INTRODUCTION

themes and approaches

This book is an attempt to account for and interpret the phenomenon
of caste in the Indian subcontinent. It deals primarily with the period
from the mid-eighteenth century to the present day, though the ®rst
two chapters explore the spread of castelike norms and values in the
age of the great sixteenth- and seventeenth-century Indian dynasts.
Of all the topics that have fascinated and divided scholars of south
Asia, caste is probably the most contentious. De®ned by many
specialists as a system of elaborately strati®ed social hierarchy that
distinguishes India from all other societies, caste has achieved much
the same signi®cance in social, political and academic debate as race in
the United States, class in Britain and faction in Italy. It has thus been
widely thought of as the paramount fact of life in the subcontinent,
and for some, as the very core or essence of south Asian civilisation.
There is of course an enormous body of academic writing on caste.
Studies by anthropologists and other social scientists provide a wealth
of closely observed ethnographic detail; many propose sophisticated
theoretical interpretations. So, given the notorious sensitivity of this
terrain, what is the case for an attempt to explore it from an historical
perspective?
In recent years historians have broken much new ground in the
study of political and economic change in the subcontinent, both
before and during the colonial period. But caste, which is best seen as
a meeting ground between everyday Indian life and thought and the
strategies of rulers and other arbiters of moral and social order, tends
to provoke more heated debate than almost anything else in the
specialist literature.
It has been common since the days of British rule for both
historians and anthropologists to refer to India as a `caste society', and
to treat the values of so-called caste Hindus as an all-pervading
presence in Indian life. Since the 1970s, however, there have been
commentators, both within India and abroad, who have accused these

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india
earlier specialists of massively overstating the importance of caste.
Some have gone so far as to question the very existence of an ancient
pan-Indian caste system, dismissing the idea of caste society as a
fabrication of colonial data-collectors and their of®ce-holding Indian
informants. This often perplexes newcomers to the ®eld when they
read about the many important Indians, including Mahatma Gandhi
and other past and present politicians and social reformers, for whom
caste was and is a real force in Indian life, and certainly much more
than an orientalist's `imaginings'.1
The subject of caste throws up other dif®culties as well. Those
unfamiliar with the ®eld often complain that even the best modern
historical studies make little effort to explain what they mean by caste,
despite the fact that these works so often refer in passing to such
mysterious phenomena as jati and varna, Backward and Forward
Castes, Brahmanism and non-Brahmanism, purity and pollution,
untouchability and outcasting, caste movements, casteism, `caste wars',
and much more.
Furthermore, in dealing with such major historical events as the
1857 Mutiny-Rebellion and the anti-colonial `freedom struggle', the
literature often identi®es groups of Indians by speci®c regional caste
titles, often without making clear whether this kind of group af®nity
truly overrides individual decision-making in times of crisis. Not
surprisingly, many readers wish to understand more fully what is
meant when they read that, in 1857, there were areas where `the Jats'
remained loyal, while `the Rajputs' and `the Gujars' rebelled; or that in
the 1920s, `the Patidars' of the Gujarat region joined Gandhi in acts of
resistance to British rule. They read too of how Gandhi and his
powerful opponent B. R. Ambedkar clashed in the 1930s over the
issue of how best to `uplift' India's millions of so-called untouchables.
Those to whom these terms and concepts are unfamiliar will rightly
want to know what an endowment of Jat, Patidar, Brahman or
`untouchable' caste identity actually entailed at these times. Further-
1 See Inden 1990. Such works as Dirks 1992a, 1992b, Cohn's `The census, and objecti®ca-

tion' in Cohn 1987, Appadurai 1992 and Breckenridge and van der Veer 1993 improve
greatly on studies which treat Western orientalist ideas in isolation, especially in suggesting
that colonial rule had the effect of turning such `constructions' into lived reality. (See also
Washbrook 1975.) The present volume shares these historical perspectives but argues that
while colonialism deserves much emphasis, so too do the many changes which were
underway well before the British conquest. Furthermore, much weight will also be given to
factors promoting the assertion of caste values in the years since Indian Independence in
1947.

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 introduction
more, did such af®nities change, or did they remain constant and
immutable when so much else was changing in India's culture and
material environment? Readers of both historical and anthropological
works have good reason to ask whether caste is to be seen in any sense
as an ancient or primordial essence of Indian life. Should such be the
case, how is this to be reconciled with what many historians now say
about the ¯uidity and dynamism of the pre-colonial state systems and
economies?
By the same token, non-specialists sometimes ®nd even the most
stimulating anthropological discussions of caste hard to reconcile with
accounts of nineteenth- and twentieth-century nationalist politics.
Those reading these historical treatments of the early nationalist era
®nd once again an emphasis on subtly overlapping af®liations of
religious community, class and regional or linguistic af®nity. They are
therefore taken aback when they then turn to works by those
anthropological theorists for whom Indian life and thought are
represented in an apparently very different way, featuring ®xed and
arbitrary schemes or structures of caste identity.

issues and premises

This study will argue that caste has been for many centuries a real and
active part of Indian life, and not just a self-serving orientalist ®ction.
Yet it will also seek to show that until well into the colonial period,
much of the subcontinent was still populated by people for whom the
formal distinctions of caste were of only limited importance as a
source of corporate and individual lifestyles. This would include much
of Bengal, the Punjab and southern India, as well as the far northwest
and the central Deccan plain.
Of course long before the age of European expansion, these and
other regional societies knew norms and conventions which named,
grouped and sometimes ranked people by order and function. There is
much debate about the nature of these social forms as they emerged in
India's medieval kingdoms, and this study will not attempt a detailed
reconstruction of these usages in the distant past.2 It is clear though
that some at least of these diverse and ¯uid ideas and practices

2 For one such reconstruction see Inden 1990: 213±62 on castes as political assemblages

or `subject-citisenries' within medieval Indian kingdoms.

