From Hierarchy to Ethnicity: The Politics of Caste in Twentieth-Century India 9781108489904, 9781108779678

Caste and ethnicity have been crucial in shaping the discourse around identity politics in modern South Asia. This book

948 64 4MB

English Pages 272 [286] Year 2020

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

From Hierarchy to Ethnicity: The Politics of Caste in Twentieth-Century India
 9781108489904, 9781108779678

Citation preview

From Hierarchy to Ethnicity Caste identities have been fundamental in shaping Indian politics, the loyalties, the structure of the party system, and the distribution of public resources. This book explains when and how caste identities became politicized, and how the nature of caste itself changed over time. The way in which caste is understood is dependent on time and place; some understand it as an integrated system based on subtle gradations of hierarchical difference, while others perceive it as a set of bounded groups similar to ethnicities in other countries across the globe. Using a wide range of historical data sources, the book shows how these identities evolved from the colonial period to the present. It describes how uneven economic development shifted the set of groups and regions where caste is important, while the spread of mass politics and the corresponding decline of patrimonial politics led to a decline in the hierarchical nature of the system. While colonial rule established some of the structural conditions of these changes, the enthusiasm with which caste activists responded to these conditions fluctuated from group to group. Even after these changes, the caste system in India differs from ethnic politics in other parts of the world, in ways that reflect its hierarchical history. The theory discussed in this book foregrounds how identity politics can vary not just between countries, but between groups, and how the modern conception of the ethnic group is the product of a long, and highly contingent, historical process. Alexander Lee teaches at the Department of Political Science, University of Rochester, New York. His research focuses on the politics of South Asia, and historical political economy. His most recent publication is Development in Multiple Dimensions: Social Power and Regional Policy in India (2019).

From Hierarchy to Ethnicity The Politics of Caste in Twentieth-Century India

Alexander Lee

University Printing House, Cambridge CB2 8BS, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, vic 3207, Australia 314 to 321, 3rd Floor, Plot No.3, Splendor Forum, Jasola District Centre, New Delhi 110025, India 79 Anson Road, #06–04/06, Singapore 079906 Cambridge University Press is part of the University of Cambridge. It furthers the University’s mission by disseminating knowledge in the pursuit of education, learning and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781108489904 c Alexander Mark Lee 2020 ! This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2020 Printed in India A catalogue record for this publication is available from the British Library ISBN 978-1-108-48990-4 Hardback Cambridge University Press has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication, and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

To Rebecca, Matt, and Nathan

Contents List of Tables List of Figures Acknowledgements 1. 2. 3. 4. 5. 6. 7.

ix xi xiii

Introduction Explaining Identity Activism Caste in Historical Context Caste in the Census of India The Causes of Ranked Rhetoric Caste since Independence Conclusion

1 29 56 83 111 132 169

Appendix A: Data Appendix B: Statisical Tests Appendix C: Additional Tables and Figures Bibliography Index

185 191 219 242 260

Tables 4.1 4.2 4.3 4.4 5.1 B.1 B.2 B.3 B.4 B.5 B.6 B.7 B.8 B.9 C.1 C.2 C.3 C.4 C.5 C.6 C.7

Census caste classification schemes by year, 1891–1931 Aggregate number of petitions by year Number of petitions by literacy (pooled) Number of petitions by caste status Disaggregated types of caste petition by year Main results: logistic regression with petition as dependent variable Alternative measures of participation Province level robustness checks: logistic mixed effects regression with petition as dependent variable Caste level robustness checks: logistic mixed effects regression with petition as dependent variable Traditional institutions and social ranking: sequential logistic regression Number of petitions by caste status Institutions and social ranking: simple models Robustness checks: sequential logistic regression Robustness checks: sequential logistic regression Census petitions in India, 1901–1931 Summary statistics Ranked petitioning by province, 1901–1931 Additional alternative models: petitioning Additional alternative models: segmentary petitioning Traditional institutions and social ranking: sequential logistic regression Traditional institutions and social ranking: sequential logistic regression

86 87 94 100 114 195 197 199 201 206 210 212 214 216 219 227 229 230 231 232 234

x

List of Tables

C.8 Participatory institutions and social ranking: sequential logistic regression C.9 Traditional institutions and social ranking: Heckman selection model C.10 Robustness checks: subsamples C.11 OBC reservations in India: a summary

236 238 239 241

Figures 1.1 3.1 4.1 4.2 4.3 5.1 6.1 6.2 6.3 6.4 6.5 6.6 6.7 B.1 C.1 C.2

The path to identity activism Average male literacy rate by year in India, 1901–1931 Proportion of castes petitioning for a name change, 1901–1931 Rate of petitioning by 1901 literacy rate and years, 1901–1931 Predicted rate of petitioning by male literacy rate Density plot of ranked and unranked petitioning groups by land control Caste associational membership in India, 2011 Caste category representation among north Indian MPs, 1952–2004 Literacy by caste category in north India, 1920–1994 Literacy, political involvement, and caste polarization among Bihari Yadavs, 1960–2017 Caste associational membership and education in India, 2011 Untouchability practice in India, 2011 Successful ‘backwardness’ petitions, 1993–2014 Kernel density plot of ranked and unranked petitioning groups by public force employment Kernel density of petitioning and non-petitioning groups by literacy rate Proportion of castes petitioning for a name change by province, 1901–1931

10 70 87 93 98 119 138 140 142 144 145 156 165 209 219 229

Acknowledgements Even by the standards of academia, this project has been an extraordinarily long time in gestation, during which I have incurred many intellectual debts which I can only gesture at trying to repay. This project was begun at Stanford University, and I would like to thank David Laitin, James Fearon, Steven Haber, Saumitra Jha, Thomas Blom Hansen, Lisa Blaydes, Karen Jusko, Bethany Lacina, Avidit Acharya, Jessica Gottlieb, Amanda Robinson, and Milan Vaishev for their insightful comments. I would also like to thank the Center for Democracy, Development and the Rule of Law and the O’Bie-Shultz Dissertation Writing Fellowship in International Studies for their financial support. Special thanks go to Eliana Vaquez for her consistent kindness and understanding. This book is based on two extended periods of field research, supported by the American Institute for Indian Studies, the University of Rochester Faculty mobility fund, the Stanford South Asia Center, and the Stanford Graduate Research Opportunity Fellowship. Numerous other individuals provided me with help and support during my time in India. An inadequate and incomplete list would include Kunal Dutt, Paul Kenny, Peter Samuels, Maria Kornienko, Berenice Guyot-Rechard, Urmilla Singh, Jameson Karns, Riyad Koya, Neelanjan Sircar, Giles Verniers, Ashok Yadav, Achin Vanaik, and the staff of the Nehru Memorial Library and National Archives of India. I would like to thank Iswarya Subbiah, Neha Tariyal, and Steffanie Laloo for their dedicated and expert research assistance. This book achieved its final form in the exception intellectual environment of the political science department of the University of Rochester. I would like to thank my colleagues there for their kindness and example, and the department itself for the financial support. I would

xiv

Acknowledgements

like to thank Bing Powell and Gretchen Helmke who read or commented upon the manuscript. Jack Paine was a tower of strength throughout this period. I would like to thank Julia Doe for her support during the writing stage. At Cambridge University Press, I would like to thank Sohini Ghosh and three anonymous referees. The book is better as a result of their comments. My mother, father, and sister have also been by my side throughout. My heartfelt thanks to them.

1

Introduction One caste, two strategies In the year 1880, Samuel Sargunar, a deputy registrar in the revenue department of Chingleput district in the far south of India, published a small pamphlet entitled Dravida Kshatriyas. The book is concerned with the social status of the Shanans, a Tamil caste (or jati) traditionally associated with the disreputable occupation of palm liquor production, but many of whose members had recently become prosperous through their involvement in trade or (as in Sargunar’s case) the colonial bureaucracy.1

1

Throughout this book, the English term ‘caste’ will be used as a synonym for the Hindi term jati, reflecting common practice in South Asia. The majority of Indians are conscious of belonging to a jati, of which there are several thousand within India as a whole, several hundred within a given state, and usually one or two dozen within a given village cluster. Jatis are defined by endogamy, common stories of origin, and by (widely varying) restrictions on social contact between groups. Most jatis also possessed at one time a traditional occupation, and the relative status of jatis is often defined by its associated occupation. Many jatis are also associated with a single region and religious affiliation, though this is not always the case. Some tribal groups and ‘communities’ of non-Hindus are occasionally considered to be the functional equivalents of jatis as primary identity units, especially in political contexts. The English term ‘caste’ is sometimes used to describe two other categories of identities. Varnas are the categories into which society is organized in the Sanskrit texts that form the sacred books of Hinduism. In order of prestige, they are: the Brahmins (priests), the Kshatriyas (warriors), the Vaishyas (traders), and the Shudras (farmers and craftsmen). An informal fifth varna is composed of the so-called untouchables. In practice, varnas serve as legitimating super-categories to which jatis seek to attach themselves. While varnas are important in how Indians think about the caste system, there is

2

From Hierarchy to Ethnicity

Sargunar’s argument was that a terrible historical mistake had occurred: the Shanans, instead of being liquor traders, were really kings and warriors, the ancient rulers of all of south India, and had gained their current bad reputation due to a revolt of the ‘servants’ against their natural Shanan masters (Hardgrave 1969: 81–84). The natural solution was for the Shanans to reclaim their former status by readopting the habits of high caste Hindus. Over the next three decades, wealthy Shanans enthusiastically took Sargunar’s advice. A series of books and genealogies ‘proved’ that the word ‘Shanan’ was a Tamil synonym for king (Hardgrave 1969: 82–87). Shanans petitioned the colonial census authorities three times to allow members of the group to be recorded as Kshatriyas, and when their petitions were refused many still managed to do so, despite warnings that this would depress the numbers of their own group and raise the numbers of the upper castes (Francis 1902; Molony 1912; Boag 1922). Some Shanans began to wear the sacred thread (the traditional symbol of Hindu orthodoxy), hire Brahmin priests to perform their ceremonies, practise vegetarianism, discourage widow remarriage, and even tie their dhotis and wear their hair in the upper caste fashion (Hardgrave 1969: 112). Shanan weddings became lavish displays of self-assertion, costing thousands of rupees, with the grooms carried on palanquins by other castes, a traditional mark of kingship. At the same time, the wealthy reformers were at pains to de-emphasize their links to those Shanans who remained involved in palm liquor production. Not only did wealthy urban Shanans cease marrying and dining with poorer ones but they also created a system of kangaroo courts to punish with beatings those found to be selling liquor (Hardgrave 1969: 106, 137). By the 1920s, however, a group of younger Shanan activists, led by W. P. A. Soundrapandian, began to question every element of this approach considerable regional variation in the numerical presence of groups claiming linkages to the three higher varnas, and there is considerable variation in social status within the Shudra category. For statistical and redistributive purposes, social scientists (and, more importantly, the Indian state) group jatis into categories, such as ‘scheduled castes’, ‘other backward classes’, and ‘upper castes’. While these groupings of somewhat similar groups are sometimes referred to as castes, they are secondary to jati as a focus of political identification (Huber and Suryanarayan 2016), and do not have the long history of social construction characteristic of jati identities.

Introduction

3

to caste self-assertion. They argued that instead of trying to advance themselves within the caste hierarchy, Shanans should reject this hierarchy entirely. Brahmin priests and upper caste hairstyles were discarded. Wealthy Nadars began holding public banquets (where all attendees ate the same food) to emphasize their solidarity with their poor coethnics, as well as holding and financing scholarships for their education through the Nadar caste association, the Nadar Mahajana Sangam (Hardgrave 1969: 170–181). Nadars abandoned ostentatious weddings with Brahmins in favor of ‘self-respect’ weddings officiated by representatives of the Sangam (Templeman 1996: 72, 73). All these activities occurred simultaneously with the expansion of the political involvement of the Sangam (notably in the campaigns of Soundrapandian himself for the provincial legislative council and local district board), and the demand for affirmative action in government jobs. Over the next few decades, under a variety of party and ideological labels, elite Nadars would use the institutions and group consciousness developed in this period to win considerable political power in Tamil Nadu, much of it at the expense of the high caste Hindus they had earlier tried to emulate. The Nadar experience was atypical only in the quality of the archival record. In much of the colonial world, the decades before independence saw a rapid increase in the political importance of ascriptive identities among nascent political elites, with groups large and small forming organizations, petitioning government bodies, and distributing propaganda. For instance, the colonial period saw Yoruba elites in Nigeria begin to organize their political conflicts around ancestral cities (Laitin 1986: 120–123) and the formation of the Malay ethnic category in Malaya (Shamsul 2001). Especially in India, the late colonial period was a golden age of caste activism, during which hundreds of caste sabhas (associations) were formed in all regions of the subcontinent: between 1901 and 1931, 1,130 petitions were filed with the census authorities for a change of caste name. Even more interesting than the general rise in ethnic or caste consciousness was its uneven distribution across groups, with many individuals disdaining narrow ethnic appeals in favour of the broader rhetoric of imperial loyalty or incipient nationalism. Even seven decades after independence, the caste identities mobilized during this period remain central to political behaviour in India, with elections featuring political parties relying on mobilizing caste ‘vote banks’ (Chandra 2004). Prominent examples include Mayawati’s Chamars in Uttar Pradesh, Laloo Prasad Yadav’s Yadavs in

4

From Hierarchy to Ethnicity

Bihar, and Hardik Patel’s Patels in Gujarat. Similar types of identity politics (both violent and non-violent) are found in many poor countries (Horowitz 1985). The changing strategies of Shanan activists also underscore a basic change in the way ascriptive identities are conceived that has occurred in many countries over the past two centuries. While most scholars today think of ethnic groups as ‘conceptually autonomous’ categories, there were (both in South Asia and elsewhere) many cases of groups that relied on external legitimation and emphasized their similarities to high-status groups over their own distinctive characteristics – where upwardly mobile members of poor groups sought to assimilate the values and behaviours of rich ones rather than challenge them. Such ‘ranked’ identities (Horowitz 1985) were common in many parts of the world before the Industrial Revolution. In India, where ranking was highly salient during the colonial era, the gradual evolution of a very different ‘ethnified’ view of identity was one of the key events of the twentieth century (Jaffrelot 2000), creating an additive, voting block identity politics that resembles in certain respects ethnic politics in other parts of the world. This book describes the causes of the upsurge in caste activism that has occurred in India over the past century, and the strategic choices made by caste activists from upwardly mobile poor groups as to what role caste should play in their political careers. This resolves itself naturally into two questions. First, why do some identities become the focus for elite activism? Second, why do some activists participate in maintaining existing ranked identity systems by rejecting opportunities to create a conceptually independent identity of their own? Finally, it will show how the differential and occasionally ranked nature of identity mobilization in modern India has influenced its society and politics.

Caste as a puzzle Non-scholars, particularly outside of India, often do not know what to make of caste. Many features of the caste system seem to set it apart from the forms of social difference with which Europeans and Americans are familiar, including the very large numbers of groups, the religious legitimation, the non-visible markers of difference, the subtle hierarchical relationships between groups, and the ties to the traditional occupation. This has led many to conclude that caste is distinct from ethnicity, should be analysed within a different analytical framework from ethnicity, and is unique to the

Introduction

5

Indian scene. This conception of caste as exceptional is closely related to the idea, popular in the nineteenth century, that caste would become less salient once India became more ‘modern’ (K. Marx 1853). Caste, however, has much more in common with other forms of identity than a casual view would suggest. Virtually every aspect of caste identity has a parallel elsewhere in the world. Non-visible ascriptive differences can still be powerfully motivating, as any visitor to Northern Ireland can attest. Similarly, religious principles have been used to justify a wide variety of systems of ethnic division, including American slavery (Fox-Genovese and Genovese 1987). In fact, many societies outside India feature ranked, religiously legitimated, and occupationally associated forms of stratification that closely resemble the caste system. These include the Burakumin minority within Japanese society; the Haratin minority among Maghrebi Arabs; the division between nobles, herdsmen, holy men, and artisans within Tuareg society; and systems of clan ranking among the Amhara people. Many further examples could be cited, particularly within sub-Saharan Africa, where craft knowledge or slave origin is the defining feature of many minority groups (for example, Larick 1991). Similarly, caste has failed to fade away over time, and caste identities play a robust political role in both colonial and post-colonial India (Rudolph and Rudolph 1967; Chandra 2004). These generalizations ignore the enormous variations in the way caste identities have been expressed, both over time and between groups. The political role of caste identities has changed over time and has always been very uneven between groups and regions, with the mobilization by a single leader around a caste party being the exception rather than the rule. Similarly, in the past century, India moved from a system where ranked ideas of identity enjoyed a prominence probably unique in the world to one where most mobilization is on an unranked basis. This is not to say that caste identities in India are now conceptually identical to racial identities in the United States or ethnicities in Uganda. Three unique features of caste politics in twenty-first century India stand out as especially puzzling from a comparative perspective. One is the extraordinary diversity of caste identities: India is thought to have well over four thousand discrete jatis, and jatis that make up more than 10 per cent of a state’s population are considered exceptionally numerically powerful. Calculated on a jati basis, India is almost certainly the most ethnically diverse country in the world. Second, while identity and socioeconomic status (SES) are highly correlated in many countries, this correlation

6

From Hierarchy to Ethnicity

is especially strong in India, which has the third highest identity-class association in the world (Baldwin and Huber 2010). Finally, while many caste identities form the basis for politicized and consolidated ‘vote banks’, this is not true of all identities, particularly at the extremes of the socioeconomic spectrum. Understanding the way in which caste identities have developed historically is essential to understanding not just the way in which caste has become a more ‘normal’ political identity but also the ways in which it remains unique. Identity politics is a well-explored topic. Why do some groups engage in identity-specific political activity (or mobilize) while others do not? Why, for instance, is Scottish identity more politically salient than Welsh? Or Yadav identity more politically salient than Kahar or Bania identity? The causes of the rise of identity-based activism in the twentieth century have been the focus of scholarly discussion formidable in both quality and quantity. In the past few decades, authors such as Chandra (2004, 2012), Posner (2004, 2005), and Lacina (2014, 2017) have developed sophisticated theories as to how social identities are formed and become politically relevant. They stress the instrumental and constructed nature of identity. Individuals choose particular identities to ‘activate’ or ‘mobilize’. Their choices reflect a desire to gain resources, either by forming ethnic blocs large enough to succeed in a political competition (R. Bates 1983; Wimmer 1997; Posner 2005; Lacina 2017) or to gain other types of material benefits from the state (E. Weber 1976; Cassan 2015). For these authors, the mobilization of Nadar identity in the 1920s and 1930s is readily explicable and indeed a textbook demonstration of what ethnic politics should look like. The organization-building, the elision of internal differences, the emphasis on common traits, the demand for transfers, and the gradual co-optation of fraternal organizations for electoral ends all reinforce the impression of identity activism as just another instrumental political tactic, albeit a somewhat sticky one. All these behaviours parallel the types of mobilization strategies pursued by ethnic groups in Africa, Europe, and Southeast Asia, and have many affinities with the development of national identities in all parts of the world (Gellner 1983; Anderson 1994). The Nadars’ behaviour in the earlier period, however, is a puzzle for existing theories of ethnic politics. Why should elites emphasize their differences with members of their own group, their most obvious potential political supporters? Why should their activism rely so heavily on the external legitimation of the Hindu tradition? And why should they, rather

Introduction

7

than claiming a unique history, emphasize their similarity to other groups, even to the point of denying themselves a separate identity? For contemporary European students of the caste system such as Risley (1892), and for later structuralist scholars such as Dumont (1980 [1966]), the answers to these questions would have seemed either obvious or irrelevant. Caste, to these scholars, was fundamentally different from ethnicity or tribe. Castes were subordinate parts of a larger whole, defined by a single cultural and ideological tradition, itself defined by the Sanskritic classics and the primacy of the Brahmin caste. Castes were arranged in a hierarchy from clean (and high status) to dirty (and low status) based on their adherence to a set of normative behaviours, and this hierarchical positioning was what defined group boundaries. This understanding of caste as a pervasive aspect of Hindu civilization was also influential among Indians, including both those who saw the caste system as a social good (Yogananda 2003 [1946]) and those who saw Brahmin power as deeply illegitimate (Ambedkar 2014 [1936]). The structuralist model of caste, however, has a great deal of difficulty accommodating change. If social hierarchy was really a fundamental aspect of Indian civilization, how could the Nadars so blithely defy it in the 1920s and 1930s, even as they clung to their caste identity? In fact, if hierarchical ideas were as fundamental to caste as the structuralists claimed, even the Nadars’ earlier attempts to climb the ladder seemed to embody a worrying degree of fluidity in group status (Francis 1902). Moreover, structuralists had only vague explanations for why a hierarchical identity system would emerge in the first place, particularly after the discrediting of the racial explanations current in the colonial era. The shortcomings of the structuralist approach were the starting point for a wholesale critique of the literature on identity and colonialism, most associated with the work of Nicholas Dirks (2002, 1993) but also found in the work of other Indianists (Cohn 1987; S. Bayly 1999; Gupta 2000), and scholars of other parts of the world (Laitin 1986; Berman 1998). These accounts, echoing broader constructivist trends in the social sciences, emphasized the role of the colonial state and the forms of knowledge it developed. British officialdom, in this view, turned fluid and contested concepts (such as caste) into rigid and hegemonic ones. The colonial state used its institutions, particularly the census, to create a set of rigid, mutually exclusive, categories from a far more complex pre-colonial reality. These studies parallel the large existing literature on how states can shape the development of both national and ethnic identities (E. Weber 1976;

8

From Hierarchy to Ethnicity

Brass 1985; Brown 2003; Miguel 2004; Singh and vom Hau 2016) and on the long-term influence of colonialism on the political and economic patterns of developing societies (Acemoglu, Johnson, and Robinson 2001, 2002; Banerjee and Iyer 2005; Iyer 2010; Lee and Schultz 2012; Lee 2017). As we will see, there is formidable evidence that South Asian states were closely involved in shaping social identities both before and during the colonial period. However, several aspects of the explanation remain puzzling. If colonialism promoted a novel hierarchical form of caste identity, what are we to make of groups (such as, laterally, the Nadars) who both embraced caste identities and derided the hierarchical aspects of the caste system? How can we explain why some elites energetically embraced caste identities while others did not? And how can we reconcile the sustained and enthusiastic engagement of many Indian elites with the concept of caste with the half-hearted, temporary, and remote nature of the colonial interventions hypothesized to cause them?

This book This book builds on the ethnic politics, structuralist, and constructivist literatures to explain the changes in identity politics that occurred in twentieth-century India. Like the ethnic politics literature, it shows the strategic motivations of elites in making specific identities salient. Like the structuralist school, it shows that in many times and places, identity is not equivalent to category and involves a strong hierarchical element. Like the constructivist literature, it shows the relative flexibility of both caste identities and caste hierarchy positioning, and the role of the colonial state in shaping the forms caste activism took. These insights are the basis of a new theory of both ethnic mobilization in general and ranked mobilization in particular. It supplements existing theories of group mobilization that focus on group size and state policy by showing that the socioeconomic status of the group has a non-linear impact on mobilization. It contributes to the discussion of ranked identities by showing both how they differ from a simple correlation between power and identity and how this differing mode of constructing identity stems from the structure of political distribution and the differing incentives of individuals in patrimonial political systems and modern democracies. The theory provides an explanation for why South Asia has historically been so permeated by ranked identities relative to other parts of the world. It suggests that the hierarchical elements of caste systems, far from being

Introduction

9

unique to India, are merely an extreme manifestation of trends found in most developing countries. These trends are of more than historical interest: although ranked identity is now very uncommon in the macropolitical sphere and in urban areas, it has left India two major legacies – a very high level of identity diversity and a high degree of correlation between identity and socioeconomic status, both factors widely thought to have negative effects on economic development and social conflict (Easterly and Levine 1997; Alesina, Baqir, and Easterly 1999; Alesina and Ll Ferrara 2005; Miguel and Gugerty 2005; Baldwin and Huber 2010; Huber and Suryanarayan 2016). The argument shows how the uneven spread of caste mobilization and its focus on jati explains several interesting elements of contemporary Indian politics, including the coexistence of a widespread disdain for ‘casteism’ with an equally widespread use of caste as a mobilizing identity (Assayag 1995), and the failure of the economic ‘rise’ of certain traditionally poor caste groups to lead to improvements in descriptive representation (Jaffrelot 2003). In showing how certain group identities became politically important, it provides the background for accounts of how politicians from specific ‘dominant’ caste identities have been able to exercise a controlling influence over public policy in most Indian states (Frankel and Rao 1989; Srinivas 1994; Lee 2019). Much of the empirical basis for this project is historical and qualitative in nature, reflecting the difficulties in collecting quantitative evidence on events in the pre-colonial period, and the reluctance of both the Indian government and private organizations to collect and publicize quantitative data on caste identities after independence. In the colonial period, however, this approach can be supplemented with a large panel dataset of the petitions filed by caste groups with the Indian census authorities, a common form of caste activism. The explicit goal of these petitions was to change the way in which the census referred to the group – a goal in which they were almost invariably disappointed. However, petitions provide a window into the complex processes of identity formation that are usually hidden from the historical record. In particular, they represent an index of the presence of an activist group and the rhetoric of that group and allow, for the first time, the construction of a measure of the ranked rhetoric. The panel structure of the data enables comparisons of petitioning behaviour within groups or categories of groups, a crucial factor given the many plausible cultural and historical differences between castes.

10

From Hierarchy to Ethnicity

While the empirics are focused on India, the theory has implications for a wide variety of periods and cases – for example, it provides an example for the much-debated divergence in the forms which racial identity takes in the United States and Latin America (Desdunes 1907; Degler 1971; Hickman 1997; Telles 2004), which it implies reflect differences in political institutions rather than cultural ones. Overall, it suggests that the ‘modern’ concept of an ethnic group as an unranked identity is a product of specific, historically determined, institutional circumstances.

The argument This book focuses on a set of nested choices made by elites, in particular the elites from traditionally marginalized groups, regarding their political involvement. They can remain politically uninvolved, become politically involved while emphasizing a broad identity dimension, or choose between ranked and unranked forms of narrow identity activism. These nested decisions are summarized in Figure 1.1: conditional on identity activism, individuals must choose whether to emphasize ascriptive identities in their political appeals and whether any such appeals should be ranked or unranked. The Nadars, for example, became involved in politics in the nineteenth century and choose to focus their political efforts on their narrow caste identity rather than Tamil nationalism or political Hinduism. In the early twentieth century, the form that this activism took shifted from a ranked strategy (tied to the traditional hierarchy) to an unranked one. The meaning of these terms, and the reasons they made these choices, are discussed below. Figure 1.1 The path to identity activism

Introduction

11

Why identity politics? Individuals possess an almost limitless number of descent-based social attributes, the possession of which divides individuals into ethnic categories (Laitin 1998; Chandra 2012). These attributes are organized into dimensions of closely related traits, sometimes assumed to be mutually exclusive. In India, for instance, the dimension ‘caste’ includes several thousand individual jatis. However, at any given time, the number of salient or activated dimensions is much smaller than the number of possible dimensions.2 Despite the constructed nature of social identities, there is no doubt of their importance. Large bodies of work have shown the influence of identities, even fairly recent and artificial ones, on voting (Ordeshook and Shvetsova 1994; Chandra 2004; Ichino and Nathan 2013; Carlson 2015; Huber and Suryanarayan 2016), conflict (Fearon and Laitin 2000; Wilkinson 2006; Cederman, Weidmann, and Gleditsch 2011), and public goods provision (Easterly and Levine 1997; Alesina, Baqir, and Easterly 1999; Miguel 2004; Banerjee and Somanathan 2007; P. Singh 2015; Singh and vom Hau 2016; Lee 2018a), though most of these accounts do not directly address the question of where these identities come from. In some cases, the activation of a dimension implies the activation of all its component categories (Posner 2005; Chandra 2012). The activation of the ‘ancestral city’ dimension, for instance, made Yoruba from all ancestral cities value this attribute (Laitin 1986), while the activation of language in Malawi made both Chewas and Timbukus value their identity more (Posner 2004). However, in practice some categories tend to be more salient than others, even within an activated dimension. While caste is a politically salient identity dimension in India, it is more important to the political alignment of some groups than others. This may reflect other groups’ preference to align on an alternative dimension, such as religion. Alternately, it may simply reflect groups that are not very politicized or do not use ascriptive traits as a basis for their political decision-making.

2

Most of these accounts have tried to explain the salience of broad identity dimensions (such as ‘caste’ or ‘race’) rather than the categories within those dimensions (‘white’ or ‘Brahmin’). Chandra (2012) notes that ‘a change in identity dimension, furthermore, is typically seen as predicting perfectly which category is likely to become activated’.

12

From Hierarchy to Ethnicity

Identity shift does not occur spontaneously and is closely associated with the actions of activists and of the parties and associations that they control. While elite position-taking is not always associated with mass change, it is often a necessary precondition for it. Activists actively campaign to convince citizens that particular identities are salient: imagining a shared history for the group, defining who is legitimately a member, and promoting certain types of actions as desirable for members. These ideas provide the vocabulary and rules by which ordinary people express their identities. For this reason, most existing work on caste politics has tended to focus on the growth and messaging of caste parties or caste associations rather than individual behaviour (for example, Rudolph and Rudolph 1967; Kothari 1970; Jaffrelot 2000; Chandra 2004). Accounts of identity shift outside of India have also focused on activism (Laitin 1986). Note that this does not necessarily mean that the identities ‘take’ – that ordinary individuals find them meaningful. In many cases, particularly historical ones, answering this question is impossible due to a lack of reliable public opinion or census data. However, in many contexts, the success of these campaigns is demonstrable in the form of increased voting and identification along group lines. One noticeable form that such an activist campaign can take is the formation of a specifically ethnic political party. However, ethnic and nationalist campaigns can also be organized by ostensibly non-political associations, individual leaders, or informal groups. In colonies, where party formation was difficult, such non-partisan activist groups played a leading role in identity politics, though they were quick to take advantage of those electoral opportunities that did present themselves. Even in the post-independence period, where identity-based parties have become possible, many identity politics projects are still pursued by individuals or factions within the context of larger political parties. This type of associational activity is especially important for groups that are not able to employ the state as an ally. Many scholars have traced the origins of identity politics to the policies of the state, either in favouring some identities over others or by creating the vocabulary in which such identity projects could be expressed (Brass 1974; E. Weber 1976; Laitin 1986, 1998; Scott 1998; Dirks 2002; Luong 2004; Peisakhin 2010; Cassan 2015). A state might establish benefits that incentivize the adoption of certain types of identities (Cassan 2015) or set up an educational system that inculcates certain types of identities (Darden and Grzymala-Busse 2013). The Tanzanian state, for instance, is generally thought to

Introduction

13

have suppressed the expression of ethnic identities and encouraged the development of a broader national identity than neighbouring African countries (Miguel 2004). While such conscious programmes of state identity politics are more common among post-independence states, there are also well-attested examples from the colonial era, such as the Yoruba in Nigeria (Laitin 1986) and caste groups in Punjab (Cassan 2015). The other common factors cited in the existing literature as explanations for mobilization are potential group size (assumed, at least as a methodological convenience, to be exogenous) and democracy. Some contemporary authors have emphasized the importance of democracy, parties, and elections in promoting identity mobilization (Chandra 2004; Wilkinson 2006; Eifert, Miguel, and Posner 2010). In this view, ethnic entrepreneurs shape identities in such a way as to create minimal winning coalitions within the population. Ethnicity should thus become more salient during election season, or when the political system becomes more competitive. One natural extension of this idea is that large groups should mobilize more often than small groups, since a large group is politically more viable than a small one (Kasfir 1979; Chandra 2004; Posner 2004, 2005; Rao and Ban 2007). Members of small categories, in this view, will seek to join larger categories (or redefine the categories) rather than mobilize a category of below minimal-winning size (Posner 2004, 2005). However, these theories do not explain why identity shift sometimes occurs within authoritarian regimes or why ethnic activism is often observed among small groups that have little or no chance of winning an election on their own.3 This book supplements these accounts by focusing on the role of education, measured at the group level. While many authors have argued that education influences identity through the content or language of instruction (E. Weber 1976; Posner 2003; Darden and Grzymala-Busse 2013), even politically neutral education can have impacts on identity politics. As groups grow more educated, they are more likely to produce individuals with the literacy, sophistication, and disposable time necessary 3

Some authors have emphasized that identities that are highly visible can easily become the basis for distributional decisions, or may have increased cognitive salience (Alcoff 2006). This might, for instance, explain why ethnicity is often more salient than class in poor countries (Chandra 2004). However, such theories cannot explain the expansion of identities, such as caste, where members are often not readily distinguishable from each other physically or behaviourally.

14

From Hierarchy to Ethnicity

to become involved in politics. Increasing levels of education should thus have a strong initial impact on individual politicization, which should in turn be strongly (though imperfectly) related to group-level politicization. This argument has many affinities with the literatures on the causes of nationalism (Deutch 1969; Gellner 1983; Anderson 1994), the growth of the nation state more generally (M. Weber 1947; Bendix 1977), and informal discussion of ‘backward’ groups (Horowitz 1985) which link the socioeconomic causes of identity change to the broader phenomena of social modernization. However, these accounts fail to explain why the effects of social change are not apparent in all cases, and why the initial stages of modernization are often accompanied by a resurgence of ‘traditional’ or sub-national identities. In fact, the effect of education is not linear. As middle-status groups grow more educated, their members become more likely to possess the resources and contacts necessary to be politically successful outside of their own group. Elites balance the negative returns of being associated with a particular group (and alienating other potential supporters) with the positive returns of having a limited but reliable support base. Politicians with a potentially broad appeal are less likely to attempt to activate their narrow ascriptive identities, instead preferring to emphasize broader ascriptive identities or de-emphasize ascriptive identities entirely. Elites from poor groups, by contrast, try to construct narrow categories in which they will not have wealthy and talented political rivals. This is a modification of Posner’s (2004, 2005) argument: while elites do seek to shape identities to maximize the size of their political constituency, this dynamic is balanced by the desire of rent-seeking elites to be the leading figure in a particular constituency. This claim fits what we know about the backgrounds of caste politicians in modern India such as Mulayam Singh and Mayawati or in colonial India such as Sir Chhotu Ram and W. P. A. Soundrapandian. They are members of the first educated generation of an upwardly mobile group, who found in caste mobilization a road to political success that their own modest contacts and credentials would have been unlikely to have brought them and they competed with the Brahmin elite on their own terms. Members of the most educated groups, therefore, may be involved in politics – perhaps even overrepresented – but their group’s identity will not be publicly emphasized, as group members prefer to project identities with a broader appeal. In colonial India, the elites of the most educated groups tended to disdain caste-based position-taking and were

Introduction

15

correspondingly predominant in the Congress and the colonial bureaucracy, which emphasized identification with broad constructs such as nation and empire as the focus of loyalty. In post-independence India, upper caste groups are both less likely to use caste rhetoric and to vote as a block and more likely to support parties based on non-caste social identities, such as the Bharatiya Janata Party and the various communist parties.

Why ranked identities? When predicting identity change, the ethnic politics literature has made a set of implicit assumptions about what ethnic identities look like. In particular, it assumes that the most important aspects of identities are the ways they define group members and non-members. In the language of Abdelal et al. (2006), they focus on ‘constitutive norms’, the rules that define group membership. However, there are other aspects of group identity: Abdelal et al. mentioned ‘relational comparisons’ (views and beliefs about other identities or groups) and ‘cognitive models’ (worldviews or understandings of political and material conditions and interests). This book focuses on an aspect of identity that is relatively understudied in the political science literature: ranking. Classic descriptions of ethnic politics have divided ethnic identity systems into two ideal types: ranked identity systems (in which groups are defined by relationships of superiority and subordination to each other) and unranked identity systems (in which groups are conceptually autonomous) (M. Weber 1958; Horowitz 1985). This distinction is based on the fact that in some identity systems, such as the caste systems of India and the racial systems of the early colonial New World, ethnic group relations are fundamentally structured around status inequality. Although the differences among these types of identities are widely acknowledged among political scientists, there has been little or no theorizing on their functions or causes. In particular, we know very little about why many subaltern groups participate in ranked systems defined by others, and why norms of ranking have gradually declined in many countries.4 4

Like the discussion on mobilization earlier, this discussion will focus on variations within a given identity dimension. The main portion of the argument takes for granted that colonial India was a society in which religiously legitimated ideas of social hierarchy existed as an ideological possibility.

16

From Hierarchy to Ethnicity

While the differences between ranked and unranked identities are many and subtle, they are especially different in their approach to inter-group relations. Unranked identities, the ‘normal’ type in the existing literature, emphasize the differences between groups. Unranked groups are ‘conceptually autonomous’: each could exist without the others, and there is no consensus on which group is superior. Ranked identities, by contrast, emphasize not differences, but values. Some groups are considered normatively superior to others, and all other groups attempt to imitate their behaviour, or even assimilate into the higher group. Rather than distinguishing in and out, ranked identities distinguish high and low. They thus tend to emphasize the relational aspects of identity over its constitutive aspects: up and down over in and out. These ideas may or may not be tied to some larger ideological or religious project. Similarly, ranked and unranked identities also differ in their effect on intra-group relations. Within unranked groups, all members of the category are formally equal, and group leaders are at pains to de-emphasize previously salient divisions within the category. Within ranked identity groups, there may be considerable internal variation based on adherence to the norms of the ranking system, and this may lead to subtle patterns of status differentiation even within groups.5 Ranking has been traditionally seen as characteristic of caste identities and as the major difference between caste and other forms of ethnic identity (for example, Dumont 1980). However, ranking norms are not confined to India. For instance, the Burakumin people of Japan have an identity defined (at least in the minds of others) by ideas about pollution and status similar to the Indian model (Amos 2011). Similarly, while racial divisions are often thought of as a rigid binary, outside of the twentieth-century United States they have encompassed a more subtle and value-driven gradation between black and white (Desdunes 1907; Degler 1971; Hickman 1997; Telles 2004). Over time, ideas of social hierarchy gradually became a less important element in the ascriptive social difference than it had been earlier. The recent trend in the political science literature has been to emphasize the

5

For reasons of expositional convenience, I will at times refer to these ideas as being ‘Sanskritic’, although they have many origins other than the Sanskrit corpus. This definition of ranked identities differs somewhat from the definition provided by Horowitz (1985). See Chapter 2 for a more thorough discussion of Horowitz’s approach.

Introduction

17

similarities between caste and other types of identity (Chandra 2004: 18, 19) or to emphasize the contestation of norms of ranking by traditionally low-status groups (Gupta 2000). However, the decline of ranking extended well beyond India, and encompasses shifts such as the gradual shift towards a sharply dyadic view of racial identity in the United States, the decline of cosmopolitan elite identities such as ‘Ottoman’ in favour of national ones, and the decline of status-based caste distinctions in twentieth-century Japan. Within India, this change in the nature of caste has been widely noted (Rudolph and Rudolph 1967), and we have a number of accounts of how caste has been ‘substantialized’ (Barnett 1977), ‘ethnified’ (Jaffrelot 2000), ‘politicized’ (Assayag 1995), or ‘culturized’ (Natrajan 2011). In practice, this difference mirrors two distinct approaches to lower caste activism. During the early twentieth century, many Indians participated in activist programmes that combined modern associational forms and formalized ideology. Some caste activists emphasized their distinctiveness from other groups and rejected or de-emphasized the traditional caste status ordering, while other groups chose to emphasize their hierarchical superiority to other groups, a process which often led them to adopt the values and identities of traditionally high-status caste groups, even to the point of denying their own group’s independent existence. Among Indians, this latter strategy is referred to as ‘Sanskritization’ (Srinivas 1956, 1966), while the unranked alternative has attracted names such as ‘Mandalization’. Where do these differences come from? This book suggests that the popularity of these approaches differs across time and countries, due to differences in the political system. Unranked identities help build a homogenous support group for an aspiring leader among his peers and co-ethnics. They are thus politically useful in societies where power is distributed based on popular support: democracies or quasi-democracies. The strategy will be especially attractive among larger groups, where the gains from forming a voting bloc are the largest. Ranked identities, by contrast, help a leader build patron–client ties, both with the elite above him and with clients below him, with the rituals of ranking mirroring and reinforcing hierarchical political relationships. In the same way that the ideology of ethnic pride or nationalism legitimizes and organizes social and political solidarity, ranked identities legitimize and organize social and political difference. These ties are thus most useful in patrimonial societies, where power is distributed based on personal connections within the elite. Within countries, ranked identities

18

From Hierarchy to Ethnicity

are more popular in rural areas, and in areas with more informal state institutions, the areas where informal patron–client ties are more important in structuring political interaction. This explanation has several advantages over existing accounts of the ranked nature of identity politics in colonial and pre-colonial India (or, in older work, theories of the ‘cause’ or ‘origin’ of caste). First, this explanation can be applied outside India, unlike theories based on racial differences or a remote history of Aryan conquest (Risley 1892; Leopold 1974), the importance of economic and occupational specialization (Dubois 1906; Freitas 2010), and the close association of caste within Indian culture and civilization (Dumont 1980 [1966]). It also challenges accounts that focus on the role of the colonial state (Cohn 1960; Srinavas 1966; Gupta 2000; Dirks 2002). As Dirks (2002) and Cohn (1987) pointed out, colonial preoccupations (both scholarly and political) with caste identity shaped the process of Sanskritization. However, colonial policies and ideologies, since they affected India as a whole, can at best provide only a partial explanation for Sanskritization. Moreover, given the emphasis on ranking in most colonial accounts of caste, such theories have difficulty with the fact that many caste groups rejected or ignored Sanskritic categories, a trend that has only accelerated since independence. While ideas about ranking (and the broader language of the Sanskritic caste system) were widely known in early twentieth-century India, their appeal seemed to differ considerably among social groups.

Implications of the theory Any book about identity mobilization using Indian data faces a major challenge in generalizing the findings to the rest of the world. Most notably, the embrace of ranking common in most popular depictions of the caste system seems to make caste sui generis. However, some scholars have moved towards the other extreme and treated caste as being similar to ethnicity, and caste politics as having similar motivations and dynamics as ethnic politics in other parts of the world. The argument discussed above can be seen as mediating between these two approaches to the external generalizability of India’s caste politics. On the one hand, it acknowledges that many aspects of caste identities, particularly before 1947, appear influenced by ranking in very profound ways, while similarly acknowledging that the caste activists of

Introduction

19

twentieth- and twenty-first-century India have little time for these norms. In fact, it shows how and why ranking became less important over time: how caste became ethnicity. While this process is not complete, it has been recent enough that both views of the caste system are consistent with the behaviour of specific groups in the twentieth century. By developing for the first time an explicit theory of what ranked identities are and how they emerge, the argument puts the emphasis on social ranking in the political ideologies in pre-1947 India in a comparative context. It shows how they stemmed from a confluence of factors that was unique in its strength in early modern India: a patrimonial political system that rewarded the development of rich and carefully ordered political networks, and a level of wealth (at least among the elite) high enough to support ritual specialists dedicated to articulating these orders. Where these conditions held in a slightly less intense form, as in many other parts of Asia and the Americas in the early modern period, ranked distinctions that bore a striking resemblance to India’s also emerged. Even if we accept this line of argument, it is unclear why the developments of the colonial and pre-colonial periods are still of interest. After all, if both India and the rest of the world have converged to a situation of fully mobilized unranked categories that closely parallels existing theories of ethnic politics, what is the point of discussing how things were different in the 1820s, or even in the 1920s? This book shows, however, that the uneven, often ranked mobilization of the colonial period has had tangible consequences for modern India. Some of these represent elements of the colonial pattern that never really went away, such as the preferences for members of the most educated castes for political ideologies (nationalism, Hindu nationalism, communism, and so on.) that attempt to activate social identities larger than jati, and the persistence of ranked rhetoric (and discrimination based on ranked distinctions) in areas with little exposure to state authority, particularly villages in more isolated parts of the subcontinent. Other effects are more indirect, though possibly more consequential. In particular, India, relative to other parts of the world, is notable for having very high levels of both social diversity (probably the highest levels in the world, if jati is considered to be the relevant social category) and economic inequality between groups. Both these facts have been shown to be potentially important for the political economy of countries and the ability to generate a political consensus to supply public goods

20

From Hierarchy to Ethnicity

(Easterly and Levine 1997; Alesina, Baqir, and Easterly 1999; Miguel and Gugerty 2005; Baldwin and Huber 2010). Both these facts, however, have obvious connections to the long history of ranked identity mobilization in India. A system where economically advantaged individuals sought to raise their social status by exiting their own groups would, over time, lead to the creation of a society with a large number of small groups closely associated with particular rungs of the economic ladder, an effect that would persist long after ranked rhetoric vanished from high-level politics.

Measuring activism The theory suggests that while identity expression is a product of economic factors, their content is a product of political ones. These two dynamics provide some working hypotheses on why ranked identities have been more prevalent in South Asia than in many other parts of the world. Identity politics in pre-colonial India was influenced by two stylized facts: the relatively high levels of wealth among political elites and the weak and unstable nature of their political authority. The first factor gave some social groups the educational and economic resources to construct highly articulated identity systems, while the second factor made it inevitable that these identity systems were largely hierarchical in form. In the colonial period, this picture was altered by the (slow) expansion of education and the (slower) expansion of political rights. This led economically prominent groups to use their new-found education to strive to improve their status within existing Sanskritic concepts of social ranking. However, hierarchical mobilization was displaced by unranked mobilization in areas that held elections. Over time, the intensification of these trends has made unranked caste mobilization very common in post-independence India, almost erasing the memory of the ranked approaches that preceded it. Colonial India was thus a period where education, patrimonialism, and democracy (the independent variables of this book) showed considerable variation across years, groups, and regions. The colonial period saw a large number of groups gain the social conditions for mobilization, while the political system was an incongruous mixture of patrimonial and democratic elements. There was a correspondingly high level of variation in the outcome variables. Caste-based activism varied from the apathetic to the enthusiastic, while caste activists took diametrically opposed approaches to the pre-existing system of caste ranking.

Introduction

21

In post-colonial India, the trends of the colonial period accelerated. Primary education became widespread even among groups and regions that previously had very low rates of educational attainment. At the same time, India became a parliamentary democracy, where the ability to win votes became the ultimate criterion for the distribution of political power. These trends led to a reduction in the diversity of ideological strategies pursued by groups – caste-based mobilization became virtually universal (outside of the most highly educated groups) and unranked rhetoric spread widely. The most visible manifestations of these trends, the ‘rise’ of the Other Backward Classes (OBC) castes to political power in northern India and the success of explicitly caste-based parties, have been widely discussed (Jaffrelot 2003; Chandra 2004). These temporal changes in the levels of the independent variables influence the nature of the empirical evidence presented here. In the pre-colonial and post-colonial periods, we should expect limited variance in the outcomes of interest; not coincidently, these periods also have tended to produce only limited quantitative data at the caste level. For this reason, while the book will describe developments in both these periods, it is difficult to make comparisons between groups, particularly on a large scale. To do this, it is necessary to examine the colonial period.

The colonial data This book makes use of a new dataset of Indian caste groups’ interactions with colonial census authorities. This data provides crucial evidence on both the existence of activist groups and the content of their rhetorical strategies: a window into a world of private, vernacular identity activism on which we have little direct evidence in most colonial countries. Read together with other information from the census and the available archival materials, they give us a view into how Indians interacted with the new ideas about caste that were becoming common in the colonial era. Using the panel structure of the data, we can make a rich set of comparisons: between different groups with similar traits, between the same group in different states, and between the same group in different years. This approach allows the book, unlike much of the existing literature on ethnic politics, to isolate economic changes from the many fixed cultural and social differences between groups. For instance, why does caste identity seem to have been more salient for Kayasths in the United Provinces than Kayasths

22

From Hierarchy to Ethnicity

in Bihar? And why did the Shanans experience such a dramatic reversal in their approach to ranking in the twentieth century? The group-level data supplement existing empirical accounts of ethnic or caste mobilization, which focus on the rise or fall of the salience of particular categories of identities (‘religion’, ‘tribe’, ‘caste’, ‘region’, and so on). Unlike these approaches, these data allow us to see which groups are mobilized within an identity category (caste) that is already potentially salient. Furthermore, they allow us to test effects that might be impossible to test in a small-n setting, such as the non-linear effect of socioeconomic status. In the panel dataset, attempted caste mobilization by elites is measured through petitions submitted by caste organizations to the colonial census authorities demanding a change of name, a common strategy of caste activists in this period. While they are an imperfect and partial measure of group activism, these petitions provide evidence about the existence and goals of non-state political activists whose behaviour is usually difficult to study in a comprehensive way due to a lack of source material. While petitioners represent a narrow subset of the members of any given caste, they do indicate the existence of a politically aware elite that took its caste identity seriously: in fact, the vast majority of petitions appear to have been submitted by formally organized caste associations, and petitioning is the best available index of the existence of such an association. The petition data also allow an examination of groups’ embrace of ranking. To get at this question, the dataset classifies petitions using the propensity to adopt upper caste group names. Dissolving one’s distinctiveness in the high-status group is in some ways a prototypical goal of a ranked system: Rough equivalents in less-ranked contexts would involve Welsh communities petitioning to be reclassified as English, Roma petitioning to be reclassified as Romanian, or Iraqi Kurds petitioning to be reclassified as Arab. Education is measured using caste-level literacy rates taken from the census of India. While this measure has a number of limitations (discussed later), it is the best available measure of group education and (given the absence of individual data from this period) of the presence of an educated elite within the group. Participatory and patrimonial institutions are measured by the reach of government employment (taken from census data) and the spread of elected local institutions (taken from an original panel dataset of district and local board elections). Data on participation in the Indian National Congress and the colonial bureaucracy allow us to see which groups were prominent in political arenas not tied to caste.

Introduction

23

The patterns in the quantitative data generally support the theory. Group-level literacy is positively related to petitioning, but very high levels of literacy are associated with lower levels of petitioning. However, these highly literate groups dominate contemporary arenas of non-caste-based political action, like the colonial bureaucracy and the Congress Party. These findings remain constant when we compare within castes, provinces, and years. These findings paint a picture not dissimilar to what we see in India: caste identity was strong among upwardly mobile middle-status groups, while the educated elite favoured the broader appeals of religions or nation. Among petitioning groups, hierarchical rhetoric is dominant among landed groups and groups with few state employees – the groups most exposed to patrimonial institutions. Unranked rhetoric is dominant among large groups in areas with electoral institutions – the groups that stand to gain the most from competition based upon group numbers – though elections have no effect on smaller groups. These findings are robust to controlling for some of the more obvious alternative predictors of ranked mobilization, such as ascribed caste status. It should be noted, however, that the modest number of petitions means that these comparisons have a smaller sample size than the models of mobilization and are correspondingly less robust. These findings match the overall patterns we see in petitioning. Overall, petitions rise over time, but proportionally fewer of them seek upper caste names. This parallels larger trends in colonial India towards (relatively) higher levels of political participation and (relatively) higher levels of literacy. Both these trends were especially strong in southern and western India and in areas that were under direct British rule, and both the move towards petitioning and the de-emphasis on ranking were especially marked in these areas.

Limitations of the colonial data Like any quantitative study, the petitioning tests trade depth for breadth, sacrificing deep knowledge of particular groups in return for the gains from analysing a wide variety of cases. The debate between qualitative and quantitative research methodologies is an old one, with persuasive arguments on both sides. Without taking an absolute position in this debate, it is worth mentioning two reasons why a quantitative approach is worth pursuing in this case. The first of these is the sheer dominance of small

24

From Hierarchy to Ethnicity

n-studies, both qualitative and quantitative, in the literature on identity.6 Many authors examine the competition between two identity dimensions within a single group (Laitin 1986), or variation between a small number of groups within a single region (Miguel 2004; Posner 2004). This approach limits the extent to which we can assess variation in identity mobilization within a single dimension, or the extent to which we can identify non-linear or second order effects. Second, in many cases the process of identity activism, particularly historical activism, is poorly served by the archival record and is impossible to survey. Most caste associations were private, often ephemeral organizations with varying levels of institutionalism. As a result, the source base for an in-depth historical study of colonial Indian caste activism in the early twentieth century is quite poor. Even the most careful attempts to write a ‘history from below’ of Sanskritization would be dependent on the material collected by the colonial state, and would naturally focus on the interactions of these activists with the state or on post-independence material. The Census of India was a pioneering work of data gathering, but for that reason, much caution is required in interpreting the data (Barrier 1981). Among the most important limitations from our perspective was a non-transparent, arbitrary, and inconsistent set of policies for defining caste groups for the tabulation, failure to tabulate caste-level data for many caste-years, and inconsistent definitions of literacy and occupation across years. Even formatting the caste census data in a consistent manner across census years is a difficult problem (Conlon 1981). Chapter 4 explains these issues in greater detail, but it should be noted that an extensive set of robustness checks are implemented to ensure that these problems do not affect the reported results. Among many others, the results are robust to the exclusion of particular census years, all castes with substantial population fluctuations, castes affected by definitional changes, and census years with low data quality.7 6 7

Lacina’s (2017) account of language politics in India is a partial exception. A related issue is that individuals were free to move from caste to caste, as long as they could convince census-takers to record them under a different name. While the census bureaucracy made determined efforts to combat this practice (for example, Edye 1922: 151), there was a certain amount of individual migration from caste to caste (Cassan 2015). This movement is in accord

Introduction

25

The unit of analysis is the ‘census’ caste, which reflects a categorization by the census authorities of the various names that individuals gave to the census takers. This does not imply that these names exactly correspond to some underlying primordial social reality. However, though the census categories did not necessarily represent ‘real’ groups of people, they did represent categories of people to whom a label was applied. The question in this book is what use the people in these categories made of them. In some cases, they sought to challenge them and move into a different category, while in others they sought to reify and inhabit the category. The census categories should be understood more as potential ideological hooks than as real communities (though in some cases, they perhaps were both). Both the theory and empirics focus on activists. Petitioning and the organizational work that led to petitioning were the product of a small, educated minority, both within the groups involved and within Indian society as a whole. Millions of Indians experienced caste in different ways, often ways that were less intellectual and more closely related to patterns of violence and social domination. However, the actions of these activists are worth studying because they shaped subsequent political events. While names such as ‘Yadav’, ‘Adi-Dravid’, and ‘Bhumihar Brahmin’ began to be adopted among a small circle of caste activists in the colonial period, they have, since that time, become central parts of the identity of millions of people (Jaffrelot and Kumar 2012). Similarly, ideas about the relationship between these names, other groups, and the caste hierarchy that emerged in this period influenced the process of group consolidation and alliance formation (vertical, horizontal, and so on) that occurred after independence (Rudolph and Rudolph 1967). Precisely because they were the first people to think of caste using the new vocabulary of the colonial state, the choices that these elites made influenced all who came after them, just as the provision of missionary education in Africa influenced subsequent identity politics (Ranger 1984).

with the theory, in particular the idea that elites adopting ranked versions of caste identity would attempt to merge into castes above them. However, individual-level identity change would potentially substitute for the type of group-level identity change captured by the petitioning variable. However, the empirical analysis shows that these strategies tended to complement rather than substitute for each other.

26

From Hierarchy to Ethnicity

While petitions were filed by individuals, all the census data were collected at the group level. This means that the submitters of the petition were not necessarily representative of the group as a whole, and group-level data do not describe the distribution of traits within the group. This is an unavoidable problem, given the failure of the census to collect detailed information on the actual submitters. Fortunately, many of the theoretical predictions are unaffected by this problem, since they concern factors at the regional level. Even testing theories about the influence of education using group-level data requires the relatively modest assumption that petitioners in more educated castes are in general better educated than petitioners in less-educated castes.

The post-colonial world While the interest of colonial officials in caste divisions had important effects on the strategies of activists, the withdrawal of that interest had serious effects for scholars of caste in India. The Indian government immediately ceased disseminating (though not, in all cases, collecting) statistics at the jati level. Even the dissemination of information at the category level (especially the strength of the OBC category) remains politically controversial. Survey data collected by non-governmental bodies can also be problematic in this regard, with collectors even refusing to tackle the problem of categorizing the thousands of confused responses to jati questions, or refusing to make any data available beyond basic tabulations. For these reasons, the quantitative analysis of the changes in caste politics in the post-1947 period represents a difficult problem. However, several conclusions are possible from the limited data available. First, the political mobilization of jati identities has remained a very common electoral strategy in India, and has, if anything, intensified over time. Caste-based parties have become very popular in some regions, as has voting along caste lines, the use of caste in political rhetoric, and the strategic use of reservation policies as distributive tools. This mobilization has had measurable consequences for both descriptive representation and public policy. Newly educated groups used mobilization along caste lines to enter legislatures and party leadership in increased numbers, while both these groups and the more established groups have used their disproportionate political power to influence public policy in ways that favour their group, a point made strongly in Lee (2019). Second, this mobilization has not been even in nature. Many social groups, especially at the extremes of the social scale, do not vote as blocs,

Introduction

27

are represented (at least explicitly) by no parties, and are often outspoken in their denunciation of caste politics and caste-based distribution. The least educated groups remain under-represented in politics, despite belonging to social categories whose more mobilized jatis have become more powerful. Caste mobilization thus remains, in an important sense, incomplete. While many Indians condition their political participation along caste lines, others do not, and are critical of the entire idea of caste as a relevant social division. Finally, the twentieth century had seen a precipitous decline in ranking as a political ideology. Deference to upper caste groups, at least in the political sphere, has collapsed, and even sub-political behaviours associated with ranking, such as the practice of untouchability, are increasingly confined to rural areas. In fact, the rise of reservation policies has created a countervailing set of incentives: groups now have very good reason to try to appear as ‘backward’, the result of which has been a strange race to the bottom that parallels many aspects of the colonial experience, with groups (many of whom had claimed high-caste status two generations previously) submitting petitions and applying political pressure in a desperate attempt to be officially deemed disadvantaged. The spectacle of relatively wealthy groups such as Jats and Patels blocking roads to assert their own backwardness shows how dramatically the politics of caste have changed over the past century. While the object of the game may still be the acquisition of state resources, the way in which this goal has influenced social identity has undergone a complete transformation.

Plan of this book Chapter 2 develops the theoretical argument of the book. This involves describing both the existing theoretical debates on the origin of ethnic and caste identities, and developing two novel hypotheses about why some identities become salient. It then describes in depth the differences between ranked and unranked identities, and discusses where these differences come from. Chapter 3 discusses the historical background of caste identities in the colonial period. It describes social conditions in pre-modern India, with a particular focus on the two independent variables, state strength and economic growth, and explains how socially ambitious elite groups articulated hierarchical identities both in Hinduism and in other Indian

28

From Hierarchy to Ethnicity

religions. It then explains the changes in caste identities that occurred in the colonial period, giving several detailed examples of caste politicization. Chapter 4 describes the colonial Census of India, which is both a source of quantitative data for the project and one of the factors proposed in the literature as directly causing changes in the development of caste identities. This involves a detailed discussion of the relationship between the colonial census and caste, and examples of how Indians responded to the census. It then discusses the distribution of petitioning behaviour across groups, noting how it illustrates the non-linear pattern predicted in Chapter 2. For reasons of space, some of the detailed discussion of the census data and the extensive robustness checks of the main results are included in the appendices. Chapter 5 discusses the evidence for the second set of hypotheses, on the origins of ranking. After discussing how ranked rhetoric manifested itself in the colonial census data, it shows how this rhetoric diminished over time, paralleling the rise of political participation. To show how the argument extends outside India, it also discusses the development of racial identities in the United States and Brazil, showing how the rise of political participation in these countries has been associated with a decline in ranked rhetoric. Chapter 6 continues the story after Indian independence. It shows how the spread of literacy led to the mobilization of caste groups that had previously been unmobilized, and the full integration into the political process of groups that had previously been marginal. Similarly, the democratic character of the new state institutions led to a dramatic reduction in the use of ranked rhetoric and the presence of ranked norms in the political sphere. These trends were intensified by the emergence of policies that allowed the reservation of jobs and educational opportunities for members of specific caste groups, which created both additional incentives for caste-based mobilization and strong disincentives for groups to portray themselves as being of high status. Chapter 7 concludes with a discussion of the broader importance in the form political identities take, and how the historical prominence of uneven mobilization and ranked rhetoric has influenced contemporary India. It argues that India’s high levels of diversity, inter-group inequality, and uneven group mobilization are directly traceable to the uneven, ranked mobilization of the colonial period. It also shows how these processes relate to the larger historical processes that have affected both India and the developing world over the past two centuries.

2

Explaining Identity Activism What explains variation in the articulation of social identities? Why do people with many other potential uses of their time, spend it forming ethnic organizations, formulating elaborate theories of who is and is not a coethnic, and writing pseudo-scholarly histories of their group and its glorious past? Moreover, why do these organizations, theories, and histories differ so much from each other? This set of questions can be thought of as presenting two problems: First, do individuals engage in identity activism (rather than non-political activities or types of political involvement that do not engage with a specific identity)? Second, what form does it take? In particular, is it ranked or unranked? This chapter suggests solutions for both these problems. It proposes that, consistent with the existing political behaviour literature, the predictors of activism are economic: the existence of an educated class capable of political involvement and the opportunity cost of non-identity activism. The predictors of ranked activism, by contrast, are political : the type of state institutions in place, and the interactions of those institutions with the attributes of groups. The chapter begins with discussions of two key concepts: identity activism and social ranking. It then discusses why rational individuals might choose to use these concepts for their own advancement, and what factors would keep them from doing so.

What is identity activism? Activism as attempted mobilization A large literature in comparative politics has argued that individuals possess a wide ‘repertoire’ of traits, the exact content of which varies from society

30

From Hierarchy to Ethnicity

to society (Laitin 1985; Posner 2004; Chandra 2012). Common categories of traits include skin color, occupation, religion, language, tribe, and region of residence. At any given time, only a few of an individual’s traits are ‘relevant’ or ‘activated’ for the purposes of political decision-making. These are the traits that individuals consider as important when making decisions about who to vote for, whether to engage in political violence, and whether to enter into patron–client relationships. In India, for instance, caste, religion, and language identities have all been strongly politicized, while skin colour remains politically (though not socially) unimportant. Individuals, to say the least, receive a great deal of advice from outside political actors about which of these identities they should use in their political decision-making. Elites conduct elaborate campaigns to get individuals to emphasize particular identities, which may take the form of propaganda, the formation of social organizations, the selective distribution of resources, and even the selective use of political violence (R. Bates 1974; Horowitz 1985; Chandra 2004; Eifert, Miguel, and Posner 2010). However, it should be noted that not all mobilization efforts by elites are successful, and that some identities promoted by elites do not become widely influential. There are a variety of technologies available to activists to pursue these goals, and to reach people who possess the trait that the activists wish to emphasize, but for whom it is not particularly relevant. They may form associations that serve as a forum for building consciousness among members (Rudolph and Rudolph 1967; Banda, Kayira, and Robinson 2017). They may publish material about the importance of the identity, and attempt to circulate that literature (Gellner 1983; Anderson 1994), use the media or public sphere to spread messages sympathetic to their cause, and actually form parties to contest elections (Chandra 2004). The case of Hindu identity in India is instructive. While over 80 per cent of Indians practise some form of Hinduism, in the nineteenth century this identity was often subordinated to other identities (including, as we shall see, caste) in the political sphere (Freitag 1989). In the twentieth century, however, a group of activists sought to change this. They founded a series of organizations, notably the Rashtriya Swayamsevak Sangh (RSS), dedicated to encouraging ‘Hindutva’ (‘Hinduness’) among both members and Hindus more broadly (Hansen 1999; Jaffrelot 1999). These groups also carried out social service activities designed to emphasize the importance of Hindutva to people (such as the Scheduled Tribes) who had previously been considered marginal to the Hindu community (Thachil 2014).

