CASTE

Origin, Function and

Dimensions of Change

SUVIRA JAISWAL

MANOHAR
1998
 First published 1998

© Suvira Jaiswal, 1998

All rights reserved. No part of this publication may be

reproduced or transmitted, in any form or by any means,
without prior permission of the author and the publisher

ISBN 81-7304-238-1

Published by
Ajay Kumar Jain for
Manohar Publishers & Distributors
2/6 Ansari Road, Daryaganj
New Delhi 110 002

Laser typeset by
P.K. Goel for
Aditya Prakashan
F 14/65, Model Town II
Delhi 110 009

Printed at
Rajkamal Electric Press #
B 35/9, G.T. Karnal Road Indl Area
Delhi 110 033
 To
the memory of
my parents
 Contents

PREFACE IX

1. INTRODUCTION 1
On defining the institution o f caste
Evidence o f Vedic texts on patriarchy
a n d social hierarchy
Caste character o f later Vedic varna
Proliferation o f jatis, the emergence o f
segmented identities in the varna structure
The class role o f caste and the elite

2. CASTE AND GENDER: HISTORIOGRAPHY 32

Some sociological theories on the
origin o f caste
Historical writings o f the
post-independence period
The brahmana, ksatriya a n d
hrahma-ksatra categories
The vaisya a n d the sudra
Roots o f untouchahility
Women, kinship a n d caste

3. STRATIFICATION IN RGVEDIC SOCIETY:

EVIDENCE AND PARADIGMS 132
Appendix: Mystifying the Aryans 188
viii CONTENTS

4. SOCIAL STRATIFICATION IN EARLY

BUDDHISM AND THE CHANGING
CONCEPT OF GRHAPATI/GMIAPATI

5. CASTE AND HINDUISM: THE CHANGING

PARADIGMS OF BRAHMANICAL INTEGRATION

BIBLIOGRAPHY

INDEX OF SANSKRIT, PAU AND TAMIL WORDS

INDEX OF ORIGINAL SOURCES AND AUTHORITIES

AUTHOR INDEX

SUBJECT' INDEX
 Preface

The pertinacity o f the institution o f caste continues to pose

serious problems in the restructuring of Indian society into
a more egalitarian system by eliminating the traditional practice
of discrimination on account of birth and gender. To understand
the reasons of its continued stranglehold, it is necessary to unfold
the contextual nature of its dynamics, which has regulated
social relations and shaped consciousness of its constituents.
The present work is an attempt in that direction. It has grown
out o f some o f the articles I had written over the years on
the them e o f social stratification and these have been
incorporated here with some additions. Chapter 1 and much of
Chapter 4 have been especially written for the monograph.
The work begins with a critique of the current theories of
caste system, w hich locate its essence in endogam y, and
argues that the present morphology of caste is the result of
transitions and transformations the institution had to undergo
in specific social contexts through centuries of its existence;
but its origins are embedded in the processes of patriarchy
a n d state fo rm a tio n from w h ic h it cannot be d e lin k e d .
Endogamy is seen not as a borrowing or survival of pre-Aryan
or tribal practices but as intrinsic to the process of stratification
and establishment of a patriarchal society. It is argued that
Vedic rituals provide enough evidence to justify this assertion.
I have tried to dilineate the origins o f the varna caste structure
from its early Vedic beginnings to its development into a pan-
Indian phenom enon and argued that the separation of the
brahm a and the ksalra elite categories and ascription of higher
status to the former, which feature forms the bedrock of Louis
X PREFACE

D um o u n t’s Homo W erarcbicus and is crucial for his ideological

interpretation of caste, has its roots in the ecology of the early
Vedic cattle-keepers. India has had an unbroken history from
the early Vedic times, and hence religious and cultural symbols
may suggest continuity but their real significance has undergone
fundamental changes.
The work examines the historical specificities which led to
the emergence of the varnas and their crystallization into castes.
The process of the construction o f segmented identities within
each varna category, which accommodated regional divergences
and allow ed sufficient flexibility to suit politico-economic
requirements, is also indicated. It is pointed out that in many
areas the role of the agent in spreading the varna system was
played not so much by the members of the brahmana caste as by
the dominant non-brahmana community or the ruling elite, which
benefited from the notions of hierarchy. This circumstance is, in
my opinion, largely responsible for the phenomenon of ‘unity
in diversity'. These developments had important implications at
the level of social relations as well as theory. Caste ideology has
undergone significant conceptual modifications over the centuries
without, however, abandoning its basic principles. Thus, for
example, the meaning of the term ‘sudra’ shows considerable
variation both in time and space.
Chapter 2 of the monograph is devoted to a re-evaluation of
historical and sociological writings o f the post-Independence
period on caste; and among other things, it seriously contests
certain ideological interpretations of the roots of untouchability.
Chapter 3 along with its appendix examines the evidence and
paradigms used for the study of stratification in Rgvedic society
and subjects the avant-garde concepts of ‘lineage society’ and
‘lineage mode of production’ to a critical analysis. It is pointed
out that the exploitation of biologically determined junior’ age
and sex groups by the elders of the same lineage is qualitatively
very different from the exploitation o f junior lineages by senior
lineages in a stratified society, where kinship is merely a metaphor
for class. 71-16 importance of this distinction is to be kept in mind
while reconstructing a picture of Rgvedic society. The chapter
also discusses the question of ethnicity of the Aryans.
 PREFACE xi

The fourth chapter deals with the social stratification in

early Buddhist sources and highlights the changing concept of
g rhapati/g ahapati category. It is shown that the early Vedic
grhapati was not an ordinary householder but a leader of the
extended kin-group which constituted a unit of production as
well as consum ption, and as such he, along w ith his wife
grhapatnt, was responsible for its ritual and material needs.
But in the early Buddhist sources the grhapati emerges as an
im portant entrepreneur w ho organizes the cultivation of
large tracts of la n d w ith slaves a n d hired la b o u r and is
an important tax payer. He is not an ordinary householder
or peasant and the category is not to be confused with that
of grhastha. There is a gradual degradation of the grhapati
class in the subsequent epoch, w h ich may be connected
with the processes leading to the emergence of segmented
identities within the broad category of the Vaisya varna. The
paradigm of varna/ja ti is well established by the early centuries
of the Christian era.
The final chapter argues that it was the brahmanical paradigm
of social integration formalized as caste which came to define
the H indu identity as it emerged in confrontation with the
‘o th e r’ in m edieval times. H ence, not only the religious
and the social are closely intertwined in Hinduism but caste
continues to be its constitutive element and its identification
mark. Failure to confront directly the social reality of multiple
fragm en ted identities and attem pts to u n ify them only
through religious-cultural symbolisms ultimately defeat the
lofty aims of social movements, as in the case of the Arya
Samaj, w hich despite its theoretical denial o f caste could
no succeed in establishing an egalitarian, universal Arya
brotherhood.
Substantial portion of Chapters 2 and 3 were published in the
In d ia n H istorical Review, Vols. VI, XVI and XX. Much of the
Chapter 5 had appeared in the Social Scientist, Vol. XIX, No. 12
(Dec. 1991). Permission to reprint these with additions, wherever
necessary, is duly acknowledged.
I hope the work would be of interest to students of history
and sociology as well as to general readers.
 PREFACE

I wish to thank Shri P.N. Sahay, Librarian, Indian Council of

Historical Research for prompt assistance in locating the books
and journals. The staff of the JNU Library and CHS DSA Library
also deserve my thanks. I am particularly thankful to my student
Ranjan A nand for preparing the b iblio g raphy and to my
publishers for bringing out the book expeditiously.

15 Ja n u a ry 1998 S u v ir a J a is w a i
 1
Introduction

C a s t e id e n t it ie s h a v e s u r f a c e d as
a powerful force in contemporary Indian politics and demands
for redressal of the inequalities and exploitation engendered by
this old institution have stimulated much fresh thinking in
academic circles on the question of the essence and dynamics of
caste. It is often assumed that a caste mentality is embedded in
‘the Indian psyche’. Hence even as traditional notions of its
integration with religion, morality and law are being increasingly
challenged (and even repudiated in modern circumstances), the
caste structure continues to survive as a salient feature of Indian
society.1This inference is further strengthened by the studies2 of
the Indian Diaspora where despite the absence of notions of
hierarchy and hereditary occupational specialization— features
intrinsic to the traditional caste system— the morphology of caste
is seen to prevail ow ing to the ‘separation’, ‘repulsion’ or
recognition of ‘difference’ of one caste from another. Castes retain
separate identities but are related to each other as constituent
units of a wider Hindu community. Thus, it is claimed,1empirical
studies have dem onstrated the in ad e q u a cie s o f earlier
Indological-sociological formulations of caste as a hierarchical
social system rooted in a religious principle that im puted
inherently pure or polluting status to social groups, legitimized
by the doctrine of karm a. Theories which looked upon the
institution as a system devised to ensure harmonious functioning
of a non-competitive, interdependent process of production,
which obviated economic, class-conflicts in a pre-capitalist social
formation are also found to be inadequate. The concept of caste
has been reformulated and Dipankar Gupta defines it as a ‘form
2 CASTE