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india
pre®gured what we now know as caste; indeed the names of many
individual castes as well as other elements of the terminology used in
contemporary caste life derive from these earlier regional schemes and
groupings. But this certainly does not mean that a single static system
of caste has dominated Indian life since ancient times, despite the fact
that a reverence for certain generalised caste ideals is extolled in
important scriptural writings. Nor did the emergence of the varying
castelike observances of the medieval realms translate directly into the
very different forms of so-called caste society which anthropologists
observe today.
These current manifestations of caste are now far more generalised
across the subcontinent than was the case in former times. The book's
aim is therefore to show that caste as we now recognise it has been
engendered, shaped and perpetuated by comparatively recent political
and social developments. The initial premise is that even in parts of the
so-called Hindu heartland of Gangetic upper India, the institutions
and beliefs which are now often described as the elements of
`traditional' caste were only just taking shape as recently as the early
eighteenth century ± that is, the period of rapid regional state-building
which accompanied the collapse of Mughal rule and the expansion of
Western power in the subcontinent.
Furthermore, from the early nineteenth century onwards, British
rule signi®cantly expanded and sharpened these norms and conven-
tions, building many manifestations of caste language and ideology
into its structures of authoritative government. It was Indians as much
as Britons who took the initiative in this process, even though the
impact of these moves was all the more compelling because it was
supported by the apparatus of an increasingly powerful colonial state,
and also by the effects of India's involvement in a Western-dominated
global market economy.
Ironically, the practices of representative government, which
became more deeply rooted in British-ruled India than in any other
part of the non-white colonial world, served further to enhance the
importance of caste af®nities in the political arena. Both for early
participants in electoral politics and to a signi®cant extent in the
period since Independence, caste has been an effective tool and
resource for the creation of common interests across the boundaries of
region, language, faith and economic status.
The argument here is not that Indians have somehow lacked the
4

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 introduction
capacity to develop `modern' political allegiances. On the contrary, it
has often been avowedly `modern' men and women who have taken
the lead here, discovering that by embracing caste principles, or by
imposing them on others, one may gain an extraordinarily ¯exible
resource in uncertain times. On the one hand, the assertions of caste
have made it possible to build broad allegiances which breach India's
many boundaries of region, faith, language and economic status. Yet,
at the same time, caste principles have often provided the means of
excluding, disempowering or subjugating others. This has proven to
be of great advantage in situations where other differentiations ± those
of class, for example ± may be far less effective than an assertion that a
group or individual is of alien or inferior caste. This may go far to
explain why consciousness of caste differentials has not altogether
given way in contemporary India to other markers of social difference
± for example, those of class, colour, language or occupation ± even
though in many situations considerations of caste may overlap or be
partially supplanted by any or all of these.
These manifestations of a more consciously castelike social order
became increasingly apparent in the turbulent environments of the
later Mughal realm, as well as those of the eighteenth-century post-
Mughal kingdoms. This explains the book's somewhat arbitrary
starting point of 1700. Of course the making or remaking of caste in
the forms that we see both in the colonial period and today was a
long-term process which cannot be pinned down to speci®c dates.
Even so, the book will attempt to show that the later eighteenth
century in particular was a period when India's regional societies
underwent profound and complex changes which tended to give more
Indians than hitherto a stake in this `traditional' caste order.
The reasons for this are extremely diverse, and no single book can
encompass all the ways in which caste and castelike identities were
shaped, debated, attacked and contested even in the relatively recent
past. There will, however, be an attempt to identify the most decisive
of these changes, and to write about them comprehensibly, avoiding
the use of abstruse technical jargon wherever possible. At the same
time, the book will seek to build on the best of the existing empirical
and theoretical literature. But, like the other New Cambridge His-
tories, this volume was commissioned as an interpretive synthesis
rather than a survey. So what can it achieve that has not already been
done by other specialists?
5

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india
First, it will seek to draw on interdisciplinary perspectives in an
attempt to bridge the gaps that often divide historians from social
scientists in the treatment of caste. Particular emphasis will be given to
the work of anthropologists: this discipline's special skills, and its
insights into the values of the small-scale community, can and should
be drawn on in the attempt to explore both changes and continuities in
the experience of caste. Secondly, the book will attempt to frame its
questions along rather different lines from those pursued in other
studies. In particular, using both historical and anthropological per-
spectives, it will ask why caste has so evidently mattered to so many
Indians, why it has aroused so much debate both within and outside
the subcontinent, and why its norms have been so widely acted on in
so many areas of economic, political and religious life, both in recent
times and in the more distant past.
The aim here is certainly not to defend caste. Nor is it the intention
to offer an all-embracing theory of caste, or at least not the kind of
theory proposed by those social scientists whose goal is to identify a
single principle such as purity, power or orientalism with which to
explain caste experience, regardless of regional or historical context.
This does not mean that the book will disregard the rich but
confusingly diverse theoretical literature, though it will give priority
to those approaches which treat caste as a dynamic and multidimen-
sional reality of Indian life, rather than an orientalist ®ction or
monolithic cultural code. The underlying premise then is that caste is a
topic that can and should be explored by those seeking to grasp the
complexities of both past and present life in the subcontinent. Indeed,
given the vast array of empirical and theoretical studies that have
contributed so much in recent years to the disciplines of Indian
anthropology, sociology and history, the time is certainly ripe for an
attempt at synthesis and interpretation.
What then of value judgements? Generations of well-meaning
observers have denounced caste as a source of dehumanising inequal-
ities and enfeebling social divisions. And it is true that in recent times
especially, caste has been for many Indians a system of oppression
comparable with the racist doctrines of apartheid, or the worst abuses
of European serfdom. But it is impossible to understand its full effect
on Indian life if we see caste only as a scheme of social and material
`disabilities'. On the other hand, nineteenth- and early twentieth-
century critiques of caste did have a powerful impact on colonial
6

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 introduction
policy, and on the ways in which Indians themselves have come both
to understand and to experience the phenomenon. Many modern
south Asians are ®ercely disparaging about caste, dismissing it as a
legacy of a backward and inegalitarian past. Yet this kind of internal
scrutiny was itself a major factor in the shaping of present-day caste
society. Its effects are most visible today in a number of far-reaching
provisions of India's post-Independence Constitution. Though now
widely contested, these empower the state to advance or `uplift' those
of its citizens who are de®ned as `backward' or collectively deprived
on the grounds of low-caste birth. Ironically, as will be seen in the
book's ®nal chapters, the implementation of these provisions has
played an important role in perpetuating rather than eliminating the
claims of caste for many Indians.
More broadly, this study seeks to show that both before and after
the end of British colonial rule, the perceptions and writings of both
Indian and foreign observers contributed directly to the shaping of
caste as a `system', both in the distant past and in more recent times. In
other words, caste was and is, to a very considerable extent, what
people think of it, and how they act on these perceptions. Far from
being a static re¯ection of received codes and values, caste has been a
dynamic force in Indian life and thought: it has been embodied in
what people do and say at any given moment about the conventions
and values which they de®ne as those of `caste society'.
This does not mean that the book will seek to reduce caste to the
realm of imagination or `discourse'.3 For centuries, south Asians have
found ways to make caste or castelike identities serve them in changing
and often threatening circumstances. As a means of coping with a
diverse and unpredictable social and physical environment, the titles,
symbols and lifestyles of caste have proved to be remarkably durable
and adaptable. So if caste is neither an orientalist ®ction nor a shameful
crime to be disguised or ignored in discussing India's history, it must
be a ®t subject for historical exploration. It is this which the volume
will attempt to provide.

