PEASANT STATE AND SOCIETY

IN MEDIEVAL SOUTH INDIA

BURTON STEIN

DELHI
O X FO R D U N IV ER SITY PRESS
O X FO RD NEW YORK
1980
 Oxford University Press
OX FO R D L O N D O N GLASG OW
N E W Y O RK TO R O N TO M ELBOURNE W E L LIN G TO N
NA IRO B I D A R E S SALAAM C A P E T O W N
K U ALA LU M P U R SIN G A PO R E H O N G K O N G TO KY O
D ELH I BOMBAY CA LC U TTA M ADRAS KA RA CH I

© Oxford University Press 1980

Filmset by All India Press, Pondicherry

printed by R ajbandhu Industrial Co., New Delhi 110027
and published by R. Dayal, Oxford University Press
2/11 Ansari Road, Daryaganj, New Delhi 110002
 Preface

This book is somewhat perverse in theoretical and historiographical

senses. Being about peasants without lords who established and
m aintained great states, enduring and complex local institutions,
and elaborate religious institutions, it is a study which denies, or
at least questions, certain of the strongly held conceptions of what
peasant society is in contem porary social science. Being a book
about early South Indian society viewed at the interface of what
is usually considered the ‘little’ and the ‘great’ tradition, a study
which looks upward and downward from this interface, as it
were, and further, a study which is concerned with social and
cultural change, it is also a work which denies, or seriously ques­
tions, m uch o f the extant historiography on early South India.
The purpose o f this study was never conceived as historiogra­
phical, th at is, an essay upon the established interpretations of
some part o f the medieval past o f South India. Yet, it has become
such a study. Issues in South Indian history which have acquired
a com fortable and com forting acceptablity am ongst many
students are found not to stand well before the queries this study
poses. The nature o f the political system and the state have been
found to be wrongly conceived by scholars of the p a s t; institutions
such as the peasant village, its locality context and rural groups
have been found to be poorly characterized, not so much on a
factual basis, but in the constructions put upon the extant evidence.
This has required the re-examination o f much that has been written
and what, for some readers, will seem a heavy burden of tedious
recapitulation o f old arguments and evidence. The cause of cum ula­
tive scholarship makes any other m ode o f exposition less than
responsible, but an inevitable consequence for some readers will
be the presence o f too much detail and for others the appearance
of disrespect, even arrogance, about what previous scholars in
this field have accomplished.
This would be an understandable but unfortunate conclusion.
This study began as an effort to discover in the published scholar­
ship on medieval South India a basis for dealing with, or con-
textualizing, the changes in South Indian society attendant upon
British rule. It was originally conceived as an exploration in pre-
British history meant to set the stage for changes in South Indian
society during the nineteenth century. That this exploration should
have resulted in a radical reworking of the pre-British past was
unanticipated when it began some twenty years ago; this rather
substantial reform ulation of many aspects of pre-modern his­
toriography was never construed as an end in itself. On the con­
trary, what has occurred here is a reform ulation o f certain aspects
o f South Indian history in response to perceived inadequacies
o f interpretation as the present writer, with his interests and
background, has looked at the field. The debt to previous writers
is obvious throughout this book, and, to m ake this sense of obli­
gation most explicit, the book is dedicated to that handful of
scholars o f the South Indian past upon whose efforts the present
book firmly rests.
M ore than acknowledging the debt to previous workers in the
field is necessary, however. F or years, the writer has received
research funds which have perm itted long and valuable sojourns
in India and England. A Ford Foundations Foreign Area Training
Fellowship first made it possible to conduct dissertation research
in South India in 1956-7. A Fulbright Research Fellowship made
another year of archival research possible in 1961-2, and a supple­
mental grant from the Social Science Research Council permitted
time in London for work in the India Office Records and Library
as well as the British Museum. Two valuable further years (1966-8)
in South India were supported by grants from the American
Institute o f Indian Studies, again with a supplemental award
from the Social Science Research Council, and, in 1973, a grant
from the American Council of Learned Societies made possible
a year o f research in London. Finally, another American Institute
of Indian Studies grant enabled the writer to spend three months
in Tamil N adu in 1975. T o these several organizations and to the
libraries, archives, and other research facilities in which their
grants made it possible for the writer to work, he expresses his
gratitude. To the learned epigraphists of the Governm ent of
India in Ootacam und and, later, in Mysore, special thanks are
extended; without their patient assistance and instruction, little
of this would have been possible.
But, to attem pt to acknowledge the contributions of all from
whom something valuable to this work has been obtained would
be impossible. M any who read this study will discover, or redis­
cover, their contributions, some to their delight and others to
their chagrin. F o r an author to pause reflectively on the obligation
to the community o f scholars in his field — those of the past and
the present — is a rich experience. Dedication o f this book to
th at com m unity o f Indians and non-Indians, to professors and to
students, is the only fitting way to acknowledge that obligation.

H onolulu, Hawaii Burton Stein

M ay 1979
 Contents

Preface V

Note on Transliteration and Pronunciation xi

Glossary xiii
Abbreviations XV

Introduction 1
I South India: The Region 30
II Form ation o f the Medieval A grarian Order : Brahman
and Peasant in Early South Indian History 63
III Peasant M icro Regions: the Nadu 90
IV The Corom andel Brahmadeya Village 141
V Right and Left Hand Castes (yalangai and idangai) 173
VI The Transition to Supra-local Integration in the
Twelfth and Thirteenth Centuries 216
VII The Chola State and the A grarian Order 254
VIII The Vijayanagara State and Society 366
Bibliography 489
Index 519

MAPS „ . page
Jacing
1-1 The South Indian Region 30
IV-1 Brahmadeya Distribution c. a . d . 1300 150
VII-1 The Chola Macro Region c. a . d . 1300 286
VII-2 Tondaimandalam Zonal Divisions c . a . d . 1300 310

F IG U R E S
VIII-1 Temples of Tamil Nadu, a . d . 1300-1750, by Mandalam,
Deity and Period 460
VIII-2 Proportions of Temples of Tamil Nadu, by Deity,
for Four Periods, a . d . 1300-1750 461
 TABLES
III-l Areal Variation of Selected Nadus a . d . 1300 93
III-2 Selected Large and Populous Cholamandalam Nadus by
Area, Villages, and Relation to Water Supply 95
III-3 ‘New Nadus' (First References) in Cholamandalam and
Naduvil-AWw, a . d . 850-1300 96
VII-1 Distribution of Chola inscriptions and Brahmadeyas by
Selected Madras Taluks to 1915 308
VII-2 Segmentary Areal Groupings in the Central Tamil
Plain during Chola Times 312
VIII-1 Village Endowments to the Tirupati Temple and
Types of Tenure by Donor Groups, 1509-68 430
VIII-2 Monetary Endowments to the Tirupati Temple,
1509-68 431
VIII-3 Temples of Tamil Nadu, a . d . 1300-1750, by Deity,
Mandalam and District 456
VIII-4 Temples of Tamil Nadu, a . d . 1300-1750, by Deity,
Period and Mandalam 458
VIII-5 Temples, a . d . 1300-1750, by Modern Taluks
Arranged by Mandalams 462
VIII-6 Valangai-Idangai Designation 474
 Note on Transliteration and
Pronunciation

Certain Indian-language terms have not been transliterated. These

include proper nouns such as the names of castes (e.g. Reddi),
kings (e.g. Rajaraja), places (e.g. Tanjavur), and authors and titles
o f Indian-language texts (e.g. Civakacintamani by Tiruttakkadeva).
Also included am ong the untransliterated words are those well-
established enough in English usage to appear in the Concise
O xford Dictionary 6th.ed. (e.g. vihara, sastra, puja). Finally, a
selection o f frequently used, technical terms have been separately
glossed (q.v. ‘G lossary’); after having been transliterated in their
first use in the text they are simply italicized thereafter.
The transliteration of Sanskrit words has followed Monier-
Williams, A Sanskrit-English Dictionary (see ‘Abbreviations’) and
Dravidian words generally have followed the D.E.D. augmented
by the M .U .T .L. (see ‘A bbreviations’) for Tamil words. There
are two exceptions to this. The first involves Sanskrit borrowings
into the Tamil where the orthographic system o f the latter is
peculiarly unadaptive and results in awkward, at times unintelligi­
ble, constructions (e.g. brahmdisava, the principal annual festival
o f a temple, is rendered by the M .U .T .L. as piramdrcavam) ; in
such cases, Sanskrit transliteration is followed to which is added
an tn as occurs in earlier usage (hence: brahmdtsavam; and gra-
mam, village, for kiram am ; prasddam, consecrated offering, for
piracatam; prabandham, poetic composition, for pirapantam).
The other deviation from the D.E.D. and M .U .T.L. schemes
for Tamil follows the practice of rendering the voiceless stops
p, t, k, in initial positions o f a word as voiced stops b, d, g, when they
occur between vowels, and rendering the hard palatal t as d. T h u s:
parambu, revenue account, not parampu; tiruccendur, a temple
town, not tiruccentur; and kongu, a large territory in Tamil country,
not konku. However, in Tamil and other Dravidian languages,
long o r double consonants occur (e.g., ndttar, those o f the country
or locality); these and kk, nn are pronounced as in the English
ho/ tea, blac& cow, ten wights. The letters r and r are trilled, and
the long consonant rr is pronounced with a hard trill (th u s: kurram,
a local territorial unit, is pronounced like kootrum ; similarly, n r
is pronounced ndr as in the English laundry.
 Glossary

The following selection of terms are formally transliterated on

their first occurrence in the text and thereafter appear in italic
script only. After each transliterated term, the language and a
brief definition are given. W here appropriate, the location in the
text, where a full discussion o f the term occurs, is also given.
Standard glossaries for the m aterial in this book are: M U T L ;
S .I.T .L .v . 3, pt 2, following p. 1610; D.C. Sircar, Indian Epigraphical
Glossary, M otilal Banarsidass, 1966; T.V. M ahalingam, South
Indian Polity, 2nd ed., University of M adras, M adras, 1967,
p. 421 ff (see ‘Abbreviations’).

amman. Tamil. G oddess; divine lady; q.v., ch. VI.

bhakti. Sanskrit. D octrinal principle o f salvation through devo­
tion; q.v., ch. I.
brahmadeya. Sanskrit. Gift to B rahm ans; specifically, a grant of
village income and management to B rahm ans; q.v., ch. IV.
devaddna. Sanskrit. G ift to gods; specifically, the endowment of
income from land or a village to a tem ple; q.v., ch. I.
gangavadi. Tamil. Territorial unit, like mandalam (q.v.) in K arna­
taka.
ghatika. Sanskrit. Centre for advanced studies; same as Tamil
salai (from S anskrit: said),
kallar. Tamil. A people of southern Tamil country.
kammdlar. Tamil. Artisan-traders, usually including five groups,
hence also pdhcdlar.
kurram. Tamil. Local territory, especially in tondaimandalam
(q.v., mandalam).
mandalam. Sanskrit. Large territorial units o f which the principal
ones are : tondaimandalam, colamandalam, kongumandalam, pandi-
mandalam.
maravar. Tamil. A people o f southern Tamil country.
mat ha. Sanskrit. Sem inary; educational centre.
nddu. Tamil. Local territory; nattar, those of the territo ry ; q.v.,
ch. III.
nayaka. Sanskrit. Title of chief; nayankara, status o f chief; q.v.,
ch. VIII.
nu{o]iambavadi. Tamil. Large territory, like mandalam (q.v.).
periyanadu. Tamil. Big-nadu (q.v.); a territory comprising several
nadus; q.v., ch. VI.
prasasti. Sanskrit. Eulogy; eulogistical preamble to inscriptions;
m eykklrtti in Tamil.
sabha. Sanskrit. Assembly; specifically, the assembly o f a brah­
madeya (q.v.) village; also mahasabha, q.v., ch. IV.
sangam. Sanskrit: samgha. Community o f Buddhist or Jaina
sectarians; specifically, such a community or academy o f Jaina
literary scholars in fifth century M adurai,
wr. Tamil. Village and assembly o f a village; urdr, those of the
village.
valangai-idarigai. Tamil. Right and left hand; specifically, castes
o f the right and left division; K annada: balagai-edagai.
 Abbreviations

A .R .E . India, D epartm ent o f Archaeology, Annual

Report on Indian Epigraphy. Published under
slightly different titles and in different places
from 1887.
D.E.D. A Dravidian Etymological Dictionary, T. Burrow
and M.B. Emeneau, Clarendon Press, Oxford,
1961.
E.C. Epigraphia Carnatica, 16 v., Governm ent Press,
M ysore/Bangalore/M adras, 1889-1955.
E.I. Epigraphia Indica, Archaeological Survey of
India, Calcutta/Delhi, 1892.
H.C.G.E.S.I. Historical and Cultural Geography and Ethnology
o f South India with Special Reference to Cola
Epigraphs, B. Suresh Pillai, Deccan College
Ph.D. Thesis, Poona, 1965.
H .S.I. A History o f South India from Prehistoric Times
to the Fall o f Vijayanagar, Oxford University
Press, M adras, 1955.
I. A. The Indian Antiquary, M adras, Bombay, 1872-.
I.E .S.H .R . The Indian Economic and Social History Review,
Delhi, 1963-.
f.H .Q . The Indian Historical Quarterly, Calcutta, 1925-.
I.M .P . A Topographical List o f the Inscriptions o f the
Madras Presidency, Collected till 1915 with Notes
and References., V. Rangacharya, ed., 3 vols.,
Governm ent Press, M adras, 1919.
J.A .H .R .S. Quarterly Journal o f the Andhra Historical Re­
search Society, R ajahm undry, 1926-.
J.A.O .S. Journal o f the American Oriental Society, New
Haven, 1843-.
J.I.H . Journal o f Indian History, Trivandrum , 1921-.
J.R .A .S. The Journal o f the Royal Asiatic Society o f
Great Britain and Ireland, London, 1834-.
M .U .T.L. Tamil Lexicon; Published under the Authority
 o f the University o f Madras, 6 v., University of
M adras, M adras, 1936.
Q.J.M.S. The Quarterly Journal o f the M ythic Society,
Bangalore, 1910-.
S. 1.1. South Indian Inscriptions, Archaeological Survey
of India, M adras/Delhi, 1890-.
S.I.T.J. South Indian Temple Inscriptions, 3 v., T.N. Sub-
ram aniam , ed., Governm ent Oriental Series,
M adras, 1953-7,
T.A.S. Travencore Archaeological Series, 9 v., T ri­
vandrum , 1910-41.
T .A .S.S.I. Transactions o f the Archaeological Society o f
South India, M adras, 1955-.
T.T.D.I. Tirumalai-Tirupati Devas'thanam Epigraphical
Series; Report on the; Inscriptions o f the Devas-
thanam Collection and 6 v. o f texts, S. Subrah-
m anya Sastry, ed., M adras, 1930-8.
Wilson Glossary H .H . Wilson, A Glossary o f Judicial and Revenue
. Terms, M unshiram M anoharlal [reprint], Delhi,
1968.
 Introduction

This study was begun many years ago in an effort to establish the
foundation for an analysis of m odem agricultural development
in the M adras Presidency of the British period and its successor
administrative units in independent India. The task proved to be
an extended one, and the results presented here are far less complete
than was naively planned at the outset. As originally intended,
the benchm ark analysis of the South Indian agrarian system was
to come up to the nineteenth century. This has not proved possible.
Essentially what is presented is an analysis of the South Indian
agrarian system to the end of the Chola period, the late thirteenth
century; to this has been added a discursive final chapter suggesting
trends to the seventeenth century.
Partly, the inordinate time required for the preparation of this
study resulted from the need to master a vast and complex corpus
of evidence. It was supposed at the outset that this evidence had
already been m ore or less satisfactorily synthesized by three gene­
rations of scholarship on the medieval period of South Indian
history. This supposition proved an incorrect one. Denigration
of this published work is not intended by the last observation
and by references below to viewpoints in the extant published
work against which reservations and, even more blunt, disagree­
ments are lodged. By com parison with other parts of the Indian
sub-continent, the literature on medieval South India possesses
distinguished scholarly qualities. And, it may be added that while
one reason for this relative excellence has been the richness o f the
inscriptipnal records o f the South Indian medieval age, the obverse
side of this richness has been an enormous task of analysis and
synthesis. Repeatedly in the discussion below, grateful cognizance
is taken o f the accomplishments of the author’s predecessors.
Still, the purposes of cumulative scholarship — the serious exami­
nation and re-examination of the findings of one’s predecessors
— has resulted in an overall critique of the established historio­
graphy on medieval South India. It is also to be recognized that
the present w riter’s interests and his somewhat different theoretical
assumptions have led to the raising of questions and the elucidation
of findings which are antagonistic to some of the established his­
toriography. if in nothing else, the concentration upon peasant
society and culture v/ould have assured this dissonant relationship
with respect to much that has been written on medieval South
India.
The peasant societies of medieval South India are more ancient,
durable and, yes, elegant than most in the world. Yet, the rich
historical literature on South India affords few descriptions of
these peasant societies and alm ost none relating peasantry to other
social elements in the development o f Indian society and culture.
References to agricultural techniques, landholding, crops, m arket
organization, and relations among peasant agriculturists and
others, including formal political authority, may be found scattered
throughout the historical literature, but only rarely is systematic
attention given to any o f these m atters save that o f land revenue.1
In too many m onographs and, o f course, in the m ore general
histories of medieval South India, it is assumed that the relationship
existing between the web of social and cultural phenom ena com ­
prising medieval South Indian peasant society and other sectors
of the social order was obvious and unchanging. It is as if events
and the historical process o f which they are part moved past or
around agrarian arrangem ents, just as in the charm ing Indian
tradition o f the peasant who worked his field while a battle raged
nearby because he knew that warriors were obliged, by their dharma,
to respect the cultivated field! Similarly, authors have seen fit to
treat with all agrarian m atters in a chapter usually entitled ‘land
revenue’ in which it is often stated that the ruler or dynasty under
study obviously recognized the im portance of agriculture and took
considerable care to follow injunctions about the proportion of
revenue to be collected and the compassionate remission of revenue
during difficult times. At the same time, however, the ‘state’
or the ‘king’ is supposed to have collected revenue from microsco­
pically small plots!
The naïveté o f this conception o f the relationship between land
and society has been responsible for much poor history.. To reduce
agrarian relations to a m atter of land revenue and then simply to
1 M orris D. M orris and Burton Stein, ‘The Economic History o f India', Journal
o f Economic History, v. 21, June, 1961, pp. 179-201, provides a general critique of
this characteristic of Indian historiography.
inventory a mélange of taxes, as is often done, is to dismiss a whole
range of vital questions pertaining to the material basis of medieval
South Indian life. To be sure, these questions are not necessarily
the m ost im portant with which an historian can occupy himself.
However, the failure in the existing historiography on early South
India to seriously address the m atter of agrarian organization
has had two obvious consequences. First, it is not possible from
the existing writing to form ulate a progressive or cumulative under­
standing o f the South Indian agrarian order with reference to
which more recent developments could be understood. Second,
and obviously, without serious consideration to the agrarian
arrangements of the medieval age, many other aspects of medieval
society — m ost notably political arrangements — have been poorly
analysed.
The present study attem pts to remedy these deficiencies. It also
attem pts to provide the foundation for a broad, new interpretation
of medieval South Indian history based upon a central concern
with peasant society and culture. The spatial context of the study
is first considered and criteria set forth according to which the
‘South Indian region’ is defined. C hapter II is a speculative essay
on the origins o f the peasant agrarian system prior to the Chola
period. The next three chapters take up salient institutions of
peasant society from the tenth to the thirteenth centuries: the
basic ethnic-ecological unit, the nâdu; the cultural and agrarian
significance of Brahm an settlements, brahmadëya ; and the bifurca­
tion o f middle and low ranked social groups into the dual division
of ‘right-hand’ and ‘left-hand’ peoples, valangai and idangai.
C hapter VI outlines a set o f significant changes which can be
identified in the late Chola period of the twelfth and thirteenth
centuries. These changes and the processes which generated them
provide the basis for the society and culture of the Vijayanagara
period, the fourteenth to the sixteenth century; this latter period is
discussed in the final chapter, C hapter X III. The nature o f the Chola
state occupies a lengthy C hapter VII. The conception o f the
Chola state and its successors as ‘segmentary’ is offered as an alter­
native model o f political organization to that which presently
exists because it more satisfactorily orders the existing evidence on
governance and because it incorporates a theory of political orga­
nization which is consonant with the kind of peasant society
presented in the evidence on medieval South India.
 The central problem of this study is that of explicating the
agrarian basis o f South India during the Chola age and later, it
has already been stated that a broader concern had originally
informed this study. The outline of this original conception has
been published elsewhere.2 In justification o f the present, more
limited, study, it may be said that the Chola period is the necessary
starting place for any longitudinal interest in the agrarian system
of the Tamil plain and its extensions into the interior upland of
m odern Salem, Coimbatore, portions o f the Arcots, and major
portions o f the K arnataka plateau and the A ndhra plain. From
the ninth to the twelfth centuries, the foundation of a system of
agrarian organization was established which appears to have
endured through m ost o f the subsequent career of the agrarian
order to the nineteenth century.
The duration and spatial scope o f the C hola state ranks it as
one of the great states o f India’s pre-m odern age. Moreover,
thousands o f stone and metal inscriptions o f the Choias have been
noticed in the annual reports o f the Epigraphical D epartm ent o f the
Governm ents o f M adras and India, and those o f several princely
states. The texts o f about 3,500 Tamil inscriptions of the Chola
period have been published. Possibly half of these inscriptions
deal with land arrangem ents of some sort. In the sheer number
of historical inscriptions dealing with agrarian arrangements, the
several dynasties of medieval South India would — and do —
assume m ajor importance. But one goes beyond this to observe,
finally, that during the C hola age we are afforded the first view
— though it is not always a clear view — of how wealthy and
powerful peasants, Brahm ans, great chiefs and kings, and to some
extent those dependent upon these leaders of the society o f the
time, shaped a highly variegated landscape to their distinctive
purposes. A nd the arrangem ents established regarding land
during the Chola period persisted into the m odem age notwith­
standing political, social, and cultural developments which trans­
formed many crucial aspects o f South Indian life.
The interaction o f Brahm ans and localized peasant folk consti­
tuted the prim ary cultural nexus o f medieval South Indian peasant
society. M ythologically and ideologically, Brahm ans traced their
- ‘integration of the Agrarian System of South India' in Robert E. Frykenberg
(ed.). Land Control and Social Structure in Indian History, University of Wisconsin
Press, Madison, 1969, pp. 175-217.
origins to the pure Gangetic land o f Aryavarta. Gotra names
preserve this element o f spatially pure substantiality. The research
of m odern scholars, m ost o f whom have either been Brahmans or
Europeans partly disposed to Brahm anical views, has tended to
accept the m ythology and ideology o f Brahmans as strangers —
outsiders — to South India. At least that has been true until
quite recently when it became impolitic in South India to make
such claims. However, except in the sense that most peninsular
peoples are descendants o f an archaic m igratory folk who made
their way into the peninsula, there is no validity to the claim o f a
prim ordial ethnic separateness of Brahm ans from m ost others of
the southern peninsula. In fact, non-Brahm an castes of the twen­
tieth century often claim an equally remote origin. Thus, during
the medieval period and subsequently, Brahmans are best seen
as peninsular folk who, in the m anner of Brahmans everywhere
in India, m aintained and transm itted some portion o f the totality
of esoteric and dharmic knowledge which aggregatively is the
Brahm an tradition in India.
Certain portions of th at total tradition have always been regarded
as uniquely southern, such as the White Yajur Veda school and
what is perhaps the most significant o f the intellectual and ritual
elaborations of the Brahm anical tradition — e.g., Vedanta and the
Panchatantra agamic temple tradition. Medieval dharm a texts
also recognize that the Brahm ans o f southern India differ from
northern Brahm ans in some o f their customs. This, in a somewhat
facile way, is seen as a consequence o f cultural interaction; e.g.,
Brahmanical cultural traditions are seen to have transformed
Dravidian ‘folk’ culture. But, ‘folk’ is hardly an appropriate
modifier to apply to the often great elegance and depth of the
Dravidian culture revealed in Classical Tamil poetry. According
to this earliest evidence, both traditions modified each other.
When the relationship between peasant carriers of the ‘folk’
cultural traditions o f South India and the Brahm an carriers of
something called the ‘great tradition’ o f India is considered, it is
an inescapable conclusion that the distinction is wholly cultural:
it is ideas, n o t distinctive origins; it is principles, not ethnicity,
which defines the nature of the interaction of the ‘great’ and ‘folk’
traditions of South India.
N or can the relationship between what tends to be regarded as
Dravidian ‘folk’ traditions and the Brahmanical ‘great’ tradition
be dichotomized in the structural categories ‘u rb a n -ru ra l’. That
has been touched upon above. Brahm ans are no m ore people
o f cities than non-Brahm ans are carriers o f the folk tradition. In
any case, cities — distinctively urban socio-cultural contexts —
are relatively unim portant in ancient India at least from the deep
antiquity o f H arappan times. D. D. Kosam bi spoke o f ‘rustic’
Brahm ans o f the post-G upta age in northern India in the sense of
being a m etam orphosis o f an earlier urban condition.3 Even
granting this conception o f a fundam ental disjunction in India’s
ancient historical period — the decline of its cities — by medieval
times it is no longer possible to m ake the kind of distinction which
occurs in m any other peasant societies, and is considered to be
generic. The city for K roeber and Redfield and others who have
discussed the cultural aspect of peasant societies has been seen as
a necessary pre-condition o f peasant society; that other ‘p art’
to which peasant society and culture are linked. The low order of
im portance o f urban areas in medieval South India has implica­
tions beyond the cultural aspects o f peasant society, o f course. The
political system and the economic order took a different shape in
this non-urban society. B ut it is especially significant with respect
to the cultural interaction o f peasants and others that the principal
custodians of the Sanskritic learned tradition in South India lived in
rural settlements with peasants. N ot only were Brahm ans ethni­
cally indistinguishable from their peasant contem poraries, they
were their neighbours.
An inevitable outcom e o f millenia of social and cultural pro­
pinquity and interaction o f Sanskritic and non-Sanskritic cultures
in South Indian rural settlements was a single cultural system.
Until the thirteenth century perhaps, there were still attem pts to
segregate certain cultural elements as Sanskritic and others as
appropriate to the Dravidian language traditions of peninsular
India. However, there could be no effective segregation of either
the language or o f specific traditions conveyed in these languages.
From late Classical times, the interaction of ‘folk’ and ‘high’
traditions (sometimes stated as desi and mdrga respectively) was
evident-as a reciprocal process involving mutual transfers.4 A t­
3 D .D . Kosam bi, ‘The Basis o f Ancient Indian H istory’, Journal o f the American
Oriental Society, 75, no. 4, Oct.-Dec. 1955, p. 236.
4 See V. R aghavan, 'Variety and Integration in the Pattern of Indian Culture',
Far Eastern Quarterly (Journal o f Asian Studies), 15, no. 2, August 1956, p. 504.
tempts in the twelfth and thirteenth centuries to m aintain a semb­
lance of cultural com partm entalization within what had become a
single cultural system resulted in the crises of the Saiva Siddhanta
movement and the sectarian divisions among Vishnu worshippers.
While nominally theological, the issues in these sectarian crises
included the appropriateness — indeed the completeness — of
Tamil as a liturgical vehicle and the participation of non-Brahman
votaries as im portant ritual functionaries and teachers. After
the thirteenth century, the notion that Sanskrit and its Brahman
adepts were to be the sole custodians of a high tradition while
Tamil, or some other Dravidian language, and non-Brahmans
were to be the custodians of folk traditions could no longer be
explicitly sustained.
The shared cultural systems of medieval South India may not
have included all who lived in the southern peninsula, but it did
encompass all who lived in the peasant settlements of the major
agrarian tracts of the southern peninsula and therefore included
Brahmans, non-Brahm an cultivating groups, artisans and traders.
Even landless, agrestic labourers in peasant localities shared in
the general cultural system, if only negatively in suffering the
opprobrium of inferior, ‘polluted’ status, a condition which was
acceptable then as it has been until recently in exchange for the
survival benefits of being part of a relatively prosperous agrarian
economy.
A final element in the relationship between Brahmans, as the
transm itters of high cultural traditions, and the non-Brahman
peasantry, upon whose support Brahmans depended and with
whom Brahm ans lived as neighbours, is political and ideological.
In the spatially circumscribed world of the South Indian peasantry,
as in the locality systems of peasants everywhere, great kingly
authority and the ‘state’ may be thought to be remote. This re­
moteness would not be seen as a flaw, an imperfection, but a predic­
table feature in a pre-m odern political system lacking the technology
to close great distances.
Such an implication or conclusion is explicitly denied here.
As the title of this study announces, there is something called
‘the peasant state’ however dissonant such a conception may appear
to modern scholars of peasant societies. Two features — indeed,
defining attributes — of peasant societies for most scholars pre­
suppose a conceptual and empirical separation of peasants from
those who rule them. One of these features is scale. Peasant
societies consist of small, local, often isolated communities; they
are physically as well as culturally distant from the ‘m ajor’, ‘high’,
or ‘im portant’ institutions, hence, they are as K roeber proposed,
part-societies and part-cultures. The other feature is explicitly
political, that is peasants occupy an ‘underdog’ position with
respect to the more powerful and dom inant. Thus, Fallers speaks
of a peasant society as ‘one whose prim ary constituent units are
semi-autonomous local communities with sem i-autonom ous cul­
tures’.5 The qualified autonom y o f peasant produces salient
boundary characteristics such as to define ‘insiders’ and ‘out­
siders’ :
On the one hand there is the local community . . . [hostile to outsiders,
sharing rights in land, submitting to and supporting institutions of local
government and social control] and on the other hand there is a hierarchy of
patrimonial or feudal relations of personal superiority and responsibility. . .
and subordinate dependence which link the local community and the wider
polity.6
In the light o f these perceived features o f peasant societies, the
idea o f a peasant state is at least paradoxical, if it is no t contra­
dictory. However, this is what the state in medieval South India
appears to have been.
The medieval South Indian state is here considered a ‘segmentary
state’ in which political authority and control were local in several
crucial ways. The scope of the constituent units of the state was
limited to well-defined and persistent ethnic territories; its chiefs
were, in most cases, leaders or spokesmen o f the dom inant ethnic
groups of the local territory; and, perhaps most unique in South
India, corporate bodies representing the interests of various folk
of a locality participated in the public business of the locality.
Peasant localities in medieval South India were not uniform,
varying in their com position, wealth, and elaborateness in accord­
ance with constraints or opportunities presented by nature to the
agrarian potentialities o f the age. W hen and where there was
reliable moisture for irrigated agriculture, the largest concentra­
tions of population were found, the greatest wealth, the most
social and cultural elaborateness. There, too, were the settlements
5 'Are African Cultivators to be Called “ Peasants” ?', Current Anthropology,
v. 2, no. 1, April, 1961, p. 108.
6 Ibid., p. 109.
of Brahmans. Localities less favoured by nature, had less. How­
ever, arching over the resource variation among localities given
by nature and hardened into valued practices and loyalties, was a
Brahman paradigm, a framework for the attainm ent of appropriate
society. The most authoritative interpreters of this ideological
framework as well as the ‘gatekeepers’ to access to status and
authority were Brahmans. Power, order, legitimacy within the
locality world of the South Indian peasantry were mediated by
Brahmans. This was m ore true in those localities o f wealth and
prosperous agriculture and trade than in those peripheral tracts
where people survived on hazardous, rain-dependent agriculture.
But, to some extent, it was true everywhere there was sedentary
agriculture.
The culture of caste as an element in the character of Indian
peasant societies does not end with the relationship of Brahmans
and non-Brahm an peasants. Caste principles pertain to kinship,
marriage, and occupation as well as to m atters directly involving
Brahmans. In South India, the relative isolation, or autonomy,
of peasant localities was reinforced by kinship and marriage
practices. The spatially restricting preference for cross-cousin
or m aternal uncle-niece m arriage has been a well-recognized
difference between N orth and South Indian m arriage systems.
In the former case, with its different consanguinity rules, affinal
networks spread over a wider area. The factor of difference between
the m arriage networks o f N orth and South India, in terms of this
spatial question, has been estimated as great as ten in recent times.7
In this, caste, m arriage and kinship codes can be seen as supporting
one o f the generic characteristics o f South Indian peasant society
— its localness and autonom y.
However, caste principles may be considered as antagonistic
to other presumed attributes o f peasant society. One is the egali­
tarianism o f the peasant community. F or some scholars who have
worked on peasant societies, the asymmetry which characterizes
the relationship of the peasant community and the powerful
outsiders to whom went a portion of peasant production contrasted
with the presumed egalitarian and cooperative ethic o f the peasant
7 M .N . Srinivas and A.M . Shah state that the m arriage field in South India for
rural castes is about twenty or thirty villages whereas that o f N orth India is 200 to
300 villages. ‘The M yth of the Self-Sufficiency of the Indian Village’ in The
Economic Weekly, 10 September 1960, p. 1376.
community. The contrast is easily overdrawn, resulting in a failure,
often, to recognize stratified relationships within many peasant
societies as well as giving expression to an idealization of social
harmony and human decency which is presumably endangered,
even destroyed, by the encroachment upon peasant societies of
archaic bureaucratic forms or the social, economic, and psycho­
logical disruption of m odern industrialized society.
Clearly, peasant societies vary in the degree to which internal
stratified relationships exist, but in most, and perhaps all, there
appears to be some of this. Caste principles of hierarchy and
inequality while contributing an im portant specific quality to
Indian peasant society cannot be thought to raise serious doubts
about whether this was a peasant society. To be sure, there is more
to the distinctiveness o f the Indian variant o f peasant societies
than the explicitness of its rules of hierarchic social relations. The
vulnerability of low caste landless labourers and many artisans
to crushing status deprivation and to economic exploitation
are real enough whether one accepts wholly or partly the recent
literature on the ‘Hindu jajmani system’ and some studies of
contem porary Indian village life8 or whether one deals with the
often fragmentary evidence of the past. However, even in India
where caste codes often enjoin rigorous forms of segregation and
ranking within a peasant locality, the exigent requirements of
peasant life generate in India, as elsewhere, forms of economic,
social, ritual, and political cooperation and interdependence
among the diverse parts of a peasant community. Until there is
more systematic comparative research on peasant societies and
better means for assessing relative degrees of inequality within
peasant societies, it may be taken that Indian peasant societies
are not essentially aberrant in the inequality of their internal
constitutions; these peasant societies fit within what must be,
and perhaps one day will be, a dimensional schema along with
other features of peasant societies in the world.
8 General discussions: T.O. Beidelman, A Comparative Analysis o f the Jajmani
System, Association of Asian Studies, M onograph, no. 8, Locust Valley. N.Y.,
1959; Louis D um ont, Homo Hierarchicus : The Caste System and its Implications,
Weidenfeld & Nicolson, London, 1970, ch. IV, pp. 92-109. Village studies: McKim
M arriott, Village India: Studies in the Little Community, University of Chicago
Press, Chicago, 1955; M.N. Srinivas, India's Villages, Asia Publishing House,
London. 1960; A ndré Béteille, Studies in Agrarian Social Structure, OU P, Delhi,
1974.
 A nother presumed generic feature of peasant societies which
may be considered to test the degree to which caste principles
can be reconciled with most definitions of this sociological type
is the ‘multi-dimensionality’ of the peasant household. What
Shanin and others appear to mean by this is that the peasant house­
hold is not simply, nor solely, involved in agricultural production
though this is, to be sure, its m ost im portant livelihood function
in the local as well as the general economy. The peasant household
in most places also produces many of its own tools and household
implements; it produces handicraft goods for its own consumption
and for exchange, and in other productive ways creates around the
family farm a high degree o f m aterial self-sufficiency. Correlated
with this ‘multi-dim ensional’ aspect of the generic peasant house­
hold is the presum ption of a long apprenticeship which younger
members of the peasant family m ust undergo under the tutelage
of their elders. Among other consequences of this lengthy training
period, generational divisions — the emphasis upon seniority
— are strengthened. W hat, then, is to be made o f the division of
labour m aintained in Indian caste codes?
Classically stated, caste culture segregates and ranks (at times
in an inexplicably arbitrary m anner) occupational specialists in
the finally graded m anner characteristic o f Indian schema of other
kinds. M oreover, during the medieval period of South Indian
history, an agrarian settlement was considered a ‘proper’ village
only when there was a full complement of occupational specialists,
‘the eighteen castes’. These included not only those service castes
performing pollution related tasks, such as the washerman, the
barber and, of course, the sweeper, but also scribal functionaries
who actually enjoyed relatively high ritual rank; there was also
a set of persons whose skills and productive functions were obviously
im portant to a peasant settlement, e.g. potters, smiths, carpenters,
and weavers.
The contrast of the Indian caste division of functions and the
presumed ‘m ulti-dim ensional’, or non-specialized, character of
peasant households generally is easily overstated. M ost peasant
communities everywhere in the world displayed some or con­
siderable specialization among peasant households, and the
presumed ‘m ulti-dim ensionality’ o f the peasant household in
most places differed only in degree from the complex, code-en­
joined occupational system in India. The English peasantry of
the thirteenth century is estimated to have been virtually landless
to the extent of one-quarter to one-third of its number, and a large
num ber were non-agricultural specialists.9 It is in this sense of
specialization within a general framework o f a peasant society
that Firth and others have called fishermen and merchants, ‘pea­
sants’. There is reason to assert, too, that the occupational com­
pleteness of caste rules was not realized in many cases. Pollution
norms required that high caste persons have access to washermen,
barbers, and others whose services were necessary for the m ainten­
ance of the pure state required for those of high ritual status. But,
beyond those few services, the prescribed array o f occupational
specializations was probably seldom realized. N or was this
necessarily a breach of code injunctions. Even Brahmans were
permitted to cultivate, if necessary, and where other specialists
were not available — as must have occurred — a condition of
necessity would have been considered to exist for Brahmans as
well as for ritually less fastidious non-Brahmans. A corresponding
relaxation of occupational restrictions can have been anticipated.
As in many aspects of caste culture, the fully elaborated codes
are to be understood as the logical and optimum condition, not the
existential condition under which Brahmans, especially, but others
too, were supposed to live.
For the historian, and particularly the Asian or Indian historian,
the concept of ‘peasant society’ is of the greatest significance.
Beneath the level of vast empires and great cities in Asia and India,
with their patent vicissitudes and often exaggerated competence
to control enormous territories, rural cultivators and those associa­
ted with them have historically accounted for the overwhelming
mass of the population. They have provided the basic wealth
and m anpower which sustained civilized institutions. This remains,
perhaps dismayingly, true in the present. F o r many o f the problems
with which historians concern themselves, it is no longer possible
to ignore these m ajor parts of polities and societies nor, what
am ounts to the same thing, to assume that the rural, peasant world
was inert, passive, and unchanging. C ontem porary social science

