhattisgarh. During the British period Bastar was over 13,000 square miles, or roughly the size of Belgium. It had a population (1901) of 306,501. It was, according to the British, one of a number of 'jungle kingdoms' – so named because of its lush, dense forests and its large population of adivasis. The state was unique in that the majority of its population were tribals, specifically Gonds.

279 The rulers of the state, however, were not tribals7 – they were Rajputs who fled

Warangal from Muslim invaders some time in the 14th century. The local tribes consider the raja of Bastar to be the high priest of Danteshwari, a local goddess; the power of the raja is embodied in the Sword of Danteshwari, which is held by the royal family. The princely state was known for its unique celebration of the

Dasera festival. The raja is 'abducted' by tribals on the 11th day of Dasera and then returned to the throne the next day, a ritual which symbolizes the historically close linkage between the adivasis and their king (Gell 1997).

During the pre-colonial period Bastar was a part of the Mughal and then

Maratha Empire. Due to its geographical inaccessibility, however, it always retained a high level of isolation. The state was heavily forested – for example, according to 2001 figures, 54.77% of is covered by forest. Overall, this region is one of the most heavily-forested areas in India – it is part of the well known forest region.

When the British finally broke the power of the Marathas in central India in 1818 they subsequently entered into a political relationship with Bastar for the first time. In 1865 the kingdom was designated a Feudatory State, essentially meaning it was relegated to a lesser position than most other princely states of similar size. In fact, the raja was not entitled to any gun salute. In 1922 the Chief of Bastar requested a dynastic salute of 9 guns but was refused.8 Bastar was governed as part of the Central Provinces Administration. The state now

7 There were a number of tribal kingdoms with tribal rulers in Orissa, for example. 8 BL, IOR/R/1/4/26, 1922

280 comprises the present districts of Bastar, Dantewada (South Bastar) and Kanker

(North Bastar).

Hyderabad State was originally part of the Mughal Empire, brought under control by the Emperor Aurangzeb during his 17th century military campaigns in the Deccan. As imperial power began to fracture a century later the Mughal governor of the region Asaf Jah established his own kingdom in 1724. He styled himself the Nizam-ul-Mulk (Ruler of the Realm). Though the Jah Dynasty still continued for decades to claim itself merely a client state of the Mughals – even continuing to send tributes to Delhi – it had essentially become an autonomous kingdom. It was the largest princely state in India, roughly the size of France

(82,698 miles), with a population (1901) of 11,141,142. Gonds and Bhils were the major tribes in the state. Hyderabad enjoyed a 21-gun salute (the highest) and, unlike almost all other princely states, utilized its own currency and had its own postal system.

Hyderabad was governed by a Muslim dynasty but the vast majority of the population was Hindu. Unlike Bastar, after 1858 Hyderabad retained a high level of internal autonomy. This originally was a reward for its loyalty during the

Rebellion of 1857. Administrator John Bruce Norton noted about the relative peace in Southern India during that time: “...if Hyderabad had risen we could not escape insurrection practically over the whole of Deccan and Southern India”

(Majumdar 1951: 646). Before the rebellion, Hyderabad's Resident General

Fraser had written to Lord Dalhousie suggesting more active intervention in the

281 internal affairs of the state, which he believed to be hopelessly corrupt. After the rebellion, however, the British tone changed considerably. They signed a fresh treaty with the nizam, gave him over £13,000 worth of presents, canceled 50 lakhs of his debt, and in 1861 presented him with the Order of the Star, the highest recognition for princes in India (Regani 1963: 322-3). In 1918 he was also uniquely styled His Exalted Highness (H.E.H.) and declared a “Faithful Ally of the British Government”.9

In 1926, after continual displeasure with nizam Osman Ali Khan, the

British did attempt a set of serious political reforms in Hyderabad. They forced the appointment of British officers to several key posts in the nizam's government, including within the Revenue and Police departments. But even after these reforms British administrators were frustrated with their overall effect. The 1946 Hyderabad Political Notebook (Top Secret) noted:

“[The 1926 Reforms], however, by no means ended the struggle between the Government of India and the Nizam to curb the latter's despotic tendencies. On the contrary, the Nizam was always on the watch to gain complete control again and thus he set the ball rolling once more in the beginning of 1927 when he proposed to constitute the new [Executive] Council as to leave it weak and merely an instrument to carry out his behests...”10

Furthermore, the nizam often paid lip service to British requests by passing official firmans (royal edicts) but not enforcing them. For example, the British demanded the creation of local councils in Hyderabad, but (just like the

Executive Council) the nizam “so deprived them of power and resources that they 9 BL, IOR/R/1/5/66: 1946, pp. 1; contrast this with Warren Hastings' remark in the 1780s on nizam Asaf Jah III, who was not “at any period of his life distinguished for personal courage or the spirit of enterprise” (Forrest 1910: 58) 10 BL, IOR/R/1/5/66: 1946, pp. 21

282 were politically ineffective” (Elliott 1974: 31). One Resident, Sir Arthur Lothian, as late as the mid-1940s complained that “there has been no real change of heart on the part of the Nizam”11 and noted that he had a “proclivity for short-circuiting the Resident...H.E.H. is still at his old tricks...”.12

The Telangana region, which is the main area where Naxal conflict occurs today, formed about fifty percent of the area of Hyderabad State; it was located in the northern, Telugu-speaking region bordering , Chhattisgarh, and

Orissa. It comprises the modern districts of Adilabad, Hyderabad, Khammam,

Karimnagar, Mahbubnagar, Medak, , Nizamabad, Rangareddy, and

Warangal. Map 6 below is a colonial-era map highlighting Bastar and Hyderabad and their surroundings.

11 Ibid., pp. 103 12 BL, IOR/R/1/1/4203: 1944

283 Map 6: Bastar and Hyderabad State, 190313

II. Caste and Tribal Violence in the Red Corridor

Bastar and the Telangana region of Hyderabad today are two of the major centers of the Naxalite insurgency. Map 7 below showcases the so-called 'Red

Corridor' of Naxalite violence in India, and both Chhattisgarh and Andhra

Pradesh are located right in the heart of the bloodshed, comprising many of the

“Severely Affected Districts” designated by the Indian government.

13 Bastar is slightly difficult to see; it is located just to the northeast of Hyderabad State (it is also yellow), within the Central Provinces.

284 Map 7: The Red Corridor, 2007

In the next two sections I present qualitative interview evidence in order to provide general information about the conflict in both regions.

2.2 Caste and Tribal Violence in Contemporary Bastar and Hyderabad: Qualitative Evidence

Dantewada is today the deadliest battleground for Naxal violence in India.

Using figures from the WITS database, Dantewada from 2005-9 experienced a staggering 516 deaths and 472 injuries in Naxalite violence. During this period the three districts which comprise the former Bastar State (which have a total population of only slightly over 2 million people) experienced 1171 casualties14

14 This is roughly half the casualties experienced by Bombay from Hindu-Muslim violence over

285 from Maoist violence, or 35% of the total number of Maoist casualties in India.

In April of 2010 Dantewada was the site of one of the deadliest Naxal attacks in recent memory – 76 members of the Indian police were killed in an ambush.15

To examine the contemporary violence in the region, as well as gain insight about the proximate causes of conflict, I utilized a number of interviews with respondents, primarily those located in Raipur, the capital of Chhattisgarh. First, respondents were asked about the composition of the Naxalites. A local newspaper editor summarized the situation as follows:

“Only a small number of tribals are Naxalites, but the Naxalites are almost all tribals. They are mostly Scheduled Tribes. Scheduled Castes might be involved, but they are not the major part of the movement – it is mostly tribals. These groups have provided a fine hunting ground for the Naxals.”16

There are no precise figures on the composition of the various Naxalite groups in the region. A Hindustan Times piece from June 25th 2005 on Chhattisgarh noted that tribal groups are “considered the backbone of the ultra-left movement”.17

Another respondent from a Chhattisgarh government body dedicated to the uplift of adivasis also confirmed this:

“They are all STs, but you must know that the leaders of the movement are not actually from this region. They come from Andhra Pradesh, Bengal, or Delhi.”18

Respondents were also asked about the causes of contemporary violence.

The aforementioned newspaper editor noted:

the period 1950-1995. 15 “75 security personnel killed in Chhattisgarh Maoist ambush” Times of India, April 6th 2010 – Note: the number killed was later revised upward to 76 16 Interview in Raipur, April 19th 2011 17 “Naxalites Meet to Analyse Tribal Revolt Against Them” – DPL, 24, IB, 210 18 Interview A in Raipur, April 12th 2011

286 “The points that the Naxalites make are valid. They talk about underdevelopment and poverty, and that is all valid. Everyone knows that development [projects have] filled the coffers of big business and the government.”19

A local Congress Party politician noted a similar view:

“It is about revenge. The main cause of the violence is that tribals are seeking revenge for the way they have been treated historically. Many historic injustices have occurred here in Chhattisgarh.”20

A Forest Department respondent also commented:

“Underdevelopment is the main cause of the violence. If you go to Bastar, if you go talk to the police, you will see that. The adivasis here are quite backward economically.”21

An academic who taught at a local agricultural college also remarked:

“The Naxalites are mainly fighting against the police and the Forest Department. Those are the two main groups that they really hate. They want to control the forests without interference.”22

The Telangana region of Hyderabad has also become a major site of

Naxalite violence in contemporary India.23 Using figures provided by the WITS dataset, Telangana districts experienced 183 casualties from the Naxalite conflict from 2004-9. While these figures are relatively small – especially compared to

Bastar – since the 1940s this area has been a continual site of Maoist violence.

Furthermore, Andhra Pradesh has, as an aforementioned interview respondent from Raipur noted, provided much of the leadership of the Naxalite movement, and the state has been a gateway for Naxal violence throughout the entire region.

19 Interview in Raipur, April 19th 2011 20 Interview C in Raipur, April 12th 2011 21 Interview in Raipur, April 11th 22 Interview B in Raipur, April 12th 2011 23 Telangana has also been the site of a major separatist movement. For more on this, see and Sharma 1979.

287 Several interview respondents in Hyderabad were questioned about the

Naxalite violence in the region. One respondent at the Andhra Pradesh Tribal

Commission noted about its causes:

“The situation for tribals here in Andhra Pradesh has definitely improved – it is much better than it was just a few decades ago. You can look at literacy rates to see that. The biggest remaining problem for tribals in this state is unemployment and land restrictions. The amount of land that they have is the same as decades ago, but their population has increased. So you have a lot of population pressures on a static amount of land.”24

The respondent further commented:

“The areas that border Chhattisgarh are the worst-hit areas for violence. The tribals that join the movement have no income, no livelihood, and they embrace the communist ideology.”

Another high-ranking official in the Tribal Commission stated about the history of violence in the region:

“The British were awful for tribals in this area.25 And really, the nizam didn't care about the adivasis very much prior to the 1940s. After that, he employed [Austrian ethnographer Christoph von Fürer-] Haimendorf26 and started to change his policies toward tribals in the area.”27

The first official noted along similar lines:

“Where the princes ruled there were better conditions for tribals. The nizam was genuinely concerned about tribals in his state.”28

Overall, the quantitative figures and qualitative evidence compiled from elite interviews in both case studies highlights that the primary cause of the immense tribal and caste conflict in Bastar and Hyderabad today is the

24 Interview in Hyderabad, February 15th 2011 25 The respondent was talking specifically about the areas of Andhra Pradesh – mainly the coastal regions – which were ceded by the nizam to the British. 26 Haimendorf will be discussed in further detail in section 3.4 27 Interview in Hyderabad, February 18th 2011 28 Interview in Hyderabad, February 15th 2011

288 socioeconomic backwardness of adivasis in the region. This has led to an immense amount of 'bottom-up' armed conflict, initiated by STs and SCs themselves. But because both regions were princely states during the colonial period this is a surprising result given that most rajas advanced protectionist policies for tribal groups, thereby preventing the occurrence of violence. What makes Bastar and Telangana unique among princely states? In the next section I delve into the colonial history of both regions to address this central question.

III. The Colonial Roots of Caste and Tribal Violence in the Red Corridor

Like all of the other contemporary ethnic conflicts discussed in this dissertation, the origins of the immense caste and tribal violence in Chhattisgarh and Andhra Pradesh lie in the colonial period. Both of these territories experienced a long history of caste antagonism and tribal revolt during colonialism. In the ensuing section I highlight that the British and their policies were directly responsible for promoting caste and tribal conflict in Bastar, as the state was essentially governed as a British province from the 1850s onward. I argue further that corresponding violence in Hyderabad was caused by the landed elite of Telangana, known as the dora, who had been given proprietary rights over the land like the zamindars of British India. This singular act created the same type of landlord oppression seen in British districts throughout India, and was the root of the Telangana Rebellion which began in 1946.

289 3.2 Bastar State Policies on Low Castes and Tribals

Once the British had come into contact with the 'jungle kingdom' of Bastar they began to interfere in its internal administration almost immediately, at first under the pretext of preventing the tribal tradition of human sacrifice. An official inquiry in 1855, however, showed that human sacrifice had never actually been a local tradition at all. The reporting British officer wrote that it was “pleasing to find that there did not exist...a tradition of human sacrifices. In the low country it was said that these hill tribes never sacrificed human beings and for once the account was strictly true.”29 A more likely cause of intervention was the fact that

Bastar had extremely large iron ore deposits as well as other precious minerals and forest produce (Sundar 1997: 8-9). The British laid a direct claim on the forests of Bastar, something which they did not do in other princely states (Ibid.:

92). In Jaipur and Travancore, for example, the British placed new commercial burdens on princely forests, but they did not take them under direct control.

Alfred Gell notes that prior to the arrival of the British “...the tribal population enjoyed the benefit of their extensive lands and forests with a degree of non- exploitation from outside which would hardly be matched anywhere else in peninsular India” (1997: 435). Nandini Sundar highlights that prior to British rule there was not even a recorded state forest policy in Bastar (2001: 24).30 29 NAI, Home Department, Public, April 1855, #47 – Note: the name of the author of the report is illegible. 30 Were there tribal revolts in Bastar prior to the arrival of the British? There do appear to have been a number of tribal conflicts in and around the area – for example, the Halba Rebellion in 1774 and the Paralkot Rebellion in 1825. However, there are two important points here. First, even these conflicts were due to British influence in the region, although though they did not directly control Bastar yet. Second, the British and the Marathas worked together in

290 Over time, British influence in Bastar increased and by 1876 they governed the state directly, the raja ruling in name only. An important consideration here is whether Bastar was unique in its high level of British interference. Did all princely states see the same level of British meddling? In Chapter 1, for example,

I noted that many scholars have treated princely rulers as powerless and subservient to the British. However, I argue that the level of interference in

Bastar was unique compared to that of other states. First, as I will detail later, the

British dismissed three straight rajas of the state under the pretext of

'incompetence'. Second, during the 56-year period from 1891-1947, the British governed Bastar directly under British administrators for 42 of them. Third, the

British even governed the state through Englishmen, something which they usually did not do, preferring instead to use British Indians from the Provinces.31

Finally, I noted earlier that the British laid a direct claim on the forests of Bastar.

There is one further point which indicates the unusual extent to which the

British interfered in Bastar. As I noted earlier, the British refused to give Bastar any gun salute. This was despite the fact that the state covered 13,000 square miles and had a population of over 300,000. But other much smaller states received gun salutes – take for example, the tiny states of Mandi and Faridkot, both located in Punjab. Mandi was only 1,200 square miles, with a population

decimating these territories. As D. Banerjea notes about the Halba Rebellion, for example: “The presence of Maratha forces and the terror caused by the East India Company...precipitated the rebellion” (2002: 12). About the Paralkot Rebellion he writes: “The presence of the Marathas and the British threatened the identity of the Abujmarias and they resisted this through organising the rebellion...” (Ibid.: 12). 31 See Ramusack 2004: 196.

291 (1901) of 174,045; Faridkot was only 642 square miles, with a population (1901) of 124,912. Yet both of these states received an 11 gun salute. Bastar, so much larger and more populous, received nothing. The amount of British interference in the princely state then seems to be different from most other areas during the post-1857 period. While the rest of the princely states experienced a gradual loosening of control by the dawn of the twentieth century, Bastar experienced the opposite.

Bastar State experienced two major tribal rebellions during colonial rule: in 1876 and 1910. The cause of the first rebellion was trivial enough: the arrival of the Prince of Wales to India. The diwan of Bastar attempted to arrange a meeting between the Prince and the raja. The adivasis, however, interpreted this as an attempt by the British to abduct the raja and within hours they mobilized in large numbers and prevented him from leaving the state. Though traditionally referred to as a 'rebellion', the 1876 conflict in reality was relatively minor and featured very little bloodshed. W.B. Jones, Chief Commissioner of the Central Provinces, summarized the incident in a confidential report from 1883:

“In March of 1876 a disturbance broke out at Jugdalpur, the origin of which has never been quite satisfactorily explained. The immediate occasion of the outbreak was the Raja's setting out for Bombay to meet...The Prince of Wales. The people assembled in large numbers and compelled him to return to Jugdalpur. Their ostensible demand was not that he should not go, but that he should not leave behind the then Diwan Gopinath Kapurdar (a Dhungar, shepherd by caste) and one Munshi Adit Pershad (a Kayeth in charge of the Raja's Criminal Court), whom the people charged with oppression...They simply demanded that the two men mentioned above should be sent away...”32

32 NAI, Foreign Department, A-Political-I, January 1884, #117-125, pp. 6-7

292 Were the adivasis rebelling against the raja? The British were clearly very skeptical. An officer sent to investigate in 1882 disconfirmed the idea, noting that

“Relations between Raja and subjects generally [were] good, very good...”33

Commissioner Jones also noted that “...the insurgents committed no violence and professed affection for the Raja.”34 At worst, the adivasis were upset with the raja's choice of appointees. But another main cause of the disturbance was creeping British influence in the state. For example, Jones noted that the adivasis earlier in the year had felt antagonized by new Christian missionaries who arrived in Bastar.35 New traders, immigrants, and colonial policies all combined to create a rising sense of embitterment among the tribal population in the state.

After the death of raja Bhairam Deo in 1891 the British began to penetrate the princely administration ever more steadily. As the raja's son Pratap Deo was only six years old at the time the British directly administered the State for the next 16 years. This direct control over Bastar really also continued after Pratap

Deo became raja in 1908. Extra-Commissioner Rai Bahadur Panda Baijnath acted as Superintendent during the last four years when Pratap Deo was a minor, and then continued to act as diwan after the raja finally took the throne. E.A. De

Brett, Officer on Special Duty in Bastar, wrote about Pratap Deo's lack of power, noting that he was “bound in all matters of importance to follow the advice of his

33 NAI, Foreign Department, A-Political-I, January 1884, #117-125, pp. 13 34 Ibid., pp. 6-7 35 Ibid., pp. 16

293 Diwan and has never taken an active part in State affairs.” The Chief

Commissioner of the Central Provinces concluded later that “the Diwan was the virtual ruler of the State”.36

The 1910 rebellion was much more violent and widespread than its predecessor. One of the chief instigators of the conflict was Lal Kalendra Singh, the first cousin of the raja and a former diwan himself. He had been angling for a return to power after he had been removed by the British due to 'incompetence'.

He mobilized the adivasis by declaring that if he were returned to the throne he would drive the British out of Bastar completely. A contemporary report from a

Christian missionary living in Bastar, Reverend W. Ward, sheds some light on the rebellion:

“In the second week of February we first heard of the unrest among the Aborigines south of . Vague rumours were afloat but none of a very serious nature. On the 18th a Christian living among the Prajas – Aborigines – came to me with the story that the Prajas were all armed and were moving toward Keslur, where the Political Agent, Mr. E. A. DeBrett, I.C.S., was camping, to make known their grievances...A branch of a mango tree, a red pepper, and an arrow were tied together, and sent to all villages in the State. The mango leaves stand for a general meeting; the red pepper, a matter of great importance is to be discussed and that the matter is necessary and urgent; the arrow, a sign of war.”37

The entire state rose in rebellion and the initial British force of only 250 armed police was quickly overwhelmed. For weeks, looting, robbery, and arson plagued the entire kingdom. By the end of February additional troops from Jeypore and

Bengal had arrived and the rebellion was finally put down. Hundreds of prisoners were taken, including Lal Kalendra Singh, who was expelled from the state and 36 NAI, Foreign Department, Secret – 1, 1911, #34-40, pp. 3 37 BL, IOR/R/1/1/415: 1910

294 later died in prison.

The British conducted several inquiries into the causes of the 1910 rebellion. The Chief Commissioner of the Central Provinces summarized the

British government's position in a December 1910 report that stated bluntly:

“...from an examination of the evidence before them the Government of India were of opinion that a too zealous forest administration might not improbably be the main cause of the discontent of the hill-tribes.”38

B.P. Standen, Chief Secretary to the Chief Commissioner, wrote at the same time that “...[t]he alleged grievances of the people are...the result of a weak rule combined with gradual over-elaboration of the administration – a combination which must sooner or later give rise to dissatisfaction.”39 De Brett also conducted an inquiry on the rebellion and discerned eleven main causes, ranking chief among them “the inclusion in reserves of forest and village lands”.40 L.W.