3 The term `discourse' is being employed here in the crude though widely used sense of

purely cognitive or unconscious operations without connections to an active social or

political domain.

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india

de®nitions and principles

The English word caste has come to be widely used in south Asia,
even by speakers of vernacular languages, though many equivalent
terms for human orders or `communities' exist in the subcontinent's
regional languages. (On the origins of the term in English usage, see
Chapter 3.) Today, as in past centuries, most Indians who would
classify themselves as Hindus (and also many non-Hindus) are likely
to be at least broadly familiar with two distinct concepts of corporate
af®liation: the jati (birth group) and the varna (order, class or kind).4
The term caste is commonly used to refer to both of these. Both
may be used of non-Hindus; they sometimes designate distinctions of
species or kind amongst gods, animals and even inanimate objects and
substances.5 Nevertheless, both now and in past centuries, the term
jati has most often been used for the units of thousands or sometimes
millions of people with whom one may identify for such purposes as
marriage. There are thousands of titles associated with speci®c jatis in
different parts of the country. A few such titles ± most notably Rajput,
Chamar and Jat ± have come to be quite widely recognised; most will
be unfamiliar to people outside a limited geographical area.
In contrast to this profusion of jatis or birth groups, the concept of
varna involves a scheme with only four divisions. Thus what would
now be called Hindu society is conceived of as being divisible into
four very large units which transcend speci®c regional associations.
This scheme is propounded in a variety of widely revered Hindu
sacred scriptures (see below, pp. 13±14). It has been most commonly
understood as a ranked order of precedence, with the four varnas or
idealised human callings appearing in the following order:
. the varna of Brahmans, commonly identi®ed with those ful®lling the callings
of priests and spiritual preceptors;
. the varna of Kshatriyas, usually associated with rulers and warriors, but also
including seigneurial landed groups;

4 These usages include such regional vernacular terms as qaum, sampraday, samudi and

jati. Like other English terms made familiar through colonial administrative practice,
`community' is still widely employed in both English and the vernaculars. It is often a
reference to ethno-religious origin, as when newspapers refer euphemistically to Hindu±
Muslim riots as `clashes of two particular communities'. It is also a term for caste origin,
often with an implication that such a `community' shares an inherited moral mandate to
promote common interests by coercive means. (See Chapters 8 and 9 below.)
5 Sharma 1975; Marriott and Inden 1977.

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 introduction
. the varna of Vaishyas, often identi®ed with commercial livelihoods, though
associated with other producers and wealth-creators as well;
. the varna of Shudras or servile toilers.

So-called untouchables (and also the hill and forest populations who
are now commonly called `tribals') occupy an ambivalent place below,
outside or parallel to this varna scheme. The titles of these four
archetypes or orders, and the hierarchy of ranked callings and moral
endowments which characterise them, are de®ned in ancient religious
scriptures which became increasingly well known both before and
after the British conquest. It is important to note too, however, that
there are many widely revered sacred texts and doctrines which
devalue or condemn caste principles. (See Chapter 1, below.)
In the words of the anthropologist R. S. Khare, the concept of jati
refers to the experience of caste in the `concrete and factual' domain of
everyday social life, as opposed to the `ideal and symbolic . . .
archetypes' which are embodied in the concept of varna.6 Once caste
or castelike norms have come to be widely shared in a given region, a
reference to jati can therefore identify people in a very minute and
precise way; the designations of varna evoke vast and sweeping
generalities. While one would expect to ®nd at least a rough match
between the two, there has often been much dispute about the precise
order of merit among the various jati populations of a given region.
Furthermore, people of different doctrinal traditions and social cir-
cumstances have attached differing degrees of importance to these
schemes of caste. Indeed all these conceptual principles, and the ways
in which people have acted on them, have been far more diverse and
¯exible than has often been thought, both by academics and by
would-be reformers of caste.
For all this ¯uidity, it is still the case that certain basic ideas
subsuming both jati and varna were shared by at least some people in
the subcontinent well before the colonial period. The underlying
premise here, which is still widely known today, is that those who
would nowadays call themselves Hindus are born into ®xed social
units with speci®c names or titles. Such a unit is one's caste or
`community'.7 And, insofar as individuals and kin groups recognise
the claims of caste, these embody something broader than the notion
6Khare 1983: 85.
7This is the sense in which the term jati is generally used, though without necessarily
overriding its meaning as a reference to broader species-like groupings. (See note 4 above.)

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india
of a common kin or blood tie. Indeed caste is widely described by
anthropologists as a notion of attachment which bundles together a
given set of kin groups or descent units. Both in the past and for
many though not all Indians in more modern times, those born into a
given caste would normally expect to ®nd marriage partners within
these limits, and to regard those outside as of unlike kind, rank or
substance.
Furthermore, both in the past and today, those sharing a common
caste identity may subscribe to at least a notional tradition of common
descent, as well as a claim of common geographical origin, and a
particular occupational ideal. Neither now nor in past centuries would
an individual claiming Brahman parentage have been obliged to follow
a priestly or preceptoral livelihood. Nor would a man professing
princely Rajput descent automatically expect to wield a sword. Yet
such claims have often conveyed recognisable messages to other
Indians. In particular, those claiming Brahman or Rajput descent
would de®nitely not expect it to be thought that their ancestors were
humble labourers or providers of menial service, as would be the case
for an individual identi®ed by a low-caste jati designation such as
Paraiyan or Chamar. (On the important topic of women's caste status,
see below, especially Chapters 1 and 3.) Above all, the concept of caste
has come to imply both boundaries and collective or corporate rank.
In theory at least, civilised `caste Hindus' should regard it as wrong
and unnatural to share food or other intimate social contacts with
those who are radically unlike them in caste terms. In theory too, the
central characteristic of `caste society' has been for many centuries the
hierarchical ranking of castes or birth groups. The implication here is
that to be of high or low caste is a matter of innate quality or essence.
This is what is said in many scriptural codi®cations of caste ideals; in
real life, these principles have often been widely contested and
modi®ed. Nevertheless, even people who came to reject caste princi-
ples either recently or in the more distant past are at least likely to
have been familiar with these notions of corporate moral essences or
qualities, meaning that in `caste society', gradations of rank and
precedence are innate, universal and collective. The implication of this
would be that all who are born into so-called clean castes will rank as
high, pure or auspicious in relation to those of unclean or `untouch-
able' birth, regardless of wealth, achievement or other individual
circumstances.
10