9 E.A. Kosminsky, Studies in the Agrarian History o f England in the Thirteenth

Century, Blackwell, Oxford, 1956, p.209. Kosminsky observes the ‘very abundant
and variety of names indicating non-agricultural work reminds us forcibly that
our sources [their exclusive concern with cultivation] . . . give a far from full and
very one-sided picture o f the English village.’ (p. 229.)
research as well as increasingly sophisticated historical studies in
India and elsewhere make this impossible. The essence of the
peasant condition is and has ever been constant and significant
interaction with other parts of the social order.
Historical studies of agrarian arrangements in India are perhaps
am ong the earliest o f the ‘modern scientific history’ undertaken
in Asia. The first serious and systematic studies o f agrarian relations
and problem s occurred over a century ago when British adminis­
trators, bent upon better understanding, or perhaps simply justify­
ing, revenue policies in British Indian territories, began to explore
the development of land relations, tenure, and related m atters of
the past. These studies were published as bureaucratic reports
and minutes, as personal memoirs or, as in M adras, as parts of the
first district m anuals in the late nineteenth century. These were
later utilized by such scholar-adm inistrators as B. H. Baden-Powell
in his m onum ental, three-volume Land Systems o f British India.1®
W ithout denigrating these pioneer efforts which sought to make
sense of what appeared to many officials to be chaos, even during
the late nineteenth century, the results of their inquiries led as often
to a distortion as to a clarification of historical agrarian relation­
ships. R ationalizing existing practices, principles, and policies in
the face o f m ounting criticism from a small group o f educated
Indians as well as from critics in England could scarcely have led
to the best analysis of historical landed relationships. Yet, much
of the past record o f these relationships was clarified and, without
doubt, some o f the m ost felicitous m om ents in the career o f the
British bureaucracy of nineteenth century India may have been
when some few Englishmen screened fragmentary agrarian records
through their intelligent and often sympathetic experience.
Studies of historical agrarian relations — narrowly by British
adm inistrator-scholars and somewhat more broadly by historians
of India — have inevitably been strongly coupled with land revenue
organization. British adm inistrators, concerned with revenue often
above all else, created this stultifying legacy, and it has been per­
petuated by nationalist Indian historians concerned with attacking
British policies. Revenue is but one aspect o f agrarian organiza­
tion, but in the absence o f effective, bureaucratic government — a
condition not attained in India until the nineteenth century —
10 Published in 3 vols, OU P, London, 1892.
revenue transfers beyond quite restricted localities was an un­
im portant m atter. Even under the Mughals, the m ost effective
pre-modern government in India, regular and routine assessment
and collection o f land revenue for remittance to the ‘state’ was
limited in a variety o f ways. However, in contrast to revenue m at­
ters, the broad historical relationship between agrarian organiza­
tion and peasant society was a central one until very recent times
in India. This is so true that it may be wondered that with the
considerable interest in agrarian studies involving revenue, land
tenure, agricultural trade and debt, scholars should not have
engaged m ore directly the web o f peasant social and cultural
relations with the land.
In part, at least, the reason for the neglect m ust be recognized to
stem from the vagueness, the potential omnicompetence, of the
concepts ‘peasant society’ and ‘agrarian system’. Real as cows,
manure, m arket places and prices, tax collectors, and cultivators
themselves may have been, the ways of analysing these and other
agrarian m atters are artifacts of the analyst. The system of agrarian
relationships examined by the economist concerned with the pro­
duction of a single unit of agricultural production, a field or farm,
is different from that of the economist concerned with the output
of all grain producers; and different from these two are the rela­
tionships examined by the rural sociologist interested in voluntary
associations among rural families or an agronomist concerned with
plant diseases. Each will have a conception o f agrarian relation­
ships appropriate to his inquiry; each will examine a portion of
the totality.
Daniel Thorner writing on Indian land reform problems in
The Agrarian Prospect in India states that, ‘the agrarian structure
is, after all, not an external framework within which various classes
function, but rather it is the sum total o f ways in which each group
operates in relation to other groups’. Notwithstanding this com­
prehensive definition, Thorner is concerned with only a part o f the
totality o f relations he speaks of, namely those related to the politics
o f contem porary land reform .11 The problem is not restricted to
India. The Soviet scholar, E. A. Kosminsky, in his Studies in the
Agrarian History o f England in the Thirteenth Century dealing
with the m anorial system, could set aside issues obviously connected
11 Daniel T horner, The Agrarian Prospect in India, University Press, Delhi, 1959.
with his problem — agricultural techniques, climate, soils, and the
relationship o f livestock to cultivation — on the basis that they
did not bear centrally upon his topic of investigation.12 And so it
must be for any investigator; the choice of concepts and the mean­
ings which are given to them must be suited to the problem at hand
and the requirements of the clearest possible exposition of findings.
The idea o f the agrarian system in the present work focuses upon
peasant society and culture in relationship to land. Along these
lines, work on the M ughal period is the best in the small class of
Indian historical studies dealing with historical agrarian rela­
tionships. By far the best is that of Irfan Habib, The Agrarian
System o f Mughal India ( 1556-1707). One of the group of able
scholars at Aligarh Muslim University, H abib sought to critically
examine ‘not only [the]. . . land revenue adm inistration. . . but
also [the] agrarian economy and social structure’.13 Though he
corrects and amends earlier work on this period, notably that of
W. H. M oreland’s Agrarian System o f Moslem India,14 H abib is
essentially concerned with the M ughal power structure rather than
the relationship o f land and land control to the economy and social
structure o f the time. Appropriately however, he discusses dif­
ferent groups o f the landed population in N orth India upon which
the century and a half o f M ughal supremacy was based : an oppress­
ed peasantry, various kinds o f local intermediaries, especially zamin-
dars, and the M uslim w arrior élite. M oreland’s work, for all its
pioneering significance, was more narrowly focused than H abib’s,
being a discussion of the importance of revenue assignees upon
whom the main burden of agrarian control rested in all but a few
years o f the long period o f Muslim dom ination over the North-
Indian countryside. The idea o f the agrarian system within which
both Habib and M oreland frame their inquiries reflect their con­
centration upon political analysis. In both cases, the study o f the
agrarian system was essential for understanding power relations
in this best-organized and m ost powerful o f the states o f medieval
India. And, if H abib is correctly understood here, his argument
is that the demise of the M ughal state in the early eighteenth century