Explaining Identity Activism

31

Finally, Hindu nationalist intellectuals produced a vast pseudo-scholarly and popular literature, reinterpreting Indian history in such a way as to emphasize the importance of the Hindu–Muslim cleavage (Hansen 1999).1 The effectiveness of such an identity campaign is vastly enhanced if it can manage to gain control of state institutions. An activist group with the Ministry of Education on their side can not just press its tracts on believers but also require that every schoolchild read them. Identities based around a language can force every child to learn the version that they prefer. For this reason, private groups are particularly important in areas, like in most of the colonial world, where identity activists are not in the position to use state institutions to spread their ideas. This is not to say, however, that colonial states were unconcerned with identity politics. On the contrary, the colonial period saw the reification of identities that had previously been non-existent or inchoate. In Benedict Anderson’s words, ‘Map and census thus shaped the grammar which would in due course make possible “Burma” and “Burmese,” “Indonesia” and “Indonesians.” But the concretization of these possibilities, concretizations which have a powerful life today, long after the colonial state has disappeared, owed much to the colonial state’s peculiar imagining of history and power’ (Anderson 1994: 189). Laitin (1986) provides a more concrete example of how colonial rule made relevant previously unimportant identity dimensions, a process referred to in other contexts as the ‘invention of tradition’ (Ranger 1997). Whether such invented traditions were either deliberate consequences of divide and rule policies or the unintended result of colonial policies of measurement, they had the effect of establishing an entirely new set of ‘concretizations’ for identity politics. However, such interventions remained peripheral to the more important effect of colonialism: the failure to cultivate national identities. During the nineteenth century, most independent states were working to promote national identities among their citizens (for example, E. Weber 1976; Darden and Grzymala-Busse 2013). Colonial states (with some reason) viewed local nationalism as a potential challenge to their own pan-national empires (Gellner 1983) and actively discouraged any such development. The development of national and even regional identities was thus on the shoulders of local elites (Hodgkin 1956; Chatterjee 1993; 1

Prerna Singh (2015) makes a similar argument regarding the rise of regional identity in certain southern Indian states.

32

From Hierarchy to Ethnicity

P. Singh 2015). While some post-colonial states have been able to use the apparatus of the state to promote national identities as an alternative to narrower ‘ethnic ones’ (Brown 2003; Miguel 2004), many others have failed to reshape identity politics in this way (Mamdani 1996).

Types of identity A central distinction in the ethnic politics literature is between dimensions of identity and particular identity categories that position individuals within those dimensions (Chandra 2012). In the United States, race is an identity dimension, while black and white are identity categories within it. In India, ‘caste’ is an identity dimension, within which there are thousands of individual castes or jatis. Identities may nest within each other, or cross-cut each other. In India, caste identities are frequently nested within religious and regional identities. Members of the Karan caste, for instance, are overwhelmingly also Oriya and Hindu. However, this is not always the case: the Jat caste, for instance, contains Hindu, Muslim, and Sikh individuals, and people from several northern states. The ethnic politics literature has tended to focus on the choice individuals face between dimensions, either cross-cutting or nested (Posner 2004; Dunning 2009; Chandra 2012). However, even within dimensions of traits there is considerable variation in the extent of politicization, often independent of group size. In modern India, some jati identities, such as that of the Chamars, are strongly politicized, while others, such as that of the Dhobis, are not. This book focuses on this ‘within-dimension’ variation, developing a theory of why some jati identities are predictive of political behaviour while others are not. Identity categories differ from each other in their potential size. Nations, for instance, usually include everyone within a given territory – within Denmark, ‘Dane’ is the only category within the ‘nation’ dimension, though not all category members identify with it equally (Shayo 2009). Some other identities are almost as comprehensive: in the Danish example, most Danes would be considered white and Christian (though they may not find these traits to be politically or socially salient). On the other hand, many categories are very small, with castes in India or tribes in African countries being the canonical examples. The competition between broad and narrow types of identity will be discussed further in the next section.

Explaining Identity Activism

33

Ranked versus unranked identities Identity assertion has often been accompanied by groups rejecting (or embracing) older norms of hierarchical inter-group relations. Since Horowitz’s (1985) discussion of the subject, political scientists have been aware in a general way that some ‘ranked’ identities are closely related to hierarchy in ways that other identities are not. However, since Horowitz’s discussion is brief, and my approach differs from his in several respects, it is worth defining ‘ranking’ in some depth.

What is ranking? The most obvious aspect of ranked identities is that ranked ascriptive identities correspond, at least in theory, to differences in social status, though not necessarily to differences in political and economic power. In Max Weber’s view: A ‘status’ segregation grown into a ‘caste’ differs in its structure from a mere ‘ethnic’ segregation: the caste structure transforms the horizontal and unconnected coexistences of ethnically segregated groups into a vertical social system of super- and subordination.... Ethnic coexistences condition a mutual repulsion and disdain but allow each ethnic community to consider its own honor as the highest one; the caste structure brings about a social subordination and an acknowledgment of ‘more honor’ in favor of the privileged caste and status groups. (M. Weber 1958: 185)

In practice, this means that in ideal-type ranked systems of identity, subaltern groups have no social elite, or a small dependent one. In ideal-type unranked systems, by contrast, each ascriptive group has an elite that considers ‘its own honor as the highest one’. While the elites of ranked systems are typically wealthier and more powerful than members of other groups, Weber’s formulation of ranking also requires the existence of entrenched, salient, widely accepted, ideas about the relative value of different social attributes. The sophistication of these ideas varies widely. At its most developed, ranking can be supported by an elaborate, religiously legitimated, and widely acknowledged set of status distinctions. At the other extreme, the social ranking may shade into a set of inchoate stereotypes about members of outgroups (as ‘lazy’, ‘dirty’, and so on) of the type that exist even in non-ranked identity systems (Verba, Ahmed, and Bhatt 1971).

34

From Hierarchy to Ethnicity

The relationship between ranked and unranked identities is thus a continuum rather than a sharp difference. In practice, of course, social status and ethnicity will rarely be perfectly aligned. By military skill, business acumen, or luck, individuals from lower status groups will gain wealth and power greater than what the ideology of the ranked system assigns to them. There are many examples of such divergence in ranked societies: India has many cases of impoverished Brahmins and wealthy peasants, and the racially divided societies of the eighteenth-century Americas had noticeable populations of both poor whites and rich blacks. If a society sees relatively little social change over time (as Weber claimed that India did), such anomalies will be rare. But in fact, both ranked social systems in general and in India in particular have produced numerous examples of low-status social groups whose members attained considerable power and wealth, even in the pre-colonial period (C. A. Bayly 1983; Kolff 1990; S. Bayly 1999). If such ‘anomalies’ are allowed to accumulate, the basis of ascriptive ranking might well collapse, as the nouveau-riche groups become restive at the gap between their high economic and political position and the low social status assigned to them in the ranked system, and adopt forms of identity that reject ranking. However, these elites can also decide to assimilate into the hierarchical system and change their position within it – choosing not to lead their old ascriptive category but to exit it. By behavioural changes, advantageous marriages, and political muscle-flexing, previously low-status elites can manage to ‘become’ high-status within a few generations (S. Bayly 1999). Such advancement reinforces the general correspondence of category and power rather than weakening it. The decisions of newly rich elites to emphasize social ranking are thus a key factor in ensuring ranking’s continued relevance. This discussion of identity ‘systems’ would seem to assume, as Weber and Dumont did, that there is unity within a society on the approach that these new elites take. In some cases, of course, it is reasonable to generalize in this way. Arguably, ancient India and twenty-first-century America fit this description, as ranked and unranked concepts of identity (respectively) were or are present among all major social groups. However, as we shall see, there are periods in which different groups make different sorts of ideological claims, and where ranked and unranked social theories coexist. In these periods, a holistic ranked ‘system’ thus reflects an ideological ideal rather than a practical reality. This book focuses on how groups choose ideological strategies within such mixed societies. As such, it focuses on the questions of

Explaining Identity Activism

35

what causes individual groups to participate in ranked systems that already exist, rather than the larger questions of why these systems emerge. The embrace of ranked and unranked identities should also be distinguished from participation in political institutions controlled by the elite (what Rudolph and Rudolph [1967] call ‘vertical mobilization’) and rejection of those institutions (‘horizontal mobilization’). In the Indian case, many lower caste politicians who publicly disdained ideologies of caste ranking have nevertheless been willing to participate in the upper-caste-led Congress Party, one of the Rudolphs’ key indices of vertical mobilization (Jaffrelot 2003). Similarly, while there is considerable variation in the degree to which party and ethnicity correspond (Huber and Suryanarayan 2016), it does not follow that members of multi-ethnic parties necessarily see themselves as subordinate to (or even ethnically similar to) their allies. Horowitz (1985: 22) conceptualizes the difference between ranked and unranked identities slightly differently, as lying in ‘the coincidence or non-coincidence of social class and ethnic origins’. Taken literally, this would mean that identity ranking is identical to inter-group economic inequality. However, this would be a dramatic simplification of the definition used in the previous literature on the issue, and that in other parts of Horowitz’s book. If ranking really were a matter of class, many wealthy minorities, such as the Chinese in Malaysia, Christians in Syria, and Iranians in the United States, would be considered among the highest ranked groups in their societies. For this reason, this version of the ranking concept will not be stressed here, though the relationship between class and ethnicity is an important theoretical issue in its own right (Huber and Suryanarayan 2016).

Category and distinction Ranked and unranked identities differ from each other not just in their relationship to social status but also in how they conceive of themselves. Political identity has two elements: what I call a theory of category (who possesses particular identities and who does not) and what I call a theory of distinction (which types of identities are normatively desirable and which are not). Theories of category sort human beings into groups, while theories of distinction sort groups into the desirable and the undesirable.2 While the 2

This language closely mirrors Abdelal et al.’s (2006) division between ‘constitutive norms’ (the rules that define group membership) and ‘relational comparisons’ of group value.

36

From Hierarchy to Ethnicity

term ‘theory’ would suggest a sophisticated intellectual apparatus, many of these theories tend to be inchoate attitudes or schemas rather than well-developed ideologies. While all projects of identity mobilization try to make identity politically important, they differ significantly on their strategic emphasis on these two elements. Unranked identities have well-developed theories of a category but weak theories of distinction. While they erect well-articulated boundaries around the group, they make only weak claims about the relative status of group members or the relationship of the group to other groups. The tendency of unranked mobilization is thus absolutist. Having taken a particular trait or identity as important, it attempts to put this trait or identity at the centre of political relations. Those that possess the mobilized attribute are considered to have something important in common. Those that do not possess the mobilized attribute are considered as the other, defined less by their own good or bad qualities than by their lack of membership in the ingroup. The distinction between ingroup and outgroup individuals, and the preference for the former, is thus fundamental for unranked identities. This distinction has become the basis for a long-running debate in psychology and behavioural economics on the causes and dynamics of ethnic prejudice (see, for instance, Allport 1954; Brewer 1999; Chen and Li 2009). Ranked mobilization is focused on creating a theory of distinction rather than a theory of category. It develops a theory of what traits (both ascriptive and non-ascriptive) are normatively good, and uses this to award social value to both coethnics and non-coethnics.3 In some instances, like the Indian caste system, this project also involves a theory of proper personal interrelationships, in which lower status individuals are expected to defer to higher status ones. While in unranked systems, differences in identification are concentrated at the ethnic boundary line, in hierarchical systems these differences are scattered throughout the system. While the differences between individual subgroups may be very small (and hence relatively permeable), the overall differences in identification from one extreme of the hierarchy to the other may be very large. By extension, while unranked systems offer only two types of political relationships (coethnic and non-coethnic), ranked systems offer an almost infinite menu of inter-group relationships based on varying degrees of superiority and inferiority. In colonial Latin America, for instance,

3

For the classic version of this argument, see Dumont (1980).

Explaining Identity Activism

37

a white person would treat a light-skinned African or Indian far better than a dark-skinned one. In the traditional Indian caste system as imagined by Dumont, a Brahmin would have a subtly different relationship to a Jat cultivator than to a Chamar leather worker, and be far more willing to enter a political or social relationship with the former, due to his membership in a ‘clean’ caste. The practical differences between the norms of ranked and unranked systems can be seen in Ahuja and Ostermann’s (2016) study of marriage partner choice in India. Scholars of ethnicity might expect individuals to prefer to marry within their own group, since those outside the group will be perceived as different, and there is indeed evidence for this type of behaviour in India (Banerjee et al. 2013). However, Ahuja and Ostermann also find evidence that members of all groups prefer (at least in theory) upper caste partners, prioritizing distinction over the difference. Similarly, many cultures with ranked systems have traditions of hypergamous marriage across group boundaries, between upwardly mobile members of the subordinate group and downwardly mobile members of the superordinate group.

Internal and intra-group relations One notable aspect of unranked identities is that they have a tendency to elide internal cultural differences within the mobilized group. This is the result of their fixation on the inter-group boundary – any attempt to further divide the group would make its external boundaries less relevant. In practice, this within-group fraternite takes the form of persuading poor group members to abandon local identities and behaviours in favour of the normative behaviours of the group elite. Examples might include Eugene Weber’s French peasants abandoning their provincial dialects for standard French, or attempts by Islamicist parties to discourage the practice of ‘heterodox’ forms of Islam. At the same time, elites must be persuaded to abandon attempts to link themselves to a more cosmopolitan elite identity and instead link themselves to the interests of their poorer coethnics. The result, in successful unranked identity projects like the nineteenth-century European nationalisms, is a world of rhetorically homogenous groups, marked by a single set of behaviours and attributes, which may come to be interpreted as a distinct ‘culture’ (Natrajan 2011). Besides assigning differential social value to other identity categories, ranked identities tend to subdivide and assign differential value to

38

From Hierarchy to Ethnicity

individuals within their own category. This is a natural consequence of assigning normative worth based on actual behaviour: Inevitably, some individuals will less closely approximate the behavioural ideal of the group than others. This leads to subtle but important divisions in hierarchical status, and thus in political identity, between members of the same ethnic category. As we shall see, one important element in the development of unranked caste identities in India is de-emphasis of previously salient intra-jati differences. In making unranked identity claims, groups may define themselves through their distinctiveness from other groups; they tend to view themselves as ‘conceptually autonomous’ (Fearon and Laitin 2000). This means that group identity, legitimation, and boundaries are based on a common history, culture, and set of myths rather than on their relationship to other groups, and that they are capable of existing as a social whole without either the deference of lower groups or the patronage of higher ones – they are ‘incipient whole societies’ (Horowitz 1985). In addition, conceptual autonomy implies that groups do not rely on myths or ideologies produced by other groups – at least in theory, their existence is legitimated by them themselves, rather than a group of outsiders. The search for conceptual autonomy means that the development of hierarchical identities is closely associated with the rejection of external ideologies and social hierarchies, such as cosmopolitan imperialism for Gellner’s Ruritanian nationalists and Sanskritic hierarchy for some Indian caste activists. By their nature, ranked identities require the existence of other groups to make them meaningful. The Brahmin’s superiority is established by the inferiority of the groups around him and would be meaningless in a society comprised entirely of Brahmins. The desire for ethnic homogeneity found among unranked identities is entirely alien to the hierarchical mindset, which accepts (and, indeed, welcomes) the existence of other groups as long as they maintain an inferior status. By the same token, ranked identities are incapable of developing the peer relations with other groups that are common among segmentary identities. Societies in which ranked identities are common are thus characterized by formal deference and resource transfers across groups, in contrast to the unstable and rivalry-prone relations of segmentary groups (Horowitz 1985: 29–32). One of the most notable differences between pre-modern identities and their successors was the tendency of these identities to be ranked. While premodern identity often seems fluid and ambiguous to us, this in fact

Explaining Identity Activism

39

reflects the primacy of distinction over category. Gellner (1983), for instance, describes a hypothetical transition from a world where a cosmopolitan ‘megalomanian’ elite rules over a poorly conceptualized ‘ruritanian’ peasantry, to one in which a small group of ruritanian intellectuals creates a ‘national’ high culture, pushing out the local elite. In the language of the theory, his transition corresponds to the increasing salience of theories of category (the hardening of the previously fuzzy megalomanian/ruritanian boundary) and a weakening of theories of distinction (the rise of the previously subordinate ruritanian identity).

The origins of identity activism Where do identities come from? The last two sections discussed sets of choices that potential activists can make. They can choose to emphasize an identity trait (or not) and choose to emphasize norms of ranking (or not). This ignored the question of where these identities come from. Why do activists have the option of talking about caste (as a dimension) or the Yadav caste (as an identity)? And if ranking emphasizes social distinctions, who decides what these distinctions are? While between-trait variation has often been neglected, the process by which identity dimensions and languages of identity emerge is the subject of extensive literature. The best work seems to suggest that caste identities were socially constructed through a historical process, often occurring during the spread of mass literacy in the late nineteenth and early twentieth centuries (Berlin 1972; Anderson 1994), though perhaps on the basis of identities that existed previously (Armstrong 1982). Rival ‘primordialist’ views of identities as innate are now widely discredited (Chandra 2012). The debate over constructivism within the social sciences overall is paralleled by the long-running debate within Indology over the origins of caste identities, although many participants have remained indifferent to the implications of the argument outside South Asia. One of the most influential perspectives in this debate, the structural theory of caste, can be traced back to Max Weber (1947) and Louis Dumont (1980), though similar views can be found in other authors (Bougle 1958; Leach 1969). According to these scholars, caste is the expression of a distinction between the pure and the impure that runs through Indian civilization and is inextricably entangled

40

From Hierarchy to Ethnicity

with the Hindu religion and South Asian culture. Since he saw caste as an ideological phenomenon, Dumont was dismissive of earlier perspectives that saw caste as a result of cooperative economic arrangements (Dubois 1806) or a relic of the racial divisions of ancient India (for example, Risley 1891). The superiority of the Brahmin, according to Dumont, came from tradition and cultural hegemony, rather than economic productivity or Aryan descent. In the larger perspective of the social sciences, Dumontian structuralism is a very extreme formulation of primordialism. Not content with declaring social groups the product of shared behaviours (which are, at least in theory, changeable), Dumont argues that they are expressions of cultural attitudes thousands of years old. Moreover, since these cultural attitudes are a reflection of the civilization as a whole, they are equally present in every group, and at every time period. While superficial changes in ‘form’ might occur, the more important ‘substance’ of the hierarchical order, such as the principle of purity, was an innate aspect of Hinduism, and thus inalterable. Dumont’s work generated a strong reaction (Srinivas 1966; S. Bayly 1999; Gupta 2000; Dirks 2002). These scholars argued that hierarchical, Sanskritic caste identities are very much a product of the modern era. While not denying the prevalence of hierarchical norms in pre-colonial society, they emphasize the plurality and mutability of these norms, and the degree to which ideas of status were shaped by the possession of political power. Paying particular attention to the activities of the census, they argue that colonialism played an important role in essentializing and reifying these fluid divisions, creating caste consciousness where none had existed before (Dirks 2002). In this view, caste as we see it is as much a product of a Foucaultian drive to state knowledge as a fixed part of Indian society. As Dirks put it: ‘Under colonialism, caste was thus made out to be far more – far more pervasive, far more totalizing, and far more uniform – than it had ever been before.’4 This echoes Laitin’s (1986) work on the role of the state in incentivizing identity change in Nigeria. The revisionists’ critique of Dumont’s views on the social ranking was even harsher than his views on identity. If Dumont thought that hierarchy was part of the Indian soul, these authors again emphasized the importance of colonialism, especially the near obsession of nineteenth-century Indianists 4

Srinivas (1966: 492) did not share this emphasis on colonialism, feeling that the urge to Sanskritize was inherent in Indian society, and that any changes in the colonial period were only quantitative in nature.

Explaining Identity Activism

41

and ethnographers with the ranking of identities (for example, Risley 1892; Thurston 1909). This was contrasted with the many documentable instances of status ambiguity and negotiation in the pre-colonial era (S. Bayly 1999). Srinivas (1966), focusing on the post-independence period, showed that outward adoption of the norms of upper caste behaviour (for which he coined the term ‘Sanskritization’) changed considerably among some groups over even very short periods of time. These authors are sometimes interpreted as saying that caste identities were created by colonialism, despite Dirks’s explicit denial that caste ‘was simply invented by the too clever British’. Instead, they hold that colonialism was successful in ‘systematizing India’s diverse forms of social identity’ (Dirks 2002: 5), a process similar to the creation of Anderson’s ‘totalizing classificatory grid’. The contributions of these scholars are numerous. The rigidity of caste identities is indeed more myth than reality, a point that has been forcefully made in recent years by a series of quantitative studies showing considerable shift in individual-level identity in India (Rao and Ban 2007; Cassan 2015). The colonial state was also closely involved in the collection of knowledge about caste, and many of those collecting this knowledge appeared to have taken a rigidly empirical view of caste divisions that glossed over the more ambiguous history of caste in pre-colonial societies (Dirks 1993). As we shall see, there is also considerable evidence that the activities of colonial officials provided a stimulus to caste mobilization, and that the identities that emerged in the twentieth century were substantially different from what came before. Note, however, that this perspective cannot explain variations in the spread of the new identities across regions and groups and the gradual decline of ranked rhetoric. ‘Colonial rationality’, at least in its ideological aspects, varied little across regions. The Census (the Indian institution most closely associated with the reification of caste) was one of the few colonial offices with a relatively uniform set of policies and procedures. While colonialism may have provided the political entrepreneurs of India with a set of ideological and institutional tools, it cannot explain the substantial differences in the alacrity with which they made use of them. As we shall see, activists often ignored concepts and traits that they found politically unhelpful, despite official encouragement to do so. They also often disagreed with the identity to which colonial ethnographers assigned them, choosing to affiliate with rival (though equally ‘official’) identities. The rest of this chapter explains how they made these choices.

42

From Hierarchy to Ethnicity

Barriers to political action Why, then, did some groups begin to talk about identity while others did not? A simple way to answer this question is that groups with no educated individuals will be unlikely to see activism. To see why, consider the correlates of political and social activism in general. A very consistent finding in the political behaviour literature is that education is associated with higher levels of political participation and social activism, as it provides the leisure for individuals to engage in political activities, the connections to enter political circles, and the skills necessary to publicize oneself or one’s cause to others. For these reasons, studies in a wide variety of social contexts have found that more educated individuals are more likely to participate in politics at all levels (Huntington 1969; Verba, Schlozman, and Brady 1995; Lee 2011), and that ethnic activism is common in the urban sector (R. Bates 1974). Some groups will have a higher concentration of such individuals than others, giving them a head start in any form of political action. Low levels of group education, in this view, serve as a barrier to political mobilization, because they mean that the elite that would lead it does not exist.5 There are even more direct reasons for why literacy should be associated with political participation. The activation of supralocal politicized identities in this period was closely associated with the rise of literacy (E. Weber 1976; Anderson 1990). There are two reasons for this association. First, most of the actions required in an activist campaign – the circulation of petitions, the publication of spurious histories, the founding of newspapers – required literate authors. The ability to read and write also lets elites discover external ideas about identity. Groups with poorly educated populations should thus have a much smaller elite capable of discovering ideas about ethnic identity, applying them to their own situation, and publicizing their cause. Second, literacy provides not just the leaders of an activist campaign but also the audience. Without writing, even the most energetic ideological campaign had little chance of influencing those beyond the immediate social 5

This perspective is similar to some classic modernization accounts of political conflict (for example, Huntington 1969). It differs from them in that it relates socioeconomic variables to the need to activate specific types of identities, rather than arguing that education increases instability by increasing the number of political players.

Explaining Identity Activism

43

circle of its leaders. With writing, such a campaign could influence people across large distances, and across subgroups that may previously have had little contact with one another. This gives us: Hypothesis 1: Higher levels of education should be associated with higher levels of identity mobilization. At the group-level, the level of literacy associated with mobilization is not fixed, since the relationship between the social traits of groups overall and the social traits of the individuals within these groups is imperfect. A generally illiterate caste may include a few highly educated individuals interested in political activism, while a generally literate caste may have not enough very highly educated individuals with sufficient interest. In practice, however, there is usually a high degree of correlation between group-level levels of literacy and the proportion of the group with high education.

Incentives to political action The discussion of the barriers to identity activism has temporarily ignored the question of why elites would want to engage in such activity in the first place. The ethnic politics literature has argued convincingly that the primary goal of individuals who engage in ethnic activism is to gain political power and the economic resources that political power brings (Laitin 1986; Posner 2004; Chandra 2012). This is the ‘instrumentalist’ view of political mobilization (Carlson 2015). This book will also assume that activists are primarily rent-seeking, though these rents can come both from holding political office or from being recognized as the leader of a defined group. Later, we will see that there are numerous examples throughout Indian history of individuals who gained power by posing as the representative of a caste group. Moreover, the role of self-interest in activism is emphasized in virtually all accounts of caste activism. Carrol (1978: 249) notes that ‘the rhetoric of “caste” activists and publicists was designed less to elevate the “caste” through some “social mobility” drive than to carve out for the publicists themselves a constituency on whose shoulders they could personally climb into positions of prestige and power’. Depending on the polity, power can be distributed through democratic elections, by informal contacts among the elite, or by some combination of the two. If identity politics plays no role, the rewards of office will be distributed based on ‘objective’ criteria, such as an individual’s wealth, charisma, intelligence, family position, or social network size. While the

44

From Hierarchy to Ethnicity

competition on such criteria advantages some, it disadvantages others. As Ernst Gellner argued, the disadvantages are particularly acute for the first generation of educated people from traditionally poor groups, who have a far less useful set of skills and contacts than those from more established groups. Such individuals have an obvious incentive to create other criteria on which to base political competition. Social identities fill this need for an ‘extra advantage’. If elites can successfully convince others that certain identity markers are important, they can gain an advantage in political competition, since individual voters and power brokers will tend to favour others who share identity traits with them. Crucially, an activated identity enables a would-be politician to appeal to individuals with whom he is not personally acquainted, and with whom he may have little in common. It thus provides a way for new elites to short-circuit the appeal of traditional elites. In order to do so, however, they must convince potential voters or clients to share their view of what constitutes a salient identity. There are several possible reasons why ordinary voters or clients might be more willing to support people with similar social identities, even if they are not going to gain political office themselves. If members of the same ethnicity have homogenous tastes in policy, individuals may use identity as a heuristic for common preferences (Alesina, Baqir, and Easterly 1999). Individuals could also have superior information about coethnics (Miguel and Gugerty 2005). If individuals tend to favour coethnics in making distributional decisions (perhaps because they believe them to be supporters), this could lead to a self-reinforcing cycle where members of the same ethnicity are likely to play a cooperative strategy, with each group favouring those who favour them (Habyarimana et al. 2007). While all these potential mechanisms are grounded in self-interest, individuals could also have a cognitive bias towards members of their group, which leads them to favour group members even when they have no material interest to do so (Tajfel 1974). The ordinary group member is thus far from a passive consumer of activist propaganda: she chooses to emphasize the identities elites are promoting because she believes that it will unlimitedly make her more likely to receive benefits from these elites. The group member can also choose to ignore activist propaganda or never hear of it, leading to the failure of the mobilization attempt, though in a group with sufficient literacy (H1) this is less likely.

Explaining Identity Activism

45

The potential benefits of identity politics to poor-group elites can be illustrated through the career of contemporary Indian lower caste politician Laloo Prasad Yadav, chief minister of Bihar (either personally or through his wife) from 1990 to 2005 and union railway minister from 2004 to 2009. Relative to the upper caste elites who dominated Bihar’s politics for much of the independence era, Laloo had many political disadvantages: he came from a poor family with no established network or large landholdings, and never earned a graduate degree. While fluent English has historically been a marker of social status in India, Laloo spoke an idiomatic and accented Hindi. However, within his own Yadav community, Laloo’s BA in veterinary medicine and student political achievements put him among the educated elite. By convincing Yadav voters that he represented their interests (through actions such as hiring many Yadavs for government positions), Laloo was able to create a ‘vote bank’ that gives him continued political influence, despite his unpopularity among many non-Yadavs (Witsoe 2013). To summarize, the development of a political identity can lead to the development of a mutually beneficial relationship between activists and ordinary groups members, with the activists providing benefits (material or symbolic) in return for support (material or symbolic). The benefits can include freedom from discrimination at the hands of police or public officials, or privileged access to employment (through formal quotas or less formal favouritism). Note that, while successful activists themselves might benefit from the policies that they enact, the other returns of being in the office are much greater. A politician from a disadvantaged group, for instance, might be able to see her children benefit from educational quotas for these groups, but this benefit would seem much less personally advantageous relative to (say) the ability to decide who obtains government contracts. However, this does not remove her incentive to provide goods to members of her own group. In fact, politicians frequently enact policies (such as anti-poverty policies or preferences in menial jobs) from which they are highly unlikely to benefit directly but which might motivate ordinary members of their groups to support them. How do activists overcome collective action problems? Since the elite activism discussed here is the work of a small group, there is no need to secure the involvement of the majority of the members of the category as a whole initially, though their eventual adoption of the identity is the end goal of the policy. However, collective action problems could potentially exist among potential activists, as other elite members could potentially free

46

From Hierarchy to Ethnicity

ride on the efforts of a handful of activists, consuming the political benefits of being part of a defined ethnic category without performing the difficult work of propagandization. However, these collective action problems are overcome by the fact that the small group actively involved gains a disproportionate share of the political benefits. By their agitation on behalf of a group, they become identified as the leaders of that group, and thus gain a disproportionate share of any benefits granted to that group. Others elite members anxious to take advantage of these benefits will be frustrated by the strong positions within group organizations held by the more established activists.

The well-educated For members of well-educated groups, the calculus was different. Relative to politically ambitious individuals from less-educated groups, politically ambitious individuals from the most educated groups are likely to possess rare social traits such as a professional qualification, a college degree, and fluency in English, at least if we accept that the distribution of these traits is correlated with the distribution of education. Note that it is theoretically possible for a group to be literate without having highly educated members, and vice versa. However, this appears to have been relatively rare in colonial India – at the 1931 census, group literacy and group literacy in English are correlated at 0.7 among province-castes. In the 2012 Indian Human Development Survey, group-level literacy was correlated with the group-level rate of secondary education at ρ = 0.78.6 These highly educated activists must choose whether or not to emphasize their identity in politics. While emphasizing identity will help win the support of those within their own category, it will make it less likely that those outside the category will support them. Consequently, these elites face a tradeoff: the certainty of leading a small group versus the lower probability of leading a large one. For poorly educated or connected politicians, this tradeoff is easily resolved: since they would be unlikely to be successful on a broad stage, they will be likely to choose to emphasize the smaller one. The highly educated, however, may be successful even in a broader context. To the extent that education and socioeconomic status make them politically appealing to voters outside their own group, these individuals 6

Reported caste names used. Minimum of 50 individuals in sample.

Explaining Identity Activism

47

have no incentive to engage in the costly mobilization of ascriptive traits that will only narrow their appeal.7 Instead, they will seek to play down these identities, and emphasize their own valence advantage, or emphasize an identity dimension that will place them in a category with many potential supporters, such as ‘nation’. This argument differs from Posner’s (2004, 2005) perspective on strategic identity formation. In Posner’s conception, individuals always want to be part of a minimal winning coalition of identity groups in order to receive benefits from their coethnics holding office. In this theory, by contrast, the incentive to be part of the winning coalition is modified by competition for leadership position within coalitions that favour the educated. This second competition gives poorly educated leaders incentives to mobilize identities that are politically suboptimal.8 Similarly, Shayo (2009) argues that individuals will always choose high-status identities, but does not address the countervailing incentive to being the most advantaged individual within a particular category. The types of choices made by such highly educated leaders can be seen in the career of Laloo Yadav’s predecessor as chief minister of Bihar, Jagannath Mishra (chief minister 1975–1977, 1980–1983, 1989–1990). Not only did Mishra come from a landed family and have a PhD (he was a professor before entering politics) but he also came from a political family with excellent connections in the central government: his older brother had been minister of the railways in the union government and a confidant of Prime Minister Indira Gandhi. Mishra was a Brahmin, but never emphasized his caste in his public pronouncements or founded a caste party, preferring (in the early part of his career) the Congress Party and a vague Indian nationalism, though his governments were dominated by the upper castes (Blair 1980). For men like Mishra, for whom connections and family status would bring success independent of the ties of caste, publicly emphasizing caste ran the risk of alienating many constituents. 7

8

In a democratic context, the most skilled politician will always gain more votes by not mobilizing a narrow ascriptive identity. It is possible (cf. Posner 2004) that a politician whose group is a majority of the population would prefer an ascriptive appeal in order to diminish the size of her coalition, already a winning one in a plurality setting. However, this condition does not apply to any Indian jati in any state. This assumes that the leader of even a small group receives rents from leadership even when his group is not in power.

48

From Hierarchy to Ethnicity

In the colonial period, there were several potential broad identities available to highly skilled individuals who wished to become involved in politics. They could either become involved in the colonial state itself (and in the imperial or loyalist ideologies associated with the colonial state) or become involved in the political challenge to the colonial state (and in the nationalist ideologies associated with these movements). Both these identities offered access to potential constituencies much larger than those associated with a single ethnic group. The relationship between education and identity politics was potentially self-reinforcing: the concentration of these highly skilled individuals in the colonial state and in the anti-colonial movement meant that the less skilled elites of lower status groups had all the more incentive to focus on narrow ethnic politics. The highly educated dominated broader identities and mainstream institutions, and the emergent elites sought to use narrower forms of identity politics to gain an opening. We should thus expect that higher group literacy should be associated with higher levels of identity mobilization, but very high levels of group socioeconomic status should be associated with lower levels of identity mobilization. However, the elites of the most educated groups will be heavily involved in forms of political action that involve broader identities. This additional observable implication of the theory suggests that very high levels of group education should be associated with participation in non-identity-based political participation. This gives us: Hypothesis 2: Very high levels of group socioeconomic status should be associated with participation in non-identity-based politics and low levels of participation in identity-based politics.

The origins of ranking Only once activists have chosen to engage in identity activism can they make a decision about whether to embrace social ranking. In addition, choices about ranking are trivial for a small set of groups at the top of the traditional status distribution, who are unambiguously advantaged by ranking. The key actors are thus the members of ‘low ranked’ groups who receive socioeconomic gains – either political, economic, or educational – that make them competitive with the traditional elite and give them the resources

Explaining Identity Activism

49

to engage in political action.9 When confronted by a ranked system that confined them to an inferior ritual position, these elites faced a variety of options. As Horowitz puts it: Theoretically, a ranked system can move in one of four directions. Subordinate groups can attempt to displace superordinate groups; they can aim at abolition of ethnic divisions altogether; they can attempt to raise their position in the ethnic hierarchy without denying the legitimacy of that hierarchy; or they can move the system from ranked to unranked. (Horowitz 1985: 34)

In colonial India, as in many other historical situations, the first two of these options were foreclosed. The economic and military dominance of the upper castes (not to mention the colonial state) made the violent revolution necessary to overthrow the ranked system a remote possibility. Similarly, the entrenched position of ascriptive divisions in the political strategies and self-images of high-status groups made the bilateral renunciation of these divisions similarly unlikely, despite the support of idealistic reformers such as Mahatma Gandhi. Realistically, therefore, low-status elites faced two options: climbing the hierarchical ladder or emphasizing more separate, unranked, conceptions of identity. This section describes how they might choose between these options. To the extent that existing work has hypothesized on the origins of social ranking, it has tended to identify ranking as characteristic of more traditional, less advanced, societies, and hypothesized that ranking will tend to fall apart over time, due to ‘“the spread of universalistic, egalitarian and achievement-oriented values”, by international contact, and education’ (Horowitz 1985: 32). This ‘simple modernization’ theory of ranking corresponds to the large literature emphasizing the links between industrialization, state building, and the development of modern national and subnational identities (Hechter 1971; E. Weber 1976; Gelner 1983). However, it is not necessarily the case that education and modernization always lead to the decline of ranking. As we shall see in the Indian case, it is perfectly possible that groups or individuals who benefit from the modernization process will use their gains to try to improve their status in 9

While these elites may have a formal or informal position of leadership within their ascriptive category, the theory does not require that this be so, and in fact many of the groupings that promote ascriptive identities are fairly small and unrepresentative.

50

From Hierarchy to Ethnicity

the traditional system, closing the status gap between themselves and the old elite without denying the existence of the hierarchy itself. If this view is correct, we should expect modernization, defined as access to education or improved social opportunities, to be correlated with ethnic activism in general, but have no relationship to whether this activism embraces ranking.

Patrimonialism Hypothesis 3: Groups linked to informal and patrimonial institutions should adopt ranked identities. Political power is distributed differently in different types of societies. In some societies, often referred to as patrimonial, power is distributed informally through networks of patronage and traditional obligation, with little concern for the needs and rights of ordinary individuals (M. Weber 1947; Bendix 1977). While this distinction was developed to explain the decline of European feudalism, the concept of the patrimonial society also fits the highly personalized politics of pre-colonial India (Cohn 1962; Fox 1971).10 In a patrimonial society, political power can be earned in two complementary ways: by gaining the favour of one’s superiors at higher levels of the political system, and by building up a support base of clients at lower levels. In Cohn’s case, the talukdar s (large landowners) of Benares needed at the same time to maintain good relations with both the raja of Benares and the village-level elites from whom they collected taxes. The raja in turn had to manage his talukdars while at the same time propitiating regional leaders like the nawab of Oudh. In neither case did the interests of citizens enter into political calculations: the stock-in-trade of a politically ambitious man was his contacts and reputation among other members of the elite. Patrimonial elites have a strong incentive to politicize identities that will ease their political relationships. An obvious response would be to adopt the same identity as one’s political superiors, thus benefiting from coethnic favouritism. However, this course is uncommon for two reasons. First, elites in patrimonial societies must think downwards as well as upwards, and the same identity that might be an advantage with one’s superiors 10

Indeed, some later Weberians have argued that pre-colonial India was a ‘purer’ example of patrimonialism than European feudalism. See Bendix (1977).

Explaining Identity Activism

51

might be a disadvantage with one’s inferiors. The raja of Benares, for instance, might gain a political advantage in his relations with the nawab by adopting (Shia) Islam, but such a move would create a major division between him and his (Hindu) talukdars. Second, patrons do not necessarily wish to be on equal terms with their clients. On the contrary, they may wish to use identity to emphasize their own legitimacy and the obligations owed them by their clients.11 Ranked identities fill this need. Instead of emphasizing groups’ divisions (which might allow some inferiors to become either too familiar or too distant), ranked mobilization emphasizes subtle distances in individual status. These differences define the patterns of obligation in patron–client relationships and enable an ambitious politician to be at once the humble supplicant and the gracious benefactor. In consequence, in patrimonial societies, a ‘high’ ranking often confers very tangible political advantages. High ranked individuals may be able to have physical access to a ruler denied to low-status individuals, as they did in Qing China (Rawski 1998), pre-colonial Tamil Nadu (Dirks 1993), and Bourbon France (Blanning 2003). High status might also become a requisite for certain types of political participation. Most pre-independence Indian village panchayat s, for instance, limited participation to members of ‘clean’ castes (Jeffrey 2001), while the ‘mean people’ of imperial China were denied the right to take the imperial examination (Brokaw 2014). Even when low-status individuals are able to interact on formally equal terms with high-status individuals, generally accepted differences in status may shape the outcome of the interaction. Bourdieu (1984), for instance, underlines the role of visible markers of social status in determining power in interpersonal interactions. Levinson and Brown (1979) note major linguistic differences in the interactions of lower caste and upper caste individuals, a finding that echoes the well-known difficulties of women in asserting themselves in spheres with pervasive gender stereotypes (Twenge 2001). Patrimonial institutions face powerful competition from the formal institutions of modern states. According to Weber, the extension of state authority leads to the decline of the political importance of hierarchical networks, which can now easily be circumvented: in Dirks’s (1993) phrase, the crown becomes ‘hollow’. Similarly, Templeman (1996: 83) links the rise 11

On such relations in contemporary rural India, see A. Chakravarti (2001) and Shah et al. (2006).

52

From Hierarchy to Ethnicity

of formal institutions to the decline of local caste panchayats: ‘The threat of appeals to government courts and officials, and of lawsuits against local associations, has been another factor reducing the effectiveness of these once powerful institutions.’ This echoes broader literatures on the formalization and depersonalization of authority that coincided with the growth of modern states (Scott 1998; Foucault 2012). As such formal institutions become more widespread, hierarchical ties become increasingly irrelevant. Like many Weberian concepts, ‘patrimonial society’ is an ideal type, and few societies fit it fully. In practice, most societies, particularly colonial ones, contain areas and groups that are primarily influenced by patrimonial institutions, while other areas and groups are more strongly influenced by the formal state. Within such mixed societies, we should expect groups whose elites have substantial patrimonial networks to find hierarchical mobilization relatively attractive, and the reverse to be true of groups whose elites are closely linked to the state and thus tend to have a more formalized relationship to political authority. Landed or weakly governed groups are more likely to direct their aspirations into winning a better position in the patrimonial system than groups with stronger links to the formal political economy. How can one improve one’s ranking in such a system? Two common strategies are the adoption of behaviours considered typical of high-status groups and individuals and the attempt to gain acceptance of the status change from external status legitimators. Over time, a combination of behaviour change and external legitimation will change the general social estimation. In the next chapter, we will discuss several examples of successful group status improvement in pre-colonial India, and the specific behaviours and group of legitimators that they used. However, similar patterns can be seen in other societies with ranked identities. For instance, in colonial Latin America many wealthy mestizos were able, given time and a suitable genealogy, to become accepted over time as white (McCaa et al. 1979). While individuals could seek to improve their hierarchical status as individuals, collective action is also common. Decisions about status-seeking made by individuals have implications for their relatives and coethnics. Status gains by individuals reflect well on people in their social or kinship network, and the gains of individuals are sometimes devalued by the low ritual position of those around them. In the colonial era, the development of new forms of information and publicity rendered these externalities

Explaining Identity Activism

53

more marked and thus made collective action for status improvement more common. Nearly all groups have something to gain from improving their ranking. However, members of the highest status groups have no reason to mobilize to gain hierarchical status, since they have it ex ante. They will concentrate on frustrating the attempts of other groups to gain a status that would devalue their social position. The behaviour of patrimonial elites has many similarities to the ‘neopatrimonial’ behaviour of modern power brokers in the developing world (Chandra 2004). They differ, however, in their relationships to formal institutions (Eisenstadt 1973). Neopatrimonial relationships are based on the distribution of state resources, and thus usually involve individuals with formal or informal relationships to the state or the party system. Patrimonial systems, by contrast, exist in deinstitutionalized spaces like eighteenth-century India, where individual power and status came almost entirely from their network. In such systems, the external legitimation of social status is obviously more important than in areas where power is more institutionalized. Put another way, while a rural power-broker in modern India still operates within a complex network of patronage ties, his power within these networks is determined by party or state titles as well as the network itself, and thus requires a smaller investment in ritual legitimation.

Participation Hypothesis 4: Within areas with participatory state institutions, large groups should be more likely to mobilize along unranked lines than small groups. Over the course of the nineteenth and twentieth centuries, patrimonial societies were gradually replaced by more institutionalized political systems. In such systems, one’s political position depends on one’s ability to gain the support of ordinary people (Bendix 1977). In many such societies, offices are apportioned democratically, and politicians must gain the support of voters. If politics becomes a competition for popular support, individuals from traditionally poor groups will be at a disadvantage relative to their more established peers due to their lower levels of money, contacts, and political skills. One way in which they can overcome these disadvantages is by attempting to politically activate social identities that they share with a large segment of the population. To the extent that these identity

54

From Hierarchy to Ethnicity

divisions are perceived as relevant, politicians from within the group will gain an advantage in competing with outsiders, either because constituents expect that coethnics will favour them in the distribution of resources or because voters gain some more abstract benefit from seeing a coethnic in power. Unranked mobilization thus gives politicians a tool for creating constituency groups to which they have privileged access. Ranked identities, which emphasize internal differences rather than internal solidarity, will be much less advantageous to new elites in such a setting, since they emphasize cultural differences between new elites and their prospective coethnic supporters. These incentives to unranked mobilization are shaped by the electoral system. Most electoral systems impose some sort of penalty on small groups,12 and in majoritarian systems this disadvantage may be very substantial. Since, in this theory, the ultimate goal of ethnic constituency-building is to win elections, it will not be equally attractive to all groups. Elites, being primarily motivated by the rewards of office, are unlikely to cultivate a constituency if it is so small that even a very successful mobilization effort would leave them in a weak position within the electorate as a whole (Posner 2004, 2005). The potential opportunities opened up by a more participatory politics are thus only attractive to groups that are large in size. Given that changing one’s ascriptive traits is difficult or prohibitively costly, the leaders of small groups will prefer to work along hierarchical or non-identity lines, since the creation of hard boundaries would divide them from most of their potential supporters. The leaders of a group with half a per cent of the population, in other words, may find the electoral rewards of ethnic mobilization so small that they will prefer to take their chances within a ranked system. Unranked identity development should thus be more common among larger groups in areas where the institutional context makes numbers politically valuable. Importantly, neither of these factors in isolation should influence mobilization strategy. In a non-democratic context, large subaltern groups are just as marginalized as small ones, while even in a democratic system very small groups have no realistic chance of political power.

12

While the relative sizes of groups may vary considerably over time, this chapter will assume that they are costly to shift, that a priori group size is thus still a predictor of group strategy.

Explaining Identity Activism

55

It is important to note that patrimonialism (H3) and participation (H4) are usually negatively correlated, since broad participation creates political alternatives to the traditional elite. However, this might not be so in every case: a group of traditional landlords might suddenly be exposed to elections, or a rationalized bureaucratic dictatorship might distribute power in ways that conform to neither archetype.

Conclusion This chapter suggested that the forms of activism we observe are the result of three nested decisions by activists: to become involved in politics or not, to become involved in identity politics rather than some other political project, and whether the resulting identity campaign should emphasize hierarchical values. This is summarized in Figure 1.1. The first two of these decisions is conditioned by education, specifically the presence of an elite (for the participation decision) and their outside options (for the identity politics decision). This produces the inverted U relationship predicted by Hypotheses 1 and 2. The rhetoric of these claims, however, is determined by the political situation: activists who wish to advance through patrimonial connections emphasize hierarchy, while those seeking to advance through participatory politics emphasize segmentary identities. These hypotheses represent an advance on current speculation on these questions, much of which has used relatively invariant variables like state policy or culture to explain complex patterns. Work that does focus on the group-level has tended to focus on group size rather than these, in some ways more basic, socioeconomic and political factors. In the next chapters, we will apply these ideas to a case where all these factors varied substantially: early modern India.

3

Caste in Historical Context Chapter 2 predicted that caste mobilization should be rare in societies where education is rare, and rise as education rises. Similarly, ranked rhetoric should be nearly universal in societies with patrimonial political systems, but fall as these systems become more participatory. This chapter will apply these ideas to the modern history of India, in two phases. First, it will show how hierarchical ideas were important in pre-colonial India, though often less rigid in form than they became later. This emphasis on ranking reflected the patrimonial structure of the political system. Second, it will show how identity articulation, while confined to a few relatively wealthy groups, was advanced in India relative to other parts of the world, reflecting the wealth enjoyed by some segments of the pre-colonial elite. Later sections describe how this process continued in the colonial era. It discussed the increase in education during the colonial period, and shows, as Hypothesis 1 would predict, that this increase in education was associated with an increase in caste activism. It will also discuss the role of the colonial state in promoting caste consciousness, though the discussion of the intervention often thought the most important, the colonial census, will be taken up in Chapter 4.

The background to pre-colonial identity politics Chapter 2 argued that the level of social hierarchy within a group or political system can be explained by two factors: education (which enables political involvement) and exposure to patrimonial political structures (which turns such involvement in a hierarchical direction). In this section, we will discuss the available evidence on the levels of these two independent variables in pre-colonial India. While pre-colonial South Asia was in general quite

Caste in Historical Context

57

poor, it offered substantial levels of economic returns to members of certain privileged groups. This relatively sophisticated economy contrasted with a political system that was quite unstable, and where politics centred around shifting and informal alliances among landholding and military elites. As we shall see, this combination of concentrated wealth amid political instability corresponded to a system in which the mass of the population remained quiescent while the elite energetically pursued hierarchical status. Compared to Europe, pre-colonial India was a poor region and appeared so to Western observers who arrived in the seventeenth century, who noted the cheapness of living, and the generally low level of consumption, particularly of protein (Braudel 1981: 187–226). These observations are backed up by modern statistical estimates, which show India to have had a per capita income about half that of Europe, and slightly below that of China, though about 25 per cent larger than that of sub-Saharan Africa (Maddison 2006). These figures reflected the reality of India’s ancien regime political economy, where subsistence agriculture remained the dominant mode of production and there was no sign of the improvements in agricultural techniques that had gradually increased European crop yields in the early modern period. There are several factors that distinguish seventeenth-century South Asia from the autonomous village economies of seventeenth-century Africa. The first was the prevalence of money, especially silver, in ordinary economic transactions. In many areas of the world, notably eastern and southern Europe, early modern tenants paid their rent and taxes in kind, using a portion of their crop. But in Mughal India, following Akbar’s revision of the revenue system in the late sixteenth century, most taxes were assessed in cash (Habib 1982: 236–240). Since all but the smallest landlords received silver from their sub-tenants, the polities of South Asia were able to operate on a cash basis, paying their officials annual incomes fixed in silver and ordinary mercenaries a daily wage (for Mughal examples, see Richards 1995: 66–75). To expedite the flow of specie (much of which was imported from Europe), there arose groups of bankers and traders, such as the Marwaris of eastern Rajasthan and the Jagat Seth family of Bengal, who prospered in long-distance trade, money transfer, and money lending (Raychaudhuri 1982: 325–359). This financial liquidity meant that wealth could be easily transferred between members of the elite, and accounts for the ease by which rulers such as the Mughals transferred resources among their supporters, and the ease with which the first generations of East India Company officials made easily portable fortunes.

58

From Hierarchy to Ethnicity

The second factor exhibiting the relative economic and cultural sophistication of the pre-colonial ancien regime was the existence of a sizable urban population. In 1500, Vijayanagar was the second-largest city in the world, with a population of half a million, and western India was dotted with small ports (S. Jha 2013). This urban sector drew its strength from South Asia’s resilient long-distance trade, both internal (carried on the backs of camels) and maritime (carried on Arab dhows and European ships). Cities served as points of transshipment and distribution and the headquarters of most Indian merchants. Some cities also became centres of industrial production, particularly of the cotton cloth that was South Asia’s leading export. These urban concentrations, and the trade on which they flourished, were important sources of revenue for states. In 1644, for instance, the Mughal emperor collected 250,000 rupees in revenue from merchants at Surat, and far more on the value of these goods as they were transported inland (Hasan 1993). The disposable income that was available to Indian rulers was thus quite large and could be used relatively flexible. The consequences of this financial power are obvious in the material culture of the period: elites poured out money on palaces, forts, jewellery, and clothing, making the courts of even minor Indian rulers deeply impressive to European travellers. However, a part of the revenue was available for the pursuit of ideological goals, such as the patronage of religious and scholarly elites, either through direct gifts of cash or the granting of tax-free inam land that provided rents in cash (for example, Dirks 1993: 324–357). Whether Muslim (and Persian-literate) or Brahmin (and Sanskrit-literate), the existence of these elites legitimated the ruler and provided him with a means to express his ideological goals. The relationship between these literate elites and the state would prove crucial for the development of political identity in India. Despite its relatively advanced economy, India boasted an unstable and poorly institutionalized set of polities. The instability began at the top and sprang from the relative openness of the Indo-Gangetic plain, a broad sweep of territory without internal boundaries to stop foreign invaders or strengthen regional state builders. North India was dominated by empires, and the ruler of these empires changed frequently as new external actors bid for power. The Delhi Sultanate went through five dynasties between 1206 and 1526, with each change of dynasty (and many changes of ruler) attended by war. The Mughal Empire managed (with several exceptions) to bring a kind of peace to the area between 1526 and 1707, but the subsequent collapse of their power made Hindustan once again an area of conflict

Caste in Historical Context

59

between rival internal and external actors, the Afghans, Marathas, and British. In the more mountainous south of the subcontinent, the geography favoured the growth of smaller polities, and the area did indeed produce several stable small states, notably Travancore and Mysore. However, the instability that characterized the north was also common in the south, often caused by the intervention of northern powers in southern affairs, as in the sustained campaigns in the Deccan under Aurangzeb and Mohammed binTughluq. Few dynasties, even the powerful Vijayanagar Empire, managed to last more than two centuries.1 While external threats bulk large in the historical record, these empires additionally faced the challenge of suppressing challenges from below, particularly from the landed kin groups who controlled land revenue collection at the village level. Not only did these groups possess the information and legitimacy needed to collect taxes locally but they also frequently possessed the military power to challenge the state themselves. Chris Bayly (1983) and Dirk Kolff (1990) portray the eighteenth-century Indian countryside as a world infested with armed men, where the division between banditry, tax collection, and rebellion was often unclear. Rulers had to control this reservoir of military power and political ambition at all costs, whether they took the form of military defiance or the mere forgetting of inconvenient tax obligations. It was the gradual withdrawal of support from these groups, as much as military defeat in the Deccan, which led to the collapse of Mughal power (Richards 1993: 282–297). The decline of such dynasties not only freed these upwardly mobile peasant groups to prosecute their own internal feuds but it also enabled a few of them to dream of upward mobility, and some of the new dynasties of the eighteenth century, such as the Jats and the Marathas, came from these militarized cultivating groups. The instability of these empires was matched by their low level of institutional development. Like many pre-modern sovereigns, the Mughal emperors and the Vijayanagar rajas did not possess bureaucracies outside their immediate secretariats. Military power was raised indirectly, by granting tax revenues (jagir s) in return for the promise to collect the taxes from the village zamindar s and appear on command with a specified number of horsemen. With revenue collection, the administration of justice, and the organization of troops all farmed out to private individuals, the Indian ruler was less the head of government than an entrepreneur of power, 1

On political instability in the Muslim world more generally, see Blaydes and Chaney (2012).

60

From Hierarchy to Ethnicity

using the threat of force by one portion of his followers to ensure the compliance of the other portions. Even with the exceptional resources of the Mughals, symbolized by their immense siege train, it was very difficult to keep these officials in line through the threat of force, without the long process of acculturation and surveillance that ensures compliance in modern bureaucracies. Rulers, for this reason, sought to create ideological ties between themselves and the ruled as a way of ensuring their behaviour in an environment with limited direct monitoring. Before passing to what these ideological ties looked like, it is worth comparing the polities of pre-colonial India to their contemporaries in early modern Europe. Relative to South Asia, Europe is impressive both in the stability of its political institutions and in their relatively high level of development. While Europe certainly suffered from wars, outside the central European core these wars were far less likely to destroy the polities completely. On the eve of the French revolution, France, Britain, Spain, and Portugal all were ruled by the distant descendents of the men who had ruled them in the eleventh century, and none had suffered any appreciable territorial loss during the intervening seven hundred years. The same record would be inconceivable in India: the Mughal emperor who ruled Delhi in 1789 was a descendent of the seventh family to rule the area since 1206 and ruled an area that barely extended beyond the city’s outskirts. Similarly, the states of Europe had gradually begun the process of building up rationalized bureaucracies, such as the French intendants and the British Admiralty Board (Brewer 1989). By a complementary process, these polities had also gradually reduced the private exercise of public functions such as justice, tax collection, and military leadership. Their autonomy usurped, the aristocracies of these states gradually evolved from seigneurs to courtiers, and from leaders of private warbands to officers in rationalized armies. While the European states of the eighteenth century were far from perfect, they thus could generally exercise some of their power through institutions. In India, rulers had to make far more use of the traditional skills of patrimonial politics: the manipulation of their inferiors through a judicious mixture of carrots and sticks.

Caste politics in pre-colonial India Ranked identities require that some groups be considered ‘better’ than others and that others seek to emulate their behaviour. In pre-colonial India, as in other places, the superior group ‘proved’ their superiority

Caste in Historical Context

61

by adopting behaviours considered religiously desirable, often drawn from Hindu beliefs. However, the process of acquiring status was intensely political, and involved rulers in projects of social credentialing that had little to do with religion.

A social climber’s guide to Hinduism All religions define some behaviours as normatively desirable or undesirable. In most religions, these include precepts such as respect for others, or prohibitions on killing or stealing. In addition, many religions also proscribe or prescribe behaviours with a less obvious connection to ethical beliefs, such as fasting on a certain day, making a pilgrimage to a certain place, or wearing specific types of clothing. By themselves, there is nothing about these requirements that is inherently hierarchical, and many of them were originally egalitarian in intent. However, over time the ability to perform many of these tasks has become associated with economic or political status. In medieval Europe, for instance, the Christian responsibility to give alms became an important facet of the aristocratic lifestyle, with noblemen employing professional almoners to distribute food and money with ostentatious lavishness. Similarly, due to the difficulty of undertaking the haj, the performance of this obligation has become a marker of status within some Muslim societies (Donnan 1989). Hinduism, being a religion of plurality, has many normative opinions on human action. Many of these relate to a discomfort with death and decay found in many religions.2 The handling of meat, excrement, or animal skin is thus considered degrading in many Asian cultures, though in an agricultural society many individuals will have to deal with such unclean materials on a daily basis. In India, groups specializing in tanning and scavenging have traditionally been of low status, while groups that are able to maintain a high level of cleanliness (ritual and physical) have usually been considered of high status. Another set of prescriptions relate to the devotional practices described in Hindu sacred texts, such as wearing of the sacred thread, full-time prayer and meditation in certain life stages, and the recitation of Sanskrit texts. Since widow chastity is seen as desirable, it is often considered prestigious 2

See, for instance, ‘Chinese Undertakers Trying to Overcome Prejudice about Their Profession,’ 7 June 2012, BBC News, available at https://www.bbc.com/ news/av/world-asia-china-18358999/chinese-undertakers-trying-to-overcomeprejudice-about-their-profession, accessed 30 September 2019.

62

From Hierarchy to Ethnicity

to maintain widowed in-laws, rather than having them remarry (Yeatts 1932: 340). While these behaviours could, in theory, be open to all, the time and literacy that they require also tend to confine their practice to the elite. A final aspect of normatively Hindu behaviour is giving protection, support, and respect to Hinduism’s priestly group, the Brahmins, a commitment that parallels the normative value attached to the support of priestly or monastic groups in Christianity and Buddhism. Brahmins (or Pandits) do not have a monopoly over Hindu religious expression: as numerous scholars have noted, attempts to conceive of a form of Hinduism without Brahmin hegemony are as old as the religion itself (for example, Gupta 2000). It is difficult to contest, however, that Brahmins have held a unique and strong influence over the Hindu tradition in all areas of the subcontinent, and that deference to Brahmins and Brahmin-sanctioned religious practice is deeply ingrained among a wide variety of South Asian social groups. This position has enabled Brahmins to establish themselves as arbiters of correct social behaviour for these groups. This arrangement has the potential to be mutually beneficial. Politically powerful groups, for instance, can gain prestige through the granting of money, land, or patronage to Brahmins, while even ordinary people can give Brahmins alms or hire them as priests. The status of Brahmins as legitimators-for-hire thus added an additional route for movement in the social hierarchy, since those anxious to improve their status must not only change their behaviour, but propitiate an external, and often mercenary, ascriptive group. There are numerous examples of this pattern. O’Malley (1912: 440) noted that ‘the methods pursued by the castes who desire to obtain a higher status follow a more or less stereotyped plan. One of the first steps is to obtain favorable vyavasthas or rulings from complaisant Pandits [Brahmins]’. Similarly, Nadar activists trying to prove their ‘clean’ status in colonial court at the turn of the century presented evidence that they maintained Brahmins to serve in their caste ceremonies and temples (Hardgrave 1969: 108).3

Caste and Hindu kingship Within Hinduism, therefore, there exists a long list of behaviours that could be used to separate social ins and outs, which range in everyday importance 3

The court disallowed the evidence on the grounds that the particular group of Brahmins used had a ‘reputation for avarice’.

Caste in Historical Context

63

from the simple expenditure of money involved in feeding Brahmins to the behavioural disruption involved in adopting strict vegetarianism. If these behaviours were merely intended to mark economic status, they could be acquired through the market, with prosperous groups paying complacent Brahmins to legitimize their gains. However, since religious status signalled political status in patrimonial politics, they were regulated by every pre-colonial Indian state. The goal of these rulers was two-fold: to monopolize socio-political status for themselves while at the same time providing rewards for their political followers. The first goal, particularly in south India, was often accomplished by ostentatious patronage and protection of temples. Some have held these gifts to be a necessary aspect of Hindu kingship, and that ‘without endowment, the king would cease to place himself in an active relationship with the redistributive powers of the deity, and will fail to acquire the honour constitutive of authority’ (Appadurai 1981: 71). Indeed, nearly every major Indian temple was established and endowed by a ruler, who in return was allowed a role in serving the deity and in administering the temple. However, even when the physical and institutional infrastructure of temples was less important, as it was in medieval north India, a ruler could acquire legitimacy by patronizing Brahmins as a class, through the distribution of alms, the patronization of literate Brahmins within the ruler’s court, or the granting of inam lands to Brahmin families. Among rulers of uncertain origin with land to distribute, like the eighteenth-century Shah kings of Nepal, this last mechanism reached enormous proportions, with about a third of the country being granted to Brahmins (Regmi 1976). Even in the more settled southern kingdom of Pudukkottai, Brahmin inam land constituted 11 per cent of the whole, in addition to the 23 per cent held by temples (Dirks 1993: 412). While such public charity brought the giver prestige, its more valuable side effect was control. If the Brahmins gained prestige as the groups most closely associated with sacred ritual, there were many other roles in the religious process available. As Appadurai and Breckenridge (1976: 198) note: Such a share [panku] would be composed of the right to offer service to the deity, either through endowment or through some prescribed religious function; the right to move the resources allocated for the specific ritual event, the right to command the relevant persons evolved in the actualization of the given ritual; the right to perform some single part of a complex ritual event; and, finally, the right to worship the deity, simply by witnessing the ritual.

64

From Hierarchy to Ethnicity

Such a variety of possible roles necessarily involved a set of rules as to who could perform the more desirable ones, and indeed temples and villages developed elaborate rules as to who would receive ritual honour at the festival (see the examples reproduced by Dirks 1993: 293–295). There are two notable features of such arrangements: First, they tend to place the ruler in a very prestigious ritual position, though the exact relative position of rulers and Brahmins varied locally. Second, such intricate lists were productive of disputes, and indeed temple access remains a highly politicized issue in south India. But while modern disputes are decided in the courts or by elected officials, in pre-colonial times such conflicts were adjudicated by the ruler. This enabled the ruler to use the ritual power of the Brahmins as a source of political patronage, rewarding loyalty with status while peripheralizing opponents. The ruler’s role as the source of honour extended well beyond the religious sphere. All rulers maintained courts, whose rituals and institutions, like those of contemporary Europe, were intended to enhance the prestige and centrality of the ruler. The ruler was also able to prioritize the status of other groups, rewarding loyalty while penalizing dissent. In pre-colonial Pudukkottai, the Kallar rulers granted the most prestigious posts at both the court and village level to members of their own caste, making careful distinctions among ‘major’ and ‘minor’ lineages. At the same time, they were careful to offer some prestigious positions, such as those in the royal bodyguard, to Akampatiyars, while excluding from office members of the Maravar caste (potential rivals because of their dominant position in neighbouring areas) (Dirks 1993: 243–268). Through such grants of symbolic favour, the raja was able to build the loyalty of lower level elites by enhancing their position relative to groups even lower in the hierarchy, like the occupational and labouring castes. Hindu rulers not only controlled the relative prestige of castes but also adjudicated disputes about caste membership. This was particularly important in the case of the ruler himself. Hindu scripture assigns the task of ruling and protecting Brahmins to members of the Kshatriya varna, and rulers were thus anxious to be considered Kshatriya. The most famous case of such pretensions was that of the seventeenth-century Maratha ruler Shivaji, whose caste was considered by some to be agricultural rather than aristocratic. Much of Shivaji’s ostentatious protection of Brahmins and promotion of Brahmins in court positions has been attributed to his desire to win legitimacy as a Kshatriya (Gordon 1993). The difficulty that Shivaji

Caste in Historical Context

65

had in having his claims recognized testifies to the potential power of such a classification. As a Kshatriya, Shivaji would have won an important measure of prestige in dealing with lower level elites in Maharashtra, while at the same time eclipsing the claims of the Mughal Empire. If the social structure of India had remained stable, as Weber and Dumont argued that it did, it would be difficult to disentangle the political aspects of these caste hierarchies from the cultural practices of particular groups. However, pre-colonial Indian history was not stable, as new groups gained economic and political power, and sought to translate these gains into a recognized social status that could solidify their political position. In the pre-colonial era, these groups could be divided into two broad categories: peasant groups that were able to rise through military skill and literate groups that were able to gain access to the growing bureaucracies of early modern Indian states. Several examples of such upwardly mobile social groups are discussed below, including the Jats and Bhumihars and the Bihari Kayasths. All these groups benefited from the unstable political situation in early modern northern India, gaining employment as tax farmers and mercenaries of the large empires (Muslim, Maratha, and British) contending for power in the region. In all these groups, the rise to social and economic power was accompanied by an increase in the adoption of the traditional norms of the caste system: ‘clean’ behaviours, the protection of Brahmins, wearing the sacred thread, and making claims to membership in the three higher varnas. While colonial accounts treated these groups as poseurs whose pretensions to high status were rejected by ‘native public opinion’, they, in fact, reflected the general pattern of eighteenth-century politics: if status gave power, power also, eventually, gave status.