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 introduction

theories and debates

The key problem for Indian social science has been to decide whether
caste should actually be seen in these terms, that is, as a coherent
system of thought and practice rather than an orientalist ®ction or a
miscellany of essentialising `discourses'.8 This in turn leads to the
question of what exactly comes into people's minds when they
differentiate between one another in caste terms. How distinctive are
these markers of difference? Are they truly unlike those of other
strati®ed social systems, where differences of status would seem to be
so much more readily reducible to straightforward material matters,
that is, to differentials of economic class, colour, education or religious
af®nity?
One might not think that caste differentials are so very distinctive,
requiring special explanations which treat the difference between high
and low castes as being fundamentally unlike the forms of strati®ca-
tion that distinguish the rich from the poor, or the dominant from the
weak and subordinated `subaltern'. After all, people of low-caste
origin are often signi®cantly poorer, less well educated, more inclined
towards unprestigious forms of `folk' religion, and even physically
darker-skinned than those claiming superior caste rank. None of
these, however, is invariably a feature of caste difference. Some other
basis of differentiation does seem to come into play, above all in cases
where there would appear to be no evident material basis for a claim of
caste superiority.
Both in the past and to a signi®cant extent today, the deprived
`untouchable' and the very poor individual of `clean' caste may appear
to be indistinguishable in economic and other material terms. Yet
there is still something real and important that divides them, not just
in the abstract, but in the bitter realities of everyday experience.
Similarly, Brahmans and other `clean' or high-caste groups and
individuals may often be found in deprived material circumstances.
Yet such people will not lose the intangible but widely recognised
quality that de®nes them as higher in caste terms than those who may
be richer, better educated and even more politically in¯uential than
they are, but who will still be seen as their inferiors by the standards of
`traditional' caste ideology. This, however, raises the question of

8 For an in¯uential treatment of this issue see Inden 1990.

11

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india
whether such perceptions have differed at the top and bottom of the
scale, either in the distant past or in more recent times. There is much
debate on these two related issues, that is, whether `modernity' has
modi®ed or undermined caste values, and whether those deemed to be
low-born in caste terms have accepted or contested the jati and varna
principles which de®ne them as unclean or otherwise inferior.9
Among social scientists, the most compelling modern interpreta-
tions of caste are those which have sought to resolve these problems
by combining ethnographic ®eldwork observations with an analysis of
sacred scriptures and other normative texts. The anthropologist Veena
Das has been a particularly eloquent and innovative champion of this
technique. She has thus rejected the approach of the empirical anthro-
pologists who studied caste in the 1950s and 1960s, notably F. G.
Bailey, for whom the learned abstractions of Hindu scripture were an
irrelevance to the life and thought of the ordinary `caste Hindu'.10
For Bailey and many of his contemporaries, `traditional' caste was
to be found in India's villages, and the villager's mental universe was
one of practical material realities. `Caste Hindus' worshipped Hindu
gods, but the logic of their social relations did not stem from the
values of those codes and scriptures which proclaimed the superiority
of `pure' Brahmans over worldly men of wealth and power. These
were merely a disguise or ex post facto rationalisation for the realities
of material advantage and disadvantage. Caste in Bailey's view was
therefore not a unique moral or religious system. It was merely a more
elaborate form of the social strati®cations to be found in many other
societies: the true basis of the distinction between those of low and
high caste was differential access to political and economic resources.
For Veena Das, as for other important commentators of the past
twenty years, texts do connect with this wider world of everyday
town and village life, and have done so for many centuries. It is
notable too that anthropologists who study `caste' norms no longer
con®ne themselves to non-literate village environments. T. N. Madan,

9 Moffatt 1979 argues for `cultural consensus' between those of low and high caste; see

Weber 1958. For opposing views, see Berreman 1967, 1971; Gough 1973; Mencher 1974;
Omvedt 1980; Juergensmeyer 1982; and works by historians of the Subaltern Studies school,
e.g. Chatterjee 1989. Other important contributions include Freeman 1986; Lorenzen 1987;
Randeria 1989; and DelieÁge 1988, 1989, 1992. And see O'Hanlon 1985. (The distinction
between those of low and high caste is now widely seen as being conceived in terms other
than or in addition to those of ritual purity and impurity: see note 18 below.)
10 See Das 1982, also Tambiah 1985; compare Bailey 1957, 1960.

12

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 introduction
R. S. Khare, Andre BeÂteille and Jonathan Parry are among the most
highly regarded of those who have taken the study of both `caste
Hindu' and apparently caste-free life and thought to environments of
complex urban modernity. This blend of textual and ethnographic
approaches has thus opened up a domain of norms and values which
would otherwise have remained hidden from view, but which are now
widely seen in anthropology as determinants of thought and action
both within and beyond the world of caste relations.11
India's earliest expressions of caste ideals can be found in the vast
body of sacred writings known as the Vedas. These texts are thought
to have been compiled between 1500 and 1000 BC, though it was in
relatively modern times that the Vedas, especially the great invocatory
sequence known as the Rg Veda, were extolled by in¯uential sage-
reformers as the de®ning core of Hindu faith and worship. One of the
most famous sections of the Rg Veda describes the primordial act of
blood sacri®ce from which the gods created the four human varnas.
The victim in this cosmic creation story is the thousand-eyed Purusa,
the ®rst created man. From the dismembered fragments of the
sacri®ced Purusa came each of the four varnas:
When they divided the Purusa, into how many parts did they arrange him? What
was his mouth? What his two arms? What are his thighs [loins] and feet called?
The brahmin was his mouth, his two arms were made the rajanya [kshatriya, king
and warrior], his two thighs [loins] the vaisya, from his feet the sudra [servile
class] was born.12

The sanctity of caste is extolled too in the Bhagavad Gita, the great
exposition of spiritual teaching which is contained within the ancient
Mahabharata epic. Without caste, says the Gita, there would be
corruption of humanity's most precious standards of domestic honour
and sexual propriety:
. . . when lawlessness prevails, . . . the women of the family become corrupted, and
when women are corrupted confusion of castes arises. And to hell does this

11 See e.g. Madan 1991, 1992, 1993; Khare 1984; Be  teille 1991a, 1991b, 1996; Parry 1980,
1981, 1985, 1994. These are certainly not simplistic portrayals of caste as an all-pervading
essence of Indian culture; see also Kolenda 1983, 1986; Fuller 1992. M. N. Srinivas (1965,
1969, 1989) gave the ®eld the important though now much modi®ed concept of Sanskritisa-
tion, an historical process of upward group social mobility through the embrace of high or
`Sanskritic' (as opposed to local or popular) forms of Hindu social and religious practice,
thus allowing caste society to be seen as mobile and ¯uid rather than static and in¯exible.
12 Rg Veda 2.2.1.1, quoted in Radhakrishnan and Moore 1957: 19; see O'Flaherty 1988:

27±8.