12 Op. cit.. p. viii.

15 Irfan Habib, The Agrarian System o f Mughal India, Asia Publishing House,
London, 1963. N ote changes in his argum ent in subsequent publications.
14 W .H. M oreland, The Agrarian System o f Moslem India, W. Heffer & Sons.-
Cambridge, 1929.
was the result o f a crisis in agrarian relations in which a substantial
portion o f the peasantry, alienated by Mughal officials, threw in
its lot with pre-M ughal locality leaders.15
Peasant agrarian relationships are aspects of social and cultural
systems; they are hum an adaptations to the natural environment
within a social and cultural framework. In the Indian sub-con­
tinent alone, one must speak of a very large num ber o f historical
peasant societies and cultures over India’s long history. This is to
underscore the truism o f Indian diversity from an early time.
Differences in languages and cultural traditions as well as in the
patterns of social relationships among Indians combined with the
differential constraints and opportunities afforded by the variety of
environmental settings of the sub-continent as a whole or any of
its sections — all of these would have contributed to the great
mosaic of cultural regions of medieval times and later. Among these
one would expect to find different patterns of development and
change resulting from the quite uniqu" historical courses through
which each went. However, this remains a largely theoretical
proposition, since only the most preliminary work has been done
on the classification of agrarian systems of modern India and even
less on historical agrarian systems.16 The latter task awaits two
scholarly developments. The first of these is more research on the
history o f the middle period, especially social and economic history
of the sort which H abib has done for the Mughal period. But of
equal, if not prior, importance, thought must be given to the for­
m ulation for all or most of the sub-continent of those social and
cultural factors which, along with ecological factors, shape agrarian
systems whether contem porary or historical.
The m ajor factors are generally recognized, and some have
already been briefly touched upon here. In economic terms, the
salient feature is cultivation of the soil and occupations related
to that — animal husbandry, artisanship to provide the basic
implements for agriculture and for peasant household use as well
15 H abib, op. cit.
16 Notice should be taken o f the w ork o f D r Chen Han-Seng provisionally en­
titled ‘Agrarian Regions of India and Pakistan’ and based upon the R eport of the
Royal Commission on Agriculture in India. Though forty-five years out o f date,
the addition of some more recent evidence and the proposed publication of a sepa­
rate volume of m aps m ake this project, being carried out in cooperation with the
École Pratique des Hautes Etudes (Sorbonne), im portant.
as to provide goods for more general exchange. These attributes
are shared by other kinds of economic orders — primitive ones,
archaic collectivist ones, and some m odem ones, hence it is difficult
to speak of a ‘peasant economics’ as distinctive. Some scholars
have, in fact, sought to differentiate significant aspects of peasant
economic operations.17 The difficulty of delineating something
called ‘peasant economics’ arises from the large and diverse num ber
of peasant relations, past and present, of which we have often
disturbingly fragm entary evidence. In part, too, the difficulty
arises from the fact that ‘peasant’ and ‘peasantry’ are not essentially
economic, but sociological categories. It is the nature of his social
relationships, where he lives and with whom he interacts, that most
authoritatively identifies the peasant. N otwithstanding these pro­
blems of definition, it is possible to speak of peasant economic
orders as those in which the m ajor forms of wealth are generated
from cultivation of the land by family-sized units and primarily
for their subsistence and where industrial and service activities
are relatively insignificant except as an adjunct to the cultivation
of land within circumscribed localities.
How land is m ade to yield this wealth varies in South India.
It may come from swidden tracts o f relatively sparse population,
or from intensely irrigated agriculture in relatively densely popula­
ted tracts, or some interm ediate varian t; this is culturally determined
within the possibilities of the natural setting of particular places
and times. How wealth, or ‘surplus’, is apportioned among culti­
vators, am ong groups of non-cultivating dependents, and between
cultivators and those who exercise political power over them are
critical m atters for those of the society as well as for those who wish
to understand that society. In economic terms, peasants are not
easily distinguished from others for whom land is the m ajor source
of wealth. Always and everywhere, peasants, like any producers
must deploy a portion o f their production to m aintain and replace
their population, tools, houses, seed, and livestock in order to
17 The critical works on this question of whether it is necessary to consider a
special m ode o f analysis of peasant economics remain: A lexander Chayanov,
The Theory o f Peasant Economy, ed. Daniel T horner, tr. from the Russian by
R.E.F. Smith, American Econom ic Association, Hom ewood, 1966; K aroly Polanyi,
et al., Trade and M arkets in Early Empires (Free Press, Glencoe, 1957) which argue
the affirmative; contra: R aym ond Firth and Basil S. Yamey, Capital, Saving, and
Credit in Peasant Societies, Aldine Press, Chicago, 1965. See M arshal Sahlins,
Stone-Age Economics, Aldine, Chicago, 1972.
survive. This is essentially a family responsibility in a peasant
society and may include several households within a peasant settle­
ment. Similarly, peasants must, like others who live in communi­
ties, invest some portion of their time, effort, and wealth in the
continuation o f the social collectivity o f which they are part. This
contribution appears to be a composite one o f personal participa­
tion and a transfer o f family wealth. Finally, peasants, like others,
must make their terms with those whose superior power commands
tribute or taxes, rent or interest. This may be done by the individual
peasant household or, perhaps more often, by the peasant commu­
nity as a whole. In a peasant society, the recipients of this payment
are often themselves local figures with whom some degree of face-
to-face interaction exists even if, as in some cases, that local person­
age is an agent for or beneficiary o f some less proxim ate authority.
W hat distinguishes the peasant solution to the problems of m ain­
tenance and replacement of productive capability, of the preserva­
tion o f the community, and of meeting the share of the demands
from external authority is that the scope of these activities is highly
localized. In consequence, one o f the m ost serious and characteristic
problems for the peasant community is that of balancing its internal,
local requirements against the external demands o f powerful
warriors, agents of a powerful state, or members of protected
mercantile groups.18
External liens upon peasant production are the consequence of
asymmetrical power relationships, and for some analysts, or for
some problems, the terms o f power between the peasantry and the
external world constitute the most im portant characteristic of
peasant society. W ithout denying the im portance of this condition
under which all peasants necessarily live, undue stress upon exo­
genous pressure has led to the distortion of two other aspects of
the political condition of the peasantry. The first is the degree
to which peasants manage m ost o f their own community problems
without external interference. Thus, onerous, and at times grati­
fying, and always im portant as the relations with powerful out­
siders often are for its welfare, and ultimately for its survival,
the peasant community can and does retain control over m ost of the
essential activities upon which its society depends. And, except
in the m ost effectively organized state system — a condition which
ls These issues are clearly discussed in Eric W olf, Peasants, Foundation of
M odem A nthropology Series, Prentice Hall, Englewood Cliffs, N.J., 1966, p. 15,
is only infrequently achieved until very recent times — the residual
competence o f the peasant com munity ensures that the relationship
between the peasant locality and those powerful enough to demand
a portion o f its wealth does not leave the peasantry entirely disarm­
ed. This is the second, often underestim ated, aspect o f the political
context o f the peasant. Through coalitions o f various kinds, the
ability of peasants to cope with the many kinds of externally power­
ful can be formidable. The ultim ate weapon of the peasant is also
his m ost difficult to use. T hat is, to abscond, to desert the lands and
jurisdiction o f an overlord whose demands have come to exceed
the benefits o f the protection o f his overlordship. R esort may be
taken with other locality overlords under more promising condi­
tions, or relief may be sought in forest and hill tracts with all of
their menaces. Desertion o f established homesteads and lands,
however modestly endowed these may be, can never be lightly
undertaken, of course, nor can the other means which peasants
may have for countering excessive demands. Yet, in South India
there is early and continuous evidence of peasant movement of
which only part can be attributed to political causes. Outside
the deltaic regions, cultivating folk appear frequently to have
moved, often great distances, to improve their prospects. The
actual or potential mobility of peasants for whatever reason,
would compel any great or small overlord to come realistically to
term s with the requirem ents o f the peasant community. Thus,
while the asymmetrical power relationship remains an essential
aspect o f peasant society, it m ust be seen as qualified, as a balancing
of advantage and disadvantage between the peasant community
and the external world.
A ttachm ent to some narrow locality is a, and for some purposes,
the salient characteristic o f the peasant social system. This is at
once a source o f strength and weakness in economic and political
terms. The web o f relationships beginning with those within the
peasant family and extending, through kinship groupings and
coalitions, well beyond the peasant village constitutes strength
because it provides the peasant with a durable nexus for cooperation
and resource pooling at critical times. W hen there is personal
m isfortune or com m unity danger, these relationships may be
relied upon to blunt the full im pact of misfortune upon any single
family. The assurance provided by this interwoven texture of
social, economic, and ceremonial relationships o f the peasant
locality exacts a price, of course. Part o f the labour and produce
o f each household m ust be expended in m aintaining these relation­
ships; individual and family decisions regarding welfare would
weigh internal household requirements against outstanding obliga­
tions to others within their large social universe.
Studies of such a complex universe may, perhaps m ust, con­
centrate on only a part o f the network of significant relationships
within the peasant community or on the nature and dynamics of
linkages between the peasant com m unity and other parts o f the
social and cultural order. According to the question being resear­
ched, stress can be placed upon any viable aspect o f the totality
o f relationships which comprise peasant society, where the test of
viability is the m ost plausible explanation of what happened, or is
happening, and why.
Historians o f Indian culture have typically concerned themselves
— when they have addressed the problem s of the peasantry — with
the external linkages between the peasantry and the larger world
o f which it is a part. Too often, the historian’s level o f analysis has
seemed to require an assum ption about the nature o f the Indian
peasantry which, if not incorrect, reduces the probability o f under­
standing the complex and changing society o f peasants. T hat is,
the assum ption that the peasantry in any part of the sub-continent
is a single, undifferentiated, conservative (‘traditional’ is often
used) part o f the social system in an alm ost constantly defensive
or com pliant posture in relation to the aggressive world around it.
In this tw o-part model o f peasant and other, there are few distinc­
tions drawn am ong different groups comprising a peasantry at any
time and place, and there is an alm ost complete neglect o f the pro­
cesses governing the constant inter-relationships between the two
spheres. G ranted that such simplifying assumptions, or sets,
about ‘the peasantry’ are frequently required by the nature o f the
evidence, it is possible at certain times and in certain parts of the
historical Indian cultural sphere to m ake finer discriminations about
peasant life and society by attem pting to perceive the system of
relationships between peasants and others and among peasant
groups themselves, as it were, ‘from the bottom ’.
Focusing upon the peasant com m unity in its relations with
others in the social order from this level requires a m ore complex
model of peasant society than m ost historians are accustomed or,
apparently, prepared to contemplate. From the outset, it is neces-
sary to envisage, and to the extent that evidence permits, to recons­
truct a social system composed o f a variety o f groupings in a place
which is defined by a common set o f bonds o f these groupings to
some particular tract of land. T hat is the arena.19 In India, while
such groups tend to have quite different social positions, goals,
and styles, and often different contacts beyond the community,
reflecting the stratified and segmented nature of Indian social
organization, there is a strong com m itm ent to the village and
locality as in other peasant societies. Corporateness is the con­
sequence o f jo in t concern with shared problem s within the commu­
nity and w ithout; yet there may be quite different perceptions of
internal or external pressure by the various strata and segments o f a
peasant community. Each must, however, come to terms with
that pressure within the fram ework of a com m on community
response.
External pressures themselves vary. Some pressures upon the
peasant community m ay generate internal responses by the com ­
munity and leave durable internal referents as m arkers or traces of
intrusions from beyond the community. Others m ay not. Thus,
a pillaging army in a peasant locality constitutes a more formidable,
immediate threat to welfare than a small band of warriors living
permanently in some nearby hills. The former may raid villages,
take all they can, and leave. No adaptive response is likely to be
effective or even relevant. However, the small w arrior band which
cannot be adequately defended against nor extirpated m ust be
dealt with through means which satisfy the demands o f the threaten­
ing warriors w ithout simultaneously incurring excessive dislocation
within the community. Each of these dangerous confrontations
may leave its m ark: in devastated fields, dead defenders of the
village, or inadequate stock or seed against the needs o f the im­
mediate future, in the first case, and in the other case the consequen­
ces may be m ore durable such as a new or modified set of land
arrangements, a modification o f village and locality leadership,
and perhaps new village and locality names.
To deal with the complex internal structure of South Indian
peasant societies and the equally complex linkages among peasant
localities, the concept o f ‘pyramidal segm entation’ has been adop-
19 The term ‘arena’ has been used by F.G . Bailey to provide a spatial dimension
to sociological relationships. Tribe, Caste, and Nation: A Study o f Political Ac­
tivity and Political Change in Highland Orissa, OU P, Bombay, 1960, pp. 269-70.
ted in the present work. This concept has been found m ost useful
in the discussion of the political organization o f the Chola and
Vijayanagara states because o f the m anner in which political and
social elements were com bined in medieval South India.
‘Pyramidal segm entation’ refers to persistent com binatorial
patterns among social elements which are distinct and often
opposed. Such distinct elements are regarded analytically as
social ‘segments’. These elements are parts of a social whole,
which ultimately extended to the peoples of the sub-continent,
to which the term ‘Indie’ may be applied; they are also parts o f the
many cultural regions into which medieval India was partitioned.
In short, they are differentiated elements of a single, universal
moral system. In the usage of medieval dharm a texts, these seg­
ments, as generic categories, were called kula (agnatic or cognate
kinsmen), srenl (occupational groups consisting o f several castes),
and gana or puga (territorial assemblies).20 Such segments com ­
bined to m ake up the num erous local social contexts o f medieval
South India, and these segments also combined, or massed, to
form supralocal com binations, o r pyramids, hence the term ‘pyra­
midal segm entation’.
W ithin peasant localities (nadus) o f C hola country, as discussed
in C hapter III, these social segments would include various groups.
Among these were cultivators who may have settled in a locality
as subordinate or client groups of the dom inant peasant community
called nattar; other groups may have consisted o f non-cultivating
people such as herdsmen or artisan s; yet others m ight be persons
and groups assimilated into the expanding sedentary agriculture
of the time (often as dependents of established local groups) from
marginal tracts where they might have been swidden cultivators
or even hunters. M any o f such groups could occasionally combine
as massed groupings o f right and left castes (valangai and idangai),
an im portant feature of supralocal social organization in many
parts of.South India from about the eleventh century; this is dis­
cussed in C hapter V. The m ost significant aspect of the massing
of primary local segments as supralocal formations was politi­
cal; it is the potential massing of local segments that provided the
foundation of states in medieval South India,
20 See the discussion of these term s and their variant meanings in P.V. Kane,
History o f Dharmasastra (Ancient and Medieval Religious and Civil 'Law), 2nd,
ed., B handarkar Oriental Research Institute, Poona, 1974, v. 3, pp. 280-i.
 The Chola and Vijayanagara states of medieval South India
are seen as ‘segmentary states’ in the discussion of chapters seven
and eight. This designation is adopted from a conception of politi­
cal organization most comprehensively dealt with by the African
anthropologist Aidan Southall.21 Seen by many scholars as ‘inci­
pient states’ or ‘proto states’, Southall proposes th at segmentary
states comprise a m ajor type o f political organization not only in
Africa, where his work was done and where the concept o f the
segmentary state has greatest currency, but in other cultural
contexts as well: European and Asian. In the segmentary state,
the parts or segments o f which the state is composed are seen as
prior to the formal state; these segments are structurally as well as
morally coherent units in themselves. Together, these parts or
segments comprise a state in their recognition o f a sacred ruler
whose overlordship is of a m oral sort and is expressed in an essen­
tially ritual idiom. It is precisely in such terms that South Indian
kings are best perceived, and it is for this reason that the concept
o f segmentary state so commends itself for the present analysis.
M ost scholars o f Indian social structure have acknowledged the
segmentary character o f localized Indian societies just as they
have acknowledged the profound im print of peasant institutions.
But, some o f these same scholars have treated such structural
realities — of the past as well as of the present — as secondary
considerations, as epiphenomena, before the com m anding holistic
culture o f caste, of Hinduism and, during the pre-m odern era at
any rate, o f great monarchical states. Here, the language of
‘segm entary’ or ‘tribal’, o f ‘structurally relative’ (rather than
absolute) groupings in society; o f ‘complementary opposition’
or ‘pyramidal segm entation’ would seem to have no place. Here,
the culture of caste, o f purity and pollution, of hierarchy, of systems
of salvation, religions based upon highly, sophisticated textual
sources, and o f sacred kings closely identified with the great gods
of those religions would appear to make unnecessary such arcane
concepts drawn from analyses of distant Africa. And yet, this is
not true.
The theory of South Indian kingship as articulated in medieval
law texts (dharmasastra) and other literary works as well as in the
norm ative language o f medieval inscriptions, especially the often
21 A.W. Southall. Alur Society, W. Heffer, Cambridge, 1956.
lengthy inscriptional preambles, speak of sacred kingship, not of
bureaucratic or constitutional monarchy as often construed by
South Indian historians. And, concrete historical evidence of the
activities of royal figures — whether these are military (and the
king is seen as a great conqueror), or religious (and the king is
seen as the greatest o f devotees o f gods whom he protects and upon
whom he confers rich gifts), or whether in the m ore rare royal
intervention as the upholders o f law or dharm a — these are the
activities o f sacral and incorporative kingship. South Indian
kings were essentially ritual figures except in the often circums-*.
cribed core territories o f their capitals where they commanded
and managed resources and men by virtue of their compelling
coercive power (k?atra). They are the m ost im portant symbols of
the sacred, m oral order to which all men m ust belong and, as such,
theirs is a sacred and m oral authority (dharm a) beyond the limited
territory of their ksatra. Given such a conception o f kingship,*:
only a segmentary political order — one bound together alone by
the com m on allegiance o f many chiefs to a sacred centre — can
be appropriate.
A constant factor in this analysis is technology. Techniques
and devices with which South Indian peasants m anipulated their
environment in order to produce food and other valuable products
appeared to have remained unchanged over the several centuries
covered. Swamp cultivation o f rice utilizing bullocks for plough­
ing and riverine irrigation were well established in very early
South India. The greatest hydraulic works in some parts o f the
South Indian m acro region may date from before the earliest
period examined here. According to some historians, m ajor
irrigation works were constructed in the Kaveri prior to the Cholas.
The m ost impressive o f these was the G rand Anicut, seventeen
miles below m odern Tiruchirapalli, a m asonry dam over 1000
feet long, up to 60 feet thick, and 18 feet high.22 M ore technically
reliable estimates o f the date o f the G rand Anicut place it in middle
or late Chola times. It is notable that new, substantial riverine
construction here did not occur again until late in the nineteenth
century. As to ‘tanks’, the em banked reservoirs which cover a
major portion of the plain and part of the South Indian upland
tracts, British irrigation specialists of the nineteenth century
22 F.E. M organ, 'Irrigation’, Southern India, ed. Som erset Payne, Foreign and
Coionial Com piling Publication Co., 1914-15, L ondon, p. 285.
frequently commented upon the virtual completeness with which
surface irregularities had been exploited for this purpose long
before their time and found little scope for the construction of
new ones.23 The appearance of unchanging technique in the
utilization o f land — a limited set of implements and a fixed
repertoire o f crops and cultivation practices — results, in part,
from a paucity o f evidence. However, that is not entirely the case,
for there is considerable direct evidence that South Indian ‘peasant
ecotypes’ were not basically altered from perhaps the ninth to the
nineteenth century.
A peasant ecotype24 can be defined as a system consisting of
two sets o f agricultural relationships within a range o f ecological
possibilities. The first involves energy transfers am ong organic
elements including hum ans and animals and the stored energy of
crops which nourish men and animals. The second set o f relation­
ships involves inorganic agencies, or devices, which men have
used to assist in energy transfers. These include: implements,
fences, machines for raising water from wells or reservoirs, and
constructions for storing or draining water from cultivatable
fields. Organic and inorganic elements in com bination with such
complex factors as soils, topography and, m ore transitorily,
water-balances tend to become relatively persistent systems which
are related to social and cultural concom itants either assumed by
a particular ecotype or the consequence of it.
The hazards of m uch peasant agriculture invariably produce
defensive agricultural strategies which m ake it appear to be un­
changing. Proven techniques, tools, seed varieties, and timings
of certain operations are transm itted over time in songs and proverbs
so that a catalogue of agricultural proverbs can probably be com­
pared to the Sanskrit literary genre, sütra, in being an outline
of reflective and orally transm itted knowledge. An ecotype is
thus not a ‘n atural’ system, but a socio-cultural sub-system to be
understood within a larger socio-cultural context.
23 R eports of M ajor R.H . Sankey on M ysore irrigation in 1866 stated that
59.7 per cent of the princely state was under tank irrigation and estim ated that
there was one tank and one village in every square mile: cited in H erbert M. W ilson.
Irrigation in India (U .S. House o f Representatives, The Executive Documents o f
the First Session o f the Fifty-Second Congress, 1891-92), Governm ent Printing
Office, W ashington, v. 18.
24 The brief discussion of ecotypes by Wolf, op. cit., pp. 18 ff., remains one of
the best and m ost accessible.
 There exists no ecotypology for South India, contem porary or
historical, and to develop one for modern South India would
require elaborate statistical and field studies. Extant historical
evidence would probably not support the form ulation of an ecoty­
pology for any time earlier than the eighteenth century, and then
only fragmentarily. To the extent that there has been an effort
to classify general cultivation systems in the literature on South
India, a simple threefold classification inherited from British
revenue usage has been relied upon: wet, dry, and ‘garden’ agri­
culture. Eleventh century Chola inscriptions refer to three slightly
different categories, namely dry land (kdttarambam or puncai,
modern ‘punjah’) wet land (mrarambam) irrigated by a small
tank or riverine channels, and swidden plots (kummari) of ‘forest
people’ (vedar).25 These designations are gross indicators of
basic ecotypes, and their lim itations are obvious.
British revenue and settlement records, even cursorily studied,
yield a long list of types of ‘wet’ cultivation, ‘dry’ and ‘garden’
cultivation for any South Indian district or part of a district.
There are, for example, fundam ental differences am ong perennially
irrigated riverine tracts as those o f the Kaveri and Godavari-
Kistna deltas, those tracts served by channels of m inor tributaries
of these rivers, those irrigated tracts dependent upon the catchment
of run off from m inor streams in tanks, and, finally, those tracts
which depend for their irrigation upon tanks which are wholly
rain-fed. Similarly, dry-crop o r ‘garden’ cultivation carried out
at the fringes of perennially irrigated lands and therefore affected
by a higher water table is quite different from the full and hazardous
dependence upon rainfall alone as occurs in many parts of South
India. W hat constitutes ‘garden cultivation’ in British revenue
usage is perhaps the m ost variable category of all. Sometimes
‘garden’ refers to ‘semi-dry’ cultivation as in parts of Chingleput
and Coim batore where one type o f rice cultivation cycle begins
as a dry crop and changes to a wet crop if and when water availabi­
lity permits. At other times, ‘garden’ m ay be used to designate
25 K.V. Subrahm anya Aiyer, (H istorical Sketches o f Ancient Dekhan, M adras,
1917, v. 1. pp. 351-2) discusses a Tamil inscription from K olar District which has
received considerable attention. The text o f the inscription is found in E. C., v. 10,
M ulbagal Taluk, no 49a, with the Tamil transliteration on p. 102 and an English
translation on p. 86, in pt 2 of the volume. In the Tamil text, line 17 refers to punsai
and line 18 to eri-kil, or ‘un.der the tan k ’.
very extensive holdings as in parts o f Chingleput, Salem, Coimba­
tore, and the Arcots where a limited am ount of assured moisture
is available for some part o f a crop cycle. Still at other times
‘garden cultivation’ describes very small holdings adjoining house
sites everywhere.
It is not only that ther.e are many kinds o f ‘wet’, ‘dry’, and ‘garden’
cultivation in all parts o f South India, but com binations of these
in particular settings defy by their variety even the m ost tentative
classification. The presence or failure of an irrigation source
with its direct effect upon water balances in the locality can permit
or prohibit forms o f dry or garden cultivation dependent upon
well w ater; the presence or absence of livestock affects manure
availability and the practice o f com pacting the germinating seed
in some forms o f dry rice cultivation. And so on; there are an
alm ost endless com bination o f interdependencies for which there
is evidence in contem porary South India. N or is this the end of
the m atter: each o f these com binations sets up unique, decision
situations for the cultivator.
Agricultural strategies are drastically different according to
whether a tract is perennially watered, partially watered as in well
or tank irrigation, or completely dependent upon rainfall. In
the older riverine irrigated areas o f the deltas, a basic strategy of
cultivation could be followed alm ost invariantly. W here special
conditions or problem s existed, solutions could be routinized as
in the productive, low-lying areas o f Tanjavur and Tenali in G untur
where stagnant water presented special problems. Or, in the wet
rice cultivation o f K olleru (Collari) Lake in K istna district and
along the banks of m any o f the larger tanks elsewhere on the C oro­
mandel plain, cultivators developed plant varieties capable of
rising with the filling o f the lake or tank after the south-west m on­
soon rains, and they developed techniques for harvesting from
boats. G arden cultivators relying upon irrigation from wells for
part o f their overall moisture requirements are faced with a wholly
different strategic regime. They are required regularly to rotate
crops so as to provide the optim um balance of soil use and nutrition.
But m ore im portantly, they have h ad to establish contingency
plans for the timings o f certain cultivation systems of the garden
class. That is, where, some irrigation could be relied upon, there
was a wide range o f options open to the cultivator, and decisions
were constantly required. Where cultivation strategy may appear
 m ost crucial is where there is no reliable source of moisture and
operations are entirely rain-dependent. In such situations, ironically
cultivation strategy is most likely to be defensive to the point
of drastically limiting possible options. Here is found a reliance
upon hardy millets, sown at a high seed rate (ratio o f output to
seed input), so as to maximize a return under the m ost difficult
circumstances. Therefore, the general classification o f ‘dry’ cul­
tivation, in which there m ay appear to be the m aximum scope for
and need o f optimizing strategies because there is too little water
available for the practice o f highly routinized swamp cultivation,
becomes almost as routinized as fully irrigated cultivation precisely
because there is insufficient moisture available to permit anything
but defensiveness in the dry cultivator. Cultivators o f the inter­
mediate category, ‘garden’ crop cultivators, with their experience
and modest stock o f wealth, were perhaps the m ost peripatetic.
Even these gross variations in the strategic com ponent in South
Indian agriculture suggest certain significant social and political
concomitants. In tracts where there was little need to make constant
judgem ents about when to plough and what to plant, where cultiva­
tion operations were relatively secure from the vagaries of rainfall
or, conversely, relatively defensive and invariant because o f them,
the im portant managem ent decisions were quite different from
those o f cultivators for whom even a limited supply o f water was
reasonably assured. Where conditions of adequate m oisture supply
and adequate storage facilities permitted swamp cultivation of
rice, cultivation practices were sufficiently routine to be left to low
status labourers for the most part. Here, the crucial issues for de­
cision were those involving the m aintenance and expansion of
lands for which the requisite annual seventy-five inches o f water
could be obtained. In m ost tracts of this kind, an élite without
cultivating skill or experience could dom inate local life effectively.
Examples of this are the Brahm ans o f the C hola period and forest-
bred warriors, such as the kallar or maravar, who came to control
delta land during the Vijayanagara period. Where, as in parts
o f the upland areas o f Coim batore, Salem, and parts o f the K ar­
nataka, conditions favoured reliable water availability for what
has been called ‘garden’ or ‘semi-dry’ cultivation, social and
political patterns have tended to reflect the dominance of groups
who retained, even as they do today, close managem ent of cultiva­
tion operations. Here, a different élite structure emerged from
(hat of the deltaic areas, thus constituting a significant variant of
South Indian society and culture. In those tracts where moisture
was so scarce or unreliable as to make doubtful the survival of even
the most hazardous settlement of agricultural communities, agri­
culturists remained, materially at least, at the margins of the m ains­
tream o f South Indian life until means were discovered of converting
these lands to productive wet or garden cultivation. These means,
in many cases slowly, became available during the fifteenth and
sixteenth centuries.
Such social and cultural factors associated with various ecotypes
can only rarely be docum ented in the earliest periods examined in
this study; the relationships generated by them m ust therefore
be assumed. Thus, in holding that agrarian technology remained
unchanging over the course o f almost a thousand years and that
the array o f m ajor ecotypes remained fixed in the same period,
there is no assertion that the full complement of medieval ecotypes
is perfectly known or understood. T hat kind o f assessment lies in
the future. F or the present argument, this assum ption provides
a. way o f focusing on the social and cultural factors which operated
over a portion of th at long period of a thousand years to give the
South Indian agrarian system its shape and continuity. If there
were significant changes in technology or in the repertoire of basic
ecotypes o f South India within the period, if new ways of combining
organic and inorganic factors related to agriculture were developed,
then these left no trace in the evidence currently available.
 South India : The Region

South India as considered here is a complex, composite region

consisting of diverse physical, social, and cultural components.
Definition o f its distinguishing characteristics constitutes a crucial
and, often, difficult problem. To a large extent, the difficulty is
conceptual. That is, delimiting the distribution of some element
or related elements which distinguish one segment of the time-
space continuum from another requires both adequate and relevant
distinguishing elements chosen to constitute the region. The ade­
quacy or relevance of the elements according to which a region is
defined are related to. and are alone justified by the problem at
hand. Naively conceived spatial units o f study have been called
‘traditional’ o r ‘historical’ regions, and it is in a naïve sense that
terms like ‘Bengal’ or ‘A ndhra’ or ‘M aharashtra’ have frequently
been used. The choice of such spatial units may of course be per­
fectly adequate and relevant if properly defined. But they ought
not to be selected for the same reason that men climb m ountains
— because they are (or at some time were) there.1
The difficulty of treating broad regions as units of study as done
here in the case o f ‘South India’ is obvious. If it were not obvious,
a review o f the early historiography on the Deccan would make it
so. Since R.G. B handarkar,2 around the turn of the century, the
‘Deccan region’ has received the serious attention of historians
who have studied the vast, highly differentiated expanse as a
single spatial entity. The presumed validity o f treating the entire
peninsula as a unit of study arises prim arily from physical facts.
1 The author acknowledges the benefits to his thinking about the issue of regions
and regionalism of the sem inar convened at D uke University, M ay 1966 and the
subsequent publication by R obert I. Crane (ed.), Regions and Regionalism in South
Asian Studies: An Exploratory Study, M onograph no. 5, Program in Com parative
Studies on Southern Asia, D uke University, D urham , N orth Carolina, 1967.
2 R.G. Bhandarkar, Early History o f the Deccan (Collected W orks), Bhandarkar
O riental Research Institute, Poona, 1972.
 100
kilom eters