of differentiation wherein the constituent units of system justify

endogamy on the basis of putative biological differences which
are semaphored by the ritualization of multiple social practices’.4
This definition according to him gives the 'essence' of a system
composed of discrete categories and not a continuous hierarchy.
Thus recent changes in the caste system have led sociologists to
revive what Dumont had termed the ‘atomistic’5view of caste, with
the rider that although discrete, castes do not exist in isolation but
form part of a system which gives them meaning and sustains their
existence. Legitimation and perpetuation of endogamy become the
basic characteristics, the ‘essence’ o f caste, in this perception.
Celestine Bougie’s6precise definition of caste epitomizing its major
features into three saliences, occupational specialization on a hereditary
basis, hierarchical status gradation, and ‘repulsion’, that is, separation
of each social group from the others through commensal and connubial
restrictions, was reduced by Louis Dumont7to one: ‘hierarchy’ deriving
from the opposition of the pure and the impure. For, according to
D um ont other features of caste were subsumed within this basic
principle. The trend is now to regard the feature of ‘repulsion’
or ‘difference’ or ‘division’ as the key concept, supposedly
maintained through ‘hyper-symbolism’, a cluster of characteristics
differentiating each caste from the other in social and ritual
matters but not occupation, the criterion laid d o w n in the
Indological works.8
However, the history of the caste system shows that belief in
‘putative biological differences’, which are expressed through a
ritualization of divergent social practices, has not acted as an
im pedim ent in transcending the rules of endogamy and the
formation of new castes when material conditions bring together
fam ilies of diverse caste origin but similar socio-economic
background. The formation o f the Kayastha caste in early
medieval times is a case in point, as literate professionals drawn
from different varnas/castes crystallized into a caste of scribes.
For a correct understanding of the dynamics of the caste system
we must pay attention not only to ‘repulsion’ or ‘fragmentation’
of castes but also to the processes of fusion which allow this
institution to continue and even strengthen itself as social,
political and economic circumstances change. For example, in
 INTRODUCTION 3

the overseas context, in Trinidad varna categories have come to

replace caste as the endogamous unit and status referrent.v No
doubt endogamy is basic to the morphology of caste but for its
origin and sustenance one has to look beyond hypersymbolic
manifestations and other ideational explanations which merely
beg the question by making it an attribute of the Indian mentality.
As we shall try to show, endogamy evolved gradually and
acquired rigidity with the growth of patriarchy in a varna-based
class society.
A major problem is that even Marxist historians who regard caste
as class on a primitive level of production1" have ignored the role of
patriarchy and subjugation of women in its ideology and rules of
endogamy. Endogamy is looked upon as a borrowing or survival of
pre-Aryan or non-Aryan tribal practices. D.D. Kosambi writes that
the fusion of tribal elements into society at large lies at the very
foundation of the caste system;" Irfan Habib concurs,u suggesting
that when tribes were absorbed they brought with them their
endogamous customs and this happened ‘only after the division of
labour had reached a particular level of development within the “general
society”'.MPerhaps this means that after stratification had emerged
in the form of varna divisions, assimilation of tribal groups led to the
institution of an endogamous caste structure. Kosambi is more specific
with regards to the time frame and he traces the origin of endogamy
to the incorporation of Aryans and pre-Aryan Harappans in one civil
society.
I have shown elsewhere" that the views of D.D. Kosambi on
the origins of the caste system are more in the nature of tentative
probings. They are contradictory and difficult to sustain in the
face of rigorous analysis. The fact that in later times incorporation
of tribal groups in to the ‘general society ’ m eant the
transformation of tribes into endogamous castes merely shows
that assimilation could take place only on the terms and patterns
of caste society. The assumption that the social structure as a
w hole became stratified into endogamous units owing to the
entry of tribal groups is perilously close to the racial explanation
of caste so vigorously propounded by Herbert Risley.11Kosambi's
perception of the unchanging nature of Indus valley civilization
has been rightly criticized'^1 by Morton Klass, whose own
4 CASTE

hypothesis regarding the origin of caste explains endogamy as

fossilization of a prehistoric South Asian aboriginal practice. The
work of Morton Klass has been acclaimed17 as a major study
providing a materialist explanation of the origin of caste and
hence deserves a detailed scrutiny.
Klass raises a pertinent point in his debate with sociologists
( 'the apostles of synchrony or even achrony ’), asking whether it is
possible to know what the caste system is without first asking how it
came to be. He emphasizes tine interdisciplinary nature of the problem.
His own search, however, leads him beyond all documented evidence
to four or more millennia ago, to ‘totemic’, ‘equalitarian’ clan groups.
Klass suggests that in prehistoric times South Asian regions were
inhabited by a galaxy of equalitarian endogamous social groups, each
internally characterized by full and undifferentiated membership and
constituting a single ‘marriage-circle1, in which prosperity and
misfortune were shared by all. Many of these endogamous societies
were composed of equalitarian exogamous segments or clans. Initially,
all were at the gathering and hunting stage, with no significant economic
specialization or exchange of goods and services. In the following
millennium food production began in the favourable ecological
zones, with the cultivation of rice or hard grain, and plants and
animals were domesticated. The new technology provided these
areas with ‘absolute surplus’, that is, the ability to produce
continually more than what was required for subsistence. This
developm ent placed those corporate groups, w ho were in
possession of cultivable land in an advantageous position vis-a-
vis those who did not have such land. The latter began to
exchange their labour and services for access to cultivable land
and crops. Since both the possessors and non-possessors were
structured in egalitarian clans, the network of exchange had a
corporate character. Labour and services were provided not on
individual but corporate basis with prosperity and misfortune
being shared equally by all members of the clan. Thus a hierarchy
of corporate groups developed ow ing to unequal access and
control of economic resources. This led to a transition from ‘clan
to caste’ or from ‘Bear to Barber’, borrowing the language of
Levi Strauss. Klass traces not only caste endogamy but even
ja jm a n i relations to prehistoric times. According to him the
 INTRODUCTION 5

system emerged not in any one specific region but over the entire
subcontinent almost simultaneously. For, while wheat and other hard
grains were being cultivated in the north-western part of the
subcontinent, rice was cultivated in the eastern and southern regions.
The caste system came into existence not in Bengal or the Malabar
Coast or the Indus Valley, but over the entire subcontinent' about the
same time.18
No doubt Klass is right in emphasizing the important role of
differential access to basic resources and economic inequality of
corporate groups in the emergence of the caste system; but his
assumption that the system originated in prehistoric times when
tribes living in ‘unfavourable’ zones migrated in search of fertile
land and crops and almost voluntarily entered into subordinate
or 'service' relations with the communities in possession of the
basic resources and practising new technology, is not only
conjectural but goes against the well-documented pattern of
agricultural expansion in ancient as well as more recent times.
The introduction of agriculture by neolithic-chalcolithic peoples
on the subcontinent also involved clearing the prim eval
vegetation and forests, which w ould destroy the habitats of
hunters and gatherers. The latter had to come to terms with the
new way of life by either adopting the new technology or
becoming marginalized as menial labourers or predators sticking
to their earlier way of life. The process accelerated with the advent
of iron technology. For instance, in the Mewar region of Rajasthan
a section of Bhils has adopted agriculture and become the peasant
caste of Gamits, which no longer has any social interaction with
the Bhils of the hills and forests. This is a classic example of the
disintegration or fission of a tribal community with one of its
segments transforming itself into an endogamous caste, not
because the segment had been earlier an endogamous marriage-
circle, but because it integrated with a stratified, fragmented caste
society which practised endogamy and as such provided it a
separate niche.19
Moreover, it may be pointed out that agricultural surplus
became available in several tribal regions of the subcontinent
without their developing a caste system, untill the introduction
of brahmanical culture and ideology. To give one example, recent
6 CASTE

studies suggest 2,1 that the Assam plains had a tribal peasantry
consisting of the Mikirs, Kukis, Khasis,and Kochch-Kacharis who
practised cultivation on a permanent basis in the pre-Ahom
period; but their assimilation in the caste system apparently took
place with the ‘creation of a dom inant class o f brahm in
landholders’ and penetration of brahmanic ideology.
Further, Klass’ hypothesis o f ‘clan to caste’ does not explain the
emergence of the priestly caste of brahmanas from the ‘clan’ stage.
They could hardly have constituted a separate marriage-circle’ in
prehistoric societies, unless we attribute to them sufficient advance
in the specialization of services and exchange of goods to maintain a
non-food producing ‘marriage-circle’. To sustain a specialized
endogamous ‘clan' or marriage-circle’ of priests, and not just one or
two priestly lineages of a tribal community, there has to be sufficient
absolute surplus’ available. On the other hand, if die formation of the
brahmana caste is explained as coining together of the exogamous
segments’ of prehistoric endogamous communities and their
crystallization into a sacred caste, this would imply fission and fusion
connected with occupational specialization and stratification. Such a
development would hardly be confined to priestcraft alone, particularly
as stratification was not just a matter of difference’ but the
consequence of growing control and manipulation of the sacred and
temporal domains by a few tribal lineages. Klass is certainly correct
in endorsing the view of Barth that die value system or ‘the cognitive
changes follow upon die social and interactional changes’/ 1but diis
does not explain the existence of a large caste of brahmanas from
the very beginning; the caste system can hardly antedate the caste of
brahm anas!22
In fact the evolution of the caste system cannot be delinked
from the emergence of patriarchy, class divisions, and state; and
as this did not happen at the same time all over the subcontinent,
one cannot speak of its simultaneous appearance in different
regions of the country. I have argued that regional variations in
the system may be partly explained by the time lag. The argument
which locates its essence in endogamy overlooks the fact that
occupational specialization and hierarchical gradation along with
the suppression of women as a class have played a no less crucial
role in the formation of caste society and in regulating its internal
 INTRODUCTION 7