Caste and Muslim kingship Pre-colonial Indian politics may seem an odd place to look for the origins of the caste system, given that for large portions of this period much of India was governed by Muslims, who came from a religious system where the normative behaviours of Hinduism were meaningless. From 1206 to the mid-eighteenth century, the north Indian plain was dominated by two Muslim empires, the Delhi Sultanate and the Mughal Empire, while peripheral areas such as Bengal, the Deccan, and Gujarat were either part of these empires or ruled by independent sultanates.

66

From Hierarchy to Ethnicity

The survival, and indeed flourishing, of caste under Muslim rule testifies to the political imperative to hierarchy in a patrimonial political system. Muslim rulers operated from a position of weakness like Hindu rulers, and faced the same need to win the allegiance of mercenary soldiers and armed revenue payers (often the same people). One common response of Muslim rulers was to simply adopt the functions of Hindu kingship, including the patronage of temples and Brahmins. In the city of Varanasi, for instance, the Mughals patronized local holy men and institutions, leaving a permanent mark on the physical space of the city (Desai 2007). In such situations, the Mughal ruler turned the Muslim elite into the functional equivalent of a dominant caste group, with the Central Asian aristocracy becoming the dominant subcaste. Many Hindu elites, such as the Rajputs of Rajasthan, eagerly sought the patronage of Muslim rulers, giving them daughters in marriage and accepting honourary titles in a manner very similar to their own relationships with lower level jagirdar s or zamindars (Richards 1993). In their construction of internal political hierarchies, Muslim rulers did not confine themselves to the adoption of Hindu cultural forms but also used elements of their own culture that were conducive to social hierarchy. In particular, the Persian-influence court culture of South Asian Muslim polities used elaborate rituals to emphasize the status of the ruler, the shadow of God on Earth, who commoners were forbidden to see in full face. By regulating access to the imperial person, the court not only distributed contacts but also graded the prestige of members of the elite. A similar end was served by the grant of honourary offices or jagir ranks, which often conferred neither responsibility nor income but were coveted for the status they conferred (Richards 1993: 66–68). Like their Hindu counterparts, the Muslim rulers also derived legitimacy from the public patronage of religion, in their case either the ulema (religious scholars) or the saints of the sufi orders, particularly the Chishti order (closely associated with the Delhi Sultanate) and the Naqshbandi order (closely associated with the Mughal Empire). The mosques of major cities and the tombs of prominent saints became vehicles for conspicuous government spending, which emphasized the spiritual prowess of the ruler. The identity politics of Muslim courts thus mirrored those of Hindu ones: lower level clients sought to use identity to be aligned with (though perhaps not formally equal to) their patrons. This meant, of course, that a political

Caste in Historical Context

67

realignment on a higher level could lead to identity change on a lower one. Guha (2003) tells one such story: Thus the Gond tribal raja of Deogarh in Central India converted to Islam and took the name Bakht Buland. In 1719 his grandson received a military appointment. He is described as ‘Sheikhzada by caste...’ But the Mughal empire collapsed, the Deogarh/Nagpur kingdom was taken over by the Marathas, and the family found it expedient to return to regional alliances and ethnic roots. So, in the mid nineteenth century the lineage maintained hypergamous marriage links with ‘Hindu’ Rajgond chiefdoms and was regarded as ‘pure Rajgond.’

Islam in South Asia extended far outside the ruling elite. For many poor South Asians, conversion to Islam offered an opportunity to improve their social and political position. By creating a religious bond with their new rulers, members of previously marginalized groups were able to gain a higher status than they would have previously had with their Hindu rulers, though this by no means implied that they would become the ruler’s equals. The doctrinal simplicity of Islam, and the relative weakness of its priestly group, made conversion less costly in many cases than a determined attempt to gain status within the existing system. For this reason, conversion to Islam (as opposed to immigration) was concentrated in areas of Muslim rule and among groups of relatively low or ambiguous status (Hardy 1979). Similar processes influenced conversion to Sikhism and Christianity, the other major South Asian religions, both of which were associated with non-Hindu rule and upwardly mobile lower and middle caste groups. For Jat peasants or Pariyan sweepers, the growth of a regime led by rulers of another religion provided an opportunity to at once shed their low status within Hinduism and gain formal religious equality with their rulers. In the case of Sikhism, this meant the articulation of an ideology which legitimized rule by Jats and disclaimed Brahmin or Kshatriya superiority, while Christianity promised converts privileged access to the European public sphere. In neither case did conversion challenge the nature of caste, which remained keenly felt by individuals and often recognized in the structure of religious institutions, but it enabled these castes to behave differently relative both to their rulers and to other groups (Puri 2003). Non-Hindu regimes thus failed to challenge the logic of hierarchical identities, but rather added alternative ways of articulating social hierarchy, and potential competitors for Brahmins as arbiters of legitimate social behaviour.

68

From Hierarchy to Ethnicity

The size of groups The emphasis on social ranking in pre-colonial India created a set of ideas about the superiority and inferiority of social groups. However, it also affected the structure of the groups themselves. Relative to the linguistic and religious groupings that form the basis of ascriptive division in many areas of the world, jatis are both small in size and closely identified with specific occupations, leading to both a high overall level of diversity and a high coincidence between jati and social status. These traits correspond to the needs of individuals in a ranked identity system. A ranked ascriptive identity must correspond in some degree to the adoption of the ideal behaviours of the system. As we have seen, these behaviours are potentially numerous, and the gradations in status among individuals are thus numerous as well. To capture these distinctions with any accuracy, identity categories must thus be both relatively small (since the differences in status within any large group will be so great as to make the group name meaningless as a label) and closely correlated with economic position (since economic position will have a substantial effect on behaviour). Neither language nor religion satisfies either of these conditions, since a territorial linguistic group or a large religion will invariably contain such a wide variety of social types as to be meaningless as a clue to political status. For this reason, the emergence of economic differences between individuals could easily lead to the creation of new groups. Francis (1902: 196) describes this process: In many castes, for example, sub-divisions may be found which are taking to the Brahminical customs of infant marriage, vegetarianism, and so on, while the main body of the caste adheres to its original Observances. . . . Sometimes these changes are sufficiently persistent to result in the gradual evolution of entirely new castes. The Jatapus afford an example. They were originally Khonds, but they have now given up eating beef, and taken to infant marriage and the worship of the Hindu gods, and are practically a distinct caste.

Even jatis, with their relatively small size, were often too large to capture the complex status distinctions of the hierarchical system as it existed in the seventeenth and eighteenth centuries. As we will see in the case of the Jats, this led some groups to focus on the subcaste or lineage as the centre of hierarchical status-building. In other groups, it might lead

Caste in Historical Context

69

to a split within a caste. The Chamar caste, for instance, split when some of its members abandoned tanning for weaving (Roy 1994). Similarly, new occupations could produce new caste-like groups. Certain units of the British Indian Army, for instance, became endogamous groups that held themselves aloof from their former castes (Farwell 1989: 172). Similarly, the Mahishya caste of Bengal is often thought to be based on those members of the Kaibartta caste who abandoned fishing for agriculture (Gait 1901: 441). Such splits reinforced the overall logic of the system by reinforcing the extent to which jati identity provided a meaningful clue to an individual’s hierarchical position.

The background to colonial identity politics Social and economic conditions In pre-colonial India, levels of literacy were generally quite low. In these circumstances, attempts to emphasize or change the relative status of caste identities were evolutionary and concentrated within the elite. In the post-independence period, the rapid growth of literacy even among the very poor and an even more rapid development of mass political awareness has led to a large-scale mobilization. The colonial period contained neither of these extremes, but was characterized by moderate shifts in education, unevenly distributed throughout the country. Like many developing societies, colonial India saw a modest secular trend towards increasing literacy rates, though from a very low initial level. Some of the causes of this trend included a growing economy, the beginnings of a state-supported school system, urbanization, missionary activity, the increasing cultivation of cash crops (which led to windfall profits for certain producers), and the development of a vernacular language press, which was often associated with a simplification of the written language. This trend towards increased education concealed considerable internal variation, across both jatis and regions. It is beyond the scope of this book to review the causes of this variation in detail, especially given that many of them are closely correlated. In general, gains in education were the product of both a group’s traditional position and the degree to which colonization impacted that position. Land-cultivating groups tended to do well under the permissive land revenue settlements of the latter Raj, even if they had not earlier been well educated, or part of the higher aristocracy.

70

From Hierarchy to Ethnicity Figure 3.1 Average male literacy rate by year in India, 1901–1931

Source: Author’s calculations from Census of India, various years.

Similarly, groups traditionally associated with trading or government did well as those sectors expanded. Other groups, particularly those associated with declining traditional crafts like weaving, saw little or no improvement in their educational position. Thanks to the meticulous record-keeping of the colonial government, we have much better data on the socioeconomic status of caste groups in the colonial period than we do for any other period of Indian history. Some contemporaries noted that this increase in education might be associated with changes in caste activism, and that the newly educated elites of previously poor castes might be particularly prone to such activism. Indeed, Srinivas (1956: 492) briefly noted the importance of colonial social changes for Sanskritization: ‘Groups which in the pre-British days had had no chance of aspiring to anything more than a bare subsistence came by opportunities for making money, and having made money, they wanted to stake a claim for higher status.’ In many cases, colonial policies changed the relative economic positions of groups, creating new elites, and weakening the economic positions of old ones. This effect was a natural product of the transition to the more institutionalized colonial state, which involved selecting personnel for the new bureaucracy and military. In many cases these individuals came from similar backgrounds to traditional elites, but in other cases they did not. In the case of the colonial military, policies shifted over time, as late-nineteenth-century recruiting officers increasingly disfavoured lower

Caste in Historical Context

71

caste individuals and Hindi speakers in favour of the so-called martial races of the Punjab, Afghanistan, and Nepal (Farwell 1989). Similarly, colonial tax policies altered existing patterns of power and wealth. The colonial land revenue settlements often altered traditional practices and property rights, either concentrating property and taxation rights in large landowners (often termed the zamindari system) or in ordinary cultivators (often termed the ryotwari system).4 Even within land tenure systems, colonial tax policy had significant social effects, with high tax assessments in the early nineteenth century often leading to non-payment, forced land sales, and the displacement of traditional elites (Metcalf 1979). Lee (2017) shows that colonial administrators were only likely to transfer formal or informal power to pre-colonial elite groups if they were secure militarily. This dynamic was based on a weighing of relative costs: while the local elite could provide a cheap and experienced group of administrators to the colonial state, they were also its most threatening potential military rivals, and the most likely collaborators of European rivals. As a consequence, in areas annexed in times of war in Europe, pre-colonial elites have low levels of wealth today relative to other groups, while in areas annexed in times of peace in Europe pre-colonial elites retain a more substantial economic advantage. Similarly, in indirectly ruled areas the caste of the ruler tends to be wealthier than other groups.

(Slowly) declining patrimonialism and (slowly) increasing participation At the same time India was changing economically, it was also changing politically. The first of these was the decline of traditional political authorities, particularly the larger ones. The courts and petty principalities of India were either completely replaced by the colonial bureaucracy or marginalized and ‘hollowed’ (Dirks 1993). Meanwhile, the late colonial era saw small and hesitant steps towards the introduction of democratic government, particularly at the local level. While the British conquest of India did not alter every aspect of the pre-colonial political system, the changes were very real. Many of the most 4

These systems are ideal types, concealing considerable internal variation. See Iversen et al. (2013) and Baden-Powell (1892).

72

From Hierarchy to Ethnicity

powerful pre-colonial courts – Murishibad, Delhi, Nagpur, Pune, Arcot, Lahore – disappeared completely, ending their ability to act as arbiters of group and individuals status and ending any incentives Indians may have had to curry favour with them. Moreover, within the context of the Pax Britannia there was no longer any possibility of establishing new polities, as the Marathas and Jats had in the instability of the eighteenth century. Even princes and zamindars who were able to retain comfortable positions within the new regime saw their power diminished. While many zamindars, especially in the permanently settled areas of Bengal and Bihar, saw their economic position improve under colonialism, they gradually lost their armies and judicial powers (Metcalf 1979). Many zamindars responded to these changes by moving to the big cities and living as rentiers, reducing their ability to act as local power brokers. While the princes with subsidiary treaties retained some measure of autonomy, even they had to give up significant autonomy to the British Resident and administration (Ramusack 2004). The administration that replaced the native rulers was the colonial bureaucracy, a bureaucracy that over time came to resemble the rationalized bureaucracies of European nations. The recruitment of British officers for the highest posts in the army and civil service meant that supreme political authority was held by individuals outside traditional status orderings, and often with little understanding of them. Even lower bureaucratic posts were allocated to those with European educations, through competitive procedures (Bose 1993). A young man eager for political advancement would thus find traditional markers of social status increasingly less useful as the colonial era progressed. Later chapters will present quantitative evidence that this decline in patrimonial politics was associated with both an increase in caste activism and a rejection of hierarchical norms. However, contemporary officials also noticed this association. In particular, they noted that the princely states, where traditionally high-status groups retained at least the forms of the patrimonial system, were much less affected by caste activism. Cole (1932: 123) waspishly noted that the desire for caste self-advancement ‘is perhaps not so prevalent in the Indian states as in British India, for in the former any undue precocity for social recognition outside and accepted sphere would be severely dealt with as it was in England up to the end of the 19th century.’ Meanwhile, the colonial government began to allow democratically elected legislatures an increasing level of influence. This trend began at

Caste in Historical Context

73

the local level and worked its way to the centre. An elected element was added to district boards in an 1882 reform, with the proportion of elected members being increased over time (Dodwell 1929: 520). The Morley–Minto Reforms of 1909 added elected members to provincial legislative councils, and their proportion and role were gradually expanded until in 1935 they were given ‘responsible government’, with control of most provincial functions. None of these institutions should be understood as democratic in a modern sense. In the 1935 election, only 12 per cent of the population could vote, and in previous elections the figure had been only 2.7 per cent (Chiriyankandath 1992). However, even these figures were dramatic relative to the modest number of individuals with any political role under the Raj before this. While peasants and labourers were still excluded, among the new urban middle class that was so central to caste activism these elections created a completely novel type of political contestation. These elections were keenly contested, despite the limited franchise. In most local councils, there were no parties and little attempt at programmatic positioning: elections depended on caste affiliations, and the distributional policies of councils varied with the caste composition of the membership (Chaudhary 2009; Suryanarayan 2016). Caste associations also ‘targeted’ district board elections, striving to elect members of their own castes (Rudolph and Rudolph 1960: 16–17). The link between caste associations and voting will be discussed in greater depth below.

Colonial caste activism Caste associations One of the most important aspects of caste activism in colonial India was the formation of caste organizations – the caste sabhas. These sabhas were not the first associations dedicated to regulating the details of caste identity. Earlier, caste brotherhoods or panchayats had existed in most villages and towns, and under the patrimonial administration of early colonial India their political role had been substantial. In villages where a single caste owned most of the land, the caste panchayat formed the local government (Indian Police Commission 1905; Lewis and Barnouw 1958; Srinivas 1960). All took on charitable and dispute-resolution functions within the caste, and often outside it as well. Nadar panchayats, for instance, were resorted to by members more commonly than the ordinary courts system, and

74

From Hierarchy to Ethnicity

could sometimes impose their jurisdiction on other castes of lower status (Templeman 1996: 83). This type of caste organization persists to this day in some areas, such as the khap panchayats among the Jats of Haryana (Pradhan 1966; Kaur 2010). The new sabhas differed from the old associations in their membership, which was generally educated and urban, and in their broader geographical spread and higher levels of institutionalization. Often, the formation of these sabhas involved an explicit rejection of the former practice. Chris Bayly (1971: 306) notes that ‘these associations sometimes sought to coerce the parochial brotherhoods to adopt measures of “westernization” or to arrange for the more equable distribution of patronage such as marriage funds and educational charities’. Turner (1933: 551) claims that [sabhas] are essentially different from panchayats. Whereas the latter are ancient and indigenous institutions of Hindu society dealing each with only one caste or even sub-caste, with a very limited jurisdiction, usually the village, and concerning itself with specific breaches of caste rules and conventions on which it adjudicates and imposes punishments when guilt is established, a sabha or mahasabha is essentially a modern product, the result of Western concepts of associations, societies and ‘corporations aggregate.’

The sabhas were much more solidly organized than the local caste panchayats that they superseded. Their constitutions, still preserved in the National Archives of India, record their often highly elaborated structure, with procedures for the election of officers, the holding of annual sessions, the handling of funds, and the affiliation of local groups. Blunt (1912: 346–347) described one such organization based on his experiences as census superintendent: The Gaur Maha Sabha is a central body of 100 elected members whose term lasts a year. The president is elected by these 100 and his term is a year. Its jurisdiction extends nominally to the whole of India but has little influence where there are no local sabhas. It endeavors to reduce expenditure at births, marriages, and funerals, to inculcate the true principals of religions and the observances of religious rites, and to encourage the study of Sanskrit.

Arnold, Jeffrey, and Manor (1976: 356–357) describe the operations of the Sri Narayana Dharma Paripalana Yogam, the association of the Irava caste of Kerala and Tamil Nadu:

Caste in Historical Context

75

By 1914 the Yogam had 1,200 paying members and assets of Rs 1 lakh. Its orientation was chiefly religious: it maintained 59 temples and strove to abolish expensive and crude ceremonies and to propagate a simplified, monogamous marriage. From 1904 the association had its own Malayalam monthly, Vivekodayam, and it had close connections with Kerala Kaumudi, an Irava weekly founded in 1921. By the 1930s Kerala Kaumudi had a circulation of about 4,000.

As these examples make clear, many activities of the sabhas were cultural and social rather than directly political. Some activities would come under the heading of social service and self-help: raising money for scholarships, trying to discourage expensive weddings, and providing a network to meet prospective spouses. The Rawani (Chandravanshiya Kshatriya) Sabha included a fairly typical list of goals in their constitution: -Rendering pecuniary assistance to promising young Chandravanshiya Kshatriya students to enable them to prosecute their studies. -Publication of a Patrika [journal] of their own or any book or pamphlet containing subjects of interest to the community or likely to promote their interests. -Appointment of itinerant Upadeshaks to popularize the work of the Mahasabha.5

All of these functions, however, had implications for group boundaries and group self-perception. Compiling membership lists, for instance, had the indirect effect of defining who was a proper group member and who was not. Organizing banquets and weddings had the side effect of reducing previously salient boundaries among subgroups within the caste, many of whom would previously have hesitated to dine with each other, let alone form marriage ties. The reduction in the cost of weddings also tended to reduce the intra-group social difference, since costly weddings (and large dowries) were a traditional way of the demonstrating the social superiority of the wealthiest caste members (V. Rao 2001). In many cases, the literatures and activities of the sabhas promoted the adoption of behaviour patterns (and, as we will see later, caste names) previously typical of the upper castes. The sacred thread, or upanaya, was widely promoted by caste sabhas, as were vegetarianism, child marriage, and 5

National Archives of India, Home Department, Public Branch, F. 45/6/31-Pub.

76

From Hierarchy to Ethnicity

bans on widow remarriage. All these were the types of normative Hindu behaviour that had characterized upwardly mobile pre-colonial groups. The list of goals of the Gaur Mahasabha’s constitution reflect the influence of such ranked norms: (a) To promote the physical, intellectual, moral, social and material welfare of the Gaur Brahman community. (b) To bring about unity and organization amongst Gaur Brahmans and thereby endeavor to lead them to act up to their duty. (c) To encourage good practices by removing evil customs and to spread education, especially the knowledge of Hindi and Sanskrit amongst Gaur Brahmans. (Quoted in Turner 1933: 552)

Not surprisingly, this type of behaviour was disdained by traditionalists, and by colonial observers. William Francis ICS (1902: 131) petulantly commented that ‘a caste does not enhance its real position by wearing threading, marrying its children as infants, and giving itself a high-sounding name. It can obtain far more honourable distinction by educating its members and elevating their lives’.

Who were the caste activists? The leadership of the sabhas was generally composed of the well-off, urban, and educated stratum within the caste. The surviving membership list for one organization, the Pradham Bhumihar Sabha, gives us some sense of the profiles of activists in a relatively well-off caste. Of the ten officers, four were zamindars (large landlords), two were lawyers, and one was a college president. Of the fifty-three ordinary members, ten were zamindars, seven were lawyers, and one was a senior civil servant.6 The Gaur Mahasabha required ordinary members to pay an annual subscription of a quarter of a rupee, but members of the Representative Committee had to pay six rupees, about four times the daily wage of a skilled urban worker (Turner 1933; Roy 2005). Most scholarly accounts have also noted the elite nature of these organizations. Arnold, Jeffrey, and Manor (1976) and Washbrook (1976) also emphasize the involvement of the wealthy in south Indian caste sabhas, 6

National Archives of India, Home Department, Census Branch, A Progs, December 1900, Nos. 24–27.

Caste in Historical Context

77

even among quite poor castes. Chris Bayly (1971: 306), notes that ‘in these organizations professional men who could make use of their literacy and province-wide professional contacts came to play a more significant role’ than they had in traditional panchayats. In the first years of the twentieth century, the wealthy members of the sabhas often concentrated on using them to advance their own social status and political power, while paying little attention, even rhetorical attention, to the majority of caste members. Such efforts at elite self-promotion and network building resembled the efforts of pre-colonial social climbers more than modern ethnic activism. Arnold, Jeffrey, and Manor (1976: 363–364) describe the problems of caste associations in Mysore before 1920: . . . the gulf separating the urban notables for the rural bulk of the Vokkaligas and Lingayats placed enormous difficulties in the way of the associations. They always had trouble in raising funds, and both failed to develop popular support in the state. The membership of either association never exceeded a few hundred of the mainly urban elite of wealth and education. Most members came from Bangalore and its environs, and the distribution of scholarships favored some parts of the state over others. A major reason for the associations’ failure in fund-raising and enrollment was the preoccupation of their leadership with public posturing and the bestowal of whatever patronage they could wean from the authorities on their personal networks of clients.

Starting in the 1920s, however, many sabhas began efforts to engage the whole of their potential caste constituency. Washbrook (1976: 151) notes that many castes pursued this turn towards ‘horizontal’ caste connections. Arnold, Jeffrey, and Manor’s (1976) discussion of Tamil Iravas and Lele’s (1989) discussion of Marahrastrian Marathas provide other examples of this trend. In all cases, this involved an increase in the organizational footprint of the caste and an attempt to remove previously salient sub-caste barriers.

Caste associations in politics These changes in the ideology of some sabhas occurred contemporaneously with the expansion of their political role. In the 1920s, especially in the south, caste sabhas formed the principal contestants in local elections (Gould 1987) and were the protagonists of bitterly fought (and well-publicized) local conflicts over temple entry and lower caste attempts to assert themselves symbolically (Sobhanan 1985). Castes encouraged voting for castemen in much the same way they encouraged interdinning.

78

From Hierarchy to Ethnicity

In a few cases, in fact, sabhas were directly formed for political purposes. The Gounder caste association, for instance, was founded by Vellingiri Gounder to publicize his activities as a member of the legislative council (Arnold Jeffrey and Manor 1976: 376). The close relationship between caste identity, local elections, and individual political careers can be seen in the careers of W. P. A. Soundrapandian, the ‘uncrowned King of the Nadar community’. Soundrapandian co-founded and dominated the central association of his caste, the Nadar Mahajana Sangam, serving for four years as general secretary and for seventeen years as vice president. Using his Nadar electoral support, Soundrapandian was able to have a very successful political career, serving for seventeen years on the Madras legislative council representing the Justice Party, and for six years as a district board president. Like many jati-based political entrepreneurs, Soundrapandian did not form a separate political party, but used his jati ‘vote bank’ to build a position within a larger political organization. For the new lower caste politicians, the obvious political goal was to transfer resources from other, historically wealthy groups to their own. In southern and western India, this meant attacking Brahmins, who dominated the local bureaucracy and the upper levels of local society (Irschick 1969; Omvedt 1976; Washbrook 1976). To the extent that every lower caste in these regions had the same enemy, they had a strong reason to ally, the more so since no group in these regions could form an electorally dominant coalition on their own. The natural consequence was the development of large ‘non-Brahmin’ or ‘anti-Brahmin’ movements, with a lower caste base, most notably the Justice Party, the dominant political force in the Madras Presidency in the 1920s. Though certain of these movements began to develop features of a more positive regional nationalism (P. Singh 2015), in this period they remained loose coalitions of caste groups united by a common enemy. However, there was no reason why the newly created caste voting blocks had to align with similar groups. In many cases, caste politicians found it more expedient to ally with upper caste elites, and in particular the Congress Party, clearly a coming force at the national level. Marathas in Bombay, Vokkaligas in Mysore, and Patidars in Gujarat are among the most prominent examples of castes with organizational and institutional affiliations with the Congress Party that predate independence (Jaffrelot 2003). In all cases, however, there remained a tension between

Caste in Historical Context

79

these groups and the upper caste Congress establishment, a division that would often have consequences for post-independence politics.

The role of the colonial state The British presence in India was originally commercial in intention, and the first generation of East India Company administrators had little intention of encouraging the types of hierarchical identity politics practised by their Indian predecessors. The result was a set of policies that often ignored traditional social distinctions. Military and bureaucratic recruitment, for instance, was at first conducted with no formal regard for traditional patterns, leading to the promiscuous creation of new elite groups (Lee 2017). However, gradually colonial scholars became more and more interested in the ascriptive divisions of Indians, and of caste in particular. During the last century of the Raj, British ‘orientalists’ (a term coined in the early years of company rule) collected an enormous mass of information on their Indian subjects, ranging from comprehensive linguistic surveys (Grierson 1906) to studies of Sanskrit texts (Jones 1824) to measurements of the head shapes of people from different castes (C. Bates 1995). The greatest monument to colonial scholarship on caste was a series of reference books on the ‘castes and tribes’ of particular provinces (Risley 1892; Thurston 1909; Rose 1919), a set of texts that remains an essential source on the topic. The causes for this turn in colonial scholarship, and their relationship to attempts by the colonial state to divide Indians and depoliticize Indian society, remain an object of scholarly interest (Inden 1990; Dirks 2002; Dodson 2007). The most direct way in which the colonial state associated itself with this scholarly turn was through the census, which was associated with not only a need to differentiate groups from each other but (in the 1901 census) a plan to rank groups relative to each other. The outcomes of these policies is discussed in Chapter 4. However, it is worth mentioning several other areas in which colonial policy influenced caste identities. In all of these areas, it tended to make the possession of particular identities economically valuable, though these groups were a small number within India as a whole. The first set of policies concerned military and police recruitment. Ever since the 1857 rebellion, colonial India remained acutely sensitive to the possibility of military revolt. As a result, army recruitment policies sought to identify loyal individuals who would at the same time make good soldiers. Before 1857 the army had been recruited widely, and included

80

From Hierarchy to Ethnicity

many southerners and some men from lower caste backgrounds, with the largest portion of the Bengal Army being upper caste Hindus from Bihar and Uttar Pradesh. After 1857, however, all these groups tended to be regarded as either disloyal or cowardly. By contrast, Marathas, Rajasthani Rajputs, Pathans, Punjabi Jats (of all religions), and Nepali Gurkhas had all both remained loyal during the mutiny and were considered ‘natural’ soldiers (Macmunn 1933). By the 1930s, these groups made up the vast majority of the Indian Army (Farwell 1989; Jha and Wilkinson 2012). The on-the-ground consequences of this change are difficult to measure, but the presence of a large employer making caste identity a key qualification for employment might easily be imagined both to make membership of those groups more attractive and to make even non-beneficiaries conscious of caste differences. It should be noted that while the colonial state practised discrimination in recruitment to the military and police services, for most of the colonial period it did not discriminate in recruitment to the other bureaucratic services. While Carrol (1978) presents some very limited evidence for informal discrimination against Kayasths in hiring, the existence of this policy was denied by higher officials, though Kayasths were officially banned from the army and police. British favouritism towards the agrarian castes of Punjab went beyond military recruitment. In the first decade of the twentieth century, British officials became concerned that Punjabi farmers were overly indebted, and that this was leading to the alienation of land to moneylenders (Darling 1925). While many solutions for this problem were possible, the Punjab government viewed the problem through the lens of caste: the unsophisticated ‘natural’ cultivating groups were being dispossessed by ‘parasitic’ merchant castes. The solution was the Punjab Alienation of Land Act (1901), which banned the acquisition of land by non-members of specific (listed) agricultural castes. Cassan’s excellent (2015) study shows that the Act had real effects on the ground. Predictably, people began to reclassify themselves as members of the agricultural castes. More subtly, the Act (and the question of its maintenance or expansion) had a major influence on political identification, with the Unionist Party (the largest party in colonial Punjab) being composed of a pan-religious alliance of agricultural groups (Talbot 2013). A final type of caste-based policy that emerged in the colonial era was the ‘reservation’ of government jobs, educational places, and legislative seats

Caste in Historical Context

81

for members of traditionally marginalized caste groups, an example of the type of compensatory discrimination founds in many other nations. Unlike the other policies discussed earlier, caste reservation in the colonial era tended to be primarily advocated not by British officials but by Indian politicians, especially members of the ‘backward’ groups themselves. Of the two major administrative units with reservation policies, Mysore (1918) was a princely state, and in Madras (1927), the policy was advanced in the elected legislative assembly. In both states, the primary targets of reservation were Brahmins, who were overrepresented among the educated and in the bureaucracy (Irschick 1969; Mathur 2004: 19–27). The creation of reservations would eventually create very strong incentives for individuals to characterize themselves as backward. This process began early: the Nadars Mahanjan Sabha, which had petitioned to have the Nadars removed from the list of depressed castes in 1918, petitioned to have them added to the list of backward castes in 1935 (Hardgrave 1969: 141). In modern India, the politics of backwardness have created incentives both for individual reclassification and the reclassification of whole groups as backward (Jenkins 2003). Caste reservation remains an important feature of the politics of post-independence India, encouraging caste mobilization and discouraging internal differentiation, since all members of a ‘caste’ benefit from reservations, and discouraging ranked mobilization (which would undermine a case for backwardness). In the colonial era, however, reservation was only beginning to become one of many tools by which the colonial state emphasized the importance of caste divisions.

Conclusion Throughout Indian history, identity has been closely linked to political and economic changes. Groups that gain in education and general economic status use that wealth and education to create forms of identity politics that will emphasize their power. Increases in education, like those that some groups experienced in the colonial period, are closely associated with increases in castes activism, such as the formation of sabhas. There are two forms of activism that the newly wealthy can pursue. One is hierarchical, emphasizing the high-status of the group and its similarity to traditional high-status groups. This may even lead, as in the Nadar case, to the newly wealthy trying to split from the larger group. The other

82

From Hierarchy to Ethnicity

de-emphasizes ranking, emphasizing the unity and identity of the group, as most caste movements do now and as the Nadars eventually did. The political environment determines the choice between these forms. In the pre-colonial era, most political entrepreneurs adopted hierarchical tactics, creating the intricate ideologies of social ordering that Dumont and Risley so eagerly analysed. In the colonial era, many followed their lead, especially in areas like the princely states where the traditional political order was strong and the colonial state weak. However, in certain regions and castes activists began to de-emphasize ranking in favour of more egalitarian forms of caste identity. The discussion of caste sabhas in this chapter is drawn from a few cases where this archival and/or historiographical evidence is particularly strong. While these cases do indicate the importance of caste activism in this period, many questions remain unanswered. Are the trends drawn from these cases typical of India as a whole? And are they robust to more consistent measures of key concepts such as education and activism? To answer these questions, the next chapter will examine a manifestation of caste activism that extended across India: the interactions of caste activists with the Census of India, and the peculiar ideas of the colonial officers who ran it.

4

Caste in the Census of India Much of the activity of ethnic activists, in particular their social and educational activities, occurred in a private, vernacular language world that was difficult for outsiders to observe even at the time. Variation in the level of importance attached to social ranking is similarly difficult to measure in practice, since much of the expression of ranking norms is cultural and takes place outside formal state structures. The role of the census in colonial India provides an interesting exception to this pattern: quite unintentionally, the colonial state created a means by which caste activists could register their activities and aspirations, and a language in which those aspirations could be expressed. The census assumed the role of a forum for caste claims when superintendent H. H. Risley decided to organize the census returns by hierarchical caste status. While Risley’s goal was to generate data supporting his own theory of the racial origins of caste, the classification change had dramatic consequences: many lower and middle caste elites anxious to maximize their hierarchical status organized to petition the census authorities for a new caste name, typically one that linked them to a ‘higher’ caste or to one of the three higher varnas of the Sanskritic caste hierarchy. Other caste elites also petitioned for new names, to distance themselves from the ranking system entirely, by disassociating themselves from names that suggested subordination. The census, of course, was not a neutral participant in this process. Simply by asking Indians about their caste, the census authorities potentially made caste more salient to individuals than it had been previously. Moreover, the census’s simple categories (and in some periods interest in a single hierarchical dimension) tended to delegitimize and override previous, perhaps more ambiguous, identities (Dirks 2002). However, as we shall see, there was considerable variation in the way in which caste activists used these new concepts and categories.

84

From Hierarchy to Ethnicity

Caste in the census While statistical enumerations of India had long collected information about caste (Guha 2003), the connection between caste and the census changed in the early twentieth century through the work of H. H. Risley, who became superintendent of the Census of India for the 1901 census. Risley was determined to use the census as an opportunity to prove his own theory on the origins of caste, which he thought reflected enduring (and physically measurable) racial differences between the Aryan ‘conquerors’ of ancient Indian and the Dravidian natives. He decreed that the census tables should list castes not by occupation, but by ritual ‘precedence as determined by native public opinion’, from Brahmins to untouchables. This precedence was to be determined by the provincial census superintendents in consultation with specially appointed committees of Indians. The criteria which Risley laid down for determining status are notable for their heavy focus on the Sanskritic tradition in general and Brahmins in particular, judging that status should be determined by . . . the facts that particular castes are supposed to be modern representatives of one or other of the original castes of the theoretical Hindu system; that Brahmans will take water from certain castes; that Brahmans of high standing will serve particular castes; that certain castes, though not served by the best Brahmans, have nevertheless got Brahmans of their own whose rank varies according to circumstances; that certain castes are not served by Brahmans at all but have priests of their own; that the status of certain castes has been raised by their taking to infant-marriage or abandoning the re-marriage of widows. . . . (Risley 1903)

Most, though not all, of the provincial superintendents followed Risley’s edict. Even when the committees and superintendents were left to themselves, this was a controversial and error-ridden procedure: The discussion of the relative rank of the different castes aroused an extraordinary amount of ill-feeling and jealousy between some of the castes whose position was disputed and in more than one instance the committees appointed to report on the subject professed their inability to come to a decision. (Gait 1901, 369)

More annoyingly for the census officials, many sabhas saw the policy as an opportunity to make a point and bombarded the provincial superintendents

Caste in the Census of India

85

with petitions demanding a new name, a request often implying some claim to a different caste status. No part of the census aroused so much excitement as the return of caste. There was a general idea in Bengal that the object of the census is not so much to show the number of persons belonging to each caste, as to fix the relative statuses of subsequent castes and to deal with claims of social superiority. Some frankly regarded the census as an opportunity that might fairly be taken in order to obliterate caste distinctions. . . . Hundreds of petitions were received – their weight alone amounts to 1.5 maunds[143 pounds] – requesting that they be known by new names, be placed higher in the order of precedence, be recognized as Kshattriyas, Vaisyas etc. Many castes were aggrieved at the position assigned them, and complained that it lowered them in public estimation. . . . One of the first steps is to obtain favorable vvavashthas or rulings from complaisant pandits. These refer to the present occupations and manner of life of the caste, and quote verses from ancient works to show that they are like those of the varna from which the caste claims to be an offshoot. (O’Malley 1912: 440)

By the time of the next census in 1911, Risley had retired, and census superintendents were not anxious to revisit the controversial system of ranking castes by ‘precedence.’ In general, they returned to the 1891 practice of classifying castes by traditional occupation. Even this system, however, became controversial, since traditional occupations often carried strong connotations of honour or dishonour within the traditional caste system, a practice reinforced by the way in which superintendents ordered the occupations. As a result, there was a gradual move towards grouping castes either alphabetically or alphabetically with the ‘depressed castes’ listed separately. Table 4.1 shows these trends by province. It is notable that, despite the attention paid to it in the historical literature, the career of the census as an orderer of all castes in the Risleyan mould was very brief. However, even as census officials thought they were withdrawing from adjudicating caste distinctions, caste activists continued to submit petitions at an increasing rate. Table 4.2 shows the upward trend over time in petitions, from 16.2 per cent of jatis in 1901 to 31.15 per cent in 1931. Overall, 23.2 per cent of caste-years had some sort of petition for a name change. (The number of castes changes over time because some census superintendents did not record the petitions they received.) These trends are shown graphically in Figure 4.1. The increase in petitioning long after census ranking had been abandoned in most parts of India provide powerful

Table 4.1 Census caste classification schemes by year, 1891–1931 Province Assam Bengal Bihar United Provinces Punjab Rajputana Bombay Central Provinces Hyderabad Central Indian Agency Mysore Madras Kashmir Baroda Travancore Gwalior Berar Western Indian Agency

1891 Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational

Source: Census of India, various years.

1901 Precedence Precedence Alphabetical Precedence Other Precedence Precedence Precedence Precedence Precedence Occupational Precedence Precedence Precedence Alphabetical Precedence

1911 Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational Occupational

1921 Occupational Occupational Alphabetical Occupational Occupational Occupational Other Occupational Occupational Occupational Occupational Alphabetical Occupational Occupational Occupational Occupational Precedence

1931 Other Depressed Castes Alphabetical Alphabetical Alphabetical Alphabetical Depressed Castes Occupational Occupational Depressed Castes Alphabetical Alphabetical Alphabetical Occupational Other Alphabetical Depressed Castes

Caste in the Census of India

87

Table 4.2 Aggregate number of petitions by year 1901

1911

1921

1931

Total

No Petition Petitions

134 27

239 55

267 67

228 113

868 262

Total

161

294

334

341

1,130

Source: Author’s calculations from Census of India, various years. Note: The unit of observation is the province-jati-year. The number of province-jati-years varies from year to year because some census superintendents did not record the petitions that were submitted to them. Figure 4.1 Proportion of castes petitioning for a name change, 1901–1931

Source: Author’s calculations from Census of India, various years.

evidence against the claim that petitioning was simply a response to Risley’s peculiar approach to caste classification. An even more persuasive argument against the argument that caste claims were entirely ‘caused’ by the practice of enumeration was the fact that enumeration on caste lines was not a nineteenth-century innovation. Guha (2003) has persuasively argued that the enumeration of caste groups in India has a long history, one that not only predated Risley but also predated British rule, being a common practice of the Mughal and Maratha regimes. In many cases, this enumeration involved adjudicating status within the aristocracy (as described earlier), but the process affected more humble groups as well. Guha further cites numerous incidents in which caste appeared to have been salient in pre-colonial times independent of enumeration.

88

From Hierarchy to Ethnicity

Despite these facts, Risley was certainly blamed by his successors for increasing their workload and for making the census more central to caste activism than it would otherwise have been: All subsequent census officers in India must have cursed the day when it occurred to Sir Herbert Risley, no doubt in order to test his admirable theory of the relative nasal index, to attempt to draw up a list of castes according to their rank in society. He failed, but the results of his attempt are almost as troublesome as if he had succeeded, for every census gives rise to a pestiferous deluge of representations, accompanied by highly problematical histories, asking for recognition of some alleged fact or hypothesis of which the census as a department is not legally competent to judge. (Hutton 1932: 433)

The attitude of the census officials towards the petitions and stacks of supporting documentation they received varied widely, though virtually no claims were accepted fully. Some officials, particularly those who had been schooled in traditional Sanskritic Indology, became angry in their rejection of these claims. These movements are contrary to the teachings of Manu, who classes falsely asserting oneself to be of too high a caste in the same category with breach of trust and incest. . . . (Francis 1902, 131)

Other officers took a more urbane (but no more generous) view: These claims were pressed with the greatest persistence, sabhas and mahasabhas often being formed for no other purpose, treatises being published, and eminent counsel being briefed. The course I followed was to correspond with the sabhas, to listen to the eminent council, and not to read the treatises; and then to instruct the enumerating staff to enter in the caste column the name by which a man’s caste was known to his neighbours. (Edye 1922, 151)

Overall, before 1931 only six petitions (out of 149 submitted) were granted. The decisions made by the census officials appear not to have been based on the social or political power of the groups in question, or even the historical or scholarly merit of their claims, but rather whether or not including the name would cause confusion between different castes (Blunt 1911: 323). Table B.4 in the Appendix B shows that differences in petition granting rates do not affect the results.

Caste in the Census of India

89

As time passed, however, opinions on petitions became more tolerant, and officials began to evolve a compromise that would record a caste by its traditional name with their new one in parentheses, or the new one with the old one in parentheses. This eventually became a national policy for the 1931 census, though not all provinces appear to have followed it, and the official tone was one of resignation rather than enthusiasm. Probably, it [petitioning] was not so intensive or insistent as at present, [so] the then census commissioner found it possible to pay no heed to the claims put forward and adhere to a no change policy. Traditions were recalled, puranas were quoted, opinions of great men were ferreted out and judgments of the high court were extracted, all purporting to justify the change and refute the arguments for no change. Some changes in names have therefore been adopted and this have been done in accordance with the decision of the conference of census officers held at Delhi on the 7th and 8th of January 1931. (Pillai 1932, 336)

What did the petitioners want? We have only limited information on who submitted the petitions. In all but a tiny handful of cases (discussed later), petitions appear to have been submitted by members of the concerned group. Although a few census reports mention petitions submitted by individuals, the overwhelming majority of discussions of submitters are in the plural, and many petitions specifically mention the sabha that submitted them. The fragmentary data that are available show that the majority of petitions were submitted by organized groups. In the United Provinces (UP) in 1931, a year for which information on petitioners was collected, 62.5 per cent of sample petitions were submitted by caste sabhas already registered as associations, while another 33 per cent were submitted by large informal groups formed for this purpose, many of them local caste panchayats (Turner 1932). It seems reasonable that the existence of a census petition usually indicates that a caste association or network of caste activists existed. As we have seen, such an organized and aggrieved group would almost certainly be conducting activities other than petitioning, many of which are described in the petitions themselves.

90

From Hierarchy to Ethnicity

There was no approved format for petitions, and a few of those in the National Archives of India are simple one-page letters written in Hindi, while others take up dozens of printed pages in English.1 The details of the petitions varied widely, from handwritten notes to masses of supporting documentation. Nor was there any standard means of submitting petitions. While some appear to have been mailed, others were presented in occasionally strident public confrontations with district and census officials. Mukerjea (1922: 325) noted that ‘the Barias of Padra were particularly obstreperous about their claims. They pursued me wherever I went with a ponderous dossier of documents and genealogical trees. . . .’ Petitioning the census of authorities might seem a useless act, since given the dismissive attitude of colonial officials, the chance of a petition being granted was extremely small (only 7.1 per cent were granted, and only 4.4 per cent if we exclude Madras in 1931). However, the high rate of rejection should not make us think that petitioning was in any way futile or irrational. Petitioning was a public act, planned in public meetings, made on stamped paper, and presented in public. Petitioners did gain a small probability of acceptance of their claims to a new name, which would lend legitimacy to their mobilization efforts and ideological claims. Relative to our own time, where citizens have numerous means of publicizing their opinions – websites, demonstrations, press releases – the public self-expression of British Indians was severely constrained, both legally and technologically. Petitions, drafted in public meetings and presented by large deputations, served as a way to publicly register group claims. Carrol (1978) argues that petitioning was motivated by the desire to gain employment in the colonial bureaucracy. However, there are two major objections to an explanation founded on direct colonial favouritism. First, the British did not have preferences for members of specific castes, let alone high-status castes, in bureaucratic occupation – Carrol’s only example was the disowned policy of a local collector. Second, given the virtually universal policy of denial, it is unclear if petitioning was a rational way of seeking

1

While most petitions appear not to have been preserved, I was able to review some petitions (either from the Central Indian Agency or Bihar) in the National Archives of India. For examples, see Home Department, Public Branch, F. 45/6/31-Pub., and Home Department, Census Branch, A Progs. December 1900, Nos 24–27.

Caste in the Census of India

91

colonial favour. The contrast to later reservation policies (where access to actual jobs was explicitly at stake) is very notable.2 Unlike the other rhetorical modes employed by sabhas, like pamphlets, petitions have several advantages for scholarly study: They were easy to produce, even for poor groups, and were preserved by the colonial state rather than lost or scattered. For this reason, they represent a unique source on the types of claims made by groups, though it should be noted that in many cases they were part of broader strategies of identity-based activism. The need for publicity was one reason petitioning was an attractive strategy for promoters of all types of identity projects. However, another factor was also in play: the presence of a system of hierarchical norms within the census system (implicit or explicit), which seemed to challenge their claims. For groups emphasizing ranking, census classifications would be a refutation of their claims to be part of or close to a high-status group, all the more damaging for its ostensible objectivity. For leaders attempting to emphasize their separateness from the Sanskritic system, classification in a subordinate position or under an insulting name was an equally public slap in the face, and an endorsement of the subordination that they denied. Petitions were the only way these elites had of confronting and challenging colonial attempts to impose a fixed version of the traditional hierarchy. Conversely, many petitions sought to erase their group’s existence as a separate category as far as the census was concerned. For instance, Molony (1912: 159) notes that ‘the Badagas of the Niligri Hills have put forward a claim to rank as “Mountain Brahmins”’. Edye (1922: 151) remarks: ‘As before, a large number of castes put forward claims to be classed as Brahmans or Rajputs whose claims are not admitted by the general community.’ For such groups, the mere fact of being listed as a separate group was an affront to firmly asserted beliefs, one that required a public challenge. As we have seen, petitioning was not necessarily significant in and of itself – the formation of a caste sabha was probably far more important in the long term than the circulation of a petition. However, petitioning serves 2

As noted above, the colonial state did hand out some social benefits, such as access to land (Cassan 2015) and military recruitment. However, there is little evidence linking these benefits to petitioning. Punjabi castes petition at lower rates than the rest of the country, as do the high caste landed groups most likely to wish to enter the colonial army, and the main results are robust to the exclusion of Punjab, and, indeed, any one province or caste category. In the statistical models in the appendix, this sort of caste favouritism is accounted for by the inclusion of caste, year, and province fixed effects.

92

From Hierarchy to Ethnicity

as a visible indicator that activist efforts of some sort were underway within the caste, and of what rhetorical strategies they employed. The centrality of the census to many accounts of Sanskritization in this period makes this a very natural measure of mobilization from a theoretical perspective. In addition, unlike other types of caste activism that did not involve interaction with the colonial state, the records of petitioning are relatively well preserved. Caste sabhas used the census as a forum to air their aspirations and ideological projects, and to protest the roles to which they were assigned in the census’s existing caste schema. The result was a remarkable record of both the presence of caste activists (as measured by petitions) and their approach to ranking (as measured by the goals these petitions announced).

Trends in petitioning It is worth examining some of the basic trends in the incidence of petitions. Petitioning activity is measured for each jati-province-census year. Therefore, ‘Kayasths in Bihar in 1921’ are treated as a separate observation from ‘Kayasths in Bihar in 1911’ and from ‘Kayasths in Bengal in 1921’. Jati-provinces which never had more than 0.5 per cent of a province’s population are excluded, as are province-census years that did not record petitions. In the data appendix (Appendix A), I discuss in greater detail the construction of the dataset, and the complex patterns of missing data that result.3 The most widely available measure of education is the male literacy rate of each caste, which the census authorities calculated on the generous basis that anyone capable of writing his name in any language was literate. 3

The census authorities often changed the names by which they referred to castes. These changes do not appear to be correlated with petitioning, and controlling for or excluding these names changes does not affect the reported results. The primary data used here are drawn from the published reports of the Census of India for the years 1891, 1901, 1911, 1921, and 1931. Before 1891, the census did not cover all of India, while after 1931 it did not include information about jati. Since petitioning began in 1901 in response to Risley’s policy change, the 1891 caste-provinces are not included in the main results. In general, provincial superintendents noted the petitions in the introduction to the caste section of the report. Usually, only the caste and the name claimed was recorded, though in some cases the superintendent also noted the group making the claim, and some details on their justification.

Caste in the Census of India

93

Literacy is a reasonable predictor of the existence of the kind of educated class that the theory posits as necessary for caste politicization, though some models in the appendix use alternative measures. Note that literacy is measured at the group-level, which posed difficulties for testing the second part of Hypothesis 1. While a group with even a small literacy rate obviously has some literate members, it does not necessarily have many highly educated members. The census data, which has only literacy as a useful measure of education, does not allow us to capture the distribution of education within the groups.4 However, it appears reasonable that a group with a high level of literacy would also tend to have a high proportion of highly educated individuals. This is borne out by recent survey data: in the Indian Human Development Survey, group-level literacy was correlated with the group-level rate of secondary education at ρ = 0.78.5 In Chapter 3, we saw that literacy increased over time in India, at the same time as petitioning increased. Figure 4.2 shows the changing probabilities of petitioning over time relative to the initial (1901) literacy rate. As the theory predicts, the trends for rich and poor castes are different. Castes with low levels of literacy petition with increasing frequency in later decades, consistent with the hypothesis that the higher levels of literacy that some of these castes gain over time give them a sufficiently large elite to initiate the mobilization process. Among groups with high initial levels of literacy, the number of petitions stays constant over time. Figure 4.2 Rate of petitioning by 1901 literacy rate and years, 1901–1931

Source: Author’s calculations from Census of India, various years. Note: ‘Poor’ jatis are those with less than 5 per cent male literacy in 1901; ‘rich’ jatis are those with greater than 5 per cent literacy in 1901. 4 5

English literacy appears to be very poorly measured (Barrier 1981). Reported caste names used. Minimum of 50 individuals in sample.

94

From Hierarchy to Ethnicity Table 4.3 Number of petitions by literacy (pooled)

Literacy

Caste–years with No Petition

Caste–years with Petition

Total

Petition Per Cent

Less than 1% 1–5% 5–10% 10–20% 20–30% 30–40% 40% + Missing

115 254 73 99 60 36 57 174

14 80 31 34 38 14 15 36

129 334 104 133 98 50 72 210

10.85 23.95 29.81 25.56 38.78 28.00 20.83 17.14

Total

868

262

1130

Source: Author’s calculations from Census of India, various years.

The inverted U relationship between literacy and petitioning is fairly obvious in the raw data. Table 4.3 gives the number of petitions in relation to the literacy rate of the caste. Castes at the lowest levels of socioeconomic status (below 1 per cent literacy) petition in only 10.85 per cent of jati-province-years, while those immediately above them (between 1 per cent and 5 per cent literacy) petition 24 per cent of the time. The rate of petitioning is highest among castes with between 20 per cent and 30 per cent male literacy – 38.8 per cent of these castes seek to change their name in any given year. At the highest levels of literacy (above 40 per cent), the rate diminishes, with only 20.8 per cent of castes submitting petitions. This pattern can be seen graphically in Figure C.1 in Appendix C, which shows the kernel density functions for petitioning and non-petitioning castes by literacy rate. The non-petitioning castes are concentrated at very high and very low levels of literacy, while the petitioning groups are relatively numerous at intermediate levels of literacy. Petitions were not recorded for many caste groups, even those that had levels of literacy in the intermediate range where the practice was most common. As Table 4.3 shows, even at the levels of literacy with the highest rates of petitioning, some 60 per cent of caste-province-years did not have a petition recorded. There are a variety of possible explanations for why petitioning was not universal, some of which have already been mentioned. Even groups with a sizable number of literates might not have active caste associations, whether because of a lack of charismatic leadership, geographical dispersion, or opposition from traditional caste elites. Even groups with caste associations might not choose to petition the government,

Caste in the Census of India

95

whether because they were deterred by the low rate of acceptance or could not achieve a consensus on what name to claim. In particular, some group members might be satisfied with the ethnonym given them by the census. If the name assigned lacked an unpleasant connotation or already embodied the group’s aspirations, it might be embraced by political activists. The Kunbis of the Marathi speaking districts, for instance, were returned as ‘Maratha Kunbis’ at all four Bombay censuses, thus recognizing their claimed link to the larger Maratha caste and making petitions unnecessary for this otherwise politically active and upwardly mobile group.6 Similarly, there was variance in the degree to which groups had internal traditions of a caste status higher than the one assigned to them in the census tables, an obvious motivator for petitioning. The provincial volumes of the Castes and Tribes series records origin myths for 117 sample castes, of which 85 per cent lay claim to some form of honourable status, though this is obviously a small minority of the sample.7 To account for this type of group level difference, the regression models will make comparisons within jatis, rather than across them. Another explanation for the low rate of petitioning is the fact that the social and political context of petitioning varied from place to place. Petitioning was a device that required some level of familiarity with the census bureaucracy, and thus tended to become more common over time as the census became routinized. Similarly, petitioning tended to thrive in contexts where there was some level of formal or informal political contestation, which both incentivized political organization and provided an assurance that petitioning would not lead to violent retribution by the upper castes. Rajputana and the Central Indian Agency, groupings

6

7

In the caste of the Kunbis, census policy reflected an unusual degree of acknowledgment of the fluid nature of caste identities on the ground, noting that ‘in the case of the Kunbis and Marathas of the Deccan and Kokan it seems certain that the boundaries of the Maratha caste are hazy, and that well-to-do Kunbis assume and retain without opposition the Maratha name’. This in turn reflected internal debates within the census organization about the proper status of the caste (Sedgwick 1922: 183). Twenty-nine per cent claim descent from gods or servants of gods, 14 per cent claim descent from mythological or historical kings, 15 per cent to be the illegitimate offspring of members of the upper three varnas, and 27 per cent to be descendants of members of the upper three varnas who lost status through an accident or in order to serve others.

96

From Hierarchy to Ethnicity

of princely states ruled by traditionalist upper caste dynasties, had very low rates of petitioning (12.6 per cent and 11.7 per cent), while Bengal, the most politically restive part of India in the early twentieth century, had the highest rate of petitioning (47.5 per cent). Venkatachai (1932: 213) noted that ‘at present the census in the [Central Indian] states is looked on with indifference, and it excites no curiosity or rivalry on the part of caste organizations which, as a matter of fact, do not exist at all’. The need to account for this type of variance is the major reason why more complex statistical models are necessary.

Regression models A major obstacle in the empirical study of political identities is the unobserved variation between groups. Some group identities may be better developed than others, or some groups may benefit from state policies. In Ghana, for instance, Asante ethnic entrepreneurs were able to use the symbols, networks, and affinities developed by the pre-colonial Asante kingdom, while other groups had to develop ethnic identities without this advantage (Paine 2018). In the context of caste, a particularly concerning problem is the ascribed status of the caste, since low-status castes might have a particularly urgent imperative to attempt to challenge existing identity forms. The solution adopted here is to make comparisons within rather than across ethnic groups. This within-group variation is of two types: First, members of the same groups in different regions may have different levels of socioeconomic status, and thus potentially different political outcomes. Second, a given group’s socioeconomic status may change over time. This is particularly true in modernizing societies like colonial India, where relative education levels were changing. If we were to compare levels of mobilization within groups over time, different types of unobserved variation would be problematic. Economic, social, and political conditions change rapidly over time, potentially affecting the ways in which ethnic politics expresses itself. In the Indian case, the period of petitioning coincided with the growth of the nationalist movement, a world war, the growth of a variety of Hindu revivalist and nationalist movements, and changes in census policy, any one of which might bias estimates of the effect of education on caste mobilization. As a solution to these two problems, this book supplements the raw data with a series of more complex statistical models. The panel of fixed effect models estimates the marginal effect of the increase in the literacy

Caste in the Census of India

97

of a particular jati-province-year (that is, Yadavs in Bihar in 1921) on the probability of members of that jati-province-caste petitioning the census. To capture the predicted inverted U relationship, both literacy and squared literacy are included. All models account for variation between castes (through caste fixed effects), variation at the year level (through year fixed effects), between provinces (through province fixed effects) and over time within jati-province-years (through the inclusion of a variety of control variables). A thorough description of these models – the data used, the coding rules used to construct the variables, and the results – is included in the appendix. The curvilinear relationship between group literacy rate and petitioning is apparent in the regressions in Table B.1, shown graphically in Figure 4.3: literacy has a positive association with petitioning, but very high rates of literacy (proxied in the models by literacy squared) have a countervailing negative association. The effect of the literacy variable is not only statistically significant but substantial in substantive terms: for an untouchable caste in Baroda in 1901, moving from a literacy rate of 1 per cent to 11 per cent increases the estimated probability of petitioning by 7 percentage points (from 16 per cent to 23 per cent) and a further literacy increase from 11 per cent to 21 per cent is associated with an additional 5.6 percentage point increase in the probability of petitioning (to 28.6 per cent). However, the effect of literacy becomes negative as it goes higher, as a caste with 41 per cent literacy has a predicted probability of petitioning of only 24 per cent, and a caste with 61 per cent literacy will petition only 7 per cent of the time.8 As an alternative, some alternative tests use the percentage of workers who are owners or managers of industrial firms. These models produce nearly identical results to those that use literacy, indicating that both are measuring the same underlying differences between ‘advanced’ and ‘backward’ castes. While many of the highly literate castes are small, the results are not a product of either absolute or relative group size (which are controlled for in all the models). In fact, the positive impact of group numbers on petitioning is modest and only inconsistently statistically significant. This is unsurprising, given the many very small castes (with less than

8

These substantive effect estimates refer to the mixed effect model, Model Four of Table B.1.

98

From Hierarchy to Ethnicity Figure 4.3 Predicted rate of petitioning by male literacy rate

Note: Based on the predicted values from Table B.1, Model Two.

1 per cent of the province’s population) who were able to submit petitions.9 One common explanation for the caste system is that it represents a system of economic complementarities, with groups providing each other with specialized services through the jajmani system or other similar arrangements (Dubois 1806; Freitas 2006). What if this interpretation is correct, and economic specialization is an important motivating factor for individuals choosing to remain within the caste hierarchy? Traditional employment is a reasonable measure of exposure to these complementarities, since a group concentrated in its traditional occupation is more likely to give and receive a large portion of its income through cooperative exchange, if only because it has a less diverse array of buyers and sellers within the group. However, the models find that while traditional employment has a positive association with petitioning this does not influence the relationship between literacy and petitioning.10 9

10

Castes with less than half a per cent of the province’s population are not included in the dataset. This effect also implies that caste mobilization is not driven by the need to position the group in new economic sectors, since petitioning is associated with remaining in the traditional occupation. Similarly, group involvement in trade is not correlated with petitioning, and its inclusion as a control does not alter the reported results.

Caste in the Census of India

99

Alternate forms of political participation If the elites of very highly literate groups are not engaged in caste politics, what are they doing? Chapter 2 proposed that the elites of educated groups spend their time engaging in other forms of political action. In fact, we should expect them to be over-represented in the political sphere as a whole, given the advantages conferred on them by their education. In the context of colonial India, what might such activities be? Two possibilities stand out. The first is involvement in the colonial bureaucracy. Given the small number of British administrators, Indian bureaucrats enjoyed considerable power and opportunities for personal advancement in colonial India. Prospective bureaucrats were socialized within an environment that emphasized loyalty to the King-Emperor but also dispassionate bureaucratic objectivity (Potter 1996). At the caste level, bureaucratic involvement is measured as the proportion of male workers from each caste-province-year who were officers in the public administration, taken from the census reports. As with other caste-specific measures, the willingness of census superintendents to enumerate the occupation of every caste was dependent on budget and personal inclination, and is missing for many cases. The data show that, much as contemporaries believed, the distribution of officers was skewed heavily towards the traditional educated castes. Among government officers in UP in 1931, 39.4 per cent were Brahmins, while 22.3 per cent were Kayasths or Bania. As we shall see, this concentration of public employment among a few small castes would contribute to the pressure for the implementation of caste-based affirmative action in the post-independence period. Another major forum for elite Indians in this period was the Indian National Congress. Before the 1930s, the Congress was more of a talking shop than a political party, holding annual sessions that anyone who could afford the train fare was free to attend. To the extent that it agreed on anything, the Congress agreed on projecting a ‘national’ identity and disparaging caste and regional divisions. In 1917, the Congress Lucknow session passed a resolution condemning untouchability, a bold stance for the time. The opposition between caste and nationalism has been widely noted, particularly in the work of Nicholas Dirks (2002). We are fortunate in having complete records of attendance at early Congress sessions, including, in many cases, the occupation and caste of attendees. As an index of Congress membership, I use the proportion of

100

From Hierarchy to Ethnicity

Congress attendees from the province for that year who were from the caste, normalized by the provincial population in thousands. Like the bureaucracy, the Congress was dominated by the traditional educated castes: among the 91 delegates from UP attending the 1911 session, 74 were Brahmins, Kayasths, or Banias. As this brief description of the data should indicate, the relationship between group-level literacy and these measures of non-identity-based political participation is simple: Both participation in the Congress and in the colonial administration were concentrated among the small set of castes in each province with very high levels of literacy. This pattern is also seen in the regression models in the appendix, which show that the curvilinear pattern found for petitioning does not extend to the Congress and the bureaucracy, but that both show a strong association with high levels of literacy.

Alternative hypotheses One of the most basic alternative hypotheses is that petitioning is driven by the pre-existing levels of social status ascribed to groups. Groups with a high level of status within the traditional caste hierarchy might have a lesser incentive to change their caste name or to take other steps for their political advancement. However, all the reported models include either fixed effects for the jati, or a set of dummy variables indicating the various levels of the caste hierarchy. The reported result should thus be seen as showing the effects of divergence in literacy among groups of similar ritual status. In fact, there is considerable variance in petitioning, even among groups with a similar social status. Table 4.4 shows the number of petitions by caste status level. The major empirical pattern is the virtual absence of Table 4.4 Number of petitions by caste status Caste Status Untouchable/Dalit Lower OBC/Shudra Upper OBC/Shudra Intermediate/Dominant Upper/‘Twice Born’ Brahmin Total

Castes with No Petition

Castes with Petition

Total

Petition Per Cent

200 322 143 101 62 39 868

45 108 61 30 17 1 262

245 430 204 131 79 40 1130

18.37 25.12 29.90 22.90 21.52 2.50

Source: Author’s calculations from Census of India, various years.