Reynolds, another officer stationed in Bastar, noted:

“The proposal to form reserves was not finally sanctioned until June 1909 and action giving effect thereto must therefore be nearly synchronous with the rising. In his telegram of the 17th March 1910 the Chief Commissioner stated that one of the objects of the rising was the eviction of foreigners. I believe it to be the case that in connection with the exploitation of the forests Messrs. Gillanders, Arbuthnot and Company, who have a contract in the State, have introduced a large number of workmen from Bengal...the [tribes] resent the introduction of these foreigners. It is not unnatural.”41

All of the contemporary reports pointed to the same causes – the excessive influence of the British, and new forest policies. Sundar also found that most of the main participants in the 1910 rebellion were from areas that suffered the most

38 NAI, Foreign Department, Secret–1, 1911, #34-40, emphasis added 39 Ibid. 40 NAI, Foreign Department, Secret–1, 1911, #34-40 41 BL, IOR/R/1/1/415, 1910, pp. 15

295 under new colonial land revenue demands (1997: 98). Despite the admission to an 'overzealous' forest administration, British policy in Bastar after 1910 did not change substantially in the wake of rebellion. They continued to sign various forest mining agreements or renewals of previous agreements – in 1923, 1924,

1929, and 1932. These agreements also continued to encroach on forest lands and steadily displace the adivasi population. Forests also continued to be reserved by the government – as late as 1940 the Administrator of the Bastar State wrote to the Political Agent of Chhattisgarh States that “ Most of them [adivasis] dislike the proposals for forest reservation...However if these areas are not reserved it will be impossible to reserve any good teak forests in the Zamindari. (It is a most unfortunate fact that the best teak areas and the thickly populated, well cultivated Maria [Gond] villages coincide).”42

The British also continued to directly govern the state through various machinations, although this too had been disastrous in 1910. In 1922 Rudra

Pratap Deo died without a male heir and his daughter, Profulla Kumari Devi, was placed on the throne as a child. One British administrator noted: “She is about eleven years of age and no reference is made as to her eventual fitness to rule, but this is unimportant as she could always rule through a Manager or Dewan”.43

Bastar therefore experienced yet another minority administration. Then in 1936 when the Maharani of Bastar died suddenly of surgical complications in London the British installed her eldest son Pravir Chandra Bhanj Deo on the throne

42 NAI, , F. Files, 1940, #F-6-19/40(M), emphasis added 43 NAI, Foreign and Political Secretary, Internal, 1922, #319-I

296 although he was only seven years old at that time. The Maharani's husband, raja

Prafulla Bhanj Deo, who was the first cousin of the ruler of the nearby

Mahurbhanj State, had been passed over as a possible successor. This was an attempt to continue directly ruling the state instead of turning over power to the

Queen's consort. The colonial administrators in charge of the guardianship of

Pravir Chandra were even themselves confused as to the justification behind his minority administration. Administrator E.S. Hyde commented:

“I am not altogether clear what is meant in this case by guardianship...It would, however, be of assistance to me and my successors if our position could be defined. It is certainly an unusual and somewhat delicate one, for normally when a Chief is a minor his father is dead.”44

R.E.L. Wingate, Joint Secretary to the Government of India, Foreign and

Political Department, noted that passing over Prafulla for the throne was against the Queen's wishes before she passed away:

“It is her desire [the Maharani] that Profulla should have the title of Maharaja and that he should share her role as Ruling Chief, being co-equal with her and succeeding her as Ruler in the event of her death before him, her son not succeeding to the gaddi until his death.”45

Despite this, Prafulla – who had been educated with high marks at Rajkumar

College in Raipur46 – was deemed “exceedingly vain and filled with self- conceit...he is a man of very questionable moral character and completely unstable”47, and was denied the throne. Prafulla had also been very popular with tribal groups in Bastar. E.S. Hyde noted a meeting between adivasis and Prafulla

44 NAI, Eastern States Agency, D. Files, 1936, #D-51-C136 45 BL, IOR/R/1/1/2703: 1935, pp. 4 46 BL, IOR/R/1/1/2703: 1935, pp. 2 47 Ibid., pp. 2

297 in 1936 after he had been passed over for control of the kingdom:

“First of all the Mahjis told Prafulla that they had confidence and trust in him and that he was their 'mabap' [mother and father]; to this he replied that he could do nothing for them, that he had no powers. He was willing to do anything for them but...he could do nothing.”48

Even before the death of the Maharani in 1936 there had been a movement to install Prafulla as the hereditary raja, in “joint rulership” of Bastar with his wife; later came an attempt to at least establish a Council of Regency and make him the

Regent.49 Both movements were squashed by the British. They believed that

Prafulla was responsible for several anti-British pamphlets which had appeared over the past several years in newspapers throughout India. Administrators noted with some reservation, however, that “there is no actual proof as the printer's name is absent from the pamphlets.”50 The British eventually even removed

Prafulla as the guardian of his children and deemed that he should not be allowed to enter Bastar State.51

In sum, the British found fault one way or another with almost all of the occupants of the throne of Bastar. Lal Kalendra Singh was removed as diwan because colonial authorities came to realize he was “totally unfit to be trusted with any powers”.52 Rudra Pratap Deo was a “ very weak-minded and stupid individual...considered unfit to exercise powers as a Feudatory Chief”.53 Prafulla

Bhanj Deo was an agitator, unstable, and even needed to be removed from his

48 NAI, Eastern States Agency, D. Files, 1936, #D-51-C136 49 “To Seek Life of a Recluse”, The Statesman, February 9th 1937 50 BL, IOR/R/1/1/2973: 1937 51 BL, IOR/R/1/1/2805: 1936 52 NAI, Foreign Department, Secret – 1, 1911, #34-40 53 BL, IOR/R/1/1/922: 1919

298 own children. And by the dawn of independence colonial administrators were already beginning to have serious doubts about the abilities of his son, Pravir

Chandra, heir to the throne. The result of this excessive British meddling was an increase in peasant discontent and, ultimately, the birth of tribal revolt in Bastar.

3.3 Post-Colonial Caste and Tribal Reforms in Bastar

The achievement of Indian independence provided no relief to the adivasis of Bastar. The new Congress government almost immediately after independence began agitating against the new ruler of the state, Pravir Chandra

Bhanj Deo, exactly as the British had done previously against its erstwhile rulers.

The first step came in 1953 when the Madhya Pradesh54 government had the prince's property taken from him and placed under the Court of Wards. They argued that this was necessary because Pravir Chandra was insane. To be sure, the prince was by most accounts an enigmatic and bizarre man. As early as 1936 his British caretakers noted that “he has always been delicate.”55 Home Minister

G.B. Pant in 1960 wrote in a letter to a minister: “Some people say that he was almost an idiot. I cannot say if that is absolutely correct; but there is no doubt that he is erratic and whimsical.”56 His father Prafulla described him as a “young puppy whom the British have ruined”.57

Despite some seemingly legitimate complaints against Pravir Chandra due to his strange behavior, the new Indian government also made numerous

54 Chhattisgarh was formed out of Madhya Pradesh in 2000. 55 NAI, Eastern States Agency, D. Files, 1936, #D-51-C136 56 NAI, Ministry of Home Affairs, Political III, 1961, #5/5/61-Pol. III., Vol 1, pp. 315 57 NAI, Ministry of States, Political (B) Section, 1951, F.26(23)-PB/51

299 frivolous claims against him. The Secretary of the Ministry of States complained in May 1953 that “...the Maharaja had now grown an enormous beard and his hair had come down right up to his waist. The nails of his fingers are very long.

He looks just like a Sanyasi [renouncer]...His is not a presentable appearance...”58. This was apparently taken as further evidence of his mental instability. About a subsequent meeting the Secretary also wrote: “He [Pravir

Chandra] said that he has taken to the practice of Yogic exercises. I suggested that he was too young for that and that he had better marry and live a decent family life.”59

Pravir Chandra was embittered by losing his property. In 1955 he formed the Adivasi Kisan Mazdoor Sangh, a political organization which was partly created to help restore his land, but also pressed for better policies for villagers in the state. In 1957 he was recruited by the Congress Party (apparently despite the fact that he was insane) to stand for election. He viewed this as another opportunity to have his property returned. However, Congress would not relent and the shaky alliance ended quickly. After that the Indian government began to work toward removing Pravir Chandra from the Bastar throne, and they intended to replace him with his brother Vijay Chandra. In their internal memos they make a clear link to the past in pursuing this line of action:

“The adivasis have seen and read the articles appearing in certain news-papers regarding the Maharaja's derecognition. They have taken a serious view and are stirring up agitation...There was a similar move at the time of the death of his

58 Ibid. 59 NAI, Ministry of States, Political Branch, 1953, #18(4)-PB/53 (Secret)

300 grand-father Shri Rudrapratap Deo and the adivasis stirred up a violent agitation, but the British Government was wise enough to put his mother on the Gaddi. History will repeat itself now.”60

The policy of declaring those they did not like insane was well-worn among the British. This was certainly the case for several of the pre-independence rulers of Bastar. In a letter to Lord Curzon in 1899 Lord George Hamilton, Secretary of

State, wrote:

“...I felt that, if ever it became necessary to take so strong a step as deposition [of a prince], you would be less likely to frighten the Native Princes generally if you took that step, not on a plea of misgovernment but of insanity.” (Ernst and Pati 2007: 103, emphasis added)

The Indian government was not above using the same methods. Even the evidence of insanity the post-independence state used mirrored that of the

British. For instance, in 1920 raja Rudra Pratap Deo was accused of 'mal- administration' and was even briefly banned from entering his state when he returned from a trip abroad. The main reason was because he, on three occasions, refused to meet with the British Resident stationed in Bastar. He apologized, stating that a family member of his had been sick at the time. He was also surprised by the British over-reaction.61 Similarly, the post-colonial Indian government attacked Pravir Chandra for the same reason. The Secretary of the

Ministry of States recounted a conversation with the raja in 1953:

60 NAI, Ministry of States, Political Branch, 1953, #18(4)-PB/53 (Secret), pp. 136, emphasis added 61 NAI, Foreign and Political Department, Deposit-Internal, 1920, #54

301 “I told the Maharaja that he had acted very improperly in not paying due respect to the when the latter had visited that part of Madhya Pradesh.62 The Maharaja in reply disowned any desire whatsoever to be disrespectful but said that his inability to be present at the President's arrival was due to his illness. He was then down with high fever.”63

Pravir Chandra's refusal to meet the President (legitimate or not), like that of his grandfather, was taken as further evidence that he should be removed as ruler of

Bastar. Congress finally succeeded in doing that in 1961 when he was replaced by his brother Vijay Chandra.

There were two major post-colonial conflicts in Bastar. They were notable in that they both featured the raja and the adivasis on one side and the new

Indian government on the other. The first occurred in 1961. After his deposition in that year Pravir Chandra was briefly arrested for anti-government activities, which led to the adivasis besieging the police station where they believed he was being held (he was not actually being held there). For several days Bastar was gripped in a state of panic. Huge protests rocked the state and the new raja was unable to quell the disturbance because the adivasis refused to accept him as their rightful king. Thousands of signatures were collected to restore Pravir

Chandra to power and G.B. Pant bemoaned that the “Adivasis still continue to cherish their traditional feelings of respect and loyalty to the erst-while

Princes...”64

The second conflict occurred in 1966. On March 25th of that year Pravir

62 Bastar was originally located in Madhya Pradesh; Chhattisgarh did not exist at this time. 63 NAI, Ministry of States, Political-B, 1953, #18(4)-PB/53 (Secret) 64 NAI, Ministry of Home Affairs, Political III Branch, 1961, #5/5/61-Poll-III., Vol. 1, pp. 315

302 Chandra Bhanj Deo was gunned down by Jagdalpur police on the steps of his own palace. Sundar recounts the extraordinary tale:

“Police guards [trying to prevent a procession by the adivasis] armed with rifles, instead of the usual lathis, were posted at the four main entrances of the palace...An incident with a prisoner who was under trial and being escorted to the policy lock-up in the palace premises accidentally sparked off an encounter between the people assembled in the palace grounds and the police, with the people shooting bows and arrows and the police responding with tear gas and bullets. One policeman died from an arrow wound and another was injured...The police then chased the adivasis through the gardens and fired at them in front of the two palaces. Pravir Chandra, who had come out to see what was happening, was shot and killed along with other people. By about 12:45 pm, most adivasis had taken shelter inside the two palaces...” (1997: 218-19)

Though the police and the government claimed that it was adivasis who had started the disturbance, most were in fact armed only with bows and arrows. A subsequent investigation by Justice K.L. Pandey also discredited this theory and found the police to blame (Ibid.: 218-19). To this day the so-called “police action” is highly controversial and it is widely believed that Pravir Chandra's death was a political assassination. Though only a small number of adivasis probably died, rumors still abound that hundreds or even thousands were killed.65 The adivasis in the former Bastar State today continue to venerate Pravir Chandra. March 25th has since been styled “Balidaan divas” or “The Day of Sacrifice”.66 It is very telling that both the British government and the new Congress government had a plethora of complaints about one ruler of Bastar after another – in fact, the only

65 See for example the Amrit Sandesh, March 25th 2007 article “Ambassador of the Revolution and Martyr Pravirchandra Bhanjdev Spilled His Blood in Sacrifice for Bastar” in which the author suggests that “the total number of deaths to this day remains a mystery, as several people claim that hundreds or even thousands died” (Hindi, my translation) – DPL, 24, 1B, un-numbered 66 “Maharajah Pravirchandra Bhanjdev, Messiah of Tribals” (Hindi, my translation) Highway Channel, March 2009 – DPL, 24, 1B, un-numbered

303 group not to complain were the adivasis whom they ruled.

The new Indian government also did not see fit to reverse many relics of the previous British administration. Foremost among them was colonial forest policy. Just as the inclusion of forest and village lands as reserves was a major cause of pre-independence rebellions, the post-independence government continued and even exacerbated this policy. From 1956-1981 one third of the total amount of forest felled in the Bastar district was for development projects

(Sundar 1997:198). It is due to the continuation of colonial-era policies and the failure of reform that a political space for the Naxalites was created in Bastar.

They became an important political force when they entered the Bastar area in the early 1980s, mobilizing villagers around their economic grievances. Many of the earliest Naxalites came from the movement in Andhra Pradesh, and they began a 'Go to the Village' campaign in the Bastar area to enlist tribal support

(Navlakha 2010). The two main groups operating there now are the CPI-ML and the People's War Group. Two of the main initial recruiting grounds for the

Naxalites in Bastar were hostels and schools, especially special schools for adivasis.

On the surface, there appear to be many, many attempts at reform in

Bastar. There are an absurd number of overlapping organizations in Bastar district, for example: Community Development Programme, Community Area

Development Programme, Whole Village Development Programme, Drought

Prone Area Programme, Hill Area Development Programme, Intensive Rural

304 Development Programme, Tribal Area Development Programme, Intensive

Tribal Development Programme and the Bastar Development Authority. Yet, the root causes of discontent in the area – land, agriculture, and poverty – remain under-addressed (EPW 1989: 2237). In 1989 the Commandant of the Bastar

Battalion charged with stopping the Naxalites famously admitted that “our battalion was raised to solve a problem it can not solve” (Ibid.: 2241). In the end it is fitting, considering how little has changed, that today the name of the main

Naxal front organization in Bastar is Adivasi Kisan Mazdoor Sangathan, almost the exact name of Pravir Chandra's tribal organization formed in 1955.

3.4 Hyderabad State Policies on Low Castes and Tribals

The roots of the contemporary conflict in Hyderabad also lie in the colonial history of the region. In 1946 a massive peasant revolt began in Telangana - a deadly harbinger of the caste and tribal violence which plagues the region today.

What accounts for the long history of conflict in Telangana? Unlike the Bastar case, excessive British influence cannot be used here as an explanation for the bloodshed. However, despite the internal autonomy of the nizam, he actually had very little control over the Telangana region specifically (Regani 1963, Elliott

1974, Reddy and Sharma 1979). His power, immense in Hyderabad city itself, projected weakly into the hinterland of his massive kingdom. The best indicator of this is the fact that he received only one-fourth of the land revenue assigned to the powerful landlords of Telangana (Wood 1981: 33). These deshmukhs (large landlords) of Telangana essentially operated an autonomous state. I. Thirumali

305 notes: “The deshmukhs...practically usurped the authority of state [sic] and kept

'the whole countryside...in their possession'”.67 But were the landlords of

Telangana an important part of the nizam's administration? P. Sundarayya, for example, refers to the landlords of Telangana as the “pillars of the Nizam's aristocracy in the rural area” (1973:4). However, the relationship between the nizam and the deshmukhs in Telangana was adversarial – as I highlighted in my chapters on Jaipur and Travancore, princely rulers quite often had long, contentious power struggles with their nobles. Telangana was no exception;

Sarojini Regani, for example, writes that “[m]any of the Zamindars [in

Telangana]...took advantage of the ineffectualness of the Nizam's army and carried on depredations in the countryside” (1963: 214). Similarly, Peter Wood, writing about the failure of the deshmukhs to pay their appropriate revenues, notes that the nizam's minister Raju Chandu Lal “lament[ed] that the larger zamindars were...feeling secure...in their independence...” (1981: 33).

The source of contemporary Naxalite violence in Telangana has its roots in the brutal oppression of the landed elite of the region, known specifically as the dora. The power of the dora was unique – they were originally revenue agents of the nizam but in the late nineteenth century, due to a very consequential set of reforms initiated by diwan Salar Jung, they were given formal proprietary rights

(patta) over the land. In this sense, the dora were very similar to the zamindars of British India, and Telangana suffered from many of the same problems faced

67 The second part of the phrase is a direct quote of A.G. Hankin, Inspector General of Police in Hyderabad.

306 by the landlord areas in the provinces.

Note that this administrative reform by Salar Jung was sharply divergent from the land policies of the other princely states examined in this dissertation.

The princes of Jaipur and Travancore had pursued very different strategies in dealing with the nobles in their kingdom. In Jaipur, the darbar, especially in the early 20th century, began to take a more direct approach in reforming noble estates and bringing them under tighter control. The Jaipur example is telling because, like Hyderabad, Jaipur's rajas only controlled a small portion of the land, but they still made attempts to curb the power of nobles. Similarly, in

Travancore the raja was more aggressive even earlier, bringing almost all of the land held by nobles under the direct control of the darbar.68 In Hyderabad, however, (perhaps due to its immense size) Salar Jung followed a very different set of policies, empowering the nobles and turning the dora into landlords. The impulse behind this action was essentially similar to that of British administrators in the provinces, who had hoped that by creating landlords native territories would be better-governed and more agriculturally productive.

Ultimately – also mirroring the results in British India – the creation of a powerful class of landlords in Telangana backfired, resulting in new levels of rural repression and, subsequently, rebellion.

68 Furthermore, even in noble-controlled lands in Jaipur and Travancore it was not clear that nobles themselves claimed a proprietary right over the land – in Jaipur, for example, some jagirdars assigned jagirs within their own kingdoms, essentially turning over the land via a hereditary grant to someone else. The zamindars in the princely states often remained what they had been prior to the arrival of the British: revenue agents and not private proprietors.

307 The dora were high castes – mainly Reddis – who lived in large forts

(gadis) and controlled the countryside. Just one dora family in Nalgonda, for example, controlled 160,000 acres of land (Pavier 1974: 1413). They were known throughout Hyderabad State for their brutality; they were infamous, for example, for their sexual exploitation of peasant women. They also enforced feudal systems of bonded labor (vetti or begar): most poor families had to send one man to serve in this capacity for the . Not only were they the main landholders in

Telangana, but they were also the main lenders of credit, which meant that the peasantry was almost wholly dependent upon them for their livelihood.

The status of peasants in Telangana was extremely backward. Thirumali writes about some of the more egregious indignities:

“The labouring people were not permitted to wear shirts, chappals [sandals], 'talapaga' (headware), 'tilak', or ornaments...They had to shave their heads clean, leaving only a tuft and of course they never dared to sport prominent moustaches. And every sentence the labourers spoke started with 'dora I am your slave, I touch your feet'....” (1992: 480)

One peasant woman interviewed during the Telangana conflict in 1947 highlighted how the depredations of the dora extended far beyond the nizam's rule: “The Nizam had fixed the amount of the paddy [for levy] but first the deshmukh would take three maunds [82 pounds] per acre for himself...” (Roosa

2001: 67). Brutal landlord rule in Telangana was the main initial cause of the peasant violence in Hyderabad State. The deshmukhs were the “prime targets”

(Thirumali 1992: 477) of the rebels. David Ludden echoes this point, noting that

“the ossification of this landed aristocracy became the context for the Telangana

308 peasant revolution in the 1940s” (1999: 157).

The brutality of the dora, however, is not meant to excuse the nizam himself from some portion of the blame for the agrarian conflicts in Hyderabad.

He privately controlled more than 10% of the state as his private lands (sarf-i- khas). On top of the revenue he appropriated from these areas he also received 7 million rupees a year from the royal coffers simply for his personal expenditures.

But the fact remains that he had little control over the landlords of Telangana.

Would the nizam having more control of this region have mattered in limiting the conflict? His rule was certainly less severe for the peasantry than that of the dora.

For example, taxes on irrigation were ten times higher in Telangana than in the nizam's directly-controlled lands (Sundarayya 1973: 8). Telangana peasants were the most backward and economically depressed in the entire state.

Furthermore, the nizam also enforced surprisingly progressive policies toward adivasis, many of whom were at the center of the rebellion in Telangana.