intercommunal relationships.23If endogamy alone out of its three

defining characteristics24 endures in contemporary times, the
phenomenon needs to be explained with reference to changing
relations of production in a changed material milieu.25
Recently the theory of the Dravidian origin of caste has been
revived on two grounds: First, it is assumed26 that the scheme of
tinais mentioned in the Sangam literature represents a caste-like or
‘proto-caste’ stage as it speaks of five different types of environmental
zones peopled by divergent communities practising different modes
of production, depending on the nature of their basic resources. Thus,
the inhabitants of kum nji-tinaiv/ere Kanavar, Kunuvar and Vetar
subsisting on hunting and gathering, those of thepalai or desert-land
lived by plunder and catde raids, and were known as Kalavar, Eyinar
and Maravar. The Ayar and Idiayar occupied the m ullai-tinai or
pastoral tract and practised shifting agriculture and animal husbandry,
whereas plough cultivation was the technology used by the U{avar
and To}uvar in the m arutam -tinai or fertile wetland. The fifth type
o f tinaiw zs the neital or coastal tracts peopled by Paratavar, Valayar
and Mlnavar, who depended on fishing and salt extraction. Increasing
interaction and interdependence among these socio-cultural groups
with ‘fundamentally different systems of settlement and subsistence’
is seen as resulting in the caste system in which each community
remained encapsulated, retaining its separate identity even as it entered
into intercommunity relationships.
Nevertheless, unlike the caste system, the concept of tin a i 27relates
to physiographical divisions with the inhabitants o f each region
practising a different mode of production. They are not part of the
same civil society. Economic needs did require a certain amount
of interaction (exchange of goods), but there is no status gradation
or hierarchy of tinais or of communities although the polity and
subsistence pattern of different regions does show a great deal of
uneven development with some still at the segmentary lineage stage
and the region of m arutam supported a much more complex
o rg a n iza tio n v erging on state form ation. The n o tio n o f
tinaim ayakkam suggests that at the ideological level the process of
integrating diverse communities into one cultural whole had begun,
but the absence of any notion of hierarchy and the location of
communities in different eco-zones makes this mechanism of
8 CASTE

integration essentially different from the brahmanical caste system

which had already taken firm shape in the Gangetic region when
Sangam literature was being composed. The presence of brahmanas
following Vedic traditions in the tribal chiefdoms is well attested in the
Sangam classics; and it seems plausible that the northern caste ideology
found a fertile ground in the fluid social conditions of the south to
grow and evolve into specific forms in the subsequent centuries.
The other argument in favour of Dravidian origins rests on the
assumption that the idea of pollution from certain social groups derives
from Dravidian culture which visualized the sacred as something
malevolent and dangerous, a quality that could be transmitted to all
who mediated with it. Hence contact with aboriginal shamans and
musicians was considered harmful and ‘polluting’ and to be avoided
by relegating them to a low-caste status. Women too were considered
as polluted and polluting at times, for they were deemed as possessed
with ‘sacred power' manifested in their capacity of reproduction and
a different biological rhythm. Therefore, various taboos were imposed
on them to control their sexuality and keep them in patriarchal bondage.
George L. Hart28is the principal exponent of this view, which I have
criticized in detail elsewhere" and will not repeat except to add a
couple of points. Heesterman has shown that the diksita (a consecrated
person engaged in the performance of a Vedic sacrifice), too was
conceived as possessing ‘dangerous sacredness "'"at least in the initial
stages of the development of. Vedic sacrifice. The Srauta Sutras1'
lay down that one should not accept the food of a diksita, nor wear
his garments nor touch him, not because it w ould defile the
consecrated one but because the ‘evil ( papam ) ’ or ‘guilt’ of the
diksita would fall on the one who eats his food, touches him and so
on. Heesterman is of the view that the diksita is clearly regarded as
‘impure’12and Gonda11agrees. Apparently with the growth of a more
complex society and the emergence of the brahmanas as exclusive
ritual specialists and ideologues, such ideas were inverted to suit a
hierarchical social structure and what was earlier a magical,M
‘dangerous’ and ‘im pure’ state came to be seen as the purest state,
the prerogative of the highest social group. Although Vedic sacrifice
could be performed by all the three upper varnas, the Srauta Sutras
clearly state that the consecrated person even if he belorigs to the
rafanya or vaiSya varna should be proclaimed a brahmana as long
 INTRODUCTION 9

as he remains in this state.35 Incidentally, this also shows that

the idea of sacredness as potentially dangerous and harmful was
not confined to the Dravidian ethos; traces of it are found among
‘authentic Aryans'.
With regards to women too, recent studies v‘ have shown that female
sexuality posed both ‘physical and metaphysical’ problems to Vedic
priesthood, which invented devices to minimize and displace the role
of the female in the systematization of Vedic rituals in their 3rauta
form. The notion of the female biological rhythm being inherently
impure is the outcome of a patriarchal outlook; and ethnological studies
have shown that many Indian tribes do not observe the menstrual
taboo, until the tribe is integrated into caste society.17
It appears that patriarchy was not superadded to the varna/class
structure in Vedic times but was intrinsic to the process of
stratification; and caste endogamy was not just a borrowing or
continuation of aboriginal tribal practices, but came to be
constructed in an effort to regulate and reproduce patriarchy as
well as the hierarchy of social groups. The Rgvedic evidence is
clear on patriliny and patrilocality but patriarchy is not well-
established. Hence, we come across contradictory evidence
suggesting egalitarian gender relations’” on the one hand and
attempts to restrict and deride women on the other. We are told in the
tenth m andala of the Rgveda that in ancient times women used to
go to the communal sacrifice ( samhotrd) and community festival
( sam and).m IndranI, the wife of Indra, is described not only as the
mother o f brave sons but also as the maker of rta,*' interpreted
variously as ‘law’, 'tribal law', truth’ and ‘sacrifice’. Similarly,
goddess Sinlvali is described as vispatnt, the protector or mistress
of the i'is. '' The hymn praises her as the sister of gods ( svasd ),
indicating the important position of the sister as vispatnl. That the
idea of control possession or protection was implied in the term patni
is also evident in AV, II. 12.1 which speaks of ksetrasya-patni , the
mistress of the field. A verse '2found in the seventh m andala of the
Rgveda describes a young woman (y u v a ti) going to the powerful
sacrificial fire morning and night of her own accord (sva), to offer
oblations (b a m ) and ghee (gbrtaci) with devotion ( aramati ), in her
search for wealth ( vasuyub ). A maiden growing old in the parental
home is said to be entitled to a measured portion of wealth ( bhdgaD,43
 10 CASTE

and a prayer11 is offered for both sons and daughters by the side
o f the dampati decked widi ornaments of gold. Such passages
give the impression that unlike later times, daughters were not
unwelcome and enjoyed a degree of freedom and autonomy.
Nevertheless, a few passages1’ of the Rgveda seem to present a
brotherless girl in an unfavourable light. She is described as pursuing
men in search of a husband. According to Hanns-Peter Schmidt'1'1a
girl without a brother found it difficult to get married: her son would
be claimed by her father to continue his lineage and would thereby be
lost to her husband's lineage. The institutionalization of this custom is
evident in later Vedic texts which speak of the appointment of a
brotherless daughter asputrika so that her son (putrikaputra) may
inherit her father's property, carry forward his lineage, and perform
ritual services for his maternal grandfather. Schmidt notes the
prevalence of similar customs among the ancient Greeks, Romans
and Persians despite there being certain important differences in
marriage rules, such as exogamy in India, which meant that the girl
would have to be married to an outsider and the wealth of the sonless
father would go to a different family. This accounts for aversion for
girls without brothers in the Indian sources, according to Schmidt.
He suggests that the disapproval o f law-givers such as Gautama,
Manu and Yajriavalkya, of marriage with a putrika indicates their
su pport of the claims of the extended family over the wealth or estate
of the sonless father,which attitude goes back to Rgvedic times. If
this interpretation of the Rgvedic passages is accepted, it would
mean that a woman ’s paramount obligation was to reproduce and
continue the lineage of her husband and in exceptional circumstances
that of her father, but that her sexuality was already under patriarchal
control.
Earlier, Irawati Karve17had explained Rgvedic descriptions of a
brotherless maiden’s attempts at husband-hunting by visualizing an
archaic stage in which sisters and brothers were marriage mates, die
terms ‘brother’ and ‘sister’ being used in a classificatory sense
denoting young boys and girls of the same generation in the clan.
Thus brotherless women had to look for husbands outside their clan.
She supports her arguments with examples culled from Vedic and
Puranic literature and holds that the taboo on brother-sister marriage
was later extended to classificatory brothers, the endogamous
 INTRODUCTION 11

patriarchal clan becoming strictly exogamous. In Karve’s view

the institution of putrika-putra initially involved incestual union
between father and daughter in exceptional circumstances, and
did not imply a general stage of promiscuity. Similar customs
prevailed among the commoners o f ancient Iran too.'*8 T he
process of establishing incestuous taboos unravels before our
eyes in Vedic literature’, she remarks.49
The question of the evolution of clan exogamy among Vedic Aryans
is a highly controversial one. Scholars such as S.V. Karandikar,50
Irawati Karve51 and G.S. Ghurye52 have forcefully argued that
group or clan exogamy was unknown to the Vedic Aryans when
they came to India and that they borrowed it from the indigenous
people; for clan exogamy is found among almost all tribes of the
subcontinent. Ghurye points out that no other speakers of the
Indo-European languages are known to have practised group
exogamy and the Roman gens were not exogamous units. On
the other hand Benveniste53 and Jo hn Brough54 maintain that
clan exogamy prevailed in Rgvedic times and that the brahmanic
gotra exogamy was inherited from Indo-European forebears.
Whatever may be the case, the establishment of patrilocality and
clan exogamy would no doubt have resulted in the strengthening
of patriarchal trends55 and increased subordination of women.
The Rgveda provides clear evidence of the fact that a girl had to
move out of her maternal residence on marriage and live with
her husband’s people. In my opinion the description56 of the
members of the bridegroom’s party as ja n ya h (persons belonging
to a different jana), in the Atharvaveda may suggest that the
bridegroom belonged to a different ja na. Elsewhere, I have
referred to the interpretation o f the term ja n y a m itr a by
Heesterman as a relative by marriage.57 It is interesting that in
such passages the term used is j a n a and not 'vis’, w hich is
generally interpreted as ‘clan’. It may not be without significance
that the wife is called ja m . In the context of marriage ja n a seems
to have had special significance.
Along with the growth of patriarchy, later Vedic sources provide
clear evidence for the splitting of Vedic tribes and the emergence
of varna categories through processes o f fusion and fission, both
horizontal and vertical. The Rgveda occasionally speaks of rajas
12 CASTE