Caste in the Census of India

101

petitioning among Brahmins, the groups at the top of the hierarchy, who cannot envisage improvement in their traditional social status. Otherwise, the differences among the various levels of status are relatively minor. Another concern is the possibility of individual movement. Castes, like other ascriptive identities, are not constant units: People enter and leave through birth, death, and by changing the answers that they give others. By changing their responses to census-takers, individuals and groups can move from one jati to another (Rao and Ban 2007; Cassan 2015), making jati-level social measures endogenous to a variety of social and political factors. There is considerable evidence that such movement in fact occurred: many of the caste sabhas who submitted petitions encouraged their members to return themselves under the new name. The Bihari Bhumihars (discussed later) encouraged members to return themselves as Brahmans in 1921, while in 1901 many Nadars returned themselves as Kshatriyas. This pattern raises two major concerns: that individual reclassification would substitute for the type of collective reclassification that the petitions were attempting, and that movement in and out of groups would skew at statistical measures. However, there is reason to think that these issues are not important in practice. The census authorities were anxious to keep individuals within their ‘proper’ caste, and took steps to minimize the effect of such movement. The danger of inaccuracy arising from these claims is not however very great. In the first place these claims were generally resisted by the enumerators, who had as large a share of human nature as anyone else. In the second place the claimants in all cases had put me in possession of the name they wished to use. And to restore the popular name in the course of tabulation was no difficulty. (Edye 1922: 151)

As a result, only a few castes show negative population growth from year to year (which would be the natural result of individuals leaving castes through reclassification), and most of those are concentrated in Punjab (the focus of Cassan’s 2015 study). Excluding these cases from the regression models does not alter the reported results. The regression analysis rules out a number of additional plausible alternative hypotheses. It is possible that the variation we see in petitioning is not the product of variation in the spread of literacy, but rather variation in preexisting cultural conditions. In particular, many observers have argued that the caste hierarchy is weaker in the south due to the ‘incomplete’

102

From Hierarchy to Ethnicity

varna system, with the far south having only a few Brahmins and no groups universally acknowledged as Kshatriyas or Vaisyas (Kothari 1970; Frankel and Rao 1989). Much of this variation is accounted for by the district and state fixed effects. Moreover, adding a control for the proportion of the population with a higher ascribed social status does not alter the results. What if assertion by lower caste groups is driven by broader movements towards social reform within Hinduism? Of particular concern is the role of the Arya Samaj. The Samaj was a social service organization strongest in north India, whose ideas have several times been described as Hindu Protestantism, Hindu fundamentalism, and Hindu reformism. The group’s leadership set itself against the caste system, at least in its existing form, favouring instead some sort of purified varna division (Bakshi 1991). In general, we should expect the rapid growth of the Arya Samaj to have a negative effect on petitioning, which represented the kind of jati-based division of Hindus that the Samaj was at pains to avoid. While this relationship is apparent in the regression models, province level Arya Samaj population is negatively associated with petitioning, which has no effect on the literacy results. Another important ideological influence on some lower caste elites in this period, Christianity, might also influence the results. The Christian missionaries who arrived in India in the nineteenth century were in theory opposed to caste distinctions and the caste hierarchy, though they never succeeded in eradicating caste consciousness among Indian Christians. More notable was the missionaries’ role in promoting primary education, particularly among their primary converts, lower caste groups in southern India. It is possible that Christianity influenced caste mobilization by making lower caste groups hostile to the existing hierarchy (Jaffrelot 2003). While increases in the Christian population are associated with increases in petitioning, the inclusion of this variable does not alter the literacy results. While the census is conceived here primarily as an arena in which castes made rhetorical claims, it is certainly possible that this arena was not neutral – that official policy influenced which groups made claims. Much of this variation, of course, is subsumed in the caste fixed effects: to the extent that colonial policy or scholarly discourse favoured the aspirations of particular castes, this is already accounted for in the regression models. However, as we have seen, several aspects of census policy differed between provinces and over time, notably the way in which castes were classified

Caste in the Census of India

103

and (to a lesser extent) the rate at which petitions were granted. Models in the appendix include these measures of census policy, which are shown now to influence the main results.

Illustrative examples The quantitative results abstract away from the actual trajectories of caste identities in this period, including such important details as the continuities between the colonial and pre-colonial periods, and the role of charismatic individuals. This section will describe the experiences of two groups who experienced noticeable changes in their socioeconomic status in both the pre-colonial and colonial periods: the Bihari Kayasths and Hindustani Jats. In each case, social and political gains were intimately associated with the development of political identities designed to reinforce and perpetuate those gains. For the Kayasths this meant (at least in Bihar) a general avoidance of anything resembling identity politics. By contrast, the Jats organized a strong caste activist movement with a distinct Sanskritic flavour. In both cases, this reflected their educational position: The Kayasths were by far the most educated group in their region, while the newly educated Jat elite was small relative to those of other groups.

The Bihari Kayasths The Kayasths are a north Indian caste traditionally connected with writing and scribal services. The Kayasths occupied a slightly ambiguous position in north Indian society, being considered generally ‘respectable’, but to be Shudras rather than members of the three upper varnas. Kayasths are found in Bengal, Bihar, and UP, and the political behaviour and social status of the group varies from province to province. The socioeconomic rise of the Kayasths in Bihar was linked to the pre-colonial tendency to mortgage or sell revenues to tax farmers. Banking and revenue farming were not at all inconsistent with involvement in the ruler’s own administrative machinery, and many revenue farmers had their origins in the secretariats of rulers. Some of these Indian bankers were figures of major political significance, like the Jagat Seths, who played a key role in the British penetration of Bengal, or Sheo Lal Dube, the banker who farmed the revenue for much of Bihar in the late eighteenth century (Cohn 1960). Most, however, were men of purely local significance

104

From Hierarchy to Ethnicity

and focused on using their position in the revenue hierarchy to gain local influence, particularly zamindari rights, which were felt to be more secure from official expropriation than cash (C. A. Bayly 1983: 166–170). The Kayasths were one of the groups that benefited the most from these developments. Unlike many revenue farmers, their background was in bureaucracy rather than trade: distrusting Brahmins, many Muslim rulers drew their administrators from among the Kayasths, notably Emperor Akbar’s finance minister Toder Mal, who completely overhauled the Mughal tax system. The major Muslim rulers in the eastern portion of the Gangetic plain, the nawabs of Oudh and the sultan of Bengal, granted extensive tax farming rights to Kayasths in Bihar and Bengal, and many non-Kayasth revenue farmers employed them as administrators (C. A. Bayly 1983: 166–167). Tax farming, combined with private trade, enabled some Kayasths to become very wealthy, and many acquired substantial landed bases. This trend only continued in the colonial period, for the East India Company’s weak bureaucracy and insatiable demand for revenue made them even more dependent on Kayasth tax farmers and intermediaries than their predecessors had been. Some of this wealth was invested in the creation of political status. Kayasths were traditionally considered Shudras, and thus were in theory inferior to Brahmin priests and Rajput rulers. This meant that the Kayasths do not represent a caste with anything to gain in terms of the traditional caste hierarchy. Indeed, the traditional exclusion of the caste from upper caste status might well provide exactly the kind of grievance against which other groups petitioned during the colonial period. Kayasth elites worked energetically to reduce this ritual distance, by adopting the behaviour patterns of the established elites, in some cases becoming more ostentatious than the Brahmins. By wearing the sacred thread, avoiding meat, and cloistering their widows, the Kayasths hoped to liken themselves to Kshatriyas, and acquire the legitimate right to exercise the power they were gradually accumulating (Risley 1891: 438–445). The economic and educational position of the Kayasths further improved during the colonial era. Their bureaucratic tradition meant that the Kayasths were the first caste in Bihar to take advantage of the opportunities opened up by the spread of English education and the colonial civil service. This head start enabled the Kayasth elite to dominate Bihari politics in the first decades of the twentieth century. In 1911, the Kayasths had 1.2 per cent of the Bihari population, but 32.5 per cent of the gazetted

Caste in the Census of India

105

civil service officers and 47.9 per cent of the local delegates to the Indian National Congress. However, the Kayasths of Bihar never emphasized their caste identity, instead claiming to represent the interests of all Biharis or all Indians, either within the idiom of bureaucratic service or Gandhian nationalism. Prominent Kayasth politicians, such as Krishna Ballabh Sahay (revenue minister and chief minister after independence) or the social activist Jayaprakash Narayan, tended to stress their nationalist and socialist credentials rather than their caste identity. While there were Kayasth caste associations in Bihar, the province represented the centre of the ‘anti-Conference’ faction of Kayasths who opposed the reformist policies of the national sabha (Carroll 1979). Similarly, in electoral politics Kayasth politicians did not function as a homogenous caste bloc, but rather as a set of personalist factions (S. S. Jha 1972; Jaffrelot 2003). As the group with the best access to Western education in an extremely poor society, the Kayasth elite did not need caste sabhas or petitions to reinforce their power.11

Hindustani Jats The Jats are a group of peasant cultivators in northern India, and are the largest agricultural group in both Punjab and in the Hindi-speaking area around Delhi (Hindustan). The Jats had originated as a pastoral group with a weak traditional Brahminical Hinduism (Asher and Talbot 2006: 270), though as the group became more settled many of its members became practitioners of Islam and Sikhism in Punjab and Hinduism in Hindustan. At the same time that they were being drawn into agriculture, Jats were being drawn into the political and military networks of the Mughal Empire, the dominant power in the region (S. Bayly 1999). The Mughals mistrusted the Rajput landholding class of Hindustan, and conducted a campaign of ethnic cleansing in the countryside, replacing Rajputs with loyal Jats (Kolff 1990: 13–15). While Jats did not occupy the highest ranks of the Mughal administrative hierarchy, they were well represented at the lower level, at the rank of the 11

Bihari Kayasths did become involved in several cases before the Patna High Court (1884, 1927). However, as Carrol (1978) points out, the litigants in these cases were not representative of the group, and had incentive growing out of their own financial interests.

106

From Hierarchy to Ethnicity

zamindar. The zamindar was a man or a family who undertook to pay the government’s revenue demand from a village or group of villages, passing on both his own contribution and those of his neighbours. Many zamindars used this position of arbitrage to become petty feudal lords, with small forts and bands of armed relatives (Kolff 1990: 1–9). Jat lineages that held this position were able to demand a relatively high level of deference from their Hindu neighbours, who considered them ‘clean Shudras’, a dramatically better position than that granted to previously similar north Indian pastoral groups such as the Gujjars. The decline of the Mughal Empire offered these local Jat elites the potential for upward mobility. This was first manifested in a series of zamindar-led peasant revolts against Mughal tax demands in the early eighteenth century. In the latter half of the century, the more aggressive of these petty landlords had evolved into petty rulers, often supplementing taxation with raiding (see C. A. Bayly 1983: 35–73). These Jat rulers faced the same task of legitimation that their Rajput predecessors had faced, without the high ritual that had helped the Rajputs overawe the rural lower classes. The Jats developed two solutions. In the first, some Jat rulers rejected Hinduism entirely for Sikhism, though this move did not imply either the destruction of Jat group consciousness or the end of their pretensions to sovereignty over other groups. The second method was to imitate the Rajputs in the patronage of Brahmins. Despite a tradition of egalitarianism within the Jat community, the rulers of eighteenth-century Jat states such as Bharatpur actively emphasized their own caste superiority (C. A. Bayly 1983: 49–50). In this as in other cases, attempts to improve the status of the ruler within the hierarchical system were associated with an increase in internal differentiation within the Jat category, as the royal lineage sought to separate itself from its less powerful coethnics. More subtle, and less well studied, is the process by which Jat lineages claimed positions of prestige within villages (Lewis and Barnouw 1958). The annexation of Punjab and Hindustan in the early nineteenth century might have been thought to have reduced the economic position of the Jats, but in fact the opposite was the case. Jats (Hindu, Muslim, and Sikh) were the most important of the ‘martial races’, probably making up at least a third of the entire British Indian army.12 To guarantee the loyalty of these 12

Some ambiguity was introduced by the collective categories ‘Punjabi Muslims’ and ‘Sikhs’.

Caste in the Census of India

107

soldiers the British introduced a variety of policies designed to benefit rural Punjabis. A generous tax assessment at the time of annexation, coupled with a land tenure system that placed revenue responsibilities on the village headmen or panchayat, reinforced the power of local Jat elites (Lee 2017). The Punjab Alienation of Land Act (1901) gave the Jats and members of other agricultural castes a monopoly right to purchase land (Cassan 2015). A robust programme of irrigation investment created large new agricultural areas that were preferentially allocated to soldiers and existing cultivators (Ali 1979). Finally, the ‘Punjab system’ of administration attempted to discourage the growth of a native bureaucracy and emphasize informal ties between British officers and local elites (Yong 2005). This favouritism had discernible economic effects: Punjabi Jat male literacy more than doubled between 1891 and 1931, while the proportion of lawyers, doctors, and teachers increased by a factor of six between 1911 and 1931.13 These changes in the social position of the Jats were paralleled by the growth of Jat caste consciousness. In 1907, the Jat Mahasabha was founded under the influence of the Arya Samaj in both UP and Punjab, with many of its constituent organizations being influenced by the Arya Samaj. The Mahasabha, like many other caste organizations, promoted education (sometimes in its own schools) and Hindu religious ceremonies (N. Datta 1997). To plan and promote these goals, the Mahasabha held annual conferences (adhiveshanas) at cities scattered across northwestern India, with addresses by British governors and wealthy zamindars seated prominently on the rostrum. Sir Chhotu Ram, the founder of the Mahasabha, hyperbolically declared: ‘Earlier the Jats were all fragmented, but now they were brought together. They began to perceive all their institutions as exclusively Jat. The Jats became conscious of their sense of unity’ (Akil Bhartiya Jat Mahasabha 2017). However united the Jats were, the existence of a Jat pressure group did Chhotu Ram’s career no harm. Besides his role in the sabha, Ram founded a newspaper, the Jat Gazette (1916), and a school, the Jat Anglo-Sanskrit School (1913) (N. Datta 1997). Ram used these organizations as a base for his campaign for the Punjab Legislative Council, of which he would remain a member until his death in 1945, rising to

13

Census-reported literacy for Male Hindu Jats was 1.3 per cent in 1891 and 3.5 per cent in 1931. The proportion of lawyers, doctors, and teachers among male workers rose from 0.02 per cent in 1911 to 0.19 per cent in 1931.

108

From Hierarchy to Ethnicity

agriculture minister (1923) and revenue minister (1935). Ram would be one of the leading figures in the Punjabi Unionist Party, which promoted the interests of farmers (many of them Jat) against urban interests. His opposition to the Congress (which he had quit in 1920) was rewarded by the British with a knighthood – quite a rise for a village boy from Rohtak (Chowdhry 1984). The Jats provide an example of how a change in the economic position of a group led to a change in its identity, and how the patrimonial structure of pre-colonial politics encouraged this change to emphasize social ranking. In the colonial period, the colonial favouritism (and the corresponding rise in group education) resulted in another outpouring of Jat identity politics, one less influenced by ranking.

Outliers While socioeconomic status explains a good deal of the variance we observe in petitioning, it does not explain all of it. It is worth examining some cases where petitions were filed on behalf of groups that might be expected to have a low probability of petitioning. There are several cases of petitions for groups that had very low levels of literacy – including four petitions from province-caste-years with male literacy rates of less than half a percentage point. When this occurs, it often appears to be a result of social heterogeneity within the social category: a very poor caste might, particularly in large and heterogeneous provinces, include a small elite group that had the education to file a petition. This reflects the overall limitations of using the imperfect census figures as a measure of literacy, and group level literacy as a measure of elite socioeconomic status. This certainly seems to have been the case of the Chamars of UP, who had the second lowest literacy rate of any petitioning group. In this case, the petition was submitted by members of the group who had abandoned the group’s traditional occupations of leatherwork and scavenging and taken up the somewhat more remunerative occupation of weaving,14 and they petitioned not on behalf of all Chamars, but to have Chamar weavers counted as a separate group (a type of petition that the census authorities tended to view more favourably than petitions on behalf of the whole 14

The male literacy rate of the Julaha weaving caste, whose name the petitioners claimed, was 2.2 per cent

Caste in the Census of India

109

group) (Blunt 1912: 323). Only in 1924 would a caste sabha claiming to represent the whole caste form and file a petition on behalf of the whole caste at the 1931 census (Lynch 1969; Pai 2002). By this time, the number of literate males in the group had tripled (to 0.6 per cent, at least 16,000 men). Moreover, even in 1931, the leaders of Chamar activism were drawn from a tiny segment of the group in western UP that had settled in urban areas and made money in the shoe business (Lynch 1969).15 In many poor groups with similar aggregate levels of literacy but without such a spatially concentrated educated elite, this type of activism would have been unlikely. The jati-province with the highest literacy rate that petitioned was the Kayasths of the UP – in fact, the group petitioned every year after 1911. This reflected a high underlying level of caste activism: UP Kayasths established a series of sabhas (the provincial Kayastha Deputation and the national Kayastha Conference), and a caste history claiming Kshatriya status was published at Lucknow as early as 1877 (Prasad and Dusre 1877; C. A. Bayly 1971). In addition to their activities in the census, UP Kayasths expended considerable effort in trying to have their claims recognized by the law courts, with indifferent success (Carroll 1978, 1979). The comparison with the non-petitioning Kayasths of Bihar is instructive. Given the political situation they faced, UP Kayasths could not assume that a non-caste based form of political competition would automatically favour them – unlike in Bihar their high literacy rates did not give them an educational monopoly. In UP, Kayasths were not the only educated group, and faced strong competition from the numerically large Brahmin population. At the 1911 census, there were approximately similar numbers of English literate Brahmins as Kayasths in UP, but 63 per cent more in Bihar and Orissa, while the disproportion of population was 10 Brahmins to every Kayasth in UP and only 3.75 to one in Bihar and Orissa.16 As Carroll’s (1978) description of Kayasth mobilization shows, this more competitive political environment was the major motivator for caste activism. As the introduction of the 1909 edition of Prasad and Dusre’s caste history puts it, ‘Government service formerly very much sought after 15

16

Of the six districts where Chamars filed petitions in 1931, five were in a clump along the state’s western border (Turner 1933: 530). This figure includes the related Oriya Karan caste with the Kayasths, and the literacy figures includes only districts where caste-specific figures were tabulated.

110

From Hierarchy to Ethnicity

by the Kayasthas has opened its doors to everybody from the Brahman downwards to the vilest bhangi or chamar and, as rival claimants for posts, they are interested in breaking up this monopoly by vilifying Kayasthas’ (quoted in Carroll 1978: 235). Kayasth activists in UP were not slow to trace attempts to oppose their activities to change their status in the census or the law courts to the jealousy of Brahmins of Kayasth success under the Mughals, often manifested in the form of Hindu revivalist movements (Bellenoit 2017: 177).

Conclusion Colonial caste activism was a phenomenon with social and economic causes. Rising levels of education created the elite that makes caste mobilization possible, but higher levels of wealth and education made it unnecessary. This non-linear pattern supports the idea that socioeconomic status influences identity not just by empowering political actors, but by altering the relative benefits of different forms of political actions. While the results shown here generally support the common theory that population size has a strong effect on identity mobilization, they supplement it with an understanding of the economic conditions that make such mobilization possible and desirable for elites. As such, the result belongs squarely within the tradition of theories discussing the effects of economic modernization, though it modifies these theories by showing that the effect of social change on groups is decidedly non-linear, and does not involve the fading of ‘primitive’ identities like caste. Just as striking is the sizable variation in the extent of mobilization in the colonial period, which serves as a corrective both to theories of a static caste system and to theories that emphasize the role of the colonial state. While some caste identities were politically salient, others, particularly at the extreme ends of the social scale, did not become the basis for political action until after independence.

5

The Causes of Ranked Rhetoric The degree to which identity activism embraces ranking is even less understood than the origins of activism itself. In particular, the self-denying nature of ranked identity mobilization is difficult to square with existing theories of what an ethnic identity should look like. The professed goal of ranked petitioners, the complete assimilation of their own group into the higher status category, seems to fly in the face of the incentives proposed by constructivist theories of ethnic politics such as Posner (2004) and Chandra (2004). This chapter describes some basic patterns in the distribution of ranked petitioning, particularly its relationship to the local political environment and the potential position of the groups within that environment. The set of cases on which we can test these hypotheses quantitatively is a limited one. In order to choose between ranked and unranked rhetoric, a group must first choose to mobilize. In the case of the colonial census data, this means that the analysis is confined to the set of groups that had already petitioned. These cases are analysed within the context of a hierarchical logistic regression model, which takes into account the previously made choice to petition. The results show that unranked rhetoric, and the more ‘ethnic’ vision of identity which it implies, is a product of the position of groups within the regional political system. Landowning groups and groups with limited contact with the state bureaucracy tended to emphasize hierarchical themes, much as their ancestors had done in pre-colonial times. Large groups in areas with democratic elections, by contrast, did not emphasize hierarchy, prefiguring the type of non-Sanskritic caste politics that became more common after independence. These results are then illustrated by examples of castes that, depending on political circumstance, either remained committed to hierarchical rhetoric or transitioned away from it. These results are obviously specific to the Indian context: no other country has such an elaborate set of ideas to order social groups, and

112

From Hierarchy to Ethnicity

the specific vocabulary of the varna system is unique to South Asia. However, this chapter shows that both the use of ranked rhetoric and the association between unranked mobilization and political participation are found in other cases as well. Analysing the modern historical experiences of Brazil and the United States, it shows that in the nineteenth and early twentieth centuries ascriptive identities often were influenced by ideas of ranking – frequently in ways that would seem strange to those familiar with ‘modern’ ideas of race and nation. However, in all these cases, increases in political participation were associated with the abandonment of ranked rhetoric and the movement towards ‘ethnic’ concepts of social difference.

Ranking in the census data Types of petition As noted in Chapter 2, there is no sharp disjunction between ranked and unranked forms of rhetoric, and in early colonial India, upwardly mobile groups deployed a wide assortment of ideas, many of which reflected some level of engagement with ranking. One behaviour, however, is so characteristic of ranked systems and so uncharacteristic of unranked systems that it will serve as the basis for making a quantitative distinction. This is the complete or partial denial of membership of the low-status group, and instead claiming to be part of the high-status group. In this situation, lower caste elites are so focused on potential status gains that they refuse to recognize a category difference at all between them and the high-status group, denying that their category exists, and making further attempts to develop an unranked identity based on this category becomes much more difficult. To measure the incidence of these self-denying caste elites, I divided petitions that were submitted into those making ranked and those making unranked claims, based on the name claimed in the petition, and associated each petition with a particular caste as defined by the census authorities. I classified as ranked those petitions in which the claimed name included either the name of a specific higher status caste (for example, ‘Saini Rajput’) or the name of one of the higher varnas, which therefore made a claim to higher status within the caste system at the expense of asserting a completely separate identity. This could include either claiming complete identity with the varna (asking the census authorities to use the term ‘Brahmin’ instead of the existing caste name) or using the varna name as an adjective modifying the caste name (for example, ‘Bhumihar Brahmin’). By tying

The Causes of Ranked Rhetoric

113

themselves to the traditional Sanskritic hierarchy, these petitions emphasize improving group status in the ranked system over their own independence. Census authorities often treated these ranked petitions as a joke – how could members of poor occupation castes hope to be confused with Brahmins or Rajputs? However, these claims reflected the reality of life in India up until the twentieth century: a wealthy and persistent group could slowly gain recognition as members of a higher caste through behavioural changes and strategic marriages. Ranked petitions reflected a desire to either use the census to advance such efforts or stop the census from undermining them. The 1921 Bombay census report notes just such a process of creeping identity change: The inhabitants of two villages in Broach district wrote protesting against being entered in the schedules as kolis of Gujarat and claiming to be rajputs and mentioning among other things that their marriage market had been spoiled. Quite possibly, had there been no caste column in the census schedules, these villages would have been known and accepted as Rajput villages in a comparatively short time. (Sedgewick 1922: Appendix C, x)

The petitions coded as unranked are those that make no such claims to belong to another caste or varna. As such, they are more heterogeneous than the ranked petitions. Most commonly, they reject a previous name thought to be insulting or associated with a traditional occupation considered degrading. Frequently, these new names also had positive connotations: The Goalas, for instance, claimed the name Yadava, which implied descent from Yadu, a mythical king. In other cases, formerly untouchable castes sought to be consolidated with others under the ‘Adi’ prefix, indicating that they were the original inhabitants of the country. What they have in common is that they reject the traditional social position of the group without attempting to join or associate themselves with the traditional elite. This does not mean that the unranked petitions reject the hierarchical worldview entirely. While a few contain robust denunciations of Brahmin domination, many of the few available in the National Archives of India are shot through with Sanskritic rhetoric, and boasts about caste ‘cleanliness’. Many of these petitioners no doubt sought to improve the position of their own caste without destroying the system as a whole (Jaffrelot 2000). However, by envisaging their group’s existence (and high-status) outside the three highest varnas, the unranked petitions refused to recognize the ‘monopoly of honour’ of the upper castes, and did not seek to merge with them. These ideas

114

From Hierarchy to Ethnicity

prefigure the varna-less justifications for caste status that have become common after independence (for example, Michelutti 2004). Table 5.1 shows the distribution of types of petitions by year. All types of petitions increase over time, and ranked petitions are always more common. However, the proportion of unranked petitions increases over time, parallel to the changes that were altering Indian politics in these years. Table 5.1 Disaggregated types of caste petition by year 1901

1911

1921

1931

Total

No Petition

134

239

267

228

861

Ranked Petitions Ranked Petition Ranked Petition by Subgroup

17 5

33 10

50 2

65 3

165 19

Unranked Petitions Unranked Petition Unranked Petition by Subgroup Untouchable Caste Seeking Merger

2 3 0

9 2 0

9 1 2

17 2 10

37 8 12

Other Petitions Merged Untouchable Caste (‘Adi-’) Conflicting Petitions Previously Granted Petition

0 0 0

0 2 0

1 2 0

3 12 1

4 16 1

161

294

334

341

1130

Total

Source: Author’s calculations from Census of India, various years. Note: The unit of observation is the province-jati-year. The number of province-jati-years varies from year to year because some census superintendents did not record the petitions that were submitted to them.

Measuring institutions Two key concepts discussed in Chapter 2 were the degree to which political power was exercised through formal bureaucratic means and the degree to which people were actively involved in politics. Both these concepts are fairly abstract and difficult to measure by quantitative means. Both, however, exhibited considerable variance within colonial India. The colonial state was neither a patrimonial successor of the Mughal Empire nor a modern nation-state, but a peculiar hybrid. In isolated rural areas and in the indirectly ruled princely states, local elites collected taxes and administered justice very much as their ancestors had done, while in the cities and towns of the directly ruled areas there existed local elections and mass-based political movements every bit as organized as those in contemporary Europe. This variance provides an opportunity for studying

The Causes of Ranked Rhetoric

115

political identity in a context where the ‘pre-modern’ and ‘modern’ were not chronological opposites but existed in uneasy proximity. No other period in Indian history exhibits an equal degree of variation in political participation and institutional form. To measure the relative exposure of castes to patrimonial and formal state institutions, I use two measures drawn from the occupational section of the census. The first is the percentage of males in a caste cultivating land or owning land, which should be associated with close links to patronage systems and weaker links to formal institutions. This is in part a function of the spatial distribution of rural groups, since even strong states have difficulty in projecting themselves in rural areas. More generally, rural areas tend to be affected more slowly by social changes than urban areas, and thus more likely to preserve older patterns of patronage, political culture, and obligation. Finally, control over land in colonial India involved a cultivator in an intricate network of social obligations, both to the labourers who did the actual work and to the network of intermediaries (bureaucratic, aristocratic, or village-level) who collected taxes and administered justice in rural areas. The management of these obligations involved linking oneself to patrons and the development of clients, and descriptive accounts of this period emphasize the importance of patron–client ties among rural Indians (Stokes 1978). Land tenure arrangements in colonial India were complex and varied not only between princely states and directly ruled provinces but also between provinces in which large landlords held revenue rights and those where these rights were vested in the cultivators or the village community (Banerjee and Iyer 2005). While the former system had once made village-level elites less secure, by our period colonial tenancy laws had made the larger tenants more autonomous everywhere (Stokes 1978). However, to ensure that these regional differences are not driving the results, all the main models use province fixed effects. To measure the direct exposure of the caste to formal state institutions, I use the proportion of male workers employed in the army and police, taken from the census. This was chosen rather than public employment as a whole because the majority of people classified as ‘public employees’ by the colonial census were village servants, who were typically closely associated with the patronage networks of village elites.1 Since in traditional theories 1

While these two subcategories are separated for India as a whole, they are not separated within castes.

116

From Hierarchy to Ethnicity

(for example, M. Weber 1947) the spread of rationalized bureaucracies is associated with the decline of traditional political linkages, we should expect groups with strong links to the bureaucracy to be the first to find patrimonialism irrelevant. In other words, in the context of a modern state, groups with many bureaucrats possess a much more viable alternative to traditional social networks than landed and uneducated groups do.2 Politics under the colonial state was always a limited business, as the most important decisions were made by European bureaucrats. There was, however, a small but expanding level of autonomy granted to local electoral institutions and an expanding sphere of protest against colonial rule. I use two measures of participation, the spread of democratic elections and the spread of the Congress Party. Neither of these proxies captures democratic participation in the full sense: the franchise in colonial India was limited to property owners, and the participation in the nationalist movement was most common among elites. Both, however, represent opportunities for new elites to challenge political arrangements based on heredity and wealth. The presence of electoral institutions is measured by the percentage of elected members of district and subdistrict boards, collected from the government’s Report on the Working of the District Boards for each province-census year. These boards were local entities, primarily concerned with sanitation, roadbuilding, and public education. Originally, district board members were appointed by the district collector. Gradually, however, the bodies expanded in size and began to include elected members, first as a minority and then as a majority. Despite their modest powers, membership in the only elected offices available to Indians conferred considerable prestige, and many future national politicians served at this level. The timing and extent of the change from elected to appointed boards varied greatly across provinces and appears to have been driven in part by local-level demand. The British were confident that granting elected seats would diffuse nationalist demands and coopt their leaders, and thus tended to expand 2

The groups employed by the colonial state are decidedly not a random sample of castes, since colonial policy favoured certain specific, often high-status, castes in military and bureaucratic recruitment. For this reason, the primary specifications include controls for caste status. It should be noted, however, that petitioning was uncommon among the ‘martial races’ and upper castes, most of whom did not petition at all. There are no petitions in the sample from the groups usually identified as ‘martial’ by the colonial state, and upper castes make up only 6.8 per cent of petitions.

The Causes of Ranked Rhetoric

117

elections in areas where there existed a politically aware Indian elite. Due to this complex causal pattern, the effect of electoral institutions should be interpreted very broadly, as an indicator of participatory politics. To measure participation in the nationalist movement, the models use the province’s average per capita attendance at the annual sessions of the Indian National Congress, collected from the Congress’s annual reports. At this stage of Indian history, the Congress was less a political party than a discussion forum for politically aware Indians, and it admitted virtually any well-dressed individual who wanted to attend. Despite its elite character, Congress attendance thus should be a reasonable proxy for the number of people in a province committed enough to the nationalist movement to invest a week of their time and a long train journey. In fact, the annual attendance figures seem to tally relatively well with trends in a political activity described in the secondary literature. Bengal, for instance, sees a peak in attendance around 1911, during the political turmoil surrounding the partition of the province.

Results Did the decrease in patrimonialism and the decline of participation discussed in Chapter 3 lead to a decrease in ranked petitioning? Table 5.1 certainly suggests that it does – the increased incidence of provincial elections is temporally correlated with the rise of the relative incidence of unranked petitioning. Moreover, this rise was more marked in some areas than in others (Table C.3). In particular, there was interesting variation between the various princely states, with the more ‘progressive’ princely states that had allowed legislative elections (Mysore, Travancore) having relatively high rates of unranked petitioning, while the more authoritarian princely states had much lower rates. Rajputana and Hyderabad, in fact, did not have a single unranked petition throughout the entire period. The more detailed predictions of Chapter 3 are tested by a set of regression models of petitioning groups, discussed in detail in the appendix. It is important to note that these models are not estimating the probability that groups will petition the government. Rather, they estimate the probability that already mobilized caste identities will emphasize ranking in their rhetoric, conditional on having already petitioned.3 Like the models in 3

Elites may be involved in politics even if their identity remains unmobilized – the Indian National Congress, for instance, built a durable

118

From Hierarchy to Ethnicity

Chapter 4, these models control for many of the more obvious confounding effects, including year, province, ascribed caste status, and group size and literacy. The results of these models provide limited support for the influence of patrimonialism on activism. Group land control has a robust and statistically significant positive association with ranked petitioning. The effect of land control is fairly large substantively: holding all other variables at their means, increasing a caste’s percentage of cultivators from 10 per cent to 20 per cent would increase the probability that a submitted petition is ranked from 46.7 per cent to 57.1 per cent. This pattern is clearly visible in Figure 5.1, which shows the distribution of both unranked and ranked petitioning groups relative to their level of land ownership.4 Group employment in the army and police has a more negative association with ranked petitioning, which is statistically significant in some models but not in others.5 The models also provide even stronger support for the relationship between elections, large groups, and petitioning. Both local board elections and the presence of Congress activists are positively associated with unranked petitioning. However, this effect is only present among large groups: in fact, the direct effect of local elections on the petition type (ignoring group size) is tiny and statistically insignificant. By contrast, the effect among large groups is substantively massive: holding all other variables at their means, increasing the proportion of local board seats elected from 20 per cent to 50 per cent (a typical increase for the 1920s) would cause the predicted probability that a group with 10 per cent of the population would emphasize ranking in their petition to decrease from 75.4 per cent to 24.9 per cent.

4

5

political coalition with significant lower caste support without using explicit caste rhetoric or conceding significant authority to lower caste elites. This chapter focuses on the rhetoric of those activists who remain outside such ‘vertical’ coalitions (Rudolph and Rudolph 1967) and choose to emphasize jati as a politically relevant identity. The smoothed relationship between the two in estimated non-parametrically using an Epanechnikov kernel function. The inefficiency of these estimates stems from the skewed distribution of employment in the police and army. Only 15 per cent of jati-province-years have a public force employment rate of greater than 1 per cent, and 56 per cent have a rate of less than two tenths of 1 per cent.

The Causes of Ranked Rhetoric

119

Figure 5.1 Density plot of ranked and unranked petitioning groups by land control

2

Landownership by Type of Petition

0

.5

Density 1

1.5

Unranked Ranked

0

.2

.4 .6 Landownership

.8

1

kernel = epanechnikov, bandwidth = 0.0903

Source: Author’s calculations from Census of India, various years.

Illustrative examples The relationship between education, politics, and identity is clearly apparent in the cases of the Bihari Bhumihars and the Tamil Shanans/Nadars, who both experienced major social gains in the colonial era. Their choice of ranked or unranked strategies reflected the political circumstances in which the groups were operating: in the early colonial era, this meant a ranked strategy for both groups, while in the late colonial era the rise of electoral politics in the Madras Presidency led the Shanans towards a more ‘ethnic’ approach.

The Shanan/Nadars The Shanans of Tamil Nadu were a large caste traditionally associated with the harvesting of coconuts and the production of toddy liquor, though in practice the majority of Shanans worked as ordinary tenant farmers. The caste was not considered a prestigious one in Tamil society, and Shanans shared with other lower castes a set of humiliating markers of their low ritual status, most notably a ban on Shanan women to cover their upper body. The Shanans were not a homogenous caste and may have originally formed at least five related jatis (Templeman 1996: 26–28).

120

From Hierarchy to Ethnicity

Despite the general poverty of the community, there existed a literate Shanan elite associated with tax collection and small trade. There even existed a wealthy sub-caste, the Nadans, who owned whole villages and adopted the manners of local aristocrats (Templeman 1996: 18–29). Like pre-colonial elites elsewhere in India, the Nadans strove to claim status within a ranked identity system, claiming descent from the ancient Pandyan dynasty, secluding their women, and patronizing local temples (Hardgrave 1969: 29–31). This elite used traditional caste panchayats to monopolize dispute resolution within the caste, and at times to try to intervene in the affairs of other castes as well (Templeman 1996: 82). These institutions were funded by a private tax system, the mahamai. This elite grew in size in the nineteenth century, as the growth of the south Indian economy improved the position of tenants, and Christian missionaries expanded access to education among traditionally poor groups (Hardgrave 1969: 43–55). While some escaped the most hated aspects of the group’s low ritual status by conversion to Christianity, on the whole, the social aspirations of the Shanan elite were channeled into attempts to gain a more prestigious position within the caste system through traditional means. This process often turned violent, most notably in the Upper Cloth Controversy of 1857–1858, when Shanan women in the state of Travancore won the right to cover their upper bodies, and the Sivakasi riots of 1899, where the demands of wealthy Shanans to be admitted to temples led to violence between them and the members of the Maravan caste (Sobhanan 1985). The riots were followed by a contentious lawsuit in which the Shanans attempted to assert their case to ‘clean’ status in court. Two notable aspects of this early mobilization were its reference to traditional Hindu norms and its refusal to engage with poor Shanans. In the minds of the Shanan elite, they were a class of warriors who had been degraded by contact with toddy cultivation, and their primary goal was to assimilate themselves to the practices of upper caste Hindus. They petitioned the census authorities to have non-toddy-tapping Shanans listed as Kshatriyas or to list the Nadans as a separate caste entirely (Hardgrave 1969: 109). Other attempts to improve the social status of the Nadar community involved the adoption of normatively desirable Hindu practices such as the wearing of the sacred thread, the abandonment of toddy tapping, and the feeding of Brahmins. Wealthy Shanans also conducted ostentatious weddings involving being carried on palanquins (symbolic of their claims to royal status). When the government of Madras

The Causes of Ranked Rhetoric

121

listed the caste as ‘depressed’ and thus worthy of receiving affirmative action benefits, Shanans petitioned the authorities to have them reclassified as ineligible. Predictably, the Shanan elite disdained their toddy-tapping coethnics. Hardgrave (1969: 106) notes: ‘These nadars held themselves superior to their caste fellows in Tinnevely. Intermarriage was prohibited and interlining discouraged.’ The activist’s main contact with poor Nadars were adversarial, as their attempts (enforced by violence and kangaroo courts) to ban the sale of toddy (Hardgrave 1969: 137). Although such a ban would have improved the status of the groups as a whole, it would have deprived many poorer Shanans of their livelihood. This Sanskritic and ranked concept of Shanan elite identity eventually became obsolete, mostly through the efforts of an organization founded in 1910, the Nadar Mahajana Sangam. While the Sangam leadership was originally identified with the Sanskritizing agenda, it quickly came to be dominated by new leadership, led by men such as E. V. Ramaswamy Naicker and W. P. A. Soundrapandian, dedicated to the creation and social uplift of a united Shanan community under the name of Nadar. This was coupled, especially in the twentieth century, with the development of a massive private social service network which included not only scholarships and aid to the needy but a well-capitalized bank and one of India’s first private universities (Hardgrave 1969: 130–172). The Sangam repeatedly petitioned the colonial authorities, eventually with success, to ban the name Shanan, which it considered derogatory, and replace it with Nadar. Relatedly, they reached out to poorer Nadars and began to encourage inter-dining within sub-castes. Perhaps the most visible aspect of the transformation from external to internal forms of cultural legitimation were the changes, strongly promoted by the Sangam, in Nadar wedding ceremonies. While elite Shanan weddings had previously been ostentatious three-day affairs officiated by Brahmins, the Sangam encouraged ‘self-respect weddings’ taking only a day and officiated by a senior association member (Templeman 1996: 72–73). Similarly, the sikha topknot hairstyle, once associated with Brahmins, was also discarded. Both moves echoed larger trends in southern India in this period towards the rejection of Brahmins and ‘northern’ cultural influences (Irschick 1969). In this period, the Nadar elite also became a key player in Tamil politics. Indeed, the transformation of the Sangam’s strategy was exactly contemporaneous with the growth of Soundrapandian’s political career.

122

From Hierarchy to Ethnicity

The Sangam leader was a member of the Madras Legislative Council from 1920 to 1937, and an elected district board member in 1928–1930 and 1943–1947. Soundrapandian based both his electoral appeal and his claims to appointive office on his status as the ‘representative’ of the Nadar community. Frustrated by the small size of the community for electoral purposes, Soundrapandian sought to extend the Nadar label to other unrelated toddy-tapping communities (the Gramanis and the Tiyars). While traditionalists defeated this proposal, it indicated the degree to which Soundrapandian understood that ‘the Sangam, representing a larger social base, would have a greater bargaining position with the government’ (Hardgrave 1969: 178). Soundrapandian also sought to build alliances across caste lines, becoming closely identified with the anti-Congress Justice Party, and the broader ‘non-Brahmin’ movement. This was associated with indifference to Indian nationalism: through most of the colonial period, a pro-British loyalty oath was required for membership (Hardgrave 1969: 172). However, this would change after independence, when Nadars switched their support to the Congress (Rudolph and Rudolph 1967). The Nadars remain an important voting block in Tamil politics. The Nadar story is remarkable for the ways in which the changing political situation in Madras led the group to adopt two very different strategies within a short period. Hardgrave (1969: 179) summarizes the contradiction: During the later part of the 19th century, the Nadars, in emulating Brahmins in the process of Sanskritization, accepted the tyranny of the caste system: they desired not to overthrow it, but to secure a higher position in its ranks. Soundrapandian, reinforced by E. V. Ramaswamy Naicker, now sought to overthrow the system, to deny the legitimacy of ranking altogether, and to oppose the Sankritization of Nadar life and custom.

The experience of the Nadars can be contrasted with the complete absence of mobilization among other, poorer, low-status south Indian groups. The Madigas, the traditional leatherworking caste of coastal Andhra, suffered even more extreme forms of discrimination in early colonial India than the Nadars. Considered untouchable, they were forced to perform degrading village chores like the preparation of latrines and the burial of the dead, while earning their living as the dependent labourers of upper caste landlords. The social changes of the nineteenth century were also less kind to the Madigas than they had been to the Nadars. The Andhra region

The Causes of Ranked Rhetoric

123

was less heavily influenced by Christian missionaries than Tamil Nadu, its agricultural sector substantially less advanced, and its cities smaller and poorer. The result was that an elite group never emerged among the Madigas as it had among the Shanans. At the 1911 census, only 0.8 per cent of Madiga men were capable of writing their name, and a caste of 808,000 people recorded exactly 24 lawyers, doctors, and teachers. With no educated group to speak on their behalf, the Madigas remained politically quiescent during the colonial period, neither forming a sabha, petitioning the census authorities, nor adopting Sanskritic behaviours. Only in the 1970s and 1980s, when a class of literate Madigas had finally emerged, did Madiga identity become politically important.

Babhans/Bhumihars/Brahmins The Bhumihars are a traditionally agrarian group common in Bihar and Uttar Pradesh. The Bhumihars have traditionally been associated with agriculture, and many Bhumihars are landholders in this region. Despite this generally strong socioeconomic position, the social status of the Bhumihars was traditionally somewhat ambiguous, as they were considered by many upper caste groups (notably by Brahmin priests) to be no more than wealthy peasants. Like the other case study groups, Bhumihar political participation has been marked by their economic and political circumstances. Their high but not extraordinary levels of education have encouraged caste activism, while the consistently unstable and patrimonial nature of politics in Bihar, both before independence and after, has kept this participation strongly hierarchical in tone. Like the Jats discussed in Chapter 4, the Bhumihars gained upward mobility through the ranks of the zamindars, taking advantage of the flux and rural militarization that accompanied the Mughal conquest (Cohn 1969). In the eighteenth century, many of the major zamindars in western Bihar and eastern UP were Bhumihars, including the most powerful local ruler, the raja of Benares. These estates were some of the largest landholdings in India, and their zamindars resembled petty kingdoms. In addition to their role in revenue collection, the Bhumihars were prominent in the military sphere, both as soldiers of the regional power, the nawab of Oudh, and in the army of a rising military force in Bengal, the East India Company (C. A. Bayly 1983: 17–19). In both these cases, the rise of firearms and western drill created an opening for new military groups, as the small cavalry armies of preceding years became obsolete.

124

From Hierarchy to Ethnicity

Bihar’s distance from major centres of Sikh or Muslim influence meant that the religious exit options that were available to the Jats were inaccessible to the Bhumihar elite, who thus concentrated their ideological efforts on the cultivation of a high-status identity within the Hindu caste hierarchy. At the courts of the Bihari zamindars, rulers evolved patterns of court life and religious behaviour that emphasized their cleanliness and the links between their caste and Brahmins. Indeed, a primary feature of the ideology of elite Bhumihars in this period was their claim to actually being Brahmins who had taken up agriculture. By claiming Brahmin status, Bhumihars not only denied the claims of Brahmin and Rajput elites to their deference but also emphasized their ritual superiority to Yadavs and Kurmi cultivators who were their main potential rivals for influence at the village level (A. Chakravarti 2001). Bhumihars were well represented in the armies of the East Indian Company, but the British banned them from the colonial army after their conspicuous part in the 1857 rebellion. In addition, the general poverty of this part of rural India meant that relatively few Bhumihars were literate until the late nineteenth century. When a literate group of Bhumihars did emerge, their aspirations were very similar to those of the early Nadar elite – they sought to assimilate into the traditional upper castes. From the early twentieth century, Bhumihars attempted to become known as Brahmins, petitioning the census authorities to merge the two groups, and when that failed, pressuring census enumerators to list them as Brahmins on an individual basis. Bhumihar activists remained convinced that seeking the favour of the British was advantageous in this fight, a fact reflected in the group’s rhetoric. The Bhumihar Brahmans are that portion of the great Brahman caste which does not practice the occupation of Priesthood. . . . Nearly all the big zamindars of Bihar, as for instance Bettiah, Hatura, Likhari, Sheohar, Norhan and Modhurban, are Bhumihars. [By enumerating Bhumihars as a sub-caste] the government will also be in a position to know much more about one class of its subjects which still enjoys the greatest territorial possession in Upper India and whose annual income is still very considerable, and whose reputation and name were never tarnished by the fiendish designation of a rebel.6

6

National Archives of India, Home Department, Census Branch, A Progs, December 1900, Nos 24–27.

The Causes of Ranked Rhetoric

125

Besides establishing a school and promoting ‘unity, brotherly love, and mutual help’, The sabha’s major goal was ‘to introduce social and other reforms’. In practice, this meant the promotion of behavioural changes designed to eliminate the cultural gap between themselves and Brahmans, such as banning the remarriage of widows. This attempt at cultural assimilation continues to this day. Despite the growth of caste-based affirmative action policies that have made Brahmin status something of a liability in government employment, Bhumihar activists remain convinced that their group is indistinguishable from Brahmins.7

Ranking without Sanskrit Is social ranking, and its association with institutional factors, unique to South Asia? This section discusses several cases where identities strongly emphasized hierarchy over division. It also suggests that the development of these systems was associated with patrimonial political systems, and that their decline coincided with the rise of participatory politics. The forms that these identity changes took, from census reclassifications to the forming of associations, are strikingly similar to the Indian case.

Racial identity in Brazil Racial identities have a long history as an ‘activated’ identity for social and political purposes in the politics of many New World countries and are often used as examples in books on ethnic politics (for example, Chandra 2012). At first glance, racial identities are striking in their pure expression of the logic of unranked mobilization. In conventional descriptions of race in the United States, one can be (for example) white or black, and the boundary between them is thought to be rigid and associated with a sharply discontinuous change in behaviour, identification, and social expectations. While assertions of racial superiority are part of the legacy of the system, these assertions are strongly contested by members of all groups. Far from considering ‘white’ behaviour an object of emulation or white people a source of legitimation, members of other racial groups may disparage ‘acting white’ (Fordham and Ogbu 1986; Austen-Smith and Fryer 2005). 7

During the author’s fieldwork in Bihar, many informants told the author that Bhumihar was a synonym for Brahmin in the local dialect, and produced privately printed Bhojpuri–English dictionaries in support of this opinion.

126

From Hierarchy to Ethnicity

However, this portrayal does not apply to all societies in which race is salient. Brazil, for instance, is often thought to have a very different racial politics. Both Brazil and the southern US were shaped by plantation agriculture and black slavery, and both thus developed into societies where skin colour was an essential aspect of social identity. These societies differed, however, in the way in which these identities were expressed politically. Social usage in Brazil recognized intermediate categories between black and white such as ‘mixed race,’ ‘mulatto’, and ‘pardo’, categories defined not just by skin color but by social consensus. Harris (1964: 21–24) notes: One may speak in other words of an ideal or abstract racial ranking gradient in Brazil in which the white physical type occupies the favorable extreme, the Negro type the unfavorable extreme, and the mulatto type the various intermediate positions. . . . There are no racial groupings which can be divorced from other criteria of status. Roughly speaking, the calculus of identity involves adding up, as in a system of merits and demerits, a man’s education, his wealth, the kind of job he has, and the color of his skin and texture of his hair.

Telles (2002) also finds that interviewers tended to make racial classifications using social class, ‘whitening’ educated and well-dressed people of colour. As in the Indian case, many low-status individuals succeeded in improving their status through strategic changes of behaviour and association, often through the use of intermediate social categories such as mulatto. Ambitious black men might become mulatto, and mulattos sought to be considered white, rather than attempting to be the elites of the lower (black) status category (A. W. Marx 1998).8 While there was an elite consensus on the social value of white skin (and the set of ‘respectable’ 8

While race in Brazil has many characteristics of a ranked identity system, the dominant way to characterize Brazil’s racial politics has been negative and comparative, emphasizing the lack of entrenched racial division in Brazil relative to other parts of the Americas (Freyre and Putnam 1946; A. W. Marx 1998), rather than examining the internal logic of the ranked system. Marx’s account also differs from my own in tracing inter-American differences in identity to state policies, and the incentives of the ruling coalitions shaping those policies. As with other state-centred accounts of identity mobilization, his emphasis makes it difficult to access the causes of variation in identity strategies across groups exposed to similar state policies.

The Causes of Ranked Rhetoric

127

social behaviours associated with it), society was not divided into groups of whites and non-whites but organized along a spectrum, with large groups of racially mixed individuals striving to emulate the values and behaviour of wealthy whites (Degler 1971). At the social level, turn-of-the-century Brazilians sought to ‘solve’ their racial problems through the assimilation and ‘whitening’ of the population: one white Brazilian told Theodore Roosevelt: ‘You of the United States are keeping the blacks as an entirely separate element. . . . They will remain a menacing element in your civilization, permanent, and perhaps even after a while a growing element. With us the question tends to disappear, because the blacks themselves tend to disappear and become absorbed’ (quoted in Skidmore 1993: 75–76). As in India, the politics of ranked identity tended to complicate the politics of the census, leading to individuals and groups claiming higher status. The language of the Brazilian Institute of Geography and Statistics’ description of the 1940 census could come almost word for word from the pen of an Indian colonial census officer. In the declaration of color, departures from what would have been obtained had objective criteria been employed are conspicuous. The white grouping was swollen by the admission of numerous unqualified aspirants to that category. In the case of pretos and pardos, there was much confusion. (Quoted in Harris 1964: 22)

These differences have led some to conclude that race is less salient in Brazil than in the United States (Freyre 1970; Degler 1971). This view has in turn been criticized for understating levels of formal and informal racial discrimination (Fernandes 1969). More broadly, even if boundaries were weak, racial hierarchy was strong: nobody questioned which group was superior or which group’s values would dominate in any eventual process of social assimilation. Harris (1964: 21) remarks that ‘Negroes throughout Brazil are abstractly regarded as innately inferior in intelligence, honesty and dependability; that negroid features are universally (even by Negroes themselves) believed to be less desirable, less handsome or beautiful than caucasoid features’. If these descriptions fit Brazil in the nineteenth and early twentieth centuries, they are less accurate now, given ‘the sudden and dramatic changes in Brazilian race thinking’ that occurred in the twentieth century (Telles 2014). During the middle of the century, a black rights movement, the Movimento Negro, sought to encourage the development of ethnic

128

From Hierarchy to Ethnicity

consciousness among Afro-Brazilians and establishing a set of cultural and philanthropic institutions within Afro-Brazilian communities (Hanchard 1998). While the celebration of Afro-Brazilian culture was in itself a rejection of the hierarchical system, these activists were also explicit in their rejection of racism and white dominance (Htun 2004; Telles 2014). The most important achievement of this movement has been the creation of a set of affirmative action programmes for Afro-Brazilians, a policy that ‘sounded preposterous and highly unlikely just a few years ago’ (Telles 2014). A key part of the Black Movement strategy was to expand the use of the ‘Black’ census category. One activist stated: ‘We see that for every 100 negros, 70 reject their identity. So we must convince the negro to reject the ideology of “whitening.” And the way to start is for him to call himself a negro’ (quoted in Burdick 1992). Nobles (2000: 129–162) shows how organized activist campaigns were able to both change the census’s approach to racial classification and individuals’ responses. The theory in this book provides a simple explanation for these changes. Before the mid-twentieth century, racial status in Brazil illustrated the main characteristics of a ranked system. Identity was very strongly associated with social class: it emphasized subtle social distinctions over hard and vast differences, it created elaborate internal differences within ethnic categories, and it tended to lead to the subordinate group defining themselves through the ideological tools of the super-ordinate group. As in colonial India, people who may have been classified as part of the subordinate group in other societies often claimed membership in the super-ordinate group. As in India, differences in identity rhetoric were associated with differences in the institutional environment. Racial identity in Brazil developed in a context where national political power was exercised in an unapologetically authoritarian manner, at first by the Portuguese colonial state and later by a series of regimes dominated by urban oligarchs. At the local level, power was held by local bosses, the so-called Coronels, who kept control through a combination of violence, clientelism, and voter fraud (Roniger 1987). These political powers reflected the dominance of the rural economy by large planters, first of sugar and later of coffee (Dengler 1971). Since all these powerholders were white, it was imperative to keep on good terms with members of the white elite. Through a similar process, the gradual democratization the political system in Brazil was roughly contemporaneous with the development of an unranked system. The early years of the Movimento Negro began during

The Causes of Ranked Rhetoric

129

the expansion of political participation and disruption of rural elite rule caused by the quasi-fascist dictatorship of Getulio Vargas (1930–1945) and the subsequent democratic governments. As in India, the census served as an important forum for publicizing and legitimating both the old ranked identities and the new unranked ones. The massive expansion of racial consciousness in recent years has been contemporaneous with the restoration of democracy since 1988.

Racial identities in the United States In the previous section, the United States was cited as an example of a society with a highly segmented set of racial identities. In America, especially in the twentieth century, racial divisions are not perfectly predictive of class status, are relatively rigid, and are attended by rhetorical pledges of racial solidarity and racial independence. This pattern, so very different from traditional caste system, was discussed by Verba, Ahmed, and Bhatt (1971: 21–22), who note that while blacks usually occupy subordinate social positions and are the subject of negative stereotypes, this is in contradiction to the general ‘egalitarian’ shape of the political culture, and is vehemently rejected by African-Americans themselves. The unranked nature of racial identity in America fits well with the theory described in Chapter 2. In the United States, political identities developed in the context of a robustly democratic political culture, both at the local and state levels. African-Americans cannot be counted out as potential competitors. The result was the development of one of the world’s most rigidly bounded identity systems, with a very small amount of African ancestry necessary to place a person in the African-American racial category. While whites frequently assert their superiority over blacks, educated African-Americans usually do not seek to become white. However, this was not always the case. Racial categories in nineteenth-century America included elements that would seem strange today. First, the black–white division was not always a binary: the US census recognized the third ethnic category ‘mulatto’ from 1850 until 1930, and in 1880 included the additional intermediate categories of ‘quadroons’ and ‘octoroons’ (Hickman 1997; Hochschild and Powell 2008). This was reflected in local practice in some areas: in New Orleans, the Creole group included many wealthy ‘Gens de Couleur’ who possessed both a good social position and African descent, leading to considerable ambiguity in

130

From Hierarchy to Ethnicity

category membership and the possibility that wealthy and respectable Gens de Couleur could be accepted as white. As late as 1907, Desdunes (1907) noted the differences between the more ranked Creole approach and the more rigid and segmented approach he associated with Anglo-black activists like W. E. B. Dubois. Thus we often perceive that one makes every effort to acquire merits, the other to gain advantages. One [Creole] aspires to equality, the other to identity. One will forget that he is a Negro in order to think that he is a man; the other will forget that he is a man to think that he is a Negro.

Even outside of New Orleans, some educated African-Americans sought to ‘pass’ as white in this period (Ginsberg 1996), and there were many instances of attempts at internal differentiation among African-Americans based on skin colour and adoption of ‘white’ cultural practices (Russel, Wilson, and Hall 1992). This more ranked approach to race was associated with a less participatory approach to politics in the southern United States. Most blacks were enslaved before 1865, and even after the Civil War violence and fraud kept most blacks from participating in politics (Key 1949). As in Brazil, planters, landlords, and local political bosses held considerable power, and the ambitious sought ties to them. However, the possibility of racial ambiguity gradually faded with the patrimonial power of the planters. Interestingly, this segmentary tendency became even more notable in the aftermath of the Civil War, which by formally enfranchising blacks made the system far more participatory, and the need to exclude black elites more immediate. When African-American elites did mobilize their community, they did so in the knowledge that this community could potentially form a voting bloc.

Conclusion While limited, these results provide some preliminary evidence that there is no single way for a social identity to be expressed politically. On the contrary, its expression is highly dependent upon the local political context, and a group’s place within that context. The decision to emphasize group difference over hierarchy seems dependent on the electoral advantages that can be gained from such a move: this type of mobilization strategy is

The Causes of Ranked Rhetoric

131

common among groups with a large share of the population situated in areas with participatory political institutions. The decision to invest in a ranked identity is guided by a group’s position within more traditional power systems: landed and non-bureaucratic groups are likely to find this option more attractive than groups less tied to the traditional patronage system. This set of theories offers a new way of understanding the development of caste politics in India. In the pre-colonial period, low levels of literacy and patrimonial polities kept lower caste mobilization at relatively modest levels and within traditional hierarchical norms. The spread of literacy in the colonial period, in the context of a state that was still pre-modern, led to an explosion of ranked mobilization (‘Sanskritization’) as groups attempted to use new wealth to achieve a traditional set of social goals. In the late colonial and post-independence periods, primary education became the norm and democratic elections became pervasive. The result was an expansion of unranked mobilization: lower caste elites switched from the cultivation of their social superiors to that of their coethnics, leading to the activation of politicized caste groups with well-marked external barriers. However, the persistence of patrimonial institutions in many parts of rural India means that ranked identities have not died out entirely, and it is notable that those areas of India (such as rural Bihar) with the weakest state institutions also tend to be the areas in which traditional conceptions of caste are most entrenched. The persistence of the relationship between hierarchical norms and state weakness in the contemporary world is a promising topic for future study. The trend away from social ranking is mirrored in the politics of the developing world more generally. As democratic norms have spread and patrimonial states have declined, ranked notions of identity have become increasingly unusual, replaced by the standardized norm of the differentiated, internally homogenous, self-legitimating ethnic group. As it has done so, it has become less common for the elites of subaltern groups to assimilate into groups above them. Instead, like the elites of India’s lower caste groups, they have embraced their identity and sought to lead the groups they were born into.

6

Caste since Independence At the time the colonial census officials finally began to accede to the demands of the caste petitioners, British rule in India was clearly nearing its close – indeed, much of the electoral jockeying of the caste associations was bent on filling the political space that the decline of the Raj opened up. India’s independence changed the formal attitude of state institutions towards the caste system from curiosity to strident hostility, and the political system from a curious colonial hybrid to a democracy with universal suffrage. Over time, the policies of the post-independence state would also change the distribution of wealth and education among social groups. The development of caste identities during these decades presents a paradoxical picture. On the one hand, it was clear that certain public expressions of caste identity that had been important in the colonial period were in precipitous decline. Elite opinion in India, at least in public, has turned against the caste hierarchy as a legitimator for social and political inequality, and even as a legitimate social affinity or object of intellectual analysis. The census no longer tabulates jati populations, both eliminating it as a focus for public contestation and dramatically limiting our ability to make reliable quantitative generalizations about this period compared to the one before. The practice of untouchability, once nearly universal, has undergone a precipitous decline, especially in urban areas. Lower caste candidates routinely win elected office, and upper caste voters routinely court lower caste support, which they know they cannot take for granted. The formal advertisement of claims to high caste status is also much less common than it once was; if anything, it is now more common for castes to claim to be ‘backward’. On the other hand, despite the hopes of some reformers of the independence generation, it is clear that caste has not withered away – in

Caste since Independence

133

fact, in some respects, caste identities appear more salient now than they were in the colonial period. During the 1990s, explicitly caste-based political parties became electorally successful in some northern states, and the controversy over caste-based quotas in reservation and public employment made caste a more noticeable element of political contention it had been in the 1950s and 1960s. Caste associations remain prominent, still dabbling in electoral politics and conducting campaigns for the social betterment of the caste. At the social level, caste remains an extremely important attribute of the personal identity of many Indians, structuring marriage, business relationships, and other forms of social cooperation. This chapter shows that these conflicting trends represent the natural culmination of the economic and political trends discussed in the previous two chapters. This period of Indian history saw the expansion of education within the context of a political system that allowed an unprecedented level of participation. As a consequence, India in these years saw an increase in identity mobilization, especially of previously immobilized poor castes, and a sharp decrease in the engagement of lower caste elites with ideas of ranking. Caste identities today thus bear a much closer resemblance to ascriptive identities in other parts of the world than they did in the colonial period: Their elites strive to lead bounded, conceptually autonomous groups to political glory rather than to assimilate into higher caste groups.

The rise of caste mobilization The critique of caste In the years immediately after independence, the colonial trends towards political and social mobilization on the basis of caste intensified. This was not the result of conscious state policy. The politicians who controlled the Indian state in the years after independence were, in fact, actively hostile to caste mobilization, and indeed in some cases to the existence of caste as an ascriptive division of society. Caste was at once derided as an outdated legacy of India’s past, doomed to be swept away by the forces of social modernization, and as a dangerous source of social division within an aspirational unitary nation, analogous to the religious ‘communalism’ that had recently led to the partition of the subcontinent. In The Discovery of India, Jawaharlal Nehru remarked: ‘Probably there is hardly anyone left in India who approves of [the caste system] in all its present ramifications . . . conditions of life have changed and thought patterns are changing so much

134

From Hierarchy to Ethnicity

that it seems impossible for the caste system to endure’ (Nehru 1989: 242). In a more expansive vein, he declared: ‘We must guard against being swept away by momentary passion, whether it is religion misapplied to politics or communalism or provincialism or casteism’ (Nehru 1958: np). It is easy to condemn these changes as either empty or hypocritical, and indeed some lower caste activists did so at the time (Ambedkar 2014 [1936]). While condemnatory of certain manifestations of caste discrimination, many senior Congressmen – with Mahatma Gandhi being the paradigmatic case – believed that a reformed or purified version of the caste system would serve a positive social function (Jaffrelot 2005). Moreover, many Congressmen continued to observe caste restrictions in their everyday lives and form political factions based on caste (S. S. Jha 1972; Brass 1984). Similarly, the reform and ‘uplift’ that Gandhi imagined for the lower castes was a top-down affair, one that was hostile to the idea of lower caste self-assertion, and the idea of lower castes as a separate community. This hostility to the creation of a bounded Dalit category motivated Gandhi in his successful 1932 hunger strike against separate electorates for Dalits, even as he was willing to condone separate elections for religious minorities. Lower caste politicians, therefore, were caught in a trap: if they attempted to mobilize support on the basis of caste identity, they were condemned as ‘casteist’ – then, as now, a term of political abuse in India. But if they ignored caste, they were implicitly accepting that most decisions would be made by upper caste individuals. Whether or not it was insincere, this change in rhetoric had several policy consequences. The Indian constitution unambiguously banned both public and private discriminations on the basis of caste and allowed the creation of an extensive system of reservations in hiring and education for lower caste groups (discussed later). While these edicts were often evaded (Galanter 1969; Shah et al. 2006), they had important symbolic consequences: the authority of the state could now not be invoked to reinforce caste ranking, as it had in colonial times. Another consequence, particularly important from a scholarly point of view, was a precipitous decline in the public dissemination of information about caste in general, and about jati in particular. Already, the wartime census of 1941 had eliminated the tabulation of caste data for financial reasons, though some provinces did tabulate jati populations on their own initiative. Post-independence censuses only collected data on the presence of members of specific vulnerable communities (the Scheduled Tribes [ST] and Scheduled Castes [SC]), and only tabulated data for these

Caste since Independence

135

aggregated caste categories.1 This reticence extends to other forms of public sector data collection: political candidates, for instance, are identified by caste category (usually only SC or ST) but not by jati, while the National Sample Survey similarly reports results by category.