Foremost, in the mid-1940s the nizam hired the famed Austrian ethnologist

Christoph von Fürer-Haimendorf to undertake a series of bold tribal reform policies. These new laws were designed to, most importantly, prevent land dispossession and promote tenurial rights. This involved the creation of thousands of land deeds, or patta documents, for adivasis. The Tribal Areas

Regulation 1356 Fasli (1946) set up designated tribal lands, panchayats, and even established the supremacy of tribal law in these areas. The nizam also created the Social Service Department, which von Fürer-Haimendorf headed, to

309 create a direct linkage between the nizam's government and tribal groups. This allowed the tribals to seek redress of grievances directly from the government without having to deal with intermediaries. The employees of the Social Service

Department had at least an amateur training in anthropology and were familiar with tribal languages and customs. Finally, the nizam also took great pains to avoid adivasi land displacement during development projects. These projects, such as the Narmada Dam in Gujarat, have provoked intense conflict in modern

India. From 1923-31, during the building of the Nizam Sagar, the major dam in

Hyderabad, the nizam created an entire new village for adivasis who had been displaced by the project (Mann 1996: 310). Overall, the change in tribal fortunes was rapid. By 1946 Gonds, the major tribal group in Hyderabad, “became suddenly the 'most favoured' ethnic group in the region” (von Fürer-Haimendorf

1982: 42). The nizam's policies in Hyderabad, then, mirrored the general protective treatment of tribal groups throughout princely India. Yet, the autonomous landlords of Telangana prevented any changes from occurring in their territories, and the tribals in those regions were untouched by reforms.

Finally, the nizam also pursued some protectionist policies for low castes.

In 1951 J.H. Subbiah, member of the Scheduled Castes Federation, argued to the new Indian government for various policies that should be implemented for the advancement of SCs. Among them he noted:

310 “There was considerable waste land available in the State and the Federation felt strongly that in allotting this land the first preference should be given to the members of the scheduled castes. They should also be given monetary and other assistance to bring this land under cultivation.”69

In fact, this was exactly the policy which the nizam had implemented in the 20th century in order to attract SCs to the state (Omvedt 2006). Dalit leaders in

Hyderabad also successfully pressed the nizam to create a trust fund for the welfare of SCs, and the nizam endowed the foundation (Depressed Classes

Welfare Fund) with one crore70 rupees (Paswan and Jaideva 2002: 216). Many well-known Dalit social reformers found a home in Hyderabad State, as the nizam appointed many of them to positions within the administration: for example, B.S. Venkatrao, J.H. Subbiah, and B. Shyam Sunder. The nizam was, however, partly driven by a desire to convert untouchables to Islam, even at one point offering 50 million rupees for their wholesale conversion (Queen and King

1996: 51)! Again, however, most of these changes toward untouchables only occurred where the nizam had direct control over his territories, and the position of low castes in Telangana, for instance, remained very backward.

3.5 Post-Colonial Caste and Tribal Reforms in Hyderabad

In 1946 a peasant struggle began in Telangana. Since the middle of the

1940s the Communist Party of Andhra had been organizing a massive campaign in the villages of the region. Specifically, the communists pursued a broad agenda of agrarian reform and adopted a harder line than the moderate State Congress

69 NAI, Ministry of States, Hyderabad Branch, 1951, #1(16)-H/51 70 One crore is ten million rupees.

311 Party, which resulted in more support from depressed agrarian castes. The communists eventually succeeded in organizing some 4,000 villages together to spearhead the revolt. The refusal to pay taxes by one peasant woman in particular, Chakali Ailamma, signaled the symbolic beginnings of the rebellion.

There are two important points here regarding the colonial revolt in

Telangana. First, the earliest phase of the rebellion was directed only against the dora and not the nizam. The aims of the communists in Telangana were all specific to the region: for example, the end of bonded labor and the creation of tenurial rights for cultivators. Second, the initial phase of the rebellion was non- violent. It could perhaps best be characterized as a protest against the dora which utilized methods of non-cooperation (i.e. refusing to pay exorbitant taxes).

However, by 1947 the struggle began to take on a more violent character largely due to a campaign of repression by the dora. They often had their own private militias which they employed, and once the communists began protesting in

Telangana the landlords utilized these forces to brutally suppress them.

John Roosa (2001) has noted that there were three different phases of the

Telangana conflict: the first was a small struggle directed entirely against local landlords; the second was meant to force the nizam to accede to the Indian

Union; and the third, post-1948, was when the struggle continued against the new Indian government. By 1947 the second phase had begun once the communists saw that the nizam was unwilling to accede to the new Indian republic. At this point, the conflict in Telangana morphed into a kind of

312 nationalist struggle, or as Roosa puts it, the conflict went from a “revolt over land and grain” to a “nationalist revolt over land and grain” (2001: 58).

As British rule in India came to a close in the 1940s the nizam was suddenly faced with the terrifying prospect of being a Muslim ruler in what would be an overwhelmingly Hindu country. The protection of the British would leave the nizam in a precarious situation. He therefore refused to accede to the Indian

Union in 1947, arguing instead that he would form his own, independent state.

Furthermore, he argued that Hyderabad had always been governed as its own country – a statement which the Government of India called “an astounding assertion”.71

In 1947 the Indian government was forced to sign a 'stand still agreement' with the nizam which maintained the status quo of Hyderabad's unique political situation for one year in order to determine the future of the state and its relationship to independent India. In 1948 the Indian government, frustrated with the continuing reluctance of the nizam to accede to the union, finally annexed Hyderabad (Operation Polo). After that, the former princely state was incorporated into the new Indian republic.

The collapse of the nizam's regime in 1948, however, did not end the conflict in Telangana. Rather, the worst violence in the region came after 1948.

This alone highlights that the nizam himself was only one component – not even the most important component – of the conflict. If the rebellion had been merely

71 Government of India White Paper on Hyderabad, pp. 2

313 against the nizam then his removal from power might have ended it. In reality, the rebels fought for only one year against the nizam (1947-8); the last four years of the conflict were fought explicitly against the new post-colonial Indian government. It was only the failure of the Government of India to deliver real and lasting reform in Telangana which allowed the conflict to continue and escalate.

During the early post-independence period the new Indian state dealt with the conflict in Telangana by instituting a massive terror campaign: thousands of rebellious peasants were jailed, tortured, and murdered (Sundarayya 1973). The police even burned down entire villages where the communists had redistributed land (Roosa 2001: 83). Roosa notes that Home Minister Patel regarded

Operation Polo as an opportunity not only to overthrow the nizam, but also an opportunity to station troops in Hyderabad State that would eventually be used to crush the communists (Ibid.: 80). By 1951 the communists realized that they were outmatched and ended the Telangana conflict, signing an armistice with the

Indian government. They announced the end of the military struggle, urging the peasants of Telangana to prepare for the next election and an opportunity to beat

Congress. Mohan Ram has noted that “it was a tame surrender because the party gave the peasantry no guarantee about protecting their hard-won gains” (1973:

1029). As I will show, many of those gains proved illusory.

The post-independence Indian government did make an attempt to institute new land reform efforts in Telangana, hoping to end the root causes of the conflict there. However, the reform – even by the government's own

314 admission – was largely a failure. The Director of Intelligence's report on the process from 1952 noted:

“Although it is admitted that it [land reform] is a progressive measure which has freed about 2 lakh tenants from the landlords...it is felt that its benefits are dubious. It is feared that the Tenancy Act would result in a fall in production because many tenants would not be able to cultivate the land in the absence of loan facilities which they formerly got from the landlords and which is now being refused to them. The problem of setting the tenant on his feet, after putting him in possession of land, is impracticable as it would involve an expenditure of Rs. 60 crores. The only other results of this fragmentation would be a loss in production and tenants disposing of their land and gradually migrating to towns.”72

Another (secret) report from the same year gathered the views 'expressed by

Congress workers and other non-Communists on the Tenancy Legislation in

Hyderabad' and noted:

“The result of the implementation of the Tenancy Act will be a fall in production, because many tenants will not be able to cultivate the land, having been deprived of the loan facilities which the zamindars used to extend to them previously. Therefore, though it is no doubt a progressive step in favour of the tenant, in the larger interests of the country it is harmful. In any case, the implementation of this Tenancy Act has not caused any mobilisation of support of the people on side [sic] of the Government.”73

Furthermore, the new Congress government went further and even reversed many of the reforms which had been carried out by the communists. J. H.

Subbiah complained in 1951 that “[in] the Telangana area the communists had distributed large areas of land to members of the scheduled castes. These lands are being restored in many cases to the original landlords.”74

Furthermore, many of the progressive tribal policies instituted by the

72 NAI, Ministry of States, Hyderabad, 1952, #6(18)-H/52 73 NAI, Ministry of States, Hyderabad, 1952, #6(18)-H/52, emphasis added 74 Ibid.

315 nizam's government under the auspices of von Fürer-Haimendorf were altered and weakened in the decades after independence. Perhaps most importantly, the tribal legislation passed in the 1940s was replaced by the Andhra Pradesh

(Scheduled Areas) Land Transfer Regulation of 1959. von Fürer-Haimendorf himself noted that this had many negative effects; for example, the abolition of tribal panchayats, and the lack of field staff who lived or worked within close proximity to adivasis. The result was that “the gains achieved in the 1940s were largely lost in the 1970s” (von Fürer-Haimendorf 1982: 57). By 1987 the Andhra

Pradesh government recorded that there were still 44,379 cases of adivasi lands being occupied by non-tribals, extending over an area of 189,337 acres (Reddy

1988:1460). The situation was also similar for low castes – a report from Swami

Ramanand Tirtha's tour of Telangana found that “[Harijans] complained of harassment by the Police...They also pointed out that the new Tenancy act was not helpful to [their] interests...” Some police in the region also continued to extract begar from low castes.75

The economic situation in contemporary Telangana is still precarious.

Even today, sixty years after the end of the rebellion, 9 out of the 10 districts comprising the region are considered 'backward' by the Government of India.

Farmer suicides in the region have also been a noted problem, higher in

Telangana than anywhere else in Andhra Pradesh (Deshpande and Arora 2010:

216). In the state as a whole 10% of the rural population is still landless; in fact,

75 Ministry of States, Hyderabad Branch, 1951, #1(16)-H/51, pp. 14-16

316 36% have less than an acre of land (Balagopal 2007: 3831). Most attempts at creating tenurial ceilings and distributing the excess to the poor are rife with corruption, with much of the land never reaching its intended recipients.76 Much like the situation in Bastar, the continuing failure of development policies and reforms to change the economic situation in Telangana directly led to the rise of the Naxalites.

In 1980 founded the People's War Group in

Andhra Pradesh. Seetharamaiah had actually been a participant in the Telangana

Rebellion, and the new Naxalite movement was carried out in the spirit of

Telangana. Even before the founding of this organization, communists from

Bengal had entered Andhra Pradesh in order to expand the movement against oppressive landlords. Like Bastar, the main response by the government to the rise of Naxalism in Andhra has been a military counteroffensive. In 2004 when

Congress came to power in the state they declared that they would negotiate with the Naxals. Within almost a year, however, they turned around and outlawed all

Naxalite groups. Unsurprisingly, the violence has continued in Telangana.

Ultimately, the roots of the contemporary Naxalite movement in

Chhattisgarh and Andhra Pradesh are located in the colonial period. Unlike almost all other modern princely states, Bastar and Telangana experience an inordinate amount of violence, but princely rulers in both areas were not its primary cause. In Bastar, British administrators directly governed the state from

76 On this, one of the best treatments is K. Balagopal's three part series in Economic and Political Weekly entitled “Land Unrest in Andhra Pradesh”.

317 the mid-19th century, and implemented the same discriminatory policies toward low castes and adivasis in the state as they did throughout British India. In

Hyderabad, the Naxalite conflict has its roots in the dora, the landed elite of the

Telangana region who were given proprietary rights over the land similar to the zamindars of the British provinces. The nizam of Hyderabad enforced a number of progressive policies for low castes and adivasis, but the dora were the prime targets of the rebellion. The new Congress government in post-independence

India did little to reform the situation of low castes and tribals in either region, thus opening up a political space for the Naxalites.

IV. Religious Violence in the Red Corridor

The primary purpose of this chapter has been to describe why Bastar and

Hyderabad have been such exceptions to the argument that British rule produces caste and tribal conflict. This section, however, considers the main argument of this dissertation about religious violence with reference to Bastar and

Hyderabad. There are two central questions:

1. If Bastar was essentially a British province, does it experience lower levels of religious violence today? 2. Does Hyderabad experience a large amount of religious violence today, especially areas which were under the direct control of the nizam?

If the answer to both of these questions is yes, then it adds further credence to my general argument about the patterns of ethnic violence in contemporary India. A look into the status of both regions in the contemporary period highlights that

Bastar today is in fact an area almost entirely devoid of religious conflict. On the

318 other hand, Hyderabad is exceedingly violent – one of the most violent communal regions in all of India.

Contemporary Chhattisgarh experiences almost no religious conflict. The

Varshney-Wilkinson dataset records only 13 casualties over 45 years and none in the former Bastar State. But this is not unexpected – after all, only a minuscule percentage of the population (2%) are Muslim, so it is an unlikely area for communal animosities. This was also the case in the colonial period, as Bastar did not experience any major communal tension during colonial rule. Despite this,

British rule may have been a factor in preventing the state from developing communal animosities; as I showcase in the next section, the British appointed a number of Muslims to important posts in the Bastar administration during the colonial period, displaying the same protectionist policies towards Muslims here that they did throughout the rest of India.

On the other hand, Andhra Pradesh is rife with communal conflict today.

The areas comprising the former Hyderabad State, which now includes districts from the states of Karnataka and Maharashtra, are among the most communally- sensitive areas in all of India. Hyderabad city has been engulfed in almost continuous Hindu-Muslim riots since independence. The Varshney-Wilkinson dataset records that Hyderabad experienced 965 deaths and injuries from communal riots from 1950-1995; it is the third most riot-prone city in all of

India. In section 4.3 I highlight that the nizam and his religious policies directly created this bloody legacy of communal conflict.

319 4.2 Communal Policies of Bastar State

By the 1850s the British had essentially come to control the Bastar State.

While there is a plethora of information about their policies toward low castes and adivasis, there is precious little on their communal policies in the region.

Muslims were a very small portion of the population of Bastar State even during the colonial period. There is not much information about the state of Muslims during this era, and a serious dearth of archival material on them. Yet, Bastar – as a British territory – does still fit the existing pattern of princely states and religious violence established in this dissertation.

First, it is important to note that the Hindu rajas of Bastar also perceived of themselves in religious terms like princes throughout the rest of India. For example, as late as 1961 the Bastar raja Pravir Chandra “wanted to form a new

Chamber of Princes which would work for the establishment of a Hindu State in

India”.77 During the colonial period British administrators also worried about the high level of influence of a “small clique of influential Hindus in Jagdalpur”.78

Perhaps in response to this, the British appointed Muslims to high positions within the Bastar administration when they came to control the state.

As I noted earlier, the British had been appointing Englishmen as diwans – a fact which highlights the exception level of control they had in Bastar – but in 1919 they appointed two Muslims to the position: “Khan Bahadur Mahdi Hassan, a retired member of the Provincial Service, and...the present Dewan Khan Bahadur

77 NAI, Ministry of Home Affairs, Political III Branch, 1961, #5/5/61-Poll-III., Vol. 1, pp. 167 78 NAI, Foreign Department, Secret – 1, 1911, #34-40

320 Hafix Walayatulah, an Extra-Assistant Commissioner”.79 Even here, in a state where Muslims had almost no representation in the local population at all, they were appointed to the most important positions within the colonial administration. This lends credence to the claim that the British generally perceived of Muslims as part of the administrative elite of their empire.

4.3 Communal Policies of Hyderabad State

Unlike Bastar, there is a wealth of information about the communal policies instituted by the rulers of Hyderabad. The Muslim nizams of Hyderabad ruled the state since Mughal regional governor Asaf Jah declared himself the ruler of an independent kingdom in the early 18th century. During most of

Hyderabad's history the region had been home to a large Hindu population – they were almost 90% of the population in 1901. Akbar Hydari, Prime Minister of the State, believed that Muslim rule in Hyderabad was merely “an accident of history” (Kooiman 1998: 646). As I will detail here, Hyderabad was one of the most notorious princely states in terms of communal conflict. The nizams of

Hyderabad, especially the last sultan Osman Ali Khan, were brutal toward the

Hindu population, creating a legacy of violence which still plagues the state of

Andhra Pradesh today. Hyderabad is also an important case because the ruler of the state was Muslim, whereas Jaipur, Travancore, and Bastar were all governed by Hindu rajas. This highlights that both Hindu and Muslim rulers could practice discriminatory policies toward non-coreligionists.

79 British Library, IOR/R/1/1/922 : 1919, pp. 1

321 Scholars have debated the state of Hindu-Muslim relations in Hyderabad prior to the 19th century, but most agree that in the late 19th century the state began to exhibit a more explicitly communal character. Much of this was due to nizam Osman Ali Khan, who ruled from 1911-1948 and enforced several discriminatory policies against the Hindus of the state. Peter Wood notes that in

Hyderabad the nizam was “the very apex of authority...All government drew its legitimacy from his appointment. All elites owed their position to his favor. He was the 'universal heir' to all his subjects” (1981: 67). He was a devoutly religious

Muslim, and enforced a number of policies to undermine the Hindu population in the state. Ian Copland notes on the religious policies of Hyderabad:

“The building and repair of places of worship was subject to the jurisdiction of the State Ecclesiastical Department whose predominantly Muslim staff had a legal right to disallow projects in areas 'where the population of the followers of Islam was considerable'...On Muslim holy days, for example during the months of Ramadan and Mohurrum, Hindus were encouraged to worship indoors and all public ceremonies involving music were banned” (1988: 788-89)

Outside of discrimination toward religious practices, numerous newspaper reports from the period highlight the general inferior socioeconomic position of the large Hindu population. The Milap of Lahore editorialized in November of

1931: “How regrettable is the matter that the Hindus [of Hyderabad]...are compelled to spend their life [sic] like slaves”.80 The Dig Vijay from Sholapur commented that “Hyderabad is a State in which being a Hindu itself is an offence”.81 Another article from The Leader noted the disproportionate power of

80 APSA, Installment #47, List #10, File #82 81 Quoted in Copland 1988: 785

322 Muslims in the princely administration:

“The Executive Council is composed of eight members; of this, one is a Hindu and one an Englishman – the rest are Muslims. Every official in the executive council, drawing a salary above Rs. 225 is a Muslim. Not a single Hindu is admitted anywhere near the precincts of the highest office under Government or his Exalted Highness's household. The High Court consists of 10 judges; eight are Muslims and only two are Hindus. A vast Hindu population is deliberately treated as hewers of wood and drawers of water.”82

In 1923 Hindu organizations published an important pamphlet titled The

Grievances & Disabilities of The Hindus in Nizam Hyderabad State.83 They summarized the general position of Hindus like this:

“The deplorable condition (bordering on serfdom), of the Hindus of the Hyderabad State has got a long history behind it. For several years the Hindus in the various Departments of the Nizam's Government have been to the Nizam and his coreligionists like a red rag to the bull...having toured through his territories and after having ganged through his Furmans, his Budget, list of his donations and his numerous other activities we were convinced that the Nizam is not the epitome of Aurangzeb84 but, Aurangzeb was an epitome of the present Nizam...”85

Like the rulers of Jaipur, the nizam of Hyderabad also tried to prevent unflattering newspaper accounts of his kingdom. In 1946 the Telugu weekly Praja

Sakti was banned. Then, in response to another article in 1931 by Hindu Outlook, the nizam wrote in an internal memo that “I think our line of action in future

[sic] should be to catch papers as far as possible on the ground of defamation”.86

Several authors, however, have argued that Muslim oppression of Hindus in Hyderabad has been exaggerated. For example, Ian Copland argues that,

82 APSA, Installment #47, List #10, File #82, emphasis added 83 NAI, Foreign and Political Department, Political Branch, 1932, #207-P (Secret) – Note: Tables 5.1 – 5.4 are all from this file. 84 Aurangzeb was a Mughal Emperor (1658-1707) who was renown for his religious discrimination and destruction of Hindu temples. 85 NAI, Foreign and Political Department, Political Branch, 1932, #207-P (Secret) 86 APSA, Installment #47, List #10, File #L71d

323 despite the grievances of the Hindu population, “I find it difficult to accept that the situation in Hyderabad was one of systematic 'Islamic' oppression” (1988:

789). Both Copland and Dick Kooiman even use the exact same phrase in arguing that Hindus “dominated” agriculture87 – Copland notes that Hindus constituted

86.2% of the agricultural sector. Along these lines, Ashutosh Varshney argues that “Hyderabad before 1938 appears to have had a history of everyday Hindu-

Muslim engagement” (2002: 200).

However, a closer look into the official documents of the Hyderabad State undermine these varied arguments. First, I consider the dynamics of the countryside; even there the Hindu population was oppressed. Tables 5.1 and 5.2 present data on the power of Hindus in rural areas. The first table examines

Hindu and Muslim representation among two different types of land owners,

Jagirdars and Maktadars; Table 5.2 shows their representation among

Mansabdars.

87 See Kooiman 1998: 646, and Copland 1988: 789, respectively.

324 Table 5.1: Jagirdars and Maktadars Revenue in Hyderabad State, 1920s88 Annual Jagir Revenue Hindus Muslim 10 – 500 100 42 500 – 1,000 109 42 1,000 – 3,000 241 161 3,000 – 5,000 65 72 5,000 – 6,000 25 22 6,000 – 10,000 29 72 10,000 – 12,000 10 23 12,000 – 25,000 20 55 25,000 – 50,000 20 24 50,000 – 100,000 6 9 100,000 and above 7 12 632 534

Table 5.2: Mansabdars Revenue in Hyderabad State, 1920s89 Annual Amount of Mansab Hindus Muslims 200 – 250 32 235 250 – 300 30 179 300 – 400 46 312 400 – 500 38 202 500 – 600 24 154 600 – 900 42 237 900 – 1,200 25 146 1,200 and above 29 178 266 1,643

Considering that Hindus were over 80% of the population at this time, they by no means 'dominated' agriculture. In fact, 71% of Hindu jagirs generated revenues of less than 3,000 rupees a year compared to only 45% of Muslim 88 Figures are in rupees 89 Figures are in rupees

325 jagirs. While there were certainly more Hindus overall in agriculture, their power in the countryside was not commensurate with their population strength.