in the plural58 but this need not be construed as the structuring

of Vedic tribes into senior and junior lineages, as the category of
raja may have merely referred to the head or ‘chief’ of a clan or
lineage assembled for deliberations, w ithout im plying the
gradation of lineages. The use of rajanya, a diminutive of rajan,
in the Purusa-sukta hym n,5v how ever, clearly shows the
emergence of the kinsmen of the raja as a distinct social category,
standing above the common tribesmen, the vis or the vaisyas.
This is a conspicuously recognizable category in the Atharvaveda
and other later Vedic sources. Nevertheless, the extent to which
the principle of seniority in genealogical reckoning contributed
to this phenomenon is disputable. The Satapatha B rahm ana,
w hich belongs to the latest stratum of later Vedic literature and
makes little distinction between the rajanya and the ksatriya,
contains some passages wherein it is stated that the ksatra arose out
of the vi§*' and for the sake of victory the ksatra and the vis should
eat from the same vessel.61 These statements are interpreted as
reminiscent of earlier closer genealogical links of the rajanJksatriya
with the ins. Yet, it is significant that the term rajanya was gradually
replaced by 'ksatriya’. Whereas the former stressed kinship with the
ra ja n or ruler, the term ksatrd’1or ksatriya referred to'rulership’,
power’ or 'control' over the dominion, and suggests the emergence
of this varna category through the fusion of the rajanya lineages of
several tribes.
Brahmanical sacrificial ritual seems to have played a crucial
role in this development. R.S. Sharma points out that ‘the genealogical
superiority of the rajanya over the vis is not advanced as a ground
for claiming tributes’, and in his opinion ‘forcible methods adopted
by the descendants of the elected chiefs led to social distancing which
was frozen into genealogical ideology at a much later stage'/’1
By the close of the later Vedic period the varnas were being
clearly distinguished on the basis of the three fundamental
characteristics o f caste organization: hierarchy, inherited
occupational specialization and endogamy. Brahmana lineages
w ere system atized into g o tra- p rav ara categories in the
B audhaydna Srauta Sutra , and the Satapatha B rahm ana tells
us that a ksatriya is born of a ksatriya, a vaisya from a vaisya and
a sudra from a sudra.64 The concrete identity of brahmanas at the
 INTRODUCTION 13

time of Alexander’s invasion is attested in Greek accounts of a

country of Brachmans conquered by Alexander. K.P. Jayaswalr’5
identified this land with the ja n a p a d a named Brahmanaka
m entioned by Patarijali. However, inter-varna mariages of
hypergamous type were allowed to royalty and the three wives
of the later Vedic king at whose residences he offered oblations
to appropriate deities are supposed to have belonged to three
different varnas/1*’ Even in the time o f the Buddha, marriage
between brahmana and ksatriya families was permissible.
Jatis emerged within the varna system through fragmentation
as well as the incorporation of tribal communities w ithin a
structure which regulated hierarchy through marriage rules and
endogamy, and privileged heredity or birth in a particular lineage,
leading to the use of the term ja ti ’ for indicating membership in
a particular community/’7 Thus varnas were extended to provide
the institutional and ideological base for the growth of a wider
society. To quote Meillassoux, ‘contradictions inherent in the
co-existence and development of the rival classes’ and foreign
intrusions and conquests led to the rise of a new status system
of jatis deriving from the earlier varna system, and it was at once
more flexible and arbitrary’ and could be applied ‘to a fragmented
society by any fractions o f the dominant classes whatever their
origin’/ 18
The earliest use of the term jati in connection with a varna is
found in the Nirukla w which is regarded as pre-Panini and speaks
of a w om an of a sudra jati. It is not clear whether this is a
reference to the varna/caste, or to a tribe of the same name, for
a tribe of this name existed in the north-west down to the time of
Alexander's invasion. The transformation of this tribal name into a
generic varna category is elaborately delineated by R.S. Sharma in his
classic work Sudras in A ncient India. However, clear evidence of
the existence of distinct social groups subsumed within the Sudra
varna in a hierarchical relationship is to be seen in a sutra of Panini
which speaks7" of aniravasita sudras. Commenting on the sutra
Patanjali explains that the metal vessels used for eating purposes by
aniravasita sudras such as the Gandikas, Sakas and Yavanas may
be purified (by fire, etc.), but those used by the niravasita sudras
cannot be purified through any process,71 and Candalas and
14 CASTE

M rtapas are cited as exam ples o f the latter group. Thus

subdivisions o f the sudras and a hierarchy o f their relative
impurity is clearly enunciated in this text of the second century
bc. In the eyes of the orthodox brahmana all those who were
outside or at the margin of brahmanical society could be none
other than sudras, for the varna system was a universal concept
defining not only human but also the divine and vegetational
worlds.72Yet widely divergent social, economic and cultural levels
o f assimilating groups and material expediencies led to the
invention of the concepts of vratya and varnasamkara, that is,
formation of separate castes due to non-performance of the sacred
duty73 or because of the mixed marriages of original founder
couples; and these theoretical devices were highly successful in
e x te n d in g the varna system in to the j a t i system. These
explanations also led to a dilution or modification of the varna
concepts and we have shown74 how the notions of ‘vaisya’ and
‘s u d r a ’ acquired new m eanings in the changed material
conditions, w hich favoured a shift from the relative purity of
function to relative purity of birth implied in the transition from
varna to jati.1''
However, available evidence suggests that territorial and occupational
differences played a major role in the emergence of segmented identities,
particularly within the brahmana and the vaisya varnas, as castes
ranked within these categories generally emphasize their specialization
in a particular craft or tradition o f learning76 or their territorial77
affiliation. Adoption in the ksatriya varna was necessitated by the
arrival of new ruling and powerful groups, foreign as well as
indigenous, which required constant adjustments. The presence of
descendants of dominant ruling lineages of the later Vedic age in the
north-west is indicated by certain tribal coins of the second and first
centuries bc which carry the legend ‘of the country of rajanyas'
( rajanajanapadas ).7SThe ra/anyoswould have contributed mainly
to the formation of professional armies of the post-Vedic monarchies.
A distinction is made sometimes between the rajanyas and ksatriyas,
w hich fact suggests that the former were splinter groups of Vedic
tribes proud of their lineages, and the latter a more general category
bringing together all those who had access to power, status and war­
like professions. For example, the khattiya, brahmana and
 INTRODUCTION 15

rajanna (rajany a) are enumerated separately in a passage of

the M ajjhim a-Nikaya 79 which contrasts them with the low-born
Candalas and nesadas. The Kasika, commenting on a sutra of
Panini explains that although the Andhaka-Vrsnis were ksatriyas,
the term rajanya was applied only to those w ho belonged to
specially consecrated families.1*’ It cites the example of Dvaipya
and Haimayana, w ho were not called rajanyas e ven though they
were obviously ksatriyas.81
The ruling groups of the gana-rajyas o f the age of the Buddha
were very conscious of their ksatriya ancestry. Thus the remark
of the Buddha that although brahmanas could adopt a son who
had partial non-brahmana blood from either parent, the ksatriyas
would refuse to accept as their equal anyone w ho was not of
pure birth for seven generations on either side.82 Purity of descent
was of vital concern to the ruling class of the gana-rajyas , as
access to political power depended upon it. Thus, whereas
identity of interests and professional solidarity led to the
integration of ruling segments of tribes into a rajanya/ ksatriya
varna, excessive emphasis on the purity of lineage could also
contribute to the formation of subdivisions within its fold. The specific
lineage names of the ruling groups of the gana-rajyas are often
coupled with the term ja tii Sakya ja ti , Licchavijati, Koliya ja ti and
so on). These are ranked as ksatriya in general, but cannot be seen as
constituting ksatriya jatis'm the later sense of the term, as they were
not separate endogamous units but married freely among themselves
without notions of hierarchy or hypergamy within their own ranks.83
In this contextjatiseem s to have been used in a purely literal sense to
emphasize the hereditary status of the lineage. However Manu provides
clear evidence84of the existence of within the ksatriya varna and
the ja ti structure within the fold of the varna system seems to have
evolved a few centuries before the beginning of the Christian era.
Multiple factors contributed to its expansion over the subcontinent in
the course of time. A recent survey lists as many as 3539 communities
(which are euphemistically called sam udayasm oidm g the tenn jati)
among present-day Hindus.85
Discussing the proliferation of castes86 in early medieval times
R.S. Sharma attributes to the practice of making landgrants to
brahmanas in aboriginal areas, the function of a catayst: it led to the
16 CASTE