1-1 The South Indian Region

Thus, the northern face o f the Deccan plateau — the Vindhya -
C hota N agpur line — has been treated as an effective barrier to sub­
stantial population movement from the Gangetic plain southward.
The double impasses of the Vindhya and Satpura hills backed by
the N arbada and Tapti rivers, on the western side of the peninsula,
and the dense jungle and deeply cut landforms in the east, justify
the widely recognized status of the peninsula as a m ajor natural
region. Discontinuities in topography cannot, of themselves, how­
ever, create viable regions for the purposes of the historian or most
other scholars. There are few significant social, linguistic, artistic,
political, or administrative elements with which these topographic
discontinuities are reliably and explicitly associated. Even as
early as Ashoka, and perhaps the M auryan precursor, M ahapadm a
Nanda, Gangetic military power had moved south-westward,
and possibly south-eastward as well, to the coasts o f the peninsula
thus bypassing a large portion o f the jungle-covered territory oil
the northern and southern faces of the double peninsular barrier
and establishing outposts in the G odavari-K istna basins. These
extensions of Gangetic power were probably prom pted by the
availability o f gold, iron, and copper. Subsequently, kingdoms
of the Deccan from the Satavahanas o f the first century a . d . to the
M arathas of the eighteenth century a . d ., looked as often north­
ward, beyond the natural barriers of the plateau, as southward
for the extension of their tributary regions and for cultural contact.3
While it is true that there has been a relatively lower order of
contact and interaction between the peoples and cultures of the
Deccan (taken as the northern portion of peninsular India) and
those o f the Indo-Gangetic plain or the Corom andel coast, as
opposed to interaction am ong peoples within these latter two
spheres, historians have probably exaggerated the extent to which
differences have been the result o f the physical barriers. The
positive attraction o f the m ajor continental and southern penin­
sular river basins have been overlooked. Thus, if one focuses
3 G. Y azdani'(ed.), The Early History o f the Deccan (2 vols), OU P, London, 1960.
This otherwise valuable work is flawed precisely in that most of its contributing
authors ignore contacts between the ‘Deccan’ and the northern, continental parts
of the sub-continent, on the one side, and the southern peninsula on the other. See,
especially, the essays in v. 1 by H. Raychaudhuri, ‘The G eography of the Deccan’,
pp. 1-65 and G urty Venket Rao, ‘The Pre-Satavahana and Satavahana Periods’,
upon the nodal attraction of the two great basins o f the Ganges
and the Kaveri rather than upon the peninsular barrier to substan­
tial population movement, a different conception o f north-south,
sub-continental discontinuity emerges. This conception affects
how one will view the entire peninsula as a unit o f historical study
and the relationship of portions of the peninsula to the whole and
the rest of the sub-continent.
W.M. Day does as much in conceiving o f the sub-continent as
harbouring two perennial cores of civilization: the Gangetic plain
with its extension into the Cham bal basin, or ‘Hindu-Aryan
India', and the Corom andel plain with its extensions to the table­
lands of the interior' peninsula, or ‘Hindu-Dravidian India’.4
Each of the cores consisted of great populations and each attracted
the interests of quite distant people. W hat separated these primary
cores of civilization was not simply the upthrusting Deccan plateau,
but a broad cultural and political zone between the Kistna in the
south and the Kaim ur Range in the north. This intermediate
zone between the two, prim ary cores of civilization has its own
ancient historical career which was consistently influenced by the
developments o f the Gangetic and Corom andel cores and little
affected by natural barriers. It is not contended that this con­
ception of what is, in effect, a trizonal (north-central-south) rather
than the conventional bizonal (north-south) division o f the sub­
continent overcomes the basic difficulties o f using gross division
of this sort. However, for the problem at hand, this kind o f dis­
tinction may serve to focus attention m ore clearly upon the peren­
nially influential character of the Corom andel plain for a major
part of the sub-continent. Moreover, it is ultimately to recognize
that, for many historical purposes, it may be m ost useful to con­
centrate upon nodal regions or cores, rather than upon uniform
regions, or boundaries.
The term ‘South India’ has been used at times to designate the
entire peninsula, but that, is not its meaning here. In this study,
‘South India’ refers generally to that portion of peninsular India
south of the K arnataka watershed (excluding m odem Kerala)
on the west, and the K istna-G odavari delta on the east. Within
this portion of the peninsula, there has existed a region charac­
terized by a high degree o f sharing o f significant social, cultural,
4 W inifred M. Day, ‘Relative Permanence of Form er Boundaries in India',
The Scottish Geographical Magazine, 65, no. 3 (Decem ber 1949), p. 114.
and political elements and an order of interaction such as to con­
stitute a viable unit for the study of certain problems.
Delineating what might be called the ‘macro region’ for this
study that portion of the peninsula which lies south of an imaginary
line from about thirteen degrees north latitude, at the Western
Ghats, to about eighteen degrees of north latitude on the Bay of
Bengal, still leaves a complex, composite region. It includes
m ost o f what has been called the ‘Dravidian culture sphere' follow­
ing the linguistic usage first suggested by Francis W. Ellis in 1816
to describe a family of languages in the southern peninsula.5
Spate employs the term ‘D ravidian S[outh]’ to refer to this part
of the sub-continent and sees it as consisting of a group o f ‘perennial
nuclear regions’ o f which he lists: Kalinga country or Orissa,
A ndhra or Telugu country, Chola and Pandya parts o f Tamil
country, and the isolated south-western littoral o f Kerala or
M alabar.6 A.H. Dani has also spoken o f the portion of the penin­
sula south of the K istna River as a paleographic region.7 In social
terms the southern peninsula has also been recognized as dis­
tinctive. Irawati Karve delineates a separate southern zone of
kinship organization which includes K arnataka, A ndhra, Tamii-
nadu, and K erala;8 M arriott, in his discussion of caste ranking
among Corom andel peoples has also suggested that parts of
K arnataka and A ndhra share Coromandel characteristics.9 Such
general attributions alone do not justify the usage ‘South India’
adopted here, but they do support the definition of macro region
used here by indicating its broad cultural and civilizational corre­
lates.
W ithin th at m acro region, primacy may be accorded to the
Tamil plain as a m ajor source o f the influence of civilization.
The Tamil poetry of the first several centuries a . d . , better than
anything symbolizes this primacy, for classical, or Sangam, poetry
5 T. Burrow and M. Emeneau, A Dravidian Etymological Dictionary, Clarendon
Press, Oxford, 1961, p. v .; hereafter D.E.D.
6 O .H .K . Spate, India and Pakistan: A General and Regional Geography. Methuen
& Co., London, 1954, p. 148.
1 Indian Paleography, Clarendon Press. Oxford, 1963, p. 193.
8 Kinship Organization in India, Deccan College M onograph Series, no. U,
Deccan College, Poona, 1953. p. 180.
9 M cKim M arriott, Caste Ranking and Community Structure in Five Regions
of India and Pakistan, Deccan College Postgraduate and Research Institute, Poona,
1965, p. 53.
— in the anthologized form we have it today — are creations guided
by a highly sophisticated poetic canon, and the resultant poetic
corpus establishes the Tamil language as the most enduring,
living classical tradition in India.10 Institutions o f Sanskrit and
Prakrit learning were contem porary with the production of most
classical Tamil poems and clearly influenced this latter poetic
tradition. The very term sangam to designate this poetry derives
from a Jaina learned com m unity (sangha) established at M adurai
during the late fifth century that provided not merely grammatical
models to Tam il poets, but gnomic ones as well. Ghatikd-s and
other institutions of high learning in the Tamil plain of this time
and later attracted men from everywhere in the South Indian
macro region, even those from distant Banavasi in northern K ar­
nataka from whence M ayurvarm an, founder o f the K adam ba
dynasty, came for education in Kanchi (Kancipuram). In this
south-eastern coastal lowland, from the fifth century, a civilization
developed whose social and cultural forms profoundly influenced
people over a great portion of the southern peninsula. A substantial
part of this influence was carried by people of the Tamil plain as
the peasant agrarian system there expanded during the Chola
period. In fact, the m acro region is almost conterm inous with
the maximal extent o f the Chola overlordship, and the provenance
of Tamil language inscriptions o f this period helps to define the
m acro region as much as other evidence. Thus, a part o f the m odem
state of K arnataka — its heartland consisting of the m odern
districts of M ysore, Bangalore, and K olar — once comprised
an area o f Tamil influence called Karm andalam , a name which
persisted in Tamil usage long after the time when the territorial
name Gangavadi replaced it am ong K annada speakers. There
were similar extensions of dominance by Tamil speakers — peasant
settlers and warriors — northw ard over the A ndhra plain to the
10 Recent scholarship on this tradition includes: K. Zvelebil, The Smile o f Muru-
gan, E.J. Brill, Leiden, 1973; George L. H art III, The Poems o f Ancient Tamils,
University of California Press, Berkeley, 1975 and K. Kailasapathy, Tamil Heroic
Poetry, Clarendon Press, Oxford, 1968. D avid Ludden, in a personal com m uni­
cation, has pointed out that the exclusion of K erala from the m acro region defined
here is scarcely justified during the Sarigam Ages since m uch of the poetry of the
time was composed in Kerala. The point is certainly well taken for in this early
period, however, except for Venadu (erstwhile Travancore State) m ost of the rest
of Kerala has been as isolated from the interior peninsula as it was linked with parts
o f the Western littoral and the A rabian Sea from the tenth century a . d .
delta formed by the Kistna and G odavari rivers. M odern Nellore
was within the overlordship o f Rajaraja I and, under the last of
the great Chola overlords, K ulottunga I, the Kistna-G odavari
deita was firmly within the overlordship o f the Cholas based in
the Tamil plain. The scope o f Chola authority — which will be
discussed as ‘ritual sovereignty’ below — culm inated a century
of efforts by its rulers to incorporate within their overlordship the
potentially rich K istna-G odavari delta in which Tamil-speaking
peasants had come to represent an im portant segment of the
population. Along with this expansion o f peasant peoples of the
Tamil plain and the Chola overlordship into territories contiguous
to the coastal lowland, including portions o f m odern K arnataka
and A ndhra as well as the uplands of m odern Coim batore and
Salem, went m any elements o f culture and society. Such elements
were transform ed, o f course, especially in interaction with Andhra
and K arnataka cultures. These other Dravidian language cultures,
though less ancient or refined in literary terms, and without a
corpus of literature of classical antiquity such as existed in Tamil,
underwent continuous development during the medieval age.
Each of these South Indian cultural traditions so changed as to
constitute distinctive subcultural variations, each substantially
and, over time increasingly, different from the early Tamil culture
of the plain proper. Even after these cultural traditions had
emerged as m ature after the thirteenth century, there were continued
influences from the core of Tamil civilization in religion, for
example, while a reverse flow o f political power em anating now
from the northern territories under the Vijayanagara overlords
was carried by peasant warriors who were possibly, in many
cases, descendants of the earlier migrants to the north. Such
interactions between the core of Tamil civilization in the coastal
plain with those territories which had been extensions of the plain
produced a supracultural zone and contributed to the distinctive
macro region and directly influenced the peasant society of South
India.
Physical elements o f the m acro region are complex. The southern
portion o f the peninsula shares with the northern portion a penin­
sular configuration. Both parts o f the peninsula are linked by
the old and stable Deccan plateau form ation comprising the core
of the entire peninsula. The plateau has affected the pattern of
settlement over m uch of the peninsula because, as a result of its
geomorphological character, fertile lands capable o f supporting
relatively dense populations were scattered and isolated nodes
o f prosperity and civilization surrounded by forest clad uplands
capable of supporting small, and often predatory, populations.
Spate’s four ‘perennial nuclear regions’ of the ‘Dravidian South’
alerts one to this scatter configuration. However, within each of
these four regions — Kalinga, A ndhra, Chola, and Pandya coun­
tries — further sub-regions may be delineated consisting of small
clusters of sedentary, advanced peoples amidst forest and hill
peoples. Spatial relationships resulting from the pattern o f isolated
settlement and one other m ajor factor — the significance of the
sea — have critical historical im portance for the macro region
under discussion; these characteristics are shared by both portions
of peninsular India (i.e. ‘South India’ and the Deccan) in contrast
to most o f the continental portions o f the sub-continent.
Within ‘South India’ itself, the m ost im portant element relating
to historical agrarian relations is the Corom andel plain on the
eastern littoral, extending from the tip of the peninsula to the
northern edge o f the broad delta of the G odavari and Kistna
rivers.11 Never deeper , than one hundred miles, in the Kaveri
basin, this lowland is moulded into a complex structure by the
rocky extensions o f a broken range o f low hills that parallel the
coast, called the ‘Eastern G hats’, and by patches o f lateritic soils
and rocky m arine deposits. The Corom andel plain is traversed
by streams draining these broken hill ranges as well as those of the
more imposing highland.blocks o f the western side o f the peninsula,
the ‘Western G hats’, including the Nilgiris in the north and the
Annamalai, Palni, and Cardam om hills in the south. Peaks of
this western highland attain an elevation o f 8,000 feet. The m ost
im portant streams are the Godavari, K istna and Kaveri, each of
which forms extensive, fertile deltas; other im portant rivers are
the Penner (or N orth Penner), Palar, Ponnaiyar (or South Penner),
Tam braparni, and Vaigai. The unfolding of this plain to its full
contem porary extent, in agrarian terms, was gradual; it was not
completed until the late nineteenth century when Kaveri canal
11 Spate, op. Cit. chapters 24 and 25 contain a good discussion of the plain and
its features; also see K. Ram am urthy, "Some Aspects o f the Regional Geography
ofT am ilnad \ Indian Geographical Journal, v. 23, no. 2 (April-June, 1948),pp. 1-137;
no. 3 (July-Sept., 1948), pp. 20-61; no. 4 (Oct.-Dec., 1948), pp. 23-33; v. 24, no. 2
(April-June, 1949), pp. 28-41; no. 3 (July-Sept., 1949). pp. 1-34.
 irrigation brought the south-eastern corner of Tanjavur into
irrigated cultivation. This phased opening of the Coromandel
plain provides one o f the im portant factors of agrarian periodiza­
tion for South India, and the gradual settlement of the plain with
sedentary peasant communities is as im portant , in the historical
development of the southern part of the sub-continent as the
settlement of the Gangetic plain was in the northern part o f the
sub-continent. Both plains were prime loci of Indo-Aryan civiliza­
tion; both were the prizes to be won by warriors and cultivators
from near and far.
The full impact o f the seaward boundary o f the Coromandel
plain is yet unclear in agrarian terms. However, the influence
of the sea is very ancient and has led some scholars to speculate
that the basic ethnic com position o f the peninsula may have been
ju st as influenced by overseas immigration as the continental
portions o f the sub-continent were by ancient overland m igrations.12
Further, the impressive maritime activities o f the Pallavas and the
C'holas were but continuations o f the earlier Satavahana interest
in overseas trade.13 M aritim e trade dating from the early years
of the Christian era is dramatically, if incompletely, documented
from m ainland and insular South-east Asia.14 The cultural impact
of the Pallavas is monum entally preserved in the Cambodian
12 Noted by K.A. N ilakanta Sastri, Cultural Expansion o f India. G.S. Press,
Madras. 1959, pp. 66-7; I. Karve. ‘India as a Cultural Region', in Indian Anthro­
pology. T.N. M adan and G opaia Sarana (eds.), Asia Publishing House, Bombay.
1962, pp. 328-35; and T.N. Subram aniam , The Pallavas o f Kanchi in South-East
Asia. Swadesam itran, M adras, 1967. Associated with this is the idea of ‘Lem urian’
or ‘G ondavana’ land forms which supposedly connected South India, Sri Lanka,
Australia, New Zealand, Malaysia and East Africa. See K .K . Pillay, South India
and Ceylon, University o f M adras, M adras, 1963, pp. 5-7.
13 Classical and other references to this trade are found in K.A. Nilakanta Sastri,
Foreign Notices o f South India: From Megasthenes to M a Huan. University of
Madras, M adras, 1939. A Satavahana coin depicts a two-masted ship as do some
coins of the same period from further south. K.A. N ilakanta Sastri, Sources o f
South Indian History, Asia Publishing House, M adras, 1964, p. 63.
14 Sastri, Foreign Notices . . . , R.C. M ajum dar, Ancient Indian Colonization in
Southeast Asia, M.S. University o f Baroda, Baroda, 1955. Notice should also be
taken of the growing critical literature of the "Greater India’ view of Indian scholars
as expressed early in J.C. Van Leur, Indonesian Trade and Society, Institute of
Pacific Relations, New' Y ork, 1955, pp. 91-104 and Paul Wheatley, 'Desultory
Remarks on the Ancient History o f the Malay Peninsula’, in John Bastin (ed.).
Malayan and Indonesian Studies: Essays Presented to Sir Richard Winstedt, Claren­
don Press, Oxford, 1964. pp. 41-3. •
kingdom of A ngkor,15 and the military impact of Cholas in Sri
Lanka and the Indonesian archipelago is documented in some
lengthy and im portant inscriptions.16 W hatever is m ade of the
fragmentary facts of this early • maritim e interest and activity
ultimately, and there is yet much for scholars to do, some of the
m ore conspicuous manifestations of it may be suggested.
From an early time until perhaps the fourteenth century, the
sea presented an opportunity for South Indians to trade and to
pillage. Chinese records identify Kanchi (Conjeevaram, Chingle-
put District) as an im portant trade centre from perhaps as early as
the second century B .C .17 O ther im portant entrepots on the C oro­
mandel coast are m entioned in western classical sources; these
include K averipatnam (Rom an sources: Cam ara), Shiyali Taluk,
T anjavur; Puduchcheri (Pondicherry) (R o m .: Poduca); M arkanam
(Rom .: Sopatm a), Tindivanam Taluk, South A rcot; and Machli-
patnam (R o m .: Masalia). From these ports, South Indian m er­
chants organized in guilds (e.g. manigramattar) sailed to the
entrepots o f the K ra Isthmus or directly to other South-east Asian
ports with wares collected from throughout the southern peninsula
o f India. And to Corom andel ports came merchants from South­
east Asia whose commercial and cultural activities, if the recent
and persuasive research on early South-east Asian history is to be
accepted, were largely responsible for many elements borrowed
from South India.18 W hatever historical assessment is made on
the m ajor directionality o f influence and agents in the early trade
and cultural relations between South India and South-east Asia,
the existence o f trade centres on the coast as well as in the interior
certainly constituted points of a vast commercial network reaching
ultimately to China.
This trade system tying South Indians to a wider world persisted
15 Lawrence P. Briggs, The Ancient Khmer Empire, T ransactions of the American
Philosophical Society, v. 41, pt I, American Philosophical Society, Philadelphia,
1951.
16 Pillay, South India and Ceylon, op. cit.; K.A. N ilakanta Sastri, The History
o f SrJvijaya, University of M adras, M adras, 1949; George Coedes, The Indianized
States o f Southeast Asia, East-W est Centre Press, H onolulu, 1968 (trans. Susan
B. Cowing, ed. W alter F. Vella); George W. Spencer, T h e Politics of Plunder: the
Cholas in Eleventh Century Ceylon’, J.A .S., v. 35 (May, 1976), pp. 405-19.
17 K.A. N ilakanta Sastri, 'The Beginning of Intercourse Between India and
China', Indian Historical Journal, v. 14 (1938), p. 386.
18 Van Leur, op. cit., and W heatley, op. cit..
from the beginning of the Christian era to the fourteenth century
when the character o f South Indians’ participation in it changed.
Two developments contributed to this change. One was the deepen­
ing Muslim control over Indian Ocean trade which favoured the
commercial ascendancy o f Muslim traders from western India,
by this time under greater control of Muslims. Coromandel
Muslims, mostly converted Hindu traders, continued to partici­
pate in the Indian Ocean network, but the role of Coromandel
diminished relatively after this time while that of western India
grew.
Coupled with the extension o f Muslim control over trade lanes
between India and South-east Asia, and the diminished role of
South Indian Hindus was the breakdown o f the guild structure o f
trade in South India. From about the ninth century, wealthy
and prestigious guildsmen, trading over extensive portions of the
southern peninsula and beyond, were an im portant element within
the existing agrarian system. Itinerant guilds provided one o f the
means by which the scattered, advanced agrarian communities
of the period were linked together. Basic changes in South Indian
society after the fourteenth century diminished the ability of
itinerant guilds to m aintain their former functions. These changes
included vastly m ore powerful and larger-scale political units;
both reduced guildsmen to a congeries of localized m erchant
groups in the internal trade while Muslim traders on the coast,
better able to fit into the M uslim-dominated trade system o f the
Indian ocean, handled m ost of the external trade. Until the four­
teenth century, therefore, overseas trade m ust be seen within the
context o f the South Indian agrarian system; after the fourteenth
century, Corom andel trade and the general influence o f the sea
upon agrarian institutions was significantly diminished. 19
The sea boundary o f the South Indian m acro region attained
its greatest early politico-military significance during the Chola
period. Then South Indian soldiers established tem porary holds
in Sri Lanka and beyond. R ajaraja C hola ( a .d . 985-1014) and
his son, R ajendra C hola ( a . d . 1012-44) undertook these daring
military expeditions. Rajaraja, according to a Tanjavur inscription
of late in his reign, claimed conquests in m any parts of South