Caste mobilization before 1990 Despite these obstacles, the lower caste political mobilization that had begun in the colonial era continued into the post-1947 period. In many cases, this took the same form that it had done during the colonial era: the formation of caste associations and attempts by these associations to influence the social and political affiliation of their members. Reflecting the patterns of colonial petitioning, this associational activity was strongest in the south and west, where literacy rates were highest. The Nadar Mahajana Sangam continued to be a major player in the politics of Tamil Nadu (Hardgrave 1968, 1969). The Vanniyars, while not abandoning their claim to Kshatriya status, turned the Vanniyakula Kshatriya Sangam into an active political force (Rudolph and Rudolph 1967: 29–60). In Kerala, the Nair Service Society, previously focused on the administration of schools, became the central player in the successful 1959 overthrow of the Communist Party government (Lieten 1982). But what was striking about this associational activity was the broad range of castes that were involved. While the political power and social reach of caste associations during the colonial period had been less notable in the north than in other parts of India, and among Dalits than among other types of castes, the post-independence period saw caste associations becoming more influential there too. In the well-studied case of the Chamars (Jatavs) of Uttar Pradesh, this activism built on a colonial history of caste organization and the expansion of education during the colonial period (Lynch 1969; Pai 2002). In the case of the very poor Bhangi caste of Delhi, the small caste sabhas that had existed before independence became much more organized and effective afterwards, organizing an annual festival (1950), an educational society (1957), and a trade union for municipal scavengers (1965) (Sharma 1995). In some cases, these associational efforts centred around a caste identity that had little or no social importance in the colonial and pre-colonial 1

A 2011 attempt to collect caste data by the census was mired in controversy, and the results were never published.

136

From Hierarchy to Ethnicity

periods. Perhaps the most famous of these efforts occurred in Gujarat, where a group of politically ambitious Rajputs decided to ‘merge’ their own caste with the Bhils (a tribal group) and the Kolis (a previously low-status cultivating and fishing group) (Kothari 1977).2 As one activist told Myron Weiner, ‘We have taken all the backward peoples who are martial by nature and called them Kshatriyas’ (Weiner 1967: 96–98). The Gujarat Kshatriya Sabha conducted a series of well-publicized banquets where members of the various component jatis ate from the same table, and energetically campaigned for Kshatriya candidates, usually in opposition to the Congress. Similarly, in Tamil Nadu, the Mukkulathor community emerged as a ‘federation’ of three small castes: the Agamudayars, Kallars, and Maravars (Rudolph and Rudolph 1967). As in the colonial period, caste associations were intensely involved with the electoral process. However, they had to confront the fact that the leaders of the Congress, incontestably the dominant party, were almost exclusively drawn from the upper castes, and in particular from Brahmins and members of the urban professional classes. Of the north Indian members of Parliament (MPs) elected in 1952, a full 28.7 per cent were lawyers, as against the 19.5 per cent who were agriculturalists (Jaffrelot/CSDS MP Dataset). Teachers, journalists, doctors, and civil servants were also well represented.3 At the same election, 66.7 per cent of legislators were forward castes (and 28.9 per cent Brahmin). These trends were even more marked in the higher levels of the Congress Party apparatus. In 1950, for instance, the upper castes (who comprised only 13.7 per cent of the population) comprised 75 per cent of the membership of the Bihar state Congress committee (Jaffrelot 2003: 73). Kayasths held 12.5 per cent of seats on the Congress working committee despite being only 1.2 per cent of the province’s population. These politicians, secure in their status as the generation who had won independence, were unwilling to make space for interlopers, particularly those professing a ‘casteist’ ideology. Consequently, many of the caste associations were forced to become involved in politics outside the umbrella of the Congress Party, a fact 2

3

Shah (1957) argues that the goal of these movements was to build larger electoral blocks for electoral purposes, and in fact, it appears that these identities are less relevant for social purposes than political ones (Shah 2010). This roughly paralleled the backgrounds of political leaders under colonialism. Of the delegates to the 1923 session of the Congress, 21.3 per cent had been lawyers (Krishna 1966: 422.)

Caste since Independence

137

that often worked to their detriment in electoral terms. The Republican Party of India (RPI), for instance, was a vehicle for Mahar activists in Maharashtra (where they won only 5.4 per cent of the vote in 1962) and Chamar activists in Uttar Pradesh (where they won 3.7 per cent of the vote in the same year) (Lynch 1969; Jaffrelot 1998; Jensenius and Verniers 2017). In Tamil Nadu, rival Vanniyar caste sabhas formed the Tamil Nadu Toilers Party and the Commonweal Party (Rudolph and Rudolph 1967; Raja 2017). Other caste leaders waffled over the idea: Kshatriyas in Gujarat backed the Congress until switching to the Swantantra Party in 1962 following the defection of Kshatriya Sabha leader Narendra Singh Mahida (Kothari 1970). In Tamil Nadu, the Nadar Sangam, supporters of the Justice Party during the colonial period, switched to the Congress at independence before declaring organizational neutrality in the late 1950s when some Nadars began to support the opposition Dravida Munnetra Kazhagam (DMK) (Hardgrave 1968). In some cases, lower- and middle-status castes were able to find a place in the Congress coalition, though typically without the explicit participation of caste associations. This was true for Vokaliggas in Mysore, Nairs in Kerala, and Jats in Rajasthan. While these politicians enjoyed much more electoral success than those in opposition, they were forced to deal with an entrenched upper caste elite. Jat leader Charan Singh, for instance, spent much of his career in the Uttar Pradesh Congress being accused (and accusing his rivals) of casteism (Brass 1993). The importance of caste associations was particularly marked in the south of India. While we have no figures for caste association membership immediately after independence, Figure 6.1 shows the distribution of caste association membership in 2011, taken from the Indian Human Development Survey. Overall, membership was common but now it is universal, with approximately 8 per cent of Indians belonging to these organizations. The distribution of associational membership is quite similar to the distribution of petitioning during the colonial period, being more common in the south and west of the subcontinent than in the north – the former Madras and Bombay Presidencies, and the former princely states of Baroda and Travancore. These were, not coincidently, the areas of the subcontinent where education was most common among lower caste groups in the colonial period, due to missionary activity (Lankina and Getachew 2012) or state policy (Nair 1976).

138

From Hierarchy to Ethnicity

Kerala 32

Tamil Nadu 33

Karnataka 29

Andhra Pradesh 28

Gujarat 24

Maharashtra 27

Madhya Pradesh 23

Orissa 21

Chhattisgarh 22

Jharkhand 20

West Bengal 19

Bihar 10

Assam 18

Uttar Pradesh 09

Delhi 07

Rajasthan 08

Haryana 06

Punjab 03

Uttarakhand 05

Himachal Pradesh 02

0

Proportion Members of Caste Ass. .1 .2 .3 .4 .5

Figure 6.1 Caste associational membership in India, 2011

Note: The figures show the proportion of Hindu non-Dalit respondents who reported being members of caste associations and parties in the 2011–2012 Indian Human Development Survey.

In areas where caste associations were weak, controlled by rivals, or neutral (M. S. A. Rao 1968), politicians could not use them as a base. To replace the associations, would-be caste politicians turned to political parties, social service organizations, and informal clientelistic networks of various kinds. Haryana Yadav leader Rao Birender Singh, for instance, was foiled in his attempt to become permanent president of the Yadav Mahasabha (1967) despite his long association with the organization, and founded his own National Yadav organization as well as his own political party, the Samyukta Vidhayak Dal (SVD).4 The primacy of the parties in representing (or, in some circumstances, creating) social cleavages in India has been a much-noted feature of its politics, particularly compared to the role of non-political social organizations such as labour unions in the development of European political parties (Chhibber 2010). Even without a formal organizational basis, caste-based mobilization was one of the most potent predictors of political behaviour in post-independence India, and political leaders were expected to service 4

Singh is an illustration of the peripatetic careers that politicians with a strong caste base could have in this period. Floating his own party after independence, he then merged it with the Congress, defected to form the SVD, joined the Congress again, and then defected to the Janata Dal.

Caste since Independence

139

constituencies defined by caste. In the 1967 Indian National Election Study, only 25 per cent of respondents did not know the party favoured by caste leaders, and 82.2 per cent of voters who knew their caste leader’s preference agreed with it (Eldersveld et al. 1967).5 While the vast majority of respondents refused to answer a question on the importance of caste in influence their vote choice, only 2.1 per cent said it was not important (vs. 20.1 per cent who said it was). A remarkable 26.9 per cent of respondents said it was very important for representatives to work for the interests of their caste, while 17.9 per cent said it was somewhat important and 14.3 per cent did not know (for comparison, only 30.8 per cent thought it was very important that a member work for his own beliefs). Given the generally small size of caste groups, they must usually be formed into coalitions to win elections, either under the umbrella of a single party or by combining. The formation of such ‘caste equations’ began in this period. Such coalitions could be bound together by policy targeted at specific jatis or categories of jatis, or the ‘projection’ of leaders from specific jatis. Since the business of these coalitions was particularistic distribution, there was no necessary social ‘logic’ to these coalitions: they could combine groups of very different economic status. The ‘coalition of the extremes’ thought to be characteristic of the Congress Party during the Indira Gandhi years combined upper caste with Dalit and Adivasi voters (Jaffrelot 2003: 427–452). In Gujarat, Chief Minister Madhav Singh Solanki’s version of this coalition was successful enough to be designated by an acronym: KHAM (Khatriyas, Harijans, Adivasis, Muslims) (Wood 1987).

The ‘silent revolution’ One complicating factor to the game of combining caste groups to form political coalitions was that the relative political weight of groups was constantly changing. As in the colonial era, the expansion of education led to new groups entering the political process and improved identity mobilization and cohesion within old groups that had some previous experience with caste-based mobilization. In the south and west of India, where education had been relatively widely spread in the colonial era, 5

An additional 25 per cent reported that their caste leaders were neutral. A majority of voters were also able to identify parties preferred by ordinary members of the caste itself.

140

From Hierarchy to Ethnicity

these changes were less marked, since most of the major caste groups were already politicized at the beginning of the period. In northern India, however, the expansion of education would lead to new groups becoming politically relevant. These changes, which Christophe Jaffrelot (2003) has termed the ‘silent revolution’, would have important consequences for the caste composition of the elite and structure of the party system in these states.

0

Proportion of MPs .2 .4

.6

Figure 6.2 Caste category representation among north Indian MPs, 1952–2004

1950

1960

1970

1980 Year

Upper Castes Non−Yadav/Kurmi/Lodh OBCs

1990

2000 OBCs

Note: The lines show the proportions of various caste categories among MPs from Hindi-speaking states at each election, taken from the Jaffrelot/CSDS MP dataset. Koeris are included with Kurmis.

The most immediate consequence of this change was the change in the caste composition of legislatures. Figure 6.2 shows the changing composition of MPs sent to the Lok Sabha from northern India. The representation of OBC castes rose continuously throughout the period, while that of upper castes fell. The representation of SCs and STs, guaranteed by quotas, remained constant. Between 1957 and 1999, the share of OBCs in the legislature rose from 19.4 per cent to 46.8 per cent in Bihar, 4.7 per cent to 22 per cent in Madhya Pradesh, and 12 per cent to 25 per cent in Uttar Pradesh. These legislators gradually gained influence. Uttar Pradesh had its first non-upper caste chief minister in 1967, its first OBC chief minister in 1977, and its first SC chief minister in 1995 (Jaffrelot and Kumar 2012).

Caste since Independence

141

One of the most revealing aspects of this rise was that it was not evenly spread among OBC groups. The vast majority of the gains in representation were concentrated among a few groups that were both relatively advanced in educational terms but happened to be among the largest groups in these states – the Yadavs, Kurmis, and Lodhs – while representation from the many smaller OBC groups stagnated. In 2004, members of these three groups made up an estimated 8 per cent of the population of the Hindi-speaking states, and 16 per cent of MPs, while the rest of the OBC population made up 31 per cent of the population and 9.3 per cent of MPs.6 In Bihar, the share of MLAs from the largest OBC jati, the Yadavs, rose from 6.6 per cent to 25.8 per cent, while the proportion of MLAs from the other OBC groups only increased from 13.8 per cent to 21 per cent. While many politicians, such as Mulayam Singh and Kalyan Singh in Uttar Pradesh and Laloo Prasad Yadav and Nitish Kumar in Bihar, have campaigned on the basis of promises to improve the lot of the OBC category, the ‘rise of the OBCs’ usually reflected a narrower and more traditional dynamic: the successful use of a voting bloc based on jati to gain political power. As in the colonial period, in the ‘headcount politics’ of modern democratic India (Chandra 2007) caste mobilization has tended to favour larger groups, even if their size is not particularly large in absolute terms.7 The rise of these caste groups was associated with the expansion of education. As we have seen, the colonial government thought that an educated population was a potentially rebellious one, and invested relatively little in education, particularly at the primary level. What resources the state did direct towards education were concentrated at the secondary and university levels, which performed the important task of producing bureaucrats. In 1940, only 5 per cent of eligible children attended primary school (Lindert 2004), and little progress was made during the subsequent decade, when official attention was distracted by war, independence, and partition. After independence, the Congress government thus began to direct resources towards education, a goal that was enshrined in the directive principles section of the constitution. While this effort was not backed by the same level of administrative centralization and drive that characterized 6

7

State population figures are based on Lee’s (2019) recoding the 2005–2006 NFHS survey. Yadavs, the largest caste group in Bihar, and Chamars, the largest caste group in Uttar Pradesh, each make up only 13 per cent of the state’s population.

142

From Hierarchy to Ethnicity

the contemporary literacy drive in China, its results were nevertheless impressive. Between 1951 and 1961, the literacy rate increased from 18.3 per cent to 28.3 per cent, and by 1971 it had reached 34.3 per cent (Chari 1977). These gains brought a substantial increase in the percentage of the population that could potentially take part in politics. Predictably, the expansion of literacy was especially marked among lower caste groups. In 1961, 10.3 per cent of members of SCs were literate, and 8.3 per cent of members of STs, up from tiny figures in the colonial period. The lower castes were also catching up with the upper castes, with the SC literacy rate being 10.3 per cent (18 points behind the overall figure) in 1961 and 37.4 per cent (15 points behind the overall figure) in 1991.

0

Proportion of Population Literate .2 .4 .6 .8

1

Figure 6.3 Literacy by caste category in north India, 1920–1994

1920

1940

1960 Year Upper Castes SCs

1980

2000

OBCs STs

Note: The lines show the proportions of individuals from Hindi-speaking states literate by caste category and decade of maturity in 2011–2012, taken from the Indian Human Development Survey (Desai and Vanneman 2015). The year label refers to the first year of the decade in which respondents turned 12. Note that due to the possibility of adult literacy acquisition and higher death rates among the illiterate, this figure probably overestimates the literacy rates of older age cohorts when they were younger.

Figure 6.3 shows the trends for various Hindus of caste categories in northern India, estimated from modern survey data.8 While the literacy 8

These figures are calculated by comparing decade cohorts in a cross-sectional survey, the 2011–2012 Indian Human Development Survey, with decades being

Caste since Independence

143

of all caste categories rose rapidly during this period, that of the OBCs and SCs rose especially rapidly, closing the gap with the upper castes. This trend was especially marked among some of the more politically active OBC groups: Yadavs, for instance, went from a literacy rate of 24.5 per cent (among individuals who turned 12 before 1950) to 84.6 per cent (among individuals who turned 12 after 1980). The ethnographic literature supports the idea that it was this rise in education that allowed the members of these castes to make full use of their group’s numerical size, and to break free of patrimonial ties to traditional elites. In Anand Chakravarti’s (2001) account of a Bihar village in the 1980s, for instance, Bhumihar landowners complained about how education has made the Yadavs ‘pushy’ (unlike other lower caste groups), while Michelutti (2002, 2004) describes the rich and highly politicized group identity that developed among Yadavs during this period. The Yadavs are large enough that the association between education, political involvement, and polarization can be measured quantitatively. Figure 6.4 shows the gradual rise of literacy among the Yadavs, their rise in representation in the legislature, and their rise in representation in the party most associated with them (the Rashtriya Janata Dal [RJD] and its predecessors). Note that the figures, drawn from a 2012 cross-section, probably seriously overstate literacy in the earlier decades, since mortality is higher among the illiterates and individuals can also gain literacy as adults. Similarly, potential Yadav representation is reduced by the reserved seats, and Yadav polarization at the politician level is understated by the desire of rival parties to nominate Yadavs to break up the RJD’s vote bank. Despite this, the overall positive trend in literacy over the past half-century seems broadly in line with the rise in Yadav descriptive representation, and the consolidation of a Yadav legislative bloc measured by the gap between Yadav representation in the RJD and in the legislature as a whole. Unlike the colonial period, in the twenty-first century, we can actually observe the relationship between education and caste associational membership at the individual level. Figure 6.5 shows the relationship between individual education and caste association membership. As Figure 6.1 showed, only 8 per cent Hindus are members, though this figure is larger than the 3.5 per cent who are members of political parties and the defined based on when the respondent turned 12. This approach has been preferred to the census data cited earlier because the census does not collect data on OBC status.

144

From Hierarchy to Ethnicity

.1

.2

Proportion .3

.4

.5

Figure 6.4 Literacy, political involvement, and caste polarization among Bihari Yadavs, 1960–2017

1960

1980

Year

2000

2020

Yadav Literacy Yadavs in Legislature Yadavs in Janata/Janata Dal/Rashtriya Janata Dal

Note: The figure shows the proportion of Yadav respondents who were literate at the 2011–2012 Indian Human Development Survey (calculated based on respondents who were seven or older in that year), the proportion of Yadavs in the Bihari legislature and the RJD and predecessor parties (taken from Jaffrelot and Kumar 2012 and press accounts).

7 per cent who are members of religious organizations. However, like other types of organizational involvement, caste association membership is more common among the educated. One of the most noticeable aspects of the lower caste rise to power was that it often involved the creation of new, explicitly caste-based parties. This was not always the case. As we saw, caste politics was the basis of many state party units in the 1960s, and many leaders of the rising OBC jatis supported the various Janata and Janata Dal parties, due to their support for reservation and the entrenchment of upper caste elites within the Congress. However, these parties were multi-caste in composition, and contained a healthy number of upper caste politicians, including much of the national leadership (Jaffrelot 2003). In the 1990s, however, the party system fragmented, producing a number of parties based around a large jati, though they at times claimed to represent larger categories of castes. The most powerful of these were the Yadav Samajwadi Party (SP) and RJD, the Kurmi Janata Dal (United) (JD[U]), the Chamar Bahujan Samaj Party (BSP), and the Paswan Lok Janshakti Party (LJP). While these parties were anxious to appeal to voters

Caste since Independence

145

0

Proportion Members of Caste Ass. .05 .1

.15

Figure 6.5 Caste associational membership and education in India, 2011

No Edu.

Primary

Middle

High

Univ.

Note: The figures show the proportion of Hindu non-Dalit respondents who reported being members of caste associations and parties in the 2011–2012 Indian Human Development Survey.

beyond their core jati, they remained caste-based parties, both in being dominated at the leadership level by a single caste and receiving most of that caste’s votes. At the 2007 Uttar Pradesh Assembly election, for instance, the BSP won 85 per cent of the Chamar vote, over a third of its total, while the SP won 73 per cent of the Yadav vote (Verma 2012). The ability of these parties to spread spatially appears to be entirely traceable to variation in the population share of their core caste (Chandra 2004). In most cases, these politicians have kept formal caste associations at arm’s length, preferring to portray themselves as the personification of the group’s political destiny, or to cultivate an image of being above ‘casteism.’ However, in many cases links to older forms of caste mobilization are direct as well as implicit. Shibu Soren, leader of the Jharkhand Mukti Morcha (JMM), got his political start as head of a Santhal tribal association, the Sonat Santhali Samaj. Similarly, the leader of the national Yadav Mahasabha, Harmohan Singh Yadav, was elected to parliament by the Yadav-dominated SP, and focused the association on demands for reservation that echoed those of the political parties (Michelutti 2002). Since 2013, the Jat Mahasabha has been led by the leader of the Punjab Congress (and Patiala royal family member) Capt. Amarinder Singh (K. Singh 2017).

146

From Hierarchy to Ethnicity

Recently, at the 2017 election, the Gujarat Congress aggressively recruited Hardik Patel, leader of the pro-reservation Patel caste association, the Patel Navnirman Sena, promising both seats and limited concessions on the reservation issue (Jaffrelot 2016). While the phenomenon is most notable in the Hindi belt, jati–party linkages are found in other parts of India. Often, a state’s major castes will each be affiliated with the major parties, with smaller castes taking sides. In (undivided) Andhra Pradesh, the Telugu Desam Party (TDP) was strongly associated with the Kamma caste (which had a Kamma leader and received 59 per cent of Kamma votes and 36 per cent of the total vote in 2009) and the state Congress with the Reddy caste (which had a Reddy leader and received 59 per cent of Reddy votes and 39 per cent of the total vote in 2009) (Suri, Rao, and Reddy 2009). In Karnataka, the Janata Dal (Secular) (JD[S]) is weakly associated with the Vokkaliga caste (which has a Vokkaliga leader and received 30 per cent of Vokkaliga votes and 13.6 per cent of the total vote in 2008), while the Lingayat community is associated with the BJP (which has a Lingayat leader and received 73 per cent of Lingayat votes and 41.6 per cent of the total vote in 2008)(Shastri and Padmavathi 2009). However, the party system is less defined by caste than in the north: not only do many members of base castes vote for the ‘wrong’ party, but the base caste’s control of ‘their’ party is less absolute, since their coalition includes many members of other groups. Even very small groups are increasingly finding ways to be incorporated into the caste party system, albeit in a subordinate capacity. Sometimes this takes the traditional form of a party ‘projecting’ a second-tier leader from the caste, or by announcing policies targeted at the group. In Bihar, for instance, Kurmi JD(U) leaders promoted the career of Jitan Ram Manji, from the tiny and impoverished Mushahar caste, while at the same time announcing policies designed to benefit the Mushahars and other ‘mahadalit groups’ whose small size and poverty had previously made their identities less politically salient than the state’s two largest Dalit groups, the Paswans and the Chamars (Kumar and Somanathan 2017).9 In other cases, a leader from a small group has formed his own party, and has been able to find electoral success through seat-sharing agreements, with

9

This example shows the limitations of the strategy, since Manjhi eventually fell out with the JD(U) leadership and left to form his own party, taking his supporters with him.

Caste since Independence

147

the smaller party receiving support in a few seats in return for throwing the support of its voting bloc elsewhere to larger parties. The past two decades have seen the proliferation of such parties, both inside and outside the Hindi belt. To take only BJP allies at the 2014 election, in Bihar the Rashtriya Lok Samata Party represents the Kushwahas and the LJP the Paswans, while in UP the Rashtriya Lok Dal represents the Jats and in Tamil Nadu, the Pattali Makkal Katchi represents the Vanniyars. The transactional politics of dealing with these small parties have become an increasing part of political management in India. In classic discussions of ethnic politics and ethnic parties, it is often assumed that identity should be perfectly predictive of voting, and that elections should be little more than censuses. While some media accounts of Indian elections tote up caste strengths in just this fashion, Indian elections are obviously much more complicated than adding up caste populations. As theory suggest, even if caste were perfectly predictive of voting, many caste identities are not the subjects of sustained political mobilization campaigns, due to small size or extreme levels of education. Even among castes who do have a party or leader that claims to represent them, a variety of factors might lead to voters voting for another party. First, rival parties frequently attempt to split their opponent’s caste bloc by nominating a candidate from that caste, leaving the voter cross-pressured between the party’s brand and the appeal of coethnic voting. Second, voters may receive clientelistic benefits even from individuals with whom they do not share a coethnic tie (Thachil 2014) – while coethnicity may be helpful in forming these links, it is by no means necessary (Kitschelt and Wilkinson 2007). Finally, the ordinary factors of political behaviour may still apply: a voter may dislike the party leader of her ethnic party, dislike policies they put forth, or feel that the party would not be viable in her constituency (Chandra 2009).10 Once in office, the policies these leaders enact have the effect of favouring certain caste groups over others. The most tangible expression of this was the imposition of reservation policies, discussed in greater detail later. At a minimum, these quotas appear to have increased the level of educational and occupational attainment among members of the OBC and SC categories (Chin and Prakash 2011; Cassan and Vandewalle 2017; Lee 2018b), though some claim that these benefits have only flowed to a few more powerful jatis 10

Experimental work has found that voters are sensitive to informational cues about ‘their’ caste party, and about caste population size (Banerjee et al. 2009).

148

From Hierarchy to Ethnicity

(Puri 2008). More directly, these parties have been accused of selectively hiring members of the base caste for government jobs: in one much-cited case, Mulayam Singh Yadav made sure Yadavs (who make up 8 per cent of the state’s population) made up 720 of the 900 teachers he appointed (Jaffrelot 2003: 380). These parties also emphasize caste in their rhetoric, condemning the dominance of the upper castes and promising to transfer resources, both symbolic and tangible, to the lower castes (Chandra 2004). Like ethnic parties in other parts of the world, these caste parties have earned a bad reputation among scholars for enacting inefficiently narrow policies that favour their base group (Munshi and Rosenzweig 2013), or having high levels of rent-seeking by leaders (Banerjee and Pande 2007), though others have argued that these results are conditional on specific types of coalition-building (Thachil and Teitelbaum 2017). More broadly, scholars have blamed the salience of caste in Indian politics for low levels of public service provision and local cooperation (Banerjee and Somanathan 2007). These critiques, however, do not necessarily mean that the consequences of caste-based mobilization have been negative, since the alternative might be biased distribution by upper castes, a pattern for which there is abundant evidence (for example, Lee 2018a). Caste mobilization might even the playing field somewhat, with rotation in office guaranteeing a fair distribution of state resources. However, caste-based distribution does not obtain anything for groups which are not mobilized, or whose identity and caste institutions are so weak as to be not worth cultivating. Despite the recent trends towards the politicization of smaller caste identities, there remain several groups whose status as vote banks is less well established, in particular the smaller occupational castes, Dalit, OBC, and Muslim.11 These groups remain very poorly represented in the bureaucracy and political class, and their interests are poorly represented in the formation of public policy. Lee (2019) extends the logic of caste-based distribution to develop a general theory of regional political economy in India. He shows that specific castes have tended to dominate the politics of specific regions, and that public policy in these regions tends to reflect the interests of these castes: states controlled by large heavily landowning groups, for instance, 11

These groups provided the traditional services in Indian villages: potter, blacksmith, scavenger, weaver, and the legacy of having only a few households in each village has left them geographically scattered, poorly educated, and without access to land.

Caste since Independence

149

spend heavily on infrastructure, while states controlled by large heavily professional groups, spend heavily on social services. These powerful castes do not necessarily mobilize supporters entirely on the basis of caste, however, often preferring alternatives such as religion and class, with explicitly caste-based rhetoric being confined to numerically large but marginalized groups. These results indicate that caste politics can explain not just political rhetoric and distributional strategy, but the development trajectory of regions.

The upper castes and mobilization In Chapter 4, we saw that in the colonial period upper castes refrained from demonstrative forms of caste mobilization like petitions while dominating non-caste-based forms of political participation such as the civil services and the Congress Party. This pattern remained true in the post-colonial era. In the civil services, the imposition of quotas has gradually eroded the position of the upper castes: in theory, 50 per cent of new recruits must be from less-privileged social categories (Bhavnani and Lee 2018). However, the upper castes remain heavily over-represented. In Bihar in 2012, the four major upper castes provided 77 per cent of senior civil servants (Witsoe 2013: 85). The clash between upper caste bureaucrats and lower caste politicians has often been noted, with politicians sometimes being accused of sabotaging the bureaucracy in order to maintain control (Jha and Pushpendra 2012). Similarly, the upper castes are still, despite the ‘silent revolution’, over-represented in electoral politics in all parts of India (Jaffrelot and Kumar 2012). While they are found in all parties, they are especially likely to serve in and support parties that do not base their appeal on caste. The major national parties like the BJP, the Communists, and the Congress de-emphasize caste in their rhetoric and use patronage to bridge the gap between their upper caste leadership and their more diverse followers (Thachil 2014). The Congress Party emphasizes Indian nationalism and the charisma of its First Family, the BJP appeals to a broad Hindu identity, and the Communists to class identities. These organizations are hostile towards jati politics, which they see as divisive: the manifesto of the Communist Party of India (Marxist) (CPI[M]) proclaims: ‘The dominant class in the rural areas utilise caste affiliations to mobilise support and resort to violence to terrorise the rural poor into submission’ (Communist Party of India [Marxist] 1964 [updated 2000]), while the BJP manifesto notes that ‘India constitutes of all its people, irrespective of caste, creed,

150

From Hierarchy to Ethnicity

religion or sex’ (Bharatiya Janata Party 2018). Both the CPI(M) and the BJP, interestingly, are either hostile or ambiguously neutral towards caste-based reservation policies, while supporting economically based reservation policies. The BJP’s identity project is perhaps the most interesting upper caste challenge to the dominance of caste-based mobilization in modern India. Like the caste associations of old, the activists of the RSS, BJP, and their sister organizations have worked to create an imagined ‘Hindu’ community, complete with supporting histories, rituals, and behaviours (Hansen 1999). Like other unranked identity entrepreneurs, the RSS condemns internal divisions within the community it defines, in particular those based on caste. In practice, of course, this entrenches the role of the upper caste leadership, since the party is opposed to any form of caste reservation or mobilization. In fact, many commentators have suggested that the BJP’s great organizational campaign in the early 1990s around the issue of building a temple on the site of the Babri Masjid was an attempt to distract voters from the reservation agitation of the time: that they would use mandir (temple) to trump Mandal (the report recommending OBC reservations) (for example, Ram 2015). Despite their claims to represent all people, the major national parties are dominated by the upper castes at the leadership level. Among north Indian MPs elected in unreserved seats in 2004, 73.5 per cent of BJP MPs and 63.6 per cent of Congress MPs were upper caste or Jat, as opposed to 32.9 per cent of members from other parties (calculated based on Jaffrelot 2003). Despite their rhetoric of mass mobilization, the Communists were just as elitist: in the same year, upper castes made up 70 per cent of MLAs in unreserved seats in the party’s West Bengal stronghold (Jaffrelot and Kumar 2012). Upper caste voters were similarly polarized. At the 2012 Uttar Pradesh election, the Congress and the BJP won 70 per cent of the Brahmin vote and 60 per cent of the Bania vote, despite coming in third and fourth (to the ‘casteist’ BSP and SP) on a collective 26 per cent of the vote (Verma 2012). Note that even these parties take caste into account in their political strategies, reflecting the fact that jati is a more effective predictor of voting behaviour than either religion or caste category (Dunning 2009; Huber and Suryanarayan 2016). As noted above, they promote lower caste leaders, target patronage, and even promote reservation policies targeted towards selected castes. However, relative to their rivals, these strategies remain subsidiary to the broader dominance of an upper caste leadership

Caste since Independence

151

and a belief in the importance of non-caste factors as the criteria for political promotion.

The decline of ranking In 1947, India became, for the first time in its history, a democracy. All key decision makers at the regional and national levels were now elected, and suffrage was universal and widely utilized. This meant that parties sought to build mass bases (with the Congress Party having a clear head start) and the stakes of electoral politics were raised considerably. At the same time, the various patrimonial institutions that the Raj had preserved – zamindari land tenure, the princely states – were abruptly abolished. While this did not eliminate the wealth or political role of the former zamindars and princes, it severed their formal connection to the state (Metcalf 1967). The result was an abrupt transition from ranked to unranked rhetoric of caste mobilization. The identity mobilization campaigns that arose in the 1950s had little use for traditional ideas of ranking. Many lower caste groups embraced the critiques of the Sanskritic caste system articulated by the non-Brahmin movement in western India (Zelliot 1970b), the Dravidian movement in southern India (Irschink 1969), or the Ambedkarite movement for Dalits everywhere (Omvedt 2006). Going further than in the colonial period, some activists rejected Hinduism entirely, with many Dravidians embracing atheism and many Ambedkarites embracing Buddhism (Zelliot 1970a). Similarly, intra-group distinctions were de-emphasized when politically expedient, with Rajput landlords inter-dining with Koli fishermen to celebrate Kshatriya solidarity (Kothari 1970; Shah 1975). In the case of the Chamars of UP, the transition was sudden. In the colonial period the Chamar elite, enriched by the growth of their traditional leather trade had previously been engaged in a fairly conventional programme of Sanskritic self-promotion, donning the sacred thread and giving up drinking alcohol, eating beef, and doing leather work (O. P. Singh 2015). The caste association (founded in 1924) petitioned the census in 1931 to be classified as ‘Jatav Rajput’, with at least three books being published to support the claim (O. P. Singh 2015).12 However, in the 1950s there 12

In 1911, Chamars who had taken up the ‘cleaner’ occupation of weaving had asked to be classified separately from the other Chamars in the census of the United Provinces.

152

From Hierarchy to Ethnicity

was a sudden rejection of these attempts, and a corresponding embrace of Ambedkarite ideals regarding the fundamental evils of the caste system (Lynch 1969). Many Chamars converted to Buddhism, and those that did not ceased using Brahmin priests (Ciotti 2006). The new community leaders, associated with the RPI, while still preferring the Jatav label, were uninterested in climbing the caste hierarchy, and saw themselves as the vanguard of a broader movement for Dalit self-assertion. While the RPI would eventually fall into internal dissension, the Jatavs vote would be key to the rise of the BSP, and in allowing its leader Mayawati (a Jatav) to be chief minister four times. For landowning groups, the transition from ranked to unranked rhetoric occurred more gradually. This was particularly true in northern India, reflecting the weakness of state institutions in the north and the relative poverty of that region in the 1950s. Jaffrelot (2000) cites examples of Sanskritic claims by Yadavs from the 1950s and claims that no ‘ethnification’ occurred in the north. However, it appears that by the 1990s the Yadavs had been able to merge traditional ideas of caste primacy into new ones about democracy (Michelutti 2004), as have other similar groups (Gupta 1997). This has been associated with the decline of previously salient status differences in the caste, reflected in marriage patterns and inter-dining, leading to social ‘fusion’ within the community. While the leaders of the Yadav Mahasabha still wear turbans at their meetings (a symbol of their traditional claims to Kshatriya status) and may still claim to be Kshatriyas in certain contexts, Yadav politicians have publically been quite critical of caste-based privilege, and Yadav activists have come to emphasize ‘the religious ideology of descent’ over ‘the traditional caste ideology of hierarchy’ (Michelutti 2002). Consistent with Hypothesis 4, both these groups were relatively large, a key factor in their eventual political success. Smaller groups have either had to form alliances with other castes, or actually merge with other castes – a rare solution, but a successful one in a few cases, notably the Gujarati Kshatriyas, discussed earlier (Kothari 1970). The Gujarati Kshatriya case was interesting because it might seem at first glance to reflect ranked norms: after all, it involved a lower caste group claiming upper caste status. However, in this period even projects that might seem to advance ranked norms tended to achieve the opposite. The Kshatriya merger did not just involve Kolis adopting Rajput norms as upwardly mobile Kolis had done in pre-colonial times, but rather Rajputs giving up their claim

Caste since Independence

153

to be cleaner than Kolis. Similarly, while upper castes were usually hostile towards ranked mobilization, this merger was initiated by the upper castes, eager to win new allies (Weiner 1967). Consequently, caste identities, as they represent themselves in the political sphere in modern India, are strongly unranked in their substance. Lower caste politicians do not claim to be of high caste status, or adopt behaviours considered typical of higher castes – on the contrary, they glory in their humble origins and in symbols of caste identity. Laloo Prasad Yadav, for instance, kept dairy cows, a symbol of the Yadavs’ traditional role as herders, in the chief minister’s residence, and was publicly photographed milking them (Thakur 2000; Michelutti 2007). The traditional rules of pollution are publicly flouted during campaigns, as even upper caste politicians ostentatiously shake the hands of and eat food presented by all kinds of voters (Banerjee 2017; Personal observation in Bihar fieldwork). In a ranked system, selection into positions of power favours those who exhibit high-status traits. In modern India, while upper caste politicians are still over-represented, nobody has ever suggested that voters systematically prefer upper caste candidates to lower caste candidates from a different caste, all things being equal. If anything, it appears that voters have a slight preference for leaders of similar status when choosing among individuals from other castes, a policy paralleled by politicians in their distributional policies (Dunning 2009). However, voter preferences for candidates of their own jati are strong, and much more predictive than any other broader identity or category (Huber and Suryanarayan 2016). This would make jati fundamentally similar to ethnic politics in other parts of the world, where in-group preferences (however derived) are thought to be the most important aspect of ethnic voting (for example, Banerjee et al. 2007). Inter-caste marriage patterns represent perhaps the hardest test of whether Indian lower castes have fully rejected ranked identities for unranked ones, since it reflects deeply personal choices about social positioning. Note that while both ranked and unranked systems dislike inter-group marriages, in ranked systems some individuals from low-status groups should be open to marrying into the higher status group. This is particularly true of women, who are more likely to take on the social status of the family they marry into. Research by Ahuja and Ostermann (2016) indicates that hierarchy still maintains some purchase among a segment of lower caste Indians in this highly personal context. Lower caste brides on a dating website were approximately 60 per cent more likely to respond

154

From Hierarchy to Ethnicity

to messages from an imaginary, highly desirable, upper caste groom than upper caste women were to respond to messages from a similarly desirable imaginary lower caste groom, with those open to cross-caste marriage being drawn from the higher income end of the lower caste distribution and the lower end of the upper caste distribution. While inter-caste marriage is often considered a modern phenomenon in India, these preferences offer echoes of the hypergamous marriages of pre-modern India, where poor Rajputs accepted large dowries to marry Koli women (Singhji 1994). However, these vague preferences must be viewed in the context of the overwhelming prevalence of caste endogamy in India. Only 5 per cent of marriages in India were across caste lines in 2006, as compared to 8 per cent of marriages in the United States that were across the much smaller number of racial lines (National Family Health Survey 2005–2006).13 In 2016, the Indian Survey of Social Attitudes found that 56 per cent of Uttar Pradesh Hindus said they favour a law against the marriage of high and low caste individuals, and even in cosmopolitan Mumbai, 29 per cent approved of such a law. (Unfortunately they were not asked their preferences on inter-caste marriage in general.) Remarkably, however, this proposal was as popular among lower castes as upper castes after controlling for education: in Uttar Pradesh, 59 per cent of Hindu Dalits approved, as opposed to 52 per cent of upper castes, indicating that much of the reaction reflected a universal distaste to crossing caste boundaries, rather than an upper caste distaste for lower castes violating the rules of caste pollution. If anyone in modern India should remain attached to ranked rhetoric, it should be a small landowning caste in an area where state authority is weak. The Bihari Bhumihars, discussed in the previous chapter, fit the bill, being the largest landowning group in Bihar (a state which historically has had an administratively weak state), despite making up only 2 per cent of the population (Lee 2019). Bhumihar elites remain committed to emphasizing the Brahmin identity of the caste, with the Akhil Bhartiya Bhumihar Brahmin Samaj’s website proclaiming that ‘Bhumihars are the Brahmins who are not stereotyped and try to go along with time. . . . Other Brahmins who lived off the donations of these hardworking Brahmins like bhumihars, chitpawans, tyagis and anavails didn’t consider them equal to their social

13

However, note that only 1 per cent of US marriages were between African-Americans and whites.

Caste since Independence

155

status and consider them beneath them in social hierarchy.’14 Remarkably in modern India, they are also unapologetic about their relative social advancement.15 While the Bhumihar community has been very active in electoral politics throughout the twentieth century, several factors hint at a more ‘patrimonial’ approach. Unlike the Yadavs and Chamars, it has never been politically united, either before independence or after. Bhumihar politicians have been active in a wide variety of political parties, and the political allegiance of Bhumihar elites seems to be based on patron–client ties rather than purely caste considerations (Jaffrelot 2003: 350–386; Witsoe 2013). Second, Bhumihar elites have been extensively involved in the use of extrastate political violence, usually directed against upwardly mobile lower castes (A. Chakravarti 2001; S. Chakravarti 2010). While this violence has usually been informal and decentralized, in the late twentieth century it grew considerably in scale, with the Bhumihar Ranvir Sena mobilizing thousands of men and conducting well-publicized massacres of left-wing activists and lower caste individuals thought to be uncooperative. The localization of ranked behaviours in areas with a weak state presence is also shown in the survey data. Figure 6.6 shows the distribution of Indians who admitted to practising untouchability (ritual avoidance of Dalits) in the 2011 Indian Human Development Survey. By itself, the fact that only 21 per cent of Indians would admit to practising untouchability is remarkable, given the usual consensus that the practice was universal even in the late colonial period (Omvedt 1994). While some of the small size of this effect might be attributed to social desirability bias (since untouchability is widely disparaged in the mass media and in elite Indian culture), the percentage of Dalits reporting suffering untouchability in the previous five years is very similar (19 per cent). The figures show that untouchability is concentrated among the poorer and more isolated parts of Indian society. Untouchability is more common among rural residents than urban residents, among villagers without electricity than with it, (slightly) among the uneducated than the educated, 14

15

http://www.bhumiharmahasangh.com/achievements.php, accessed 11 August 2018. ‘Bhumihars have always been emblematic icon caste for Land control. Bhumihars had control over 45% of agricultural land in Bihar and wielded a great political strength in Pre-Mandal politics of Bihar as Brahmins all over the country’ (ibid.).

156

From Hierarchy to Ethnicity Figure 6.6 Untouchability practice in India, 2011

Large City

City

More Dev. Village Less Dev. Village

0

0

Proportion Practising Untouchability .1 .2

Proportion of Practising Untouchability .2 .1 .3

.4

(b) Electricity Access

.3

(a) Urbanization

No Electricity

(d) Landownership Proportion Practising Untouchability .2 .1 .3

No Educ.

Primary

Middle

High

Univ.

0

0

Proportion of Practising Untouchability .1 .2 .3

.4

.4

(c) Education

Electricity

No Land

Landowner

Land Rented Out

Land Rented

Note: The figures show the proportion of Hindu non-Dalit respondents reporting practising untouchability in the 2011–2012 Indian Human Development Survey. Figures B and D refer only to rural areas.

and among landowners than non-landowners. Figure 6.6(d) provides an interesting confirmation of Chapter 2’s discussion of the role of ranking in the construction of patronage networks: The type of cultivators most likely to practise untouchability are those who rent out land to tenants, rather than those who cultivate the land themselves. These findings are in some respects encouraging: all of the segments of Indian society that practise untouchability are declining in relative size – India is gradually becoming more educated, urbanized, electrified, and so on. However, that does not mean that social relations between the castes have become close. In the 2016 Social Attitudes Survey, only 58 per cent of Rajasthan residents had a Dalit friend or acquaintance, and only 26 per cent had eaten at that person’s house – though the figures were a more promising 78 per cent and 58 per cent in Mumbai.

Caste since Independence

157

Caste reservations The discussion so far has ignored one important difference between caste politics today and during the colonial era: the presence of policies of caste-based affirmative action in political recruitment, government hiring, and educational admissions (Galanter 1984; Parikh 1997; Jenkins 2003; Dunning 2009; Besley, Pande, and Rao 2013; Dunning and Nilekani 2013; Jensenius 2013; Lee 2018b). These policies have profoundly accelerated the two major dynamics of post-independence caste politics discussed in this book: the intensification of mobilization and the decline of ranked caste rhetoric. At the same time, causation flows in the other direction: politicians representing mobilized unranked caste identities have been the biggest advocates for the gradual expansion of reservation policies.

The evolution of reservation policies Quotas for specific groups in government hiring (known in India as ‘reservations’) began during the late colonial period in southern India, being implemented in Mysore in 1918 and Madras in 1927, and were being discussed (though not implemented) in Bombay in the late 1920s (Irschick 1969; Mathur 1998: 17–27).16 The target of these policies was generally Brahmins, and all ‘non-Brahmins’ were able to take advantage of them. Politically, these reservations represented an attempt to court popularity by colonial and princely governments concerned about the rise of nationalism, while at the same time dividing the mass of the population from the Brahmin leadership of the nationalist movement. Ideologically, the reservation movement (like the contemporary movement to guarantee the rights of agricultural tenants) reflected the final chapter in the long-term alienation of the colonial bureaucracy from the upper caste elites that had benefited so much from colonial rule, but were now regarded as ‘effeminate’ and treacherous (Sinha 2015). However, reservation was also pushed from below by the caste sabhas, who were energetic in lobbying for reservation and in ensuring that they were included on the list of beneficiary groups (Hardgrave 1969; Mathur 2004: I). Perhaps for this reason, the areas that 16

Mysore had reserved some posts before 1918. Several princely states in the Bombay Presidency also implemented reservations.

158

From Hierarchy to Ethnicity

adopted caste reservations in the colonial period tended to be areas where caste sabhas were very active.17 These policies became a national possibility after independence, and indeed the political ferment over B. R. Ambedkar’s agitation for separate electorates for Dalits in the 1930s had made some form of concession inevitable.18 Reservations at the state and national levels for SCs and STs were mandated by the constitution: SCs and STs were guaranteed representation in government hiring, educational admissions, and state and local legislatures proportional to their population. These reservations have never been particularly controversial in India, in part because, at least until very recently, they were of relatively little practical significance due to the low level of education among SCs and STs. Weisskopf (2004) cites figures that in 1969 only 11 per cent of the SC quota at a medical college in Mumbai was filled, and that by the 1980s, the figure had only risen to a third. Given that unfilled seats were filled by ‘general category’ applicants, these policies actually did little to redistribute opportunities. Similarly, the practical (as opposed to symbolic) effects of electoral reservations of SCs and STs appear to have been either modest or insignificant, because (as Ambedkar had feared) SC and ST legislators were forced to appeal to voters and party leaders from other groups (Dunning and Nilekani 2013). The constitution also opened the possibility that other disadvantaged groups, the so-called Other Backward Classes (OBCs), would receive reservations at some unspecified future time, though it was vague on how these groups would be defined and what benefits they would receive. The results were not unpredictable. In the areas of India where unranked caste identities had become politically mobilized in the colonial era, ‘OBC’ status was defined on a caste basis, to include many of the largest and most politically powerful caste groups, who were then given reservations. Table C.11 in Appendix C shows the size of these quotas and the dates of their introduction. Tamil Nadu, Kerala, Andhra Pradesh, Karnataka, and Maharashtra all introduced OBC quotas by 1964, while the northern states remained aloof.19

17

18

19

It is also possible that the early adoption of reservations in the south stemmed from the weak numerical position of the upper castes in this region. Ambedkar had in fact won modest reservation concession from the British during the Second World War, when the Congress was out of office. The exception was Punjab, which implemented a modest 5 per cent quota.

Caste since Independence

159

The first OBC reservations prompted a series of legal challenges, which gradually established legal restrictions on the practice that still hold today. The most important issue was the size of the quotas relative to the seats allotted to the general category. Some states, eager to accommodate politically powerful caste groups, attempted to distribute virtually all opportunities based on quotas. The Supreme Court, in M. R. Balaji vs. Mysore (1963) rejected this view, declaring that at least 50 per cent of opportunities should be allotted to the general category. While Tamil Nadu was eventually able to win an exception to the 50 per cent ceiling by constitutional amendment, Balaji remains good law, and an important brake on the scope of quotas. A second, less coherent set of cases established that determinations of ‘backwardness’ could be made on the basis of caste (see Galanter 1984). Many had assumed that backwardness would be defined based on wealth or education, allowing poor upper caste members to benefit, and a few states (Haryana, Himachal Pradesh, and Karnataka) made experiments with such classifications.20 However, the practical difficulty of determining eligibility and pressure from lower caste politicians made such policies increasingly rare. Instead of determining each individual’s backwardness, the state would instead determine the caste’s backwardness, place it on a list, and then determine whether individuals belonged to these castes. The courts retained the right to second guess the state’s judgement, and in several cases set aside the inclusion of castes on the OBC lists of castes they considered too socioeconomically advanced (see Ram Singh & Ors. vs. Union Of India 2015). Courts also came to require that demonstrably wealthy members of OBC castes (the so-called creamy layer) should not benefit from reservation, though this rule is widely evaded through the falsification of income certificates. As we have seen, in this period upwardly mobile lower caste politicians were often more powerful outside the Congress Party than within it. It was thus unsurprising that the electoral defeat of the Congress in 1977 was associated with a surge of interest in reservation policies.21 At the national level, the new Janata Party government appointed the Mandal Commission, 20

21

Heavily Muslim Jammu and Kashmir retains a class-based system. At the time of writing, the government is introducing an additional 10 per cent reservation for poor members of the upper castes. In 1980, 11 per cent of north Indian Congress MPs were from OBC groups, versus 24 per cent of non-Congress MPs (Jaffrelot MP dataset).

160

From Hierarchy to Ethnicity

whose eventual report called for OBC reservation in all levels of government and contained a draft list of castes whose members should be given OBC status (India Backward Classes Commission 1980). At the state level, there were fierce conflicts over the introduction of reservation by the newly elected Janata government, particularly in Bihar, Gujarat, Karnataka, and Uttar Pradesh (Jaffrelot 2003), with upper caste protests succeeding in defeating reservations in the latter state and limiting their scope in Bihar. In the 1980s, with the Congress in power nationally and in most states, no new states implemented reservation, though the issue continued to be a politically important one in some states. However, the defeat of the Congress Party in 1989, and the more general opening up of the party system in these years, put in power politicians with strong incentives to extend reservation. Haryana introduced reservation in 1991, in what was widely seen as a concession to the Jats (the state’s largest caste), while in Bihar Laloo Yadav extended the OBC quota up to the legal maximum. In other states, the catalyst was action at the national level. In 1990, as upper caste protestors immolated themselves on the streets of Delhi, Prime Minister V. P. Singh announced the implementation of central reservations in the central bureaucracy and in all remaining states, though the policy was only implemented in 1994 after a Supreme Court ruling. Even though all large states now have OBC reservation, political conflict over reservations remains intense. In the Indian context, reservations are assigned to specific categories of jatis (OBC, SC, and ST) and jatis are assigned to categories based on their perceived level of backwardness. This naturally made ‘backwardness’ both desirable and politically controversial (Jenkins 2003). Middle status groups seek to be classified as OBC, while poorer OBCs seek to be classified as SC or as members of a separate ‘Most Backward Caste’ category. Since there is no absolute definition of backwardness, nearly all groups have a plausible case. While the formal procedure for such decisions is that national and state level Commissions for the Backward Classes carefully investigate the socioeconomic status of each group (often commissioning surveys and ethnographic investigations), most groups with any political influence prefer to take their chances in the streets.22 In 2018, for instance, groups

22

The activity of the National Commission for Backward Classes focuses predominantly on castes too small to be noted in the original Mandal Commission list, and in notifying alternative names for listed groups.

Caste since Independence

161

claiming to represent the Jats of Haryana and the Patels of Gujarat, two relatively wealthy peasant groups, are mounting aggressive agitations to secure OBC status, combining both violent protests and threats to political parties that did not agree to their demand. In Rajasthan, after protests that shut down the state, the Gujjar community was twice (2008 and 2015) able to get the government to pass bills to give them a special quota within the OBC category, only to see them struck down in the courts (Chauhan 2015). The design of the reservation system ensures that within a given caste category only the most qualified individuals receive the benefit. Given the close association between formal qualifications and socioeconomic status, quotas thus are widely thought to give benefits primarily to the most advantaged jatis within each disadvantaged category, though the quantitative evidence on this point is ambiguous (Lee 2018b). In practice, this means that certain jatis are frequently accused of ‘hogging’ the benefits of reservations (Ram 2015). The preferred policy solution to this problem has been to subdivide the categories and create new ones – ‘most backward caste’, ‘extremely backward castes’, ‘mahadalit’, and so on. In some states, this subdivision had been carried to absurd lengths: Karnataka now has five different types of OBCs (1, 2a, 2b, 3a, and 3b), each with their own quotas. The creation of such sub-quotas is obviously a fertile area for political conflict. Another source of conflict is the scope of reservation. Until 2008, central educational institutions such as the Indian Institutes of Technology and Indian Institutes of Management were exempt from reservation.23 Given that these are widely considered the most prestigious undergraduate institutions in the country, this was a major limitation of the policies, and the implementation of reservation at these schools remains a point of bitterness among many upper caste Indians. Similarly, reservation does not generally apply after hiring, and many individuals hired through reservations are thought to be denied promotion. The constitutionality of promotion reservations has been a matter of great controversy, with the Supreme Court’s decision in Indra Sawhney vs. Union of India being modified by no less than four constitutional amendments. Following M. Nagaraj & Others vs. Union of India & Others (2006), states are now given the option of implementing promotion quotas if they can show just cause, a condition that has tied up all attempts to implement such policies in litigation. 23

These institutes had briefly implemented SC/ST reservation in the 1970s.

162

From Hierarchy to Ethnicity

Consequences of reservations The goal of reservation, or at least the stated goal, was to reduce socioeconomic inequalities between caste groups. In this it has not been fully successful: caste remains highly correlated with wealth and education within Indian society (Zacharias and Vakulabharanam 2011). However, it appears that reservations have had positive effects on the educational attainment and formal sector labour participation of the beneficiary groups (Chin and Prakash 2011; Khanna forthcoming; Cassan and Vandewalle 2017; Lee 2018b). Lee (2018b) estimates that the effect of the OBC reservations in 1994 was to increase OBC educational attainment by 0.5 years and the probability of holding a middle-class job by 3.6 percentage points. While modest in size, the results are intriguing given the link between education and identity mobilization. A caste’s educated elite could agitate for reservation which would create an educated elite of increased size which would in turn lobby for the preservation or extension of reservation benefits. Another, in some ways more remarkable, effect of reservation was to involve the Indian state closely in the delimitation and policing of caste boundaries. Dirks and Gupta accused the colonial state of reifying caste identities through the census, which every 10 years asked individuals their caste, and then classified and sorted the responses, which were then thrown away. The contemporary Indian state’s intervention, by comparison, is permanent, overt, and expensive. In order to avail themselves of the benefits of reservation, Indians must possess a caste certificate. At the 2012 India Human Development Survey, 38 per cent of Indians were members of a household which possessed a caste certificate – for a few, it was their only type of identification. These certificates are issued by the sub-district administration, who demand paperwork proving a relationship to someone with an already issued certificate and conduct field inquires of the neighbours in doubtful cases. One consequence of the certificate system was to reinforce the hegemony of a standardized taxonomy of castes, as promoted by the major caste organizations and canonized in the central and state lists. Even if a bureaucrat could be persuaded to give it, there was no point in obtaining certified membership in a group not on the list of those receiving benefits, or even of an alternative spelling or subgroup of such a group. Similarly, the certificate provided no room for ambiguous or mixed identities. In the reservation system, an individual had to choose a binary identity, and a subgroup could not claim an identity distinct from others within the same

Caste since Independence

163

category. For instance, many Uttarkhandis could be classified as either Tharu or Rajput, but now had to choose which side of an officially defined line they wished to be on. For this reason, the caste certificate process has been criticized as eliding subtle differences in individual identity and status in favour of simple categorical binaries (Jenkins 2003), though falsification of caste certificates is still common in some cases (Khanna forthcoming). In colonial times, of course, many individuals had decided that they wished to be on the ‘higher’ side of such lines, leading to the phenomena of ranked claims. Reservation, however, has reversed these incentives. Castes are encouraged to emphasize not their glorious ancestry and affinities to the upper castes, but their history of poverty and discrimination. Claiming to be Kshatriya or Brahmin creates the dangerous suggestion that the group might not always have been the victim of discrimination. This incentive to backwardness has led to a number of peculiar reversals in group rhetoric. While the Jats had historically been slowly Sanskritizing, and the Jat Mahasabha’s website still cites Risley with approval,24 the head of the Mahasabha declared in 2015: ‘Agriculture is the main source of livelihood for the Jat community. On an average, Jats have less than five bighas of land in their names. . . . For the Jats of the state, reservation is the only ray of hope’ (Dabas 2015). It might be thought that the boundary-marking aspects of reservation policy would be counterbalanced by the fact that they create common interests for sets of groups. Over time, perhaps, Kurmis and Yadavs might merge into a common OBC identity, or Chamars and Paswans into a common SC identity. Certainly, there have been attempts to do this, such as the BSP’s attempts to benefit all ‘Bahujans’ (or, more modestly, all Dalits), and the older effort to create an ‘Adi-Dravida’ identity in Tamil Nadu (Chandra 2004). However, these trends are at best modestly successful: jati appears to trump category in both distribution and voting (Dunning 2009; Huber and Suryanarayan 2016). The reasons are straightforward: within each category, groups are competitors for the benefits of reservations. In particular, poorer groups within the category may perceive reservation policy as being irrelevant to them, or wish to enact sub-quotas to guarantee a share of benefits for themselves. Such slicing has undermined most attempts at category mobilization. In Bihar, for instance, Nitish Kumar rallied the ‘Extremely Backward Castes’ against Laloo Prasad Yadav’s 24

http://www.jatmahasabha.in/p/jat-history.html, accessed 28 August 2018.

164

From Hierarchy to Ethnicity

attempt to pose as a leader of all OBCs, and ‘mahadalits’ against Ram Vilas Paswan’s attempt to pose as the leader of all SCs (A. Kumar 2013). Similarly, in Uttar Pradesh, the BJP has been able to peel voters from the poor scavenging Bhangi caste away from the BSP by capitalizing on their resentment against Chamar domination of reservations (Verma 2017). Reservations have thus accelerated the decline of ranked rhetoric. Since jatis receive the benefits of reservation as units, the process tends to de-emphasize distinctions within jatis and emphasize those between jatis: within the unit everyone is the same from the point of the policy, while the boundaries often conform to discontinuities in very real benefits – and, given the potential to create sub-quotas, virtually any jati boundary can conform to such a discontinuity. Not only are these units starkly bounded, but the leaders of these groups now have no incentive to emphasize links, even tenuous ones, to higher status groups within the Hindu caste system. If anything, the incentives now run the other way – backwardness, rather than ‘cleanliness’, is the key to accessing state resources. This continuing rush to backwardness is shown in annual reports of the National Commission for Backward Classes (NCBC).25 In many cases, a group that had petitioned the colonial state to be declared twice born returned to the state 80 years later, asking to be declared backward. In 2013, the NCBC held a public hearing on whether Lohars and Nais in Chandigarh should be certified as backward, while the census commissioner of Punjab in 1931 had received petitions from both groups to be classified as Brahmin.26 (The 2013 request was granted, the 1931 one was refused.) Similarly, Darzis in Maharashtra petitioned the commission for OBC status in 2007, even though the same group had petitioned to be referred to as ‘Bharsar Kshatriyas’ in 1911.27 Overall, as Figure 6.7 shows, the number of groups classified as Backward has remained steady since 2000, even two decades after the Mandal report’s implementation, with several dozen new castes successfully being added to

25

26

27

The commission does not publish the number of requests it receives, only the number of public hearings it holds, and many requests appear not to be immediately heard. http://www.ncbc.nic.in/User Panel/UserView.aspx?TypeID=1162, accessed 16 September 2018. Both groups were already classified as OBCs in the modern state of Punjab. http://www.ncbc.nic.in/Writereaddata/1123.PDF, accessed 16 September 2018.

Caste since Independence

165

Figure 6.7 Successful ‘backwardness’ petitions, 1993–2014

0

100

Advices 200

300

400

NCBC Advices to Central Government

1995

2000

2005 Year

2010

2015

Note: The figure shows the amount of advice issued from the National Commission for Backward Classes to the government for inclusion or amendment of the central OBC list by reporting year. The end date of reporting years in usually in March of the next calendar year, but varies from year to year. Corrections of spelling mistakes are not included.

the national list each year. One indication of the politicization of the listing process, and the general unspoken opposition of the major national parties to reservation, is the spike in listings during 1996–1999, the only period where parties other than the Congress and the BJP were in government (or playing a key role in supporting a BJP minority government). Conversely, the low number of listings in the mid-2000s reflects the choice of the Congress government not to appoint a new chairman for the commission. The other major consequence of reservation was to create a new technique for distributional politics. One goal of caste mobilization has always been to gain resources from the state, and one of the basic premises of caste politics was that the members of a politician’s vote bank would be rewarded with transfers by the politicians once they were in office (Chandra 2004). There is abundant empirical evidence for this type of exchange both in India and elsewhere. Not only do politicians selectively transfer resources to individuals from their own group (Dunning 2009; Frank and Rainer 2012; Ejdemyr, Kramon, and Robinson 2017) but the most politically powerful caste groups within states tend to win policy concessions that advance their particular economic interests (Lee 2019). However, as methods of targeting, both these types of distribution left much to be

166

From Hierarchy to Ethnicity

desired. Individual distribution by politicians or brokers, like clientelism in general, is limited by the amount of time and information available to politicians (and their difficulty in monitoring brokers) (Stokes et al. 2013). Biases in broader policy, while often effective, may ‘waste’ resources on members of other groups. A policy designed to help Haryana farmers, for instance, will disproportionately benefit Jats, but will also benefit all other farmers and not benefit urban Jats (see Lee 2019: ch. 2). Reservations solved this problem by creating a new type of transfer, one tied explicitly to identity and easy to explain to voters. Simply by passing a law, politicians could guarantee that a specific proportion of benefits went to their own group whether they were in office or out, with the bureaucracy helpfully ensuring that non-members did not slip in. Moreover, since reservations tended to benefit those in the caste who were interested in formal education or government employment, obtaining them held disproportionate attractions to the elites of poor groups, the very individuals from whom the politicians and caste association leaders of these groups were drawn. Unsurprisingly, the successful politicians of the late twentieth century (Laloo Yadav of the Yadavs, Mayawati of the Chamars, Nitish Kumar of the Kurmis) have placed reservations at the centre of their rhetorical and distribution strategies. In 1990, Laloo had campaigned for the implementation of the Mandal report, calling for ‘Mandal Raj.’ Even with reservation now an established fact, he was still on the subject during the 2015 state elections, declaring: ‘If I am a shaitan [demon] it is because I am raising the voice against the RSS–BJP conspiracy to end reservations for dalits and OBCs’ (India Today 2015). On the campaign trail in 2018, the Karnataka Congress chief minister called for a constitutional amendment to raise reservation to 70 per cent, declaring: ‘We want STs to have 7.5 per cent reservation and SCs to have 17 per cent reservation.’ He added, ‘I am not doing this for the sake of votes’ (Swamy 2017). Even a BJP politician, former UP chief minister and Rajasthan governor Kalyan Singh, urged voters to ‘give a hard slap to anyone who tries to snatch your reservation rights’ (Shukla 2018). Among leaders without a foothold in the electoral process, the equation between voting and reservation is explicit. Hardik Patel, leader of the Patel reservation movement, told voters: ‘Whoever talks of interest of Patels will come to power in Gujarat. Patel community has brought people to power and thrown many others out of power. If we fail to get OBC status, we will

Caste since Independence

167

vote against the incumbent government’ (Chakravorty 2015). Col. Kirori Singh Bainsla, head of the Gujjar Aarakshan Sangharsh Samiti, declared: ‘We would unite in our crusade for reservation and would support this election those who would identify with our demand and give us reservation.’ If every major caste group demands reservation, and the criteria for granting these demands are unclear, India would seem to be drifting towards a long-term outcome were virtually every politically influential group has its own quota. However, there is a major obstacle in this process: the Supreme Court’s restriction of 50 per cent reservation in total. Since nearly every Indian state is now at the maximum, the inclusion of any new groups can only come at the expense of existing beneficiaries. In recent years, several states, most notably Rajasthan, have tried to break the 50 per cent barrier, under intense pressure from caste agitators. Thus far, the judges have held firm. However, this promises to be an area of active conflict in twenty-first-century India, as powerful groups not on the central OBC list (such as Patels, Jats, and many Muslim subgroups) mobilize to either amend the constitution or win OBC status in the teeth of opposition from the groups currently on the list.