employment of the backward local people as landless labourers

and agriculturists in a land-based feudal system of production,
and to their classification as sudra castes by the brahmanical
law-givers. The agency of brahmanas in imposing varna-caste
categories is em ph asized , although it is recognized that
acculturation was not one-sided. Sharma explains the rise of
Tantricism as a consequence of brahmana-tribe interaction. But
in several areas the initiative seems to have come from a rising
class of tribal nobility when certain communities were exposed
to socio-economic and political stimuli from more advanced
regions and began to go through a process of state formation.
Ku!keK7 calls this ‘ksatriyaization’, which, he carefully explains,
does not mean merely imitating the ksatriya way of life but social
change ‘from above’ initiated in tribal areas ‘by the ksatriyas,
that is, zamindars, chiefs, or rajas, in order to strengthen their
legitimation as Hindu rajas in their own society and to broaden
the basis of their economic and political power’,1win other words,
by tribal nobility to further their own material interests. Kulke
points out that in many areas local languages were far more
important than Sanskrit in spreading this cultural pattern and
that the agents of Brahmanization’ were not always brahmanas.
Kulke’s remarks are based on his study of medieval Orissa, but
may apply to other regions going through the process of state
formation in early medieval times. A recent study*9 has shown
how the Rayalaseema area of south-western Andhra Pradesh,
which had been populated by hunters and gatherers, acquired a
strategic importance in the regional power struggle between the
Calukyas of Badami in north Karnataka and the Pallavas of Karici
in the eighth and ninth centuries and was thus exposed to the
influence of advanced cultures. There gradually unfolded
transition from tribe to state, leading to the rise o f tribal
chieftaincies; and the tribal communities were transformed into
a stratified society with the growth of a political nobility exercising
control over land. This new class patterned itself o n the
brahmanical model, which helped legitimize their separation from
the commonality of tribesmen and demanded their allegiance.
No less important was the fact that it gave them scope to operate
on a larger social space’. Thus, the varna/caste model of social
 INTRODUCTION 17

stratification was ado p te d by the rising elite w hich also

deliberately promoted the regional language, in this case Telugu,
as it gave a new sense of unity cutting across earlier tribal
affiliatio n s and e n a b le d them to face the ch alle n g e o f
neighbouring Calukya and Pallava powers. It is argued that at
least in the initial stages the presence of brahmanas in the region
was insufficient to establish strong bases of Sanskritic learning
and that the Jaina and Saiva sects played an important role in
the development of Telugu language and the spread of a pan-
Indian ‘great culture’. We may add that by this time Jainism too
had thoroughly imbibed the varna ideology*' and while one
occasionally comes across passages w hich challenge caste
hierarchy, caste identities are taken for granted.
Thus the adoption of the brahmanical social scheme in diverse
regional conditions resulted in different culture zones having
widely divergent structural categories of caste. Unity in this
diversity was provided by the caste ideology, which, as we have
shown, underwent certain conceptual modifications in a changed
material milieu without needing to alter its basic principles. Many
factors contributed to the multiplicity of castes. The splitting of
immigrant groups from parent bodies, crystallization of new
professional groups91into new units, differences in tribal configurations
of various regions, and historical specificities all generated fragmented
identities whose number and morphologies are astounding but reflect
centuries of historical development under the influence of brahmanical
culture. Caste ideology provided an integrating mechanism that did
not require uniformity or replication of the fourfold varna structure.92
This may be illustrated by a study of the VeHaja caste. I have
argued that owing to the changed perception of the functional role of
the vaiSya and Sudra varnas in the early centuries of the Christian
era, the land-based agricultural communities of the south were placed
in the Sudra varna and they stood next only to brahmanas. This
was due to the failure of the emergence of viable social groups to be
identified as ksatriyas and vaisyas in the specific material conditions.
By this time the brahmanical theory too had begun to measure the
relative purity and rank of a community in terms of the myth of
a n u lo m a and pra tilo m a marriages o f the original founder
couples. Hence, the Vejjajas who were landowners and tillers
18 CASTE

of the soil and held offices pertaining to land were ranked as

sudras91 but nevertheless became a status category at the regional
or sub-regional level. This caste has a very wide geographical
spread. It is pointed out* that land-based communities quite
distinct from the Vellaja have claimed VeUaja status and in the
course of time have gained acceptance and intermarried with
older VeUajar families. This is reminiscent of the formation of
the Rajput caste in the north. Thus caste ideology could even
transcend varna categories while retaining notions of hierarchy,
endogamy and functional similarities to organize varied structural
forms.
An exploitative system which has the capacity to enrol the best of
whatever origin in its own service is far more pernicious and long
lasting than one that is closed and static. Varna stratification received
strong justification from the doctrine of karma, which was an
invention of the elite95 in later Vedic times and tried to explain the
inequalities of the varna order from the view point of its own class
position. The class role of caste ideology has not been camouflaged
too well: it has been an instrument of power hierarchy; and field
studies have shown* that a correlation between caste and class may
be seen in a large measure even in modem times. This was even more
true in pre-colonial times, when caste rules had the sanction and
support of the ruling classes.97The religious cloaking of caste ideology
begins with the justification of varna divisions in later Vedic texts,
where it plays an important role in legitimizing the transition from
tribe to state. The mechanism was perfected and proved very helpful
to the indigenous ruling elite of later centuries — for cultural as well
as politico-economic reasons. Brahmanical law-givers enjoined upon
the ruler to ensure proper observance of caste duties, and inscriptional
evidence shows that brahmanized rulers took pride in championing
the varna-dharm a and actively intervened in regulating caste
hierarchy.,JKAfter all, it was a status system which could not be delinked
from the question of power.99 It has been argued""’ that political
histories of local level chiefs in pre-colonial times vitally transformed
caste structures of their region and the dichotomy of religion and
politics is ‘inappropriate at the level of ideological or cultural analysis
in Indian social thought’. To the extent that this view seeks to emend
earlier Indological interpretations of caste, which conceived it as a
 INTRODUCTION 19

static socio-cultural phenom enon ignoring the political and

econom ic underpinnings, it may be seen as an im portant
corrective. However, the recent attempt of Ronald In d e n 101 to
retheorize’"12 castes as ‘subject-citizenries’ constituting the
territorial associations of paura-janapadas , and as ‘complex
agents overlapping with one another and with royal courts’ in
an imperial formation which reconstituted or reproduced’
itself, or more accurately, its ruling class ‘as a self-ruling society’
through the annual holding of the ‘ceremonial bath’ or abhiseka
of the king,"’3 the entire polity being ‘constituted’ or ‘conducted
in the language of the major religious orders’,104 is an astonishing
exercise in obfuscation. In the name of restoring ‘agency’105 to
Indian (read ‘H indu’) people and to show that their ‘rationality’
was different from that of the Western mind it is argued that in
the Indian context socio-economic categories of analysis, the
problematic of inequitable structuring of castes and other social
groups, the systems o f production, distribution, etc., are
irrelevant. What makes sense is that which is located firmly in
the irrational: rituals, religious processions, discourses on the
divine will, and so on. But the post-modernist jargon ill-conceals
the neo-colonialist agenda which, as Shrimali has righdy pointed
o u t,"16 is to resurrect Jam es M ill’s ‘H in d u ’ and ‘M u s lim ’
periodization of Indian history by attributing the formation of
‘castes’ in something resembling their modern form to the collapse
of Hindu ‘kingship’.107Apparently Inden's argument is that during
Muslim rule castes lost their so-called ‘subject-citizenry’ character
and became autonomous and inward-looking, reduced to merely
'subjects’, as the new rulers followed a different religion; but no
attempt is made to show how dominant castes were made to
lose their political or socio-economic privileges or rights of
citizenship' with the changed religion of the ruler. By attributing
the dynamics of social change to the advent of Muslim rulers on
the scene, taking a static, unchanging view of the entire pre­
fourteenth century Indian society, conflating paura-janapadas
as caste associations10" and ignoring the question of menial castes
and tribes, Inden seeks to emphasize the Hindu-Muslim divide
rather than promote any logically consistent and scientific
understanding of India's past.
20 CASTE

However, India is a land of long continuities (which does not

mean that it has been static). The history of caste shows that not
only did it play an important role in the political economy of the
day, but it has been modifying, changing and adapting itself to
suit material conditions prevailing in the course of its long
existence. The crucial question is raised by Irfan Habib:10'-’ who
were its m ain beneficiaries? His answer is that no d ou bt
brahmanas w ould benefit to the extent that every priesthood
benefits from the success of the religious system it upholds. But
the main beneficiaries were the ruling classes as the system helped
generate larger revenues from the countryside by reducing the
cost of peasant subsistence. The repression of menial castes and
securing their structured dependence11" made agricultural labour
cheap and it also reduced the cost of artisanal products and
services; for artisan castes had a depressed status with restricted
mobility; and hereditary transmission of skills reduced the
expenses on training, etc., lowering the wage-cost as a whole.
Hence, Habib points out, while differentiation in terms of caste
was alien to Islamic law, Muslim rulers and writers never question
the inequities of the caste system. Their criticism of Hinduism
concentrates on its alleged polytheism and idol worship. ‘So
long as “petty production” remained the dominant form, the caste
system retained its inestimable value for every regime’, whether
the rulers were situated within the caste society or outside it.111
Modem industry has replaced petty production which favoured
craft-exclusiveness on a non-competitive basis. It has eroded the
notions of caste hierarchy and untouchability, and the taboos on
interdining at least in urban areas. But the prohibition of inter­
caste marriages is still observed widely, for it is not in conflict with
the capitalist mode of production. On the contrary, endogamy almost
invariably means arranged marriages on considerations of wealth,
power and status, and as such is well impregnated with the capitalist
value system. As a matter of fact in some aspects the strength of the
caste system has even increased in modem times; and caste ideology
may well be undergoing further modifications. The patron-client nexus
of the ja jm a n i type (service relationship) which existed in the
traditional village community, is being increasingly replaced by the
contractual, pecuniary and impersonal forms of exchange under the
 INTRODUCTION 21

influence of market forces, w ith the result that in times o f

adversity an individual has to depend all the more on the
members of his own caste for group support. Present day politics
too allow the elite of caste to exploit the caste-consciousness of
their castemen in order to compete with the elite of other castes
and communities for political power. Thus caste ideology has
gained strength both for political and economic reasons, in spite
of the fact that there are increasing differentiations of wealth
and status of individuals within each caste. Inter-caste relations
are now marked by cleavages and conflicts replacing the
trad itio n a l ethos that gave the c o m m u n itie s a sense o f
togetherness in the countryside. Thus caste ideology has become
a material force impeding the growth of class consciousness.