19 See au th o r’s ‘Medieval Corom andel T rade’, in Merchants and Scholars, ed.
John Parker, University of M innesota Press, M inneapolis, 1965.
India as well as Sri Lanka (Ilam andalam ) and the M aldives;20
Rajendra in the Tanjavur inscription o f his nineteenth regnal
year, boasts o f the same conquests as his father, and also his own,
perhaps most daring, raids upon K adaram (Kedah, modern
Malaysia), the northern portion of the Srivijaya kingdom .21
These expeditions are o f interest in this discussion — whatever
may be m ade of them in m ore general historical terms — because
the composition of the Chola armies reflects something essential
about the organization of contem porary society and because the
probable purposes of such expeditions suggest something essential
about the nature of Chola government and kingship. It is the
latter point which is o f some concern here.
For the Chola overlords, a prime objective was to extend their
overlordship to the full extent of the Corom andel plain from their
central position in it. In part, this was achieved through warfare
and resulted in the recognition of Chola dom inance by warriors
o f northern sections o f the plain at the K istna-G odavari delta,
or the territory o f Vengi, as well as the southernm ost portion of
the Tam ilnadu plain, o r Pandya country. But, this expansion of
the Chola overlordship under R ajaraja and Rajendra from their
core dom ain in the Kaveri basin to the entire Tamil plain and its
adjacent upland was not achieved nor was it m aintained by military
means alone. Chola kings o f the tenth to the thirteenth centuries
were exemplars of medieval South Indian kingship and models
o f appropriate rulership for chiefs o f the macro region; it was
less the might of the C hola rulers than it was their m oral appro­
priateness that provided the basis of Chola rule over the C oro­
mandel plain.
Different from this extension of Corom andel authority over
territories close to the Kaveri were the Chola actions as far afield
as the R aichur D oab to the north-west, the M ahanadi basin in the
north-east, M alaya, and Sri Lanka. These were pillaging raids
m eant to yield rich booty and nothing m ore permanent. Such
20 India, Archaeological Survey of India, South-Indian Inscriptions [S./. t.] (M adras,
1890) v. 2, pt 2, no. 65, 252-9; V.R. R am achandra D ikshitar, Selected South
Indian Inscriptions [S.S.1.1.], University o f M adras, M adras, 1952, pp. 34-44. The
same prasasti appears in other R ajaraja inscriptions of the period, viz., Ukkal
(N orth A rcot District), S.I.l... v. 3, pt 1, no. 4, 6-8, and M erpadi (N orth Arcot
District), ibid., no. 15, 22-4.
21 S .S.I., v. 3, pt 1, no. 20, pp., 105-9 and S .S . 1 .1., pp. 70-6.
far-ranging raids were a part o f the pattern o f Indian overlordship
and a central aspect o f Indian kingship. Where the realization
o f state revenue through an ordered adm inistrative system was
absent except in quite localized territories, Indian rulers, great
and small, depended for a major portion of their income on raids
upon neighbours or, in some cases, upon distant, rich, overlords.
The G upta king Sam udragupta, o f the fourth century, provided
the classic tem plate for such predatory expeditions; more modest
versions were the cattle raids o f contem porary hill peoples upon
plains villages. Often rationalized by historians as logical ex­
pansions o f territory, Chola raids, particularly those beyond
the sea frontier, have a patently predatory purpose. Detailed
knowledge about distant, small, rich, kingdoms, probably borne
by contem porary venturesome Corom andel merchants, would
have served to justify the risks o f these undertakings. The sea thus
provided the opportunity and the means to extend the region
from which wealth could be extracted beyond the peninsula itself,
and it may be suggested that the wealth won in such daring raids
was utilized to widen the Chola power base in the central C oro­
mandel region.22
The western or interior boundary of the Corom andel plain is
m ore difficult to define, consisting as it does of a great variety of
transitional physical situations. These range from slight modifica­
tions o f the plain to sharp topographical breaks caused by up-
thrusting hills and ranges. In general, however, the western defini­
tion o f the plain was m arked by intrusions of the plateau formation.
This ancient upland derives its agrarian significance not only in
marking the edge o f the coastal plain, but also in containing ex­
tensive cultivable land. However, tracts of such land are discontinu­
ous owing to the deep cutting rivers and formidable rocky stratum
bared by these rivers during their turbulent traverse to the plain.
Also, upland ecotypes are hazardous: the rainfall regime there is
adversely affected by the rain-shadow effect o f the western bioc
and the highly variable m onsoonal effects upon the peninsular
interior. M ost parts o f the interior are, however, within easy
access o f the plain at many points, and it is to the broad, fertile
22 N ote is taken of the im portant work of George W. Spencer on this question;
see his Ph.D. thesis ‘Royal Leadership and Im perial Conquest in Medieval South
India: The N aval Expedition o f R ajendra Chola I, c . a . d . 1025', D epartm ent o f
History, University of California, Berkeley, 1967.
Corom andel plain that people of the upland have been oriented
since the ninth century.
The 1,000 foot contour is a relevant indicator o f agrarian poten­
tial north and west from near Kanya K um ari (Cape Camorin)
to the Palghat Gap. West o f this line, between 8 ° 05' and 10 ° 30'
north latitude, rise the western highlands consisting o f the Annama-
lai, Palni, and Cardam om m ountains. N orth of the Palghat opening
and the Ponnani river which drains this gap area to the west, loom
the highest of the western bloc, the Nilgiris, which extend to 10 °40'
north v/here they fall to the M ysore plateau at about 2,000 feet.
N orth, beyond the Nilgiris, the 1,000 foot contour is deceptive
as an indicator for it disguises im portant socio-cultural relation­
ships. It obscures the extent to which the southern M ysore plateau
at 2,000 feet and higher, and the Coim batore plateau with its
extension across the Kaveri into the portion of m odern Salem
south of the Shevroy Hills and the Toppur divide, at between 500
and 1,000 feet, have been linked to the Corom andel plain. Access
to these uplands from the plain is attained with considerable ease
in most places, but with difficulty in others where the gradient is
steep and precarious.
Between the Corom andel plain and the bloc of highlands to the
west, crested by the Nilgiri peak of D odabetta at 8,000 feet, the
tableland constitutes a secondary agrarian core o f South India.
The uplands of C oim batore and southern Salem are ringed about
by a series of broken hill ranges at between 2,000 and 4,000 feet
— the Shevroy, Kollam alai, Pachamalai, K alrayan, and Javadi
— whose foothills m ark portions o f the western edge o f the plain.
O ther hills which rise to 2,500 feet separate portions o f these up­
lands from each other. The Baramahal, which constitutes modern
northern Salem, is separated from its southern portion by difficultly
traversed passes at 1,500 feet. Deeply cut river courses also divide
the upland internally. The Kaveri has comprised a long-recognized
internal boundary between southern Salem and Coim batore;
the Palar and Ponnaiyar too have also divided portions o f the
uplands. However, each o f these m ajor streams, besides providing
focuses of significant population clusters, have served to connect the
uplands with other, m ore densely settled tracts; the Kaveri with
the southern Tamil plain, the Ponnaiyar with the northern Tamil
plain. M odern Coim batore town occupies a strategic place on the
trans-peninsular axis o f the Kaveri and Ponnani through the
Palghat gap thus constituting one of the tenuous links between
the two coastal plains of the peninsula.
The Mysore plateau, though geomorphologicaliy different from
the C oim batore and southern Salem uplands in being more closely
linked to the core of the Deccan plateau formation, has historically
been a continuation of these interior Tamil uplands connected
with the Tamil plain. The agrarian heartland o f the Mysore
plateau has been the maidan, or ‘open country', southward of
the watershed which crosses K arnataka between thirteen and
fourteen degrees north latitude. The K arnataka heartland has
been connected with the C oim batore and southern Salem uplands,
and through these with the C orom andel plain by way of the Kaveri
in the south and the Palar in the east.23 N orth of the watershed,
an irregular line running west from Hassan to K olar, the Mysore
plateau has long been an area of marginal population and agri­
culture constituting a transitional zone to the dry Bombay-Kar-
nataka and the Deccan lavas o f M aharashtra. N orthern Salem,
or the Baramahal, has, in effect, acted as a pivotal tract connecting
the eastern portion o f the K arnataka maidan with the Coimbatore
plateau on the south, and the coastal plain in the east.
The entire upland complex o f Coimbatore, southern Salem,
northern Salem, and southern K arnataka not only share general
topographic characteristics, in contradistinction to the plain,
but climatic ones as well. It is a dry area with a rainfall pattern
variable in quantity and timing. The range of precipitation is
between about twenty-seven and thirty-five inches in a pattern
resembling neither the classic m onsoon of the west coast nor the
so-called ‘retreating South-west m onsoon’ of the coastal plain in
the east. Some parts o f this upland are subject to high unreliability
o f rainfall such as m ost o f C oim batore district where unreliability
exceeds thirty-five per cent.24
The northern boundary o f the coastal plain does no t end abruptly
at the rough hills which m ark the northern edge o f the Godavari
trough. Skirting the coast at this point is an im portant routeway
which the British called the ‘N orthern Circars’ including, in addi­
tion to the deltaic districts, the m odern districts of Vishakapatnam
23 B. Lewis Rice, M ysore: A Gazetteer Compiled fo r Government (Rev. Ed.),
v. 1, London, 1897, pp. 4-5.
24 Government o f India, vol. IX, ‘M adras’, pt IX, ‘Atlas of the M adras State’,
M adras, 1964, m ap no. 8; ‘Rainfall Reliability, 1911-1960'.
and Srikakulam and those o f Ganjam and Puri. This attenuated
extension of the Corom andel lowland, properly the A ndhra plain,
constituted a channel through which trade and people moved from
the eastern portion o f the prim ary core of Indian civilization, the
Gangetic plain, to the second such core in the south-eastern penin­
sula. There is also some evidence o f a reverse flow o f religious
influence and interaction from the Tamil plain to Orissa as in the
a . d . j 396 stone inscription from Bhubaneshwar recording a gift
of land to the Siddhesvaram atha to support ascetics from Chola,
Pandya, and Kanchi country.25 M oreover, there were times,
brief to be sure, when this northern portion of the Coromandel
plain figured prom inently in the affairs of the peninsula south of
the G odavari-K istna delta region.26
The appropriateness o f including portions o f the A ndhra and
Orissa coastal plains in the m acro region designated as ‘South
India’ presupposes non-physical criteria. These should be made
explicit. There are two such criteria which serve to delineate the
macro region being considered here. One consists of a set of
interrelated political, cultural, and social elements which differen­
tiate the region from others in the sphere of Indian culture. The
other criterion is interactional, that is, the recognition o f a qualita­
tively different order of interrelations am ong parts of the macro
region than between any part of the m acro region and places beyond
it.
The medieval South Indian political system was based upon
states which were ‘segmentary’. In this sense it cannot be differen­
tiated from the political system o f most o f the sub-continent at the
time. Localized political units were capable o f being linked,
25 M adras, Archaeological D epartm ent, Southern Circle, Annual Report on Epigra­
phy [A .R.E. j 1955-1956, p. 8 and n o .117 relate to a n a . d . 1396 bilingual inscription,
Oriya and Tamil. Full text of this: India, Archaeological Survey of India, Epigraphia
Indica [£./.,] v. 32, no. 6 (Aprii 1958), 229ff.
26 Rajaraja I according to an inscription at Tirukkovalur in a . d . 1012 claimed
the title o f ‘G ajapati' by which the overlords of Kaiinga were known, indicating
a successful expedition against O rissa; see T.N . Subram aniam , South Indian Temple
Inscriptions [S.I.T.I.] Governm ent Oriental M anuscripts Library, M adras, M adras
Governm ent Oriental Series, no. 157, ¡957, v. 3, pt 2, 14-15. D uring the Vijaya-
nagara period, K rishnadevaraya struck at the G ajapatis along the same corridor,
wresting all of the delta south of the K istna in 1512; see S. Krishnasw am i Aiyangar,
Ancient India and South Indian History and Culture, v. 2, Poona Oriental Series,
no. 74, Oriental Book Agency, Poona. 1941, pp. 136-9.
loosely and symbolically, to kings whose sovereignty might for a
time be recognized by local chieftains. Kings whose overlordships
were so recognized were in all cases effectively in control o f but a
small portion of the m acro region, but the legitimacy of their
hegemonic claims — which were ceremonial rather than real
in any case — could be recognized by those far-removed from this
core o f real power. W hai was insisted upon by those who extended
recognition to an overlord was a style of dharm ic kingship. South
Indian kings and dynasties of the medieval age, beyond the localities
o f their own power, were symbols of authority and legitimacy for
a vast num ber o f chieftains throughout the macro region. It did
not m atter that Chola rulers, for example, were Tamilians for their
sovereignty to be recognized among K annada-speaking or Telugu-
speaking chiefs. N or was it im portant that the Hoysala o r Vijaya-
nagara kings were not Tamils for their authority to be accepted
by local chiefs in Tamil country as attested by num erous inscrip­
tions. It m attered only that the idea of legitimate secular authority,
as symbolized in the concept of kingship, exist for the system of
localized, segmentary political units to function as part of a king­
dom.
These characteristics o f kingship have considerable antiquity.
From the Classical Tamil literature of seven hundred years before
the great Cholas, three kingships were recognized and, during the
medieval period, references continue to be made to the ‘three
kings’ (mu-vendar)27 in the titles of prom inent men, mu-venda-
velar. Then, and even earlier, in the time of Ashoka, the three
kingships were called: ‘Chola, Chera [or Kerala] and Pandya.’28
While there is some question about the location o f Chera kingship29
it would appear to be what is now southern Kerala or, somewhat
earlier, Travancore State. In medieval times, this area was called
venadu. A nother possibility considered by some, including Arokia-
swami, and rejected, is that Chera kingship had its locus in the
upper Kaveri basin, m odern southern K arnataka and northern
Coimbatore. These kingships of Tamil antiquity do little more than
27 [Tamil Lexicon], (University o f M adras, M adras, 1936) v. 6, p. 3332, mu-
ventar.
2f! R om ila'T hapar, Asoka and the Decline o f the Mauryas, O U P, Oxford, 1961,
pp. 251, 255-6.
29 M. Arokiaswami, The Kongu Country, University of M adras, Madras, 1956,
pp. 3-12.
recognize the principle of monarchy. It is oniy with the later
Pallavas and especially the Cholas that the medieval institution
of kingship is established, and this is a monarchical form which
does not vary whether the state is avowedly Hindu, as in the case of
the Pallavas and the Cholas, or Jain, as in the case o f the Gangas
and early Hoysalas o f southern K arnataka.
The character o f the South Indian medieval state as a form of
‘segmentary’ state is dealt with in chapter VII. F or purposes of
delineating the m acro region, it is sufficient to propose that this
form of state is perhaps best exemplified by the Chola state of
Rajaraja the G reat and his son, Rajendra, in reigns covering the
period from a . d . 985 to about 1045.30 This and other South Indian
states of the medieval period were integrated primarily by the
symbolic or ritual sovereignty which attached to the king, not by
the effective power possessed by him. In a polity where coercive
means, or military capability, was not monopolized by the kingly
centre, but possessed by every peripheral, local unit in greater or
lesser measure, there could be no centralized m onarchy in terms
of real power. However, these m any localities and locally powerful
men were linked to kings in Tanjavur in their recognition of the
ritual sovereignty acknowledged in the tens of thousands of stone
and metal inscriptions which have been discovered in the macro
region. These inscriptions do more than tell us about ritual sove­
reignty ; they are crucial com ponent elements of that sovereignty.
Inscriptions exist everywhere in the Indian cultural sphere, of
course. However, in no other parts of India, except perhaps
Bengal during the medieval period, do they appear to have con­
stituted the expressive linkage mechanism of ritual sovereignty
as they do in medieval South India. M oreover, whatever the diffi­
culties o f using inscriptional evidence,31 part o f their great value
is that they can be located in time and space with considerable
precision. M ost inscriptions are documents recording gifts (ddna-
sasana) to Brahm ans or temples from wealthy and powerful
persons o f groups o f a locality. Yet, such documents often have
an introduction, frequently in Sanskrit, referring to the reigning
king, his genealogy, conquests, and dharmic rule. This introduction
adds nothing to the gift docum ent which, in the case of most
30 F o r a discussion o f th e final years of R ajen d ra, see N ila k a n ta S astri. The
Colas, U niversity o f M adras, M ad ras, 1955, p. 221.
31 D. C. Sircar, Indian Epigraphy. U niversity o f C a lc u tta. C a lc u tta, 1965. pp. 23-30.
 copperplates is usually in another local language. The introduc­
tion or prasasti, is all-im portant, however, as a statem ent o f homage
to a great king by those locally prom inent persons who instituted
such gifts. Under R ajaraja I, Chola prasastis were standardized,
kept current on an almost annual basis, and even provided with
an identifying label of a few words, m eyklriti, which could be used
in place o f the full introductory portion of a grant. Here, then,
was a powerful symbolic medium used in the political system.
Inscriptions called upon the skills of literate and erudite men and
expressed the symbols by which the particularistic loyalties,
interests, and affiliations o f powerful local persons were merged
as a segmentary state within a spatial zone o f legitimate overlordship.
A m ore widely recognized distinctive feature o f medieval South
Indian states was the prim acy of assemblies of all kinds in the
' governance o f the num erous localized societies of contemporary
South India.32 It was an assembly o f some sort which most
consistently articulated and took responsibility for the decisions
to allocate agrarian resources to various purposes, at least from
the ninth to the fourteenth centuries. W ith regard to agrarian
resources, the polity was less one of regal, or Kshatriya raj, than
one o f assembly, or sabhd raj.
The political structure in the South Indian macro region was
distinctive by the alm ost total absence o f the Kshatriya institution.
Sastric norm s related to formal political authority and, to a signi­
ficant extent, to the actual practices followed in many parts of
northern India were predicated upon the existence o f a warrior
class which enjoyed high ritual status, its m arriage and descent
rules resulted in broad territorial groupings o f w arrior lineages,
and who possessed a m onopoly o f coercive competence beyond
that o f individual ethnic units (castes or tribes) and other corporate
entities such as village communities, guilds, and religious bodies.
Kshatriya institutions, variable though they were over northern
India, represented a w arrior élite whose authority was maintained
through extensive agnatic and affinal relationships and whose
prim ary function was that of ruling a small territory, a domain
which at times could be expanded by warfare into a large realm
under able military leadership o r could be reduced by succumbing
to the expanding overlordship o f some other regal lineage. Rarely
32 N o te d by A. S. A lte k a r, a m o n g o th ers, in A H isto ry o f Village C om m unities
in W estern India, O U P , M a d ras, 1927.
was this system of political organization capable of being moulded
into a great kingdom covering even a portion of the Gangetic
plain and then only under the quite extraordinary military capabi­
lities of a Sam udragupta or a Harsha. The norm al political condi­
tion in northern India during ancient and medieval times was its
division into a great num ber of small territories under kin-linked,
w arrior families of high status. Such was the nature of K shatriya/
raj.
The macro region of southern India was w ithout such high
status warrior groups, and it would be difficult to reconcile such
a pattern of authority with two o f the salient social structural
characteristics of the m acro region: Brahm an secular authority
and the narrow territorial segmentation of social relationships
and loyalty.
Only in Kerala did there emerge w arrior lineages whose rule
over relatively large territorial units persisted through substantial
periods. It is im portant to note that these warriors never lost
their identity as ‘N ayars’, the bulk of whom were regarded by
Brahmans at least as Sudras. N ayar ‘K shatriyahood’ has thus
been a case of enriched Sudra status. However modest an accom­
plishment, the investment o f sections o f the dom inant non-Brahm an
population with durable Kshatriya status was not replicated in
other parts of the southern peninsula except in very remote and
miniscule hill tracts. This is an im portant reason for excluding
Kerala from the macro region.
The exclusion o f K erala from the m acro region on this, or on
any, basis may appear unjustified; it would appear to present the
regional argument in a weak form even though it does serve to
emphasize an im portant discontinuity within the Indian cultural
sphere. To state this politically distinctive factor in more positive
terms, within the m acro region the most persistent feature of
political organization was the alliance between sections of Brahmans ,/
and representatives o f the dom inant land-controlling population
o f Sudra rank. This constituted a political arrangement consistent
with and contingent upon the high secular authority of Brahmans
acting corporately as assemblies (sabhas ) and an arrangem ent!
basically different from that of northern India where Kshatriya i
rule existed. In the latter case, power relations and conflict were
the exclusive concern of recognized warrior lineages struggling
for dominance among themselves within a territory or coping with
the armed intrusion of warriors from neighbouring territories.
The crucial element for victory, apart from individual military
ability, was the capacity of a warrior leader to mobilize a large
contingent o f armed kinsmen and their retainers to defend against
or to aggress upon a neighbour. In the South Indian m acro region
during the earliest period surveyed here, the political context was
also confined to a small territory in which Brahm ans and high
(sdtvik) non-Brahm ans shared authority over the population of
lower caste people, whose m ajor political function it was to dispose
of the corporate interests expressed by various kinds of groups in
formal assemblies.
The political culture of the m acro region until the fourteenth
century would have perm itted no other dom inant political arrange­
ment than that of sabha raj. Large and populous villages of the
macro region provided an efficient and legitimate means of govern­
ing the affairs of diverse social groups through assemblies w'hose
entire constituencies inhabited a single territorial unit. The terri­
torially segmented structure of social relationships among numerous
social fragments and the absence o f Kshatriyas form the basic
determinants o f the region’s political organization.
Rule through assemblies may have been inherently m ore peaceful
than K shatriya-dom inated areas elsewhere, but violence and
warfare occurred. Defence agains* hill and forest people was a
regular problem in this period and later, and predatory wars against
neighbours did occur along with daring pillaging expeditions far
from the region itself. The great Brihadesvara temple o f R ajaraja I
in Tanjavur, for example, was constructed and m aintained through
demands by R ajaraja upon villages throughout the Kaveri delta
core of Chola power as well as from ‘the booty in the conquests
of Chera and Pandya kings. . .’ and the Chalukya king Satyasraya
as described in the ‘Larger Leiden Plates’ o f R ajaraja.33 Such
warfare tended to enhance the prestige of a few warriors, and this
was an im portant secondary objective o f the activity, but it also
brought fame and fortune to the corporate groups who made up the
armies led by such warriors — soldiers o f the left and right divisions
of castes, certain artisan groups, guilds — thus it strengthened
the vigorous corporate structure of South Indian political relations
until at least the fourteenth century.
53 E .I., v. 22, no. 34.
 From the fifteenth to the eighteenth century, the political structure
o f the m acro region was altered by the introduction of a new order <
of w arrior control in all of its parts. W arriors from the northern
fringe portions o f the m acro region established themselves in
m any parts of the southern portions and, where these ‘n o rth ern s
m en’ ( vadugan ) did not establish their control, there arose locality
leaders from within to end the era of sabhâ râj. Yet, the warriors
who came to dom inate territories previously under the control
of the assemblies o f Brahmans and other corporate groups in
South Indian society did not establish themselves as Kshatriyas !
whom they may superficially have resembled. While destroying '
the system of corporate assemblies, the new w arrior élite preserved
the older Brahm an-Jôi-Sudra alliance. The latter groups provided
a ready pool of collaborators with the new order, and they were
rewarded by enhanced functions as village and locality leaders.
As for Brahmans, the new warrior élite of South India sought and
obtained a required m oral standing or legitimacy in return for
support of new sacred temple centres of Brahmanical authority.
For all of their considerable efforts however, m ost new warriors
failed to establish either durable or extensive networks o f kinship
and stability within the social structure o f the m acro region.
K allar and M aravar rulers o f many parts o f the southern peninsula
and Tanjavur were exceptional in this respect, but their political
networks conformed to pre-existing clan-like control of many of y
these parts o f South India. However, neither achieved high ritual
status, though they often claimed high non-Brahm an (i.e. Vellala)
rank.
When the British extended their territorial power in the south
during the eighteenth century, this shallow-rooted w arrior élite
was found not to constitute a landed class through which Company
objectives of stability, peace, and tribute could be realized. Called
‘ poligar’ by the British (Tamil: pâlaiyakkârar, or ‘m en o f military
encam pm ents’) these warriors had not developed the client relation­
ships of, say the Oudh talukdars or that group o f landed
agents who were called ‘zam indars’ in Bengal; they had failed to
create for themselves an integrated place within the social structure
o f the south essential for the intermediary functions which the
British required. In the face of British expansion, ‘poligars’ could
only fight to retain their unstable overlordships and thereby risk
annihilation, or they could forego their prized positions and accept
the uncertain status of village renter or later, zam indar under the
Company. M any warriors chose the former, and the late eighteenth
century witnessed the liquidation o f this class of low caste military
leaders which had failed to alter the earlier structure of society
sufficiently to provide viable places for themselves during the
course of four centuries when they had power. The Brahman -
high non-Brahm an alliance thus remained intact to almost 1800.
Perhaps, as in none o f the m ature sub-cultures o f the Indian
sub-continent in medieval times, the Sanskrit language and ideas
derived from its texts were balanced by non-Sanskritic cultural
elements in South India. Social, religious, and political categories
from Sanskrit sastras and puranas were utilized in the inscriptional
and literary sources o f the age, often when these categories had
little sociological, ritual, and doctrinal, or political reality. How­
ever, in the m acro region under study here, and especially among
the Tamils, there was a prom inent place given to the languages,
symbols and usages of the indigenous, Dravidian cultures as well.
To state the m atter in this way is to suggest two independent
traditions — Sanskrit and Dravidian — which is a false concep­
tion, for even in classical Tamil literature the two are so inextricably
interwoven as to defy disaggregation into autochthonous, interact­
ing phenom ena. Yet, an effort was m ade to preserve some degree
o f formal independence o f the two traditions as shown in the
bi- and sometimes tri-lingual inscriptions of the medieval period
and the tendency in inscriptions — the m ajor public documents
of the time — to use Sanskritic and Dravidian language forms.
Related to this balance of Sanskritic and Dravidian elements
was the place o f Brahm ans as the m ajor cultural arbiters of the
culture of the m acro region. The high visibility o f Brahmans in
their distinctive settlements (brahmadeyas ) over m uch of the
Tamil and A ndhra plains, in their religious and instructional roles,
and in the significant political duties they carried out, are noted
immediately below and elsewhere in this study. W hat may be
emphasized at this point is that the cultural role o f Brahmans in
the medieval South as here delineated is not that o f an exogenous
influence upon essentially D ravidian societies o f Tamilians,
Kannadigas, or Telugus. Brahmans were as integral to these
respective traditions as were non-Brahmans. As prim e custodians
of Sanskrit knowledge, they are easily identifiable, of course,
but it is im portant to recognize that Brahm ans as learned indivi-
duals perhaps most perfectly incorporated the balance of Sanskritic
and non-Sanskritic elements which are given saliency here.
Brahmans were not alone in this ability, however. It becomes
increasingly clear that from am ong the dom inant non-Brahm an
peoples of Tamil country at least there were an impressive num ber
at any time who were adept in Sanskrit learning as well as, with
Brahmans, custodians and transm itters of non-Sanskritic cultural
forms. And, while the evidence o f tension between Brahman and
non-Brahm an savants and religious teachers becomes manifest
in the thirteenth century, and lays the foundation for some o f the
conflict between the two in the twentieth century, the tension is
not that between maintainers o f an indigenous culture against
external intruders, but largely that o f cultural variants and their
upholders seeking the greatest favour from those in a position
to- support them.
There are, finally, social forms which give distinctiveness to the
macro region designated as South India here. While sastric
models of Indo-A ryan society are in no part o f India perfectly
realized, historically or at present, the social structure o f the South
Indian m acro region has long provided a unique and persistent
-variant o f Indian society.
The uniqueness o f this variant of Indian social structure is
based upon three persistent and related characteristics: the great <
secular authority and significant secular functions of South Indian
Brahm ans; the dual division of lower social groups; and the
territorial segmentation o f all social hierarchies in the South Indian
macro region.34
B rahm an secular authority over the entire period o f this study
appears to stem from two conditions in South Indian society.
Brahm ans were strongly entrenched in the localities where they
lived as a result o f the prestige which attached to their sacerdotal
functions, and this prestige was backed in many places by the v
power which they possessed by virtue o f their direct control over
land and those dependent upon land. In the m atter o f prestige
associated with sacral functions, South Indian Brahmans did
not differ from priests in m ost other parts o f India. This is strongly
suggested by caste ranking in all parts o f the sub-continent. There
is no reason to suppose that the high rank ubiquitously enjoyed by
54 M a rrio tt, Caste R anking . . . , pp. 45-53, notes these factors as does K arve,
op. cit.
 priestly castes during the last century, when data for reasonably
reliable ranking has been available, was different in earlier times.
However, in respect to Brahman locality power associated with
'< land control, South India appears quite unique. In no other
portions o f the sub-continent did elaborate and powerful Brahman
villages, brahmadeyas , exist as they did in the Corom andel and in
many parts o f the contiguous tableland during the period.35 Simi­
larly, in no other parts o f the Indian cultural sphere were there so
;f many Vedic temples with substantial control over endowed villages,
, devaddna, as in South India from about the fourteenth century
! to the eighteenth century. South Indian temples of the medieval
period were unique in the degree to which they provided the means
for Brahm an temple functionaries to exercise not only ritual
primacy over all other castes and religious institutions, but also in
th at temples were the headquarters of bhakti sects through which
organizations the religious allegiances and ritual activities of most
Hindus were ordered. The relatively comprehensive ritual, social,
and agrarian functions of medieval South Indian temples resulted
from many factors.
As compared with H indu temples elsewhere in India during this
period, the fact of being outside of the control o f Muslim power,
and possibly reacting to the threats of that power, was very im por­
tant. However, ritual developments within bhakti Hinduism were
an equally im portant reason for their flourishing condition in
South India. As custodians of these religious centres, Brahmans
were in a position to enjoy great secular authority, compensating,
to some degree, for the contem porary decline in influence o f Brah­
man villages under the altered political conditions o f the fourteenth
century and after. Finally, the early period o f British rule in
South India provided opportunities for the maintenance of powerful
secular positions by Brahmans. As beneficiaries o f early British
land policies in South India, Brahm ans were invested with pro­
prietory rights over some o f the richest, best irrigated lands in the
presidency of M adras. Simultaneously, Brahm ans became the
most dependable servants o f British officials in both collectorate
and provincial offices of the Presidency. In these ways they were
able to protect their ancient secular authority and, in some ways,
even to enhance it by the early nineteenth century. Brahman
35 T h is p o in t is stressed by A ltek ar, Village C om m unities in W estern India.
secular authority was not uniform throughout the m acro region;
it was greater in places like Tanjavur than in Salem and Coimbatore.
But, in contrast to the secular authority and functions o f Brahmans
elsewhere, it is a most striking phenom enon in South India.
The second persistent characteristic o f the social structure o f the i
South Indian macro region was the bifurcation o f lower social
groups. The terms for this dual division of lower castes are not
consistent over the macro region, being referred to by the right
and left hand designations am ong Tamil and K annada speakers
and by these, as well as by sect designations (Vaishnavas corres­
ponding to the right-hand division and Saivites corresponding to
the left-hand castes elsewhere). N or is the composition o f this
fundamental division easily specifiable because of variations
in time and place.36 However, in m ost cases the right-hand castes
(Vaishnavas am ong lower Telugu castes) are associated primarily ^
with agricultural production and local trade in agricultural com­
modities while left-hand castes are associated with mobile artisan
production and relatively extensive trade in non-agricultural 'com-'1
m odities.37 These divisions of lower castes appear to operate
as supra-local systems of alliance from which are excluded the
two other, and most powerful, social strata of South Indian society,
Brahmans and high non-Brahm an castes.
For at least the last five hundred years it has been possible to
identify a tripartite horizontal division of all South Indian castes
in most parts of the m acro region. These are Brahmans, respectable
agricultural castes of acknowledged high ritual rank, and lower
castes, the last group of castes being divided into the bifurcated
36 M a rrio tt, in Caste R a n k in g . . . . uses the term 'fa c tio n ’ m ost often in his dis­
cussion o f this stru c tu ra l ch aracteristic, an d only rarely th e term ‘alliance’ as on
p. 51. A n excellent, recent reco n sid eratio n o f the m atter is fo u n d in A rjun A p-
p adurai, ‘R ight an d Left H an d C astes in S o u th In d ia', The Indian Economic and
Social H istory R eview , v. 11, nos. 2 an d 3 (1974), pp. 216-60.
37 This is su p p o rted from a w ide range o f evidence: C.S. S rinivasachari, ‘T he
O rigin o f the R ig h t an d Left H an d C aste D ivisions', Journal o f the Andhra H is­
torical R esearch So ciety. IV (1929), pp. 77-85: J. H. N elson, The M adura C ountry:
A M anual, A sylum Press, M a d ras, 1888, pp. 4 -7 ; N . S u b b h a R eddi, ‘C om m unity
C onflict A m o n g the D epressed Classes o f A n d h ra ’, M an in India, v. 30, no. 4 (1950),
pp. 1-12; G o v ern m en t o f M ad ras, Proceedings o f the Board o f Revenue, 1796, 16,
20 and 27 J u n e ; 14 a n d 18 Ju ly ; 4, 8, 22, an d 25 A ugust. M anual o f A dm inistration
o f the M adras Presidency, v. 3, ‘G lo ssary ’, G o v ern m en t Press, M a d ras, 1893,
pp. 1036-7.
 groupings. Prior to the fifteenth century, and possibly as early
as the eleventh century, the dual division of lower castes was an
im portant factor in South Indian society, but the deeply competitive
and conflict-ridden nature of relations between the two divisions
of the Vijayanagara period and later appears to have been largely
absent. Why and under what conditions the change from the
relatively peaceful and integrated relations of the earliest period
to the violent and competitive relations occurred is discussed
below. But, whether peaceful or violent, the dual division of
lower castes is a distinctive m arker o f the caste organization of
the macro region.
Territorial segmentation o f society and culture in the South
Indian macro region is the third characteristic of social structure,
the distribution of which contributes to its definition as a region.
This characteristic relates to the m arked extent to which social
groups have tended to m aintain a low order of significant and
y persistent relationships with groups at any substantial distance
from their locality as com pared to the often extensive network
o f group relations m aintained in other parts o f the sub-continent.
Locality loyalties and parochial relationships o f all kinds are
attributes of peasant society and culture to which the Indian
cultural sphere is not excepted. However, m arriage and descent
<systems in South India operate in quite narrow territorial arenas.
This is related to the preference for cross-cousin marriage and
maternal uncle-niece m arriage within narrow marriage ‘kindreds’
as well as to the tendency for village and village clusters to be more
populous in South India than elsewhere. Also related to the
territorial segmental feature is the precocious development of
locality institutions by the Pallava period which provided the
institutional framework through which the numerous nuclear
territories of the Corom andel lowland functioned.
Territorially differentiated units of culture and society are
deeply imbedded in the earliest Dravidian culture of the Classical
period. According to one of the Classical works of this early
period of Tamil literature, possibly the third century a.d., social
groups in Tamil country were divided into five situational types
on the basis of natural sub-region and related occupational patterns.
The Pattupattu enumerates these territorial segments (or tinai)
as follows:
 1. M aruta m akkal or tribes of ploughmen (ulavar) inhabiting
fertile, well-watered tracts (panai) and living in villages
called ur;
2. Kuravar m akkal or hill people who are foresters, make
charms, and tell fortunes and may come out of the forest
to work in the p a n a i ;
3. M ullai m akkal or pastoralists, also called ay dr (cowmen),
kovalar (shepherds), and idaiyar (cowherd or shepherd);
4. N eytal m akkal or fishing people living in large coastal
villages called pattinam or small ones called pakkam ; and
5. Palai m akkal or people o f the dry plains called eyinar ,
maravar, and vedar who are hunters of both the dry plains
and the forest.38
This five-fold division of ancient Tamil speakers is interesting for a
variety of reasons. As a richly elaborated poetic scheme, it provides
a pool of images which gives these poems much of their expressive
power. Beyond that, as a description of spatial categories of Tamil
subcultures they are im portant cultural concepts. Finally, these
categories suggest a ranking postulate com parable to the varna
concept elsewhere in the Indian cultural sphere. It is clear from
Classical literature that the people of the first category, those of the
panai who lived by the plough enjoyed a special place in the affection
of the greatest Classical poets who were, for the m ost part, men
of the ‘thriving soil’ of the panai-39 It is easy, of course, to make
too much of this element o f ‘ancient’ Tamil culture. F or one reason,
the chronology of the literature containing these cultural categories
is still controversial, with one group o f scholars favouring the
period from the sixth to ninth century and another larger group
favouring the earlier period.40 Also, the relationships between the
38 S. V ith ian ath an , ‘T he P a ttu p a ttu : A H isto rical, Social, and L inguistic S tu d y ’,
un p u b lish ed P h.D . thesis, U niversity o f L o n d o n , S chool o f O rien tal an d A frican
Studies, July 1950. T hese term s have been exam ined by n u m ero u s scholars: see
V. K a n a k a sa b h a i. The Tam ils Eighteen H undred Years A go. H igginbotham & Co.,
M ad ras, 1904; S rinivasa Iyengar, op. c it.; N. S u b ra h m an iam , Pre-Pallavan Tam il
Index, U niversity o f M a d ra s. M ad ras, 1966 an d his Sahgam P olity, A sia P ublishing
H ouse, M a d ras. 1966.
39 V ith ian ath an , op. cit., p. 140. T h a t S angam poets w ere in m ost cases m en o f
the panai is suggested in the a u th e n tic a te d o r p u tativ e a u th o rs, com pilers, and
p a tro n s o f th e an th ologies as p rovided in S u b ra h m an iam , P re-Pallavan In d ex,
pp. 1-16.
40 N ilak an ta S astri, Sources o f Indian H isto ry, pp. 54-7; Z velebil, S m ile o f
Murugctn, ch. 3, pp. 23ff.
culture of a people, as reflected in a body of poetry collected and
selected as anthologies centuries after being composed, and actual
social patterns may be easily distorted.
These political, cultural, and social attributes and their dis­
continuous distribution over the southern peninsula permit the
delineation of an historically significant macro region. However,
there is another criterion which is im portant, an interactional or
transactional one.
It is proposed that a portion of the southern peninsula may be
demarcated on the basis of persistent and im portant interrelation­
ships over most of the medieval period. In political, cultural,
and social terms all of Tamil country and the southern parts of
K arnataka and A ndhra may be seen as bound together by the
movement of peoples of all kinds — from Brahmans to the most
vulnerable of landless folk — cult practices, and shifting patterns
of overlordship. The outcome of these diverse interactions was a
region which, while complex in language, some aspects of social
structure, and cultural forms, was a uniformity which sets it off
from other, physically contiguous territories. It is moreover clear
from these historical interactions that the Kaveri basin — the
seat of Chola power — constituted a core to which most of the
southern people of the Indian cultural sphere looked for sources
of political, cultural, and social developments in their own, often
distant, homelands. Finally, the boundaries of the macro region
which found the Kaveri as its prime locus are conveniently dem ar­
cated by the Chola overlordship in the time of R ajaraja I and
Rajendra I. At the zenith of Chola power, that part of the southern
peninsula which fell under the Chola overlordship was the macro
region. This should not be understood to mean that the macro
region was created by Chola conquest, but rather the opposite.
The extensive overlordship claimed by the great Chola rulers and
validated by hundreds of inscriptions of the time of these rulers
was made possible in large part by the broad agreement about the
legitimacy of certain symbolic relationships. The Chola. king was
a ritual centre, a political, cultural, and social reference point for
the organization of medieval South Indian society.
Using the attributional and interactional markers discussed
above, it is possible to delimit the m acro region of ‘South India’
as understood here from other parts of the southern peninsula.
Specifically, this study will not deal with what is now most of the
state of Kerala, nor with Telingana, nor with the northern portion
of what is now K arnataka state or what the British called ‘the
Bom bay-K arnatak’.
The west coast o f the southern peninsula has long constituted
a special problem for scholars of South India. Its isolation behind
the high scarp of the Western Ghats has resulted in well-recognized
discontinuities with respect to its social structure, its culture,
and its settlement patterns.41 Yet, however unique it may be —
even considering the diversity of India — its society and culture
possess elements of Dravidian India, and it cannot be summarily
dismissed. Two factors weigh heavily in excluding most of the
west coast o f the peninsula from the m acro region under study
here. First, though modern and medieval Kerala is a part of the
general South Indian cultural sphere, its interactions with the
Coromandel lowland and with interior uplands o f m odern Coim­
batore and Salem as well as with southern K arnataka have been
of a low order and sporadic. Physical separation has never been
overcome except in the case of the Venadu tract of the extreme
south which was in continuous and close contact with other parts
of the m acro region.42 The second factor is that the historical
evidence of this isolated portion of the peninsula is too meagre
and insufficiently synthesized at this time to permit analysis o f its
peculiar agrarian characteristics for any time before the nineteenth
century.
Exclusion of Telingana is based upon similar considerations.
41 K arve, op. cit., p. 252, states, . . K erala represents a lan d o f isolation w here
ancient cu sto m s h av e been preserved an d w here im m ig ran ts soon lost c o n ta c t
w ith th eir h o m elan d s an d m ade stran g e a d a p ta tio n s to the custom s o f the native
p o p u latio n , th u s them selves a d d in g to the peculiarities o f the la n d ’. D r Jo a n
P. M encher m akes the sam e p o in t in term s o f settlem ent an d ecological factors
in her ‘K erala an d M a d ra s: A C o m p arativ e S tudy o f E cology an d Social S tru c tu re ’,
E thnology, V, no. 2 (A pril 1966), pp. 135-71.
42 T h is seem s very clear from the w o rk o f T. K. V elu Pillai ( The Travencore
S ta te M anual, Vol. / /, H istory, G o v ern m en t o f T rav an c o re, T riv an d ru m , 1940,
pp. 50-118) w hen discussing the p eriod from the n in th cen tury to the tim e o f R avi
V arm a K u la se k h a ra in the early fo u rte e n th century. It is also clear from the case
o f the im p o rta n t V en ad u (so u th ern T rav an c o re) Saivife tem ple o f S ucindram
w here T am il B rah m an s m ain tain ed c o n tro l th ro u g h their c o n tro l over th e village
assem bly o f the place until the fo u rtee n th c e n tu ry when they w ere replaced by
N a m b u d ris; m o reo v er there w as no sta te interference w ith the o p eratio n s o f the
tem p le un til th e m iddle o f th e sixteenth century. See K. K . Pillay, The Sucindram
T em ple, K alask sh etra P u b licatio n (A dyar), M a d ras, 1953, pp. 153, 167.
This tract between the relatively fertile trappan lands of the Bombay-
K arnatak, seat of the ancient power of the Chalukyas and Yadavas,
and the rich deltaic lands of Vengi provided such a poor basis for
settlement of agricultural peoples that it was not until the middle
o f the twelfth century that something like an organized state
emerges from the confusion of warring m inor chiefs.43 This state,
when it does come into existence is a doubtful version of the seg­
m entary state system o f other parts o f the m acro region. The
Kakatiya kingdom o f Telingana was not an operative state except
for the late part o f the reign o f Ganapatideva ( a . d . 1199-1261)
when the congeries o f previously independent m inor chiefs of
the area recognized the overlordship, or the ritual sovereignty,
of this king. It appears quite certain from the records of the
period that w ithout the vigorous military activities of G anapatideva
even this brief period o f overlordship could not have been achieved.
Thus the Kakatiya ‘kingdom ’ existed as a recognized polity for
too brief a time before the Muslim conquest in the fourteenth
century to inspire confidence that it was really a ‘state’ at all,
A final reservation about the inclusion of Telingana is that the
inscriptional evidence pertaining to the ‘kingdom ’ shows little
of the balance of Sanskritic and non-Sanskriiic elements charac­
teristic of the inscriptional records of other states of the southern
peninsula. During the twelfth and thirteenth centuries, Sanskrit
forms are dom inant in the glimpses we are permitted of this area;
there is no evidence of self-conscious Telugu culture as one finds
contem poraneously in the neighbouring Vengi region. In time,
a balance of Sanskritic and Telugu elements does emerge, even
through the veneer of Muslim forms, but the process is a slow one.
Telingana, in short, was a shatter region during almost all of the
medieval period; it serves to m ark one of the boundaries of the
m acro region of medieval South India.44
Reservations about the inclusion of what during the British
period was called ‘B om bay-K arnatak’ are based upon attributional
and interactional factors. W hat are now the K arnataka districts
of Bidar, Gulbarga, Bijapur, Belgaum, Raichur, and parts of
D harw ar and Bellary were in medieval times linked to the Deccan
43 N . R a m e san , Copper P late Inscriptions o f Andhra Pradesh G overnm ent M useum ,
H yderabad, G o v ern m en t o f A n d h ra P ra d esh , H y d erab ad , 1962, v. 1. p. 96.
44 N. V en k ataram an ay y a a n d M. S o m ase k h arasarm a, 'T h e K ak a tly a s of W aran-
g al’ in Y azd a n i, op. cit., v. 2, pp. 575ff.
cultural sphere rather than South India. Kannada is, to be sure,
a Dravidian language on the basis of which the districts mentioned
above were joined with the princely state of Mysore in 1956.45
However, there is a well-recognized and major dialectical break
between northern and southern K annada districts, the ‘D harw ar’
and ‘M ysore’ dialects respectively.46 The dialectical isopleth is
said to lie along the T ungabhadra River, but that is an imprecise
designation.
This dialect division conforms to a basic division of the Mysore
plateau by a watershed ‘Crossbelt’ which traverses the entire
Karnatak area and divides the drainage basin of the Kistna system
to the north and that of the Kaveri to the south.47 A ridge line
runs irregularly across that portion of modern K arnataka state
below the Tungabhadra, beginning at around 12°30' north latitude
near the 76 degree meridian and extending to about 78 degrees
where it merges with a complex river system formed by the northern
and southern Penner and the Palar rivers. N orth of this ridge
line, the drainage is northerly toward the K istna; south of the line,
the drainage is to the Kaveri, and this area constitutes the upper
Kaveri basin. The eastern extension of this divide merges with
the riverine systems of the Tamil plain.48
The dialectical variation of K annada follows this topographical
division and so do other im portant cultural elements. In the
introductory chapter on cultural geography in A nanthakrishna
Iyer’s M ysore Tribes and Castes, these cultural discontinuities
are noted and roughly m apped. They include: the incidence of
Tamil inscriptions,49 the distribution o f Vira Saivite sectarians,50
and the distribution o f the m ost num erous castes o f K arnataka.51
^LSpate, op. cit. p. 645, cites the 1931 C ensus to rep o rt th a t K a n n a d a speakers
on a d istrict basis w ere: M ysore d istrict, 93% ; H assan, C h ik m ag alu r, Shim oga,
C h arw ar, B ijapur, an d B angalore, 75-90% ; C h itald ru g an d R a ic h u r. 72% an d
64% respectively; an d B ellary an d G u lb arg a, 49%.
46 W .C. M c C o rm ack , Kannada: A C ultural Introduction to the S poken S ty le s o f
the Language, U niv ersity o f W isconsin, M adison, 1966, p. 3.
47 L .K . A n a n th a k rish n a Iyer, T he M ysore Tribes and C astes, M ysore U niversity,
M ysore, 1928-35, v. 1, p. 84 an d m ap on p. 80; R ice, M yso re: A G azetteer Com piled
for Government, p p .4-5.
4« Ibid., ch. 2, pp. 8 Iff.
« Ibid., p. 115, Fig. 8.
5° Ibid., p. 120, Fig. 12.
51 Ibid., pp. 123-4, 126; Figs. 17-20, 21, 23, 24, 27-32.
That these cultural elements are not abruptly discontinuous in
their distribution needs no emphasis, of course; however, what
is significant about the distributional gradient between the northern
and southern portions o f the princely state o f Mysore (i.e., prior
to the addition of the Bom bay-K arnatak districts in 1956) is that
much of K arnataka is best regarded as a transitional zone between
the Kaveri core of the macro region which is delimited here and
the Deccan culture area whose southern edges touch the Kistna
basin.
A social element which dem onstrates the transitional character
of K arnataka and supports the exclusion of the Bom bay-Karnatak
is the attenuation of the dual division o f castes into groups of the
right and left hand as one moves northw ard from the southern parts
of K arnataka. This characteristic social structural m arker of the
macro region is noted in early nineteenth century reports on
Mysore and continue to be referred to in the censuses of the late
part of the century. Oddly, there are no references in medieval
inscriptions to the right hand (halagey) and the left hand (edagey)
castes and, while these term s are used during the nineteenth century,
the census reports rem ark that the term phana, was preferred to the
reference to ‘hands’. The ‘Eighteen Phanas ’ referred to castes of the
right hand, and the ‘Nine Phanas’’ to those of the left hand. Ac­
cording to the Census R eport of 1901, the word phana was a corrup­
tion of the K annada word, banna, itself a corruption of the Sanskrit,
varna.52 In the same 1901 Census, it is noted that the dual division
utilized ‘eighteen’ and ‘nine’ as conventional numbers since all
but a handful of non-Brahman castes were to be regarded as either
of one or the other phana, and o f those not included in the two
conventional divisions, it was noted that some numerically signi­
ficant castes declared themselves to belong to a new ‘twelve phana’
category. Thus, in 1901 this division o f castes in Mysore appears
to have become a confused category.
However, earlier, in 1891, a larger num ber of persons reported
themselves as affiliated with one or the other of the divisions
(approximately one-half of the population), and these affiliations
were tabulated on a district basis. It is strikingly clear that the dual
division was more significant in the Kaveri oriented districts of
52 Census o f India, 1901, v. 24, M ysore, pt 1, 'R e p o rt’ com piled by T . A n an d a
R ow , B angalore, 1903, pp. 508-10.
Mysore, Bangalore, Tum kur, Kolar, and Hassan than in the
northern districts of the then Mysore princely state: Shimoga,
Kadur, Chitaldrug.53 Moreover, the acknowledged leaders of the
‘Eighteen Phana' group (balagey , or right hand castes), the Banajiga
in or seriously diminished incidence o f the distribution regional
traders, were concentrated in the southern Kaveri oriented, dis­
tricts of Mysore state.54 This is further confirmation of the break
markers along the K arnatak watershed. N orth of this watershed,
there begins a transitional zone which crosses the Tungabhadra
into the Deccan. Thus, the ‘Bom bay-K arnatak’ is treated here
as another shatter-zone boundary of the macro region.