On the other hand, the small Muslim agricultural population generally enjoyed more privileged agrarian positions.

Furthermore, the state of Hindus declined significantly once the British government began to devolve authority back to princely rulers in the early 20th century. This is similar to the position of Muslims in Jaipur and Travancore, as pent up communal animosity exploded in both areas in the 1930s and 1940s.

Consider Hindu and Muslim literacy rates in Hyderabad. Table 5.3 indicates that while Muslim literacy continued to expand slowly from 1881-1921, Hindu literacy actually declined over the same period.

Table 5.3: Hindu and Muslim Literacy, Hyderabad State, 1881-1921 Year Hindu Literacy Rate Muslim Literacy Rate 1881 2.97% 4.94% 1891 3.35% 6.16% 1901 2.53% 5.46% 1911 2.29% 5.88% 1921 2.59% 8.89%

Table 5.4 below similarly highlights the communal makeup of the princely administration: most of the high salaried positions were taken by Muslims.

Roughly 70% of all gazetted posts in the administration belonged to Muslims

(Kooiman 1988). Communal representation in the state was so backward that the

Government of India in its White Paper on Hyderabad condemned “the basic

326 structure of the Civil Service of the Hyderabad State...[it] maintains a farcical system of recruitment, which assigns privileged position to certain sections and is based on unmasked favoritism and nepotism” (1948: 3).

Table 5.4: Salaries of Employees of Hyderabad State, 1920s90 Annual Salary Hindus Muslims 300 – 600 1,884 3,380 600 – 900 302 718 900 – 1,200 217 474 1,200 – 1,500 81 148 1,500 – 1,800 85 204 1,800 – 2,400 44 172 2,400 – 3,000 27 110 3,000 – 3,600 33 137 3,600 – 4,200 6 44 4,200 – 5,000 9 59 5,000 and above 51 205 2,739 5,651

All of these accumulated facts and figures suggest that the oppression of

Hindus in Hyderabad was not overstated, but rather a very real fact of politics in the state. Varshney's argument about interethnic engagement is also not borne out by closer examination. Even prior to the late 1930s communal relations in

Hyderabad were poor. It was, contrary to his description of everyday engagement, a society “where social intercourse between communities...was rigidly regulated and tightly restricted” (Wood 1981: 72). Varshney is also incorrect in noting that there were no riots in the city prior to 1937: they occurred

90 Figures are in rupees

327 in 1924, 1929, 193291, and 1933 (Copland 1988: 784).

There is one final counterargument made by scholars of colonial

Hyderabad: they write that the Muslims of the state were not united – rather,

Islam was deeply factionalized in the kingdom (Copland 1988: 789). For example,

Sunnis in the state continuously worried that the mercurial nizam Osman Ali

Khan would convert to the Shia sect. While the factionalization of Muslims was certainly a major internal divide, this summarizes the situation only until the late

19th century. Communal animosities began to grow at that time due to an influx of

Muslims from north India, brought into Hyderabad to serve as administrators by

Salar Jung. These new immigrants “introduced a militant Muslim consciousness from North India” (Elliott 1974: 31), and many had been part of the Aligarh movement. Over time, the Sunni-Shia divide receded as the Hindu-Muslim divide grew. As the 20th century dawned Hyderabad had become an explicitly

Islamic state.92

As the end of British rule in India neared at the close of the 1940s, the political situation in Hyderabad deteriorated considerably. For his part, the nizam had by the eve of independence essentially abdicated power in his kingdom. By 1946 most of his time was devoted to determining the future of

Hyderabad once the British granted India independence. The collapse of the

91 Copland does not mention the 1932 riot – see NAI, Foreign and Political Department, General Branch, 1932, #444-G 92 This immigration formed the source of major friction in Hyderabadi politics in the 20th century. The mulkis were indigenous inhabitants of the state, whereas the new Muslim administrators from north India were non-mulkis. See Leonard 2003: 370-71.

328 nizam's legitimacy in Hyderabad was clear – as the Muslim ruler of an overwhelmingly Hindu state he feared the repercussions of an independent

(Hindu) Indian republic. In the midst of this political crisis the nizam turned to

Kasim Razvi and the Majlis-i-Ittehad-ul-Mussulmeen in Hyderabad.93 The

Ittehad was a hard-line Islamic faction which tried to unite the Muslims of the state against Hindus. As time wore on the Ittehad became the de facto rulers of

Hyderabad. In November of 1947 Kashinathrao Vaidya, a leader in the satyagraha movement in Hyderabad, wrote to that “The Majlis is sure to carry out its programme and murder, blood-shed, arson, pillage and butchery will be the inevitable features. The Majlis want to seize power by hook or by crook and it will not spare the Nizam even.”94 The Ittehad, to this end, also claimed that all power in Hyderabad was vested not in the nizam but rather in the Muslim people of the state. An October 1947 article from Praja Patrika, a

Telugu weekly, noted that the nizam was “a puppet in the hands of Ittihadul-

Muslimeen. This Association is threatening the Nawab. If he is not going to heed to their advice they will start an armed revolution. The Nawab is frightened. He bowed down before the Ittihadul Muslimeen. Today, the Government is being run by this Association.”95

As the conflict in Telangana gained traction from 1946-7, the Ittehad sensed an opportunity to strengthen its position in the state and began sending

93 For more on the Ittehad see Elliott 1974 and Pernau 2000. 94 NAI, Ministry of States, Political Rehabilitation Branch, 1947, #2(5)-PR/1947, pp. 82, emphasis added 95 NAI, Ministry of States, Political Rehabilitation Branch, 1947, #2(5)-PR/1947

329 militia groups called into the countryside to suppress the communist revolt. By crushing the rebels the Ittehad hoped that this might rouse the

Muslims of the state and strengthen Hyderabad's bargaining position regarding integration with India. For example, V.P. Menon noted that when the nizam tried to meet with him and other delegates from India to discuss integration the

Ittehad organized some 25,000 people to surround the meeting place in opposition (Menon 1956: 329). These Razakars to a large extent operated as anti-Hindu killing squads as most of the peasantry were Hindu. This is evident from witness statements from the period.96 Gurubsappa Gavappa Patil, a villager from Achaler in Osmanabad, noted in 1948: “We hindus in Hyderabad are entirely left in the hands of cruel and ruthless Razakars, where wanton cruel acts know no bounds, and are beyond description.”97 In October 1946 British administrators also wrote to the diwan of Hyderabad, Mirza Ismail98, asking why a group of Hindu pilgrims had been attacked in Hyderabad city.99

Since independence, the roles in Hyderabad have reversed; now the

Hindus have become the dominant group in the state and Muslims have been relegated to poverty and backwardness. The riots which have occurred in post- independence Hyderabad have often focused on 'payback' or 'revenge' for the legacy of Muslim domination during colonialism. As early as 1951, the

96 See APSA, Installment 70, List 4, Serial #35, and Installment 70, List 6, Serial #191 97 APSA, Installment 70, List 6, Serial #2 98 Ismail was, of course, also the diwan in Jaipur – he served in that position before moving on to Hyderabad State. 99 APSA, Installment 70, List 4, Serial #35

330 Government of India reported that over 300 mosques in Hyderabad State had been damaged and 163 had been converted into Hindu temples.100 Communal animosities, therefore, are still a direct result of the changes which occurred during colonialism.

V. Conclusion

This chapter, the third and final qualitative chapter of this dissertation, examined two 'deviant cases' – two princely states, Bastar and Hyderabad, which experience an immense amount of caste and, especially, tribal conflict in the contemporary period. Princely rulers throughout India advanced protectionist policies toward adivasis; therefore, what led to the violence in these regions?

In Bastar, I highlighted that the British came to control the state directly by the 1850s. While the British interfered to some extent in every princely state, colonial interference in Bastar was unique: British administrators brought the forests under their direct control, dismissed several rajas as incompetent, and replaced them not with British Indians, as was generally the custom, but

Englishmen. Once the British came to control the state they pursued the same discriminatory policies toward low castes and tribals here as they did throughout the rest of the subcontinent, although at times even more brutally. Consequently, tribal revolts in Bastar became more frequent and eventually consumed the state.

In Hyderabad, I showcased that the nizam, reigning supreme in the capital of his massive kingdom, did not have power that extended into the countryside.

100 NAI, Ministry of States, Hyderabad Branch, 1951, #1(16)-H/51, pp.

331 The landed elite of Telangana, known as the dora, were the initial targets of the

Telangana Rebellion, and they were like the zamindars of British India in that they had been given proprietary rights over the land in the 19th century. In his dominions, however, the nizam enforced a number of progressive policies toward adivasis and low castes, but these reforms did not affect the rest of the state.

These policies of the nizam are nevertheless important because they highlight one of the principal arguments of this dissertation: that colonial rulers often had sharply divergent ways of thinking about disparate ethnic groups. Adivasis in

Hyderabad, a small and weak group, became a most-favored ethnicity whereas

Hindus, who formed almost 90% of the population, were brutally repressed.

Looking over both cases in the post-colonial period, I then argued that the new Government of India did not effect reforms in either region. In Bastar, the post-independence government followed their colonial predecessors in uncanny ways; for example, in keeping state control over forests, and dismissing a raja,

Pravir Chandra Bhanj Deo, under the pretext of mental illness. In Hyderabad, the

Government of India reversed most of the land reforms initiated by the communists who had spearheaded the Telangana Rebellion. Today, the region is still the most backward in the entire state, and also the site of frequent caste and tribal conflict.

Finally, I then also briefly discussed the state of religious conflict in both areas. As Bastar was essentially a British province, I showed that the lack of communal violence in the region may have been due to the British policy of

332 appointing Muslims to important positions in the princely administration, despite the fact that the population was overwhelmingly Hindu. In Hyderabad, on the other hand, nizam Osman Ali Khan – especially in his directly-controlled areas – enforced a number of discriminatory policies against Hindus. The legacy of this discrimination still plagues the city today in the form of constant communal animosities.

The next and final chapter of the dissertation is the conclusion. In it, I reassess the primary arguments from the various preceding chapters, and I also sketch out a test of the comparative implications of this project, examining whether my argument about colonialism and patterns of ethnic violence in India is applicable to other post-colonial states around the world. I argue that because the 'Indian model' of colonialism was later exported to colonies in various states in Africa and Asia, this dissertation can help explain patterns of ethnic violence in a number of post-colonial polities round the world.

333 Conclusion The Indian Model of Colonialism

“...surely the oddest political set-up that the world has ever seen.” – Francis Wylie, Political Advisor to the Viceroy, on bifurcated colonialism in India, 19701

“The existence of the Raj...made possible British imperial conquest, control, and governance across a wide arc of territory stretching from Africa to eastern Asia. Further, with India placed at its center, the British Empire has to be imagined differently… [it] is not just one among many British colonies, or a 'periphery' of the capitalist system, or a land of 'subalterns' struggling to be free. It is in addition a nodal point from which peoples, ideas, goods, and institutions – everything that enables an empire to exist – radiated outward.” – Thomas Metcalf2

This dissertation project has endeavored to explain why ethnic conflicts in multi-ethnic states revolve around one ethnic identity rather than another. Why is it the case, for example, that religion may form the axis of conflict in one region of a plural state, but language forms the axis of conflict in other areas? What explains these patterns of ethnic conflict? This is an under-addressed question of importance for a number of multi-ethnic states around the world struggling to limit ethnic bloodshed. This dissertation answered this central question through an investigation of India, among the most ethnically diverse states in the world.

Specifically, I analyzed variations in three different types of ethnic conflict – caste, tribal, and religious violence.

The principal argument I put forward in this project is that patterns of ethnic conflict in contemporary India can be traced to political developments

1 Quoted in Kooiman 2002: 17 2 2007: 1

334 from the colonial period, specifically the different ethnic policies implemented in areas the British directly controlled (provinces), and areas which remained under the control of native kings (princely states). Through a quantitative study of 589

Indian districts, as well as qualitative archival work and elite interviews conducted in six carefully selected case studies, I found that former British provinces experience more contemporary caste and tribal conflict whereas former princely states experience more contemporary religious conflict. I argued that this dichotomy in ethnic violence outcomes was due to two factors: first, the disparate way in which British administrators and princely rulers constructed ethnic policies (Figure 1); and second, the failure of the post-independence

Indian government to reform the vestiges of its colonial past.

Figure 2: Ethnic Politics in Colonial India BRITISH PROVINCES PRINCELY STATES

High Castes Dominant Religious Groups

Low Castes / Tribals Minority Religious Groups

Favorable Policies Discriminatory Policies Protectionist Policies This conclusion is organized into four main sections. In the first section I summarize the major arguments presented in all of the chapters of this project.

In the second, I present the major academic contributions of this project, as well

335 as several social policy implications. In the last two sections I then switch to the comparative implications of this dissertation; in the third section I focus on the construction of a specific 'Indian model' of colonialism, and in the fourth and final section I examine two other colonial states which were heavily influenced by the British experience in India: Burma and Malaysia. I end by considering whether ethnic conflict outcomes in these two states can be better understood by extrapolating insights from the colonial period in India.

I. Summarizing the Arguments

In the Introduction I laid out the basic puzzle driving this dissertation: explaining patterns of ethnic conflict in multi-ethnic states. There are a variety of ethnically plural states in Europe, Asia, and Africa in which ethnic conflicts have taken thousands of lives, yet understanding why one axis of conflict takes precedence over another has thus far presented a major challenge for social scientists. In contrast to existing rationalist accounts which emphasized the importance of ethnic group sizes (Chandra 2004, Posner 2005), I argued that historical legacies are the key determinant of conflict patterns. I then turned to a critique of existing literature, particularly a very recent and exciting research agenda on Hindu-Muslim riots in contemporary India. While this literature has illuminated fascinating aspects of modern violence, I highlighted two major shortcomings of existing work: a lack of historical depth, and an exclusive focus on religious violence. This is a problem for two reasons: first, ethnic violence in

India is not simply a modern phenomenon; and second, religious violence

336 appears to be less prevalent in the country today than caste and tribal conflicts, which have not been analyzed in depth. My dissertation, by taking a longer historical time frame and considering three different forms of ethnic conflict, therefore significantly advances the existing literature. I then proceeded to detail the methodology of this project, which utilized a mixed-methods research design.

I conducted both a quantitative analysis of 589 districts, as well as 15 months of archival work and elite interviews in six different case studies in India.

Part I of the dissertation focused broadly on theory and quantitative evidence, and consisted of Chapters 1 and 2. The first chapter laid out in extensive detail an historical theory of colonialism and ethnic conflict in India.

Beginning with a broad treatment of the rise of the British in India, I explained why colonial administrators were not able to take control of the entire country as they had in many other colonies. The Rebellion of 1857, led initially by disaffected native soldiers and backed by a number of princely states, abruptly stopped the expansion of British colonialism in its tracks. This rebellion also fundamentally locked in place the boundaries between British India and princely India. This critical historical juncture had long-term effects for patterns of ethnic conflict in the country, even into the modern period. British rulers in the provinces and native kings in the princely states had very different conceptions of ethnicity, which I defined as a useful heuristic for ordering a complex social world. In the provinces, British administrators were concerned primarily with caste, understanding it as the central organizing principle of Indian society. They then

337 enforced policies which benefited high castes, but discriminated against low castes and tribals. At the same time, the British were a secular power, and pursued a policy of religious neutrality, meaning that they enforced protectionist policies for religious minorities, primarily Muslims. Ethnic policies in the princely states, on the other hand, were the complete opposite of British India.

Princely rulers derived their legitimacy from religion, and religious identities formed the basis of princely society. This view was held by native kings themselves, but was also reinforced by British administrators for two reasons: they believed that traditional India was defined by religion, and they wanted to create a clear dichotomy between modern, secular provinces and traditional, religious native states. Princely rulers generally enforced policies which benefited coreligionists and discriminated against non-coreligionists. At the same time, princely rulers pursued protectionist policies for low castes and tribals. This dichotomy in ethnic policies created the main fault lines of conflict in India. I then ended Chapter 1 by discussing why these patterns persisted into the modern period. Foremost, I highlighted that the post-independence Indian government actively chose to keep large parts of its colonial inheritance intact. Furthermore, I argued that despite a wealth of reform efforts, most of them were ultimately stalled or not implemented, and that through a mechanisms ethnic conflicts stemming from the colonial period were reproduced in the modern context.

The second chapter tested this historical theory using a quantitative study of 589 modern Indian districts. Using a variety of datasets and regression

338 models, I found that British rule correlated positively with caste and tribal conflict but negatively with religious conflict, while controlling for a number of alternative arguments. In other words, British rule led to more caste and tribal conflict, but princely rule led to more religious violence. In substantive terms, my statistical models showed that being a former British province increased Naxalite casualties (caste and tribal) by 228%, whereas being a former princely state increased Hindu-Muslim riot casualties by 52%. I also briefly tested some mechanisms associated with caste and tribal conflict, finding that British rule correlated positively with increased caste antagonisms and landlessness, therefore providing evidence for the importance of two key drivers of conflict which I posited in Chapter 1.

The third chapter then marked the start of Part II of the dissertation, which dealt overall with three carefully selected qualitative pairs of case studies.

In chapter 3 I presented the results of fieldwork conducted in the first pair of cases: Jaipur and Ajmer districts in Rajasthan. These districts constituted a controlled historical comparison – a longitudinal comparison of two territories that are similar in most respects except for one key independent variable: the nature of colonial rule. While Jaipur and Ajmer are similar in terms of geography, language, ethnic group composition, etc. they differ in their historical experience.

Jaipur remained a princely state during colonialism whereas Ajmer came under the direct control of British administrators, the only state in the whole of the territory of Rajputana to do so. I further argued that Ajmer became a British

339 province not for any reasons related to ethnic violence, but rather because it was placed in the center of Rajputana and along an important trade route.

Then, using evidence collected from a number of data sources, as well as elite interviews in both districts, I highlighted that in the post-independence period Jaipur experiences more religious violence, whereas in Ajmer, caste forms the main axis of contemporary ethnic conflict. After considering several existing arguments, none of which could explain the particulars of these two cases, I then delved into the historical experiences of both territories. Using a variety of archival material I showcased that the historical theory I laid out in Chapter 1 applied to Jaipur and Ajmer. The Hindu rulers of Jaipur pursued policies that benefited Hindus, discriminated against Muslims, and protected low caste and adivasis. In neighboring Ajmer, however, British administrators pursued policies that benefited high castes, discriminated against low castes and tribals, and protected Muslims. It was these divergent policies which constituted the colonial roots of conflict.

The fourth chapter then presented the results of fieldwork conducted in

Kerala, this time a state-level comparison of the British north (Malabar) and the princely south (Travancore). These two areas also constituted a controlled historical comparison – British administrators themselves believed that the boundaries between Malabar and Travancore were an 'accident'. Again, north and south Kerala are incredibly similar territories in terms of geography, language, ethnic group composition, etc. But the two cases are also unique – I

340 presented extensive historical evidence that the British were interested in conquering Travancore but, for a variety of reasons, were unable to do so.

Travancore exactly fit the model of a territory in which the British would be interested – unlike most other princely states, it was a coastal, agriculturally productive area. Therefore, the fact that Travancore remained independent from

British rule was a quasi-random historical outcome, and the Malabar-Travancore cases are an even stronger comparison to overcome endogeneity problems.

First, I analyzed patterns of ethnic conflict in contemporary Kerala using a variety of data sources as well as qualitative interview evidence. In Malabar, caste and tribal identities make up the dominant narrative of conflict. The Communist

Party of Kerala (a lower caste party) has its regional base in the north, and this area is the only part of the state which has experienced tribal conflict. In

Travancore, on the other hand, the BJP and RSS has its main headquarters in this region. Caste conflict is minimal, but religious violence– involving Hindus,

Muslims, and Christians – has remained a persistent problem.

I then delved into the history of both regions, and the historical theory which I presented in Chapter 1 again fit the specifics of the Kerala case. In

Malabar, British administrators privileged the high caste Jenmis while discriminating against low castes and tribals. However, the British policy toward

Muslims was different from the rest of India – while initially protective of the

Muslim minority, the Mappillas, the region briefly experienced an immense amount of communal violence. I argued that this violence was actually primarily

341 agrarian caste conflict reformulated as communal tension; most of the Mappillas were recent low-caste converts, and worked primarily in agriculture.

Furthermore, their attacks throughout the late 19th and early 20th centuries were directed almost always against high caste Hindus. In Travancore, on the other hand, religion formed the basis of princely state, as the Travancore government linked with Hindus, tried actively to subvert the power of Muslims and

Christians, and advanced some of the most progressive policies in all of India for low castes and tribals. Therefore, the divergent colonial experiences of Malabar and Travancore were the main driver of ethnic conflict patterns in the contemporary state of Kerala.

The results of fieldwork from the last pair of case studies were presented in

Chapter 5. In this chapter I examined two deviant cases: the former princely states of Bastar and Hyderabad, both of which experience unusually high levels of contemporary caste and tribal conflict. The main proxy for this kind of ethnic conflict which I utilized was the Naxalite movement, a rebellion against the

Indian state driven primarily by SCSTs. Both Bastar and Hyderabad have been major centers of Naxalite conflict for decades, going back to the colonial period.