NOTES

1 Andre Beteille, ‘The Reproduction of Inequality: Occupation, Caste

and Family’, Contributions to Indian Sociology (N.S.), Vol. 25
(1991), pp. 3-28, reprinted in K.L. Sharma, ed., Social Inequality
in India (New Delhi, 1995), pp. 115-47.
2 David F. Pocock, Difference in East Africa: A Study of Caste and
Religion in Modern Indian Society’, South-Western Journal of
Anthropology, Vol. 13 (1957), pp. 289-300.
3 Dipankar Gupta, ‘Continuous Hierarchies and Discrete Castes’,
Economic and Political Weekly, Vol. 19, No. 46 (17 November
1984), reprinted in idem, Social Stratification (Delhi, 1991),
pp. 112f.
4 Ibid., p. 137.
5 Infra, Chap. 2, sec. i.
6 Ibid.
7 Louis Dumont, Homo Hierarchicus (Delhi, 1970).
8 Dipankar Gupta, op. cit., p. 131.
9 Steven Vertovec, ‘Hinduism in Trinidad: The Transformation of
Tradition in Trinidad’ in Gunther D. Sontheimer and Hermann
Kulke, eds., Hinduism Reconsidered (Delhi, 1991), p. 169.
10 D.D. Kosambi, The Culture and Civilization of Ancient India in
Historical Outline (London, 1965), p. 50; Irfan Habib, Interpreting
Indian History (Shillong, 1985), p. 17.
22 CASTE

11 D.D. Kosambi, Introduction to the Study of Indian History

(henceforth ISIH) (Mumbai, 1956), p. 25.
12 Irfan Habib, Essays in Indian History (Delhi, 1995), p. 165.
13 Irfan Habib, Interpreting Indian History, p. 17.
14 Infra, Chap. 2, sec. ii.
15 The disproportionate influence of Risley’s theory on historical and
sociological writings on caste (as well as on popular perceptions)
has been largely due to the misreading of the Arya/Dasa-Dasyu
and Aryan/Dravidian dichotomies in the Indological texts and
their concordance with the varnasamkara theory, which is the
DharmSastric view of the origin of lower ranking mixed’ castes.
However, the view that the Dasas constituted an earlier pre-Vedic
wave of the Aryans seems now to be gaining ground in academic
circles. See ibid., sec. i. For the Aryan/Dravidian issue, see Suvira
Jaiswal, 'Studies in the Social Structure of the Early Tamils’, in
R.S. Sharma and V. Jha, eds., Indian Society. Historical Probings
(In Memory of D.D. Kosambi) (Delhi, 1964), pp. 124-55.
16 Caste: The Emergence of the South Asian Social System
(Philadelphia, 1980), pp. 57-8.
17 Satya P. Sharma, 'A Materialist Thesis on the Origin and Continuity
of the Caste System in South Asia’, The Eastern Anthropologist,
Vol. 36, No. 1 (January-March 1983), pp. 55-7; Dipankar Gupta,
op. cit., p. I l l ; Gail Omvedt,Dalits and the Democratic Revolution
(New Delhi, 1994), p. 32.
18 Morton Klass, op. cit., p. 175.
19 Distinguishing between tribe and caste, Max Weber wrote that a
tribe, unless it has become a ‘guest’ or 'pariah people', usually has
a fixed territory and practises exogamy at the level of the totem,
the village, and the sibs. Endogamy exists only under certain
conditions. Caste however does not have a fixed territory and
endogamy forms the essential basis of a caste [The Religion of
India, translated and edited by Hans H. Gerth and Don Martindale
(New York, 1968), pp. 30-31
20 N. Lahiri, Pre-Ahom Assam: Studies in the Inscriptions of Assam
Between the Fifth and the Thirteenth Centuries AD (Delhi, 1991),
Chap. 4; also see idem, ‘Landholding and Peasantry in the
Brahmaputra Valley, c. 5th-13th Centuries AD’, Journal of the
Economic and Social History of the Orient, Vol. 33 (1990),
pp. 157-68.
21 M. Klass, op. cit., p. 185.
22 The point is not rebutted by the example of ancient Sinhalese
society which had caste without the presence of brahmanas, and
 INTRODUCTION 23

few untouchables, except for the Rodis who were itinerant beggars
and very small in number. It is held that the Sinhalese caste system
is ‘historically and conceptually related to the Indian' [Richard F.
Gombrich, Buddhist Precept and Practice: Traditional Buddhism
in the Rural Highlands of Ceylon (Delhi, revd. edn., 1991),
p. 3451. The structuring of Sinhalese tribes into a caste system took
place under Buddhist influence. Buddhism had no need for the
ritual role of the brahmanas and disclaimed the relevance of caste
for the pursuit of salvation,even as it accepted caste as a fact of life.
The Pali canonical literature shows upper class prejudices against
low occupations and this is reflected in the Sinhalese system which
is based on a hierarchy of occupations but not notions of ritual
impurity of social groups.
23 E.R. Leach, ed., Aspects o f Caste in South India, Ceylon and
North-West Pakistan (Cambridge, 1971), p. 10.
24 A.L. Kroeber in David L. Sill, ed., Encyclopaedia of Social Sciences
(London, 1930), p. 254.
25 Suvira Jaiswal, ‘Studies in Early Indian Social History: Trends and
Possibilities’, IHRVI (July 1979-January 1980), pp. 5-6.
26 Gail Omvedt, op. cit., pp. 34f.
27 For the tin ai concept see K. Sivathamby, 'Early South Indian
Society and Economy: The Tinai Concept', Social Scientist,
No. 29 (1974), pp. 20-37; Rajan Gurukkal, ‘Forms of Production
and Forces of Change in Ancient Tamil Society’, Studies in History,
V, 2, n.s. (1989), pp. 164-5. I had a useful discussion with
C.N. Subramaniam (of Ekalavya, Bhopal) on the subject.
28 George L. Hart III, The Poems of the Ancient Tamil: Their Milieu
and Their Sanskrit Counterparts (Berkeley, 1975), pp. 119f.
29 Suvira Jaiswal, op. cit., pp. 45-8.
30 J.C. Heesterman, ‘Vratya and Sacrifice’, Indo-Iranian Journal,
Vol. VI (1962-3), pp. 11-15.
31 Apastamba Srauta-sutra, X.13: 1-2; 15.5 quoted in ibid., p. 12;
R.N. Dandekar, Srautakosa, Vol. II, Pt I (Pune, 1973), p. 81.
32 Heesterman, op. cit., p. 12.
33 J- Gonda, Change and Continuity in Indian Religion (The Hague,
1965), p. 326.
34 Ibid., p. 322.
35 Apastamba Srauta-sutra, X. 11.5-15 quoted in Srautakosa, II, i,
p. 80.
36 See Fredrick M. Smith, 'Indra’s Curse, Varuna’s Noose, and the
Suppression of Women in the Vedic 3rauta Ritual’ in Julia Leslie,
ed., Roles and Rituals fo r H indu Women (Delhi, 1992),
24 CASTE

pp. 17- 45. For the use of sacrifice as an occasion for defining
gender-based relationships see Kumkum Roy, The Emergence of
Monarchy in North India: Eighth to Fourth Centuries B C. as
Reflected in the Brahmanical Tradition (Delhi, 1994), p. 67.
37 For example, the Koch Rajbangshis. This point was made by
Vasanthi Raman. See Vaskar Nandy and Vasanthi Raman, ‘The
Long Transition: The Case of the Koch-Rajbangshis of North-Eastern
India’, paper presented at the Seminar on ‘From Tribe to Caste',
Indian Institute of Advanced Study, Simla, 8-12 November 1993­
38 S. Jaiswal, ‘Stratification...’, pp. 29f; Irawati Karve, ‘Kinship
Terminology and Usages in Rgveda and Atharvaveda', Annals of
the Bhandarkar Oriental Research Institute (hereafter ABORT),
XX (1938-9), pp. 219f; Kumkum Roy, op. cit., pp. 246-7.
39 RV, X. 86.10.
40 Vedha rtasya tnrini, ibid.
41 RV, II. 32.7.
42 RV, VII. 1.6.
43 RV, II. 17.7. Compare this with A V, I. 14.3 where such a maiden is
called the kulapa of the rajan (or Yama?). RV, X. 179 2 speaks of
kulapas, heads of families, attending upon the chief (vajapati).
44 RV, VIII. 31.8.
45 For example, IV. 5.5.
46 Hanns-Peter Schmidt, Some Women's Rites and Rights in the
Veda (Pune, 1986), pp. 30-75.
47 Irawati Karve, 'Kinship Terminology and Kinship Usage in Rgveda
and Atharvaveda', ABORI, XX (1938-9), pp. 109-44; idem, ‘The
Kinship Usages and the Family Organization in Rgveda and
Atharvaveda', ABORI, XX (1938-9), pp. 213-14; idem, 'Kinship
Terms and the Family Organization as found in the Critical Edition
of the Mahabharata’, Bulletin of the Deccan College Research
Institute, V (1943-4), pp. 61-148.
48 Idem, ABORI, XX (1938-9), pp. 2l6f. Also see Sarva Daman Singh,
Polyandry in Ancient India (Delhi, 1978), pp. 19. 39f.
Karve is aware that the stories of Yama-Yami and the Sun-god’s
incest with his daughter may have other meanings and may not be
literal evidence for the prevalence of marriage between siblings
or promiscuity. But she says 1Kinship Organization in India
(Bombay, 3rd edn., p. 32)] that not all references can be dismissed
in this manner. For example, RV, X. 162.5, which is an incantation
to drive away the demon causing abortion goes as follows: ‘He
who sleeps with you becoming your brother, husband or lover and
who kills your progeny, him I destroy.' Apparently here the demon
 INTRODUCTION 25