Conclusion Perhaps the most cogent response to the argument that caste was an artefact of colonial policy was what occurred after colonial withdrawal. The trends that had been developing in the last years of colonial rule intensified. As more groups developed a literate elite, more groups became involved in politics, and their involvement was intimately associated with their caste identities. The best-educated groups, however, retained their colonial pattern of disdaining explicit caste appeals, remaining prominent in the Congress, BJP, and Communist parties, as well as in the colonial bureaucracy. This involvement advanced a specifically unranked vision of caste identity, with the old idea of climbing the caste ladder being replaced by the ambition of using a caste vote bank as a base for advancement in the new nation’s exuberantly participatory electoral politics. These trends were what allowed reservation to become a political reality – it is doubtful if the constitution’s promises, particularly to the OBCs, would have become reality without the constant pressure on politicians who claimed to represent these groups. However, reservation also accelerated the trends towards mobilization and unranked identities.

168

From Hierarchy to Ethnicity

It created powerful incentives for voters to favour politicians who claimed to represent the interests of their caste, and similarly powerful incentives to avoid any claims to high status. If democracy spelled the death knell for the association between ritual status and political power, reservation created what at times appears to be the opposite dynamic – a race towards the ritual bottom. As a result, as India enters the twenty-first century caste identities are stronger than ever in the political sphere, even as they become less relevant in some aspects of social behaviour. Rather than a signal of status useful to elites in the race to the top of patrimonial polities, they have become categories for the transference of state resources and the granting of electoral loyalty, a potential asset for anyone who aspires to get a ration card, send his son to university, or work as an office peon. Given the intense competition for scarce resources that characterizes Indian society and politics, it is hard to imagine it fading away any time soon.

7

Conclusion As recently as the middle twentieth century, at least a few Indians were willing to publicly argue that the caste system reflects a way of ordering human difference that was not only unchanging and unchangeable, but normatively desirable (Yogananda 2003 [1946]). At the same time, many colonial officials were willing to adopt the slightly more limited position that caste represented an ancient and fundamental aspect of Indian civilization, and that caste divisions were similarly ancient and important. Today, given the broader rise of constructivist theories of ethnicity, in the social sciences, such ideas seem almost quaint. Virtually nobody would contest that the politics of social identities in general and caste in particular have changed over time. Twentieth-century India was a society undergoing a set of political changes – from an authoritarian colonial state to a democracy, and from a society where peasant agriculture was the dominant economic activity to one where urban centres are important and many individuals are educated. We have seen how these changes were reflected in the politics of caste. During this period, caste activists successfully politicized the identity of many caste groups, beginning with a few wealthy groups in the more educationally advanced areas of the country and expanding to include even the most ‘backward’ groups in the northern states. At the same time, the rise of democratic institutions and mass politics led to a change in the nature of caste identities, with concepts of caste that emphasized social ranking giving way to those that envisioned caste groups as discrete, bounded, and conceptually autonomous, very much like the ethnic and racial groups that were emerging in other parts of the world in the same period. It is worth questioning whether this story has any relevance for Indian politics today, or for the study of identity as a whole. After all, if India ended up looking fairly similar to ‘normal’ identity systems in other parts

170

From Hierarchy to Ethnicity

of the world, what does it matter if its path getting there was long and complicated? Conversely, certain aspects of the Indian experience seem so unique and case-specific that the general applicability of the Indian case might seem questionable. This chapter answers these questions by discussing the long-term significance of both uneven caste mobilization and ranked caste rhetoric for India today. The first section summarizes the findings on how and why caste identities changed during the twentieth century. The second section discusses the long-term consequences of these trends, while the third section concludes with a discussion of the implications of these patterns for the social sciences.

Caste meets ‘modernity’ Well before the nineteenth century, many Indians, perhaps a large majority, identified with one or another of a set of identities that we refer to as jatis or castes. Most of these groups were not politicized beyond the local level in a modern sense – their members did not have the education to articulate these identities, the technology to spread their ideas beyond their immediate locality, or much of a realistic chance of winning political power in the patrimonial politics of the subcontinent. However, the high level of political instability in medieval and early modern India meant that the system was far from static. A few subgroups were able to become educated and win political power, either through military prowess, tight commercial networks, or skillful alliance building with the Muslim leaders of medieval India’s bigger polities. The identity politics that these groups practised was characterized by its embrace of social ranking. Successful individuals and social subgroups adopted the behaviours of more powerful groups, both winning their favour and subtly distancing themselves from less powerful individuals – even those with whom they might previously have had a great deal in common. Superficially, the resulting system was unchanging – groups that adopted a specific set of cultural values and behaviours and often referred to themselves as Brahmins or Rajputs occupied dominant socioeconomic positions, and other groups treated them with deference. In practice, these behaviours served as a means for elites entrench their social position, and new elites to assimilate into these groups. While these behaviours and identities were in many cases tied to the Hindu religion and the idealized concept of varna, there is no evidence that these ideas were the reason

Conclusion

171

that identities in India took on such a sharply ranked appearance, since similar spectra of subgroups emerged among other Indian religious groups and among societies outside India. Colonial South Asia was a set of societies in transition. Many parts of India possessed a small, educated middle class, but this group was tiny relative to the society as a whole. Similarly, while some parts of India possessed participatory political institutions, they were very limited by the standards of a modern democracy. In parts of the subcontinent with princely states or powerful zamindars, pre-colonial forms of authority still existed in a constricted form. This strange mixture has advantages from a research design point of view, since it means that there is considerable variation on all the factors identified in Chapter 2 as especially significant for identity mobilization. This ambiguity was reflected in the nature of colonial identity politics, where an upsurge in interest in caste identities coexisted with widespread apathy and interest in broad identities like nation and empire. Even at the end of the colonial period, there were many caste groups where virtually none of the group members were literate, let alone educated at higher levels. This fact meant that these groups had very few people capable of organizing a supralocal modern caste sabha, few people capable of confidently interacting with the colonial state, and only a small potential audience for the propaganda produced by such a group of activists. At the other end of the social scale, politicians from the most educated caste tended to see themselves as members (and, naturally, leaders) of broader and more potentially powerful identity categories than the jati. However, in the broad middle of colonial Indian society was a set of groups whose elites had enough education to form sabhas but not enough to feel that a jati-defined support base was a political obstacle. These groups established a vigorous associational life, one with well-defined social and political goals. The types of identity assertion common in this period, such as petitioning, were transitional, neither the type of intricate ritual self-promotion common in pre-colonial times nor the ethnic parties of modern times. Instead, they reflected the limited opportunities available in an authoritarian regime, and a set of ideas about social difference enthusiastically embraced by many colonial officials. While some caste elites used their new wealth to advance the traditional goals of advancement within the caste hierarchy, other groups concentrated on developing their own identity, often in opposition to hierarchical ideas.

172

From Hierarchy to Ethnicity

For groups still exposed to patrimonial politics, the education that they gained during the colonial period gave them an opportunity to do what newly elite groups had always done in India: adopt the values, and in some cases the name and identity, of traditionally high-status groups. For groups exposed to the growing electoral contestation that occurred in colonial India, such a strategy would only alienate their potential political base: the poorer members of their own group. Instead of trying to climb the hierarchy, these elites sought to become leaders of bounded and conceptually autonomous groups. Larger groups, where the electoral payoff of creating such a group was higher, were more affected by this trend. In post-colonial India, the trends in caste politics that had been apparent in the colonial period became just as ‘obvious’ and invariant features of Indian politics and society as a ranked system had seemed a century earlier. The two main features of this change were the political activation of most jati identities comprising more than a few percentage points of a state’s population, and some even smaller groups, and the almost exclusively unranked rhetoric adopted by these groups. These outcomes flowed naturally from the political and social trends of the period, in particular the increasing levels of education and the full democratization of the political system. The first of these led to activation of poorer identity groups, especially in the north of the country, whose caste institutions had been less prominent in the colonial period. The last of these led to the gradual abandonment of attempts to adopt the behaviours and identities of high caste groups, even among groups that had attempted such shifts earlier. A feature of Indian politics that had been only incidentally relevant in the colonial period further accelerated these changes: the possibility of reserving vacancies in educational institutions and government bureaucracies. This change has had sweeping direct effects on the castes and institutions in question, the consequences of which are only now beginning to be understood. The political effects, however, were also far-reaching. Even in the ordinary business of club good and patronage distribution, politicians might plausibly seek votes by promising to advance the interests of particular castes. Reservations further increased the incentives for this type of behaviour by providing an easy and credible way of delivering resources along caste lines. Furthermore, since the legal and moral case for reservations hinged on the low social status of the recipients, reservations further reduced the incentives to claim a high ranked status, leading to a

Conclusion

173

rhetorical ‘race to the bottom’ of the status hierarchy among even relatively advantaged caste groups. As a result, identity politics in twenty-first-century India, as described by Chandra (2004) and others, appears similar to identity politics in other parts of the world. The population is divided into a set of conceptually autonomous identity groups, each the product of careful development by individuals within the group. These individuals not only regard their identity as important but also use it as an important factor in their voting decisions, while politicians use them to guide distribution. Even as politicians attempt to form coalitions along varna or caste category lines, inter-jati coalitions could equally be opportunistic in nature. The unique features of the caste system are considered either archaic or an empirical obstacle to be surmounted. However, India’s history of partial identity politicization and ranked mobilization has had important consequences for its politics. These legacies are discussed in the next section.

Consequences for India The size of groups In the ethnic politics literature outside of India, one of the most commonly discussed variables is the level of ethnic diversity, usually operationalized through the Herfindahl index, which captures the probability that members of a randomly selected pair of citizens will be from different ethnic groups. The theoretical intuition behind this measure is straightforward: cooperation is easier between members of the same ethnic groups, due to common tastes, social sanctioning, or shared social networks, and societies with higher levels of diversity will thus find interpersonal cooperation more difficult (Habyarimana et al. 2007). This will, in turn, lead to higher levels of violence and lower levels of development (Easterly and Levine 1997; Alesina, Baqir, and Easterly 1999; Alesina and Ll Ferrara 2005; Miguel and Gugerty 2005). While other authors have modified this contention, arguing that the effect is conditional on the salience of the identity in question (Miguel 2004), the broad negative association and developmental achievement remains a widely held view. The creation of large identity groups is itself a product of several factors, of which the most studied are the role of the state (Laitin 1986) and political coalition building by groups (Gellner 1983; Chandra 2004). India, like many nations in the developing world, was ruled for a crucial

174

From Hierarchy to Ethnicity

two centuries by a colonial state for whom the creation of a homogeneous national identity was more threatening than useful; certainly, India never saw the sort of sustained effort to use conscription and primary education to inculcate national identities that was occurring in Europe during the same time period (E. Weber 1976). The failure of the state made the type of elite identity mobilization discussed in this book especially important. In many African countries, local political leaders worked to create fairly large identity groups, since small ones were often not electorally viable (Posner 2004). In India, as we saw, there have been some similar trends in the twentieth century towards group consolidation (Kothari 1970). However, even late into the colonial period, electoral incentives were not particularly important, and groups sought instead to rise in status within a patrimonial system. Small groups would use some gain in wealth and power to improve their own social status, creating over time a continuum of small groups each associated with a particular occupation, immigration event, or religious movement. The small size of the groups did not pose a particular political obstacle, and in fact the expansion of the groups would inevitably have led to the inclusion of some people of lower status, and a corresponding loss in status for other members of the group. While often inchoate in the pre-colonial period, the identities of these many small groups were reified by the ranked caste mobilization of the colonial period, creating hundreds of caste identities recognized by the state and with an association dedicated to its perpetuation. The result of this is that India has one of the highest levels – perhaps the highest – of ethnic diversity in the world. Conventional codings that use region and caste category already rank India as extremely diverse, behind only a dozen or so African countries (Fearon 2003). The enormous number of castes in India (the Anthropological Survey counts more than 4,000 [K. S. Singh 1998]) means that no one has ever attempted to calculate fractionalization indices for India using jati. Lee (2019) calculated that most Indian states have fractionalization measures between 0.93 and 0.97 (equivalent to the Democratic Republic of Congo, the third most fractionalized country in the world), and no state has a level of fractionalization less than 0.84 (equivalent to Ghana, the 13th most fractionalized country in the world). Nationally, this would almost certainly make India the most diverse country in the world, with predictable consequences for interpersonal cooperation.

Conclusion

175

The wealth of groups Not all scholars have accepted the claim that diversity of salient identities is automatically associated with lower levels of cooperation and public service provision. A number of persuasive recent accounts have claimed that the while the presence of multiple ascriptive groups can have negative political effects, these effects are in fact the product of economic inequality between groups (Baldwin and Huber 2010) or historical differences in the distribution of political power between groups (Lee 2018a). Where wealth and power are disproportionately concentrated in some groups but not in others, cooperation between groups is difficult, since differences in preferences align with ascriptive divisions. These arguments reinforce older accounts of the advantages of cross-cutting, relative to reinforcing, social cleavages (Dunning and Harrison 2010). Recent accounts of Indian society have emphasized that while jati is not perfectly correlated with wealth and education, there is a high degree of correlation, possibly higher than the correlations in other countries.1 In Baldwin and Huber’s (2010) dataset, India is coded as having the third highest level of between-group income inequality in the world, behind only Brazil and South Africa.2 Similarly, political power, while much more broadly diffused across groups than at any other period in Indian history, is still correlated with identity. Not only are members of the upper castes disproportionately represented among politicians, but they make up a disproportionate share of bureaucrats as well (Witsoe 2013). More broadly, most Indian regions possess ‘dominant’ caste groups which have secured disproportionate access to political power, often building on pre-colonial and colonial foundations (Frankel and Rao 1989; Srinivas 1994; Lee 2019). These high levels of association between socioeconomic status and identity are a natural consequence of the existence of ranked forms of identity for centuries. In an unranked system, social and economic change will gradually create elites even within previous poor groups. In the United States, the past century has seen many African-Americans become middle 1

2

Some authors refer to this correlation as creating a ‘ranked’ identity system in India, citing Horowitz (1985), though, like Horowitz, they do not specify the level of correlation necessary to make an identity system ranked. See Chapter 2 for a fuller discussion of the difference between these two uses of the term. Note that Brazil, as noted earlier, has a similar history of ranked identity rhetoric.

176

From Hierarchy to Ethnicity

class and even wealthy. In India, the Dalit middle class has grown in recent decades (Ahuja and Ostermann 2016). In a ranked system, however, the winners assimilated into higher groups, strengthening rather than weakening the correlation between power and identity. The fact that Kunbis are generally poorer than Marathas in Maharashtra, for instance, is a product of the fact that for centuries a successful Kunbi lineage would identify as Maratha (Deshpande 2004). In West Bengal, the fact that the Mahishya caste is much wealthier than the Kaibartas is a product of the fact that the Mahishyas were formed from the wealthy subsection of the Kaibarta community. Since occupation influenced both ranking and socioeconomic status, members of identities associated with low-status professions tended to remain poor. If they abandoned their occupation for something more remunerative, they would also gradually lose their caste identity. Since the ranked system created both the means and the incentive for successful individuals to join already successful groups, it left India a legacy of an unusually tight correlation between group identity and social status, one that is only very gradually being reversed by the emergence of wealthy individuals from traditionally low-status castes. Does this tight association between caste and status explain the high salience of caste in Indian politics and the high levels of social conflict along caste lines? Certainly, there is growing evidence that even high-level policy decisions in India are made on the basis of caste, since policies with consequences for specific castes and occupations will have positive and negative effects on specific jatis (Lee 2019). Similarly, given the laboratory evidence that caste is the most important predictor of interpersonal transfers (Hoff, Pandey, and Brooks 2010), it is not surprising that wealthy Indians are less likely to support redistribution policies than they would be if they were confident that their taxes would support members of their own group.

Casteism versus the upper castes Chapter 6 portrayed caste identities as being central to an understanding of contemporary Indian politics. Individuals are encouraged to identify with jati identities by a combination of careful activism and the promise that they will be rewarded by the distribution of state resources (either through reservation or through other means). This, in turn, encourages individuals to vote for leaders and politicians associated with this identity. While no one

Conclusion

177

would deny that ‘caste equations’ are very important politically in much of India, this account is something of a caricature. Even in the most politicized castes, individuals often vote for politicians outside ‘their’ caste or caste party. This behaviour may stem from clientelistic distribution, the existence of multiple competing leaders within the caste, the pressures of a first past the post electoral system, or the appeal of competing identities such as religion. However, perhaps the biggest problem in explaining Indian elections as the simple accretion of mobilized jatis is the equivocal position of the upper castes. As in the colonial period, members of the most educated groups disdain explicitly caste-based political involvement. While Brahmins and Banias vote for some parties more than others (in recent years, especially for the BJP and the Congress), in no Indian state is there a more or less explicitly Brahmin of Bania party, like the caste parties of northern India. Moreover, at both the elite and mass levels, the association between the upper castes and their chosen parties has been weaker than for other caste groups – upper caste politicians remain scattered among the parties, even parties associated with lower caste jatis. For instance, in Uttar Pradesh in 2012, Jaffrelot and Verniers (2012) found that upper castes made up 57 per cent of BJP candidates and 52 per cent of Congress candidates, but also 21 per cent of the Jatav-led BSP and 25 per cent of the Yadav led SP.3 While 89 per cent of Yadav MLAs were from the SP (relative to the party’s) 57 per cent of the seats overall), only 40 per cent of Brahmin MLAs and 29 per cent of Rajput MLAS were from the BJP and Congress (which collectively won 20 per cent of the seats). At the individual level, at least until the recent BJP wave elections, the upper castes in Uttar Pradesh did not vote in the same disproportionate numbers for the BJP and Congress as the Yadavs or Chamars voted for ‘their’ parties (Verma 2007, 2012). Jeffrey Witsoe’s (2013) fieldwork provides some ethnographic insight into the phenomenon. Even at the height of Laloo Yadav’s highly polarizing period in office, a sizable segment of upper caste voters and politicians were willing to cooperate with his regime for the sake of particularistic benefits. These north Indian examples are to some extent stories of failure, in that in some major parties the upper castes are forced to cede second place to members of other groups. In many other Indian states, members of the upper castes are important in all major political parties. West Bengal, where Brahmins and Kayasths are the largest groups in both the Trinamool 3

These statistics refer only to the 378 seats won by the four major parties.

178

From Hierarchy to Ethnicity

Congress and Left parties, is a notorious case, but Madhya Pradesh, Chhattisgarh, Himachal Pradesh, Kerala, and Orissa have all seen leaders from the most educated groups playing a disproportionate role in all the major parties, with lower caste vote banks serving as subsidiary parts of their coalition (Jaffrelot and Kumar 2012; Lee 2019). The lack of explicitly caste-based mobilization among the most educated groups is also apparent in the rhetoric of politicians from these groups. Many upper caste politicians are vocal advocates for broad identities: the nation (for Congressmen), the religiously defined nation (for BJP members), and the social class (for Communists) – anything, in fact, other than caste. Moreover, ‘casteism’ (defined as explicitly caste-based appeals) is largely denigrated in the upper caste media, and even more scholarly accounts have tended to attack caste-party leaders (Thakur 2000), and treat voting for these parties as a normatively concerning puzzle to be explained (Banerjee et al. 2007). This pattern – with some caste identities highly politicized but the wealthiest groups seeking some other basis for political engagement – is similar to that found in the colonial period, where the most educated groups did not petition or exhibit the degree of visible associational activity of moderately educated groups. While caste mobilization is necessary for the political success of some leaders, for members of the traditional elite it is not only unnecessary, but dangerous, since the polarization of the electorate along caste lines would cut them off from many potential supporters – and, in many cases, leave them in the minority. The presence of these elites, and the continued potency of their political appeal, mean that India is not a pure ethnic democracy of the type discussed by Posner or Horowitz.

The ‘muscle castes’ The other continuing legacy of the differential mobilization of caste identities in the twentieth century is the political gap between those castes with a long tradition of identity mobilization and those without it. The early mobilizing castes, what one informant called the ‘muscle castes’, are now very well represented in the political sphere, often to the same degree as the upper castes: this book has discussed many examples of such groups: Yadavs and Kurmis in Bihar, Jats in Haryana, Kammas and Reddis in Andhra Pradesh, and Lingayats and Vokkaligas in Karnataka. In many cases, these groups are now so well established that multiple parties and

Conclusion

179

leaders compete to be considered the true representative. In Indian political parlance, they hope to ‘consolidate’ the ‘vote bank’. The ‘dominance’ of these groups leaves some other groups under-represented in the political process. In the colonial period, very poor groups had little caste-based political engagement. In the post-independence period, the extremes for poverty seen in the colonial era are less in evidence – even the very poorest groups have a few educated members capable of forming a sabha or circulating a petition. However, these efforts are more recent and tentative than those of the dominant castes, and their hold over the loyalties of voters more tenuous. In many cases, poverty is difficult to separate from a small group size, since groups that make up only 1 or 2 per cent of a state’s population are less attractive vote banks, and less likely to reward the effort of ambitious politicians. This is particularly true if the group in question is spatially scattered, like many of occupational castes – the groups that formerly served as village potters, blacksmiths, carpenters, and so on. The lack of caste-based mobilization is reflected in the survey data, which usually find members of the ‘other OBC’ and ‘other SC’ categories much less consolidated behind a single party than any of the enumerated OBC groups (Shastri, Suri, and Yadav 2009).4 In this sense, modern India resembles the colonial era, in that caste mobilization is still to some extent non-linearly correlated with socioeconomic status. As a result, sections of Indian society that make up a very large section of the population collectively have little or no political voice. Lee and Ramachandra (forthcoming) find that in Delhi, members of OBC groups outside the most powerful landholding groups made up at least 20 per cent of the population but less than 1 per cent of corporation members. The contributors to Jaffrelot and Kumar (2012) similarly find that the ‘extremely’ backward castes, mostly occupational, are extremely under-represented among candidates. Lee (2019) argues that the interests of these groups have taken a back seat in policy decisions relative to dominant castes. Similarly, accounts of Indian villages often find members of these groups still tied in more traditional patron–client relations (Price and Ruud 2012; Anderson, Francois, and Kotwal 2015). For these groups, the promise 4

Note, however, that since these groups are not enumerated separately, we cannot eliminate the possibility that they are voting by bloc as jatis, but not as a category.

180

From Hierarchy to Ethnicity

that the decline of upper caste power would increase their political power remains, in certain key aspects, unfulfilled.

Social legacies of ranking While the main focus of this book has been on political actions, it is worth noting that the ranked system still influences other areas of social life. While the idea that upper caste behaviour is ideal and worthy of emulation has been banished from the political sphere, it still retains some influence on more intimate types of behaviour. In marriage, for instance, lower caste brides appear to prefer grooms of a higher caste to grooms from other caste groups on the same or lower economic level, though within the context of a large preference for within-caste marriage (Ahujar and Ostermann 2016). Chapter 6 showed a sizable portion of Indians, even from lower castes, still practise untouchability: 31 per cent of the upper and intermediate castes, 27 per cent of OBCs, and 10 per cent of Dalits, according to the 2011–2012 India Human Development Survey. There also remain some parts of India where the state has little reach, and patrimonial politics still continues. While lower caste members in these areas may be aware of political developments elsewhere, they are often unable to escape the coercive and economic pressures that keep them bound to upper caste patrons. The economic pressures that force individuals into clientelistic relations are now well studied: in an environment where uncertainty is high, clientelistic ties to a landlord or employer may be attractive even if they provide lower wages than other arrangements, or involve symbolic deference of various sorts (Anderson Francois and Kotwal 2015). To supplement these pressures, patrons may use physical force against individuals who reject either their economic or symbolic predominance. Shah et al. (2006) document many cases of violence against Dalits, which often reflect an attempt to defend local power against Dalit attempts to undermine it. In many cases, as in classic ranked identity projects in India, status is linked to Sanskritic ideals of cleanliness. Dalits are forbidden to use the same wells and dishes or to let their shadow touch upper caste members and are forced to perform unclean tasks such as scavenging. In some cases, however, the status markers are novel: Dalits are forbidden to have the same haircut as members of the dominant caste, or to send their children to school. While these types of dominant caste behaviours are often resisted by lower castes, in many cases lower caste

Conclusion

181

members will cooperate, at least in public; Anderson, Francois, and Kotwal (2011), for instance, found higher levels of reported community feeling in villages with a dominant caste than villages without one. Absent an intervention from urban activists, such actions remain local and often sub-political. However, they indicate that the logic of the ranked system is still valid. In an environment where personal protection and access to valuable resources depend more on contacts with elites than on voting power, the behaviours of elites become the centre of a status ordering, with the status being defined as the ability to maintain these behaviours and prevent others from doing so.

Implications for the literature South Asia The analysis in this book has several implications for how we understand caste identities. Some of these are already fairly well known. Chapter 4 replicates quantitatively some of the major qualitative findings in the literature. Far from being primordial and static, caste identity is changing and socially constructed. Chapter 5 showed that there was considerable change and inconsistency not just in the salience of caste but in the idea of caste itself. In the post-colonial era, these changes became virtually universal. Existing work on caste politics has often mentioned the relationship between the political power of castes and their level of socioeconomic status (Srinivas 1966; Frankel and Rao 1989; Jaffrelot and Kumar 2012). Chapter 4 demonstrated that differences in political engagement existed, and that education tends to increase caste-level political participation. However, it also shows that caste activism, narrowly defined, becomes less common among very educated castes, both patterns that exist into the modern period. It thus shows how many ‘dominant’ castes became dominant, or were able to translate numbers or power within villages into a voting bloc capable of bending the political system to their interests. It also explains why the electoral strategies of politically powerful caste groups are so variable, with upper castes emphasizing caste much less than members of newly educated groups. The theory also has implications for the future, suggesting that the much-discussed upsurge in caste-based party politics in India in the twentieth century may eventually give way to division on broader

182

From Hierarchy to Ethnicity

cleavages – that is, as more groups become highly educated, caste mobilization will become less valuable. There are some tentative signs that this process is occurring in India. In Delhi, India’s wealthiest state, S. Kumar (2013) argues that in recent years politics has tended to polarize along class rather than caste lines. Similarly, many of the most prominent caste parties in northern India have had difficulty making head in urban areas, which have tended to support the Congress and BJP. However, this trend remains highly speculative, and given the persistence of economic and educational inequality between castes, it appears unlikely that lower caste politicians will find it easy to compete with upper caste candidates in pure valence terms soon. Finally, the argument grapples with the element of the caste system thought to be most case-specific: its emphasis on ranking. The hierarchical aspects of the caste are often considered to be unique to the subcontinent, and this unique element of Indian civilization is often used as an explanation for particular features of India’s political economy (Lal 1988). While not questioning the importance of caste and social hierarchy in modern India, this book places these concepts in a comparative context. It shows that the emphasis on ranked identities in South Asia reflects a specific set of political circumstances that were paralleled in many other countries: a patrimonial political system embedded within a somewhat articulated economy. This theory also explains why ranked ideals are much more prominent in the subcontinent than in other parts of the world. Unlike China, early modern India possessed a political system that was decentred, unstable, and patrimonial. Unlike Africa, the Indian economy was able to support a set of ritual specialists and subtle differences in individuals’ behaviour and consumption patterns. The theory also helps bridge the gap in the literature between the ranked, status-driven contention thought to be characteristic of the caste system in accounts of the pre-1947 period (S. Bayly 1999), with the conflict between resource-hungry vote banks described in accounts of contemporary caste politics (Chandra 2004). As this political system gave way to the more participatory one we have today, there has been a decline in the ranked nature of the system, to the point where it is not unreasonable, at least for political purposes, to treat caste identities as unranked.5 Far from being unique and unchanging, caste politics have followed a trajectory paralleled in many other regions of the world. 5

This does not necessarily extend to all social situations.

Conclusion

183

Colonialism One of the most common explanations for the changes in identity politics in the nineteenth and twentieth centuries is that they reflect the policies of the colonial state, either accidental or as part of a conscious strategy of divide and rule (Laitin 1986; Cohn 1987; Gupta 2000; Mamdani 2005). This book modifies this view. While colonial institutions provided a set of vocabulary to express social differences and a forum (the census) to express them, there were massive differences in the uses to which different groups put these emergent possibilities. To paraphrase E. P. Thompson, castes were present at their own creation. The crazy quilt of caste institutions and claims in colonial India reflected the different political needs of Indian elites as much as the policies of the colonial state. This is not to say, of course, that states are irrelevant to the development of identities. There is considerable evidence that activist states can reify and shape individual identification. Indeed, the Indian state’s extensive affirmative action programmes appear to be engaging in just such an effort at the present moment. However, with a few exceptions, the colonial state was either unable or unwilling to impose a consistent policy like that being imposed in France in the same time period (E. Weber 1976). Absent a nation state, the shaping of identity in India was left to an emergent class of ethnic entrepreneurs.

Ethnic politics Much of the theory in this book draws on existing ideas in ethnic politics literature. It modifies these ideas in two ways. First, it shows that socioeconomic status serves as a scope condition for ethnic mobilization, an idea implicit in much of the literature but poorly articulated. Second, it finds that this tendency is non-linear, and people in highly educated groups tend to disdain narrow ethnic appeals. These ideas supplement the emphasis in the existing literature on group size (Chandra 2004; Posner 2004), showing that the effects of social modernization emphasized in the traditional literature on nationalism can sometimes be just as important as strategic coalition building. Given that many groups do not have politicians making claims on their behalf, ethnic politics can be an oddly asymmetric game, with some groups becoming variables in complex vote bank equations and others either unengaged in the political process and not explicitly conditioning

184

From Hierarchy to Ethnicity

their involvement on narrow identity considerations. This book has briefly discussed some of the ways in which this type of differential mobilization could affect the political process. At a minimum, it has created a set of winning and losing groups in terms of descriptive representation and policy formation, meaning that the mobilization of new identities has still not created an egalitarian distribution of power. Moreover, it creates a situation where many voters feel cross-pressured not just between rival identities, but between identity appeals and the valence appeals of elites. Further research is needed on how voters resolve these appeals, and more broadly how ethnic politics responds to attempts by elites to delegitimate it. Second, the theory moves ranking to the centre of the conversation about identity construction. As Abdelal et al. (2006) suggested, there is far more to identities than their theories of category. Many human identities are ranked, which involves not simply the correlation of wealth and identity but an entirely different way of conceiving identity, one where the division between in and out, while important, is secondary to the division between high and low. While the patrimonial institutions that produced them have faded away, the identity structures that emerged from these systems differ in important ways from identity systems that were never ranked, with potentially large consequences for development. The idea of the ‘modern’ ethnic group, with its external boundaries and internal egalitarianism, is a recent and contingent outcome, and the process by which it became the natural way to theorize human diversity is worthy of serious examination.

Appendix A

Data Measurement The unit of analysis The unit of analysis is caste-province-year, but both ‘province’ and ‘caste’ require some explanation. Colonial India was divided into areas ruled by the British government and areas in which sovereignty was delegated to native princes, usually descendants of early British allies. The directly ruled areas were divided into fairly large provinces, of which the most important were: the United Provinces, Bengal, Bihar and Orissa, Bombay, Madras, Punjab, the Central Provinces and Berar, and Assam. Four very small provinces (two of which are overwhelmingly Muslim border areas) have been ignored. The indirectly ruled areas were divided into three main groups: four large states (Jammu and Kashmir, Hyderabad, Mysore, and Baroda) that enjoyed a direct relationship with the central government, two large groups of states (the Rajputana Agency and the Central Indian Agency) which reported to a political agent appointed by the central government, and a large number of states who reported to the political department of the local British province. The policy followed here has been to report the large states and agencies separately, as the equivalent of British provinces, and include the other states with the province that controlled them. The only exception to this was the large princely state of Travancore, which published its own census report and has been treated as separate from Madras. There were several changes to provincial boundaries during this period, so I have consolidated provinces to keep the unit of analysis consistent. For the 1891 and 1901 census years, Berar has been treated as part of the Central Provinces even though it was in fact independent, and for the same years, Bihar has been treated separately even though it was in fact part of Bengal. In 1931, the recently created Western Indian States Agency is

186

Appendix A

treated as part of Bombay, and for 1911, 1921, and 1931, Gwalior state is treated as part of the Central Indian Agency. Two major challenges in the collection of the data were the multiplicity of small castes in India and the large number of alternative names (and alternative spellings) for castes in different areas. The census superintendents generally tried to keep jatis separate, but at times they differed as to what constituted an independent jati (Conlon 1981). The most usual cross-year difference is that in the later census years, officials sometimes clubbed together castes practising the same occupation but using different names and speaking different languages. The practice followed is to use the caste classifications used in 1891 (when nearly all the jatis were listed separately) and to divide aggregated groups using their relative proportions in the most recent year in which they were listed separately. The results are robust to the exclusion of these composite castes. Hyderabad in 1901 and Kashmir and Assam in all years used highly aggregated and rapidly changing groupings of castes, making it difficult to trace any continuity from year to year. These province-years have been excluded from the analysis. Sub-castes (on which we have very limited information) will be treated as part of their parent jati. This choice is largely pragmatic: There is a large amount of ethnographic and case study evidence that jati was and is the primary identity on which castes are mobilized politically, and the primary focus of individual’s identification in a local context. Note that many of the problems described by Conlon (1981) in using census caste data arise in trying to track groups considered by the colonial authorities to be sub-castes. As an economy measure, census officials did not always tabulate occupational and literacy information for all castes, and in the depression year of 1931, a few provinces (notably Madras) did not tabulate the population results for all castes. To minimize this problem, I have confined the analysis to castes with over 0.5 per cent of a province’s Hindu population at the 1891 census, among whom these omissions are less severe (these castes also tend to be much more consistently named and classified than small ones).

Coding petitions The main procedures for coding petitions are straightforward. However, some special cases require additional explanation.

Appendix A

187

Within each census year, I focused only on castes that had over 0.5 per cent of a province’s population in 1911. This avoids generalizing from tiny castes, which both tend to have a larger amount of missing data and to suffer from inconsistent census enumeration policies from year to year (Conlon 1981). The larger castes, by contrast, are relatively stable. Additional quantitative data on caste population, literacy, and occupation was collected from the tables volume of each census report. While population data is available for nearly every caste-year, some provincial census superintendents only collected occupational data on castes that they considered important. For this reason, models that include the occupational variables lose a number of observations. A complicated chain of events occurred in south India at the 1931 census. In Madras, Mysore, and Hyderabad, activists from untouchable groups demanded that their groups be not counted separately, but combined under a common name (depending on the region, Adi-Hindu, Adi-Dravid, Adi-Andhra, and Adi-Karnataka). The Adi prefix emphasized the claim that these groups were in fact the original inhabitants of south India and had been enslaved and degraded by later invaders. The census policy towards these claims varied by state: in Hyderabad, the castes were still listed separately, but in Mysore they were combined in the tables, while in Madras both the old and the new names were given as options, which respondents were free to choose between. The policy I followed in these cases was to treat all the old untouchable castes in these provinces as having submitted petitions while ignoring the new castes. The major results are robust to the exclusion of these province-years. Madras in 1931 also contains the only two cases of jatis not petitioning because their demands were fully granted in a previous year. They were dropped from the analysis. In some cases, the census reports do not mention petitions in their reports, or mention rejecting them without going into details about the castes submitting them. All castes in these province years were coded as missing data. The only instances in which all cases in a province-year were coded as zero were when the superintendent specifically noted receiving no petitions. In 1931, some sabhas began mailing petitions for reclassification to every provincial census officer in India, even in areas where they had no organizational presence. Since it is crucial for the research design that petitioning reflect some local group activity, these mailed petitions are also coded as a zero, unless there is some indication they were supported by a

188

Appendix A

local organization. Similarly, two cases where petitions were submitted on behalf of a group by non-members of a group (in both cases, the provincial branch of the Arya Samaj) are also coded as zero. In the main analysis, I have excluded non-Hindu castes and tribal groups. While caste divisions do exist among Muslims, Sikhs, and Christians, they do not have the normative meaning that they do in Hinduism, since all these religions officially disapprove of caste (Ahmad 1978). In addition, while some descent groups within the non-Hindu religions are precisely analogous to modern castes (often being the caste from which the member of the group converted) others reflect denominational divisions (‘Syrian Christian’) or fictive descent groups (‘Pathan’) that were often loosely defined both in practice and in the census enumeration. In addition, many of these groups are missing data for one or other of the main variables. Tribal groups also differ from caste Hindus in their spatial isolation and unique relationship to caste as a category (though assimilation into the caste system is an ancient pattern among ST elites).1 Including these ethnic categories in the analysis would slightly strengthen the reported results. A few petitions were submitted on behalf of subgroups of a caste group recognized by the census, requesting a name change for themselves rather than the caste as a whole. These cases are of great theoretical interest and generally indicate that the elite of the caste was attempting to split off and form a distinct group. In the analysis, these petitions are treated as being submitted by the whole caste. However, excluding these cases does not affect the reported results.

Additional variables Literacy is the primary measure of education in the quantitative models.2 The square of the literacy rate will be used to test the quadratic shape of the function. Census definitions of literacy changed slightly over the years, with 1

2

There are in fact three instances of petitions by tribal groups in this period, though the overall rate is very low, and in one instance there is evidence that they were encouraged to petition by an outside group of upper caste Hindus. There are six instances of petitioning by non-Hindu groups. In some province years, caste-level literacy was only reported for areas in which the caste was especially prevalent. The reported results are robust to the exclusion of these cases.

Appendix A

189

the 1901 census using, at least in some provinces, a more generous definition of literacy. Appendix B shows that the results are robust to excluding the 1901 data entirely. Some models include control variables for the male population for each caste,3 which I use to calculate their proportion of the provincial population and their absolute male population.4 The absolute population is included to control for the possibility that petitioning is stochastic across individuals, and that simply having a large number of people (as would even a small caste in a large province) increases the probability of petitioning significantly. The political changes that occurred at the province level were obviously correlated with the wide variety of social and economic changes taking place in colonial India. To account for such time-varying confounders, most of the statistical models include three control variables often associated with the level of economic and social modernization: the provincial literacy rate, the proportion of the population living in urban areas, and the proportion of the population engaged in public administration, all taken from the census tables.

Caste status Most of the reported models include caste fixed effects, and thus account for a wide variety of unobserved cultural attributes of groups, including the social status that colonial census takers ascribed to them. In a few models, however, it is necessary to measure this status (the ‘status quo’ against which petitioners were protesting) directly. The caste hierarchy dummies are coded using the infamous ‘tables of precedence’ compiled for the 1901 census, supplemented with information from the Castes and Tribes series for each province that were published in the same period. In all but a few cases, the classifications agree with those made by the post-independence government for the purpose of affirmative action. The status variable is constructed as an ordinal variable, with each jati being assigned to a category. From highest to lowest, the categories are Brahmins, other clean twice-born castes (‘upper castes’), high-status 3

4

Census policy was not to ask women their caste identity, which was recorded as that of their husbands or fathers. The two population variables are perfectly collinear within province-years.

190

Appendix A

cultivating castes (‘intermediate castes’, ‘middle castes’, ‘dominant caste’), low-status cultivating castes (‘upper OBCs’, ‘unclean Shudras’), lowstatus occupational castes (‘lower OBCs’, ‘artisan castes’), and former untouchables (‘Dalits’, ‘Harijans’). While the terminology varies, this sixfold classification is familiar to India scholars, as it lies at the heart of most previous work on caste politics (Frankel and Rao 1989; Jaffrelot 2003; Jaffrelot and Kumar 2009) and is the format used in most contemporary surveys. While there is much blurring at the edges, particularly among the Shudra categories, this scheme seems to capture certain important aspects of status hierarchy in India.

Appendix B

Statisical Tests Modelling petitioning Model selection The tests of Hypotheses 1 and 2 focus on variation in petitioning within castes, using a set of caste fixed effects. However, petitioning may vary by province (due to historical and institutional differences) and by year (due to time-related trends). Observations are also not independent within castes (since jatis would find it easier to organize if their caste-fellows in other provinces had already done so) or within province-years (since groups might be more likely to petition if their neighbours were already doing so). The choice of a statistical model to analyse this data presents a fairly standard set of trade-offs between bias and variance. Unfortunately, it is impossible to estimate the ‘pure’ FE model with fixed effects for jati-province and province-year. While such a model would certainly be free of all these sources of bias, the broad and short structure of the panel would mean that this model would use a large number of degrees of freedom. Moreover, this model would have a very small N due to the loss of the large number of observations that do not vary within jati-province.1 The basic specifications are thus based on a modified fixed effects model, with fixed effects for jati, province, and year. This last model differs from a conventional fixed effects model only in that the within-jati variance is not only across time but also across province (for jatis located in multiple

1

Overall, 46 per cent of the variance in petitioning is within the 381 jati-provinces, while 65 per cent is within the 188 jatis.

192

Appendix B

provinces). It thus controls for most of the cultural and social unobserved variables associated with individual’s caste. The model estimated is: Ypjt = αp + β j + γ t + δLitpjt + ρLit2pjt + θXpjt + 'pjt

(B.1)

where Ypjt is the log odds of petitioning for a given jati-province-year, αp , β j , and γ t are vectors of province, jati, and year fixed effects, and δ and ρ are the parameters of interest, and X is a vector of controls. While this fixed effects model eliminates a great deal of bias, it is still relatively inefficient due to the very large number of castes. This is especially problematic in some of the specifications that test the robustness of the covariates to the inclusion of controls (many of which are available for only a portion of the data). Most of these specifications thus use a nested mixed-effects logistic regression model with random effects for jati and province-year and fixed effects for province, year, and caste ritual status. The random effects adjust the coefficients to take into account the probable non-independence of observations within province-year and within jatis. The fixed effects take into account the direct influences of time, geography, and the traditional social status (the fixed point against which castes were trying to appeal). This model is estimated as: Ypjt = αp + β s + γ t + σ pt + τ j + δLitpjt + ρLit2pjt + θXpjt + 'pjt

(B.2)

where β s is a vector of fixed effects for each level of caste status, and where σ pt and τ j are vectors of IID random effects for each province-year and jati.

Testing H1 and H2 The curvilinear relationship between group literacy rate and petitioning is shown in Table B.1. Model One tests the simple quadratic relationship between literacy and petitioning. The literacy rate of a group is a positive predictor of its propensity to petition, but this effect reverses at high levels of literacy. Both results are statistically significant. Model Two runs the basic fixed effects model, with the key independent variables of interest, the literacy rate, and its square. The model also includes a set of variables associated with the level of modernization at the province level: the provincial literacy rate, the rate of urbanization, the per cent of male workers employed in public services, and employment in agriculture, none of which affect the value of the literacy variables, which retain their

Appendix B

193

strong quadratic relationship with petitioning. The predicted values from this model are shown graphically in Figure 4.3. Model Three of Table B.1 introduces a set of variables drawn from the occupational data in the census tables, including the percentage of caste members cultivating land and the percentage involved in public administration. Of particular interest is a third variable, the percentage of caste members engaged in the caste’s traditional occupation. The effect of this variable on petitioning is substantial and positive: traditionally employed castes are more likely to petition than non-traditionally employed castes, though whether this represents a challenge to the system or a rebellion against it is unclear. However, this variable does not affect the value of the literacy variables, which retain the same inverted U relationship with petitioning. Model Four uses a mixed effects regression and reruns the basic model using the mixed effects model. As we might expect, the more efficient model produces substantially the same coefficients with greatly reduced standard errors. The results from 1901 might be plausibly thought to be biasing the results, given the effect of Risley’s programme of classification and the fact that some census years used inconsistent definitions of literacy. Model Four reruns the mixed effects model without the 1901 data, and gets identical results. Model Seven of Table B.1 replaces the literacy variables with the rate of caste employment in industry. This is intended as a check on the validity of using literacy as a measure of the broader concept of socioeconomic status. The industrial ownership variables produce results similar to literacy: the effect of literacy on petitioning is positive but attenuated at higher levels. This result gives us some confidence that the main results are not a product of some factor unique to education. Model Eight uses an extremely conservative version of the model, which includes fixed effects for every jati-year and every province-year. These models thus control for every potential confounding time trend at the regional level, and every potential census-specific effect, such as a particularly high rate of granting or a particular emphasis on ranking by a census superintendent. Similarly, they focus only on variation in literacy within jati-years, controlling for all caste-specific time trends. Understandably, the number of observations is much smaller in this model than in the others, due to perfect prediction: the majority of observations are either within province-years where no group petitioned or within

194

Appendix B

jati-years where no group petitioned.2 Despite the much smaller sample size, the results in this model are much larger than those in the other models in substantive size, and still statistically significant at the 10 per cent level. This indicated that the non-linear effect of literacy is a pattern that persists even when we account for a very wide set of unobserved confounds. Table C.4 reproduces these results using two alternative models, an OLS model (which, while less efficient, does not suffer from the perfect prediction problems of the logit model) and a model in which all the key independent variables are lagged (to account for a possible reverse causal relationship between caste mobilization and education). These models produce results quite similar to those in Table B.1.

Participation among the highly educated H2 predicted that very high levels of group socioeconomic status should be associated with participation in non-identity-based politics. This hypothesis is tested in Table B.2. Model One of Table B.2 regresses group literacy and size on caste-specific rates of employment as gazetted (senior) government officials. The squared literacy rate of the group has a statistically significant positive effect on bureaucratic employment, the only significant caste-level predictor. The same result holds true for caste-level attendance at the annual sessions of the Congress (Model Two of Table B.2). While their slightly less advantaged peers dominate caste mobilization, the literate urban castes seem to suffer from no handicap in the political system as a whole.

Alternative explanations Education is obviously not the only factor which may affect caste mobilization. This section discusses several alternative mechanisms that may be influencing caste petitioning and tests whether they affect the results in Table B.1. One possible alternative explanation for the results is the variation of the structure of the caste system and the intensity of its norms, in particular the perceived higher salience of caste and great ‘completeness’ of 2

Note that this means that all jatis that were unique to a province are not included in this model.

Table B.1 Main results: logistic regression with petition as dependent variable Variables

(1) Petitioning

(2) Petitioning

(3) Petitioning

(4) Petitioning

(5) No 1901

(6) Petitioning

(7) Petitioning

(8) Petitioning

Population as Proportion Population 0000s

2.298 (4.656) 0.00732 (0.00450) 7.958*** (2.388) −13.78*** (4.280)

10.72* (5.544) 0.00173 (0.00569) 10.12*** (2.989) −14.37*** (4.688)

11.08 (7.013) 0.00351 (0.00710) 14.27*** (4.260) −19.39*** (6.535) −0.137 (0.75) −11.23 (9.26) 2.29*** (0.817)

11.66** (4.889) 0.00175 (0.00437) 7.123*** (2.748) −10.24** (4.212)

12.64** (5.073) 0.00101 (0.00469) 6.938** (2.901) −10.22** (4.387)

11.45** (5.736) 0.00245 (0.00511) 10.47*** (3.446) −14.66*** (5.391) 0.346 (0.687) −6.869 (8.189) 1.559** (0.640)

10.31* (5.558) 0.00519 (0.00507)

21.58 (27.94) 0.00220 (0.00213) 38.72* (21.00) −44.47* (25.65)

Literacy Rate Lit Rate Sq. Land Cultivation Rate Government Employment Rate Traditional Occ. Empl. Rate Jati Industry Owners Rate Jati Industry Owners Rate Sq.

251.0** (117.4) −12,947** (6,443) contd.

Variables

(1) Petitioning

(2) Petitioning

(3) Petitioning

(4) Petitioning

(5) No 1901

(6) Petitioning

(7) Petitioning

(8) Petitioning

Constant

−2.385*** (0.289) 891

5.375 (5.974) 891

14.72 (12.34) 579

5.184 (5.701) 891

3.784 (7.263) 794

8.554 (11.67) 579

12.99* (7.849) 628

−11.82** (5.709) 207

0.63 NO NO NO NO NO NO NO NO NO

0.75 YES NO YES YES YES NO NO NO NO

0.79 YES NO YES YES YES NO NO NO NO

0.91 NO YES YES YES YES YES YES NO NO

0.91 NO YES YES YES YES YES YES NO NO

0.90 NO YES YES YES YES YES YES NO NO

0.90 NO YES YES YES YES YES YES NO NO

NO NO NO NO NO NO NO YES YES

Observations C-Stat Jati FE Caste Status FE Province FE Year FE Prov.-Year Controls province-year RE Jati RE Jati-Year FE Province-Year FE

Standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 The first three models are time series logistic regressions with fixed effects for jati, province, and year. The last three models are mixed effects logistic regression models, with fixed effects for caste status, year, and province and random effects for jati and province-year. Only the fixed effect constants are reported. The Province-Year Controls (proportion urban, proportion in public employment, provincial population, and provincial literacy rate) are not reported for reasons of space.

Appendix B

197

Table B.2 Alternative measures of participation Variables

(1) Government Officers

(2) Congress Delegates

−0.00238 (0.00182) 4.39e-08 (2.13e-07) 9.73e-05 (0.00107) 0.00423*** (0.00142) −0.00945* (0.00521)

1.45e-05 (0.249) 0.45e-05 (0.249) −0.109* (1.79e-05) 0.407*** (0.0612) 1.018 (1.121)

Observations C Stat

674 0.49

408 0.41

Caste Status FE Province FE Year FE Province-Year Controls province-year RE Jati RE

YES YES YES YES YES YES

YES YES YES YES YES YES

Population as Proportion Population 000s Male Literacy Rate Male Literacy Rate Sq. Constant

Standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 These are mixed effects linear regression models, with fixed effects for caste status, year, and province and random effects for jati and province-year. Only the fixed effect constants are reported. The Province-Year Controls, proportion urban, proportion in public employment, provincial population, and provincial literacy rate, are not reported for reasons of space.

the caste system in northern India. In its basic form, much of this cultural heterogeneity is already accounted for by including province and jati-level fixed effects. If norms about caste are stronger in Bihar than in Tamil Nadu, or among Brahmins than among Marathas, this should be accounted for in the fixed and random effect coefficients in the models. Local variations in social status, however, might vary at the jati-province level. Model One of Table B.3 adds a variable for the percentage of the provincial population with a higher social status than the caste in question. This is intended to account for the uneven ‘spread’ of the caste system across regions of the country: due to the smaller number of high castes in

198

Appendix B

the south, some peasant castes in that region have fewer people ‘above’ them than a similar caste in the north. This variable has a small positive effect on petitioning – groups at the bottom of the local status ordering are more likely to petition than those above them – but does not affect the value the literacy variables. Closely related to this concern is the possibility that ideological efforts by outsiders could influence the willingness of groups to embrace or reject the caste system. One group in early twentieth-century India, the Arya Samaj, was especially notable in such efforts. Model Two of Table B.3 includes variables for the percentage of Arya Samajists in each province in a given year. The growth of the Samaj does seem to have a marginal negative effect on petitioning, though it does not change the reported results. Another over-time change that might affect the level of petitioning is the growth of Christianity. Variables for the percentage of Christians in the province, along with its interaction with caste status, are included in Model Three of Table B.3. While Christianity has a negative effect on hierarchical petitioning in some models, this effect does not substantially reduce the estimated effect of the literacy variables. The contrast between the north and south of the subcontinent is at the centre of much social scientific and historical work on South Asia, with the south often being portrayed as less ‘caste ridden’ than the north, though in general it seems that traditional strictures on ritual pollution were stronger in the south. Despite the presence of province fixed effects in all reported models, it is possible that these results are driven by one region. To test this possibility, Models Four and Five of Table B.3 report results with certain regions of the country excluded: Model Four reports the results excluding the non-Hindi-speaking provinces, and Model Five for the non-Dravidian-speaking provinces. Despite the smaller sample, the non-linear effect of literacy remains robust and statistically significant predictor of caste petitioning. Chapter 4 suggested that, given the high rates of rejection, the chance of success was not the primary goal of the petitioning process. What if this was not the case, and petitions were motivated by the desire for official recognition? In this situation, the dynamics of petitioning could be influenced by the dynamics of the census itself, as changes in the official attitude lead to spurts of petitioning. To test this hypothesis, Model One of Table B.4 includes a variable for the rate of petition granting for that census year. The inclusion of this variable does not affect the reported results.

Table B.3 Province level robustness checks: logistic mixed effects regression with petition as dependent variable Variables

(1) Petitioning

(2) Petitioning

(3) Petitioning

(4) No Hindi Belt

(5) No South

Population as Proportion

11.31** (4.966) 0.000203 (0.000442) 6.878** (2.726) −9.260** (4.164) 6.092* (3.372)

11.64** (4.908) 0.000167 (0.000438) 6.971** (2.740) −9.774** (4.196)

12.08** (5.362) 0.000163 (0.000453) 6.380** (2.741) −8.349** (4.203)

11.51* (5.896) 0.000560 (0.000959) 8.218*** (3.173) −10.82** (4.777)

12.87 (7.935) 0.000208 (0.000549) 7.704** (3.280) −11.01** (5.010)

Population 000s Male Literacy Rate Male Literacy Rate Sq. Local Social Status Prov. Prop. Arya Samaj Prov. Prop. Christian Prov. Prop. Christian x Caste Status

−133.3* (78.12) 43.29** (16.82) −8.205** (3.484) contd.

Variables

(1) Petitioning

(2) Petitioning

(3) Petitioning

(4) No Hindi Belt

(5) No South

Constant

2.574 (3.954) 887 1 YES YES YES YES YES YES

1.995 (4.006) 887 1 YES YES YES YES YES YES

3.510 (4.017) 887 1 YES YES YES YES YES YES

−4.801 (6.149) 497 1 YES YES YES YES YES YES

6.632 (4.927) 680 1 YES YES YES YES YES YES

Observations Number of groups Caste Status FE Province FE Year FE Province-Year Controls province-year RE Jati RE

Standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 These are mixed effects logistic models, with fixed effects for caste status, year, and province and random effects for jati and province-year. Only the fixed effect constants are reported. The Province-Year Controls, proportion urban, proportion in public employment, provincial population, provincial employment in agriculture, and the provincial literacy rate, are not reported for reasons of space.

Table B.4 Caste level robustness checks: logistic mixed effects regression with petition as dependent variable Variables

(1) Petitioning

(2) Petitioning

(3) No Splits

(4) No Combinations

(5) Petitioning

Population as Proportion

11.26** (4.933) 0.000228 (0.000439) 7.168*** (2.765) −10.07** (4.240) 1.902** (0.846)

11.27** (4.918) 0.000228 (0.000443) 7.117*** (2.737) −9.709** (4.141)

11.29** (5.282) 0.000334 (0.000457) 8.961*** (3.057) −12.93*** (4.757)

9.242 (6.030) 0.000276 (0.000588) 7.804** (3.246) −12.28** (5.175)

11.18** (4.930) 0.000181 (0.000441) 6.551** (2.731) −9.003** (4.146)

Population 000s Male Literacy Rate Male Literacy Rate Sq. Petition Grant Rate Hierarchical Census Classification Hierarchical Census Classification x Caste

−1.169 (1.094) 0.201 (0.181) contd.

Variables

(1) Petitioning

(2) Petitioning

(3) No Splits

(4) No Combinations

(5) Petitioning

Change in Population 000s Constant

Observations Caste Status FE Province FE Year FE Province-Year Controls province-year RE Jati RE

5.376 (4.311)

3.927 (4.048)

6.668 (4.205)

5.455 (4.404)

0.0973 (0.0836) 4.279 (3.962)

858 YES YES YES YES YES YES

887 YES YES YES YES YES YES

820 YES YES YES YES YES YES

832 YES YES YES YES YES YES

861 YES YES YES YES YES YES

Standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 These are mixed effects logistic models, with fixed effects for caste status, year, and province and random effects for jati and province-year. Only the fixed effect constants are reported. The Province-Year Controls, proportion urban, proportion in public employment, provincial population, provincial employment in agriculture, and the provincial literacy rate, are not reported for reasons of space.

Appendix B

203

The secondary literature on the census of India has often promoted the view that caste petitions are the products of the ideas of the census takers and of official support for a reified and Sanskritized caste system (Cohn 1987; Dirks 2002). There is reason to doubt this, since the census was one of the most nationally unified of Raj institutions, with strong national rules even in the princely states. However, there was one important area that varied both from province to province and from year to year: the arrangement of castes within the census table. The most common alternatives were alphabetical classification and occupational classifications, though in 1901 most provinces followed national policy and attempted to rank castes based on their ritual status. For the purposes of analysis, I simplified these differences into a binary variable, with a ‘1’ indicating occupational- and ‘precedence’-based classifications. Model Two of Table B.4 includes this hierarchical classification variable and its interaction with caste status. Neither of these variables has a statistically significant effect on caste mobilization, and the inclusion of these variables does not affect the values of the coefficients for the literacy variables. The main models assume that the petition made a claim on behalf of the whole group (even if it was only submitted by a handful of individuals) and that only one petition was submitted on the group’s behalf (or, if multiple petitions were submitted, that they were substantially similar in content). In fact, neither of these conditions is always the case. In some cases, a petition was submitted on behalf of a sub-caste of a jati claiming to be either a separate caste or the members of a different jati entirely. In Bengal, for instance, a sub-caste of the Kayasth caste claimed to be Kshatriyas while fully admitting the Shudra status of the remainder of Kayasths. More rarely, two groups made differing claims on behalf of the whole caste. Model Three of Table B.4 drops these split cases from the analysis, leaving the results unchanged. Chapter 4 discussed the problems presented for the coding for the caste merger movement among Dalits in south India. Not only does it mean that some groups are exiting the dataset (into the artificial ‘merged castes’) but the broad anti-Brahmin or pre-Hindu identity advanced by the campaigners made it difficult to know whether castes were actually active in a political movement or were merely ‘annexed’ by activists from other castes. In addition, the broad influence of the non-Brahmin movement in interwar south India could be seen to be an ideological confound affecting all groups. To test this hypothesis, Model Four of Table B.4 excludes the province-years

204

Appendix B

affected by the non-Brahmin movement (Madras, Mysore, and Hyderabad in 1931 and Mysore in 1921). This exclusion does not substantively affect the reported results. While the units of observation in this study, jatis, are constant over time, their compositions are not, as individuals could easily shift jati for census purposes. Despite the efforts of census officials, there are frequent references to such shifts in the census records, and such shifts are also detectable in quantitative analysis (for example, Cassan 2015). Such ground-level status shifts may serve as a substitute for large-scale collective action like petitioning. Any jati that sees a major gain or loss might also see changes in its other covariants due to the influx of outsiders. To test this hypothesis, Model Five of Table B.4 includes a variable for the absolute percentage of gain or loss that a caste had in male population over the previous census year. However, this variable appears to have no significant effect on petitioning or on the size of the literacy coefficients.

Modelling ranked petitioning The leaders of Indian castes faced primary choices in interacting with the census authorities: whether to petition for a name change at all and whether or not to choose a new name that emphasized social ranking. This decision structure is modelled as a sequential logit model with the choice of ranking language nested within petitioning cases. This means that Pr , the probability of a province-caste-year recording a ranked caste petition, can be estimated as the product of two probabilities: Pr = Pp ∗ Pr|p

(B.3)

where Pp is the probability of submitting a petition, and Pr|p is the probability of submitting a ranked petition given that a petition is already submitted. While Chapter 5 focused on the causes of identity mobilization in general, Pp , this appendix focuses instead on Pr|p , the probability of ranked mobilization among mobilizing groups. This can be estimated by using a nested logit model, with the choice of ranked or unranked petitioning being nested within the choice to petition at all: Pr|p = !

eYr

j∈n

eYr

(B.4)

Appendix B

205

where Yr are the independent and control variables of interest.3 The findings are substantively identical to those obtained by using a Heckman selection model, the results of which are reported in Table C.9. All reported models control for the caste status of the caste at the time of petition. The inclusion of these variables accounts for the possibility that the results might be driven by lower incentives to social repositioning among higher castes, since groups are compared to those with similar ex ante social status. All the models also include fixed effects for province and year, to account for the possibility of geographical or time-varying confounds. In Table 5.1, we see the changes that occurred in petitioning over time. Overall, the level of both ranked and unranked petitioning rose over time. This first order result is largely the result of changes in Pp . Since the overall incidence of petitioning is related to the socioeconomic status of groups, the gradual expansion of wealth and literacy in early twentieth-century India led to steep increases in petitioning, even among poorer and more rural groups. To study the rhetoric of petitioning independent of the changing incidence of petitioning, it is necessary to use the sequential model.4

Hypothesis 3 The effect of caste landedness on ranked mobilization is tested in Table B.5, which includes the two independent variables of interest, group control of land and employment in the army and police. Land control has a strong positive relationship with ranked petitioning, supporting the idea that ranking is more attractive to groups enmeshed in rural patron–client relationships. Note that neither variable appears to affect the overall incidence of petitions, reported in the first stage of the hierarchical model in Table C.6. As expected, employment in the army and police has a negative effect on ranked petitioning. Though the effect of police employment is statistically significant in most models (and close to significance in the others), it is much smaller than the effect of landedness, with a similar 10 per cent increase decreasing the probability of ranked petitioning by only 4 percentage 3

4

Dropping cases of no petitioning and running a simple binary model of the incidence of ranked produces substantially identical results. Readers interested in predictors of Pp should refer to Table C.6, which reports the results of the first equation of the sequential model.

Table B.5 Traditional institutions and social ranking: sequential logistic regression Variables

(1) Ranking

(2) Ranking

(3) Ranking

Caste Land Cultivation Rate Caste Army/Police Rate

3.517*** (1.186) −68.43 (36.55)

4.359*** (1.557) −76.73** (33.78)

4.796*** (1.546) −82.99** (34.44)

Population 00000s Male Literacy Rate Male Literacy Rate Sq.

(5) Ranking

−15.04 (14.24) 0.0244 (0.0699) −10.51 (7.926) 32.34 (17.64)

−14.52 (15.26) 0.0351 (0.0724) −12.17 (8.458) 33.70 (19.78)

11.99 (30.05) 0.236* (0.126) 2.219 (7.935) 10.63 (19.63)

(6) Ranking 4.698** (1.889) −52.18 (40.37) −11.59 (10.88) −219.2*** (77.56)

−2.782 (3.378) −91.44** (43.44)

Prop. Local Boards Elected Prop. Local Boards Elected x Population Prov. Congress Delegates Prov. Congress Delegates x Population Population as Proportion

(4) Ranking

188.3** (87.87) −3,136** (1,243) −49.90* (28.72) 0.352 (0.196) −0.611 (9.721) 6.512 (23.79)

40.13 (36.84) 0.621** (0.263) −0.855 (11.93) 19.43 (31.48) contd.

Variables

(1) Ranking

(2) Ranking

(3) Ranking

(4) Ranking

(5) Ranking

(6) Ranking

−0.642 (0.873)

−4.383 (20.74)

−4.792 (2.770) −0.743 (1.186) 0.0453 (21.22)

3.718 (4.375)

6.915 (8.091)

16.93 (28.08)

412 NO YES YES YES

390 YES YES YES YES

390 YES YES YES YES

587 YES YES YES YES

401 YES YES YES YES

300 YES YES YES YES

Caste Ag. Labour Rate Caste Traditional Occupation Rate Constant

Observations Province-Year Controls Province FE Year FE Caste Status FE

Robust standard errors in parentheses *** p < 0.01, ** p < 0.05 These are pooled sequential regression models, with fixed effects for caste status, year, and province, with the standard errors clustered by caste. This table only reports the results of the second equation, where ranked petitioning is the dependent variable. The corresponding results of the first equation, with petitioning as the dependent variable, are reported in Table C.6. The Province-Year Controls, proportion urban, proportion in public employment, provincial population, and provincial literacy rate, are not reported for reasons of space.