Census o f India 1891; M ysore, v. 25, pt I, ‘R e p o rt', B angalore, 1893, pp. 308-11.
54 A n a n th a k rish n a Iyer, M ysore Tribes and C astes, v. 1, p. 124, Fig. 24.
 Formation of the Medieval
Agrarian Order
Brahman and Peasant in Early
South Indian History1

An agrarian system being a social arrangement involving the uses

of land and its products, it is to those persistent and normative
relationships among social groups that one turns first. The core
of social relationships involving the land in medieval South India
vwas that between Brahmans and peasants. It is this nexus, despite
variations over time and over the complex macro region, which
provides a fundamental defining characteristic of the medieval
period. M odern conditions and a misplaced emphasis upon the
role of the medieval South Indian state have led most scholars
to consider the ‘state’, or formal governmental institutions, to
comprehend most salient aspects of agrarian relations. Such a
view is explicitly rejected here. Operations of the Pallava, Chola,
*and other states did not touch the core of agrarian relations in
medieval South India, though the political framework established
by these states had an obvious relationship with agrarian arrange­
ments o f the age. In general, the segmentary states of medieval
South India assume, rather than create, an agrarian order m ain­
tained and m anaged by dom inant peasant groups, their chiefs,
and prestigious communities of Brahmans. This has been true
from the time o f the Pallavas when the basic structure of agrarian
relations were first clearly exposed by contem porary documents.
W hat is revealed in these Pallava docum ents is an agrarian order
which endured for a millenium.
The period of the Pallavas associated with the line of Simha-
vishnu ( c . a . d . 575-900) has been identified as transitional in a
■ T his discussion, in altered form , w as first published, as ‘B rahm an an d P easant
in E arly S o u th In d ian H isto ry ’, T he A dyar L ibrary Bulletin (D r V. R aghavan F eli­
citatio n V olum e), v. 31-2 (1967-8), pp. 229-69.
num ber of im portant ways related to the development of South
Indian society and culture. Among these are: the monum ental
building, the foundation o f devotional ibhakti) sects upon the
sacred hymns o f the Saivite and Vaishnavite devotees (nñyanñr
and àlyâr ), the efflorescence of essentially rural Brahmanical insti­
tutions as loci o f Sanskrit learning and culture, and the establish­
m ent o f kingship based upon the cakravartin model o f rule over
a territory consisting o f diverse peoples. W ith these changes, an
era in South Indian history was brought to a close. The pre-Pallavan
era, though it continues to be vague in its specific social and cultural
content, was characterized by quite different and distinctive
elements. The m ost im portant of these'were orthogenetic elements
including the Tamil language, territorially segmented2 peoples under
tribal chieftains, folk religious beliefs and practices reflecting
territorially segmented cultures, and a certain degree of urbaniza­
tion in a few widely separated core areas of advanced agrarian
and commercial organization. Coexisting with these pre-Pallavan,
orthogenetic, or as sometimes called, ‘D ravidian’, elements were
heterogenetic styles o f marriage, music, and games as well as a full
complement of religious forms — Hindu, Buddhist, and Jaina
— drawn from all parts o f the sub-continent. The Classical lit­
erature3 o f the early Christian era conveys the sense o f a liveiy,
2 ‘T e rrito ria l se g m en tatio n ’ refers to a ch aracteristic o f social stru c tu re in w hich
im p o rta n t social relatio n sh ip s p ersisten tly co n fo rm to a p a rtic u la r a n d usually
circu m scrib ed territo ry . It is one o f the w ays in w hich social arra n g em en ts occur
in m an y societies, b ut th e degree to w hich such significant relatio n sh ip s as m arriag e,
caste, a n d po litical re la tio n s reflect territo rially w ithin th e In d ian c u ltu ral sphere
n o w h ere exceeds th a t o f S o u th In d ia. C f.M . M a rrio tt, C aste R a n kin g and C om ­
m u n ity Stru ctu re in Five Regions o f India and P akistan, D eccan C ollege M o n o g rap h
Series, no. 23, P o o n a, I960, especially pp. 31-6; J.S. S hahani, ‘A C o m p arativ e
S tudy o f T ra d itio n a l P olitical O rg an izatio n o f K e ra la an d P u n ja b ’, u n p u b lish ed
Ph. D . thesis, U niversity o f L o n d o n , School o f O rien tal an d A frican Studies, 1965;
F .G . Bailey, ‘C losed S tra tificatio n in In d ia ’, A rchives Européennes de Sociologie,
4, no. 1 (1963), pp. 107-24; S tephen F uchs, The G ond and Bhum ia o f Eastern M andia,
A sia P ublish in g H ouse, B om bay, 1960, especially pp. 133-41.
3 T he term ‘C lassical’, is used here in preference to sañgam , follow ing the usage
o f P ro fesso r G eorge L. H art. ‘S an g am s’, o r literary academ ies, o f w hich there were
allegedly th ree, create th e sense o f an in stitu tio n a l co n tex t w hich p ro b a b ly never
existed a n d a false im pression o f th e sources o f th e n u m ero u s p o em s w hich were
an th o lo g ized a tim e lo n g a fte r th eir co m p o sitio n . See H a r t’s ‘A n cien t T am il L ite ra ­
tu r e ; Its S cholarly P a st an d F u tu re ’, originally presen ted to th e S econd C o n ­
ference o f the Society for S outh Indian Studies. U niversity o f W isconsin, 7-9 A pril,
1970 an d p u b lished alo n g w ith o th e r p ap ers in B u rto n Stein (ed.), E ssays on South
rich, and increasingly heterogeneous cultural milieu in which
these elements — endogenous and exogenous — thrived.
There has been a tendency am ong South Indian historians to
emphasize the disjunctive character o f Pallava society and culture.
The reasons for this are understandable. First, there is the fact
that Pallava society was manifestly different from that portrayed
in the Classical works mentioned above. Secondly, Pallava ins­
titutions provided the nascent forms of those institutions with
which we identify Chola society o f R ajaraja Chola, Rajendra
Chola, and K ulottunga Chola I. Finally, there is the rather re­
m arkable paucity o f inform ation about South India between the
often vague and always poetical depictions o f life in the Classical
literature and the relatively more reliable inform ation in Pallava
inscriptions. This hiatus together with the operation of retrospec­
tive vision — that is, seeing in Pallava institutions the foundation
of Chola society — has served to emphasize differences between
the society and culture o f the Classical age and those of Pallava
times. As a result, our understanding of South Indian historical
development has tended to fix the Pallava period as one o f great
change and disjuction.
Against this view of radical change resulting from the establish­
ment of the Pallava line of conquerors from the northern edge of
the Corom andel plain, if not beyond, a different reading of the
evidence plausibly suggests a slower, less disjunctive course of
social change.
This proposed more gradual development of the society and
culture of South India is based upon a somewhat different evalua­
tion of extant evidence. The evolving social order appears dimly
in the earliest Prakrit records of the Pallavas, m ore clearly in their
Sanskrit records, and finally quite clearly in the Sanskrit and
Tamil inscriptions o f the Cholas. Such a proposal must lay heavy
stress upon the processes of accom m odation and assimilation
among a variety of forms, indigenous and foreign, to the Tamil
plain. M any of these forms are already evident in the Classical
literature. These processes continued to operate, acquiring in­
creasing cultural loads and contributing to the evolving character
of Tamil and South Indian culture.

India, T h e U n iv ersity P ress o f H aw aii, H o n o lu lu , 1975, an d H a r t’s recently p u b ­

lished, The Poem s o f A ncient T am ils: Their M ilieu and Their Sanskrit Counter­
parts, U n iv ersity o f C a lifo rn ia Press, Berkeley, 1975.
 The character o f social change inherent in this gradualist
view is more congenial to our notions of how changes occur in
most societies at most times; it also accords better with those
facts which we possess in the particular case of the pre-Pallavan
and post-Pallavan periods in South Indian history. Thus, the
view that Pallavan kingship, architecture, Brahmanical language
and culture were all the consequences of deliberate state policy
seeking to remould the social and cultural forms which we know
from the Classical works must be reassessed. Pallava military
power was sufficient to the task of achieving an extensive over­
lordship in the Corom andel plain. However, Pallava power was
not capable, in itself, of transform ing social and cultural institu- y
tions as they existed in this vast territory. To produce such re­
volutionary changes would have required of a government those
military and political techniques known in the present century,
but clearly beyond the capacity of ancient Indian rulers. N or is
there evidence of a massive invasion or m igration from the ‘N orth’,
wherever that may be considered to be.4
On the contrary, there is very good evidence that the general
process of ‘Sanskritization’5 was already well advanced several
centuries before Pallavan power was extended southward. In
4 W hile th e p resum ed origin o f the P allavas from the n o rth -w est p o rtio n o f the
su b -co n tin en t, as suggested by V. V enkayya (Annual Report, Archaeological Survey
o f India, 1906-7) an d o th ers h as been a b an d o n ed , there is still disagreem ent a b o u t
P allava origins in such sta n d a rd m o n o g rap h s as R. G o p a la n , H istory o f the Pallavas
o fK a n c h i. T h e M a d ra s U niversity H istorical Series, 3 U niversity o f M ad ras, M adras,
1928; C. M 'm akshi, A dm inistration and L ife under the Pallavas, U niversity o f M adras,
M a d ra s, 1938, a n d T .V . M ah alin g am , Kancipuram in E arly South Indian H isto ry,
A sia P u b lish in g H ouse, N ew Y o rk , 1969. Y . S u b b aray alu , w h o h a s u n d e rta k e n
a revision o f G o p a la n ’s w o rk , states his conviction th a t th e P allavas w ere indigenous
to th e A n d h ra p lain (p erso n al com m unication).
3 ‘S an sk ritizatio n ’ o r, its cognates, ‘A ry a n iz a tio n ’ an d ‘B ra h m a n iz a tio n ’ are
p ro b lem atic al term s, an d th o u g h th e re is a grow ing critical lite ra tu re on such
term s, som e convenient w ay is req u ired for referring to the in teractio n betw een
D rav id ian elem ents, o r o th e r su b -cu ltu ral v arian ts within the In d ian cultural
sphere, an d the set o f elem ents w ith w hich the ‘G reat T ra d itio n ’ o f In d ia is associated.
C f.M .N . Srinivas, ‘A N o te on S an sk ritizatio n an d W estern iza tio n ’, Journal o f
A sian S tudies, v. 15 (A ugust, 1956), pp. 481-96 an d A .P . B a rn ab as, ‘S an sk ritiza­
tio n ’, E conom ic W eekly, v. 13, n o. 15 (1961), p p . 613-18. K .A . N ila k a n ta Sastri
uses th e term in his recen t The Culture and H istory o f the Tam ils, F .K .L . M u k h o -
p ad h y ay , C a lc u tta, 1965, p. 18. H a r t in his ‘A ncient T am il L iteratu re . . . . ’, p. 33,
suggests th a t th e re is as m u ch validity in speaking o f the ‘T am iliza tio n s’ o f B rahm ans
as the S an sk ritizatio n by them .
the severa! centuries preceding the Pallavas, this syncretic culture
was linked to the dom inant position achieved by peasant society
over other and older cultural forms. The achievement of dominance
by peasant peoples over others in South India and the firm establish­
ment of those social and cultural forms reflecting and supporting
this dom inance must be considered one of the most im portant
developments in ancient South Indian history. It was a condition
which was to endure for a millenium and contributed to the identity
of one of the m ost durable peasant culture areas in history. The
Pallava period did not m ark its beginning; it was then in a relatively
advanced state of the evolutionary process around which the
present re-evaluation of the period centres.