is supposed to take the form of a person to whom the woman

would be sexually accessible, the husband, lover or brother! Griffith
excludes this hymn from his English translation, providing a Latin
translation in the Appendix.
50 S.V. Karandikar, H indu Exogamy (Mumbai, 1928).
51 Irawati Karve, Kinship Organization in India, pp. 51f.
52 G.S. Ghurye, Two Brahmanical Institutions-. Gotra and Cbarana
(Mumbai, 1972), pp. 293f-
53 Emile Benveniste, Indo-European Language and 5bcj'e/y(London,
1973), p. 303.
54 John Brough, The Early Brahmanical System ofGotra and Pravara-.
A Translation o f the Gotra-Pravara-Manjari of Purusottama
Pandita with an Introduction (Cambridge, 1953).
55 Genda Lerner,77?e Creation of Patriarchy (Oxford, 1986).
56 AV, XI.8.1.
57 S. Jaiswal, ‘Studies in Early Indian Social History...', p. 14, fn. 4.
58 RV, IX.92.6; X.97.6. Also see 1.40.8, and X.42.10.
59 RV, X.90.12; AV, X1X.6.6.
60 Satapatha Brahmana, XII. 7.3-8.
6 1 Ibid., IV. 3.3.15; see Jogiraj Basu, India ofthe Age ofthe Brahmanas
(Calcutta, 1969), pp. 115-19.
62 It is interesting that the term foa/ra/ksatriya is derived from the
root ksi, which also has the sense of ‘dwelling’ and ‘movement’
(Monier-Williams, SED, s.v. ksi, Nighantu, II..14, Nirukta, II.4; 11.21).
Thus ksaya meaning ‘dwelling’ or abode and ksiti meaning
‘habitation’ are also derived from the same root which also has
the sense of ‘to move, go'. Apparently at one time dwellings or
habitations were mobile, and those who protected them were
known as ksatra. The task of protection was associated with
power', ‘might’ and ‘governance’. For a movable dwelling see
the A V, IX. 32.4 which says Like a bride O dwelling, we carry thee
where we desire’ (vadhumiva Wa saleyatrakamam bharamasi).
6 3 R.S. Sharma, Aspects of Political Ideas and Institutions in Ancient
India (Delhi, 3rd edn., 1991), pp. 178-9.
64 Satapatha Brahmana, XIV.4.2.27, also II. 1.4.4.
65 Arrian, VI.16; Diodorus, XVIII, CII quoted by K.P. Jayaswal, Hindu
Polity (Bangalore, 4th edn., 1968), pp. 65-6; Patanjali’s
Mahdbhasya, Vol. II, quoted by Jayaswal, p. 66.
66 Jogiraj Basu, op. cit., p. 37; Suvira Jaiswal, ‘Caste in the Socio­
Economic Framework of Early India’, Presidential Address, Section
I, Proceedings o f the Indian History’ Congress, 38th Session
(December 1977), pp. 32, 34-6. Also see R.S. Sharma, Sudras in
26 CASTE

Ancient India (Delhi, 2nd edn., 1980), pp. 69-70.

67 S. Jaiswal, ‘Studies ...', IHR, VI, Nos. 1-2 (1979-80), pp. 9-13.
68 Claude Meillassoux, 'Are There Caste in India?’ Economy and
Society, Vol. II (1973), pp. 92-3.
69 Nimkta, XII. 13­
70 Sudranam aniravasitanam, Pa., II. 4.10.
71 Mahabhasya, edited by Keilhorn (Bombay Sanskrit Series, Mumbai,
1892), Voi. I, p. 475.
72 The Brhadaranyaka Upanisad speaks of the gods belonging to
the brahma, ksatra, vaisya and sudra varnas, 1.4.11-15, R E. Hume, tr.,
Thirteen Principal Upanisads (Delhi, 2nd edn., 1969), pp. 84-5.
73 Manusmrti, X.20-3- For the existence of four varnas among trees
and the animal world see C.G. Kashikar, ed., Srautakosa, Vol. I,
Pt II (Pune, 1962), p. 1156.
74 SuviraJaiswal, ‘Varna Ideology and Social Change', Social Scientist,
Vol. 19, Nos. 3-4 (March-April 1991), pp. 41-8; idem, ‘Studies ...’,
Section IV, pp. 70f.
75 Dipankar Gupta, ‘From Varna to Jati: The Indian Caste System
from the Asiatic to the Feudal Mode of Production', Journal of
Contemporary Asia, Vol. X (1980), pp. 249-71.
76 P.V. Kane mentions some brahmana castes named after the Vedic
sakhas studied by their ancestors, such as Kanvas, Maitrayaniyas,
Carakas, etc., History ofDharmasastra, Vol. II, Pt II, p. 976. Gonda
relies on D. Bhattacharya in Fundam ental Themes o f the
Atharvaveda (Pune, 1968), p. 39, for the statement that the
brahmanas of the other Vedas do not practice commensality or
connubium with the Atharvanic brahmanas (Paippaladins) of Orissa;
J. Gonda, Vedic Literature. A History of Indian Literature, Vol. I
(Wiesbaden, 1975), p. 267. For a hierarchical arrangement of
brahmana lineages named after land measures, such as Bis Biswa
(lineage with 20 measures of land) and Solah Biswa (lineage with
16 measures of land) among the Kanyakubja brahmanas, see R.S.
Khare, ‘The Kanyakubja Brahmans and Their Caste Organization’,
Southwestern Jo u rn a l o f Anthropology, Vol. 16 (I960),
pp. 348-67.
77 For brahmanas see B.P. Mazumdar, Socio-Economic History of
Northern India (from A.D. 1030 to 1194) (Calcutta, I960), p. 81.
For vaisyas, K.C. Jain, Malawa Through the Ages (From the Earliest
Time to 1305 A.D.) (Delhi, 1972), p. 485.
78 J. Allan, Catalogue of the Coins of Ancient India (in the British
Museum) (London, 1936), pp. cxviiif, 164.
79 Majjhima Nikaya (edited by Mahapandita Rahula Sankrityayana
 INTRODUCTION 27

and P.V. Bapat, 3 vols, Nalanda Devanagari Pali Series, 1958),

Vol. II, p. 447.
80 Kasikavivarana-pahjikd (edited by Srish C. Chakravarti), Vol. II,
p. 343 quoted by B.P. Mazumdar, ‘Polity of the Andhaka-Vrsni
Sangha’, D r Satkari Mookerji Felicitation Volume (The
Chowkhamba Sanskrit Studies, Vol. LXIX, Varanasi, 1969), p. 207.
81 Ibid.
82 T.W. Rhys Davids and J.E. Carpenter, eds., Digha Nikaya, Vol. I
(London, 1890), p. 97; Narendra Wagle, Society at the Time of the
Buddha (Mumbai, 1966), pp. 102-3- Compare this with the
statement in the Pancainmsa Brahmana (XIV.6.6) that Vatsa, the
son of sage Kanva from a sudra wife, proved to be better brahmana
than his other son Medhatithi. For a useful discussion of the early
Buddhist attitude to caste see Richard F. Gombrich, op. cit.,
pp. 354-8.
83 Trisala, the sister of the Licchavi raja Cetaka had married thcjnatrika
raja Siddhartha, father of thejaina Tirthankara Mahavira. Cetaka’s
daughter Chelana had married the ksatriya king Bimbisara of
Magadha. According to the Mahavastu the Koliyan and Licchavi
princes had competed with prince Siddhartha (the Buddha) for
the hand of the Sakyan princess Yasodhara. A careful scrutiny of
the terms relating to kinship and marriage in the early Pali sources
leads Wagle to conclude that ‘endogamy and commensality, the
two fundamental characteristics of modern caste’, were absent.
He further writes that marriage ‘with a non- nati that Ls, one outside
the “extended kin-group” was permissible outside the caste'
especially when it took place between the high varna-categories.
Thus marriage between khattiya and brahmana meets no strong
disapproval, unlike the marriage of a brahmana with a dasi which
was strongly disapproved. N. Wagle, Society at the Time oj the
Buddha, pp. 132-3­
84 Manusmrti, X.43-4 speaks of ksatriya-jatis in plural.
85 K.S. Singh, People of India: An Introduction (Calcutta, 1992),
pp. 24f.
86 R.S. Sharma, Social Changes in Early Medieval India (circa A.D.
500-1200), First Devraj Chanana Memorial Lecture (Delhi, 1969).
87 Hermann Kulke, Kings a n d Cults: State Form ation an d
Legitimation in India and Southeast Asia (Delhi, 1993), pp. 82-93­
88 Ibid., p. 85.
89 S. Nagaraju, ‘Emergence of Regional Identity and Beginnings of
Vernacular Literature-. A Case Study of Telugu', Social Scientist,
Vol. 23, Nos. 10-12 (October-December 1995), pp. 8-23-
28 CASTE