208

Appendix B

points. These results seem to indicate that, consistent with Hypothesis 1, investment in traditional social institutions leads to an emphasis on social ranking, while exposure to the more modern institutions of the colonial state reduces this emphasis. These effects are apparent in Model Two, which includes control for group population and literacy, along with measures of provincial social change and fixed effects for caste status. Model Three adds a set of measures of occupational structure of each caste, including the percentage of caste members working as agricultural labourers and the percentage practising the group’s traditional occupation. Neither of these variables changes the effect of land control and paramilitary employment. Another interesting aspect of Model Three is that the percentage of group members employed in the group’s traditional occupation is unrelated to their choice of ranked ideological forms. This goes against a well-known argument in the literature that traditional caste relationships are the product of economic exchange (Dubois 1806; Frietas 2010). It seems that either economic ties or economic intimidation can explain identity ranking among lower castes. Ranked petitioning is thus most common among land controlling groups with weak links to the state, while unranked petitioning is most common among land-poor groups with strong links to the state. This fact provides some support for the hypothesis that this variation stems from the difference between patrimonial and formal politics. These effects can be seen graphically in Figure 5.1, which presents kernel density plots for ranked and unranked petitions relative to land control. Figure 5.1 shows that, relative to unranked petitions, a substantially higher proportion of ranked petitions are made by land-rich groups, and a correspondingly smaller proportion by land-poor groups. Since only a small proportion of groups have a majority of land cultivators, the majority of ranked petitions, like unranked ones, are made by groups with less than 50 per cent of their workers controlling land. Figure B.1 shows the kernel density distribution of types of petitions by the logged rate of police and army employment. The large majority of both ranked and unranked petitions were submitted by castes with low levels of public force employment, reflecting the discriminatory recruitment policies of the colonial government: very few castes were considered suitable recruits. However, unranked petitioning is by far the most common mode among those castes who were outliers in their level of public force employment.

Appendix B

209

Figure B.1 Kernel density plot of ranked and unranked petitioning groups by public force employment .5

Army/Police Employment by Type of Petition

0

.1

.2

Density .3

.4

Unranked Ranked

−10

−8

−6 −4 Logged Army/Police Employment

−2

kernel = epanechnikov, bandwidth = 0.3483

Source: Author’s calculations from Census of India, various years.

One notable aspect of the results in Table B.5 is that the results hold even after we account for castes’ levels of literacy, the variable that most closely captures the type of educational opportunities and social changes usually described as ‘modernization’. It thus appears unlikely that between-group differences in ‘modernization’ are driving the empirical results. Even more interestingly, literacy has no substantial effect on ranked petitioning nor do province-level variables associated with modernization, such as the size of the provincial bureaucracy and the province-level spread of elections. The key to this puzzle is found in Table C.6, which shows the predictors of petitioning itself, independent of petition type. Literacy has a strong, robust quadratic effect on petitioning, and none of the political or patrimonial variables have any effect. These results paint a picture very similar to the findings in Chapter 5. While modernization is strongly related to ethnic activism in the aggregate, the form that this activism takes depends on the political environment. We have good reason to expect that differences in a priori status should influence the mobilization strategies of groups: As groups with a higher ritual status have a more advantageous position within the existing hierarchy, they are more likely to desire its preservation. While these patterns were accounted for by the inclusion of status fixed effects

210

Appendix B Table B.6 Number of petitions by caste status

Caste Status

No Petition

Ranked Petition

Unranked Petition

Total

Hierarchical Per cent

Untouchable/Dalit Lower Unclean Shudra Upper Unclean Shudra Intermediate/Dominant Upper/‘Twice Born’ Brahmin

200 322 143 101 62 39

12 23 21 4 0 1

20 90 47 26 17 0

245 430 204 131 79 40

62.5 79.6 69.1 86 100 0

Total

868

61

201

1,130

Source: Author’s calculations from Census of India, various years. Note: The numbers in the table represent caste-province-years. For details on the coding of the caste categories, see Appendix A.

in Table B.5, these patterns are of some interest. Table B.6 shows the rates of petitioning by caste categories, the details of whose coding are discussed in Chapter 4. Aside from the one Brahmin petition, ranked petitioning is universal among high caste groups, and less common (though still predominant) among lower caste groups. The exception is in the ‘Lower Unclean Shudra’ category, which has higher levels of hierarchical petitioning than the untouchable or ‘Upper Unclean Shudra’ categories. This is probably a result of the small size of many of the groups in this category, which reduced the incentive to mobilize on an unranked basis.

Hypothesis 4 Hypothesis 4, by contrast, posits that unranked petitioning should occur for only a particular kind of caste (large) under a particular set of circumstances (institutions that value political participation). Increasing political participation should not affect mobilization among small castes (since they will tend to be politically weak even when mobilized) and increasing levels of population should be unrelated to unranked mobilization in areas with low participation (since they cannot be brought to bear politically). However, the interaction of the two variables (corresponding to the effect of a large group population in areas with high participation) should be significant and positive. Model Four of Table B.5 adds the crucial variable of interest for Hypothesis 2, the interaction between group population and local elections.

Appendix B

211

While local elections have a little direct effect on ranked petitioning, the interaction of elections with group size is statistically significant and negative in Model Four. This indicates that when large groups in areas with democratic elections enter politics, their involvement tends not to emphasize social ranking. Model Five of Table B.5 reports the results of the same basic model, with the per capita number of Indian National Congress attendees substituted for the proportion of local officials. Despite a dramatic drop in sample size (Congress attendance data is not available for the princely states, or for the year 1931), the interaction of population with participation is positive and statistically significant. This supports the hypothesis that group size and regional politicization affect ethnicity only in association with each other. In Model Six, this effect is shown to be robust to the inclusion of the occupational variables from Model One, though the inclusion of the election variables causes the effect of police employment to become statistically insignificant though still negative. Table C.5 shows that the results are robust to a variety of alternative specifications. Model One uses an OLS instead of a logit function, and Model Three uses an OLS model using a very conservative set of fixed effects (Jati-Year and province-year). Both models produce results very similar to those in Table B.5.5

Alternative models and samples Are these results the product of the structure of the hierarchical model or a product of the inclusion of a large number of covariates? Table B.7 presents the results of a series of simple logistic regression models that include only petitioning groups and no fixed effects. The results are very similar to those in Table B.5. Castes with relatively large numbers of soldiers and policemen and relatively small numbers of landowners are less likely to seek to adopt an upper caste name than other groups. Relatively large groups in areas with a high proportion of elected legislators are less likely than other groups to seek to adopt an upper caste name. Tables C.7 and C.8 test the robustness of the results to changing the sample. Model One, intended to see if the results are robust to the exclusion of southern India, does not include Mysore, Travancore, Madras, 5

Note that this last set of fixed effects could not be estimated in the logit model, due to perfect prediction.

212

Appendix B Table B.7 Institutions and social ranking: simple models

Variables

(1) Ranking

Population as Proportion

(2) Ranking

(3) Ranking

(4) Ranking

4.233 (20.54)

−24.16*** (8.528) 0.0204 (0.0424) −8.249* (4.894) 21.31* (11.13) 1.887** (0.786) −101.6* (56.48)

0.444 (21.26) 0.134 (0.0822) −9.828 (6.177) 32.03** (15.58)

Population 00000s Male Literacy Rate Male Literacy Rate Sq. Caste Land Cultivation Rate Caste Army/Police Rate

1.241 (0.800) −60.27** (26.07)

Prop. Local Boards Elected

0.363 (1.113) −27.80 (27.13)

Prop. Local Boards Elected x Population Constant

Observations

0.366 (1.213) −58.68* (32.61)

0.786** (0.381)

1.297* (0.770)

1.691*** (0.622)

1.759* (0.944)

149

188

143

164

Robust standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 These are logistic regression models with the standard errors clustered by caste.

or Hyderabad. Model Two does not include 1931, the year when caste agitation was at its peak. Model Three does not include castes where the literacy data was calculated for less than 80 per cent of the caste. Model Four does not include castes where petitions were filed on behalf of a sub-caste. Model Five excludes both ‘clean’ and untouchable castes, thus including only groups from the middle ranking ‘OBC’ castes. Table C.7 shows that, despite the drop in sample size, the landownership results are in general robust to the alterations in the sample, despite the smaller sample size and dropping of some observations due to perfect prediction by the fixed effects. Group-level landedness and public force employment are correlated with petitioning for upper caste names. The participation results appear to be less robust: while the interacted effect of group size and provincial elections on ranked petitioning is negative, it

Appendix B

213

is insignificant in many models, and close to zero when the year 1931 is excluded. This reflects the high level of collinearity between the elections variable and the year variables. It is thus possible that some portion of the effect of election is due to unobserved factors. The conclusions we can draw from the results in Table B.2 are limited by the size of the dataset. However, the state and caste status fixed effects and literacy controls account for some of the more obvious alternative hypotheses. Tables B.8 and B.9 explore a few more alternative hypotheses, including changing census policies, Christian and Hindu missionary work, and individual movement between groups. Table B.8 tests the robustness of the results of Hypothesis 1, while Table B.9 tests the robustness of the results of Hypothesis 2. Model One of both tables shows that the results are not a product of differences in the probability of rejection or acceptance of petitions. The rate of petition granting has no effect on the incidence of ranked petitioning, and has no effect on the main results. Model Two of both tables tests the effect of the classifications used in the census table. While in some province-years the census authorities classified castes by ritual ‘precedence’, in other years castes were organized alphabetically. The use of ranked census norms might well encourage the use of ranked language in caste petitions. However, while ranked census classifications do seem to have a positive effect on ranked petitioning (particularly for higher castes), this effect does not reduce the estimated effect of caste employment, or of the interaction of caste population and elections. Model Three of both tables tests the hypothesis that inter-group movement could serve as a substitute for collective action. Instead of petitioning collectively to raise their caste status, individuals could merely reclassify themselves as members of higher castes. Such individual movement among castes is well attested in colonial India, and is readily apparent in census figures (see Cassan 2015). Model Three includes a control for the percentage change in caste membership since the last census. While caste population changes do appear to be positively correlated with petitioning, their inclusion as a control does not affect the effects of the key independent variables. Are the differences in petitioning driven by differences in politicization and the spread of the nationalist movement? Model Four of Table B.8 and Model Five of Table B.9 test this idea by controlling for the per capita attendance at Indian National Congress sessions. This variable has no effect on the effects of landedness and the size-election interaction, but it does

Table B.8 Robustness checks: sequential logistic regression Variables

(1) Ranking

(2) Ranking

(3) Ranking

(4) Ranking

(5) Ranking

Population as Proportion

−28.99** (14.61) 0.0680 (0.0701) −13.55 (8.905) 41.61** (20.43) 5.058*** (1.409) −75.18** (37.77) −7.612 (5.697)

−26.07 (14.10) 0.0555 (0.0679) −11.61 (9.061) 35.47 (20.19) 4.933*** (1.386) −72.95 (41.51)

−22.31 (13.37) 0.0586 (0.0658) −10.82 (8.817) 37.11 (20.40) 5.436*** (1.651) −80.13** (37.77)

−60.32*** (19.69) 0.161 (0.101) 5.917 (12.17) −27.70 (42.92) 5.433*** (1.843) 141.3 (107.1)

−22.69 (12.35) 0.00485 (0.0649) −7.646 (7.858) 26.08 (17.68) 3.197*** (1.199) −75.98** (36.08)

Population 00000s Male Literacy Rate Male Literacy Rate Sq. Caste Land Cultivation Rate Caste Army/Police Rate Petition Grant Rate Hierarchical Census Classification Hierarchical Census Classification x Caste Status Population Change

−0.995 (3.887) 0.398 (0.654) 1.478 (1.734) contd.

Variables

(1) Ranking

(2) Ranking

(3) Ranking

(4) Ranking −143.2** (55.65)

Prov. Congress Del.

−4.023 (16.90)

23.89 (35.82)

9.211 (28.61)

−2.743 (2.440)

−22.18 (359.6) −15.78 (69.68) −2.593 (24.70)

365 YES YES YES YES

391 YES YES YES YES

389 YES YES YES YES

208 YES YES YES YES

391 YES YES YES YES

Prov. Prop. Arya Samaj Prov. Prop. Arya Samaj x Caste Status Constant

Observations Province-Year Controls Province FE Year FE Caste Status FE

(5) Ranking

Robust standard errors in parentheses *** p < 0.01, ** p < 0.05 These are pooled sequential regression models, with fixed effects for caste status, year, and province. This table only reports the results of the second equation, where ranked petitioning is the dependent variable. The corresponding results of the first equation, with petitioning as the dependent variable, are not reported. The Province-Year Controls, proportion urban, proportion in public employment, provincial population, and provincial literacy rate, are not reported for reasons of space.

Table B.9 Robustness checks: sequential logistic regression Variables

(1) Ranking

(2) Ranking

(3) Ranking

(4) Ranking

(5) Ranking

Population as Proportion

11.15 (30.25) 0.239 (0.126) 1.359 (8.454) 12.56 (21.41) −1.995 (3.377) −91.52** (43.34) −2.774 (3.338)

24.80 (30.70) 0.283** (0.135) 3.421 (7.265) 9.582 (16.01) −1.675 (3.336) −123.9** (48.47)

20.18 (32.05) 0.235 (0.137) −0.359 (8.303) 19.24 (20.82) −3.145 (3.527) −101.7** (47.07)

7.765 (25.96) 0.250** (0.122) −0.299 (8.800) 15.16 (22.81) −2.944 (3.438) −94.08** (42.31)

11.10 (28.57) 0.224** (0.103) −0.742 (9.615) 9.417 (23.14) −0.468 (2.148) −100.1** (45.28)

Population 00000s Male Literacy Rate Male Literacy Rate Sq. Prop. Local Boards Elected Prop. Local Boards Elected*Population Petition Grant Rate Hierarchical Census Classification Hierarchical Census Classification x Caste Status Population Change

−5.188 (3.289) 1.188** (0.592) −1.412 (1.588) contd.

Variables

(1) Ranking

(2) Ranking

(3) Ranking

Prov. Prop. Arya Samaj

Observations Province-Year Controls Province FE Year FE Caste Status FE

(5) Ranking

506.5 (421.0) −77.85 (83.87)

Prov. Prop. Arya Samaj x Caste Status Prov. Congress Delegates Constant

(4) Ranking

0.871 (4.736)

2.705 (4.696)

1.998 (4.678)

4.790 (4.683)

38.03 (28.84) 3.715 (2.022)

561 YES YES YES YES

587 YES YES YES YES

556 YES YES YES YES

587 YES NO NO YES

401 YES NO NO YES

Robust standard errors in parentheses *** p < 0.01, ** p < 0.05, These are pooled sequential regression models, with fixed effects for caste status, year, and province, with the standard errors clustered by caste. This table only reports the results of the second equation, where ranked petitioning is the dependent variable. The corresponding results of the first equation, with petitioning as the dependent variable, are not reported. The Province-Year Controls, proportion urban, proportion in public employment, provincial population, and provincial literacy rate, are not reported due to space reasons.

218

Appendix B

eliminate the effect of state employment, primarily by halving the size of the sample. As with the simple petitioning model, it is also possible that the presence of the Arya Samaj might encourage the adoption of norms of social ranking. The Arya Samaj sought to recreate an idealized form of the original Hindu caste hierarchy. Model Five of Table B.8 and Model Four of Table B.9 include the provincial levels of these variables and their interactions with caste status. While the Arya Samaj does appear to have an impact on petitioning behaviour, it has no effect on petition language. Also, the inclusion of these variables does not alter the effects of the key independent variables. One final possibility is that the adoption of ranking is especially common among the so-called martial races. These were ethnic groups considered by the British especially loyal and suitable for military service, and were given preferential access to the armed forces and other government benefits. Given the sophisticated discourse of ethnic superiority that emerged around these groups, it is reasonable to think that their political identities might develop differently than other groups, a major concern given that military and police recruitment is one of my measures of state access. However, it is certain that these groups are not driving the results, because no martial groups are included in the dataset. The majority of the important martial groups were either from outside India (Gurkhas) or members of religious minorities (Sikhs, Baluchis, Punjabi Muslims, Hindustani Muslims) and are thus outside the sample frame. The only Indian Hindu groups generally considered martial were the Dogras of Kashmir and Punjab, the Rajputs of the Himalayan regions and Rajasthan, and the Marathas. Since none of these groups ever submitted a petition, they have no effect on the results shown in Table B.2.6

6

Removing these groups also has no effect on hypothesis testing about the causes of petitioning.

Appendix C

Additional Tables and Figures

Figure C.1 Kernel density of petitioning and non-petitioning groups by literacy rate

Source: Author’s calculations from Census of India, various years. Note: Distribution of pooled literacy rates, 1901–1931, by petitioning status. Table C.1 Census petitions in India, 1901–1931 Caste Name

Year

Province

New Name (if any)

Brahman Lohar (Luhar) Baria (Koli Baria) Baria (Koli Baria) Talabda (Koli Tablada)

1911 1911 1911 1921 1921

Baroda Baroda Baroda Baroda Baroda

Acharyas Panchal Lohar Thakore Rajput Rajput contd.

220

Appendix C

Caste Name

Year

Province

New Name (if any)

Dhed Lohar (Luhar) Anja Coudhari (Kanbi Anjana) Baria (Koli Baria) Hajjam (Valand) Lewa Patidar (Kanbi Lewa) Bhil Dhed Thakarda (Koli Thakarda) Kadwa Patidar (Kanbi Kadwa) Sunri Jogi Jaliya-Kaibartta Namsudra Kayastha Kalu Rajbagshi (Koch) Barui Sadgope Tiyar Barui

1921 1931 1931 1931 1931 1931 1931 1931 1931 1931 1901 1901 1901 1901 1901 1901 1901 1901 1901 1901 1911

Baroda Baroda Baroda Baroda Baroda Baroda Baroda Baroda Baroda Baroda Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal

Sadgope Goala Rajbagshi (Koch) Namsudra Pod

1911 1911 1911 1911 1911

Bengal Bengal Bengal Bengal Bengal

Teli Sunri Mahishya (Chasi Kaibartta) Napit (Hanjam)

1911 1911 1911 1911

Bengal Bengal Bengal Bengal

Dhobi Jogi Kumhar Malo Baishnab (Bairagi) Kayastha Mayra

1911 1911 1911 1911 1911 1921 1921

Bengal Bengal Bengal Bengal Bengal Bengal Bengal

Mayavasi Rajput Panchal Brahman Rajput Rajput Valand, Nai Brahman Patidar Arya, Bhil Kshatriya Vankar Thakarda Solaki Patidar Vaisya (Shahas only) Jogi Jaliya Namsudra Subcaste Kshatriya Subcaste Teli Bhanga Kshatriya Kayasth Vaishya Rajbanshi Vaisya Barujibi or Barujibi Vaisya Sadgop Vaisya Balab Gop Koch-Kshattriya Namsudra-Brahman Bratya Kshatriya or Pundra Kshatriya Vaisya Vaisya (Shahas only) Mahishya Kshattriay or Paramanik or Kayasth Satchasi Yogi Rudra Pal Bratya Kshatriya Brahma Baishnab Kshatriya Kayasta Kurmi contd.

Appendix C

221

Caste Name

Year

Province

Tiyar Rajbagshi (Koch) Sunri Barui Kamar Napit (Hanjam) Kalu Goala Mahishya (Chasi Kaibartta) Jaliya-Kaibartta Pod Malo Tanti or Tatwa Teli Shaha Jogi Tanti Barui Kayastha Pod

1921 1921 1921 1921 1921 1921 1921 1921 1921 1921 1921 1921 1921 1921 1931 1931 1931 1931 1931 1931

Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal

Sunri Bagdi Tiyar Namsudra Rajbagshi Dhobi Napit Adi-Kaibarta (Jaliya-Kaibartta) Mahishya

1931 1931 1931 1931 1931 1931 1931 1931 1931

Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal Bengal

Babhan (Bhuminyar) Sunri or Shaha Kurmi Kurmi Rajbanshi (Koch) Babhan (Bhuminyar) Sonar Gaur Khandait

1901 1901 1901 1911 1911 1911 1911 1911 1921

Bihar Bihar Bihar Bihar Bihar Bihar Bihar Bihar Bihar

New Name (if any) Mahisya, Rajbanshi Kshatriya Shaha Kayastha, Lata Baida Karmar Kshatriya Baidya Rishi Teli Vaisya Gop, Sadgop Mahisya Mahisya Poudra Kshatriya Bratya Kshatriya Vaisya Basak Vaisya Baisya Shaha Yogi Tantubaya Barujibi Kshatriya Paundra, Paundra Kshatriya Kshatriya Byaga Kshatriya Rajbanshi Namabrahman Kshatriya Vaisya Nai Brahman Mahisya, Rajbanshi Mahisya Kshatriya, Devdas Brahman Vaisya (Shahas Only) Kshatriya Subcaste Kurmi Kshatriya Bhanga Kshatriya Brahman Kanaujia Chhattri Vaisya Gop Mahisya contd.

222

Appendix C

Caste Name

Year

Province

Kurmi Koiri Kahar

1921 1921 1921

Bihar Bihar Bihar

Dosadh Gaur Babhan (Bhuminyar) Gaula (Ahir, Gaura) Hajjan (Nai) Koiri

1921 1921 1921 1921 1931 1931

Bihar Bihar Bihar Bihar Bihar Bihar

Babhan (Bhuminyar) Kurmi Barhi Dosadh Gaula (Ahir)

1931 1931 1931 1931 1931

Bihar Bihar Bihar Bihar Bihar

Kahar Berad or Bedar Darji, Shimpi Soni, Sonar Mali Berad or Bedar Soni, Sonar Koli Bania

1931 1911 1911 1911 1921 1921 1921 1921 1911

Bania Lohar Koshti Sonar Teli Kawar Ganda Teli Kalar Dhimar Lodhi Chamar

1911 1911 1911 1911 1911 1911 1911 1931 1931 1931 1931 1931

Bihar Bombay Bombay Bombay Bombay Bombay Bombay Bombay Central Indian Agency Central Provinces Central Provinces Central Provinces Central Provinces Central Provinces Central Provinces Central Provinces Central Provinces Central Provinces Central Provinces Central Provinces Central Provinces

New Name (if any) Kurmi Kshatriya Dangi Chandravanshiya Kshatriya Rajput Vaisya Gop (Orissa Only) Brahman Ahir Kshatriya Nai Brahman Kushwaha Kshatriya, Dangi Brahman Kurmi Kshatriya Vishvkarma Brah Gahlot Rajput Jadubansi Kshatriya, Yadava Chandrabansi Kshatriya Put Kunbi Bharsar Kshatriya Daivadnya Brahman Sani Kshatriya Walmiki Daivadya Brahman Rajput Brahman Bhargava Brahman Brahman Devang Vishwas Brahman Rathor Rajput (Subcaste) Tanwar Rajput (Subcaste) Binjhal Rathor Haihaya Kshatriya Kashyap Rajput Lodhi Rajput Stanami contd.

Appendix C

223

Caste Name

Year

Province

New Name (if any)

Gond

1931

Central Provinces

Kurmi Mahar Hajjam) Mangala Chambhar (Chamar) Panchal Kapu (Kunbi) Mochi (Madiga) Sutar Mang Mala Kurma Sunar Dher (Mahar) Kammala (Viswasbrahman) Vellala Shanan Arya Vaisya (Komatia Balija Shanan Komati Kammala (Viswasbrahman) Pariyan Adi-Dravida Kamsala Idiga

1931 1931 1931 1931 1931 1931 1931 1931 1931 1931 1931 1931 1931 1901 1901 1901 1901 1901 1911 1911 1911 1921 1921 1921 1921

Central Provinces Central Provinces Hyderabad Hyderabad Hyderabad Hyderabad Hyderabad Hyderabad Hyderabad Hyderabad Hyderabad Hyderabad Hyderabad Madras Madras Madras Madras Madras Madras Madras Madras Madras Madras Madras Madras

Raj Gond, Kshatriya Surujwansi Kurmi Kshatriya Kashyap Rajput Nayee-Brahman Adi-Hindu Viswas Brahman Reddi Adi-Hindu Viswas Brahman Adi-Hindu Adi-Hindu Kurmi Kshatriya Viswas Brahman Adi-Hindu Brahman Vaisya Kshatriya Vaisya Kshatriya Kshatriya Vaisya Brahman Adi Dravid

Arya Vaisya (Komati Shanan (Nadar) Visvabraham (Kamala) Visvabraham (Kamala) Cheruman or Pulayan Madiga Adi-Andhra Yadava Adi-Dravida Arya Vaisya (Komati Pariyan

1921 1921 1921 1931 1931 1931 1931 1931 1931 1931 1931

Madras Madras Madras Madras Madras Madras Madras Madras Madras Madras Madras

Vivasbrahman Arya Hihida, Setti Balija Vaisya Kshatriya Kamsala Visvabraham

Yadava

contd.

224

Appendix C

Caste Name

Year

Province

Ambattan Pallan Malla Holaya Lingayat (Includes Priests And Trades) Madiga Viwaskarma (Panchala) Neyige (Neyigi) Kuruba Adikarnataka (Holeya+Madiga) Lingayat Vakkaliga Beda Viwaskarma Yadava Gangakula (Besta) Nai (Hindu) Lohar Hindu Bania Total Kanet (Hindu) Tarkhan (Hindu Ahir (Hindu) Chuhra (Hindu) Kamboh

1931 1931 1931 1921 1921

Madras Madras Madras Mysore Mysore

1921 1921 1921 1931 1931 1931 1931 1931 1931 1931 1931 1911 1911 1911 1911 1911 1911 1911 1911

Mysore Mysore Mysore Mysore Mysore Mysore Mysore Mysore Mysore Mysore Mysore Punjab Punjab Punjab Punjab Punjab Punjab Punjab Punjab

Sunar Jhinwar (Hindu) Tarkhan (Hindu Jhinwar Kamboh Kanet (Hindu) Lohar Hindu Nai (Hindu) Chamar (Hindu) Tarkhan (Hindu Nai (Hindu) Lohar Hindu

1911 1911 1921 1921 1921 1921 1921 1921 1931 1931 1931 1931

Punjab Punjab Punjab Punjab Punjab Punjab Punjab Punjab Punjab Punjab Punjab Punjab

New Name (if any) Nai Brahman

Adi Dravid Virasiava Brahman Adi Dravid Viwas Brahman Nayanaja Kshatriya Kuruba (Arya) Lingayat Salahuv Vakaliga Naik Viswasbrahman Yadava Gangakula Khatri Brahman Enumeration Vaishya Rajput Brahman Mehtar Separation From Arains Rajput Rajput Brahman Mehra Rajput Kamboj Rajput Brahman Raja Brahman

Brahman Dhiman Brahman contd.

Appendix C Caste Name

Year

Province

Ahir Khati Sonar Khati

1911 1911 1921 1921

Rajasthan Rajasthan Rajasthan Rajasthan

Mali Kacchi Daroga, Chakar Khati Nai Mali Chakkilyan (Chakala) Nadar ( Chanan, Shanan) Kammalan (Visvabraham ) Cheruman or Pulayan Pariyan Nadar Ampattan Jat Kalwar Sonar Gujar Kayastha Murao Kurmi Gadariya Barhai Kurmi Chamar Lohar

1921 1921 1931 1931 1931 1931 1921 1921 1931 1931 1931 1931 1931 1901 1901 1901 1901 1901 1901 1901 1911 1911 1911 1911 1911

Rajasthan Rajasthan Rajasthan Rajasthan Rajasthan Rajasthan Travancore Travancore Travancore Travancore Travancore Travancore Travancore United Provinces United Provinces United Provinces United Provinces United Provinces United Provinces United Provinces United Provinces United Provinces United Provinces United Provinces United Provinces

Sonar Kayastha

1911 1911

United Provinces United Provinces

Ahir Kahar

1921 1921

United Provinces United Provinces

Kurmi Kalwar

1921 1921

United Provinces United Provinces

225 New Name (if any) Rajput Jangra-Brahman Mair Kshatriya Jangida Maithil Brahman Sani Rajput Kacchwara Rajput Rawana Rajput Jangida Brahman Nai-Brahman Sainik Kshatriya Nair Nadar Kshatriya Viswasbrahman Pulayan Sambavar Nadar Velakkirhalanayar Kshatriya Karanwal Ks Kshatriya Kshatriya Bagban Kshatriya Kshatriya Brahman Kurmi Kshatriya Subcaste Vishvakarmabansi Brahman Kshatriya Chitraguptabansi Kshatriya Ahir Kshatriya Chandra Vanshya Kshatriya Kurmi Kshatriya Bathan Vaish contd.

226

Appendix C

Caste Name

Year

Province

Sonar Barhai Bhuminyar Gadariya Lodh Kachhi Lohar Barhai Gujar Ahar Lodh Luniya Jat Chamar Bhuminyar

1921 1921 1921 1921 1921 1921 1921 1931 1931 1931 1931 1931 1931 1931 1931

United United United United United United United United United United United United United United United

Provinces Provinces Provinces Provinces Provinces Provinces Provinces Provinces Provinces Provinces Provinces Provinces Provinces Provinces Provinces

Kalwar Dhobi Koeri Teli Ahir Kumhar Lohar Kayastha

1931 1931 1931 1931 1931 1931 1931 1931

United United United United United United United United

Provinces Provinces Provinces Provinces Provinces Provinces Provinces Provinces

Kisan Kurmi Nai Gadariya Sonar

1931 1931 1931 1931 1931

United United United United United

Provinces Provinces Provinces Provinces Provinces

Bhar

1931

United Provinces

Source: Census of India, various years.

New Name (if any) Mair Rajput Maithil Brahman Bhuminyar Brahman Pali Rajput Lodhi Rajput Kacchwaha Chhatri Dhiman Brahman Dhiman Brahman Kshatriya Yadava Kshatriya Lodhi Rajput Chauhan Rajput, Thakur Jaduvanshi Thakur, Kshatriya Jatav, Jatav Rajput Brahman, Sarwai Brahman, Kshatriya Vaishya, Batham Vaishya Chhatri Kushwaha Kshatriya Vaishya Yadava, Jaduvanshi Kshatriya Rajput, Vaisya, Jaiswar Dhiman Brahman Chitra Guptavanshi Kayatha Kshatriya Rajput Kurmi Kshatriya, Kairati Nai Brahman, Brahman, Rajput Pali Rajput Mend Rajput, Vaishya, Yadubanshi Jaduvanshi Kshatriya

Table C.2 Summary statistics Variable

Obs.

Mean

SD

Min.

Max.

Varies by Jati-province-year Male Population Share Male Population 000s Literacy Rate Litrate Sq. Prop of Province with higher Status Proportion Cultivators and Landowners Proportion in Public Administration Proportion in Traditional Occupation Proportion Govt. Officers Industry Owners Rate Industry Owners Rate Sq. Prop. of Gazetted Officers Prop. of Congress Delegates Absolute change in pop.

1130 1130 920 920 1130 606 606 606 732 730 730 730 521 1079

0.028 259.99 0.129 0.044 0.054 0.390 0.019 0.452 0.00038 0.001 0 0 0.016 0.216

0.033 365.63 0.166 0.102 0.112 0.272 0.053 0.282 0.0015 0.003 0 0.002 0.08 1.090

0 1.5 0 0 0 0 0 0 0 0 0 0 0 0

0.277 3226. 0.936 0.877 0.476 0.991 0.592 0.985 0.0204 0.035 0.001 0.02 0.728 29.88

Varies by province-year Proportion Urban Proportion Literate

1130 1130

0.060 0.056

0.026 0.033

0.018 0.023

0.134 0.174 contd.

Variable Proportion in Public Administration Provincial Population Congress Attendees/Pop. 000s Logged Land Revenue per capita Prop. Brahmins Proportion Hindu Prop. Arya Samaj Prop. Christians Prop of Petitions Granted Status Based Census Classif. Prop. Local Boards Elected Prov. Agricultural Employment Rate Per Capita Land Revenue Varies by Jati Untouchable Caste Lower OBC Caste Upper OBC Caste Intermediate Caste Upper Caste

Obs.

Mean

SD

Min.

Max.

969 1130 616 758 1130 1130 1130 1130 1101 1130 0.525 0.66 23.207

0.011 27,532 0.007 22.57 0.068 0.780 0.001 0.021 0.124 0.598 0.228 0.105 9.047

0.007 16,350 0.014 9.187 0.032 0.170 0.003 0.052 0.271 0.490 0.057 0.389 7.83

0.001 2033 0.0002 7.83 0.022 0.315 0 0.001 0 0 0.933 0.839 40.7

0.023 51,087 0.067 40.7 0.131 0.959 0.017 0.315 1 1 828 1428 823

1130 1130 1130 1130 1130

0.218 0.381 0.181 0.116 0.105

0.413 0.486 0.385 0.320 0.307

0 0 0 0 0

1 1 1 1 1

Source: Author’s calculations from Census of India, various years.

Appendix C

229

Table C.3 Ranked petitioning by province, 1901–1931 Province

Prop. of Ranked Petitions

Baroda Bengal Bihar Bombay Central Indian Agency Central Provinces Hyderabad Madras Mysore Punjab Rajasthan Travancore United Provinces

0.64 0.64 0.79 0.90 0.83 0.73 1.00 0.82 0.50 0.80 1.00 0.43 0.80

Source: Author’s calculations from Census of India, various years.

Figure C.2 Proportion of castes petitioning for a name change by province, 1901–1931 BENGAL

BIHAR

BOMBAY

CIA

CP

HYD

MADRAS

MYSORE

PUNJAB

RAJ

TRAV

.2 .4 .6 0 .2 .4 .6 0

Petition Rate

0

.2 .4 .6

BARODA

1900

1910

1920

1930 1900

1910

1920

1930 1900

1910

0

.2 .4 .6

UP

1900

1910

1920

Graphs by province

1930

Year

Source: Author’s calculations from Census of India, various years.

1920

1930

230

Appendix C

Table C.4 Additional alternative models: petitioning Variables Population as Proportion Population 000s Male Literacy Rate Male Literacy Rate Sq. Constant Observations Number of year Province FE Year FE Province-Year Controls

(1) OLS

(2) Lagged Predictors

1.062* (0.592) 4.26e-05 (5.94e-05) 1.197*** (0.283) −1.448*** (0.395) 0.923 (0.706) 891 4 YES YES YES

6.460* (3.850) 0.000336 (0.000402) 8.210*** (2.449) −13.23*** (4.522) −4.172 (8.982) 711 3 YES YES YES

Standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 These models show the results for an OLS model with petitioning as the outcome (Model One) and a logit model with petitioning as the outcome and the reported predictors lagged by one census year (Model Two). The Province-Year Controls (proportion urban, proportion in public employment, provincial population, and provincial literacy rate) are not reported for reasons of space.

Appendix C

231

Table C.5 Additional alternative models: segmentary petitioning Variables Population as Proportion Population 000s Male Literacy Rate Male Literacy Rate Sq. Prop. Local Boards Elected Prop. Local Boards Elected*Population

(1) OLS

−1.284 2.849 (3.820) (2.044) −0.000420* −6.45e-05 (0.000222) (0.000141) −0.901 0.558 (0.880) (1.014) 0.400 −2.420 (1.478) (1.654) −0.0717 (0.601) 15.54** (6.458)

Jati Public Employment Rate Jati Land Cultivation Rate Constant

Observations Number of year Jati-Year FE Province-Year FE Province FE Year FE Province-Year Controls

(2) OLS

(3) (4) OLS Full FE OLS Full FE −5.262 (8.558) −0.000672* (0.000358) −6.967 (4.367) 7.746 (5.907) −2.393 (1.868)

9.101 (6.531) −0.000338 (0.000267) −11.13* (5.865) 14.55* (8.296)

24.83* (14.77) 5.240 (3.758)

47.05 (51.64)

−0.582*** (0.159)

−0.481 (1.026)

8.805 (7.252)

−1.980 (4.754)

3.255* (1.776)

2.295* (1.356)

164 4 NO NO YES YES YES

143 3 NO NO YES YES YES

164 4 YES YES NO NO NO

143 3 YES YES NO NO NO

Standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 These models show the results for OLS models with segmentary petitioning as the outcome. The Province-Year Controls (proportion urban, proportion in public employment, provincial population, and provincial literacy rate) are not reported for reasons of space.

Table C.6 Traditional institutions and social ranking: sequential logistic regression Variables Caste Army/Police Rate Caste Land Cultivation Rate Prop. Local Boards Elected Prop. Local Boards Elected*Population Prov. Congress Delegates Prov. Congress Delegates*Population Population as Proportion Population 00000s Male Literacy Rate Male Literacy Rate Sq.

(1) Petitioning

(2) Petitioning

(3) Petitioning

−14.27 (10.91) −0.130 (0.638)

−16.99 (17.91) −0.330 (0.739)

−10.64 (16.74) −0.388 (0.751)

(4) Petitioning

(5) Petitioning

−13.04 (37.45) −1.019 (0.777) −0.611 (3.281) 3.980 (29.37)

0.387 (1.796) 9.450 (19.20)

8.313 (5.520) 0.0856* (0.0479) 12.56*** (4.223) −24.93*** (9.089)

8.876 (5.499) 0.0927** (0.0450) 13.85*** (4.314) −26.47*** (8.794)

−5.369 (11.83) 0.104 (0.0737) 10.81*** (3.119) −16.60** (6.713)

(6) Petitioning

−38.58 (32.09) 392.4 (468.1) 5.832 (12.25) 0.0648 (0.0712) 10.74*** (3.468) −19.48** (7.766)

3.295 (13.72) 0.0950 (0.0853) 20.01*** (6.901) −52.04** (20.29) contd.

Variables

(1) Petitioning

(2) Petitioning

(3) Petitioning

(4) Petitioning

(5) Petitioning

(6) Petitioning

−2.112*** (0.494)

7.273 (5.734)

1.232 (1.239) 1.516*** (0.551) 5.147 (5.766)

0.434 (2.440)

−1.656 (5.077)

1.747 (5.576)

412 NO YES YES YES

390 YES YES YES YES

390 YES YES YES YES

587 YES YES YES YES

401 YES YES YES YES

300 YES YES YES YES

Caste Ag. Labour Rate Caste Traditional Occupation Rate Constant

Observations Province-Year Controls Province FE Year FE Caste Status FE

Robust standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 These are pooled sequential regression models, with fixed effects for caste status, year, and province, with the standard errors clustered by caste. This table only reports the results of the first equation, where petitioning is the dependent variable. The corresponding results of the second equation, with ranked petitioning as the dependent variable, are reported in Table Two. The Province-Year Controls, proportion urban, proportion in public employment, provincial population, and provincial literacy rate, are not reported for reasons of space.

Table C.7 Traditional institutions and social ranking: sequential logistic regression Variables

(1) No South

(2) No 1931

(3) No Partial Data

(4) No Splits

(5) Only OBCs

Population as Proportion

−44.38*** (15.59) 0.132* (0.0730) −8.451 (7.979) 22.65 (16.50) 5.436*** (1.607) −91.79** (41.64) 19.29 (31.55)

−13.92 (39.69) 0.196 (0.203) −40.48** (17.38) 125.6** (55.66) 2.453 (1.919) −104.6** (44.41) 1.862 (2.695)

−29.21* (15.95) 0.0383 (0.0722) −11.16 (8.890) 27.12 (20.86) 3.711*** (1.276) 37.11 (93.98) 2.350 (1.502)

−86.12*** (28.07) 0.704** (0.334) −0.775 (15.97) 30.93 (36.33) 6.860*** (2.195) −67.76 (50.39) 27.69 (41.00)

20.15 (27.81) −0.174 (0.119) −18.75** (7.703) 60.40*** (20.17) 1.612 (1.377) −210.2** (83.20) 4.645*** (1.657)

Population 00000s Male Literacy Rate Male Literacy Rate Sq. Caste Land Cultivation Rate Caste Army/Police Rate Constant

contd.

Variables Observations Province FE Year FE Province-Year Controls Caste Status FE

(1) No South

(2) No 1931

(3) No Partial Data

(4) No Splits

(5) Only OBCs

357 YES YES YES YES

342 NO YES YES YES

382 NO YES YES YES

366 YES YES YES YES

318 NO YES YES YES

Robust standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 These are pooled sequential regression models, with fixed effects for caste status, year, and province, with the standard errors clustered by caste. Some models do not include province fixed effects, due to high levels of perfect prediction This table only reports the results of the first equation, where petitioning is the dependent variable. The corresponding results of the second equation, with ranked petitioning as the dependent variable, are reported as Table Two. The Province-Year Controls, proportion urban, proportion in public employment, provincial population, and provincial literacy rate, are not reported for reasons of space. Model One does not include Mysore, Travancore, Madras or Hyderabad. Model Two does not include 1931. Model Three does not include castes where the literacy data was calculated for less than 80% of the caste. Model Four does not include castes where petitions were filed on behalf of a sub-caste. Model Five excludes both ‘clean’ and untouchable castes.

Table C.8 Participatory institutions and social ranking: sequential logistic regression Equation

(1) Variables

(2) No South

(3) No 1931

(4) No Partial Data

(5) No Splits

Population as Proportion

10.99 (28.41) 0.244** (0.124) 0.318 (8.562) 13.11 (21.52) −0.609 (3.939) −93.48** (42.67) 1.032 (4.762)

−99.69 (64.30) 0.422 (0.313) −8.730 (12.70) 37.41 (28.42) −39.60 (26.42) 124.8 (92.22) 12.27 (19.37)

45.88 (41.55) 0.624** (0.269) −14.10 (9.954) 54.41** (26.58) −0.855 (4.520) −234.6*** (70.40) −10.76 (7.619)

−72.87 (66.23) 0.818** (0.361) 2.899 (10.14) 4.438 (26.52) −4.491 (4.410) −37.79 (82.49) −1.269 (5.536)

45.31 (44.80) 0.0111 (0.230) −4.683 (10.47) 26.61 (25.99) 2.694 (6.312) −94.89 (77.17) −5.460 (9.250)

Population 00000s Male Literacy Rate Male Literacy Rate Sq. Prop. Local Boards Elected Prop. Local Boards Elected x Population Constant

Only OBCs

contd.

Equation Observations Province FE Year FE Province-Year Controls Caste Status FE

(1) Variables

(2) No South

(3) No 1931

(4) No Partial Data

(5) No Splits

498 YES YES YES YES

381 YES YES YES YES

333 YES YES YES YES

468 YES YES YES YES

320 YES YES YES YES

Only OBCs

Robust standard errors in parentheses *** p < 0.01, ** p < 0.05, * p < 0.1 These are pooled sequential regression models, with fixed effects for caste status, year, and province, with the standard errors clustered by caste. Some models do not include province fixed effects, due to high levels of perfect prediction This table only reports the results of the first equation, where petitioning is the dependent variable. The corresponding results of the second equation, with ranked petitioning as the dependent variable, are reported as Table Two. The Province-Year Controls, proportion urban, proportion in public employment, provincial population, and provincial literacy rate, are not reported for reasons of space. Model One does not include Mysore, Travancore, Madras or Hyderabad. Model Two does not include 1931. Model Three does not include castes where the literacy data were calculated for less than 80% of the caste. Model Four does not include castes where petitions were filed on behalf of a sub-caste. Model Five excludes both ‘clean’ and untouchable castes.

238

Appendix C Table C.9 Traditional institutions and social ranking: Heckman selection model

Equation

Variable

Ranking

Population as Proportion Population 00000s Caste Army/Police Rate Caste Land Cultivation Rate

(1)

(2)

−2.59 (1.65 ) 0.006 (0.011) −11.3*** (3.47) 0.676*** (0.137)

0.155 (3.86) 0.035*** (0.013)

Prop. Local Boards Elected

0.128 (0.245) −12.2** (5.03)

Prop. Local Boards Elected x Population Petitioning

Population as Proportion Population 00000s Male Literacy Rate Male Literacy Rate Sq. Caste Army/Police Rate Caste Land Cultivation Rate

3.01 (2.97) 0.056** (0.025) 7.21*** (2.16) −12.97*** (4.62) −5.11 (7.14) 0.138 (0.417)

Prop. Local Boards Elected Prop. Local Boards Elected* Population Observations Province FE Year FE Caste Status FE

391 YES YES YES

2.81 (6.82) 0.063 (0.03) 6.79*** (1.66) −9.90** (4.14)

0.64 (0.465) −3.71 (11.55) 539 YES YES YES

Robust standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1 These are pooled Heckman selection models, with fixed effects for caste status, year, and province, with the standard errors clustered by caste. The first panel reports results for ranked and unranked petitioning, the second panel the selection equation for petitioning versus not petitioning.

Table C.10 Robustness checks: subsamples Variables Population as Proportion Population 00000s Male Literacy Rate Male Literacy Rate Sq. Caste Land Cultivation Rate Caste Army/Police Rate

(1) British Provinces

(2) British Provinces

(3) Zam. Provinces

(4) Zam. Provinces

−58.34*** (18.21) 0.213** (0.0890) −3.452 (9.689) 27.29 (21.46) 5.513*** (1.624) −59.49 (37.95)

14.08 (31.51) 0.218* (0.130) 1.969 (6.790) 13.34 (14.43)

22.12 (42.59) −0.228 (0.221) −25.71* (15.13) 78.44* (40.13) 4.246** (1.824) 14.98 (192.1)

26.89 (30.33) 0.339 (0.309) −15.00 (9.460) 51.42** (24.92)

Prop. Local Boards Elected Prop. Local Boards Elected x Population Constant

−0.514 (2.404)

−1.074 (2.148) −90.79** (44.70) 3.638** (1.750)

−81.87 (221.3)

−1.956 (2.424) −144.9** (72.03) 2.888 (1.918) contd.

Variables Observations Province-Year Controls Year FE Province FE Caste Status FE

(1) British Provinces

(2) British Provinces

(3) Zam. Provinces

(4) Zam. Provinces

400 YES YES YES YES

657 YES YES YES YES

208 YES YES YES YES

400 YES YES YES YES

Robust standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1 These are pooled sequential regression models, with fixed effects for caste status, year, and province, with the standard errors clustered by caste. Models One and Two include only provinces with a majority of the population under direct British rule, while Models Three and Four include only provinces with over a quarter of the land under zamindari tenure in 1931. This table only reports the results of the second equation, where ranked petitioning is the dependent variable. The corresponding results of the first equation, with petitioning as the dependent variable, are not reported. The Province-Year Controls, proportion urban, proportion in public employment, provincial population, and provincial literacy rate, are not reported for reasons of space.

Appendix C

241

Table C.11 OBC reservations in India: a summary State Some Post 1947 2011 Subquota Notes Colonial Res. Began OBC Year Res. Quota Andhra Yes 1947 23% 1970 System reformed on Pradesh state reorganization 1964 Tamil Nadu Yes 1947 68% 1989 Quota expanded from 25% before 1971 and 31% before 1980 Maharashtra No 1964 32% 1964 Kerala Yes 1964 40% Punjab No 1964 12% Quota 5% before 1993 Karnataka Yes 1977 32% 1977 50% quota pre-1956, changed to class-based 1961, 35% 1977, 55’% quota 1986, 32% quota 1994 Bihar No 1978 33% 1978 Quota increased from 20% in 1992 Gujarat No 1978 27% Jharkhand No 1978 14% 1978 Quota shrunk from 33% in 2000 on state reorganization Haryana No 1991 27% 1995 Class-based res. from 1969 Assam No 1994 27% Chhattisgarh No 1994 14% Delhi No 1994 27% Goa No 1994 19.5% Himachal No 1994 20% Class-based res. Pradesh from 1970 Madhya No 1994 14% Pradesh Manipur No 1994 17% Orissa No 1994 12% Rajasthan No 1994 21% Sikkim No 1994 23% Uttar Pradesh No 1994 27% Uttarakhand No 1994 14% Quota shrunk from 27% in 2000 on state reorganization West Bengal No 1994 7% Source: Based on Mathur (2004). Note: Tripura, Meghalaya, Mizoram, and Arunachal Pradesh do not have OBC reservations; Jammu and Kashmir has class-based reservations. Union territories are not included for space reasons.

Bibliography Abdelal, Rawi, Y. M. Herrera, A. I. Johnston, and R. McDermott. 2006. ‘Identity as a Variable.’ Perspectives on Politics 4(4): 695–711. Acemoglu, Daron, Simon Johnson, and James A. Robinson. 2001. ‘The Colonial Origins of Comparative Development: An Empirical Investigation.’ American Economic Review 91(5): 1369–1401. ———. 2002. ‘Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution.’ Quarterly Journal of Economics 117(4): 1231–1294. Ahmad, Imtiaz, ed. 1978. Caste and Social Stratification among Muslims in India. Columbia, MO: Asia Books. Ahuja, Amit and Susan L. Ostermann. 2016. ‘Crossing Caste Boundaries in the Modern Indian Marriage Market.’ Studies in Comparative International Development 51(3): 365–387. Alcoff, Linda. 2006. Visible Identities: Race, Gender, and the Self. New York: Oxford University Press. Alesina, Alberto, Reza Baqir, and William Easterly. 1999. ‘Public Goods and Ethnic Divisions.’ The Quarterly Journal of Economics 114(4): 1243–1284. Alesina, Alberto and Eliana La Ferrara. 2005. ‘Ethnic Diversity and Economic Performance.’ Journal of Economic Literature 43(3): 762–800. Ali, Imran. 1979. ‘The Punjab Canal Colonies, 1885–1940.’ Ph.D. thesis, Australian National University, 1980. Akil Bhartiya Jat Mahasabha. Available at https://www.facebook.com/ jatmahasabha.akhilbhartiya. Accessed 5 June 2017. Allport, G. W. 1954. The Nature of Prejudice. Cambridge: Addison-Wesley. Ambedkar, Bhimrao Ramji. 2014. Annihilation of Caste: The Annotated Critical Edition. London and New York: Books. Amos, Timothy D. 2011. Embodying Difference: The Making of Burakumin in Modern Japan. Honolulu: University of Hawaii Press. Anderson, Benedict. 1994. Imagined Communities. New York: Verso. Anderson, Siwan, Patrick Francois, and Ashok Kotwal. 2011. ‘One Kind of Democracy.’ Working Paper, Bristol Institute of Public Affairs, University of Bristol.

Bibliography

243

———. 2015. ‘Clientelism in Indian Villages.’ American Economic Review 105(6): 1780–1816. Arnold, David, Robin Jeffrey, and James Manor. 1976. ‘Caste Associations in South India: A Comparative Analysis.’ The Indian Economic and Social History Review 13(3): 353–373. Assayag, J. 1995. The Making of Democratic Inequality. Caste, Class, Lobbies and Politics in Contemporary India (1880–1995). Pondichery: Institut francais de Pondichery. Austen-Smith, David and Roland G. Fryer. 2005. ‘An Economic Analysis of “Acting White”.’ The Quarterly Journal of Economics 120(2): 551–583. Baden-Powell, Baden Henry. 1892. The Land-Systems of British India. Oxford: Clarendon Press. Bakshi, S. R. 1991. Arya Samaj. New Delhi: Anmol. Baldwin, Kate and John D. Huber. 2010. ‘Economic Versus Cultural Differences: Forms of Ethnic Diversity and Public Goods Provision.’ American Political Science Review 104(4): 644–662. Banda, Paul C., Gift Kayira, and Amanda Robinson. 2017. ‘The Political Logic of Ethnic Associations.’ APSA annual meeting paper. Banerjee, Abhijit, Donald Green, Jennifer Green, and Rohini Pande. 2009. ‘Can Voters Be Primed to Choose Better Legislators? Evidence from Two Field Experiments in Rural India.’ Unpublished Paper. Banerjee, Abhijit, Esther Duflo, Maitreesh Ghatak, and Jeanne Lafortune. 2013. ‘Marry for What? Caste and Mate Selection in Modern India.’ American Economic Journal: Microeconomics 5(2): 33–72. Banerjee, Abhijit and Lakshmi Iyer. 2005. ‘History, Institutions and Economic Performance.’ American Economic Review 95(4): 1190–1213. Banerjee, Abhijit V. and Rohini Pande. 2007. ‘Parochial Politics: Ethnic Preferences and Politician Corruption.’ Working Paper, MIT Economics. Banerjee, Abhijit and Rohini Somanathan. 2007. ‘The Political Economy of Public Goods: Some Evidence from India.’ Journal of Development Economics 82(2): 287–314. Banerjee, Mukulika. 2017. Why India Votes? New Delhi: Routledge. Barrier, Norman Gerald, ed. 1981. The Census in British India: New Perspectives. Delhi: Manohar. Barnett, Steve. 1977. ‘Identity Choice and Caste Ideology in Contemporary South India.’ In The New Wind: Changing Identities in South Asia, ed. Kenneth David. The Hague: Mouton, 393–414. Barth, Frederick. 1969. Ethnic Groups and Boundaries: The Social Organization of Culture Difference. Boston: Little Brown. Bates, Crispin. 1995. Race, Caste and Tribe in Central India: The Early Origins of Indian Anthropometry. University of Edinburgh, Centre for South Asian Studies. Bates, Robert. 1974. ‘Ethnic Competition and Modernization in Contemporary Africa.’ Comparative Political Studies 6(4): 457–484.

244

Bibliography

———. 1983. ‘Modernization, Ethnic Competition, and the Rationality of Politics in Contemporary Africa.’ State Versus Ethnic Claims: African Policy Dilemmas, ed. Donald S. Rothchild and Victor Olorunsola. Boulder, CO: Westview Press, 152–171. Bayly, Christopher A. 1971. ‘Local Control in Indian Towns? The Case of Allahabad 1880–1920.’ Modern Asian Studies 5(4): 289–311. ———. 1983. Rulers, Townsmen, and Bazaars: North Indian Society In the Age of British Expansion, 1770–1870. Cambridge, UK: Cambridge University Press. Bayly, Susan. 1999. Caste, Society and Politics in India: From the Eighteenth Century to the Modern Age. Cambridge: Cambridge University Press. Bellenoit, Hayden J. 2017. The Formation of the Colonial State in India: Scribes, Paper and Taxes, 1760–1860. London: Routledge. Bendix, Reinhard. 1977. Nation-Building and Citizenship. Berkeley: University of California Press. Berman, Bruce J. 1998. ‘Ethnicity, Patronage and the African State: The Politics of Uncivil Nationalism.’ African Affairs 97(388): 305–341. Besley, Timothy, Rohini Pande, and Vijayendra Rao. 2012. ‘Just Rewards? Local Politics and Public Resource Allocation in South India.’ The World Bank Economic Review 26(2): 191–216. Bharatiya Janata Party. ‘Election Manifesto 2014.’ Available at https://issuu. com/rajamama/docs/bjp manifesto. Accessed 11 September 2018. Bhavnani, Rikhil R. and Alexander Lee. 2018. ‘Local Embeddedness and Bureaucratic Performance: Evidence from India.’ The Journal of Politics 80(1): 71–87. Blair, Harry W. 1980. ‘Rising Kulaks and Backward Classes in Bihar: Social Change in the Late 1970s.’ Economic and Political Weekly 15(2): 64–74. Blanning, Timothy C. W. 2003. Culture of Power and the Power of Culture: Old Regime Europe 1660–1789. Oxford: Oxford University Press. Blunt, Edward. 1912. Census of India. United Provinces. Vol. XVI Part I, Report. Lucknow: Census of India. Boag, G.T. 1922. Census of India. Madras. Vol. XIII Part I, Report. Calcutta: Census of India. Bose, Sugata. 1993. Peasant Labour and Colonial Capital: Rural Bengal Since 1770. Cambridge: Cambridge University Press. Bougle, C. 1958. ‘The Essence and Reality of the Caste System.’ Contributions to Indian Sociology 2: 7–30. Bourdieu, Pierre. 1984. Distinction. Translated by Richard Nice. Cambridge: Harvard University Press. Brass P. R. 1974. Language and Politics in North India. London: Cambridge University Press. ———. 1984. Caste, Faction, and Party in Indian Politics: Faction and Party. Delhi: Chanakya Publications. ———, ed. 1985. Ethnic Groups and the State. London: Taylor & Francis. ———. 1993. ‘Chaudhuri Charan Singh: An Indian Political Life.’ Economic and Political Weekly 28(39): 2087–2090.

Bibliography

245

Brewer, M. B. 1999. ‘The Psychology of Prejudice: Ingroup Love and Outgroup Hate?’ Journal of Social Issues 55(3): 429–444. Brokaw, Cynthia Joanne. 2014. The Ledgers of Merit and Demerit: Social Change and Moral Order in Late Imperial China. Princeton: Princeton University Press. Brown, David. 2003. The State and Ethnic Politics in Southeast Asia. London: Routledge. Brown, Penelope and Stephen C. Levinson. 1979. ‘Social Structure, Groups and Interaction.’ In Social Markers in Speech, ed. Klaus R. Scherer and Howard Giles. Cambridge: Cambridge University Press. 291–341. Burdick, John. 1992. ‘Brazil’s Black Consciousness Movement.’ Report on the Americas 25(4): 23–46. Carlson, Elizabeth. 2015. ‘Ethnic Voting and Accountability in Africa: A Choice Experiment in Uganda.’ World Politics 67(02): 353–385. Carroll, Lucy. 1978. ‘Colonial Perceptions of Indian Society and the Emergence of Caste(s) Associations.’ The Journal of Asian Studie 37(2): 233–250. ———. 1979. ‘The Seavoyage Controversy and the Kayasthas of North India, 1901–1909.’ Modern Asian Studies 13(02): 265–299. Cassan, Guilhem. 2015. ‘Identity-Based Policies and Identity Manipulation: Evidence from Colonial Punjab.’ American Economic Journal: Economic Policy 7(4): 103–131. Cassan, Guilhem and Lore Vandewalle. 2017. ‘Identities and Public Policies: Unintended Effects of Political Reservations for Women in India.’ Working Paper No. 18-2017. The Graduate Institute of International Studies. Cederman, Lars-Erik, Nils B. Weidmann, and Kristian Skrede Gleditsch. 2011. ‘Horizontal Inequalities and Ethnonationalist Civil War: A Global Comparison.’ American Political Science Review 105(03): 478–495. Chakravarti, Anand. 2001. Social Power and Everyday Class Relations: Agrarian Transformation in North Bihar. New Delhi; Thousand Oaks, CA: Sage Publications. Chakravarti, Sudeep. 2010. Red Sun: Travels in Naxalite Country. India: Penguin Books. Chakravorty, Abhimanyu. 26 August 2015. ‘Hardik Patel Leads Quota Protests: Here Are His Top 10 Quotes.’ Indian Express. Available at https://indianexpress.com/article/india/india-others/hardik-patel-leads-obcquota-protests-here-are-his-top-10-quotes/. Accessed 11 September 2018. Chandra, Kanchan. 2004. Why Ethnic Parties Succeed. Cambridge: Cambridge University Press. ———. 2009. ‘Why Voters in Patronage Democracies Split Their Tickets: Strategic Voting for Ethnic Parties.’ Electoral Studies 28(1): 21–32. ———. 2012. Constructivist Theories of Ethnic Politics. Oxford: Oxford University Press. Chari, R. B. 1977. Census of India 1971. New Delhi: Office of the Registrar General, Government of India.

246

Bibliography

Chatterjee, Partha. 1993. The Nation and Its Fragments: Colonial and Postcolonial Histories. Princeton, NJ: Princeton University Press. Chaudhary, Latika. 2009. ‘Determinants of Primary Schooling in British India.’ The Journal of Economic History 69(01): 269–302. Chauhan, Anil Kumar Dr Sangeeta. 2015. ‘Demand of Gurjars for Scheduled Tribe Reservation.’ Global Journal of Multidisciplinary Studies 4(8). Chen, Yan and Sherry Xin Li. 2009. ‘Group Identity and Social Preferences.’ The American Economic Review 99(1): 431–457. Chhibber, Pradeep K. 2010. Democracy Without Associations: Transformation of the Party System and Social Cleavages in India. Ann Arbor: University of Michigan Press. Chin, Aimee and Nishith Prakash. 2011. ‘The Redistributive Effects of Political Reservation for Minorities: Evidence from India.’ Journal of Development Economics 96(2): 265–277. Chiriyankandath, James. 1992. ‘Democracy under the Raj: Elections and Separate Representation in British India.’ Journal of Commonwealth and Comparative Politics 30(1): 39–63. Chowdhry, Prem. 1984. Punjab Politics: The Role of Sir Chhotu Ram. Delhi: Vikas Pub. Ciotti, Manuela. 2006. ‘In the Past We Were a Bit “Chamar”’: Education As a Self- and Community-Engineering Process in Northern India.’ Journal of the Royal Anthropological Institute 12(4): 899–916. Cohn, Bernard S. 1960. ‘The Initial British Impact on India: A Case Study of the Benares Region.’ The Journal of Asian Studies 19(4): 418–431. ———. 1987. An Anthropologist among the Historians and Other Essays. New York: Oxford University Press. Cole, B. L. 1932. Census of India. Rajputana Agency. Vol. XXVII Part I, Report. Calcutta: Census of India. Communist Party of India (Marxist). 1964 (updated 2000). ‘Party Programme.’ Available at https://www.cpim.org/party-programme. Accessed 11 September 2018. Conlon, Frank F. 1981. ‘The Census of India as a Source for Historical Study of Religion and Caste.’ In The Census in British India: New Perspectives, ed. Gerald Barrier. Delhi: Manohar, 103–118. Dabas, Harveer. ‘Jat Mahasabha Leaders Protest for Quotas.’ 28 December 2015. Available at http://timesofindia.indiatimes.com/articleshow/50358665.cms? utm source=contentofinterest&utm medium=text&utm campaign=cppst. Accessed 11 September 2018. Darden, Keith and Anna Grzymala-Busse. 2006. ‘The Great Divide: Precommunist Schooling and Postcommunist Trajectories.’ World Politics 59(1): 83–115. Darling, Malcolm. 1925. The Punjab Peasant in Prosperity and Debt. Pusa: Agricultural Research Institute. Datta, J. N. 1922. Census of India. Central Indian Agency. Vol. XX Part I, Report. Calcutta: Census of India.

Bibliography

247

Datta, Nonica. 1997. ‘Arya Samaj and the Making of Jat identity.’ Studies in History 13(1): 97–119. Davidson, Basil. 1992. The Black Man’s Burden: Africa and the Curse of the Nation-State. London and New York: Times Books/Random House. Degler, Carl N. 1971. Neither Black Nor White: Slavery and Race Relations in Brazil and the United States. Madison: University of Wisconsin Press. Desai, Sonalde and Reeve Vanneman. 2015. India Human Development Survey-II (IHDS-II), 2011–12. ICPSR36151-v2. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor] 31. Desdunes, Rodolphe L. 1907. A Few Words to Dr. DuBois with Malice toward None. New Orleans: n.p. Deshpande, Rajeshwari. 2004. ‘Kunbi Maratha as OBC: Backward Journey of a Caste.’ Economic and Political Weekly 3(10): 1448–1449. Deutsch, Karl. 1969. Nationalism and Its Alternatives. New York: Knopf. Dirks, Nicholas B. 1993. The Hollow Crown: Ethnohistory of an Indian Kingdom. 2nd ed. Ann Arbor: University of Michigan Press. ———. 2002. Castes of Mind. Princeton: Princeton University Press. Dodson, Michael. 2007. Orientalism, Empire, and National Culture: India, 1770–1880. New York: Palgrave. Dodwell, Henry Herbert. 1929. The Cambridge History of the British Empire. Vol. 2. CUP Archive. Dubois, Abbe J. A. 1906. Hindu Manners, Customs, and Ceremonies. Translated by Henry Beuchamp. Oxford: Clarendon Press. Dumont, Louis. c1980 [1966]. Homo Hierarchicus. Chicago: University of Chicago Press. Dunning, Thad. 2009. ‘The Salience of Ethnic Categories.’ Society for Political Methodology, Yale University Working Paper. Dunning, Thad and Lauren Harrison. 2010. ‘Cross-Cutting Cleavages and Ethnic Voting: An Experimental Study of Cousinage in Mali.’ American Political Science Review 104(1): 21–39. Dunning, Thad and Janhavi Nilekani. 2013. ‘Ethnic Quotas and Political Mobilization: Caste, Parties, and Distribution in Indian Village Councils.’ American Political Science Review 107(01): 35–56. Easterly, William and Ross Levine. 1997. ‘Africa’s Growth Tragedy: Policies and Ethnic Divisions.’ The Quarterly Journal of Economics 112(4): 1203–1250. Edye, E. H. H. 1922. Census of India. United Provinces. Vol. XVI Part I, Report. Lucknow: Census of India. Eifert, Benn, Edward Miguel, and Daniel N. Posner. 2010. ‘Political Competition and Ethnic Identification in Africa.’ American Journal of Political Science 54(2): 494–510. Eisenstadt, Shmuel N. 1973. Traditional Patrimonialism and Modern Neopatrimonialism. Beverly Hills: Sage Publications. Ejdemyr, Simon, Eric Kramon, and Amanda Lea Robinson. 2017. ‘Segregation, Ethnic Favoritism, and the Strategic Targeting of Local Public Goods. Comparative Political Studies 51(9): 1111–1143.