13

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india
confusion bring the family itself as well as those who have destroyed it . . . By the
misdeeds of those who destroy a family and create confusion of varnas [castes],
the immemorial laws of the race and the family are destroyed.13

The principles of caste as a universal law of life are further elaborated

in the Manavadharmasastra or Manusmrti, an encyclopaedic treatise
in verse on human conduct, morality and sacred obligations. This
work is most commonly known as the Laws or Institutes of the
mythical sage or lawgiver Manu; it was probably composed in about
the ®rst century AD. Here, as in the Bhagavad Gita, the focus is on
the concept of dharma. This key principle of `caste Hindu' thought is
usually understood as the code of duty, religious law and right human
conduct which de®nes the path to virtue and spiritual ful®lment for all
humankind. In the Institutes of Manu, the source of this dharma is the
will of the divine creator who gave each of the four human archetypes
or varnas a distinct moral quality, and a calling to follow. God, `the
lustrous one', `made separate innate activities' for the different orders
of humanity.14 All wellbeing and merit, indeed the preservation of the
entire created universe, depend upon this strati®ed ordering of castes.
The term `dharmic' is often applied to those ways of life which
conform to these principles of varna.
How then have modern social scientists sought to relate these
ancient scriptural ideals to the everyday life of the `caste Hindu'? The
best-known though most hotly contested attempt to construct a
textually informed interpretation of caste has been that of the French
sociologist Louis Dumont. Dumont proposed his formulation as
nothing less than a synopsis of Hindu civilisation itself, which he saw
as being animated by a unique and coherent structure of `core values'.
These he saw as conforming to the structuralist cognitive principles
elaborated most in¯uentially in the work of Claude LeÂvi-Strauss. The
view here is that social systems are underpinned by identi®able
systems of values and concepts, and that these in turn are organised
around universal cognitive and symbolic processes. In both LeÂvi-
Strauss's and Dumont's versions of structuralist analysis, these regu-

13 `The distress of Arjuna', Mahabharata 40±3, quoted in Radhakrishnan and Moore

1957: 105.
14 O'Flaherty 1991: 12. On the growth of social complexity in ancient India, which is

thought to have provided the context for this text's differentiation of ranked human classes,
see Thapar 1984 and 1992. On the idea of dharma as universalising `laws of life', see
O'Flaherty 1991: lxxvi±lxxvii, also pp. xxxv±xxxvi on dharma in the cultural synthesis
embodied in the Hindu epics (the Ramayana and Mahabharata) and the Laws of Manu.

14

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 introduction
larly recurring core patterns or operations of thought take the form of
paired binary oppositions.
Famously ± and controversially ± Dumont speci®ed the opposing
conceptual categories of purity and pollution as the ®rst in the
sequence of all-important complementary principles which, in his
theory, pervade the conscious or unconscious thought processes of all
Hindus. These are the archetypal or core principles which Dumont
held to be unique to caste, and which he claimed to have observed
both in scriptural formulations and in everday life and worship.15
Thus for Dumont, the facts of life for the Hindu villager are not the
straightforward matter of material differences on which Bailey in-
sisted. The difference between those of high and low caste is far from
being a disguised re¯ection of the ability to command material
resources. It is instead, says Dumont, the reference points of purity
and pollution which provide the important measurements of rank and
status for the `caste Hindu'. `Preoccupation with the pure and the
impure is constant in Hindu life', Dumont declares.16
Many anthropologists have found corroborating ethnographic evi-
dence for this. Referring to the Pandit Brahmans whom he studied in
their home region of Kashmir, T. N. Madan declares, `[Their] whole
way of life is pervaded by a sense of the pure, and consequently, by
the fear of impurity.'17 Other ethnographers too stress the importance
attached in everyday speech and action to these complementary
concepts of purity and de®lement, though many specialists would
now insist that the picture is misleadingly simplistic without the
addition of further conceptual polarities, most notably those of
auspiciousness and inauspiciousness.18

15 Both Dumont and his critics draw on such earlier theorists as Emile Senart (1894),

CeÂlestin Bougle (1927), Georges DumeÂzil (1957) and especially A. M. Hocart (1938) who
emphasised the ritualised or sacri®cial dimension of the caste system's occupational
specialisations. For interpretive overviews, see DelieÁge 1993 and Kolenda 1983; on Dumont's
intellectual pedigree, see Galey 1981 and Appadurai 1992; on Hocart's view of caste as a
ritualised redistribution system with principles comparable to that of the royal kingdom, see
Dirks 1987.
16 Dumont 1970: 44. For debate on Dumont's view of caste as a unique feature of

Hinduism which has nevertheless in¯uenced or `contaminated' many non-Hindus (1970:

202±12), see Ahmad 1973; Fuller 1975 and 1996.
17 Madan 1992: 109.
18 Such terms as the Hindi shuddha/ashuddha for purity/impurity, sutak for ritual

pollution caused by birth and death, and shubh/ashubh for auspiciousness/inauspiciousness

are widely used. On the debate about whether `caste society' embodies power-centred
relations of auspiciousness/inauspiciousness which exist independently of those of hierarchy