II
A m ajor feature of Pallava society and that of the succeeding
C hola age was th at it was organized into a large num ber of localities
of peasant society and culture. While the other elements of the
Pallava period mentioned above have been recognized by historians,
the fact o f a prosperous, if dispersed, peasant society has tended
to be overlooked.6 This peasant society, though dispersed, may
be called the dom inant social and economic form of the age. By
this is m eant:
(1) m ost people appear to have lived in settled agricultural
villages ;
(2) peasant agriculture provided the principal means of liveli­
hood for most of the population, directly, or indirectly, through
6 A n id en tificatio n o f the localities o f adv an ced ag ricu ltu re in th e C o ro m an d el
plain d u rin g th e P aliav a p eriod is suggested in the m ap o f B rah m an settlem ents,
M ap IV -1. T he basic tech n iq u e fo r the identification an d m ap p in g o f such places
involves the assu m p tio n th a t the brahm adeya is a valid m a rk e r o f advanced ag ri­
cu ltu ral o rg an izatio n in a locality since these villages o f su b sta n tia l B rahm an
p o p u latio n req u ired incom e for their m ain ten an ce a n d the c o n stru ctio n an d o p e ra ­
tion o f tem p les a n d m athas w hich only p ro sp e ro u s p easan t villages could m eet.
In general term s, th e d istrib u tio n o f brahm adeyas co n fo rm s to th e d istrib u tio n
o f those en v iro n m en tal factors required for ad v an ce d p easan t cu ltiv atio n , especially
h ydro lo g ical facto rs co n n ected with ta n k an d canal irrig atio n . T he p a tte rn w hich
em erges, th u s, is one o f fo u r m ajor clusters: in th e K istn a delta, cen tral T o n d aim a n -
dalam in the basin s o f the P a la r an d C eyyar rivers, the P o n n a iy a r basin, and, of
course, th e K averi b asin. S ubsidiary clusters o f the V aigai an d T a m ra p a rn i ap p e a r
to rem ain stab le u n til a fter th e fo u rte e n th century. Cf. ch. IV.
related service occupations, and peasant agriculture was responsible
for the overwhelmingly large source of wealth to the society at all
levels;
(3) the structure of social relations for m ost people conform
to those of peasant societies in general including:
(a) asymmetrical relations with those powerful enough to
dem and a part of peasant production,
(b) well-developed corporate groups within peasant settlements
and localities, and
(c) effective alliances am ongst various corporate elements.7
It is thus necessary to see the linguistic, architectural, religious, and
broad cultural and political changes with which Pallava society
have been associated against the position already achieved by the
Corom andel peasantry in order to perceive the period in proper
perspective.
M any of the peasant tracts in the Pallava heartland of Tondai-
mandalam, if not opened in Pallava times, were brought to their
full m aturity as significant agricultural territories during the
Pallava period. Then, large-scale, tank-irrigation projects were
carried out to convert the central Tamil plain from a region of
forest and hazardous dry crop agriculture to one of reliable wet
cultivation capable of supporting a dense population.8
Older riverine tracts of the southern Tamil plain had attained
this condition long before when the Kaveri and the western portion
of the Vaigai basins had become the m ajor centres of peasant folk,
and there were isolated concentrations of peasant settlements in the
central plain of equal antiquity judging from the early prominence
o f K anchi.9 The urban centres o f K averippattinam , M adurai,
and Kanchi which had earlier flourished in these older tracts of
7 A n excellent discussion o f p easan t society is th a t o f W olf, Peasants. Because
it is a general w o rk on th e subject it can n o t be expected to deal w ith the im plica­
tio n s o f caste m o re th a n cursorily, an d th ere is little a tte n tio n given to the rath er
u n iq u e p ro b lem s o f a n e x p an d in g p ea sa n t system because th e w o rk is addressed
to an th ro p o lo g ists.
8 T his significant aspect o f the P allav a p eriod is m en tio n ed by M inakshi, op.
cit., pp. 94-100; G o p a la n , op. cit., p. 155; a n d K .R . S u b ra m an ian , ‘N ew Light
o n th e P allav a P e rio d ’, M aharaja's College M agazine, V izian ag aram , v. 6. no. 1, n.d.
pp. 1-9.
9 K .A . N ila k a n ta S astri, ‘T h e B eginnings o f In terco u rse betw een In d ia an d C h in a ’,
The Indian H istorical Q uarterly, v. 14 (1938), p. 386, presents evidence for C hinese
tra d e w ith K a n c h i i n th e seco n d cen tu ry B .C .
advanced agriculture continued to exist as im portant centres,
but there was a decisive shift during the Pallava period from such
isolated territories with their cities to new agricultural tracts and new
rural centres. By the time of K ulottunga C hola I, the Coromandel
plain between the two deltaic regions of the K istna-G odavari and
the Kaveri had become a region of peasant agriculture and society,
constituting a prim ary zone of influence over the propinquitous
interior upland in the same way that the Gangetic plain did for
m ajor portions of northern India. However, even earlier, during
the Pallava period, the northern and southern portions o f the
plain were linked and the basic form ation of the greatest southern
variant of Aryan civilization achieved. Here, in the Coromandel
rural setting of the Pallava period, Sanskrit and Brahmanical
knowledge were firmly established, the bulk of the Brahm an
custodians of this knowledge lived, the saints of Saivism and
Vaishnavism spent their lives, and a large population of peasants
lent their support to the maintenance of this culture.
The basis o f close co-operation between the peasant cultivators
of the Corom andel plain and the Brahmans who lived as their
spiritual preceptors and neighbours can best be understood as an
alliance, peasants, generally, are related to non-peasants in three
broad ways. First, there are relationships with those whose power
is sufficient to successfully demand a part of the proceeds of peasant
cultivation. Second, there are more complex asymmetrical re­
lationships between peasants and those in subordinate positions.
Such subordinately related people would include non-peasant
peoples brought into the lower strata o f peasant society by the
expansion o f th at society into previously forested and dry tracts;
other subordinate people were those in peasant settlements whose
occupations were of low value or even considered antagonistic to
an expanding and increasingly powerful peasantry, such as mobile
artisan-traders were for a considerable time. Somewhat less subor­
dinately related to any localized peasant groups were various other
cultivating groups, agricultural traders, and artisans closely linked
to agricultural production. In contrast to these asymmetrically
linked relations o f peasants and others was a third set o f relations
with rural groups th at were m ore nearly symmetrical. These relation­
ships took the form o f tem porary coalitions, as when peasant groups
o f propinquitous localities com bined to achieve some common
endeavour (e.g. defence), or such relations might be more enduring
as in the case of relations with Brahmans. In this case, the term
‘alliance’ seems appropriate.
To reconcile this classification of essentially power relationships
based on political and economic factors with caste relationships
is not easy. Even in the Classical period,, there was a system of
ranked relationships among groups, and certainly, in the Pallava
period, ranked relationships according to ritual purity, as enjoined
by the sastras, were extolled. But alliances do not require equality
among the participants; an agreement about com m on ends and
mutually congenial m eans suffices. The fact of caste relationships
cannot be ignored even in early Pallava times, but such relationships
can be discussed within the classification of power relations in­
volving th e Corom andel peasantry at the time.
Relations between Corom andel peasants and those with a
sufficiency of power to demand a regular portion of peasant pro­
duction were distinctive when considered with respect to other
parts of India and other peasant societies. The idealized system
according to which such relationships operated in India was that
involving Kshatriyas. In northern India, this ideal was realized
to a large extent after the seventh century, but not so in South
India. Elsewhere in India, rural dominance was the prerogative of
groups o f warriors enjoying .high ritual status, whose locality
power was m aintained through extensive networks o f agnatically
and affinally related w arrior families. In South India there cer­
tainly was awareness of Indo-Aryan, varna -organized society
in which decisive secular authority vested in Kshatriyas. However, 5
apart from certain great royal lines of w a rrio rs— such as the
Pallavas, Cholas and Vijayanagara where cakravartin status was
claimed and a royal style m aintained.— few locality warrior
families achieved the extensive, prestigious, kin-linked organiza­
tion o f northern warrior/kingly groups.10 Why the Kshatriya
institution never challenged the secular authority o f Brahmans
and dom inant peasants, the alliance between whom formed the
10 T h ere is a basic difference, o f course, betw een the cakravartin arid K sh atriy a
ideals o f rule, an d , at best, th e tw o m odels can n o t be m o re th a n com plem entary'
th o u g h h istorically th a t balance has n ot been achieved often. In S o u th T n d ia , the
cakravartin m odel w as im p o rta n t fro m P allav a tim es, .-but K sh atriy as, th at is locally
based w arrio rs w ith ritu a l sta tu s sufficiently high to share high sta tu s w ith B rahm ans
did n o t em erge.. L acking high ascribed statu s in a society w here this w as significant,
p easan t w arrio rs o r those o f in ferio r origins w ere alw ays dep en d en t u p o n B rah m an
co -o p eratio n .
keystone o f local South Indian societies, is too complex an
issue to be considered m ore than briefly here. The persistent
im portance of quite narrow territorial segmentation of significant
social relations, which inhibited widespread m arriage networks,
appears to be involved here. Also im portant was the fact that
m ost locality warriors o f South India, in Pallava times and some­
what later, were obviously of peasant origin and derived a part
of their local authority from their continued identification as
such; many other warriors of lower social origins (e.g. hill people
on the fringes of peasant society) appeared to have been content
to rise to the status of respectable peasants. In addition, there
was no conquering élite which might seek to preserve its identity
through putative Kshatriya rank. But perhaps the most im portant
reason for the failure of a Kshatriya tradition to emerge in medieval
South India was the entrenched secular power of Brahmans.
Collaboration with would-be K shatriya warriors could not streng­
then, but only weaken Brahm an secular authority. Since Brahmans
were firmly anchored in a satisfactory alliance with localized peasant
groups and their chiefs, there would have been no inducement
for Brahm an collaboration with aspirants to Kshatriya status.
The Corom andel peasantry constituted a social system of
significantly stratified interactions. Because this peasantry was
constantly expanding and assimilating non-peasant peoples of the
forests and dry plains who were often regarded as inferior, the
m atter of superordinate-subordinate relations within the peasantry
is im portant. Though the evidence is sparse, it m ay be supposed
th at there was a constant process o f social grading am ong the
various groups o f the established peasant tracts in the Coromandel
plain in which the m ore powerful land-controlling groups sought
to differentiate themselves from other agrarian groups. According
to the late Classical works, including the C ilappadikaram (possibly
of the fifth or sixth century), there was a rudim entary hierarchical
ordering o f social groups in which ulavars, or cultivators, were
regarded as the first people. Beneath the ulavars , also called
vellâlar and kâralar, were ranked cowherds and shepherds (ciyar
and kovalar), hunters ( vëdar ), various artisan groups, armed men
(padaiyâcciar) and, in the lowliest stratum , fishermen ( valaiyar )
and scavengers (pulaiyar).n The precise nature of status dif­
11 V. K a n a k a sa b h a i, The Tam ils Eighteen H undred Years A g o , H ig g in b o th am &
C o., M a d ras, 1904, pp. 113-14; P .T . S rinivas Iyengar, H istory o f the T a m ils; From
ferences suggested in these literary works cannot however be
ascertained. Another, somewhat less ambiguous, vantage point
on ranked relations in the early peasant society of South India
occurs with the full development of that characteristic feature of
social structure; the bifurcation o f middle and lower castes into
what am ong Tamil and K annada speakers has been called castes
of the right-hand (valangai-balangai) and castes of the left hand
(idangai-edangai).12
Always in the background of these competitions among segments
of localized rural groups, was the slow accretion o f new people.
As new tracts of land were opened by peasant colonists, those who
had formerly occupied the land either fled more deeply into the
forests and hills or found a place in the new order. Places had to
be found for the latter within the expanding peasant tracts of
Corom andel because the conversion o f dry tracts o r forest to wet
field cultivation required the labour o f many. O f the new people,
some by their industry and good fortune overcame the stigma which
must have attached to their previous non-peasant way of life and
eventually acquired land of their own; others found or possessed
a skill which was needed in the new settlement; most of the new
folk, however, attached themselves to those with enough land
to com m and a following of dependents — a m ark of power and
respectability then, as now, am ong Indian peasants.13

Ill

Surpassing all other relationships within the peasant society of

this ancient period, however, was the close co-operation between
Brahm ans and respectable cultivating groups. Their alliance,
their sharing of control in the num erous localities of advanced
agriculture into which the Corom andel plain was divided, was
the distinctive social and political element up to the fourteenth
century. M oreover, there can be no question that the relationship

the Earliest Tim es to 600 A .D ., C. C o o tn arasw am y N aid u , M a d ra s, 1929, pp. 565,602.

12 T h e su b ject o f th e d u al division is discussed ab o v e, ch. V.
13 A g ricu ltu ral la b o u r g ro u p s a tta c h e d them selves to either B rah m an o r V ellala
lan d ed g ro u p s fro m a n early tim e. Cf. E. T h u rsto n , Castes and Tribes o f Southern
India, v. 5, G o v ern m en t Press, M a d ras, 1909, p. 473 an d M anual o f A dm inistration,
v. 3, p. 1037.
between Brahm ans and dom inant cultivators was entirely voluntary
and recognized as mutually beneficial. The benefits to Brahmans
are most obvious inasmuch as they are recorded in thousands of
inscriptions on the walls of temple buildings which themselves
are m onum ents of the scale of this peasant support. That the
gifts of money, of a regular portion of income from peasant villages
or o f labour involved the invocation and ostensible ratification of
the Pallava rulers, their immediate families and officials or, more
often, local w arrior chiefs, does not diminish the essentially volun­
tary nature of the gifts from locality folk who com m anded them
in the first instance. The determ ination by local people of how
their resources were to be apportioned is one of the clearest features
of Pallava times.
A nother point, which has been inadequately appreciated with
respect to peasant support of Brahmanical institutions in numerous
rural settings during Pallava times and later, is th at the towns of
the Corom andel region were bastions of Jaina and Buddhist
influence. This fact is attested by the inscriptions of such places
as well as by the accounts of Buddhist pilgrims like Hsiian-tsang.
Lavish support extended by townspeople to these heterodox
faiths, if it did not reflect hostility towards Brahmanical institu­
tions, certainly must have appeared to Brahmans as proceeding
from a dangerous ecumenical sentiment. Thus, it should be recog­
nized that for learned Brahmans and their ritual colleagues, it was
the Corom andel countryside which offered the best, if not the only,
situation for support of the m ultifarious activities associated with
resurgent Brahmanism during the Pallava period. Other benefits
to both parties in the alliance will be discussed more fully below.
The Corom andel countryside had passed through several
centuries of im portant change by the Pallava period. Peasant
societies had im planted themselves over m uch o f the plain and had
been subjected to stern challenges from non-peasant peoples.
The most distinctive characteristic of the Corom andel plain north
of the Kaveri basin, the oldest and m ost dense region of peasant
agriculture and settlement in the region, was that of numerous
and scattered peasant localities separated by large and small
tracts o f inhospitable land. Excessive slope, aridity or forest
cover and other factors would have reduced the suitability of these
tracts for peasant agriculture. In such inhospitable places, non­
peasant peoples could and did live.14 The highly variegated and
discontinuous physical character o f the southern peninsula and
the social adaptations to it had produced a set of quite stable
social forms in early South India. These are prom inent in Classical
literature where a fivefold classification o f m an-nature situations
(tin ai) was recognized.15 In the early centuries of the Christian
era, these physiographic categories may be considered as culturally
significant with respect to settings in which men of the south­
eastern peninsula lived. Then, peasant folk, ulavar , of the marudam ,
were only one among several territorially segmented social and
cultural subsystems. Each of these were different in essential
ways, yet all comprised a single general culture area with shared
linguistic and other cultural elements. By the ninth century,
peasant society had become dom inant over, without entirely
eliminating, the hunting, fishing and pastoral peoples, and without
reducing the territorially segmented social organization which
continued to ex ist— even flourish — as a structural factor of
great im portance in South India. A recurrent poetic theme accom ­
panied the rise of peasant groups and settlements. This theme was
the fear and loathing which men of the hills and dry plains inspired
in those o f the plains. The K alittokai o f perhaps the fourth to
sixth centuries refers to the maravars of the dry plains and hills
in the following term s:
14 T h e general d istrib u tio n o f plains a n d forest a n d / o r u p lan d tracts in T am il
co u n try m ay be found in: K . R a m a m u rth y , 'S om e A spects o f the R egional G e o ­
g rap h y o f T a m iin a d ’, Indian Geographical Journal, 23, nos. 2, 3, 4 (1948) and 24,
nos. 2 & 3 (1949); also see B .M . T h iru n a ra n a n , ‘T h e T ra d itio n a l L im its a n d S ub­
divisions o f the T am il R e g io n ’, in K .V . R angasw am i A iyangar C om m em oration
Volum e, G .S . Press, M a d ras, 1940, pp. 159-69. T he social aspects o f this re latio n ­
sh ip in S o u th In d ia is very u nclear still, b u t its general featu res m ay be p resum ed
to follow p a tte rn s w hich h ave been suggested in the follow ing w orks on In d ia :
D .D . K o sam b i, ‘T h e Basis o f A ncient In d ian H isto ry ', in tw o p a rts in Journal o f
th e A m erican O riental Society, v. 75, nos. 1 & 4 (1955) an d B.A. S aletore, The W ild
Tribes in Indian H isto ry, M o tilal B a n arsid as, L ahore, 1935. F o r im p o rta n t dis­
cu ssio n s o f th e p ro b lem in S o u th east A sia, see, fo r exam ple, E .R . L each , P olitical
S ystem s o f H ighland Burm a, H a rv a rd U niversity P ress, C am bridge, M ass., 1954,
a n d R o b b in s B urling, H ill F arm s a n d P adi Fields in M ainland S outh-E ast A sia,
P re n tic e-H a ll In c., E nglew ood Cliffs, N .J ., 1965.
15 P .T . S rin iv asa Iy engar, H isto ry o f the T a m ils; From the E arliest T im es to 600
A .D ., U n iv ersity o f M a d ra s, M a d ra s, 1929, p p. 3-12. a n d p a ssim ; also S. V aith ian a-
th a n , ‘T h e P a ttu p p a ttu : A H isto rical, Social, a n d L inguistic S tu d y ’, u n p u b lish ed
P h .D . thesis, S chool o f O rien tal a n d A frican Studies, U niversity o f L o n d o n , 1950.
 O f strong limbs and hearty frames and fierce looking as tigers, wearing
long and curled locks of hair, the blood-thirsty mqravars armed with
bow bound with leather, ever-ready to injure others, shoot their arrows
at poor and helpless travellers, from whom they can rob nothing, only
to feast their eyes on the quivering limbs of their victims. . . . The wrath­
ful and furious maravar . . . the loud twang o f whose powerful bow
strings, and the stirring sound of whose doubieheaded drums, compel
even kings at the head of large armies to turn their backs and fly. . .'6

In ‘The H unters’ Song’ of the Cilappadikaram a priestess chides

the maravar and eyinar hunters for failing to keep their vows to
destroy the gardens of their enemies and for ceasing their practices
of plundering passersby as a result of which hill villages suffered
and those o f the plain prospered.17
These literary references alert one to the hostility which appears
to have existed between the lowland plain people, whose power
increased in pre-Pallavan times, and the dangerous people of the
hills and dry plains. By late Classical times, wealthy and populous
peasant communities had probably succeeded in assimilating most
pastoral and fishing groups. To the latter, substantial advantages
could be offered by the peasantry even as the superior resources
of the peasants of the fertile lowland proved attractive to and
resulted in the assimilation o f hill and dry tract people on the
fringes of peasant settlements. However, a substantial population
o f non-peasant folk remained in scattered, isolated pockets.
Between the peasants and their associates and these non-peasant
people there existed the same prolonged tension as between plains
and hill people elsewhere in the sub-continent as well as in many
parts o f South-east Asia.
Though there are im portant parallels in the relationship of
peasant folk to non-peasant folk in remote hill settlements in
South India with those in other parts o f Asia, the duration of
antagonism between the two in South India appears notable.
Peasants were obliged to deal cautiously with such non-peasant
people even after peasant society and culture had become dom inant
in South India. One reason for this lengthy struggle was the fact
of dispersion. The discontinuous pattern of peasant settlement
made peasant localities particularly vulnerable to raids. Efforts
to clear hill tracts o f their fierce occupants had to be considered
16 C ited in K a n a k a sa b h a i, o p .cit., p p. 42-3.
17 C ited in R a g h a v a n , ‘N o t e s . .
circumspectly. Even as late as the sixteenth century, the great
Krishnadeva Raya counselled diplomacy and caution in dealing
with the people of the hills and dry forests.18
A somewhat ironic factor accounting for the existence of tension
between peasants and those of hills and dry plains in South India
was that most of the latter shared to a greater extent than similar
peoples elsewhere the culture of the peasantry. They were never
a people apart — to be ignored or massacred — as in parts of
South-east Asia. A measure of this shared culture has been the
ease with which non- or partially-peasantized warriors of hills
and dry plains were able to establish themselves as masters over
peasant peoples until the nineteenth century. D uring the pre-
Pallava period, when conditions were fluid and the peasant frontier
constantly expanding, physical and cultural proximity provided
the opportunity for close interaction even within a generally
hostile and competitive context.
A turning point in the relationship between the peasant peoples
of the plains and their non-peasant adversaries may have come in
the still mysterious period before the seventh century. According
to the scanty literary and inscriptional evidence relating to that
period, peasants of the Corom andel plain, especially in its southern
extremity, were subjected to the control of a people or peoples
who were recalled with terro r.19 When, for how long, by whom
and which o f the Corom andel peasantry were thus subjugated
is not clear. W hether it was a single conquering people from beyond
the Tam il plain, as has been suggested, or from within the region,
and whether the conquest was that of a single people or many,
are queries unanswerable from the extant evidence. The name
most often associated with these conquering people is kalabhra
(Pali : kalabba) of whom it was said they abused families of local
chiefs of the plain and the Brahm ans of the villages; another name
is kalavar.20 While they were different from the plains people,
18 A. R an g asv am i S arasvati, ‘P olitical M axim s o f the E m p e ro r-P o et, K rish n a ­
deva R a y a ’, Journal o f Indian H istory, v. 4, pt 3 (1925), pp. 61-88; also cited in
S aletore, W ild Tribes in Indian H istory, p. 12.
19 T he sources are review ed by N ila k a n ta Sastri in The Colas, v. 1, U niversity
o f M ad ras, M ad ras, 1935, pp. 119-21 an d include as m o st im p o rta n t, the V elvikudi
g ran t o f th e late eighth cen tu ry an d B u d d h a d a tta ’s M an u al. Cf. S rinivasa Iyengar,
op cit., pp. 436-7.
20 A ttem p ts to link the. Kajabhras, an d kalarars w ith the k allars o f the so u th e rn
T am il p lain have n ot been generally accepted-, cf. ibid., pp. 437-8, 535.
these conquering folk are, at least later, were adm ired by some
Tamil poets as upholders of some Hindu, Buddhist, and Jaina
institutions.21
This ‘K alabhra interregnum ’, as it is sometimes called, may
m ark a point where non-peasant people made their strongest bid
to control some, at least, of the lowland, peasant population; this
may have been in the very southern portion o f the peninsula since
Pandyan inscriptions are the source o f the m ost vehement
judgements of Kalabhras. W hether in Pandya country alone or
elsewhere, this onslaught appears to have been the culmination
of a long period o f non-peasant, armed resistance to the expansion
of peasant society, and it was distinguished by the attem pt of the
‘interlopers’, as N ilakanta Sastri has called them,22 to establish
their sway over peasants without, apparently, adopting substantial
elements o f Corom andel peasant culture. This attem pt failed,
and with the assertion o f Pallava w arrior control over the northern
and central portion o f the plain and that o f the Pandya rulers
over the southern Tam il plain by the late sixth century, the dom i­
nance o f the peasant people and the society which they m aintained
was never again lost. True, warriors from the hills continued to
raid and plunder peasant settlements at the fringes o f peasant
core areas, and they occasionally set up durable power am ong the
peasant people — as did the Hoysalas in K arnataka23 — but
such durable power was facilitated by the adoption of the two
elements rejected by the K alabhras: respect for and support of
Brahmanical institutions and Hinduism as well as recognition of
locality chiefs who were, in m ost cases, members of the dom inant
peasant group of the locality. After the seventh century, confirmed
in military dom inance and cultural superiority, peasant society
continued its steady encroachment upon non-peasant peoples,
at times slowly, at other times rapidly.
The full implications of the so-called ‘K alabhra interregnum ’
and the establishm ent o f Pallava power over much of the Tamil
plain have only recently begun to be appreciated from the cultural
and social point o f view. N ilakanta Sastri has stated:
We may perhaps surmise the Kalavar-kalabras were a widespread tribe
whose large scale defection to the heretical faiths [Jainism and Buddhism]
resulted in a political and social upset lasting over some generations.24
21 Ib id ., p. 535.
22 N ila k a n ta S astri, The C olas, v. 1, p. 121.
 That this intrusion upon the expanding order of peasant peoples
h ad im portant consequences seems quite clear. The reference
above to ‘heretical faiths’ is significant because the K alabhra ruler
and conqueror of M adurai, Accuta Vikkanta, is said by the writer
B uddhadatta to have.been his patron. Some of the songs celebrat­
ing this Accuta are reported by Amitasagara, a tenth century Jaina
gram m arian.25 Moreover, in the centuries between the Classical
works and the Pallavas of Kanchi, the zenith of Jaina and Buddhist
influence in South India was achieved. D uring this period, the
influence of Aryan cults, the earliest evidence of which are found
in caves of about the third to the first century B .C . bearing Tamil-
Brahmi inscriptions,26 increased in South India. This contributed to
the increasing tempo o f cultural change.27 It also appears that during
the same period these cuLts coexisted peacefully w ith each other and
with indigenous forms of religion and that the Jaina and Buddhist
sects of South India were as successful as Saivite and Vaishnavite
sects in winning the allegiance of leaders in South Indian society.28
Reference has already been made to the support of urban people.
Am ong warriors, Accuta Vikkanta K alabhra was^ probably a
Buddhist, and M ahendravarm an I, the Pallava ruler of the late
sixth century began his reign as a Jaina and is supposed to have
persecuted Siva worshippers until the time of his conversion by
the Saivite saint (nayanár) A ppar (Tirunavukkarasu). As a. result
o f the support o f the powerful, Jainas and Buddhists could boast
of a num ber of impressive monastic establishments in many parts of
South India.29 The Chinese pilgrim, Hsüan-tsang, in the middle