248

Bibliography

Eldersveld, Samuel James, Bashiruddin Ahmed, and Dwaine Marvick. ‘Indian National Election Study, 1967, 1971, 1979, 1985.’ Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2011-04-12. Available at https://doi.org/10.3886/ICPSR25402.v1, accessed 1 October 2019. Ergul, F. Asli. 2012. ‘The Ottoman Identity: Turkish, Muslim or Rum?’ Middle Eastern Studies 48(4): 629–645. Farwell, Byron. 1989. Armies of the Raj: From the Mutiny to Independence, 1858–1947. New York: W.W. Norton and Company. Fearon, James D. 2003. ‘Ethnic and Cultural Diversity by Country.’ Journal of Economic Growth 8(2): 195–222. Fearon, James D. and David D. Laitin. 2000. ‘Violence and the Social Construction of Ethnic Identity.’ International Organization 54(4): 845–877. Fernandes, Florestan. 1969. The Negro in Brazilian Society, ed. Phyllis B. Eveleth. New York: Columbia University Press. Fordham, Signithia and John U. Ogbu. 1986. ‘Black Students’ School Success: Coping with the Burden of Acting White.’ The Urban Review 18(3): 176–206. Foucault, Michel. 2012. Discipline and Punish: The Birth of the Prison. New York: Vintage. Fox, Richard Gabriel. 1971. Kin, Clan, Raja, and Rule: State-Hinterland Relations in Preindustrial India. Berkeley: University of California Press. Fox-Genovese, Elizabeth and Eugene D. Genovese. 1987. ‘The Divine Sanction of Social Order: Religious Foundations of the Southern Slaveholders’ World View.’ Journal of the American Academy of Religion 55(2): 211–233. Francis, W. 1902. Census of India. Madras. Vol. XV Part I, Report. Madras: Census. Franck, Raphael and Ilia Rainer. 2012. ‘Does the Leader’s Ethnicity Matter? Ethnic Favoritism, Education, and Health in Sub-Saharan Africa.’ American Political Science Review 106(2): 294–325. Frankel, Francine and M. S. A. Rao, eds. 1989–1990. Dominance and State Power in Modern India. Delhi: Oxford University Press. Freitag, Sandria B. 1989. Collective Action and Community: Public Arenas and the Emergence of Communalism in North India. Berkeley: University of California Press. Freitas, Kripa. 2010. ‘The Indian Caste System as a Means of Contract Enforcement.’ Stanford Institute for Theoretical Economics Seminar Paper. Freyre, Gilberto. 1970. The Masters and the Slaves: A Study in the Development of Brazilian Civilization. New York: Alfred A. Knopf. Gait, E. A. 1902. Census of India. Bengal. Vol. VI Part I, Report. Calcutta: Census of India. Galanter, Marc. 1969. ‘Untouchability and the Law.’ Economic and Political Weekly 4(1/2): 131–170. ———. 1984. Competing Equalities: Law and the Backward Classes in India. Berkeley: University of California Press.

Bibliography

249

Gellner, E. 1983. Nations and Nationalism. Ithaca: Cornell University Press. Ginsberg, Elaine K. 1996. Passing and the Fictions of Identity. Durham: Duke University Press. Gould, Henry. 1987. The Hindu Caste System. Vol. 1. Delhi: Chanakya Publications. Grierson, George Abraham, ed. 1906. Linguistic Survey of India. India: Office of the Superintendent of Government Printing. Guha, S. 2003. ‘The Politics of Identity and Enumeration in India c. 1600–1990.’ Comparative Studies in Society and History 45(1): 148–167. Gupta, Dipankar. 1997. Rivalry and Brotherhood: Politics in the Life of Farmers in Northern India. Oxford: Oxford University Press. ———. 2000. Interrogating Caste. New Delhi: Penguin. Habyarimana, James, Macartan Humphreys, Daniel N. Posner, and Jeremy M. Weinstein. 2007. ‘Why Does Ethnic Diversity Undermine Public Goods Provision?’ American Political Science Review 101(4): 709–725. Hanchard, Michael George. 1998. Orpheus and Power: The ‘Movimento Negro’ of Rio de Janeiro and Sao Paulo, Brazil 1945–1988. Princeton: Princeton University Press. Hansen, Thomas Blom. 1999. The Saffron Wave: Democracy and Hindu Nationalism in Modern India. Princeton: Princeton University Press. Hardgrave Jr, Robert L. 1968. ‘Caste: Fission and Fusion.’ Economic and Political Weekly 3 (26/28), special number: 1065–1070. ———. 1969. The Nadars of Tamilnad: The Political Culture of a Community In Change. Berkeley: University of California Press. Harris, Marvin D. 1964. ‘Racial Identity in Brazil.’ Luso-Brazilian Review 1(2): 21–28. Hechter, Michael. 1971. ‘Towards a Theory of Ethnic Change.’ Politics and Society 2(1): 21–45. Hickman, Christine B. 1997. ‘The Devil and the One Drop Rule: Racial Categories, African Americans, and the US Census.’ Michigan Law Review 95(5): 1161–1265. Hochschild, Jennifer L. and Brenna Marea Powell. 2008. ‘Racial Reorganization and the United States Census 1850–1930: Mulattoes, Half-Breeds, Mixed Parentage, Hindoos, and the Mexican Race.’ Studies in American Political Development 22(01): 59–96. Hodgkin, Thomas. 1956. Nationalism in Colonial Africa. London: Muller. Hoff, Karla, Priyanka Pandey, and Benjamin Brooks. 2010. ‘Can Social Structure Shape Preferences Detrimental to Development? Theory and Evidence from Simple Games in Rural India.’ World Bank Working Paper. Horowitz, David. 1985. Ethnic Groups in Conflict. Cambridge: Cambridge University Press. Htun, Mala. 2004. ‘From “Racial Democracy” to Affirmative Action: Changing State Policy on Race in Brazil.’ Latin American Research Review 39(1): 60–89. Huber, John D. and Pavithra Suryanarayan. 2016. ‘Ethnic Inequality and the Ethnification of Political Parties.’ World Politics 68(01): 149–188.

250

Bibliography

Huntington, Samuel. 1969. Political Order in Changing Societies. New Haven: Yale University Press. Hutton, J. H. 1932. Census of India. Vol. I Part I, Report. Calcutta: Census of India. Ichino, Nahomi and Noah L. Nathan. 2013. ‘Crossing the Line: Local Ethnic Geography and Voting in Ghana.’ American Political Science Review 107(02): 344–361. Imber, Colin. 2009. The Ottoman Empire, 1300–1650: The Structure of Power. New York: Palgrave Macmillan. Inden, Ronald B. 1990. Imagining India. Bloomington: Indiana University Press. India Backward Classes Commission. 1980. Report of the Backward Classes Commission. New Delhi: Controller of Publications, Government of India. India Today. 2015. ‘If I Am Devil, Modi Is Super Demon, Says Lalu Prasad.’ 9 October. Available at https://www.indiatoday.in/india/story/if-iam-devil-modi-is-super-demon-says-lalu-prasad-267263-2015-10-09. Accessed 11 September 2018. Indian Police Commission. 1905. Report of the Indian Police Commission and Resolution of the Government of India. London: His Majesty’s Stationary Office. Irschick, Eugene F. 1969. Politics and Social Conflict in South India: The NonBrahman Movement and Tamil Separatism, 1916–1929. Berkeley: University of California Press. Iversen, Vegard, Richard Palmer-Jones, and Kunal Sen. 2013. ‘On the Colonial Origins of Agricultural Development in India: A Re-examination of Banerjee and Iyer.’ Journal of Development Studies 49(12): 1631–1646. Jaffrelot, Christophe. 1998. ‘The Bahujan Samaj Party in North India: No Longer Just a Dalit Party?’ Comparative Studies of South Asia, Africa and the Middle East 18(1): 35–51. ———. 1999. The Hindu Nationalist Movement and Indian Politics: 1925 to the 1990s: Strategies of Identity-Building, Implantation and Mobilisation. New Delhi: Penguin Books India. ———. 2000. ‘Sanskritization vs. Ethnicization in India: Changing Identities and Caste Politics Before Mandal.’ Asian Survey 40(5): 756–766. ———. 2003. India’s Silent Revolution. New York: Columbia University Press. ———. 2005. Dr. Ambedkar and Untouchability: Fighting the Indian Caste System. New York: Columbia University Press. ———. 2016. ‘Quota for Patels? The Neo-middle-class Syndrome and the (Partial) Return of Caste Politics in Gujarat.’ Studies in Indian Politics 4(2): 218–232. Jaffrelot, Christophe and Gilles Verniers. 2012. ‘Castes, Communities and Parties in Uttar Pradesh.’ Economic and Political Weekly 47(32): 89–93. Jaffrelot, Christophe and Sanjay Kumar, eds. 2012. Rise of the Plebeians? The Changing Face of the Indian Legislative Assemblies. New Delhi: Routledge.

Bibliography

251

Jeffrey, Craig. 2001. ‘A Fist is Stronger than Five Fingers: Caste and Dominance in Rural North India.’ Transactions of the Institute of British Geographers 26(2): 217–236. Jenkins, Laura Dudley. 2003. Identity and Identification in India: Defining the Disadvantaged. New Delhi: Routledge. Jensenius, Francesca Refsum. 2012. ‘Political Quotas in India.’ Asian Survey 52(2): 373–394. Jensenius, Francesca R. and Gilles Verniers. 2017. ‘Studying Indian Politics with Large-Scale Data: Indian Election Data 1961–Today.’ Studies in Indian Politics 5(2): 269–275. Jha, Manish K. and Pushpendra. 2012. ‘Governing Caste and Managing Conflicts: Bihar, 1990-2011.’ Policies and Practices 48 (March): 4–31. Jha, Shashi Shekhar. 1972. Political Elite in Bihar. Bombay: Vora. Jha, Saumitra. 2013. ‘Trade, Institutions, and Ethnic Tolerance: Evidence from South Asia.’ American political Science Review 107, no. 4: 806–832. Jha, Saumitra, and Steven Wilkinson. 2012. ‘Does Combat Experience Foster Organizational Skill? Evidence from Ethnic Cleansing During the Partition of South Asia.’ American Political Science Review 106(04): 883–907. Jones, William. 1824. Discourses Delivered Before the Asiatic Society: And Miscellaneous Papers, on the Religion, Poetry, Literature, etc., of the Nations of India. CS Arnold. Just, R. 1989. ‘Triumph of the Ethnos.’ In History and Ethnicity, ed. E. Tonkin, M. McDonald, and M. Chapman. London: Routledge, 71–88. Kasfir, Nelson. 1979. ‘Explaining Ethnic Political Participation.’ World Politics 31(3): 365–388. Kaur, Ravinder. 2010. ‘Khap Panchayats, Sex Ratio and Female Agency.’ Economic and Political Weekly 45(23): 14–16. Key, Valdimir. 1949. Southern Politics in State and Nation. New York: Knopf. Khanna, Gaurav. Forthcoming. ‘Does Affirmative Action Incentivize Schooling? Evidence from India.’ Review of Economics and Statistics. Kitschelt, Herbert and Steven I. Wilkinson. 2007. ‘Citizen–Politician Linkages: An Introduction.’ In Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition, ed. Herbert Kitschelt and Steven I. Wilkinson. New York: Cambridge University Press, 1–49. Kolff, D. H. A. 1990. Naukar, Rajput, and Sepoy: The Ethnohistory of the Military Labour Market of Hindustan, 1450–1850. Cambridge: Cambridge University Press. Kothari, Rajni. 1970. Caste in Indian Politics. New Delhi: Orient Blackswan. Kumar, Awanish. 2013. ‘Nitish Kumar’s Honourable Exit: A Brief History of Caste Politics.’ Economic and Political Weekly 48(28): 15–17. Kumar, Hemanshu and Rohini Somanathan. 2017. ‘Caste Connections and Government Transfers – The Mahadalits of Bihar.’ Working Paper No. 270. Centre for Development Economics, Delhi School of Economics, New Delhi. Kumar, Sanjay. 2013. Changing Electoral Politics in Delhi: From Caste to Class. India: Sage Publications.

252

Bibliography

Kushner, D. 1977. The Rise of Turkish Nationalism, 1876–1908. London and Totowa, NJ: Cass. Lacina, Bethany. 2014. ‘India’s Stabilizing Segment States.’ Ethnopolitics 13(1): 13–27. ———. 2017. Rival Claims. Ann Arbor: University of Michigan Press. Laitin, David. 1986. Hegemony and Culture. Chicago: University of Chicago Press. ———. 1998. Identity in Formation: The Russian-Speaking Populations in the Near Abroad. Ithaca: Cornell University Press. Lal, Deepak. 1988. The Hindu Equilibrium. Oxford: Clarendon. Landau, Jacob M. 1995. Pan-Turkism: From Irredentism to Cooperation. Indiana University Press. Lankina, Tomila and Lullit Getachew. 2012. ‘Mission or Empire, Word or Sword? The Human Capital Legacy in Postcolonial Democratic Development.’ American Journal of Political Science 56(2): 465–483. Larick, Roy. 1991. ‘Warriors and Blacksmiths: Mediating Ethnicity in East African Spears.’ Journal of Anthropological Archaeology 10(4): 299–331. Leach, E. R. 1969. ‘Introduction.’ In Aspects of Caste in South India, Ceylon and North West Pakistan, ed. E. R. Leach. Cambridge, UK: Cambridge University Press, 1–10. Lee, Alexander. 2011. ‘Who Becomes a Terrorist? Poverty, Education, and the Origins of Political Violence.’ World Politics 63: 203–245. ———. 2017. ‘Redistributive Colonialism: The Long Term Legacy of International Conflict in India.’ Politics and Society 45(2): 173–224. ———. 2018a. ‘Does Affirmative Action Work: Evaluating India’s Quota System.’ University of Rochester Working Paper. ———. 2018b. ‘Ethnic Diversity and Ethnic Discrimination: Explaining Local Public Goods Provision? Comparative Political Studies 51(10): 1351–1383. ———. 2019. Development in Multiple Dimensions: Social Power and Regional Policy in India. Ann Arbor: University of Michigan Press. Lee, Alexander and Kenneth Schultz. 2012. Comparing British and French Colonial Legacies: A Discontinuity Analysis of Cameroon. Quarterly Journal of Political Science 7: 1–46. Lee, Alexander and Varun Ramachandra. Forthcoming. ‘Gender Quotas and Electoral Outcomes: Evidence from Indian Local Elections? American Journal of Political Science. Lele, Jayant. 1989. ‘Caste, Class and Dominance: Political Mobilization in Maharashtra.’ In Dominance and State Power in Modern India: Decline of a Social Order, Vol. 2, ed. Francine R. Frankel and Madhugiri Shamarao Ananthapadmanabha Rao. Oxford: Oxford University Press, 115–211. Leopold, Joan. 1974. ‘British Applications of the Aryan Theory of Race to India, 1850–1870.’ The English Historical Review 89(352): 578–603. Lewis, Oscar and Victor Barnouw. 1958. Village Life in Northern India. Studies in a Delhi Village. New York: Vintage Books.

Bibliography

253

Lieten, Georges Kristoffel. 1982. The First Communist Ministry in Kerala: 1957–9. Calcutta: KP Bagchi. Lindert, Peter H. 2004. Growing Public: Volume 1, the Story: Social Spending and Economic Growth Since the Eighteenth Century. Cambridge: Cambridge University Press. Luong, Pauline Jones. 2004. The Transformation of Central Asia: States and Societies from Soviet Rule to Independence. Ithaca: Cornell University Press. Lynch, Owen M. 1969. The Politics of Untouchability. New York: Columbia University Press. MacMillan, Margaret. 2001. Peacemakers: The Paris Conference of 1919 and Its Attempt to End War. London: Random House. MacMunn, George Fletcher. 1933. The Martial Races of India. Delhi: Mittal Publications. Mamdani, Mahmood. 1996. Citizen and Subject: Contemporary Africa and the Legacy of Late Colonialism. Princeton, NJ: Princeton University Press. ———. 2005. ‘Political Identity, Citizenship and Ethnicity in Post-colonial Africa.’ Key note address at Arusha Conference on ‘New Frontiers of Social Policy’. 12–15 December. Marx, Anthony W. 1998. Making Race and Nation: A Comparison of the United States, South Africa, and Brazil. Cambridge: Cambridge University Press. Marx, Karl. 1853. ‘The Future Results of British Rule in India.’ New York Daily Tribune, 8 August. Modern Publishers. Mathur, Mukat Lal. 2004. Encyclopaedia of Backward Castes: Mandal, Media and Aftermath. New Delhi: Gyan Publishing House. McCaa, Robert, Stuart B. Schwartz, and Arturo Grubessich. 1979. ‘Race and Class in Colonial Latin America: A Critique.’ Comparative Studies in Society and History 21(03): 421–433. Metcalf, Thomas R. 1967. ‘Landlords Without Land: The UP Zamindars Today.’ Pacific Affairs 40(1/2): 5–18. ———. 1979. Land, Landlords, and the British Raj: Northern India in the Nineteenth-Century United Provinces. Berkeley: University of California Press. Michelutti, Lucia. 2002. ‘Sons of Krishna: The Politics of Yadav Community Formation in a North Indian Town.’ Diss. London School of Economics and Political Science. ———. 2004. ‘“We (Yadavs) Are a Caste of Politicians”: Caste and Modern Politics in a North Indian Town.’ Contributions to Indian Sociology 38(1–2): 43–71. Miguel, Edward. 2004. ‘Tribe or Nation? Nation Building and Public Goods in Kenya Versus Tanzania.’ World Politics 56(3): 328–362. Miguel, Edward and Mary Kay Gugerty. 2005. ‘Ethnic Diversity, Social Sanctions, and Public Goods in Kenya.’ Journal of Public Economics 89(11): 2325–2368. Molony, J. Charles. 1912. Census of India. Madras. Vol. XII Part I, Report. Calcutta: Census of India. Mukerjea, S. 1922. Census of India. Baroda. Vol. XVII Part I, Report. Calcutta: Census of India.

254

Bibliography

Munshi, Kaivan and Mark Rosenzweig. 2013. ‘Networks, Commitment, and Competence: Caste in Indian Local Politics.’ No. w19197. National Bureau of Economic Research Working Paper. Nair, P. R. Gopinathan. 1976. ‘Education and Socio-economic Change in Kerala, 1793–1947.’ Social Scientist 4(8): 28–43. Natrajan, Balmurli. 2011. The Culturalization of Caste in India: Identity and Inequality in a Multicultural Age. New Delhi: Routledge. Nehru, Jawaharlal. 1958. Speeches: 1953–57. Delhi: Publications Division. ———. 1989. The Discovery of India. Delhi: Oxford University Press. Nobles, Melissa. 2000. Shades of Citizenship: Race and the Census in Modern Politics. Palo Alto: Stanford University Press. O’Malley, L. S. S. 1912. Census of India. Bengal. Vol. V Part I, Report. Calcutta: Census of India. Omvedt, Gail. 1976. Cultural Revolt in a Colonial Society: The Non-Brahman Movement in Western India, 1873 to 1930. Delhi: Scientific Socialist Education Trust. ———. 1994. Dalits and the Democratic Revolution: Dr Ambedkar and the Dalit Movement in Colonial India. New Delhi: Sage Publications. ———. 2006. Dalit Visions: The Anti-caste Movement and the Construction of an Indian Identity. Delhi: Orient Blackswan. Ordeshook, Peter C. and Olga V. Shvetsova. 1994. ‘Ethnic Heterogeneity, District Magnitude, and the Number of Parties.’ American Journal of Political Science 38(1): 100–123. Pai, Sudha. 2002. Dalit Assertion and the Unfinished Democratic Revolution: The Bahujan Samaj Party in Uttar Pradesh. New Delhi: Sage Publications. Paine, Jack. 2018. ‘Pre-colonial States and Ethnic Violence in Africa.’ Working Paper. Parikh, Sunita. 1997. The Politics of Preference: Democratic Institutions and Affirmative Action in the United States and India. Ann Arbor: University of Michigan Press. Peisakhin, Leonid V. 2010. ‘Living Historical Legacies: The “Why” and “How” of Institutional Persistence: The Case of Ukraine.’ 1 September. Available at SSRN: https://ssrn.com/abstract=1666548 or http://dx.doi.org/ 10.2139/ssrn.1666548. Pillai, N. K. 1932. Census of India. Travancore. Vol. XXVII Part I, Report. Calcutta: Census of India. Posner, Daniel N. 2003. ‘The Colonial Origins of Ethnic Cleavages: The Case of Linguistic Divisions in Zambia.’ Comparative Politics (1 January): 127–146. ———. 2004. ‘The Political Salience of Cultural Difference.’ American Political Science Review 98(4): 529–545. ———. 2005. Institutions and Ethnic Politics in Africa. New York: Cambridge University Press. Potter, David C. 1986. India’s Political Administrators: 1919–1983. New York: Oxford University Press.

Bibliography

255

Pradhan, Mahesh Chandra. 1966. The Political System of the Jats of Northern India. New Delhi: Oxford University Press. Prasad, K. and Srivastava Dusre. 1877. The Kayastha Ethnology, an Inquiry into the Origin of the Chitraguptavansi and Chandrasenavansi Kayasthas. Lucknow: American Methodist Mission Press. Press Trust of India. 2018. ‘Bainsla Urges Raje to Resolve Gujjar Reservation Issue.’ Times of India, 19 August. Available at http://timesofindia. indiatimes.com/articleshow/65462470.cms?utm source=contentofinterest& utm medium=text&utm campaign=cppst. Accessed 11 September 2018. Price, Pamela and Arild Engelsen Ruud. 2012. Power and Influence in India: Bosses, Lords and Captains. New Delhi: Routledge. Puri, Rajinder. 2008. ‘Who Gains From Reservation?’ Outlook, 29 May. Available at https://www.outlookindia.com/website/story/who-gains-fromreservation/237542. Accessed 11 September 2018. Raja, R. Saravana. 2017. ‘The Associational Basis of Vanniyar Organizations in Tamil Nadu.’ Studies in Indian Politics 5(2): 181–192. Ram, P. R. 2015. ‘Hindutva Offensive Social Roots: Characterisation.’ South Asia Citizens Watch. Web 15. Available at http://www.sacw.net/aii/ HIoffensive.html. Accessed 11 September 2018. Ramusack, Barbara N. 2004. The Indian Princes and Their States. Cambridge: Cambridge University Press. Ranger, Terence. 1984. ‘Missionaries, Migrants, and the Manyika: The Invention of Ethnicity in Zimbabwe.’ African Studies seminar paper, Africa Studies Institute, University of the Witwatersrand. ———. 1997. ‘The Invention of Tradition in Colonial Africa.’ In Perspectives on Africa: A Reader in Culture, History, and Representation, ed. Roy Richard Grinker, Stephen C. Lubkemann, and Christopher B. Steiner. London: John Wiley & Sons, 450–462. Rao, M. S. A. 1968. ‘Political Elite and Caste Association: A Report of a Caste Conference.’ Economic and Political Weekly 3(20): 779–782. Rao, Vijayendra. 2001. ‘Celebrations as Social Investments: Festival Expenditures, Unit Price Variation and Social Status in Rural India.’ Journal of Development Studies 38(1): 71–97. Rao, Vijayendra and Radu Ban. 2007. ‘The Political Construction of Caste in South India.’ World Bank Working Paper. Rawski, Evelyn S. 1998. The Last Emperors: A Social History of Qing Imperial Institutions. Berkeley: University of California Press. Risley, Herbert Hope. 1892. The Tribes and Castes of Bengal. Calcutta: Bengal Secretariat Press. ———. 1903. Census of India, 1901. Vol. 1. India: Off. of the Supt. of Govt. Printing. Roniger, Luis. 1987. ‘Caciquismo and Coronelismo: Contextual Dimensions of Patron Brokerage in Mexico and Brazil.’ Latin American Research Review 22(2): 71–99. Rose, H. A. 1919. A Glossary of the Castes and Tribes of the Punjab and North-West Frontier Province. Lahore.

256

Bibliography

Roy, Tirthankar. 1994. ‘Foreign Trade and the Artisans in Colonial India: A Study of Leather.’ Indian Economic and Social History Review 31(4): 461–490. ———. 2005. ‘Agricultural Labor and Economic Transition in Colonial India: Lessons from Wage Data.’ Economic History Seminar, Vol. 30, Yale University. Rudolph, Lloyd I. and Susanne Hoeber Rudolph. 1960. ‘The Political Role of India’s Caste Associations.’ Pacific Affairs 33(1): 5–22. ———. 1967. The Modernity of Tradition. Chicago: University of Chicago Press. Russell, Kathy, Midge Wilson, and Ronald E. Hall. 1992. The Color Complex: The Politics of Skin Color among African Americans. New York: Anchor. Scott, James C. 1998. Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed. New Haven: Yale University Press. Sedgwick, Leonard John. 1922. Census of India, 1921. Volume VIII Part 1. Bombay: Government Central Press. Shah, Ghanshyam. 1975. Caste Association and Political Process in Gujarat: A Study of Gujarat Kshatriya Sabha. Bombay: Popular Prakashan. Shah, Ghanshyam. 2010. ‘Beyond Caste Voting: Modasa, Gujarat Revisited.’ Economic and Political Weekly: 54–61. Shah, Ghanshyam, Harsh Mander, Sukhadeo Thorat, Satish Deshpande, Amita Baviskar. 2006. Untouchability in Rural India. New Delhi and Thousand Oaks, CA: Sage Publications. Shamsul, Amri Baharuddin. 2001. ‘A History of an Identity, an Identity of a History: The Idea and Practice of “Malayness” in Malaysia Reconsidered.’ Journal of Southeast Asian Studies 32(3): 355–366. Sharma, Rama. 1995. Bhangi, Scavenger in Indian Society: Marginality, Identity, and Politicization of the Community. Delhi: MD Publications Pvt. Ltd. Shastri, Sandeep, Kondaveeti Chinnaya Suri, and Yogendra Yadav, eds. 2009. Electoral Politics in Indian States: Lok Sabha Elections in 2004 and Beyond. New York: Oxford University Press. Shastri, Sandeep and B. S. Padmavathi. 2009. ‘Karnataka: The Lotus Blooms... Nearly.’ Economic and Political Weekly 44(6): 42–45. Shayo, Moses. 2009. ‘A Model of Social Identity with an Application to Political Economy: Nation, Class, and Redistribution.’ American Political Science Review 103(2): 147–174. Shukla, Srawan. ‘Slap Those Who Try to Snatch Your Reservation Rights.’ DNA India. Available at https://www.dnaindia.com/india/report-slap-thosewho-try-to-snatch-your-reservation-rights-kalyan-singh-s-advise-to-backwardcastes-2629589. Accessed 11 September 2018. Singh, Khushwant. 2017. Captain Amarinder Singh: The People’s Maharaja: An Authorized Biography. Carlsbad: Hay House, Inc. Singh, Kumar Suresh. 1998. People of India. Delhi: Anthropological Survey of India. Singh, Om Prakash. 2015. ‘Emergence of Lower Caste Associations and Their Role in the Eradication of Untouchability in United Provinces.’ International Journal of Innovative Social Science & Humanities Research 2(2): 25–32.

Bibliography

257

Singh, Prerna. 2015. How Solidarity Works for Welfare: Subnationalism and Social Development in India. New York: Cambridge University Press. Singh, Prerna and Matthias vom Hau. 2016. ‘Ethnicity in Time: Politics, History, and the Relationship between Ethnic Diversity and Public Goods Provision.’ Comparative Political Studies 49(10): 1303–1340. Singhji, Virbhadra. 1994. The Rajputs of Saurashtra. Bombay: Popular Prakashan. Sinha, Mrinalini. 1995. Colonial Masculinity: The ‘Manly Englishman’ and the ‘Effeminate Bengali’ in the Late Nineteenth Century. Manchester: Manchester University Press. Skidmore, Thomas E. 1993. Black into White: Race and Nationality in Brazilian Thought. Durham: Duke University Press. Sobhanan, B. 1985. Temple Entry Movement and the Sivakasi Riots. Madurai: Raj. Srinivas, M. N. 1956. ‘A Note on Sanskritization and Westernization.’ The Far Eastern Quarterly 15(4 [August]): 481–496. Srinivas, Mysore Narasimhachar. 1960. The Social Structure of a Mysore Village. Indianapolis: Bobbs-Merrill. ———. 1966. Social Change in Modern India. Berkeley: University of California Press. ———. 1994. The Dominant Caste and Other Essays. New York: Oxford University Press. Stokes, Eric. 1978. The Peasant and the Raj. Studies in Agrarian Society and Peasant Rebellion in Colonial India. Cambridge: Cambridge University Press. Stokes, Susan C, Thad Dunning, Marcelo Nazareno, and Valeria Brusco. 2013. Brokers, Voters, and Clientelism: The Puzzle of Distributive Politics. New York: Cambridge University Press. Suri, K. C., P. Narasimha Rao, and V. Anji Reddy. 2009. ‘Andhra Pradesh: A Vote for Status Quo?’ Economic and Political Weekly 44(39): 108–113. Suryanarayan, Pavithra. 2016. ‘Hollowing out the State: Essays on Status Inequality.’ PhD Dissertation, Columbia University. Swamy, Rohini. 2017. ‘Siddaramaiah’s Poll Promise: 70 Per Cent Reservation for SC, OBC for Government, Education Sectors.’ India Today, 6 October. Available at https://www.indiatoday.in/india/story/siddaramaiah-poll-promise-70-percent-reservation-sc-obc-in-government-valmiki-jayanti-1059340-2017-10-06. Accessed 11 September 2018. Tajfel, Henri. 1974. ‘Social Identity and Intergroup Behaviour.’ Social Science Information 13(2): 65–93. Talbot, Ian. 2013. Khizr Tiwana, the Punjab Unionist Party and the Partition of India. London: Routledge. Telles, Edward E. 2002. ‘Racial Ambiguity among the Brazilian Population.’ Ethnic and Racial Studies 25(3): 415–441. ———. 2014. Race in Another America: The Significance of Skin Color in Brazil. Princeton: Princeton University Press. Templeman, Dennis. 1996. The Northern Nadars of Tamil Nadu: An Indian Caste in the Process of Change. Delhi: Oxford University Press.

258

Bibliography

Thachil, Tariq. 2014. Elite Parties, Poor Voters: How Social Services Win Votes in India. New York: Cambridge University Press. Thachil, Tariq and Emmanuel Teitelbaum. 2015. ‘Ethnic Parties and Public Spending New Theory and Evidence from the Indian States.’ Comparative Political Studies 41: 473–494. Thakur, Sankarshan. 2000. The Making of Laloo Yadav: The Unmaking of Bihar. New Delhi: HarperCollins Publishers. Thurston, Edgar. 1909. Castes and Tribes of Southern India. Madras: Government Press. Turner, A. C. 1933. Census of India, United Province of Agra and Oudh. Part-I Report. Lucknow: Government Press. Twenge, Jean M. 2001. ‘Changes in Women’s Assertiveness in Response to Status and Roles: A Cross-Temporal Meta-Analysis, 1931–1993.’ Journal of Personality and Social Psychology 81(1): 133. Varshney, Ashutosh. 2003. ‘Nationalism, Ethnic Conflict, and Rationality.’ Perspectives on Politics 1(01): 85–99. Veblen, Thorstein. 1973 [1899]. The Theory of the Leisure Class. Boston: Houghton Mifflin. Venkatachai, C. S. 1932. Census of India. Central Indian Agency. Vol. XX Part I, Report. Calcutta: Census of India. Verba, Sidney, Bashiruddin Ahmed, and Anil Bhatt. 1971. Caste, Race, and Politics: A Comparative Study of India and the United States. Beverly Hills: Sage Publications. Verba, Sidney, Kay Schlozman, Henry Brady. 1976. Voice and Equality. Cambridge: Harvard University Press. 1995. Verma, A. K. 2007. ‘Mayawati’s Sandwich Coalition.’ Economic and Political Weekly 42(22): 2039–2043. ———. 2012. ‘Why Did Mayawati Lose?’ Economic and Political Weekly 47(18): 17–19. Verniers, Gilles. 2016. ‘The Localization of Caste Politics in Uttar Pradesh after Mandal and Mandir: Reconfiguration of Identity Politics and Party-Elite Linkages.’ Diss. Paris, Institut d’Etudes Politiques. Washbrook, David. 1976. The Emergence of Provincial Politics. Cambridge: Cambridge University Press. Weber, Eugene. 1976. Peasants into Frenchmen. Palo Alto: Stanford University Press. Weber, Max. 1947. The Theory of Social and Economic Organization. Translated by A. M. Henderson and Talcott Parsons. New York: Oxford University Press. ———. 1958. From Max Weber: Essays in Sociology, ed. Hans Heinrich Gerth. New York: Oxford University Press. Wiener, Myron. 1967. Party Building in a New Nation: The Indian National Congress. Chicago: University of Chicago Press. ———. 2007. State Politics in India. Princeton: Princeton University Press. Weisskopf, Thomas E. 2004. ‘Impact of Reservation on Admissions to Higher Education in India.’ Economic and Political Weekly 36(39): 4339–4349.

Bibliography

259

Wilkinson, Steven I. 2006. Votes and Violence: Electoral Competition and Ethnic Riots in India. Cambridge and New York: Cambridge University Press. Wimmer, Andreas. 1997. ‘Who Owns the State? Understanding Ethnic Conflict in Post-colonial Societies.’ Nations and Nationalism 3(4): 631–666. ———. 2008. ‘The Making and Unmaking of Ethnic Boundaries: A Multilevel Process Theory 1.’ American Journal of Sociology 113(4): 970–1022. Witsoe, Jeffrey. 2013. Democracy Against Development: Lower-Caste Politics and Political Modernity in Postcolonial India. Chicago: University of Chicago Press. Wood, John R. 1987. ‘Reservations in Doubt: The Backlash Against Affirmative Action in Gujarat, India.’ Pacific Affairs 60(3): 408–430. Yeatts, M. W. M. 1932. Census of India, Madras. Vol. XIV, Part 1. Madras: Census of India. Yogananda, Paramhansa. 2003. Autobiography of a Yogi. New Delhi: Sterling Publishers Pvt. Ltd. Yong, Tan Tai. 2005. The Garrison State: Military, Government and Society in Colonial Punjab, 1849–1947. New Delhi: Sage Publications. Zacharias, Ajit and Vamsi Vakulabharanam. 2011. ‘Caste Stratification and Wealth Inequality in India.’ World Development 39(10): 1820–1833. Zelliot, E. 1970. ‘Learning the Use of Political Means: The Mahars of Maharashtra.’ In Caste in Indian Politics, ed. R. Kothari. Delhi: Orient Longman, 29–69. ———. 1970. ‘Mahar and Non-Brahman Movements in Maharashtra.’ The Indian Economic and Social History Review 7(3): 397–415.

Index adhiveshanas, 107 ‘Adi-Dravida’ identity, in Tamil Nadu, 25, 163, 187 administration of justice, 59 adult literacy, 142 Africa, 5, 32, 57, 129, 174 ethnic groups in, 6 income inequality in, 175 missionary education in, 25 African-Americans, racial identity of, 129–130, 175–176 Afro-Brazilian communities, 128 Agamudayars, 136 agricultural labourers, 208 agricultural tenants, rights of, 157 Akampatiyars, 64 Akbar, Emperor, 57, 104 Akhil Bhartiya Bhumihar Brahmin Samaj, 154 Akil Bhartiya Jat Mahasabha, 107 alliance formation, process of, 25 Ambedkar, B. R., 158 Ambedkarite movement, for dalits, 151 American slavery, 5 Anderson, Benedict, 31, 41 ‘anti-Brahmin’ movements, 78 anti-colonial movement, 48 anti-Congress Justice Party, 122 anti-poverty policies, 45 aristocratic lifestyle, 61

Aryan conquest of ancient India, 84 history of, 18 Arya Samaj, 102, 107, 188, 198, 218 Babhans, 123–125 backward castes, 81, 97 extremely backward castes, 161, 163, 179 ‘Most Backward Caste’ category, 160, 161 Backward Classes Commission, 160 backwardness, 164, 165 determinations of, 159 incentive to, 163 perceived level of, 160 politics of, 27, 81 Bahujan Samaj Party (BSP), 144, 163 Bainsla, Kirori Singh, 167 Banias, 100, 177 Bayly, Chris, 59, 74, 77 Bengal, 57, 65, 85, 117, 123, 185 Bengal Army, 80 British penetration of, 103 Kayasths in, 92, 103–104, 203 Mahishya caste of, 69, 176 zamindars of, 72 Bhangi caste of Delhi, 110, 135, 164 Bharatiya Janata Party (BJP), 146, 149–150, 182 conspiracy to end reservations for dalits and OBCs, 166

Index identity project, 150 issue of building a temple, 150 Bharsar Kshatriyas, 164 Bhumihar Ranvir Sena, 155 Bhumihars of Bihar, 25, 101, 119, 123–125, 136, 154 in armies of East Indian Company, 124 associated with agriculture, 123 Bhumihar Brahmans, 124 in electoral politics, 155 high-status identity, 124 in Hindu caste hierarchy, 124 as landowners, 143 literate group of, 124 social status of, 123 socioeconomic position of, 123 as zamindars, 123 Bihar, 4, 22, 47 Bhumihars of. See Bhumihars of Bihar Kayasths of. See Kayasths of Bihar upper caste elites, 45 Yadavs in, 97 zamindars of, 72 black rights movement, 127 black slavery, 126 black–white division, 129 Brahminical Hinduism, 105 Brahmins, 84, 91, 101, 123–125, 177 expenditure of money involved in, 63 Gaur Brahman community, 76 granting of inam lands to, 63 ‘non-Brahmin’ movement, 122 patronage to, 62 patronization of, 63 priests, 3, 37, 62 religious practice sanctioned by, 62 Shivaji’s protection of, 64 sikha topknot hairstyle, 121 social hierarchy of, 62 as source of political patronage, 64

261 status as legitimators-for-hire, 62 superiority of, 40 task of ruling and protecting, 64 Brazilian Institute of Geography and Statistics, 127 Brazil, racial identities in, 125–129 Afro-Brazilian communities, 128 black rights movement, 127 categories of, 126 democratization of political system and, 128 ethnic consciousness, development of, 127–128 politics of, 127 Portuguese colonial state, 128 racial problems, 127 rural elite rule, 129 social consensus, 126 whitening, ideology of, 128 British Admiralty Board, 60 British conquest, of India, 71 British India, 114, 185 education system in, 141 European bureaucrats, 116 land tenure system, 107 policies for rural Punjabis, 107 tax assessment, 107 tenancy laws, 115 British Indian Army, 69 Bhumihars in, 124 recruitment of Jats in, 106–107 British Raj, 151 British rule, in India, 132 Buddhism, 62 Ambedkarites embracing, 151 Chamars conversion to, 152 Burakumin minority, in Japan, 5, 16 bureaucratic dictatorship, 55 bureaucratic employment, 194 Carroll, Lucy, 43, 80, 90, 109 caste activism, 20, 43, 109, 135, 181 colonial. See colonial caste activism

262 common form of, 9 golden age of, 3 upsurge in, 4 caste activists, 4, 17–18, 20, 22, 25, 38, 43, 76–77, 82–83, 85, 89, 92, 134, 169 caste associations, 3, 92, 151 in colonial India, 73–76 formation of, 73, 135 membership of, 137–138, 143–145 in Mysore, 77 in politics, 77–79 south Indian, 76 caste-based parties, 21, 26, 144, 145 politics in India, 181 caste brotherhoods, 73 caste certificate, 162–163 caste consciousness, 3, 40, 56, 107 among Indian Christians, 102 caste, critique of, 133–135 caste discrimination, 134 caste divisions, importance of, 81 caste endogamy, prevalence of, 154 caste equations, formations of, 139, 177 caste hierarchy, 3, 101, 152 coding of, 189 ideology of, 152 caste identity, 5, 78, 82, 135, 176 after independence, 9 characteristic of, 16 development of, 132 diversity of, 5 dominant, 9 ‘ethnified’ view of, 4 in India, 5, 153, 173–181 political role of, 5–6 politicization of, 148 raking of, 4 rigidity of, 41 in South Asia, 181–182 caste, in the census, 84–89 casteism versus upper castes, 176–178

Index caste-level literacy rates, 22 caste mobilization, 9, 20–21, 56, 178–179 before 1990, 135–139 consequences of, 148 critique of caste, 133–135 rise of, 133–151 silent revolution, 139–149 upper castes and, 149–151 caste–modernity relationship, 170–173 caste organizations, formation of, 22, 73–74, 96, 107, 135, 162 caste panchayats, 73–74, 120 decline of, 52 caste petitioners, 132 caste politics development of, 131 in eighteenth-century, 65 features of, 5 Hindu beliefs on, 61 Hindu kingship and, 62–65 in India, 131 Muslim kingship and, 65–67 in pre-colonial India, 60–69 size of groups and, 68–69 and social climber’s guide to Hinduism, 61–62 caste ranking, 134 ideologies of, 35 pre-existing system of, 20 caste relationships, 208 caste reservation in colonial era, 81 policies for, 91 caste sabhas. See caste associations caste self-assertion, 3, 134 caste system, in India Hindu kingship, 62–65 Muslim kingship, 65–67 origins of, 65 in post-colonial world, 26–27 puzzle of, 4–8 Sanskritic, 18

Index social distinctions of, 39 as social good, 7 census petitions, in India additional variables in, 188–189 aggregate number by year, 87 alternate forms of political participation, 99–100 alternative hypotheses, 100–103 among highly educated, 194 caste classification schemes, 86 caste in, 84–89 castes petitioning for a name change, 87 of caste status, 189–190 coding of, 186–188 collection of data for, 186 illustrative examples of, 103–110 information on petitioners, 89–92 modelling of alternative explanations, 194–204 hypothesis testing, 192–194 model selection, 191–192 participation among the highly educated, 194 ranked petitioning, 204–218 number of petitions by literacy, 94 post-independence, 134 ranking in, 112–118 rate of petitioning by caste status, 100 literacy rate and years, 93 male literacy rate, 98 regression models of, 96–98 trends in, 92–96 Central Indian Agency, 95, 185–186 Chakravarti, Anand, 143 Chamars of UP, 3, 32, 69, 108, 135, 151, 163 conversion to Buddhism, 152 domination of reservations, 164 Chishti order, 66 Christianity, 62, 67, 102, 120, 198

263 Christian missionaries, 102, 120, 123 clan ranking, systems of, 5 cleanliness, Sanskritic ideals of, 180 coalition politics, 139, 148 coethnic voting, 147 colonial bureaucracy, 1, 15, 22–23, 71–72, 90, 99, 157, 167 colonial caste activism, 110 caste activists, 76–77 caste associations, 73–76 in politics, 77–79 colonial civil service, 104 colonial data, on Indian caste groups, 21–23 limitations of, 23–26 colonial identity politics declining patrimonialism and increasing participation, 71–73 nature of, 171 role of colonial state in, 79–81 social and economic conditions, 69–71 colonial India. See British India colonial land revenue settlements, 71 colonial petitioning, patterns of, 135 colonial rationality, aspects of, 41 colonial scholarship, 79 colonial tax policies, 71 colonial tenancy laws, 115 Commissions for the Backward Classes, 160 Communist Party of India (Marxist) (CPI[M]), 135, 149–150 corporations aggregate, 74 cosmopolitan elite identities, decline of, 17, 37 dalits mahadalit, 161 self-assertion, movement for, 152 untouchability, issue of, 155 Darzis of Maharashtra, 164

264 decision-making, political, 11, 30 Delhi Sultanate, 58, 65–66 depressed castes, 81, 85 descent-based social attributes, 11 Dirks, Nicholas, 7, 18, 40–41, 51, 58, 64, 99, 162 Discovery of India, The (Nehru), 133 disposable income, 58 divide and rule policies, 31 Dogras of Kashmir and Punjab, 218 ‘dominant’ castes, 9, 175, 181 Dravida Kshatriyas, 1 Dravida Munnetra Kazhagam (DMK), 137 Dravidian movement, in southern India, 151 Dube, Sheo Lal, 103 Dubois, W. E. B., 130 Dumont, Louis, 39 views on caste system, 39–40 social ranking, 40 East India Company, 57, 79, 104, 123 economic inequality, 19, 35, 175 economic modernization, effects of, 110 educational charities, 74 education system during colonial government, 141 higher education, 43 and identity politics, 48 in India, 21 and levels of identity mobilization, 43 missionary education in Africa, 25 primary, 21 school system, state-supported, 69 Western education, 105 electoral politics, rise of, 105, 119, 133, 149, 151, 155, 167 elite identity mobilization, 174

Index ethnic activism, 13, 42–43, 50, 77, 209 measuring of, 20–21 ethnic consciousness, development of, 127–128 ethnic constituency-building, 54 ethnic division, systems of, 5, 49 ethnic group relations, 15 ethnic identity systems, 7, 15, 111 expression of, 13 ethnic mobilization, 54, 183 theory of, 8 ethnic politics, 147, 183–184 consequences for India, 173–181 ethnic pride, ideology of, 17 ‘ethnic’ segregation, 33 ethnic superiority, discourse of, 218 ethnic voting, 147, 153 extremely backward castes, 161, 163, 179 French revolution, 60 fuzzy megalomanian, 39 Gandhian nationalism, 105 Gandhi, Indira, 47 Gandhi, Mahatma, 49, 134 Gaur Maha Sabha, 74, 76 goals of, 76 Gellner, Ernst, 44 Ruritanian nationalists, 38 group consolidation, process of, 25, 174 group-level literacy, 23, 46, 93, 100 group mobilization, theory of, 8, 28 Gujarati Kshatriyas, 152 Gujarat Kshatriya Sabha, 136 Gujjar community, 106, 161 Gujjar Aarakshan Sangharsh Samiti, 167 harijans, 190 Herfindahl index, 173 hierarchical identity politics, 79

Index hierarchical mobilization, 20, 52 hierarchical status-building, 68 high caste Hindus, habits of, 2–3 Hindu caste system, 164 Hindu cultural forms, 66 Hindu fundamentalism, 102 Hindu identity, 30, 149, 203 Hinduism, 188 Brahminical, 105 Brahmins religious practice sanctioned by, 62 role of, 62 practise of, 30, 40 as religion of plurality, 61 social climber’s guide to, 61–62 social reform within, 102 Hindu kingship caste politics, 62–65 money involved in feeding Brahmins, 63 Hindu–Muslim relations, 31 Hindu Protestantism, 102 Hindu reformism, 102 Hindu religious ceremonies, 107 Hindu sacred texts, devotional practices described in, 61 Hindutva, 30 human identities, ranking of, 184 identity activism as attempted mobilization, 29–32 barriers to political action, 42–43 concept of, 29 incentives to political action, 43–46 by ‘low ranked’ groups, 48 ‘objective’ criteria, 43 origins of, 39–48, 111 path to, 10 Posner’s concept of, 47 process of, 24 ranked versus unranked, 33–39 rise of, 6

265 role of self-interest in, 43 types of, 32 well-educated groups, 46–48 identity assertion, 33, 171 identity-based activism, 6, 91 identity mobilization causes of, 204 levels of, 43 identity politics, 6, 12, 43, 55 benefits of, 45 concretizations of, 31 of Muslim courts, 66 origins of, 12 pre-colonial. See pre-colonial identity politics reasons of, 11–15 relationship with education, 48 in twenty-first-century India, 173 identity ranking, among lower castes, 208 identity shift occurrence of, 12 outside of India, 12 identity-specific political activity, 6 incipient whole societies, 38 income certificates, falsification of, 159 income inequality, 175 India Human Development Survey, 162, 180 Indian caste groups colonial data on, 21–23 limitations of, 23–26 traditional, 37 Indian civilization, 7, 39, 169, 182 Indian Human Development Survey, 46, 93, 137–138, 142, 155 Indian Institutes of Management (IIM), 161 Indian Institutes of Technology (IIT), 161 Indian National Congress, 22, 47, 78, 99, 105, 116–117, 159, 182, 211 electoral defeat of, 159–160

266 Nadars support to, 122 post-independence politics, 79 working committee, 136 Indian National Election Study, 139 Indian nationalism, 47, 122, 149 Indian Survey of Social Attitudes, 154 individual-level identity, 41 Indra Sawhney vs. Union of India, 161 industrial ownership, 193 Industrial Revolution, 4 institutional development, 59 inter-group relations, 16, 33, 36 internal political hierarchies, construction of, 66 interpersonal cooperation, 173–174 Irava caste, of Kerala and Tamil Nadu, 74 Islam Chishti order, 66 conversion to, 67 ‘heterodox’ forms of, 37 Muslim kingship. See Muslim kingship Naqshbandi order, 66 in South Asia, 67 Jaffrelot, Christophe, 140 Jagat Seth family, of Bengal, 57, 103 jagirdar s, 66 jagir s, 59, 66 jajmani system, 98 Janata Dal (Secular) (JD[S]), 144, 146 Janata Party, 159 Jat Anglo-Sanskrit School, 107 Jatav Rajput, 151 Jat Gazette, 107 jati-based division, of Hindus, 102 jati–party linkages, 146 jatis, 68 Jat Mahasabha, 107, 145, 163 Jats of Haryana, 67, 74, 103, 105–108 in British Indian army, 106 caste consciousness, 107

Index economic position of, 106 egalitarianism within, 106 male literacy, 107 martial races, 106 in Mughal administrative hierarchy, 105 OBC status, 161 political and military networks, 105 programme of irrigation investment, 107 reservation quota for, 161 revolts against Mughal tax demands, 106 right to purchase land, 107 social position of, 107 zamindars, 106 Jharkhand Mukti Morcha (JMM), 145 Kaibarta community, 176 Kallars, 136 Kamma caste, 146 kangaroo courts, 2, 121 Karan caste, 32 Kayasths of Bihar, 21, 65, 80, 92, 103–105, 109–110, 203 ‘anti-Conference’ faction of, 105 bureaucratic tradition of, 104 caste identity of, 105 economic and educational position of, 104 mobilization of, 109 political behaviour of, 103 social status of, 103 socioeconomic rise of, 103 tax farmers and intermediaries, 104 tax farming rights, 104 writing and scribal services, 103 Khatriyas, Harijans, Adivasis, Muslims (KHAM), 139 Kolff, Dirk, 59 Kshatriya varna, 2, 64, 102, 120, 136, 203 Kumar, Nitish, 163, 166

Index Kumar, S., 182 Kurmi Janata Dal (United) (JD[U]), 144 landedness, effect of, 205, 212–213 leadership, of the sabhas, 76 Lingayat community, 146 literacy adult literacy, 142 by caste category in north India, 142 in China, 142 definitions of, 188–189 group-level, 23, 46, 93, 100 male literacy rate in India, 70, 92, 98 in pre-colonial India, 69 in south and west India, 135 Lok Janshakti Party (LJP), 144 Madigas of Andhra, 122–123 Madras Legislative Council, 78, 122 Madras Presidency, 78, 119 mahadalits, 146, 161, 164 Mal, Toder, 104 Mandal Commission, 159–160, 164, 166 Mandalization, strategy of, 17 Mandal Raj, 166 Manji, Jitan Ram, 146 Marahrastrian Marathas, 77 Maratha Kunbis, 95 Marathas of Maharashtra, 176, 218 Maravar caste, 64, 136 marriage funds, 74 marriage system. See weddings martial races, 71, 106, 218 Marwaris, of Rajasthan, 57 mass political awareness, development of, 69 Mayawati, 3, 14, 152, 166 members of Parliament (MPs), 136 caste category representation among, 140

267 military leadership, 60 Mishra, Jagannath, 47 missionary education, in Africa, 25 M. Nagaraj & Others vs. Union of India & Others (2006), 161 most backward caste, 160, 161 Movimento Negro. See black rights movement M. R. Balaji vs. Mysore (1963), 159 Mughal Empire, 57–58, 65, 105, 114 collapse of, 59, 106 tax system, 104 Mukkulathor community, 136 muscle castes, 178–180 Muslim kingship caste politics, 65–67 identity politics of Muslim courts, 66 patrimonial political system, 66 Persian-influence on, 66 in pre-colonial Indian politics, 65 ulema (religious scholars), 66 Mysore, 59, 81, 185, 204, 211 caste associations in, 77 reservation policies, 157 untouchable groups in, 187 Vokkaligas in, 78, 137 Nadar caste, 78, 119–123 Hardgrave’s views on, 122 as key player in Tamil politics, 121–122 mobilization of, 6 Nadar Mahajana Sangam, 3 ‘self-respect’ weddings, 3 wedding ceremonies, 121 on weddings with Brahmins, 3 Nadar Mahajana Sangam, 3, 78, 121, 135 Nadars Mahanjan Sabha, 81 Naicker, E. V. Ramaswamy, 121–122 Nair Service Society, 135 Naqshbandi order, 66

268 Narayan, Jayaprakash, 105 National Archives of India, 74, 90, 113 National Commission on the Backward Classes (NCBC), 164 national identities, development of, 6–7, 13, 31–32, 99, 174 nationalist movement, 96, 116–117, 157, 213 nawab of Oudh, 50, 123 Nehru, Jawaharlal, 133 neopatrimonial relationships, 53 New World countries, 125 ‘non-Brahmin’ movement, 122, 151, 203–204 non-identity activism, cost of, 29 Other Backward Classes (OBC), 21, 141, 179 effects of electoral reservations on, 158–159 literacy race among, 143 Ottoman empire, 17 outliers, social hierarchy of, 108–110 panchayats, 51, 107 functions of, 73 khap, 74 Nadar, 73 Pariyan sweepers, 67 parochial brotherhoods, 74 Paswan, Ram Vilas, 144, 164 Patel, Hardik, 4, 146, 166 Patel Navnirman Sena, 146 Patel reservation movement, 166 Patels of Gujarat, 4 OBC status, 161 reservation quota for, 161 patrimonial elites, 50, 53 behaviour of, 53 patrimonial political systems, 8, 19, 56, 66, 125, 182 patrimonial society concept of, 50, 52

Index and institutionalized political systems, 53 patronage networks, of village elites, 115, 156 patron–client relationships, 30, 51, 115, 179, 205 Pattali Makkal Katchi, 147 political competition, 6, 44, 109 political identity, development of, 45, 58 political mobilization barriers to, 42–43 instrumentalist view of, 43 political power, 54 distribution of, 50 political violence, 30, 155 Pradham Bhumihar Sabha, 76 pre-colonial identity politics among polities of South Asia, 57 among religious and scholarly elites, 58 background to, 56–60 caste politics in, 60–69 in Delhi Sultanate, 58 informal alliances among landholding and military elites, 57 level of social hierarchy within, 56 in Mughal India, 57, 58 relative economic and cultural sophistication of, 58 in Vijayanagar Empire, 57–58 pre-colonial India, politics of, 50 private warbands, leaders of, 60 public banquets, 3 public charity, 63 public employees, 115 public employment, 115 caste-based quotas in, 99, 133 public force employment, 208 public policy, formation of, 148 Punjab, 101, 105, 145, 164, 218 administration system of, 107

Index agrarian castes of, 80 Alienation of Land Act (1901), 80, 107 caste groups in, 13 Jats of, 80, 106 Legislative Council, 107 martial races of, 71 Punjabi Unionist Party, 108 quotas, caste-based, 133, 149 creation of sub-quotas, 161 criteria for granting, 167 promotion quotas, 161 racial ambiguity, 130 racial divisions, of ancient India, 40 racial identities in Brazil, 125–129 development of, 28 in Uganda, 5 in United States, 5, 17, 129–130 racial origins of caste, theory of, 83 racial superiority, 125 Rajputs of Rajasthan, 66, 91, 105, 218 merger with Kolis, 152–153 Ram, Chhotu, 107 ranked behaviours, localization of, 155 ranked identities, 9 caste politics and, 60–69 category and distinction of, 35–37 decline of, 49 internal and intra-group relations, 37–39 of ‘low ranked’ groups, 48 of martial races, 218 meaning of, 33–35 mobilization of horizontal, 35 vertical, 35 modernization process of, 49 origins of, 48–55 participation in, 53–55 patrimonialism and, 50–53

269 patron–client ties, 17–18 politics of, 127 reasons for, 15–18 Rudolphs’ key indices of, 35 versus unranked identities, 15, 33–39 Weber’s view of, 33 ranked mobilization, theory of, 8, 19–20, 23, 28, 36, 51, 54, 81, 112, 125, 131, 153, 173, 205, 210 ranked petitioning, modelling of, 204–218 alternative models and samples for, 211–218 for caste status of the caste, 205 Heckman selection model, 205 ‘Lower Unclean Shudra’ category, 210 rural patron–client relationships, 205 ‘Upper Unclean Shudra’ category, 210 ranking, in census data decline of, 151–156 density plot of, 119 illustrative examples of, 119–125 measuring institutions, 114–117 petition, types of, 112–114 results of, 117–118 social legacies of, 180–181 without Sanskrit, 125–130 ranking, meaning of, 33–35 Rashtriya Janata Dal (RJD), 143–144 Rashtriya Lok Dal, 147 Rashtriya Lok Samata Party, 147 Rashtriya Swayamsevak Sangh (RSS), 30, 150 conspiracy to end reservations for dalits and OBCs, 166 rationalized bureaucracies, spread of, 55, 60, 72, 116 Rawani (Chandravanshiya Kshatriya) Sabha, 75

270 regional political economy, in India, 148 religious communalism, 133 religious ideology of descent, 152 Republican Party of India (RPI), 137, 152 reservation policies, for lower caste groups, 134, 157–167 benefits of, 164 boundary-marking aspects of, 163 in central bureaucracy, 160 in colonial period, 158 consequences of, 162–167 Court’s restriction on, 167 creamy layer, 159 creation of sub-quotas, 161 criteria for granting reservation, 167 design of, 161 evolution of, 157–161 falsification of income certificates, 159 ‘general category’ applicants, 158 goal of, 162 in IITs and IIMs, 161 implementation of, 160 imposition of, 147 Indra Sawhney vs. Union of India, 161 Mandal Commission, 159–160, 164, 166 M. Nagaraj & Others vs. Union of India & Others (2006), 161 for ‘Most Backward Caste’ category, 160 for promotion quotas, 161 quotas, caste-based, 133, 149, 158 rise of, 27 RSS–BJP conspiracy to end, 166 Risley, H. H., 83–85, 88 Roosevelt, Theodore, 127 ruritanian boundary, 39

Index Sahay, Krishna Ballabh, 105 Samajwadi Party (SP), 144 Samyukta Vidhayak Dal (SVD), 138 Sangam caste, 3 self-respect weddings, 121 Sanskritic caste system, 18, 40, 151 Sanskritic hierarchy, 38, 113 Sanskritization, strategy of, 17–18, 24 Santhal tribal association, 145 Sargunar, Samuel, 1–2 Scheduled Castes (SC), 134, 166 literacy rate among, 142–143 reservations for, 158 Scheduled Tribes (ST), 134, 158, 166 school system, state-supported, 69 self-help, 75 Shah kings of Nepal, 63 Shanans of Tamil Nadu, 22, 119–123 caste hierarchy, 3 caste panchayats, 120 habits of high caste Hindus, 2 intermarriage, issue of, 121 palm liquor production, 2, 119 poverty of, 120 Sivakasi riots of 1899, 120 social aspirations of, 120 social status of, 1 tax collection and small trade, 120 tenant farmers, 119 Upper Cloth Controversy of 1857–1858, 120 wedding ceremonies, 121 Shudras, 1, 103–104, 203 clean Shudras, 106 unclean Shudras, 190 lower, 210 upper, 210 sikha topknot hairstyle, 121 Sikhism, 67, 105–106 silent revolution, in caste mobilization, 139–149 Singh, Amarinder, 145

Index Singh, V. P., 160 Sivakasi riots of 1899, 120 social activism, 42 social and political affiliations, 135 social assimilation, process of, 127 social desirability, attributes of, 155 social division, 27, 133 social hierarchy, 7, 16, 56, 62, 66–67, 155, 182 social identity, 44 articulation of, 29 nature of, 11 social mobility, 43 social modernization, effects of, 14, 133, 183, 189 social network, 43, 116, 173 social ordering, ideologies of, 82 social organizations, formation of, 30, 138 social ranking, 34, 83. See also ranking, in census data adoption of norms of, 218 Dumont’s views on, 40 of identities, 41 origins of, 49 in pre-colonial India, 68 Sanskritic concepts of, 20, 29 social relationship, 37, 156 social service, 75 activities, 30 network, 121 organization, 102, 138 social status and ethnicity, 34 legitimation of, 53 markers of, 51, 72 socioeconomic inequalities, between caste groups, 162 socioeconomic status (SES), 5, 8–9, 46, 48, 70, 94, 96, 103, 108, 110, 160, 161, 175–176, 179, 183, 194, 205 concept of, 193

271 Solanki, Madhav Singh, 139 Sonat Santhali Samaj, 145 Soundrapandian, W. P. A., 2, 78, 121–122 South Asia, implication of caste identities in, 181–182 Sri Narayana Dharma Paripalana Yogam, 74 status-based caste distinctions, decline of, 17 status hierarchy in India aspects of, 190 race to the bottom, 173 ‘status’ segregation, 33 supralocal politicized identities, 42 talukdar s of Benares, 50–51 tax farming, 104 tax-free inam land, grant of, 58, 63 Telugu Desam Party (TDP), 146 toddy-tapping communities, 122 Trinamool Congress, 177–178 Tuareg society, 5 ulema (religious scholars), 66 United Provinces (UP), 21, 89, 185 United States, racial identities in, 5, 17, 129–130 of African-American racial category, 129 black–white division, 129 Civil War violence, 130 Gens de Couleur, 129 unranked nature of, 129 unit of analysis, 185–186 unranked identities, 17 category and distinction of, 35–37 cultural differences within mobilized group, 37 internal and intra-group relations, 37–39 political relationships in coethnic, 36 non-coethnic, 36

272 ranked identities versus, 15, 33–39 unranked mobilization, 20, 36, 54, 112, 125, 131, 210 unranked petitioning, 117–118, 204–205, 208, 210 untouchability, practice of, 27, 99, 132, 155–156, 180 untouchables, 1, 84, 122, 187, 190 upper caste dynasties, 96 Upper Cloth Controversy of 1857–1858, 120 Vaisya varna, 85, 102 Vanniyakula Kshatriya Sangam, 135 Vanniyars, 135, 137, 147 Vargas, Getulio, 129 vertical mobilization, 35 Vijayanagar Empire, 58–59 village community, 115 village economies, 57 Vokkaliga caste, 146 vote banks, 3, 6, 45, 78, 143, 148, 165, 167, 178–179, 182–183 Weber, Max, 33, 39 formulation of ranking, 33 weddings cross-caste, 154 hypergamous, 37

Index inter-caste, 153–154 marriage funds, 74 Nadar–Brahmin weddings, 3 ‘self-respect’ weddings, 3 within-caste, 180 Weiner, Myron, 136 Western education, 105 Western Indian States Agency, 185 ‘white’ cultural practices, adoption of, 130 whitening, ideology of, 126–128 widow chastity, 61 widow remarriage, 2, 76 Witsoe, Jeffrey, 177 Yadav caste of Bihar, 3, 25, 39 caste polarization among, 144 rise of literacy among, 143 Yadav, Harmohan Singh, 145 Yadav, Laloo Prasad, 3, 45, 47, 153, 160, 163, 166, 177 Yadav Mahasabha, 138, 145, 152 Yadav, Mulayam Singh, 14, 141, 148 Yoruba elites, in Nigeria, 3 zamindar s, 59, 66, 72, 76, 104, 151, 171 Bhumihars, 123 Jats of Haryana, 106–107