15

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india
Thus, many commentators, particularly Veena Das, Richard
Burghart and the `ethnosociologists' inspired by McKim Marriott,
have discerned a much wider array of cultural coordinates in `Hindu
thought' than those emphasised by Dumont.19 Indeed, even Dumont
contrasts the Hindu social being, that is, the purity-loving dharmic
`caste Hindu', with another ideal type whom he sees as a second
crucial pillar of the Hindu moral order. This is the ascetic renouncer ±
the so-called holy man or god-person ± who for Dumont provides a
complementary counterpart to the values of caste society. The ascetic
steps outside social norms to follow a path of transcendant spirituality;
this may lead to the achievement of ultimate release (moksha, libera-
tion) from the bonds of material existence.20
Other theorists, most notably J. C. Heesterman, also make much of
this seeming tension between the two domains of caste life and other-
worldly asceticism. As will be seen in Chapter 4, this is an important
consideration for both historians and anthropologists, because both in
the past and in more recent times, in¯uential doctrines of anti-
Brahmanical and even anti-caste `uplift' and `reform' have been
constructed around claims that Hinduism's highest spiritual principles
exalt `casteless' renunciation over and above the values of caste.21
Furthermore, much like Heesterman, both Veena Das and Burghart
are dissatis®ed with Dumont's simple binary oppositions, and insist
instead on the importance of additional patterns and conceptual
categories as reference points, both in scripture and in everyday
ethnographic reality. In particular, all three depart from Dumont in
emphasising the ideal of kingship and power as an independent variable
in Hindu life and thought. Burghart ®nds three rather than two spiritual

or ritual purity/pollution in Dumont's sense, see Carman and Marglin 1985; Raheja 1989;
Madan 1991; Parry 1991 and 1994: 135±8.
19 See Das 1982; Burghart 1978b, 1983a, 1983b. Ethnosociologists, whose techniques

derive from American cognitive anthropology, seek to interpret non-Western cultures using
indigenous sociological concepts, rather than those of Western social theorists. See Marriott
1968, 1989; Marriott and Inden 1977; CIS Special Issue 23, 1: 1989; Moffatt 1990; Khare
1990; compare Madan 1982.
20 On the associated doctrine of karma, the effects of past actions determining a being's

successive rebirths, with moksha as the goal of ultimate liberation from incarnate existence,
see Keyes and Daniel 1983; Fuller 1992: 245±52. On ascetic renunciation: Weber 1958; van
der Veer 1987; Burghart 1983a; Babb 1987.
21 Heesterman sees no true opposition between the Brahman-centred values of caste

society and those of the ascetic who seems to reject the norms of caste life. In classical Hindu
scripture Heesterman sees a basis for reconciliation of this `inner con¯ict': the Brahman
absorbs the ascetic's renunciatory principle and is therefore able to claim to be `in the world
yet not of it' (1985: 43).

16

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 introduction
ideals extolled as principles of supreme human and cosmic virtue and
harmony. These three elements, each of which is a path of virtue and a
source of cosmic harmony and righteousness in its own right, are
represented as follows: ®rst, the standards of dharmic caste life, as
embodied in the priestly functions of Brahmans; second, ascetic renun-
ciation; and third, the exercise of power by righteous kings, and by
those who share kingly qualities of initiative, assertion and command.
Yet for Burghart, this apparent diversity still has coherence as a
single system of thought and faith, with each of the three ideals
interpenetrating and referring back to the others. Kings glory in their
capacity to order the world through the exercise of power, yet still
shore up their claims to be fonts of righteousness by borrowing from
the characteristics of the other two ideals. In the same way, the
exemplars of the priestly and ascetic ideals absorb the key qualities
that de®ne the other two paths of righteousness.22
For Veena Das, Hindu thought involves an even more complex
array of patterned mental structures. She identi®es a whole series of
interconnected conceptual pairings: kings and their `unkingly' subjects
(commoners or Shudras); Brahmans and renouncers; renouncers and
unkingly subjects; Brahmans and kings. For each pairing, she then
identi®es a `latent' third ideal, arguing for example that the relationship
between the Brahman and the king has no meaning without the
implied presence of this pairing's shadowy silent or latent principle,
that of the renouncer. These webs of interconnecting core concepts,
which she describes as a scheme of `tripartite classi®cations with one
term latent', are for her the basis on which the diversities of Hindu
thought can be seen to achieve ultimate coherence.23
Marriott's `ethnosociology' offers yet another highly complex
formulation, treating the bonds of caste as a product of mutable, ever-
changing `coded substances', and offering a model of caste society in
which status rankings are expressed and experienced as a multidimen-
sional web of ordered ceremonial exchanges and transactions.24

22 Burghart's (1978b) sources include seventeenth- and eighteenth-century texts from the

Hindu-ruled Himalayan kingdom of Nepal. Compare Malamoud 1981.

23 Das's (1982) sources include Gujarati caste puranas (mythological `community'

histories) dating from about the ®fteenth century AD.

24 For Marriott's view of Hindu culture as `transactional and transformational', and of the

Hindu person as a ¯uid, unbounded, continually transacting `dividual' or divisible entity

composed of coded substances or essences transferred to others through marriage and other
interpersonal contacts, see Marriott 1976 and 1989.

17

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india
For all the richness and sophistication of these approaches with
their insistence on this greater diversity of conceptual reference points
and core principles, they have all remained bounded by a surprisingly
circumscribed notion of Hinduism. As a result, these formulations do
not seem to recognise the extent to which the `Hindu''s life and
thought have been intertwined with the subcontinent's other powerful
religious traditions, including those of devotional Islam, as well as
Christianity and Sikhism.25
Some social scientists have argued that caste norms are based on
ideals which are unique to Hinduism, and that `true' caste is to be
found only among those who profess the Hindu faith. Yet this
presupposes much ®rmer boundaries between ethno-religious `com-
munities' than was often the case in past centuries. Certainly, castelike
forms of rank and corporate allegiance have been very prominent in
the lives of most people who would nowadays be thought of as non-
Hindus. The dif®culty here is that so many studies of the supposedly
casteless minority faiths have played down those elements of religious
and social life which adherents of these faiths have shared with the
wider society. Yet if one looks at the millions who subscribe to India's
minority faiths ± Islam, Sikhism, Christianity, Jainism, and the osten-
sibly anti-caste neo-Buddhism to be discussed in Chapter 7 ± one
®nds both in the past and today a high level of sensitivity to the
nuances of caste, especially in matters of marriage and ritual pollution.
For adherents of the Sikh faith, the distinction between `peasant'
tillers and urban moneyed groups has long been re¯ected in an
awareness of which Sikhs are of `peasant' Jat caste origin, and which
are to be identi®ed by other jati names denoting a background in the
literate service occupations. In south India it is common to encounter
Christians who take pride in Brahman ancestry, and until recently
many north Indian Muslims identi®ed with the caste ideals of the
lordly Rajput. Furthermore, as James Laidlaw has shown, most of the
powerful north Indian traders who follow the austerely anti-Brah-
manical Jain faith are as insistent as their Hindu neighbours on the
importance of marrying within named Vaishya merchant jatis, while
simultaneously claiming descent from converts of princely Rajput
caste.26 Above all, for the members of virtually all the so-called
25 See S. Bayly 1989 on south Asian religious `syncretism'; on the historic construction of

Hinduism see Thapar 1989 and van der Veer 1988.