However, the existence of these deviant cases was explained by two facts: first, in

Bastar, the British essentially came to completely control the state by the mid-19th century. Bastar was unique in that the British were excessively involved in its internal affairs – one example of this is the fact that Bastar was one of the larger princely states in India and yet it was not given any gun salute, a kind of

342 ceremonial ranking bestowed by the British government. Prior to British arrival in Bastar there was no major history of tribal conflict in the region, yet from the mid-19th century onwards the state became the major center of tribal conflict in

India. British administrators, driven by Bastar's mineral wealth and lush forest produce, discriminated brutally against the local tribal and low caste communities. At the same time, the British protected Muslims in the area – several Muslims served in important government positions despite being a minuscule portion of the local population. In Hyderabad, on the other hand, caste and tribal conflict was driven not by the policies of the princely ruler, the nizam, but by an independent class of landlords in the Telangana region named the dora. They were, unlike landholders in most other princely states, given proprietary rights over the land like the zamindars of British India. But I showed that the nizam, in the areas he did control, pursued the general policies of princely rulers – he enforced beneficial policies for coreligionists (Muslims), discriminated against non-coreligionists (Hindus), and tried to protect low castes and adivasis. Finally, I ended Chapter 5 by noting that – perhaps more than any other pair of case studies – Bastar and Hyderabad personify the failure of the post-independence Indian government to reform its colonial past. In Bastar, for example, the new Congress government pursued, often in uncanny ways, the exact same policies of colonial rulers. Similarly, the Congress also stalled reforms in Hyderabad, preventing the minimization of violence in that area.

Overall, Chapters 3-5 provided solid empirical evidence for the primary

343 argument of the dissertation: in six varied cases spread out over very different parts of India, colonial administrators and princely rulers pursued markedly different ethnic policies. In British India, administrators were primarily concerned with caste, benefiting high castes, discriminating against low castes and tribals, and protecting religious minorities. And princely rulers were primarily concerned with religion, benefiting coreligionists, discriminating against non-coreligionists, and protecting low castes and tribals. And in all of the case studies, the post-independence government failed to effectively reform its colonial past, and in many cases actively chose not to do so, ultimately ensuring that patterns of ethnic conflict persisted into contemporary Indian politics.

Furthermore, due to a variety of mechanisms – local institutions and power structures which reinforced colonial-era ethnic divides, for example – patterns of ethnic conflict were also actively reproduced over time.

II. Academic Contributions of the Project

This dissertation project makes contributions to a number of academic literatures. I highlight three specific research programs here: studies of ethnicity, ethnic conflict, and colonial legacies. First, in this project I have shown the utility of treating ethnicity as a heuristic for reducing the complexity of the social world

(Hale 2004, 20083). Ethnic markers are more useful than other kinds of social

3 Hale's theory is applied primarily to cases in Eurasia; therefore, I view this project as a kind of test of his theory and its applicability to other cases. I find that his central argument that ethnicity is about uncertainty reduction offers a powerful explanation both for how ethnic identities were constructed in colonial India and why patterns of ethnic conflict continue to persist in the modern state. This view of ethnicity is also somewhat similar to the work of other authors, such as Rogers Brubaker, for example.

344 identities because members of the same ethnicity share myths of common origin, a sense of a common fate, symbols and culture, physical similarities, and they face reduced barriers to communication. This definition of ethnicity captures an under-addressed aspect of the concept: ethnicity connotes the way in which individuals think about and subsequently make sense of the social world.

This is a rather unorthodox way of conceptualizing ethnicity, as much of the recent ethnic politics literature has tended to adopt 'thinner' interpretations in which ethnicity is treated as primarily instrumental.4 Instead, in a push back of these dominant constructivist accounts, I argue that while ethnic identities are socially and historically constructed, because they are essentially ways of thinking about and interpreting the world they are not as easily changeable as constructivist scholars sometimes suppose. Ethnic identities (and by extension patterns of ethnic conflict) in India which were initially forged during the colonial period have in fact been highly resistant to change over time. Certain narratives of ethnicity become embedded in local institutions and power structures, and certain kinds of ethnic violence become legitimized whereas other kinds are de- legitimized (Laitin 1986).

Second, this project can also advance our understanding of ethnic violence. Empirical studies of ethnic conflict have generally focused on the causes of violence; for example, take the long-standing greed versus grievance debate

(Collier and Hoeffler 1998, 2004).5 However, this project has pursued a different

4 For example, see Chandra 2005, Posner 2005, Chandra 2012. 5 This debate is actually about the causes of civil war, but it has been applied to ethnic violence.

345 but related puzzle: explaining why ethnic conflicts revolve around one identity rather than another. This question has been analyzed previously by scholars

(Laitin 1986, Posner 2005), but Laitin's study of the salience of ethnic cleavages in Yorubaland and Posner's study of the same in Zambia came to two very different conclusions. The former focused on the impact of colonial rule in structuring ethnic conflicts whereas the latter focused on group sizes in a given area and how political coalitions form around them for instrumental purposes. In one sense, the Indian case therefore offers an opportunity to test these two competing theories and how well they apply to cases outside of Africa. Do patterns of ethnic conflict in India derive from history or merely from the calculus surrounding the constellation of ethnic groups in a particular area?

If we position these two theories side by side for comparison in the Indian context, I have found a significant amount of evidence in support of Laitin's argument. This is not to say that the size of ethnic groups or electoral systems in a region – or rationalist motivations more generally – do not matter, but rather that they do not seem to be easy or straightforward determinants of conflict. For example, almost every district in India is Hindu majority, meaning that it would always behoove members of the Hindu ethnic group to emphasize this identity to divide themselves from other religious groups and remain in the majority.

However, the Hindu-Muslim axis only forms the basis of conflict in a certain number of areas, and in other areas Hindus are bitterly divided by caste

See Fearon and Laitin 2003, Cederman et al. 2010.

346 distinctions. In yet other areas tribal or linguistic identities form the main axes of conflict. This project therefore argues that historical legacies play the main role in producing conflict; the fault lines of ethnic conflict are not simply determined by the particular constellation of ethnic groups in an area, but by the inheritances of the past.

Finally, this project can make some important contributions to the broad multidisciplinary literature on the legacies of colonialism. Generally this literature has focused on making broad comparisons: for example, comparing one kind of European colonialism with another (Blanton et al. 2001, Lange et al.

2006, Lange 2010, Lee and Schultz 2012), or comparing settler colonies versus extractive colonies (Acemoglu et al. 2001, Engerman and Sokoloff 2002).

Instead, I have focused on a relatively under-studied type of colonialism: a hybrid model which combined governance by colonial administrators and native rulers.6

As I detail later in sections III and IV, there were a specific number of colonies within the British imperial system whose governance structures were modeled directly from the Indian experience, colonies in which both British administrators and native authorities divided power. These particular colonies are, as I noted in the Introduction, pivotal to researchers because they allow us to potentially isolate the effects of colonial rule. In India, for example, we can at the very least compare the provinces, areas where the colonial footprint was pronounced, with the princely states, areas where colonial administrators had

6 See Lange 2010: 29.

347 much less power. It is not a perfect comparison, but colonies like India allow researchers the opportunity to investigate seriously the impact of colonial rule side by side with its relative absence.

2.2 Social Policy Implications

Finally, there are a number of social policy implications of this project. At first glance, however, the results of this dissertation seem to paint a rather bleak picture for stopping ethnic conflict. If the salience of particular ethnic categories are historically constructed and patterns of conflict can persist for hundreds of years even after their initial constitution, then undoing these patterns and, more importantly, stopping the violence seem like tall orders indeed. Nevertheless, there are a number of insights from this project that may assist in minimizing ethnic conflict.

Perhaps most importantly, policymakers must recognize that the fact that ethnic identities are socially constructed does not mean that ethnicity is epiphenomenal. For example, Posner's (2005) study seems to suggest that changing the institutional rules (i.e. the electoral system) can have a major impact on ethnic identification. There are two problems here. First, as I noted in

Chapter 4, party systems can often be based on ethnic cleavages and not the other way around – this certainly seems to be the case in Kerala. Second, even if ethnic identities are epiphenomenal, it is not clear that this insight can explain ethnic violence outcomes. There are a number of scholars who argue that ethnic identities are hardened by conflict (Kaufmann 1996, Byman 2000, Van Evera

348 2001). Therefore, policymakers must recognize that – in the words of Donald

Horowitz – “[t]he ethnic group is not just a trade union” (1985: 104). Ethnic identities are often deeply-held, internalized, and difficult to minimize in importance. Therefore, only once the basis behind ethnic identities are understood can ethnic policies then be formulated.

Second, and along similar lines, policymakers should also understand that the conditions under which ethnic identities were constructed, often during colonial rule or under other similarly unique historical periods, are not easily replicable today.7 Therefore, suggestions that post-colonial states simply emphasize or de-emphasize one ethnic identity over another may be unrealistic, at least without other policies discussed below.8 The power of the post-colonial

Indian government, for example, to change or manipulate ethnic identities is not comparable to the changes to ethnic identification which occurred during the colonial period. For instance, in comparing colonial state formation with state formation in Europe James C. Scott has noted that “[m]any of the same state objectives animate both [cases], but in the colonial case, the state is at once more bureaucratized and less tolerant of popular resistance” (1998: 69). The 19th century in India witnessed a massive colonial project to classify, organize, and generally order Indian society, and the result was a wholesale shift in relations

7 Along these lines, Keith Darden (2012) argues that the first identity constructed and instituted by a state during periods of mass literacy has a distinct first-mover advantage. These identities, once introduced, are hard to displace over time. 8 And even if states can do this, it is not clear it is beneficial. Wilkinson (2008), for example, argues that the post-colonial Indian state's decision to not consider claims made by minority religious groups has led to increased religious violence.

349 among various ethnic groups in society. The post-independence period, on the other hand, mostly maintained existing relationships rather than re-creating them. In short, the fact that ethnic identities are constructed does not mean that they can be easily reconstructed. Most of the research for this project suggests that ethnic affiliations persist long after their initial construction. Therefore, policymakers must deal with ethnic groups 'as they are' rather than try to re- create or change existing ethnic divides.

Finally, this project was more concerned with explaining patterns of conflict rather than addressing the proximate causes of violence, but there are some interesting points to consider in sketching out the actual determinants of ethnic bloodshed. My research indicates that some of the primary drivers of violence are socioeconomic factors, which in this case stem from legacies of discrimination. This is especially true for 'bottom-up' ethnic conflict, which is driven by minority ethnic groups. In my quantitative analysis in Chapter 2, for example, many of the control variables related to economic and social development were significantly related to conflict in my various statistical models. Similarly, fieldwork in my case studies revealed that the poor socioeconomic conditions of minority ethnic groups in a region played an important role in producing conflict. Therefore, improving the socioeconomic position of minority groups ought to be a principal concern for policymakers.

350 III. Creating the 'Indian Model'

India was the crown jewel of the British Empire, a term which referred to its vast resources and mineral wealth. But India offered much more to the British

– as Thomas Metcalf has written: “The existence of the Raj...made possible

British imperial conquest, control, and governance across a wide arc of territory stretching from Africa to eastern Asia” (2007: 1). Once India came under their control the entire Indian Ocean was opened up to British administrators.9 By controlling the Indian Ocean the British gained access to a vast new set of markets and commercial possibilities. Therefore, India is perhaps best understood as the capital of the British Empire in the east – without it, colonial expansion in a wide array of states would not have been possible.

Aside from the men who governed the Raj the British public in the early

20th century also seem to have understood the importance of India within the greater imperial system. As The Times noted on 'Empire Day' in 1911:

“India stands right across the greatest highway in the world; it is the centre of the East. Through its possession we secured our great predominance in Eastern trade, and from its shores we extended our interests to Australasia, the Malay peninsula, the Pacific Islands, and the Chinese coast...It is, therefore, in a special and peculiar sense the centre of Imperial defence.” (Times 1911: 3)

While its geographical position was pivotal to the British, India was also a unique colony for another reason: it provided a model for later colonial expansion. By this I mean that prior to their rule in India the British had generally brought territories under their direct control. But the experience of

9 On the importance of the Indian Ocean see Bose 2006.

351 colonialism in India changed the way in which later colonial expansion unfolded.

At the dawn of the colonial period the British probably hoped to extend their power across the entire subcontinent (Copland 2002: 15), but the Rebellion of

1857 suddenly upended these plans. The rebellion, which took an entire year to defeat, locked in place a hybrid system of colonial rule: direct rule in some areas coupled with native rule in others. Traditional rulers, in this case the princes, therefore came to occupy a significant position in governing India after 1857.

By the late 19th century British administrators in India had discovered the virtues of this model of colonialism which a century earlier no one predicted would exist. While this may have not been what the British ever intended, administrators came to see over time that there were a number of key benefits to this system. Michael Fisher writes:

“The perceived success of the Residency system of India later inspired officials in London, and in the field elsewhere in Asia and in Africa, with the idea that imperial expansion need not, and indeed should not, draw heavily on the limited resources of the home nation. Advocates of British expansion, and opponents of greater British involvement as well, began to support the application of an Indian model.” (1991: 1)

After the British had conquered India they turned their eyes toward other parts of

Africa and Asia – for example, the end of the 19th century witnessed the 'Scramble for Africa' among the major European powers. But the singular lesson learned from the Indian experience was that incorporating local elites was necessary (to at least some extent) in the process of state building and governing foreign territories. As the British Empire expanded into Africa and other parts of Asia the

352 British began to rely on the system of keeping a number of native rulers in power in certain parts of the country (Fisher 1991, Mamdani 1996, Metcalf 2007). This process only gained added importance in the 20th century as British imperial power declined amidst two world wars and the Great Depression.

What led to the acceptance of and even praise for this model of colonial rule? There are a few key reasons. First and foremost, it was cheap. Rather than investing heavily in governing and policing a colony, the British were able to extract resources and pursue commercial exploits with a minimal drain on their treasury. Second, many colonial states – especially India, and more generally those in Africa and Asia – contained very rough terrain, a host of new tropical diseases, and were generally not easily inhabitable for British settlers. Therefore, making large investments in manpower in these colonies was impractical and even dangerous. Third, British administrators realized that utilizing native rulers could provide a form of legitimacy which white settlers never could. As Sir John

Malcolm, the Governor of Bombay, once famously stated:

“If we made all India into Zillahs [British districts] it was not in the nature of things that our Empire should last fifty years; but that if we could keep up a number of native states without political power, but as royal instruments, we should exist in India as long as our naval supremacy was maintained.” (Assary 2009: 90)

This mentality, forged initially in India, was then later utilized in other British colonies, as the incorporation of native authorities in the process of governing became more widespread in the late-19th and early 20th centuries. Lastly, utilizing native authorities from the colony as rulers also lessened (but certainly did not

353 eliminate) the risk of rebellion against colonial rule. The British experience in

India highlighted that princely rulers, if kept in power and generally kept happy, would not then unite and attempt to overthrow the status quo. After the

Rebellion of 1857 in India, princes were devolved substantial internal autonomy and the British began to think of them not as an obstacle to colonial rule, but rather an important tool for its continuation.

It is important to recognize too that the export of the Indian model was assisted both by the export of British administrators who had served in India and thousands of Indians themselves. Many colonial administrators who had cut their teeth in India were later sent to colonies in the larger Indian Ocean region. With them came the specific experiences of Indian colonialism, especially the importance of involving traditional elites in the process of governance. For example, Lord Lugard, who is considered the father of combining direct and indirect rule in Nigeria during his tenure as Governor-General, was born in India and served in a number of military campaigns there.10

To be sure, the Indian model was not exported everywhere; in Natal in

South Africa, for example, the British kept the entire state under indirect rule.

Local conditions always determined British colonial policy. Nor was the Indian model applied indiscriminately to other colonies. For instance, in India the

British emphasized caste as the foundation of politics in the provinces, but in

David Laitin's study of Yorubaland he finds that colonial administrators in

10 See Laitin 1986: Chapter 7 for more on Lugard.

354 directly-ruled areas emphasized ancestral city membership. Therefore, the Indian model always contained slight gradients in its application to other colonies. But the divide between the 'traditional' and the 'modern' was always present, as was the use of traditional authority figures: it began in India with the princes, then in

Africa the chiefs, in Burma the Sawbwa11, and in Malaysia the sultans.

IV. Comparative Implications: Burma and Malaysia

In a particular class of British colonies, direct and indirect rule were combined together simultaneously, and governance in these states was modeled directly from a century of British colonialism in India. Two important such states were Burma and Malaysia. In the next two sections I offer a cursory investigation into whether bifurcated colonial rule in these areas might have produced differentiated patterns of ethnic conflict. To be clear at the outset, the Indian model refers mainly to a system which combined direct and indirect rule. I hypothesize that this distinction, where it existed elsewhere, had a major effect on ethnic relations, as British administrators and native authorities would be likely to think about ethnicity and ethnic policies differently. Therefore, I ultimately argue that ethnic violence outcomes across British-ruled and native- ruled territories will be different. The expectation is not that religious violence is always less of a problem in former areas of direct British rule – this is only a hypothesis based on India, but one that can be tested empirically using data on ethnic violence in other post-colonial British states.

11 A term that roughly means king in Burmese.

355 Although I focus on analyzing the significance of the Indian model with reference to two cases in Asia, this model of colonial rule was also extremely important to Africa. I have already made several references to Laitin's study of

Yorubaland, but similar distinctions between 'traditional' and 'modern' rule were institutionalized elsewhere in Africa – for example, in Kenya and Nigeria. In

Kenya (known then as the British East Africa Protectorate), British administrators allowed a larger white settler population to control the highlands, but governed other major parts of the country through the use of native chiefs12

(Throup 1985). Similarly, in southern Nigeria the British made use of traditional rulers to govern the territory (Afigbo 1972). Ultimately, however, I argue that the specific model of combining direct and indirect rule in a non-settler colony is best illustrated in the Asian cases of Burma and Malaysia.

Finally, there is one last point before I begin the ensuing discussion – it is important to note that the comparative implications of this project do not extend only to post-colonial states. While India was a key British colony and the workings of the Indian model were exported to other colonies elsewhere, this dissertation project in a broader sense is about the role of history and historical legacies in structuring and conditioning ethnic conflicts. The role of the state in promoting new ways of classifying and organizing its varied populations is the subject of a broad and important literature in comparative politics (Gellner 1983,

Anderson 1983, Scott 1998). This extends to ethnic policies; states in Europe, for

12 In Africa the term 'warrant chiefs' is often used to refer to native administrators picked by the British.

356 example, have had many national identities constructed through specific historical policies. Other ethnic groups may also have been constructed in the same way, and the role of historical legacies should be an area of focus for researchers of ethnic conflict more broadly. Also, while colonial states may have vanished from the international system, authoritarian states with similar levels of coercive capacity have not. Thomas Mullaney (2011), for example, has argued that the Chinese communist state played the key role in classifying the diverse ethnic population of the country into the 56 main ethnic nationalities existent today. As constructivists have rightly argued, ethnic identities around the world are historically constructed. To date, however, this insight has not been systematically applied to the question of ethnic violence and patterns of conflict.

In short, if ethnic groups were historically created then patterns of ethnic violence also come from the legacies of the past.

4.2 Ethnic Conflict in Burma13

Burma is an intriguing test case for the argument that the Indian model of colonialism was exported elsewhere because Burma was once itself part of the

British Indian Empire, but today is an independent polity. In this sense, it shared many similar experiences with India, but its post-independence trajectory has been divorced from that of its neighbor. Do patterns of ethnic violence in Burma, like India, still mirror the colonial period? And did Burma also refuse to tackle its colonial legacy, or did it, on the other hand, actively attempt to reform its past?

13 Also known as Myanmar; I use the term Burma throughout.

357 From 188614 until 1937 Burma was officially a province of the British Raj.

'Lower Burma' was first incorporated into Indian territory, then it was followed by 'Upper Burma'. Like India, Burma experienced a bifurcated model of colonial rule (Lange et al. 2006: 1430). Roughly two-thirds of Burma and the vast majority of its population was controlled directly by British administrators, but the remaining portion of the country was governed by native rulers. Map 8 below indicates the division between British and native Burma. The areas which constituted the former (sometimes known as Ministerial Burma) included the

Tenasserim, Arakan, Pegu, and Irrawaddy Divisions. Native Burma (also known as the Scheduled or Frontier Areas) consisted of the Shan States, Chin Hills, and

Kachin State.

14 The exact date is debatable; British interference in Burma began in 1824 with the First Anglo- Burmese War. This led to the ceding of several parts of Burma. More conflicts ensued and by 1886 Burma became an official part of the British Indian Empire.

358 Map 8: Colonial Burma15

British rule in Burma obviously mimicked the processes of state-building simultaneously occurring in India in many ways. For example, in their Burmese provinces the British abolished the traditional Buddhist monarchy, and – most importantly – tried to de-emphasize the importance of religion. Buddhism was no longer the official state religion, and colonial administrators re-organized the entire system of primary education which had hitherto been run by monasteries and provided religious instruction. Instead, the British sought to establish a secular educational system which, they hoped, would produce a modern, secular populace. Therefore, it is right to state that the British – just like in India and 15 British Burma is the white area in the middle; native states are colored and on the periphery.

359 Yorubaland – once again tried to enforce policies that 'expunged' or diminished the importance of religion.

Information on ethnic violence in Burma is still in the process of being compiled, but it is worth noting that peasant revolts in the state occurred primarily in British Burma, which, given the history of British India, is exactly what we would expect. The major peasant revolt – led by the Buddhist monk Saya

San in 1930-31 – began in directly-ruled Tharrawaddy District (Ghosh 2000).