90 Nagaraju argues that Jaina and Saiva orders at least ideologically

‘did not subscribe' to the varna-yis/j hierarchical structure and hence
had 'an easy and mutually beneficial interaction with the new
social class’. Even if these religious sects were more open than the
Vedic orthodoxy, the Jaina records of the ninth and tenth centuries
from Karnataka in fact show strong caste prejudices. The Adi
Purana of Pampa attributes the creation of the four varnas to
tirthankara Adinatha and his cakravartx son Bharata, and describes
Adinatha as forbidding the mixing of castes (vamasamkara). K.L.
Narasimha Sastri, ed. and tr., Adi Purana of Pampa (Bangalore,
1980), VIII. 64 and XV.6-12 quoted in U. Malini Bhat, Religion and
Society in Southern Karnataka in the Early Medieval Period’,
unpublished Ph.D. thesis (jawaharlal Nehru University, Delhi,
1996), p. 327. The Arikanathapura memorial of Camakkabbegal
describes her and her two sons as supporters of the caturvarna
and the Sravana Samgha, and the Cavundaraya Purana explicitly
denies artisans and other low-caste groups the right to be initiated
in the Jaina ritual vow of sanyasana. S. Settar, Pursuing Death.
Philosophy and Practice of the Voluntary Termination of Life
(Dharwad, 1990), p. 251. The Laksmesvara inscription of the reign
of Vikramaditya VI ( a d 1081) describes a distinguished pious Jaina
family of Dinakara, Rajumayya and Dudama, and speaks of
Dinakara as 'the sun in the sky of the brahmana race’. Dudama too
is a 'scion of worthy brahmanas’ (£7, XVI, p. 9). It was correct
observance of caste rules and not adherence to a belief system
which was crucial to individual identity. Even husband and wife
could be followers of different religious faiths depending on their
individual inclination,without creating discord or contradictions.
Thus, for example, the inscriptions of the Nojamba ruler
Mahendradhiraja, who lived in the last quarter of the ninth century,
describe him as a staunch 3aiva who built the temple of
Mahendresvara at Baragur (Epigraphia Carnatica, XII, old edn.,
Si. 38), but his queens Bijaya Mahadevi, Parama Mahadevi, Akkabbe
and Domabbe, were Jainas and patronized the Jaina basadis
(EC', XII, old edn., Si 24 of cab 880). Similarly Calukya Mahasamanta
Durga (tenth century) was a Saiva, and his wife Pittabbe a Jaina.
A.M. Shah writes that in Gujarat a Vania caste may have both Hindu
and Jaina members without restrictions on intermarriage, and
frequently husband and wife have different religious affiliations
(Division and Hierarchy An Overview of Caste in Gujarat, Delhi,
1988, pp. 1-39, reprinted in K.L. Sharma, ed., Social Inequality in
India, Jaipur, 1995, p. 225, n. 6).
 INTRODUCTION 29

91 For example, the caste of scribes known as Kayasthas.

92 Richard G. Fox, ‘Varna Schemes and Ideological Integration in
Indian Society’, Comparative Studies in Society and History,
Vol. XI (1969), pp. 27-45.
93 Gita Dharmapal-Frick, ‘Shifting Categories in the Discourse
on Caste: Some Historical Observations’ in Vasudha Dalmia and
H. von Stietencron, eds., Representing Hinduism (New Delhi,
1995), pp. 82-100, refers to a Tamil poem, Ererubadu, cited by
Ziegenbalg, a seventeenth-century Pietist-Lutheran missionary
based in the Danish enclave of Tranquebar. The poem is in praise
of the plough and places the sudra VeHa|a on a pedestal even
higher lhan the brahmana. She quotes:' Even being born a Brahmin
does not by far endow one with the same excellence as when one
is born into a VeNaki family.’ Such high praise reminds us of the
panegyrist of Singaya Nayaka of Akkalapundi grant, who claimed
that the sudra varna to which his patron belonged was higher than
the olher three varnas, as it came out of the feet of Visnu along
with the holy river Ganga. This only goes to substantiate my
contention that, at least in the south and in Bengal, the Sudra stood
next only to the brahmana and was not a degraded status as was
the case in the Gangetic valley, the cradle of the fourfold varna
system. It is for this reason that Ziegenbalg’s writing gives the
impression that the sudras represented the ‘genuine mass of Tamil
society’, which latter category would of course, have excluded
the outcastes or avarnas. The scheme of social gradation followed
in the Aparajita-priccha, a twelfth century work on architecture
composed in Gujarat, sticks to the conventional enumeration of
varnas in descending order. What is more significant, however, is
the fact that sudras are placed above karsakas and prakrtis.
interpreted as dependent peasants and serving castes of craftsmen
respectively. B.N.S. Yadava, Presidential Address, Indian History
Congress, 53rd Session (Warangal, 1993), p. 17, fn. 1. Perhaps the
compilation of Census Reports in colonial times made the people
of these areas more conscious of the pejorative meaning of the
term 'sudra'. Infra, Chap. 2, secs, iii, iv.
94 S. Arasaratnam, ‘Social History of a Dominant Caste Society: The
veljajar of North Ceylon (Sri Lanka) in the 18th Century ’, The Indian
Economic and Social History Review, Vol. XVIII, Nos. 3 and 4
(July-December 1981), pp. 377-91. Also see E. Thurston, Castes
and Tribes of Southern India (Chennai, 1909), Vol. VII, p. 376.
95 The doctrine of karma is propounded for the first time in the
Brhadaranyaka Upanisad (III.2), which tells us that those who
30 CASTF.

please the gods by good conduct are reborn into the three higher
varnas but those whose conduct is evil enter into a foul and stinking
womb, such as that of a bitch, a pig or an outcaste (V.10.7-K). For
the origin of this doctrine in the elite circles of the later Vedic age,
see A.L. Basham, The Origin and Development o f Classical
Hinduism , edited and annotated by Kenneth G. Zysk (Delhi,
1990), Chap. 3­
96 K.L. Sharma, ‘Stresses in Caste, Stratification: A Study of Six Villages
in Rajasthan’, Economic and Political Weekly, 4(3), 1969.
97 Infra. Also see R.S. Sharma, 'Varna in Relation to Law and Politics’;
idem, Aspects of Political Ideas and Institutions in Ancient India,
Chap. XVI.
98 The Saka rulers Rudradaman and Usavadata were great patrons of
brahmanas. For the claims of a number of kings to have regulated
the order of castes see ibid., pp. 234-5. Kulke points out that
Somadatta, who ruled in north Orissa in the first quarter of the
seventh century under the overlordship of king Sasarika of Bengal,
explicitly states in his Midnapur inscription that he followed the
laws of Manu; Anncharlott Eschmann, Hermann Kulke and Gaya
Charan Tripathi, eds., The Cult of Jagannath and the Regional
Tradition of Orissa (Delhi, 1978), p. 127, fn. 5.
99 For a critique of Dumont's view which separates ritual status from
power by interpreting brahma as denoting the former and ksatra
as denoting the latter category, see Suvira Jaiswal, ‘Varna
Ideology...’, Social Scientist, Vol. 19, Nos. 3-4, p. 44; idem, ‘Studies
in Early Indian Social History’, IHR, VI, pp. 3-7.
100 Nicholas Dirks, The Original Caste: Power, History and Hierarchy
in South Asia', Contributions to Indian Sociology, Occasional
Papers, Vol. V (1990), pp. 59-77.
101 Ronald Inden, Imagining India (Oxford, 1990).
102 Ibid., p. 227.
103 Ibid., p. 229.
104 Ibid., p. 268.
105 For an excellent critique of Ronald Inden 's work see Aijaz Ahmad,
' Between Orientalism and Historicism: Anthropological Knowledge
of India’, Studies in History, Vol. 7, No. 1 (january-June 1991),
pp. 152f. Ahmad remarks, ‘only the mindless have ever asserted
that kings and dominant castes lacked “agency", in the sense of
capacity to act purposefully to change their own circumstances, or
that the Indian was constitutionally irrational’ (ibid., p. 160).
106 K.M. Shrimali, ‘Reflections on Recent Perceptions of Early Medieval
India', Presidential Address, Section IV, Andhra Pradesh History
 INTRODUCTION 31

Congress, XVIII, Annual Session (8-9 January 1994), p. 8.

107 R. Inden, op. cit., p. 82, is wrong in assuming that castes became
autonomous structures on the decline of Hindu kingship. Long
before the advent of Muslims on the scene Manu advised the ruler
to respect the laws of castes (Manusmrti, VII.203; VIII.41), which
were already autonomous decision-making groups. For colonial
interventions undermining the traditional autonomous jurisdiction
of caste groups over their own affairs and their resistance to this
process see Kanakalatha Mukund, 'Caste Conflict in South India in
Early Colonial Port Cities: 1650-1800', Studies in History, Vol. XI,
No. 1, n.s. (1995), pp. 1-20.
108 On the character and function of paura-janapada assemblies,
R.S. Sharma, op. cit., p. 82.
109 Irfan Habib, Interpreting Indian History, p. 19.
110 Infra, Chap. 2, sec. ii.
111 Even for the earlier period there is little evidence to show that the
6aka and Kusana rulers, ‘the heretical sudra kings' of brahmanical
perception, fraternized with low-caste sudra slaves and hired
labourers owing to their non-commitment to varna ideology, as
R.S. Sharma suggests (Sudras in Ancient India, p. 235). On the
contrary their quick absorption of varna ideology is well attested.
See n. 98 of this chapter. The fact that some of them patronized
Buddhism does not prove their antipathy to the varna system. The
Buddhist texts of the early centuries of the Christian era, such as,
the Milindapanho and the Duddbacarita of Asvaghosa, are imbued
with varna ideology. Incidentally, the authc >r<>fthe Vajrasuciseems
to have been different from the author of the Buddbucarita, and
later in date. He appears to have belonged to eastern India. Whereas
the Buddbacatita accepts the norms of the varna system, the
Vajrasuci criticizes it. At any rate, although ideologically Buddhism
was opposed to caste hierarchy and as such could have had special
appeal to those who were its victims, in practice varna/caste
prejudice is evident even in the early Pali canon. The main thrust
of passages dealing with the question of caste in these texts is to
prove the superiority of the ksatriyas over the brahmanas. In Sri
Lanka where caste structuring took place under Buddhist inspiration
we are told that it was the lawful activity of the king of Kandy ‘to
ordain appropriate function to various castes: he could also degrade
certain villages or families of high caste to a lower status ... . Ralph
Pieris, Sinhalese Social Organization: The Kandyan Period
(Colombo, 1956), p. 180. quoted in Richard F. Gombrich, op. cit.,
p. 349.