26 Laidlaw (1995: 88±119) shows that while Jains and Hindus regard their faiths as

18

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 introduction
conversion faiths, the ¯uid but highly potent phenomenon of the
pollution barrier is still a living force in everyday life, with those
deemed to be descendants of untouchable caste groups often being
denied social ties with others of their ostensibly casteless faiths.27

aequalis or hierarchicus ?
With this one important caveat then, the view of caste offered in this
volume is closest to the multidimensional models of those anthropol-
ogists who have emphasised the persistent appeal of renunciatory and
kingly ideals, and their interconnections with those of Brahmanical
purity. Veena Das and Richard Burghart are among the key in¯uences
here. At the same time, there are certain debates which are pursued in
the work of other theorists, notably Dumont and his critics, which
will be important to the approach being taken here. For Dumont, the
presence of the renunciatory element in Hinduism does not alter his
central premise, which is that it is fundamental to Hindu thought to
rank all beings, all substances and all aspects of worldly social
existence by this one overarching criterion of purity and pollution.
This for him is an inherently `religious' principle. It is on this basis,
and no other, that the hierarchical rankings of caste derive their
meaning. Brahmans therefore stand at the apex of the moral hierarchy
which we call caste because they are inherently purer than the people
of every other caste.28
The great problem which Dumont thereby claims to have solved is
how to explain the role of the so-called untouchables whose presence
in `caste Hindu' communities is for him a paramount fact of Hindu
life. Dumont saw the presence of these `unclean' toilers as a funda-
mental manifestation of Hindu values. In Dumont's theory, the
distinguishing characteristic of so-called untouchables is that they and
only they must perform the tasks of ritual cleansing and pollution-
removal which he sees as indispensable for the existence of Hindus as
social beings. These are tasks which keep the `untouchable' in a
permanently unclean state, but which thereby allow those of `clean'
separate and distinct, the same castes exist among both; Hindu±Jain intermarriage is
permissible so long as the partners are of matching caste. On Muslims and Rajput lordliness,
see Chapter 1, below.
27 See Chapter 1, note 3.
28 Important discussions of Dumont and the concept of hierarchy include Kolenda 1976;

Appadurai 1986, 1992; Galey 1989.

19

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india
caste to maintain a state of ritual purity in a world which continually
surrounds them with both tangible and intangible sources of de®le-
ment and pollution. Thus for Dumont, the distinctiveness of the
supreme, pure Brahman is the complementary counterpart to that of
the inherently unclean `untouchable'.
What then are the problems arising from these claims? Many of
Dumont's most vehement critics have accused him of disguising or
even legitimising the coercive side of caste relations, particularly the
concrete realities of disadvantage as experienced by those of low and
`unclean' caste. Among those who have attempted to reassert a
material or political economy dimension to caste have been those
Marxist commentators who since the 1970s have written sympatheti-
cally about caste-based militancy involving so-called `Dalits' (ex-
untouchables). Gail Omvedt in particular has thus rejected older
Marxist views of caste as mere `superstructure' or `false consciousness'
in a world where the true realities of political economy were to be seen
as those of class-based oppression. She argues instead that the material
effects of colonialism and modern capitalism served to make caste an
authentic force in Indian life, and that any truly `revolutionary' move-
ment in India must take note of the special oppression experienced by
those who have been thus coerced and disadvantaged through the
workings of this `rede®ned caste system'.29
Some of those who have argued in these terms have at least hinted at
the idea that Dumont's picture of caste may therefore be an accurate if
only partial re¯ection of Indian reality, not in the sense of being a
timeless expression of age-old `traditional' values, but as a product of
economic and social change in comparatively modern times. Yet there
are still those who have condemned Dumont's entire theory as an
exercise in demeaning orientalism, charging him with purveying a false
stereotype of Indians as the slaves of an all-powerful `religious' code,
hence dreaming, irrational, otherworldly and devoid of the capacity to
take `modern', secular or individual initiatives. A number of Dumont's
critics, notably Mattison Mines, have found evidence of strongly
individualist values in Hindu thought.30 Above all, many see Dumont
as recapitulating Western colonial views which supposedly exalted or
even fabricated the Brahman-centred perspectives that are central to
29 Mencher 1974, 1978; Berreman 1979; Omvedt 1978, 1982; Gough 1989. See also

Meillassoux 1973; Godelier 1986; Shah 1985: 14±15.

30 Mines 1992.

20

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 introduction
his analysis, and which are also fundamental for such commentators as
Madeleine Biardeau, for whom `orthodox Brahmanism' is nothing less
than the `permanent heart' of Hinduism.31
It true that Dumont's theory treats Indians (or Hindus) as heirs to a
system of values which are radically unlike those of the West. For
Dumont, the Hindu's hierarchical judgements of human worth are
made on a basis of collectively inherited moral qualities, rather than
personal endowments or attainments. He maintains that only in
modern, secular, rational Western society has there evolved a genuine
concept of the individual. For Dumont, no such principle is possible
in `traditional' Hindu thought. The Hindu is a caste being whose
social identity derives from collective rather than individual bonds and
claims; in Hinduism, he says, only the follower of the renouncer's
path can lead a life approximating to that of the Western individual.
Thus, in Dumont's famous phrase, Indians belong to a distinct human
order or cultural category, that of homo hierarchicus. This is a broad
category embracing other supposedly `traditional' non-Western civili-
sations. Yet for Dumont the Indian variant of this hierarchical being is
unique; in their deference to the overriding `religious' values de®ned
in his theory of caste, Indians (or Hindus) are so different from other
peoples that they are almost a distinct species of humankind.32
Even Dumont's defenders have generally accepted that there are
serious ¯aws in this sweeping portrayal of the Euro-American as
homo aequalis, in contrast to the Hindu Indian who is consigned to
the category of homo hierarchicus. But the key complaint here, as far
as the historian is concerned, is the charge that Dumont makes India a
land of static `oriental' spirituality rather than action and agency. More
speci®cally, these critics say that by insisting on Brahman-centred
caste values, Dumont and those who share his views make India
appear to lack any indigenous values which might have inspired the
construction of strong states and the achievement of effective political
action, either in the distant past or in resistance to colonial conquest.33
A number of commentators have therefore mounted strong objec-
tions to Dumont's hierarchical or purity-centred picture of caste
values, since this suggests to them a claim that only with colonial rule

31 Biardeau 1992: 15; those criticising Dumont on these grounds include Inden 1990; and

see Searle-Chatterjee and Sharma 1994.