The causes of the violence also seemed to mirror the discontent of the peasants of

British India. As Robert H. Taylor notes, “[San] developed peasant support for opposing the state's restrictions on the use of timber and bamboo by the villagers, a traditional privilege that the colonial state had appropriated as a monopoly”

(2009: 199). Recall from Chapter 5 that I highlighted a major source of tribal discontent in Bastar as the British monopoly over access to forests and rich forest produce.

Native Burma, on the other hand, remained under the control of indigenous rulers, and ethnic politics was organized quite differently. First, native Burma seemed to be similar to princely India in its administrative framework. As Henry Dodwell notes about the power of native rulers:

“...major chieftains in the Shan States retain powers of life and death, and administer their native customary law, not the English codes.” (1932: 446)

In native Burma both indigenous rulers as well as the British emphasized the traditional, which in this case meant a distinct group of ethnic minority

360 identities.16

However, religion did not form the basis of native Burma as was the case in the princely states. Rather, native rulers – known as the Sawbwa – generally enforced chauvinistic policies that benefited their own specific ethnic minority group at the expense of others. For example, Karen leaders sought beneficial policies for the Karen, and Kachin leaders sought beneficial policies for the

Kachin. This bred hostility between different ethnic groups, hostility which still plagues the region to this day and has led to a number of ethnic separatist movements located throughout this area (Smith 1994: 22-23).

Therefore, the Burmese case highlights that the Indian model was adopted by British administrators in the region. British Burma and native Burma existed side by side, but ethnic politics in both regions was conducted and organized in highly disparate ways. British administrators and native rulers pursued policies in their respective regions that benefited and harmed different sets of ethnic groups; the result is that the former areas have experienced widespread tribal discontent whereas the latter territories have experienced violence organized around ethnic minority identities.

16 Martin Smith (1994: 36) states that most of these groups find the term 'tribe' to be pejorative. In general, the native Burmese – the Chin, Kachin, and Karen – are known simply as ethnic minorities.

361 4.3 Ethnic Conflict in Malaysia

During the colonial period in Malaysia the country was also divided between systems of direct and indirect rule. Two areas, the Straits Settlements and the Federated Malay States, were placed under the direct control of British administrators. The Straits Settlements consisted of Malacca, Dinding, Penang,

Singapore, and Labuan; the Federated Malay States consisted of Selangor, Perak,

Pahang, and Negri Sembilan. The Unfederated Malay States, however, were governed by native sultans with varying levels of internal autonomy (Emerson

1937). These areas consisted of Johor, Kedah, Kelantan, Perlis, and Terengganu.

Map 9 below indicates these divisions.

362 Map 9: Colonial Malaysia17

17 Taken from Hirschman 1986: 335

363 Just like in the Indian princely states, the British dispatched 'Residents' to the courts of the native rulers of these states, and there is a similar debate among historians about the power relations between Residents and sultans.

The distinction between British Malaysia and native Malaysia was reinforced via a different set of policies for both regions. For example, the British conducted the first census for their directly-ruled territories in 1871, but it did not occur in the Unfederated Malay States until 1911 (Hirschman 1987: 559). The use of the census in India, which I discussed in Chapter 1, was a pivotal means of classifying British India according to castes and tribes. The fact that official colonial policies toward ethnicity in British Malaysia were crafted 40 years before any such policies occurred in indirectly-ruled areas is an important historical fact with ostensibly long-reaching implications. Most ethnic conflicts which occur in

Malaysia today revolve around race, and it has been argued that “the residue of racial ideology continues to haunt contemporary Malaysia” (Ibid.: 570).

The census, however, was not the determinative factor in producing ethnic violence in contemporary Malaysia. In British Malaysia, it was the widespread use of immigrants from other parts of the empire that led to increased ethnic antagonisms. Chinese and Indian workers were brought to Malaysia to work in the tin and rubber trade. The use of immigrants was also bolstered by the fact that many British administrators in the region held the opinion that native

Malays were naturally lazy and indolent (Shoup 2008: 46). Therefore, administrators in British Malaysia enforced policies which were generally

364 favorable to immigrants, and these areas today have experienced periodic ethnic violence based around racial identities.

In areas where the sultans remained in control, however, immigration was not a widespread attribute of local politics, and therefore racial tensions have been minimized. However, the sultans did tend to emphasize, like the Indian princes, their religious identities. This was, also like India, further reinforced by

British administrators, who sought to uphold the sultanates as a vision of

'traditional' Malaysia. This focus on religion (specifically Islam) has not changed today. As Shanti Nair notes:

“In contemporary times, Malay royalty...have continued to be regarded by some parts of the Malay community as protectors of its privileges and special rights, and as the upholders of Islam.” (1997: 133)

Similarly, Frederik Holst has described the sultans as “a common reference point for ethnicized religious policies” (2012 :46). Therefore, I expect that the former sultanates in Malaysia experience higher levels of religious violence than the directly-ruled British areas, and this pattern would mirror exactly the situation in contemporary India.

Finally, Malaysia is also an interesting comparative case study because, since the bloody 1969 riots in Kuala Lumpur, the country has done an admirable job in limiting major spasms of ethnic conflict (Collins 1998, Sriskandarajah

2005).18 Compared to Indonesia, for example, which has experienced intermittent bloody flare-ups of ethnic violence, Malaysia has been a model for

18 Is is worth pointing out, though, that even in areas where ethnic violence is minimal (Kerala, for example), there still may be clear patterns to the conflict which does occur.

365 managing ethnic tensions. Therefore, the country and its leaders may have differed from India in tackling widespread reforms of colonial-era ethnic policies, and insights from the Malaysian case may provide lessons for the Indian government and governments of other multi-ethnic states.

This discussion on Burma and Malaysia highlights that the bifurcated nature of colonial rule in India was not only exported to other colonies, but that the British and native rulers of these colonies also created disparate ethnic policies which had long-lasting effects on the production of violence. Therefore, the legacies of British rule and the specifics of the hybrid model of Indian colonialism may be central to understanding patterns of ethnic conflict in a wide array of multi-ethnic states surrounding the vast expanse of the Indian Ocean.

366 Bibliography

Acemoglu, Daron, Simon H. Johnson, and James A. Robinson. 2001. “The

Colonial Origins of Comparative Development: An Empirical

Investigation” American Economic Review; Vol. 91(5): 1369-1401.

Afigbo, Adiele Eberechukwu. The Warrant Chiefs: Indirect Rule in Southeastern

Nigeria, 1891-1929. London: Longman 1972.

Almond, Gabriel. 1988. “Separate Tables: Schools and Sects in Political Science”

PS: Political Science and Politics; Vol. 21(4): 828-42.

Anderson, Benedict. Imagined Communities: Reflections on the Origin and

Spread of Nationalism. London: Verso, 1983.

Assary, S. Thulaseedharan. Colonialism, Princely States and the Struggle for

Liberation: Travancore (1938-1948). New Delhi: APH Publishing

Corporation, 2009.

Baden-Powell, Baden Henry. The Land Systems of British India. Oxford:

Clarendon Press, 1892.

Baig, Mra. The Muslim Dilemma in India. New Delhi: Vikas Publishing House,

Ltd., 1974.

Balagopal, K. 2006. “Maoist Movement in Andhra Pradesh” Economic and

Political Weekly; Vol. 41(29): 3183-3187.

Banerjea, D. [ed]. Criminal Justice India Series. Volume 19: Chhattisgarh. New

Delhi: Allied Publishers Private Ltd., 2002.

367 Banerjee, Abhijit, and Lakshmi Iyer. 2005. “History, Institutions, and Economic

Performance: The Legacy of Colonial Land Tenure Systems in India” The

American Economic Review; Vol. 95(4): 1190-1213.

Bates, Robert, and Rui J.P. De Figueiredo Jr., and Barry R. Weingast. 1998. “The

Politics of Interpretation: Rationality, Culture, and Transition.” Politics &

Society; Vol. 26(4): 603-642.

Bayly, C.A. 1985. “The Pre-History of 'Communalism'? Religious Conflict in

India, 1700-1860” Modern Asian Studies; Vol. 19(02): 177-203.

Bayly, Susan. 1984. “Hindu Kingship and the Origin of Community: Religion,

State and Society in Kerala” Modern Asian Studies; Vol. 18(2): 177-213.

. Caste, Society and Politics in India from the Eighteenth Century to the

Modern Age. Cambridge: Cambridge University Press, 1999.

. Saints, Goddesses and Kings: Muslims and Christians in South Indian

Society, 1700-1900. Cambridge: Cambridge University Press, 1989.

Bhattacharjea, Ajit. Kashmir: The Wounded Valley. New Delhi: USB Publishers'

Distributors, 1994.

Bijoy, C. R., and Ravi Raman. 2003. “Muthanga: The Real Story: Adivasi

Movement to Recover Land” Economic and Political Weekly; Vol. 38(20):

1975-1982.

Blanton, Robert, and T. David Mason, and Brian Athow. “Colonial and Post-

Colonial Ethnic Conflict in Africa” Journal of Peace Research; Vol. 38(4):

473-491.

368 Bohlken, Anjali, and Ernest Sergenti. 2010. “Economic Growth and Ethnic

Violence: An Empirical Investigation of Hindu–Muslim Riots in India”

Journal of Peace Research; Vol. 47(5): 589-600.

Bose, Sugata. A Hundred Horizons: The Indian Ocean in the Age of Global

Empire. Cambridge: Harvard University Press, 2006.

Bose, Sumantra. Kashmir: Roots of Conflict, Paths to Peace. Cambridge:

Harvard University Press, 2005.

Brady, Henry, and David Collier [eds]. Rethinking Social Inquiry: Diverse

Tools, Shared Standards. Lanham: Rowman & Littlefield, 2004.

Brass, Paul R. “A Reply to Francis Robinson” The Journal of Commonwealth &

Comparative Politics; Vol. 15(3): 231-234.

. Theft of an Idol: Text and Context in the Representation of Collective

Violence. Princeton: Princeton University Press, 1997.

. The Politics of India Since Independence. Cambridge: Cambridge

University Press, 1994.

. The Production of Hindu-Muslim Violence in Contemporary India.

Seattle: Washington University Press, 2003.

Brubaker and Laitin. 1998. “Ethnic and Nationalist Violence” Annual Review of

Sociology; Vol. 24: 423-52.

Burman, J.J. Roy. 1996. “Hindu-Muslim Syncretism in India” Economic and

Political Weekly; Vol. 31(20): 1211-1215.

Byman, Daniel. 2000. “Forever Enemies? The Manipulation of Ethnic Identities

369 to End Ethnic Wars” Security Studies; Vol. 9(3): 149-190.

Cameron, Adrian Colin, and P.K. Trivedi. Regression Analysis of Count Data.

Cambridge: Cambridge University Press, 1998.

Caporaso, James A. 1995. “Research Design, Falsification, and the Qualitative-

Quantitative Divide” American Political Science Review; Vol. 89(2): 457-

60.

Cederman, Lars-Erik, Andreas Wimmer, and Brian Min. 2010. “Why Do Ethnic

Groups Rebel?: New Data and Analysis” World Politics Vol. 62(1): 87-119.

Chandra, Bipan. Essays on Colonialism. New Delhi: Orient Longman Ltd., 1999.

Chandra, Kanchan [ed]. Constructivist Theories of Ethnic Politics. Cambridge:

Cambridge University Press, forthcoming.

. Why Ethnic Parties Succeed: Patronage and Ethnic Head Counts in

India. Cambridge: Cambridge University Press, 2004.

. 2006b. “Review of Brass, Production of Hindu-Muslim Violence” The

Journal of Asian Studies; Vol. 65(1): 207-209.

. 2006a. “What Is Ethnic Identity and Does It Matter?” Annual Review of

Political Science; Vol. 9: 397-424.

Chatterjee, Partha. Nationalist Thought and the Colonial World: A Derivative

Discourse? Minneapolis: University of Minnesota Press, 1986.

Char, S.V. Dasika. Caste, Religion and Country: A View of Ancient and Medieval

India. New Delhi: Orient Longman, 1991.

Charsley, Simon. 1996. “‘Untouchable’: What is in a Name?” The Journal of the

370 Royal Anthropological Institute; Vol. 2(1): 1-23.

Chaterjee, Partha. Nationalist Thought and the Colonial World: A Derivative

Discourse. Minneapolis: University of Minnesota Press, 1993.

Chhibber, Pradeep K. Democracy without Associations: Transformation of the

Party System and Social Cleavages in India. Ann Arbor: University of

Michigan Press, 1999.

Chhibber, Pradeep K., and John R. Petrocik. 1989. “The Puzzle of Indian Politics:

Social Cleavages and the Indian Party System” British Journal of Political

Science; Vol. 19(2): 191-210.

Chopra, P.N., and B.N. Puri, M.N. Das, A.C. Pradhan [eds]. A Comprehensive

History of India, Volume 3: Modern India. New Delhi: Sterling Publishers

Private Limited, 2003.

Cohn, Bernard. Colonialism and its Forms of Knowledge: The British in India.

Princeton: Princeton University Press, 1996.

. “The Census, Social Stratification and Objectification in South Asia”, in

Banerjee-Dube, Ishita [ed]. Caste in History. Oxford: Oxford University

Press, 2008.

Collier, David, and James Mahoney. 1996. “Insights and Pitfalls: Selection Bias in

Qualitative Research” World Politics; Vol. 49(1): 56-91.

Collier, Paul, and Anke Hoeffler. 1998. "On Economic Causes of Civil War"

Oxford Economic Papers; Vol. 50: 563-73.

. 2004. "Greed and Grievance in Civil War" Oxford Economic Papers ; Vol.

371 56: 563-95.

Collins, Alan. 1998. “The Ethnic Security Dilemma: Evidence from Malaysia”

Contemporary Southeast Asia. Vol. 20(3): 261-278.

Copland, Ian. 1973. “The Maharaja of Kolhapur and the Non-Brahmin Movement

1902-10” Modern Asian Studies; Vol. 7(2): 209-225.

. 1988. “Communalism in Princely India: The Case of Hyderabad” Modern

Asian Studies; Vol. 22(4): 783-814.

. 1991. “The Princely States, the Muslim League, and the Partition of India

in 1947” The International History Review; Vol. 13(1): 38-69.

. 1998. “The Further Shores of Partition: Ethnic Cleansing in Rajasthan

1947” Past & Present; No. 160: 203-239.

. 2002. “The Master and the Maharajas: The Sikh Princes and the East

Punjab Massacres of 1947” Modern Asian Studies; Vol. 36(3): 657-704.

. State, Community and Neighbourhood in Princely North India, c. 1900-

1950. New York: Palgrave Macmillan, 2005.

Connor, Walker. 1993. “Beyond Reason: The Nature of the Ethnonational Bond”

Ethnic and Racial Studies; Vol. 16(3): 373-89.

Dale, Stephen F. 1973. “Communal Relations in Pre-Modern India: 16th Century

Kerala” Journal of the Economic and Social History of the Orient; Vol.

16(2): 319-327.

. “The Mappilla Outbreaks: Ideology and Social Conflict in Nineteenth-

Century Kerala” The Journal of Asian Studies; Vol. 35(1): 85-97.

372 Darden, Keith. Resisting Occupation: Mass Literacy and the Creation of

Durable National Loyalties. Cambridge: Cambridge University Press,

2012.

Datta, Rajat. Rethinking a Millennium: Perspectives on Indian History From the

Eight to the Eighteenth Century, Essays for Harbans Mukhia. New Delhi:

Aakar Books, 2008.

Desai, Manali. 2005. “Indirect British Rule, State Formation, and Welfarism in

Kerala, India, 1860-1957” Social Science History; Vol. 3 (Fall 2005): 457-

489.

. State Formation and Radical Democracy in India. New York: Routledge,

2007.

Deshpande, R.S. and Saroj Arora. Agrarian Crisis and Farmer Suicides.

Thousand Oaks: SAGE Publications, 2010.

Dhanagare, D. N. 1977. “Agrarian Conflict, Religion and Politics: The Moplah

Rebellions in Malabar in the Nineteenth and Early Twentieth Centuries.”

Past & Present; Vol. 74(1): 112-141.

Dirks, Nicholas. Castes of Mind: Colonialism and the Making of Modern India.

Princeton: Princeton University Press, 2001.

Dodwell, H.H. [ed]. The Cambridge History of India, Volume VI. The Indian

Empire, 1858-1918. Cambridge: Cambridge University Press, 1932.

. The Cambridge Shorter History of India. Cambridge: Cambridge

University Press, 1934.

373 Emerson, Rupert. Malaysia: A Study in Direct and Indirect Rule. New York: The

Macmillan Company, 1937.

Ernst, Waltraud, and Biswamoy Pati [eds]. South Asia India’s Princely States:

People, princes and colonialism. New York: Routledge, 2007.

EPW. 1989. “Bastar: Development and Democracy” Economic and Political

Weekly; Vol. 24(40): 2237-2241.

Elliott, Carolyn M. 1974. “Decline of a Patrimonial Regime.” The Journal of

Asian Studies Vol. 34(1): 27-47.

Edmunds, David and Eva Wollenberg (eds). Local Forest Management: The

Impacts of Devolution Policies. London: Earthscan Publications Ltd.,

2003.

Engerman, Stanley L., and Kenneth L. Sokoloff. 2002. “Factor Endowments,

Inequality, and Paths of Development among New World Economies”

Economia; Vol. 3(41): 41–88.

Engineer, Asghar Ali. Communalism and Communal Violence in India: An

Analytical Approach to Hindu-Muslim Conflict. New Delhi: Ajanta

Publications, 1989.

Engineer, Asghar Ali [ed]. Communal Riots in Post-Independence India. New

Delhi: Sangam Books, 1991.

Fanon, Franz. The Wretched of the Earth. New York: Grove Press, 1961/2004.

Fawcett, F. 1897. "The Moplahs of Malabar" The Imperial and Asiatic Quarterly

Review; October 1897: 288-299.

374 Fearon, James D., and David D. Laitin. “Ethnicity, Insurgency, and Civil War”

American Political Science Review; Vol. 97(1): 75-90.

. “Explaining Interethnic Cooperation.” American Political Science

Review; Vol. 90(4): 715-35.

. “Violence and the Social Construction of Ethnic Identity,” International

Organization, Vol. 54(4): 845-877.

Fisher, Michael H. 1984. “Indirect Rule in the British Empire: The Foundations

of the Residency System in India (1764–1858)” Modern Asian Studies;

Vol. 18(03): 393-428.

. Indirect Rule in India: Residents and the Residency System 1764 – 1858.

New York: Oxford University Press, 1991.

Forrest, G.W. [ed.] Selections from the state papers of the governors-general of

India: Warren Hastings Vol. I, 1910.

Frenz, Margret. From Contact to Conquest: Transition to British Rule in

Malabar, 1790-1805. Oxford: Oxford University Press, 2003.

Furber, Holden. 1951. “The Unification of India, 1947-1951” Pacific Affairs; Vol.

24(4): 352-371.

Geddes, Barbara. 1990. “How the Cases You Choose Affect the Answers You Get:

Selection Bias in Comparative Politics” Political Analysis; Vol 2(1): 131-50.

Gell, Alfred. 1997. “Exalting the King and Obstructing the State: A Political

Interpretation of Royal Ritual in Bastar District, Central India” The

Journal of the Royal Anthropological Institute; Vol. 3(3): 433-450.

375 Gellner, Ernest. Nations and Nationalism. Ithaca: Cornell University Press, 1983.

George, Alexander and Andrew Bennett. Case Studies and Theory Development

in the Social Sciences. Cambridge: MIT Press, 2005.

Ghosh, Parimal. Brave Men of the Hills: Resistance and Rebellion in Burma,

1825-1932. Honolulu: University of Hawaii Press, 2000.

Ghosh, S.K. Indian Democracy Derailed: Politics and Politicians. New Delhi:

APH Publishing Corporation, 1997.

Ghurye, G.S. “Excerpt from Caste and British Rule”, in Banerjee-Dube, Ishita

[ed]. Caste in History. Oxford: Oxford University Press, 2008.

Gough, Kathleen. 1974. “Indian Peasant Uprisings” Economic and Political

Weekly; Vol. 9(32): 1391-1393-1412.

Gold, Anna Grodzins, and Bhoju Ram Gujar. 1997. “Wild Pigs and Kings:

Remembering Landscapes in Rajasthan” American Anthropologist; Vol.

99(1): 70-84.

Gold, Ann Grodzins, and Bhoju Ram Gujar. In The Time of Trees and Sorrows:

Nature, Power, and Memory in Rajasthan. Durham: Duke University

Press, 2002.

Guha, Ramachandra. 1983. “Forestry in British and Post-British India” Economic

and Political Weekly; Vol. 18(45/46): 1940-1947.

. 1996. “Savaging the Civilised: Verrier Elwin and the Tribal Question in

Late Colonial India” Economic and Political Weekly; Vol. 31(35): 2375-

2380.

376 . 2001. “The Prehistory of Community Forestry in India.” Environmental

History; Vol. 6(2): 213-238.

Guha, Ramachandra, and Madhav Gadgil. 1989. “State Forestry and Social

Conflict in British India.” Past & Present; Vol. 123(1): 141-177.

Guha, Ranajit. A Rule of Property for Bengal: An Essay on the Idea of the

Permanent Settlement. Durham: Duke University Press, 1981.

. Elementary Aspects of Peasant Insurgency in Colonial India. Durham:

Duke University Press, 1999.

Gupta, Dipak K. 2007. “The Naxalites and the Maoist Movement in India: Birth,

Demise, and Reincarnation” Democracy and Security; No. 3: 157-188.

Haeuber, Richard. 1993. “Indian Forestry Policy in Two Eras: Continuity or

Change?” Environmental History Review; Vol. 17(1): 49-76.

Hale, Henry E. 2004. “Explaining Ethnicity” Comparative Political Studies; Vol.

37(4): 458-485.

. 2008. The Foundations of Ethnic Politics: Separatism of States and

Nations in Eurasia and the World. Cambridge: Cambridge University

Press, 2008.

Hardgrave Jr., Robert L. 1977. “The Mappilla Rebellion, 1921: Peasant Revolt in

Malabar” Modern Asian Studies; Vol. 11(1): 57-99.

Hardy, Peter. The Muslims of British India. Cambridge: Cambridge University

Press, 1972.

Hasan, Mushirul. Legacy of a Divided Nation: India's Muslims Since

377 Independence. London: Hurst & Company, 1997.

Hasnain, Nadeem. Tribal India. Lucknow: New Royal Book Co., 2007.

Haynes, Edward S. 1990. “Rajput Ceremonial Interactions as a Mirror of a Dying

Indian State System” Modern Asian Studies; vol. 24(3): 459-492.

Heller, Patrick. The Labor of Development: Workers and the Transformation of

Capitalism in Kerala, India. Ithaca: Cornell University Press, 2000.

Herbst, Jeffrey. States and Power in Africa: Comparative Lessons in Authority

and Control. Princeton: Princeton University Press, 2000.

Herring, Ronald J. 1980. “Abolition of Landlordism in Kerala: A Redistribution

of Privilege” Economic and Political Weekly; Vol. 15(26): A59-A69.

. “From ‘Fanaticism’ to Power: Deep Roots of Kerala’s Agrarian

Exceptionalism”, in W. Pinch [ed]. Speaking of Peasants: Essays on

Indian History and Politics in Honor of Walter Hauser. New Delhi:

Manohar, 2008.

. Land to the Tiller: The Political Economy of Agrarian Reform in South

Asia. New Haven: Yale University Press, 1983.

Hilbe, Joseph M. Negative Binomial Regression. Cambridge: Cambridge

University Press, 2007.

Hirschman, Charles. 1986. “The Making of Race in Colonial Malaya: Political

Economy and Racial Ideology,” Sociological Forum; Vol. 1(2) 330-361.

. 1987. “The Meaning and Measurement of Ethnicity in Malaysia: An

Analysis of Census Classifications” The Journal of Asian Studies. Vol.

378 46(3): 555-582.

Holst, Frederik. Ethnicization and Identity Construction in Malaysia. New York:

Routledge, 2012.

Horowitz, Donald L. Ethnic Groups in Conflict. Berkeley: University of California

Press, 1985.

. The Deadly Ethnic Riot. Berkeley: University of California Press, 2001.

Huntington, Samuel P. 1993. “The Clash of Civilizations?” Foreign Affairs; Vol.

72(3): 22-49.

Hurd II, John. 1975. “The Economic Consequences of Indirect Rule in India”

Indian Economic & Social History Review; Vol. 12(2): 169-181.

Hutchins, Francis. The Illusion of Permanence: British Imperialism in India.

Princeton: Princeton University Press, 1967.

Isaac, T., and P.K. Tharakan. 1986. “An enquiry into the historical roots of

industrial backwardness of Kerala: a study of Travancore Region” Centre

for Development Studies (Trivandrum), Working Paper.

Iyer, Lakshmi. 2009. “The Bloody Millennium: Internal Conflict in South Asia:

Internal Conflict in South Asia” Harvard Business School, Working Paper.

. 2010. “Direct versus Indirect Colonial Rule in India.” Review of

Economics and Statistics; Vol. 92(4): 693-713.

Jaffrelot, Christophe. 2000. “The Rise of the Other Backward Classes in the

Hindi Belt” The Journal of Asian Studies; Vol. 59(1): 86-108.

. Hindu Nationalism: A Reader. Princeton: Princeton University Press,

379 2007.

Jeffrey, Robin. 1978. “Matriliny, Marxism, and the Birth of the Communist Party

in Kerala” Journal of Asian Studies; Vol. 38(1): 77-98.

. Decline of Nayar Dominance: Society and Politics in Travancore, 1847-

1908. New York: Holmes & Meier Publishers, 1976.

Jensenius, Francesca Refsum. 2012. “Representation, Power and Development

for SCs in India 1974-2008”, University of California, Berkeley, Working

Paper.

Jha, Saumitra. 2008. “Trade, institutions and religious tolerance: evidence from

India” Harvard University, Working Paper.

Jinnah, Mahomed Ali. The Nation's Voice, Towards Consolidation: United We

Win: Annotated Speeches and Statements April 1940 – April 1942.

Karachi: Quaid-i-Azam Academy, 1996.

Johnson, Carter, and Diana Dumitru. 2011. “Constructing Interethnic Conflict

and Cooperation: Why Some People Harmed Jews and Others Helped

Them During the Holocaust in Romania” World Politics; Vol. 63(1): 1-42.

Kalyvas, Stathis N. 2003. “The Ontology of ‘Political Violence’: Action and

Identity in Civil Wars” Perspectives on Politics; Vol. 1(3): 475-94.

. 2008. “Ethnic Defection in Civil War” Comparative Political Studies;

Vol. 41(8): 1043-1068.

Karat, Prakash. 1973. “Agrarian Relations in Malabar – 1925-1948. Part Two:

Differentiation Amongst the Peasantry” Social Scientist; Vol. 2(3): 30-43.

380 Kareem, C.K. 1977. “The Agrarian Relations of Kerala” Journal of Kerala Studies;

Vol. 4(4).

Kaufman, Stuart J. Modern Hatreds. Ithaca: Cornell University Press, 2001.

Kaufmann, Chaim. 1996. “Possible and Impossible Solutions to Ethnic Wars”

International Security, Vol. 20(4) 136-75.

Kaura, Uma. Muslims and Indian Nationalism: The Emergence of the Demand

for India's Partition, 1928-40. New Delhi: Manohar Press, 1977.

Kawashima, Kōji. Missionaries and a Hindu State: Travancore, 1858-1936.

Oxford: Oxford University Press, 1998.

Keltie, J. Scott [ed]. The Statesman's Year-Book: Statistical and Historical

Annual of the States of the World for the Year 1903. London: Macmillan

and Co., 1903.

Kesharlal, Ajmera Jain, and Jawaharlal Jain. The Jaipur Album. Jaipur:

Rajasthan Directories Pub. House, 1935.

Kooiman, Dick. 1988. “State Formation in Travancore” Journal of Kerala

Studies; Vol. 15(1-4).

. 1993. “Political Rivalry among Religious Communities: A Case-Study of

Communal Reservations in India” Economic and Political Weekly; Vol.

28(7): 287-294.

. 1995. “Communalism and Indian Princely States” Economic and Political

Weekly; Vol. 30(34): 2123-2133.

. 1998. “The Nizam’s Last Victory” Economic and Political Weekly; Vol.

381 33(12): 645-660.

. Communalism and Indian Princely States: Travancore, Baroda, and

Hyderabad in the 1930s. New Delhi: Manohar Press, 2002.

Kooiman, Dick, and Adrian Koster, Peter Smets, and Bernhard Venema [eds].

Conflict in a Globalising World: Studies in Honour of Peter Kloos. Assen:

Royal Van Gorcum, BV, 2002.

Koya, S.M. Mohamed. 1975. “Commercial Interests in the Making of British

Policy in Malabar” Journal of Kerala Studies; Vol. 2(4).

Krishna, Gopal. 1985. “Communal Violence in India: A Study of Communal

Disturbance in Delhi” Economic and Political Weekly; Vol. 20(2): 61-74.

Kulkarni, Sharad. 1983. “Towards a Social Forest Policy” Economic and Political

Weekly; Vol. 18(6): 191-196.

Kumar. A. Indian Agriculture: Issues and Prospects. New Delhi: Sarup & Sons,

2001.

Kumar, Kapil [ed]. Congress and Classes: Nationalism, Workers and Peasants.

New Delhi: Manohar Publications, 1988.

Kurien, Prema. 1994. “Colonialism and Ethnogenesis: A Study of Kerala, India”

Theory and Society; Vol. 23(3): 385-417.

Kurup, K.N.N. 1976. “British Colonial Policy in Malabar” Journal of Kerala

Studies; Vol. 3(3/4).

Kusuman, K.K. Issues in Kerala Historiography. Thiruvananthapuram: Kerala

University Co-operative Press, 2003.

382 Lacina, Bethany. Forced ruling: Violence, language, and federalism in the

governing of India, 1950–2009. Book Project, Rochester University, 2012.

Laitin, David D. 1999. “Language Conflict and Violence: Or The Straw That

Strengthened The Camel's Back” Stanford University, Working Paper.

. 2003. “The Perestroikan Challenge to Social Science” Politics and

Society; Vol. 31(1): 163-84.

. Hegemony and Culture: Politics and Religious Change Among the

Yoruba. Chicago: University of Chicago Press, 1986.

Lange, Matthew, James Mahoney, and Matthias vom Hau. 2006. “Colonialism

and Development: A Comparative Analysis of Spanish and British

Colonies.” American Journal of Sociology; Vol. 111(5): 1412-1462.

Lange, Matthew. Lineages of Despotism and Development: British Colonialism

and State Power. Chicago: University of Chicago Press, 2009.

Lieten, George Kristoffel. “The Economic Structure of Kerala in the Mid-Fifties”

Journal of Kerala Studies; Vol. 2(4).

Lee, Alexander, and Kenneth Schultz. 2012. “Comparing British and French

Colonial Legacies: A Discontinuity Analysis of Cameroon” Quarterly

Journal of Political Science; Vol. 7: 1–46.

Lee-Warner, William. The Native States of India. London: Macmillan and Co.,

1910.

Leonard, Karen. 2003. “Reassessing Indirect Rule in Hyderabad: Rule, Ruler, or

Sons-in-Law of the State?” Modern Asian Studies; Vol. 37(2): 363-379.

383 Lieberman, Evan. 2005. “Nested Analysis as a Mixed-Method Strategy for

Comparative Research.” American Political Science Review 99, 3: 435-52.

Lipset, Seymour M., and Stein Rokkan. Party Systems and Voter Alignments:

Cross-National Perspectives. New York: The Free Press, 1967.

Logan, William. Malabar Manual. New Delhi: Asian Educational Services,

1887/1989.

Long, J. Scott, and Freese, Jeremy. Regression Models for Categorical

Dependent Variables Using Stata. College Station: Stata Press

Publication, 2009.

Ludden, David. An Agrarian History of South Asia. Cambridge: Cambridge

University Press, 1999.

Lustick, Ian S. “History, Historiography, and Political Science: Multiple

Historical Records and the Problem of Selection Bias” American Political

Science Review; Vol. 90(3): 605-18.

Madhok, Balraj. Indianisation?: What, Why, and How. New Delhi: S. Chand,

1970.

Mahmood, Cynthia Keppley. 1993. “Rethinking Indian Communalism” Asian

Survey; Vol. 33(7): 722-737.

Mahoney, James. 2000. “Path Dependence in Historical Sociology” Theory and

Society; Vol. 29(4): 507-548.

Mahoney, James and Dietrich Rueschemeyer. Comparative Historical Analysis

in the Social Sciences. Cambridge: Cambridge University Press, 2003.

384 Majumdar, R.C. [ed]. The History and Culture of the Indian People. Bombay:

Bharatiya Vidya Bhavan, 1951.

Malleson, G.B. An Historical Sketch of the Native States of India. London:

Longmans, Green & Co., 1875.

Mamdani, Mahmood. Citizen and Subject: Contemporary Africa and the Legacy

of Late Colonialism. Princeton: Princeton University Press, 1996.

Mann, Rann Singh. Tribes of India: Ongoing Challenges. New Delhi: MD

Publications, 1996.

Mathur, L.P. Tribal Revolts in India Under British 'Raj'. New Delhi: Aavishkar

Publishers: 2004.

Marx, Karl. The 18th Brumaire of Louis Bonaparte. Rockville: Wildside Press,

1852/2008.

Mayaram, Shail. 1993. “Communal Violence in Jaipur” Economic and Political

Weekly; Vol. 28(46): 2524-2541.

. Resisting Regimes: Myth, Memory, and the Shaping of a Muslim

Identity. Oxford: Oxford University Press, 1997.

. “The Shudra Right to Rule: Caste, Clan, Community and State”, in

Banerjee-Dube, Ishita [ed]. Caste in History. Oxford: Oxford University

Press, 2008.

Mendelsohn, Oliver, and Marika Vicziany. The Untouchables: Subordination,

Poverty, and the State in Modern India. Cambridge: Cambridge

University Press, 1998.

385 Menon, Dilip. Caste, Nationalism and Communism in South India: Malabar

1900-1948. Cambridge: Cambridge University Press, 2007.

Menon, Shungoony P. History of Travancore. Madras: Higginbottom & Co.,

1878.

Menon, A. Sreedhara. Triumph and Tragedy in Travancore: Annals of Sir CP's

Sixteen Years. Kottayam: Current Books, 2001.

Menon, V.P. The Story of the Integration of the Indian States. New York:

Macmillan Publishers, 1956.

Metcalf, Barbara, and Thomas Metcalf. A Concise History of Modern India.

Cambridge: Cambridge University Press, 2001.

Metcalf, Thomas. Imperial Connections: India in the Indian Ocean Arena, 1860-

1920. Berkeley: University of California Press, 2007.

. The Aftermath of Revolt: India, 1857-1870. Princeton: Princeton

University Press, 1965.

Miklian, Jason, and Scott Carney. 2010. “Fire in the Hole: How India's Economic

Rise Turned an Obscure Communist Revolt Into a Raging Resource War”

Foreign Affairs; Sep/Oct.

Mishra, Narayan. Scheduled Castes Education: Issues and Aspects. New Delhi:

Kalpaz Publications, 2001.

Mohammed, U. Educational Empowerment of Kerala Muslims: A Socio-

Historical Perspective. Calicut: Other Books, 2007.

Mohanty, P.K. Encyclopaedia of Scheduled Tribes in India. New Delhi: Gyan

386 Publishing House, 2006.

Moore, Barrington. Social Origins of Dictatorship and Democracy: Lord and

Peasant in the Making of the Modern World. Boston: Beacon Press, 1966.

Meyer, William Stevenson, and Richard Burn, and James Cotton, and Herbert

Risley. The Imperial Gazetteer of India. Oxford: Clarendon Press, 1909.

Mishra, Narayan. Scheduled Castes Education: Issues and Aspects. New Delhi:

Kalpaz Publications, 2001.

More, Leena. English East India Company and the Local Rulers in Kerala: A

Case Study of Attingal and Travancore. Pondicherry: Mission Press,

2003.

Mueller, John. “The Banality of Ethnic War” International Security; Vol. 25(1):

43-71.

Mullaney, Thomas. Coming to Terms with the Nation: Ethnic Classification in

Modern China. Berkeley: University of California Press, 2010.

Mukherjee, Rudrangshu. 1990. “’Satan Let Loose Upon Earth’: The Kanpur

Massacres in India in the Revolt of 1857” Past & Present Vol. 128

Naipaul, V.S. 1971. "The Election in Ajmer" Sunday Times Magazine; 8: 17.

Nair, Shanti. Islam in Malaysian Foreign Policy. New York: Routledge, 1997.

Navlakha, Gautam. 2010. “Days and Nights in the Maoist Heartland.” Economic

and Political Weekly; Vol. 45(16): 38-47.

North, Douglass. Institutions, Institutional Change and Economic Performance.

Cambridge: Cambridge University Press, 1990.

387 Nossiter, T.J. Communism in Kerala: A Study in Political Adaptation. Berkeley:

University of California Press, 1982.

O.J. 1983. “Political Backdrop to Trivandrum Riots” Economic and Political

Weekly; Vol. 18(7): 209-211.

Ollapally, Deepa M. The Politics of Extremism in South Asia. Cambridge:

Cambridge University Press, 2008.

Omvedt, Gail. Dalit Visions. Hyderabad: Orient Longman Private Ltd., 2006.

Oommen, John. 1995. “Politics of Communalism in Kerala” Economic and

Political Weekly; Vol. 30(11): 544-547.

Page, Scott E. 2006. “Path Dependence” Quarterly Journal of Political Science;

No. 1: 87-115.

Pandey, Gyanendra. The Construction of Communalism in Colonial North India.

Oxford: Oxford University Press, 1990.

Panikkar, K.N. 1984. “Peasant Exploitation in Malabar in the Nineteenth

Century” Journal of Kerala Studies; Vol. 11(1-4).

. Against Lord and State: Religion and Peasant Uprisings in Malabar,

1836-1921. Oxford: Oxford University Press, 1989.

Paswan, Sanjay. Encyclopaedia of Dalits in India: Constitution. Delhi: Kalpaz

Publications, 2002.

Pavier, Barry. 1974. “The Telengana Armed Struggle” Economic and Political

Weekly; Vol. 9(32/34): 1413-1420.

Pernau, Margrit. The Passing of Patrimonialism: Politics and Political Culture in

388 Hyderabad, 1911-1948. New Delhi: Manohar Publishers, 2000.

Petersen, Roger. Understanding Ethnic Violence. New York: Cambridge

University Press, 2002.

. “Identity, Rationality, and Emotion in State Disintegration and

Reconstruction”, in Chandra, Kanchan [ed]. Constructivist Theories of

Ethnic Politics. Cambridge: Cambridge University Press, forthcoming.

Pierson, Paul. 2000. “Increasing Returns, Path Dependence, and the Study of

Politics” American Political Science Review; Vol. 94(2): 251-68.

. Politics in Time: History, Institutions, and Social Analysis. Princeton:

Princeton University Press, 2004.

Pillai, T.K. Velu. The Travancore State Manual, Volume I. Cochin: Kerala Books

and Publication Society, 1940.

Ponnumuthan, Selvister. The Spirituality of Basic Ecclesial Communities in the

Socio-Religious Context of Trivandrum/Kerala, India. Rome: Gregorian

University Press, 1996.

Posen, Barry. 1993. “The Security Dilemma and Ethnic Conflict”, in Michael E.

Brown [ed] Ethnic Conflict and International Security. Princeton:

Princeton University Press, 1993: 103-24.

Posner, Daniel. 2003. “The Colonial Origins of Ethnic Cleavages: The Case of

Linguistic Divisions in Zambia” Comparative Politics; Vol. 35(2): 127-146.

. Institutions and Ethnic Politics in Africa. Cambridge: Cambridge

University Press, 2005.

389 Prakash, B.A. 1988. “Agricultural Backwardness of Malabar during the Colonial

Period: An Analysis of Economic Causes” Social Scientist; Vol. 16(6/7): 51-

76.

Price, Pamela. “Democracy and Ethnic Conflict in India: Precolonial Legacies in

Tamil Nadu” Asian Survey; Vol. 33(5): 493-506.

Putnam, Robert. Making Democracy Work: Civic Traditions in Modern Italy.

Princeton: Princeton University Press, 1993.

Queen, Christopher S., and Sallie B. King [eds]. Engaged Buddhism: Buddhist

Liberation Movements in India. Albany: State University of New York,

1996.

Radhakrishna, Meena. Dishonoured by History: Criminal Tribes and British

Colonial Policy. New Delhi: Orient Longman Ltd., 2001.

Radhakrishnan, P. Peasant Struggles, Land Reforms and Social Change: Malabar

1836-1982. New Delhi: Sage Publications, 1989.

Rai, Mridu. Hindu Rulers, Muslim Subjects: Islam, Rights, and the History of

Kashmir. Princeton: Princeton University Press, 2004.

Ram, Mohan. 1973. “The Telengana Peasant Struggle” Economic and Political

Weekly; Vol. 8(23): 1025-1032.

Ramusack, Barbara. The Indian Princes and Their States. Cambridge: Cambridge

University Press, 2004.

Ravindran, T.K. Institutions and Movements in Kerala History. Trivandrum:

Charithram Publications, 1978.

390 Reddy, G. Ram and B.A.V. Sharma. Regionalism in India: A Study of Telangana.

New Delhi: Concept Publishing Company, 1979.

Reddy, N. Subba. 1988. “Depriving Tribals of Land: Andhra Move to Amend

Land Transfer Laws” Economic and Political Weekly; Vol. 23(29): 1458-

1461.

Regani, Sarojini. Nizam-British Relations 1724 – 1857. New Delhi: Concept

Publishing Company, 1963.

Richter, William. 1971. “Princes in Indian Politics” Economic and Political

Weekly; Vol. 6(9): 535-542.

Robinson, Francis. 1977. “Nation Formation: The Brass Thesis and Muslim

Separatism” The Journal of Commonwealth & Comparative Politics; Vol.

15(3): 215-230.

. Separatism among Indian Muslims: The politics of the United

Provinces' Muslims 1860-1923. Cambridge: Cambridge University Press,

1974.

Roosa, John. 2001. “Passive revolution meets peasant revolution: Indian

nationalism and the Telangana revolt” Journal of Peasant Studies; Vol.

28(4): 57-94.

Rosin, Thomas R. 1978. “Peasant Adaptation as Process in Land Reform” (pp.

460-495) in Vatuk, Sylvia [ed]. American Studies in the Anthropology of

India. New Delhi: Manohar Press, 1978.

. 2010. “Touring Ajmer-Marwara: The Tripartite Drainage at the Ajmer

391 Saddle” Working Paper.

Rubin, Barnett. Feudal Revolt and State-Building: The 1938 Sikar Agitation in

Jaipur. New Delhi: South Asian Publishers, 1983.