32 Dumont 1970; see Madan et al. 1971.
33 See e.g. Dirks 1987; Raheja 1988a, 1988b; Quigley 1993.

21

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india
did the subcontinent ®nally acquire forceful but `derivative' models of
polity and statecraft.34 It is indeed true that for Dumont, the logic of
caste makes those who are collectively pure by birth and essence
superior to those who are endowed with mere worldly power.
Although kings and other embodiments of the Kshatriya's active,
lordly qualities may stand supreme in the material order, Dumont
maintains that the Brahman as priest-preceptor derives his status from
a source which is beyond and superior to the concerns of the mere
material plane.
Dumont's famous phrase for this is the assertion that in caste
society power is invariably `encompassed' by status.35 Thus he
proposes yet another key set of binary conceptual oppositions:
following on from his complementary pairings of purity/impurity,
Brahman/untouchable and renouncer/man-in-the-world is his insis-
tence on a radical disjunction in Hindu values between priesthood and
secular power. In other words, unlike other societies which possess
the cognitive capacity to recognise and exalt individual prowess and
achievement in the worldly sphere, Dumont argues that in the Hindu
social order, the worldly achiever and doer of active this-worldly
deeds performs a less exalted task than that of the `pure' and therefore
superior Brahman.
Nicholas Dirks and Gloria Goodwin Raheja in particular charge
Dumont with having overlooked much strong ethnographic and
textual evidence which would radically reduce the importance of
Brahmans and Brahmanical values in Hindu thought and social life.
These critics propose a view of caste relations emphasising action,
initiative and concepts of power deriving from indigenous cultural
concepts and categories, rather than the `religious' values of purity and
hierarchy proposed by Dumont. In this view it is rulers, and those
who exercise king-like power through the command of men, land and
other material resources, who stand at the apex of India's scheme of
moral order and values.36 Thus in the centuries before colonial rule,
Brahmans are to be seen as little more than technicians, performing

34 Chatterjee 1986.
35 Dumont 1970: 76±9.
36 Dirks's formulation (1987) which emphasises indigenous notions of royal gifting

proposes a view of power which, unlike Bailey's, does not depend on simple economic or
material differentiations; Raheja (1988b) insists too on the mechanism of the gift (dan) which
transfers inauspiciousness and thereby asserts and con®rms the power (or `ritual centrality',
rather than hierarchy) of dominant, king-like landed groups.

22

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 introduction
their specialist rituals as subordinate servants of kings and other men
of power.37 Even in present-day village life, it is the Kshatriya-like
qualities of landed elites, rather than Brahman-centred purity values,
which are seen by Raheja as structuring the relations between Hindus
of differing caste rank.38
These attempts to downplay or even dismiss the signi®cance of
Brahmans and Brahmanical caste values go against the grain of much
that is familiar both from the historical record and in contemporary
Indian life. The social scientists who will probably have the most
enduring impact on the ®eld are therefore those who have taken
Dumont's formulations seriously rather than dismissing them
altogether.39
At the same time, however, India's enormous complexity and
historical dynamism must make any quest for a single model or
formula of caste a deeply frustrating experience. Herein may lie the
great advantage of exploring these issues historically. The central
premise of this volume will be that through historical perspectives, it
may be possible to reach a view of caste which captures much of the
plurality and multiplicity of Indian life and thought. From an histor-
ical vantage point, it may indeed be possible to show that none of the
divergent theoretical interpretations outlined above can be refuted
absolutely. On the contrary, each of them may actually be correct for
some if not all Indians, at least for limited periods, and in at least some
areas of the subcontinent, either recently or in the more distant past.
Those arguments which de-emphasise the Brahman and identify the
princely warrior as the lodestone of the social order work best for the
pre-colonial military states which took shape in comparatively remote
frontier regions, well away from the great centres of high Hindu
culture and worship. Yet it cannot be convincingly inferred from this
that the role of the Brahman was merely a ®ction promulgated by
ancient law-givers, and then seized upon by British of®cials and
academic orientalists. On the other hand, far from illuminating the life
and thought of an unchanging or primordial `traditional' India,
Dumont's theory is probably best understood as a description
37 Quigley 1993.
38 Raheja 1988a.
39 Among the many strengths of these works is that they treat Dumont's category of

purity as an element of a dynamic world shaped by the force of the political and the
`religious', rather than an abstract or independent entity. Notably successful examples are
Parry 1974 and 1994; and Fuller 1979 and 1988.

23

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002
 caste, society and politics in india
capturing some of the rapid and complex changes which were
becoming increasingly active in Indian society just before and during
the colonial period.
Chapters 1 and 2 therefore explore the changes in political,
economic and religious life which helped to spread castelike ways of
life in many areas of the subcontinent in the period immediately
preceding the colonial conquest. Chapters 3 and 4 deal with thought
and `discourse': Chapter 3 assesses Western colonial perceptions of
caste, and Chapter 4 considers the understandings of caste that
animated debates among Indians themselves, concentrating particu-
larly on in¯uential social, religious and political commentators of the
later nineteenth and early twentieth centuries. Chapter 5 returns the
discussion to the domain of everyday life. Its aim is to explore the
changing experience of caste in the age of high colonialism. The
emphasis here will be on the increasing rigidities of the so-called
pollution barrier, by which is meant the formalisation of social barriers
separating those of superior caste from the lowest or most unclean and
inauspicious caste groups, in circumstances of growing con¯ict in both
town and countryside, in the period before the First World War.
Turning next to the power and resources of the twentieth-century
Indian state, Chapters 6 and 7 focus on the electoral arena and the
impact of modern political institutions on the experience of caste, ®rst
in the late colonial era, and then in the period since Independence. In
the two ®nal chapters, an attempt is made to ask in what ways and to
what extent contemporary Indians are still affected by the norms of
caste. Chapter 8 considers the practical realities of caste in the late
twentieth century, asking how far the experience of caste has truly
changed or been supplanted by other forms of solidarity and moral
obligation in everyday life. Chapter 9 explores the painful and
controversial phenomenon of so-called `caste wars' in present-day
India.
It goes without saying that the interpretations offered here will not
please all of the ®eld's contending experts. Nevertheless, on the
assumption that these are important and compelling matters to explore
and comment on, it is to this array of thorny issues that we now turn.

24

Downloaded from https://www.cambridge.org/core. Deakin University, Australia (Journals), on 29 Aug 2017 at 20:30:18, subject to the
Cambridge
Cambridge Core terms of use, available Histories Online © Cambridge University
at https://www.cambridge.org/core/terms. Press, 2008
https://doi.org/10.1017/CHOL9780521264341.002