Rudolph, Lloyd I., and Susanne Hoeber Rudolph. 1960. “The Political Role of

India’s Caste Associations” Pacific Affairs; Vol. 33(1): 5-22.

. 1966. “Rajputana under British Paramountcy: The Failure of Indirect

Rule” The Journal of Modern History; Vol. 38(2): 138-160.

. In Pursuit of Lakshmi: The Political Economy of the Indian State.

Chicago: University of Chicago Press, 1987.

Rudolph, Susanne Hoeber, and Lloyd I. Rudolph. Essays on Rajputana. New

Delhi: Concept Publishing Company, 1984.

Sarma, E.A.S. 2006. “The Adivasi, the State and the Naxalite: Case of Andhra

Pradesh” Economic and Political Weekly; Vol. 41(15): 1434-1437.

Saumarez Smith, Richard. 2008 “Excerpt from From Village to Community”, in

Banerjee-Dube, Ishita [ed]. Caste in History. Oxford: Oxford University

Press, 2008.

Scott, James. Seeing Like a State: How Certain Schemes to Improve the Human

Condition Have Failed. New Haven: Yale University Press, 1998.

. The Art of Not Being Governed: An Anarchist History of Upland

Southeast Asia. New Haven: Yale University Press, 2009.

Scuto, Giuseppe. 2008. “Caste Violence in Contemporary India” Delhi University,

Working Paper.

392 Sekhon, Jasjeet S., and Rocío Titiunik. 2012. “When Natural Experiments Are

Neither Natural nor Experiments” American Political Science Review;

Vol. 106(1): 35-57.

Shani, Ornit. Communalism, Caste and Hindu Nationalism: The Violence in

Gujarat. Cambridge: Cambridge University Press, 2007.

Sharma, C.L. Ruling Elites of Rajasthan. New Delhi: M.D. Publications Pvt. Ltd.,

1993.

Shoup, Brian. Conflict and Cooperation in Multi-Ethnic States: Institutional

Incentives, Myths, and Counter-Balancing. New York: Routledge, 2008.

Simhadra, V.C. Ex-Criminal Tribes of India. New Delhi: National Publishing

House, 1979.

Sikand, Yoginder and Manjari Katju. 1994. “Mass Conversions to Hinduism

among Indian Muslims” Economic and Political Weekly; Vol. 29(34):

2214-2219.

Singh, Godwin R. In Search of Communal Harmony. New Delhi: I.S.P.C.K.,

1985.

Singh, Hira. 2003. “Princely states, peasant protests, and nation building in

India: the colonial mode of historiography and subaltern studies” Social

Movement Studies; Vol. 2(2): 213-228.

Singh, Prerna. 2010. “We-ness and Welfare : A Longitudinal Analysis of Social

Development in Kerala, India” World Development; Vol. 39(22): 282-293.

SinghaRoy, Debal K. 2005. “Peasant Movements in Contemporary India:

393 Emerging Forms of Domination and Resistance” Economic and Political

Weekly; Vol. 40(52): 5505-5513.

Sisson, Richard. 1969. “Peasant Movements and Political Mobilization: The Jats

of Rajasthan.” Asian Survey; Vol. 9(12): 946-963.

. The Congress Party in Rajasthan: Political Integration and Institution-

Building in an Indian State. Berkeley: University of California Press, 1971.

Somarajan, C.N. 1983. “A Survey of Land Reforms in Kerala Up To 1956” Journal

of Kerala Studies; Vol. 10(1-4).

Smith, Donald Eugene. India as a Secular State. Princeton: Princeton University

Press, 1963.

Smith, Martin. Ethnic Groups in Burma: Development, Democracy and Human

Rights. London: Anti-Slavery International, 1994.

Sriskandarajah, Dhananjayan. “Development, Inequality and Ethnic

Accommodation: Clues from Malaysia, Mauritius and Trinidad and

Tobago” Oxford Development Studies; Vol. 33(1): 63-79.

Steinmo, Sven, and Kathleen Thelen, and Frank Longstreth. Structuring Politics:

Historical Institutionalism in Comparative Analysis. Cambridge:

Cambridge University Press, 1992.

Stern, Robert W. The Cat and the Lion: Jaipur State in the British Raj. Leiden:

Brill Publishers, 1988.

Stokes, Eric. The Peasant and the Raj: Studies in Agrarian Society and Peasant

Rebellion in Colonial India. Cambridge: Cambridge University Press,

394 1978.

Streeck, Wolfgang, and Kathleen Thelen [eds]. Beyond Continuity: Institutional

Change in Advanced Political Economies.

Streets, Heather. 2001. “The Rebellion of 1857: Origins, Consequences, Themes”

Teaching South Asia: An Internet Journal of Pedagogy; Vol 1(1).

Sundar, Nandini. 2001. “Debating Dussehra and Reinterpreting Rebellion in

Bastar District.” Journal of the Royal Anthropological Institute; Vol. 7(1):

19-35.

. Subalterns and Sovereigns: An Anthropological History of Bastar,

1854-2006. Delhi: Oxford University Press, 1997.

Sundarayya, P. 1973. “Telangana People’s Armed Struggle, 1946-1951. Part One:

Historical Setting” Social Scientist; Vol. 1(7): 3-19.

Sunderland, Jabez T. India in Bondage: Her Right to Freedom. Calcutta: Prasabi

Press, 1924.

Tambiah, Stanley J. 1990. “Presidential Address: Reflections on Communal

Violence in South Asia” The Journal of Asian Studies; Vol. 49(4): 741-760.

Tandor, Rajeshwari (ed). A Case for Conservation of Tribal Heritage &

Environment of Tribal Areas. New Delhi: INTACH Press, 2005.

Taylor, Robert H. The State in Myanmar. Singapore: National University of

Singapore Press, 2009.

Thelen, Kathleen. How Institutions Evolve: The Political Economy of Skills in

Germany, Britain, the United States, and Japan. Cambridge: Cambridge

395 University Press, 2004.

Thirumali, I. 1992. “Dora and Gadi: Manifestation of Landlord Domination in

Telangana” Economic and Political Weekly; Vol. 27(9): 477-482.

Throup, D.W. 1985. “The Origins of Mau Mau” African Affairs; Vol. 84(336):

399-433.

Tod, James. Annals and Antiquities of Rajasthan or the Central and Western

Rajpoot States of India. London: Oxford University Press, 1829/1920.

Urdal, Henrik. 2008. “Population, Resources, and Political Violence: A

Subnational Study of India, 1956-2002” Journal of Conflict Resolution;

Vol. 52(4): 590-617.

Van Evera, Stephen. 2001. “Primordialism Lives!” APSA-CP; Vol. 12(1): 20-2.

Varma, S.C. The Bhil Kills. New Delhi: Kunj Publishing House, 1978.

Varghese, T.C. Agrarian Change and Economic Consequences. Calcutta: Allied

Publishers, 1970.

Varshney, Ashutosh. 2001. “Ethnic Conflict and Civil Society: India and Beyond”

World Politics; Vol. 53(3): 362-398.

. Ethnic Conflict and Civic Life: Hindus and Muslims in India. Yale: Yale

University Press, 2002.

Veer, Peter van der. Religious Nationalism: Hindus and Muslims in India.

Berkeley: University of California Press, 1994. von Fürer-Haimendorf, Christoph. Tribal Populations and cultures of the Indian

subcontinent. Leiden: Brill Publishers, 1985.

396 . Tribes of India: The Struggle for Survival. Berkeley: University of

California Press, 1982.

Wal, S. [ed]. Child Labour in Various Industries. New Delhi: Sarup & Sons,

2006.

Washbrook, D.A. The Emergence of Provincial Politics: The Madras Presidency,

1870-1920. Cambridge: Cambridge University Press, 1976.

Weiner, Myron. Sons of the Soil: Migration and Ethnic Conflict in India.

Princeton: Princeton University Press, 1978.

Wilkinson, Steven I. 2008. “Which group identities lead to most violence?

Evidence from India”, in Kalyvas, Shapiro, and Masoud [eds]. Order,

Conflict, and Violence. Cambridge: Cambridge University Press, 2008.

. Votes and Violence: Electoral Competition And Ethnic Riots in India.

Cambridge: Cambridge University Press, 2004.

Wood, Conrad. 1974. “Historical Background of the Moplah Rebellion:

Outbreaks, 1836-1919” Social Scientist; Vol. 3(1): 5-33.

. 1976. “First Moplah Rebellion against British Rule in Malabar” Modern

Asian Studies; Vol. 10(4): 543-556.

. “Peasant Revolt: An Interpretation of Moplah Violence in the Nineteenth

and Twentieth Centuries”, in Dewey, Clive, and A.G. Hopkins [eds]. The

Imperial Impact: Studies in the Economic History of Africa and India.

London: Athlone Press, 1978.

. The Moplah Rebellion and its Genesis. Delhi: South Asia Books, 1987.

397 Wood, John R. 1984. “British versus Princely Legacies and the Political

Integration of Gujarat” The Journal of Asian Studies; Vol. 44(1): 65-99.

Wood, Peter. “Vassal State in the Shadow of Empire: Palmer's Hyderabad, 1799-

1867”, Ph.D. Dissertation, University of Wisconsin, 1981.

Young, Crawford. The African State in Colonial Perspective. New Haven: Yale

University Press, 1994.

Zagoria, Donald S. 1971. “The Ecology of Peasant ”

American Political Science Review; Vol. 65(1): 144-160.

Zakaria, Rafiq. Rise of Muslims in Indian Politics: An Analysis of Development

From 1885 to 1906. Bombay: Samaiya Publications Pvt. Ltd., 1970.

[Other Sources]

5th Report from the Select Committee on the Affairs of the East India Company

Appendix No. 13 Political Survey of the Northern Circars – James Grant,

Esq. to Warren Hastings, Esq. 1786

Crime in India (2010). National Crime Records Bureau, Ministry of Home

Affairs.

Government of India White Paper on Hyderabad, 1948.

(IDS) Indian Department of Statistics, Commercial Intelligence Department.

Statistics of British India: Part 3. New Delhi: Superintendent Government

Printing, 1908.

398 India and the Durbar: A Reprint of the Indian Articles in the 'Empire Day'

Edition of the Times, May 24th, 1911. London: Macmillan and Co., 1911.

Ministry of Home Affairs, Government of India. Report of Working Group of

National Integration Council to Study Reports of the Commission of

Inquiry on Communal Riots, 2007.

Sachar Committee Report, 2006.

The Rajputana Gazetteer: Volume II. Calcutta: Office of the Superintendent of

Government Printing, 1879.

[Map Sources]

Map 1: http://dsal.uchicago.edu/reference/gaz_atlas_1909/pager.html?

object=26

Map 2: http://dsal.uchicago.edu/reference/gaz_atlas_1909/pager.html?

object=40

Map 3: http://www.rajasthantravel.org.uk/gifs/map-of-rajasthan.jpg

Map 4: http://dsal.uchicago.edu/reference/gaz_atlas_1909/pager.html?

object=47

Map 5: http://www.mapsofindia.com/maps/kerala/kerala.gif

Map 6: http://en.wikipedia.org/wiki/File:Central_Provinces_India_1903.gif

Map 7:

http://upload.wikimedia.org/wikipedia/commons/9/90/India_Red_Corr

399 idor_map.png

Map 8: http://www.jaunted.com/files/16133/Burma_Map_385.jpg

Map 9: Hirschman 1986: 335

400 Data Appendix

Archival Research:

While conducting research for this project I visited 6 archives and several libraries in India, the UK, and the US. The archives were: National Archives of

India, British Library, Jaipur City Palace Archives, Deshbandhu Press Library,

Andhra Pradesh State Archives, and Kerala State Archives. Most of these archives had some kind of an electronic database, and I searched for any number of keywords that dealt with my research. I have detailed the precise records for archival documents which were utilized in this dissertation, and if these numbers are presented to the archivists in these various institutions then the same material can easily be re-accessed.

Elite Interviews:

During the course of this project I conducted elite interviews with roughly

75 respondents spanning my six case studies. Note that not all of the interviews made it into the final dissertation draft.

Interviewees were chosen in the following manner: first, I wanted to ensure a large cross-section of respondents. For example, I wanted to get a government perspective, but also the perspective of non-governmental organizations. The respondents generally fit into the following categories: local government officials, NGO officials, police officers, religious leaders, journalists, and academics. At the end of each interview session I also asked respondents for

401 the names of other individuals who might be able to assist me on a particular issue, and this method snowballed into a greater number of total interviews.

Interviews generally took between 30 minutes and an hour; sometimes longer interviews took almost two hours to complete. They were generally conducted in the respondents' offices. Several interviews were also conducted over the phone, either from within India or from the United States, and they are labelled as such. In a later section of this appendix I provide a basic questionnaire which I utilized for these interviews. Because respondents were often pressed for time, I did not always adhere strictly to the questionnaire. In fact, most of the time the interviews became open-ended and conversational. Other times I was unable to complete an interview and therefore not all questions were asked.

The interviews were – in every state except Kerala – conducted in either

Hindi (or what English speakers of Hindi might call 'Hinglish') or, less often,

English. All interviews in Kerala were conducted entirely in English. I did not utilize a voice or video recorder for interviews (and most interviewees would probably not have consented to being recorded). Instead, I took detailed longhand notes during the interview and then immediately typed them up after returning home. Sometimes quotes used in this dissertation, if not taken verbatim, were either very close, or based on impressions from the interview.

Because this project was considered somewhat controversial in India the identity of almost all the interviewees has remained confidential. All respondents were told at the outset of the interview that their names would not be used and

402 that any identifying information about them would not be utilized either. This was done to ensure their privacy and candor.

Aside from formal elite interviews, I also conducted a number of interviews with everyday people living in each of the six cases. For example, I went to local chai stalls and held informal chats with patrons. These interviews were used to provide background information, or simply collect general impressions from citizens of a given area.

The basic questionnaire I used as the prompt for interviews is attached below:

403 Basic Questionnaire

1. Can you tell me about your background? What do you do?

2. How long have you worked [with this organization]? What does it do?

3. What is the general state of relations between caste groups in this area? Can you tell me about the position of low castes here?

a. Geographic

b. Economic

c. Educational

d. Political

e. Historical

f. How do powerful castes compare to low castes? Is there a large disparity

between them? How do they interact?

4. Does this area experience a lot of caste violence?

a. Can you provide details of recent episodes?

b. What are its causes?

c. Who initiates the violence?

d. Does most of the violence occur around election time?

5. Are there civil society groups that band together low castes and high castes in this area? If so, what are their names?

6. What is the general state of relations between religious groups in this area?

7. Does this area experience a lot of religious violence?

8. Are there civil society groups that band together different religious groups in

404 this area? If so, what are their names?

9. How does this area compare to [other case study] on these issues?

405 Quantitative Analysis:

Chapter 2 displayed the results of a quantitative analysis of ethnic violence in 589 modern Indian districts. There are a number of items pertaining to the analysis which will be elucidated upon here. First, I detail how I coded the main independent variables for the study. The process is described below:

Coding Districts as British or Princely:

1. The first step is to visit the particular district’s website; this can be found at the database maintained by the Indian government, at: http://districts.gov.in/

2. The second step is to determine the particular district’s current geographical coordinates. This information should be available on the website under a

‘Geography’ (or similarly titled) section. For example (Kurukshetra district in

Haryana), the information would look like this: 29.52 to 30.12 latitude and 76.26 to 77.04 longitude.

3. The third step is to find the history of the district; this is often available on the district website under a ‘History’ or ‘About the District’ section. This section will usually include information about when the district came under the control of the

British or whether it remained a princely state. This is the main information which allowed me to code the district as a ‘province’ or a ‘princely state’.

a. The main measure of the independent variable was a dummy variable. I

code a district as 1 if it was a province and 0 if it was a princely state.

b. The second measure of the independent variable – the number of years

a district was under the control of the British – was also based on

406 information provided either by the district website, or historical

information (see #4 below). For example, a district like Ajmer in

Rajasthan would have a value of “129” because it came under the control of

the British in 1818 (1947-1818 = 129). All princely states were coded as

zero.

4. The fourth step is to match the modern district with its colonial-era district. I first visit the Imperial Gazetteer (1909) which was published by the British government as a collection of information about every district of the British

Indian Empire (i.e. it was a compilation of all the numerous District Gazetteers).

This source is available online through the University of Chicago at: http://dsal.uchicago.edu/reference/gazetteer/

5. The fifth step is to find the colonial-era district’s geographical coordinates. This way I can determine whether or not the modern district roughly correlates to the colonial-era district.

6. The sixth step is to double-check the history of the district. If the district website says, for example, that the British were ceded the territory in 1765 then the Imperial Gazetteer should also state the same.

Additional notes on this point:

a. Many of the current districts match up with a colonial-era district of the

same name. When this is not the case (if, for example, the current district

was a tehsil of an erstwhile colonial district) then I record information

for the previous district.

407 b. Determining when a district came under British control is also often

difficult; sometimes, for example, the British might have invaded and

annexed two-thirds of a current district. The other one-third might have

been annexed ten years later. In this case I try to record the earliest date at

which the British gained political control over a majority of the district.

Dependent Variables:

There were various dependent variables used in the statistical analyses.

Below I highlight some basic information about the datasets which I utilized.

WITS: These conflict data were collected by the National Counterterrorism

Center, and the program's homepage is located at: http://www.nctc.gov/site/other/wits.html. All specific details about the dataset are located there. All data utilized in my statistical analyses were restricted to the

Maoist conflict in India, for the years 2005-9. Data on conflict were ultimately taken from local, national, and international press reports. The reports on attacks were geocoded, enabling them to be mapped onto current Indian districts. The

Colonialism and Caste/Tribal Conflict Regression displayed on page 119 summarizes my findings.

IHDS: The Indian Human Development Survey was conducted by researchers at the University of Maryland, and the homepage is located at: http://ihds.umd.edu .

All specific details about the survey and its contents are available there. The

IHDS (household) data were collapsed using Stata into district level figures. This dataset was used in constructing two dependent variables (level of trust between

408 jatis, and percentage of full-time workers engaged in agricultural labor). The

Colonialism and Caste/Tribal Mechanisms Regression displayed on page 122 summarizes my findings.

Varshney-Wilkinson: The Varshney-Wilkinson dataset records Hindu-Muslim riots in India from 1950-1995. Information on the dataset is available at: http://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/4342. All specific details about the dataset are available there.

In Chapter 2 I used the dataset as the main dependent variable for my regression analysis on colonialism and contemporary religious conflict. I went through the Varshney-Wilkinson dataset and created pooled figures for riot casualties per-district over the 45 year period. So I compiled, for example, the total number of riot casualties in Hyderabad District from 1950-1995. The

Colonialism and Hindu-Muslim Conflict Regression displayed on page 125 summarizes my findings utilizing the Varshney-Wilkinson dataset.

Furthermore, in Chapters 3 and 4 I presented data on Hindu-Muslim riots in contemporary Rajasthan and Kerala. I utilized the Varshney-Wilkinson dataset as the baseline for these figures, although these numbers sometimes proved to be problematic (especially in the Kerala case).

First, in Rajasthan, the Varshney-Wilkinson dataset presented riot metrics which were, as I found, largely correct. The only other riot in Jaipur or Ajmer about which I could find information was a 1990 riot in the former district (see pg. 142), although I was unable to find figures on casualties.

409 The Varshney-Wilkinson figures presented on riots in Kerala, however, were deeply flawed. As I noted on page 215, the dataset recorded only 16 injuries from Hindu-Muslim riots in Trivandrum. This is an extremely unlikely result –

Trivandrum is, in fact, the deadliest city for communal violence in Kerala.

Furthermore, much of the rioting in Travancore is between Muslims and

Christians. Therefore, I did additional research into the pervasiveness of religious riots throughout Kerala in the post-independence period. I found the following additional information, and I display the data source as well:

Communal Riots in Travancore, 1950-1995 Year Location Casualties 1968 Vizhinjam 6 1982-3 Trivandrum 8 1983 Nilikkal 50 1992 Poonthura 5 1992 Trivandrum 39 1995 Trivandrum 38

Sources:

1. 1968 riot (Madhok 1970: 105)

2. 1982-3 riot (“Rioting Continues in Southern India”, New York Times,

January 1st, 1983)

3. 1983 riot (“Stone Cross Unearthed in India Ignites Christian-Hindu

Dispute”, Hinduism Today Magazine, October 1st, 1983)

4. 1992 riot (“UDF govt negligence led to '92 riots: panel”, DH News

Service, September 19th, 1996)

5. 1992 riot (“6 slain in Hindu-Muslim clashes over temple site in north

410 India”, The Toronto Star, July 21st, 1992)

6. 1995 riot (“Trial on Vizhinjam riots begins”, The Hindu, August 8th,

2008; and Ponnumuthan 1996: 98)

To ensure that I was not simply arbitrarily increasing the number of riots in

Travancore, I also researched the number of casualties from communal incidents in Kozhikode, the center of rioting in the north. The Varshney-Wilkinson dataset does not count an additional 8 injuries from a 1954 riot, which I then added to my calculations. Other than this, however, I was not able to find any additional information for riot metrics from Kozhikode or other northern cities. For example, I searched LexisNexis for “Trivandrum Riot” and was able to find information on the 1992 riot from an article from The Toronto Star. There was, on the other hand, nothing when I looked for “Calicut Riot” or “Kozhikode Riot”.

Finally, I am still not using any figures for the Trivandrum riot of 1986, as

I was not able to find any information on casualties. There were also probably riots in the coastal areas of Trivandrum in 1972 and 1981 (Ponnumuthan 1996:

98), but no information was available for these incidents either.

411