23 J.D .M . D e rre tt, The H oysatas, O .U .P ., M a d ras, 1951, p p . 16-19.

24 N ila k a n ta S astri, C ulture an d H isto ry . . ., p. 19.
25 N ila k a n ta S astri, The Colas, p. 120 a n d his Culture and H istory . . ., p. 19; also
S rin iv asa Iyengar, op. cit., p. 528, dates B u d d h a d a tta in the late fifth century.
26 K . Z velebii, The Sm ile o f M urugan, E .J. Brill, L eiden, pp. 140-1.
27 S rinivasa Iyengar, o p .cit., p. 18.
28 E vidence o f this is m assive: cf.M .S . R am asw am i A yy an g ar a n d B. S eshagiri
R ao , S tudies in South Indian Jainism , H oe & C o., M adras, 1922 a n d the w orks o f
o f P.B. D esai. It is o f in terest th a t p erh ap s th e earliest, im p o rta n t T am il inscription
reco rd s a n en d o w m en t to a Ja in te a c h e r: K. Z velebii, T am il in 550 A .D .: A n Inter­
pretation o f Early Inscriptional Tam il, O rien tal In stitu te, P rague, 1964.
29 A ppax him self, as a Ja in a m o nk, before his reversion to Saivism , served in
a m o n aste ry at C u d d a lo re, S o u th A rcot. S cholars have p o in ted to the significance
o f th e n u m b er ‘63’ used first by Ja in as, th e n by Saivites for th e m ost h o n o u re d saints,
an d the tra n sfo rm a tio n o f niches o f Ja in a shrines to h o ld Saivite shrines w hen the
seventh century, commented on such establishments in various
parts of the Corom andel plain, and the latter reported that his
own Buddhism was giving way to Digam bara Jainism .30 These
accounts, plus numerous inscriptions, including early Tamil
Brahmi ones and others of the sixth century, bear eloquent testi­
mony to the widespread support enjoyed by ‘heretical’ centres
from the powerful and the wealthy.
Under the circumstances of apparently peaceful competition
among the religious cults in pre-Pallava times, the ‘heretical’
sects of Buddhism and Jainism were in thriving condition. Scholars
have noted with surprise that Hsiian-tsang, in a . d . 642, seemed
completely unaware of the bhakti revival which was being carried
forward by Saivite and Vaishnavite hymnists. To assume that
his indifference to this movement was a consequence of this Chinese
pilgrim’s prejudice for or a preoccupation with Buddhism may be
justified, but he did report the relative and growing success of
Jainas — and may have faithfully reflected the favour which
Jainism found with warriors prior to and during the early years
of Pallava rule in South India.31 There is also good reason to
believe that the worship of Jaina goddesses (yakshinis) inspired
popular support.32
It hardly seems possible to separate the ‘K alabhra interregnum ’,
or what is perhaps better understood as the extension of non­
peasant, w arrior control over the plains, from the favour shown
by these warriors to Jainism. Rather, it seems appropriate to
suggest that the attractiveness of Jainism was that it permitted
a warrior to achieve legitimacy and ‘A ryan’ respectability without
necessarily accepting the elements of contem porary peasant culture
with which Hindu sects had become associated at the time. Among
such elements would have been the high secular place accorded
to Brahm ans and the primacy of sedentary peasant agriculture.
Certainly, the correlation of non-peasant warrior power with

build in g s w ere ta k e n o v er by th e latter, as at M a d u rai. N ila k a n ta S astri, Culture

and H isto ry . . . . pp. 109-10 an d R am asw am i A yyangar, o p .cit., p. 78. A lso see:
P.B. D esai, Jainism in S outh India, Ja in a S am skrti S am k ak sh a S angha, S h olapur,
1957, an d R. C ham pakaS akshm i, 'Jain ism in S o u th In d ia ’, u n p u b lish ed M. L itt.
thesis. D e p a rtm e n t o f H isto ry , U niversity o f M a d ras, 1958.
i0 N ila k a n ta S astri, Culture and H istory . . . , pp. 113-14.
31 M in ak sh i, o p .cit., pp. 213-38.
32 T h is is arg u ed by P.B. D esai, o p .cit., pp. 38-40, 72-4.
Jainism compels a consideration of Jainism not so much as a
‘heretical’ sect but as an ideological element in the critical period
of struggle between the militarily formidable non-peasant people
of South India and perhaps peasants of dry, and mixed dry and
wet ecotypic situations in South India against the increasingly
prosperous and im portant agricultural people of the riverine
plains. Just as the ascetic or monastic figure of the Jain monk
symbolized one form o f sacredness and the grhastha Saivite or
Vaishnavite Brahm an symbolized an alternate form o f sacredness,
the m orally complete and ritually independent Jaina king re­
presented one kind o f royal legitmacy and the ritually dependent
Hindu king, with his Brahm an purohita and royal sacrifices re­
presented another kind.33 The bitterness and violence with which
sectarian controversy erupted by the seventh century suggests
an ideological com ponent of considerable importance.
During the seventh and eighth centuries, the tolerant relations
am ong religious sects in South India had clearly come to an end.34
M ahendravarm an I persecuted Saivites until his conversion; then
he turned on Jainas.35 Later, the Saivite saint Sam bandar who
converted the Pandyan ruler, is celebrated in an annual festival
at the M inaksi temple of M adurai which commemorates the
impalement of 8,000 Jaina heads at the young saint’s urging.36
Still later, in the eighth century, Nandivarm an II Pallavamalla,
an ardent Vaishnavite, carried out persecution of Jainas and Bud-
33 T his statem ent is fully elab o rated in an essay by the a u th o r entitled, ‘All the
K ings’ M a n a : P erspectives on K ingship in M edieval S outh In d ia ’, in a volum e
o f essays, en titled Kingship and A uthority in South Asia, under the ed ito rsh ip o f
Jo h n F . R ich ard s by th e C enter o f S o u th A sian S tudies o f the U niversity o f W isconsin
in 1978. T he term s 'id eo lo g y ' and 'id eo lo g ical' inevitably present difficulties.
Ideology is ta k e n to be rh eto ric, or persuasive speech w hich p ertain s to m oral issues
in social arra n g em en ts. It is therefo re a form o f m oral reaso n in g intended to p rovide
a basis for a d ju d ic a tin g conflicting claim s or ju stifying p a rtic u la r arra n g em en ts
in a society. Ideology c o n tra sts w ith theology in its concern w ith conflictful relations
a m o n g m en ra th e r th a n a m o n g gods o r co n cep tio n s o f deity.
34 T his p ro p o sitio n seem s generally accepted am o n g P allava h isto rian s even
th o u g h M in ak sh i, op. cit., p. 170, could state after review ing the evidence o f P allav a
p ersec u tio n : '. . . P allava m o n arch s as a class [?] w ere to leran t to w ard s these
religious sects.'
35 S u b ra m an ian , 'N ew Light . . p. 7.
36 R e p o rte d by N ila k a n ta S astri, Culture and H istory . . . , op. cit., p. 110. H e
co u ld n o t cred it the n o tio n th a t the saintly S am b an d ar conceived this h o rro r, but
ack now ledged th a t the festival continues to be celebrated.
dhists, and his contem porary, the Vaishnavite hymnist Tirumangai
is said to have plundered the Buddhist vihara at the town of Naga-
pattinam using the golden image to finance the construction of
walls around the principal shrine at Srirangam and other benefices.37
A nother aspect of this complicated interplay of religious acti­
vities and power relations throughout the Corom andel plain
during the pre- and early Pallava period is that of the bhakti
movement. Hymns o f Saivites ( Tevdram) and Vaishnavites (Ndldyira
Prabandham) were the works of those from all social strata, from
Brahman to untouchable. At a time when Siva and Vishnu worship
was apparently still dom inated by Brahman votaries of the jnaha-
yoga tradition, these works reflect an impressive cult devotionalism.
Between the chaste religion of jnahayogins and the ‘excesses’ of
such' Saivite cults as the Pasupatas, Kapalikas, Kalamukhas,
and others,38 bhaktas of the hymnal tradition presented a religion
apparently suited to the peasant society which was achieving
supremacy over the non-peasant peoples.39 Theirs was a religious
tradition well-rooted in the devotional faith of peaceful people
o f the plain. It had other advantages as well. It was congenial
to Brahm an religious leaders in its philosophical presuppositions,
and it offered a powerful theological and ideological counter to
Jainas and Buddhists. Indeed, many of the Saivite hymns condem n­
ed Jainas and Buddhists and the N dldyira Prabandham of the
Vaishnava bhaktas castigated Jainism .40 The attack upon Jainism
and Buddhism attained its philosophical culm ination in the work
of Sankara by which time, for all practical purposes, the victory
of the new, devotional orthodoxy over Jainism and Buddhism
had been assured as a consequence o f the assimilation of folk
religion and by the crucial shift o f popular support to the puranic
cults of Siva and Vishnu.
The tension between peasants and non-peasants, the ‘Kalabhra
interregnum ’ as an event in that prolonged and hostile relationship,
and the ideological significance of religious controversy in connec­
tion with the competition between peasant and non-peasant powers
in the Corom andel plain are all factors which appear to be related
37 Ib id , pp. 112-13 an d M inakshi, op. cit., pp. 170-1.
38 D .N . L o ren zen , The K apalikas and K a la m u k h a s; Two L ost Saivite Sects, U n i­
versity o f C a lifo rn ia P ress, Berkeley, 1972.
■
w N ila k a n ta S astri, Culture and H istory . . . , pp. 11 1-12.
40 M in a k sh i, op. cit., p. 170.
to the peasant-Brahm an alliance of the Pallava period. To the
extent th at scholars have concerned themselves with Brahman-
peasant relationships heretofore, two different kinds of explana­
tions have been favoured. The first has presumed that the con­
quering Pallavas instituted not only political-m ilitary changes
assuring ascendancy over the people o f Tondaim andalam , but
simultaneously they set into m otion the linguistic, architectural,
and comprehensively ‘Brahm anical’ changes of which mention
has been made. It is further implied that the Pallava state used
force, or the threat of force, as well as its com m and over the wealth
of the territory, to establish landed Brahman settlements throughout
the central Corom andel plain. An alternative line of explanation
advanced to explain the manifestly voluntary nature of peasant
support to Brahm ans is that peasants, recognizing the special
learning, piety, and probity of Brahmans, subordinated their
institutions such as the peasant village assembly (ür) to those of
Brahmans, such as the Brahm an assembly (mahásabhd) and allo­
cated a substantial portion of their wealth as well.
Both of these explanations are somewhat flawed. If Pallava
rulers were powerful enough to have brought about the changes
attributed to them — a highly doubtful proposition in itself —
why would they have perm itted such complete self-government,
including the right o f Brahman villages to dispose of not only their
own resources but those of dependent villages as well? Were the
Pallavas so different from other Indian warriors and overlords
that resource-control m eant less to them, or were they more deeply
pious than others ? Could not their piety have been served as well
by m aking direct endowments for specific Brahmanical under­
takings without surrendering such a substantial degree of political
and social control? This was done in other times in South India
or in other parts of India. As to the putative piety of Coromandel
peasants, it is clear that there had existed for several centuries
prior to the establishment of Pallava power a sympathetic rela­
tionship between peasants and Brahm ans; certainly, Brahmans
had lived among and upon the generosity of the peasantry in those
centuries, and even if Brahm ans shared religious functions and
m aterial rewards with others, including ‘heretical’ teachers and
non-Brahm an ritual functionaries, they held an esteemed place
in Corom andel society. M oreover, there is no convincing evidence
that the Pallavas ever achieved sufficient power or organization
to assert their authority effectively over the powerful local institu­
tions of Brahmans and peasants of which we have considerable
evidence.
Since an alliance is a voluntary association to achieve particular
ends, the most plausible answer to the question of why an alliance
existed — or why people or groups who were not constrained to
act together, did so — is in the identification of the interests which
appear to be satisfied by the alliance. The Brahm an-peasant alliance
of the Pallava period was based upon the convergence of im portant
interests which came to exist between those who cultivated the
land along with their dependents and those who by their sacral
functions possessed a powerful ideological capability. The bene­
fits were mutual and alliance durable. For Brahmans, the ad­
vantages were the following. Jaina and Buddhist sectarians had
positioned themselves strongly in the towns of Coromandel and
among many warriors, especially those beyond the core areas of
peasant society where tribal organization remained im portant.
The remains of early Jaina shrines in many ancient peasant villages
of the plains suggest, moreover, that the Jaina goddess (yakshihi)
cult was held in high favour by many peasants. Brahmans, on
the other hand, were perhaps most securely situated in the many
peasant villages and localities of the plain and had already estab­
lished close relations with dominant peasant groups many of whose
ritual requirements they met. The expansion o f peasant society
before the seventh century provided Brahmans with the most
promising basis for maintaining and extending their position.
Moreover, adoption of devotional, temple-centred forms of
ritual by Brahmans required a new scale of support which peasants
of the plains could best provide. In this connection, one cannot
overlook the relationship between the ritual requirements of the
devotional religion and the need for large concentrations of Brah­
mans in the Corom andel type of brahmadeya which makes its
appearance during Pallava times. Apart from towns, and these
were relatively few and inhospitable to Brahmans for a long time,
wealth sufficient to support devotional temple-based Hinduism
existed only in prosperous peasant villages. F or reasons of their
own, dom inant peasant groups were ready to deepen the existing
relations with Brahmans supporting the Brahman claims to ritual
primacy and the norms of stratified interaction required by caste.
The corresponding interests of the Coromandel peasantry
appear to be the following, Having come through a period when
non-peasant power seriously threatened their security — of which
K alabhra control, especially in Pandya country, was perhaps only
the clearest example — peasant groups may have felt considerable
need for greater ideological coherence as a means of unifying the
diverse segments o f peasant society against similar threats in the
future. Some protection was certainly afforded by Pallava rule
after a . d . 550, and the Pallavas had broad peasant support in
recognition of that protection. Beyond this threat of conquest,
the leading cultivating groups were also faced with the need to
assimilate new people to the peasant order and to preserve their
own place of ascendacy in that order. The means for accomplishing
these objectives could be found in the system of stratification
according to ascribed ritual purity in which the respectable and
powerful cultivators were acknowledged to be next only to Brah­
m ans in m oral standing. They were accorded the status o f satvik,
or men o f a respectable way o f life, and thus distinguished from
lower orders o f the population. W hen added to the economic
and m ilitary position o f Vellalas and other similar cultivating
peoples, Brahmanical caste relations provided the means for
internal regulation of established peasant localities as well as new
tracts open to peasant occupation. Changes in prevailing Brah­
manical religious forms to those more congenial to the devotional
religion of peasants made the alliance easier to achieve and sustain
— a factor which should perhaps be given greater weight in our
understanding of the bhakti movement.
For the Corom andel peasantry, the gains in the alliance were
extremely im portant. Devotional sects provided expanded op­
portunities for participation in H indu ritual for at least the highest
and middle status groups living in villages around brahmadeyas.
One cannot speak of this as a new activity, for Brahmans had lived
am ong peasants for centuries ministering rituals associated with life
cycle and village ceremonies. However, the shift o f religious ritual
increasingly to Brahmanical temples sheltering Vedic gods and the
elaboration of devotional forms must have been addressed as
much to the religious requirements and beliefs of this peasant
population am ong whom Brahmans lived as to the canons of
Brahm anical orthodoxy. Ancient folk deities were assimilated
— Seyon, or M urugan, of the hill regions (kurihji) became identified
with Subrahm anya, and May on. the black god of the pastoralists,
was readily transform ed into K rishna;41 Agam ic prescriptions
were altered or created to facilitate the inclusion by Brahmans of
many new folk elements and the partial exclusion of others such
as animal sacrifices and excessively erotic religious customs.
Devotional religion from the early Pallava period onwards,
insofar as it involved the sacral activities of Brahmans, did not
include the participation of all segments of the peasant population.
Specifically, it excluded those whose status was below the broad
and complex strata denom inated by Brahmans as ‘Sudra’. As
a social category, Sudra has been even less relevant in South India
than elsewhere in the Indian cultural sphere, but with respect to
ritual participation it continued to have currency. In the religious
activities of Pallava times, the first signs of what was to become the
enduring pattern of social stratification in South India became
evident. Three strata, comprised of Brahmans, respectable or
‘clean’ non-Brahmans, and low castes, seem to have been recog­
nized for ritual purposes, and each o f these was a complex ag­
gregate of groups with some mobility between strata. While each
strata, even that of Brahman, could not be considered closed,
corporate entities such as those idealized in the varna conception,
the m iddle strata was perhaps the m ost complex and least stable.
Occupying this level were m ost o f the peasantry with some claim
to land o r to an agrarian-skill which was valued, and ranking within
the category was so competitive and indeterminate that the dual
division into right and left-hand castes became an im portant
structural ancillary. It appears that the right-hand division of
castes — of whom we learn something during the Chola period
— were those whose relations to the peasant agrarian system were
most direct, including, as it did, cultivators and those artisans and
merchants directly involved in the agricultural economy. The
left-hand division appeared to provide an association of groups
less directly involved with agriculture. As suggested above, the
m ost im portant cultivating groups, such as the Vellalas, Reddis
and Kammas, sought to remain above the dual division, along with
Brahmans. To accomplish this would have required more than
landed wealth alone; it required a special relationship with Brah­
mans based upon ritual opportunities not shared by other non-
Brahmans. In later medieval times, especially among Srivaishnavas,
41 S rinivasa Iyengar, op. cit,, pp. 76-7, 355, 6!2.
such people often achieved considerable, if insecure, prominence
in temples, and there were some castes of non-Brahman Saivites
about whom little has been written, that enjoyed a special relation­
ship with Brahmans for substantial periods. Given the system of
ritually protected stratification which caste has been, and given
the profoundly greater evidence of Brahmanical influence during
and after the Pallava period, it appears probable that the most
powerful and established cultivating groups in any of the numerous
peasant localities o f Corom andel would have sought to use the
prestige of association with Brahmans as a means of stabilizing
their position with respect to other groups within the localized
peasant societies of the plain.

IV
This analysis of the Brahm an-peasant alliance of the Pallava
period and the presumed reasons for its existence has sought to
emphasize factors which are perhaps inadequately appreciated
by many Indian historians. These factors are mundane, and
because the alliance conception proposed here may be seen to
distort or to pervert what most historians of South India continued
to view as a purely religious development, the proposal requires
clarification.
There are probably few historians o f early South India who would
disagree that an enduring and symbiotic relationship between
dom inant cultivators of the most productive and prosperous
zones of agriculture and Brahmans had come into existence in
pre-Pallavan and Pallavan times. N or would there be disagreement
that this relationship deepened and elaborated through subsequent
historical periods. But these same scholars might cavil at the idea
that this relationship could be called an ‘alliance’ or that bhakti
Hinduism of the South Indian macro region, in any im portant
sense, reflected the quite worldly concerns of peasants and priests.
Religious change, especially such fundam ental change as the
displacement of sacrificial forms of the Brahmanical religion and
Jainism and Buddhism, tends to be seen conventionally as the
evolution of or a dialectical process involving an ancient religious
devotionalism of Tamils in relation to ancient theological or
philosophical conceptions long present in Brahmanical thought.42
42 See the fo rth co m in g essay by F riedhelm H ard y , ‘Ideology and C u ltu ral C on-
Such a perception of religious change as independent of social
contexts and processes is generally difficult to consider; in India
this notion of religious thought and institutions as com partm en­
talized, as somehow isolated from the nexus of social, political,
and economic relationships, is certainly doubtful, if not wholly
unacceptable. Therefore, it is quite essential to specify how and
why religious and non-religious factors interacted to shape both
religious and non-religious ideas and institutions in early South
India.
In emphasizing the ideological com ponent of Hinduism as a
factor in the alliance of Brahmans and peasants, it is not denied
that quite genuine piety m otivated the actions of many peasants
nor that many Brahm ans whole-heartedly subscribed to the popular
and devotional tenets of bhakti faith. However, in seeking to
understand the factors involved in the balanced and durable
alliance of Corom andel Brahm ans and peasants, piety appears
a weak analytical reed. For, when the Brahm an-peasant alliance
changed after the fourteenth century, are we to suppose that piety
ceased or changed? Where, as in Hinduism, we have a religion
which is not com partmentalized but comprehends a ‘total way
of life’, we must expect to find in religion symbols and dispositions
much which we do not regard as essentially religious. We should
expect to find injunctions about basic patterns o f livelihood,
power, social relations, and those systematically related, morally-
valenced, and publicly-expressed (in inscriptions) secular ideas
which constitute the ideology of any people. And when, as in
Pallava times and later, the premier religious functionaries are
found to be linked to powerful peasant groups, the ideological
com ponent of religion should assume a high order of priority
in our understanding of agrarian relations. Religion was made
to serve the Brahm an-peasant alliance which constituted the under­
pinning of localized, self-governing territories in South India
from the eighth to the fourteenth centuries. In the succeeding period
o f agrarian organization, when M uslim pressure upon the northern
edges o f the Corom andel plain created the conditions for the
expansion of the highly m artial Telugu warriors of the Vijayanagara
period, religion again served to provide an essential cohesion to
texts o f the S rivaisnava T em ple' in a sym posium on S outh In d ian tem ples (edited
by B u rto n Stein) by The Indian Econom ic and Social H istory R eview , v. 14 (Jan .-M arch
1977), pp. 119-51.
different fundamental relationships in South Indian society.43
Prevailing interpretations o f the Pallava period as one o f signi­
ficant, disjunctive change have tended to ignore the steady ex­
pansion o f peasant society and culture from the Classical period.
This expansion over the entire Coromandel plain and the ultimate
emergence of powerful peasant peoples fundamentally altered
relationships among peoples of the society of South India as des­
cribed in the early poems o f the Tamils. In particular, it set the
peoples o f the river plain and their subordinate allies, the pastora-
lists and fishing folk, against the non-peasant peoples of the forests
and dry uplands, m atching the superior populations and wealth
of the former against the military capabilities o f the latter. It is
suggested that the ‘K alabhra interregnum ’ may be recognized
as a period of m om entary success on the part o f the non-peasant
peoples seeking to achieve control over the plains of the Southern
peninsula and that the establishment of Pallava rule may be taken
as the ultim ate trium ph o f the wealthy plainsmen. It is further
proposed that the social and cultural changes associated with
Pallava rule in Tondaim andalam and eventually the entire Tamil,
K arnatak, and Telugu plains are to be understood as a part of the
effort of the peasant people o f this macro region to achieve and
consolidate that victory. To this end, the Corom andel peasantry
supported Pallava kings as cakravartins who were rulers over a
territory not over a people as in older forms of kingship and who
were committed to the prosperity and protection of the peasantry.
Similarly, the South Indian peasantry supported the development
of devotional Hinduism not simply from reasons o f piety or respect
for its Brahm an functionaries but for the contribution religion
could make to the ideological cohesiveness of the plains people.
Great settlements of Brahmans became centres of the revitalized
religion just as they served as centres of the peasant culture o f the
plains in other ways, and these centres, for all o f their ostensible
dom ination by learned and pious Brahmans displayed many
signs of the association with the peasantry, including the devotional
folk elements of the religion and the continued strength o f peasant
43 Ib id . A n o th e r essay in the tem ple sym posium is A rju n A p p a d u ra i's ‘K ings,
Sects a n d T em ples in S o u th In d ia : 1350-;700 a .d .’ ; th is essay deals w ith T elugu
w a rrio rs’ successful in te g ra tio n in to tem ples o f T am il co u n try .
corporate group life. Situated in the most developed peasant
tracts o f the Coromandel plain, these great Brahman settlem ents
could not but show themselves to be bastions o f plains culture
reflecting the close collaboration between Brahmans and peasants.
 Peasant Micro Regions :
The Nadu

The m acro region defined above was comprised o f numerous

distinctive locaiities which increased in the period from the ninth
to the twelfth centuries with the expansion o f peasant society.
Varying according to ecoiogicai conditions, proximity to established
older cores o f peasant settlement, and their social composition,
these micro regions were the enduring and basic units of South
Indian peasant society. M ost predated the Cholas and most,
while retaining their identity and organization, were capable, at
times, of being aggregated as larger territorial entities under
various kinds of overlords. Occasionally, this massing o f locali­
ties was carried out by members o f the Chola royal family, especially
in the Kaveri delta, but more often by chiefs, themselves locality
notables and closely linked to the dom inant peasantry o f the
place. From the Pallava period until quite recent times, these
micro regions constituted the basic ‘arena’— to use F.G. Bailey’s
term — of the functioning peasant society.'- That is, the peasant
locality was an interactional region defined by relatively dense
interrelations am ong social groups with common interests in some
tract of cultivable land.
Nadu is the term which designated the micro region. It is also
used in the Chola period to designate an assembly of the micro
region. Etymologically, nadu refers to agricultural land in con-

1 Q u o ted from F .G . Bailey, ‘P olitics a n d Social C hange in O rissa' in Politics

and S o ciety in India, ed. C .H . Phillips, L o n d o n , 1962, a n d co m m en ted upon by
B aru n D e, ‘A H isto rical P erspective on T heories o f R é g ionalisation in In d ia’, in
Regions and Regionalism in South A sian Studies: A n E xploratory Study, ed. R o b e rt
I. C ran e, P ro g ra m in C o m p a ra tiv e S tudies on S outhern A sia, D u k e U niversity,
M o n o g ra p h no. 5, D u rh a m , N .C ., 1966, pp. 54-5. S ee also B ailey's discussion in
his, Tribe, Caste, and N ation : A S tu d y o f Political A ctivity and P olitical Change in
H ighland Orissa, O .U .P .. B om bay, 1960, pp. 269-70.
trast to kudu, a Dravidian word for forest or other land not suited
to cultivation .2
In Classical poems, nadu appears to have been consistently
contrasted to hill and forest country as in the Pumnanuni where
a C hera king is said to rule over the hills, the forests, and the nadu,
and where the C hola K arikala is said to have ‘'made nadu (nadu
akkl) by ‘clearing forests’ (kadu konru />.3 M ost nadus are identified
by the name of some village from as early as the Classical period.
In Tondaim andaiam another term., kurram , is used along with
nadu. Kurram , in some inscriptional contexts appears to be
equivalent to nadu though this is the subject of some controversy
am ong scholars .4 The term kbit am also occurs in T ondaim an­
daiam and may not have been a simple equivalent o f nadu ; kottam
appears to designate physical sub-regions o f the central Tamil
plain m arked by the topographically prom inent distribution of
hills to which the root of the word, kot-, may refer .5
Nowithstanding differences among scholars regarding possibly
m eaningful distinctions among these various terms, there is general
agreement that the territorial term nadu has primacy as the funda­
mental building block of rural organization during the Chola
period.
Long ignored by historians o f the C hola period, the nadu has
recently been the subject o f a serious and able study by Y. Subba-
rayalu, a substantial portion of whose m onograph on Chola
country from a .d . 800 to 1300 is devoted to this territorial unit.
‘The N ad u ’, Subbarayalu writes, ‘is the very key to political
geography o f the C hola country. It was a very im portant limb
o f the adm inistrative system o f the period . . . and it was the basic
unit o f the then agrarian society.’6 Treating the territorial unit
kurram as synonymous with nadu, Subbarayalu has identified
and m apped 140 in Cholam andalam and sixty-five in the territory
immediately north o f Chola country, Naduvil-nadu, as of about
2 D .E .D ., no. 3012, p. 242, 'natu .
3 Y. S u b b aray alu , P olitical G eography o f the Chola C ountry. T am il N a d u S tate
D e p a rtm e n t o f A rch aeo lo g y , M ad ras, 1973, p. 32.
4 B. S uresh, ‘H isto rical an d C u ltu ral G eo g rap h y an d E th n o g ra p h y o f S outh
In d ia w ith Special R eference to C ola E p ig ra p h s’, u n p u b lish ed Ph. D . T hesis, D eccan
C ollege, U niv ersity o f P o o n a , 1965, p. 30 (h ereafter: H C G E S I). A lso N ila k a n ta
Sastri, The Colas, p. 465; S u b b aray alu , op. cit., pp. 19-20.
5 S uresh, H C G E S I, p p . 25-6.
6 S u b b a ra y a lu , o p . cit., p. 19.
a .d . 1300.7 He has aiso identified more than 1,300 villages within

these two sub-regions of the Tamil plain of which 1070 were in

the 140 nadus of C holam andalam and the remainder in the
adjoining sixty-five nadus of N aduvil-nadu.8 In addition to the
perennial region o f the Kaveri basin, Subbarayalu has also delinea­
ted 203 nadus in the central Tamil plain of Tondaim andalam .
Here, between the northern and southern Penner rivers, were
located 638 villages during the Chola period. Other territories
similarly treated are: Kongu country in the interior upland of
m odem Salem and Coimbatore of Tamil N adu with thirty-two
nadus and 106 villages positively located; Pandim andalam (modem
M adurai) had ninety-five and 490 respectively; Gangavadi (ganga-
pddi or mudikondaehola-mandalam as it was also called) in southern
K arnataka provides evidence of eleven nadus and twenty-five
villages and in several other smaller tracts, Subbarayalu has
located twenty nadus and fifty-eight villages. Thus, by about
a . d . 1300, it is possible to locate with considerable certainty 556
localities in a substantial portion of the peninsular macro region
based upon the certain location o f 2,620 villages mentioned in
contem porary inscriptions.
Maps prepared by Subbarayalu also indicate two related charac­
teristics of the nadu which emphasizes its essential agrarian func­
tions. The first is the relationship of size and proximity to reliable
water sources. Nadus of the Chola country vary in area from ten
to three hundred square miles. Considering only those nadus
whose boundaries can be fixed with a high degree of certainty,
sixty-eight in Chola country and fifteen in Naduvil-nadu, the
following variation is discovered.
Larger nadus were located on the infertile, poorly-watered
margins of the riverine plain according to the maps prepared by
Subbarayalu. In such relatively inhospitable tracts there were
fewer villages and people than in the deltaic portions o f the plain
with its greater availability o f moisture. Their boundaries reveal
a second predictable characteristic related to the agrarian functions
7 H e h as also included the in freq u en t term s kandam a n d vattam (used in five cases)
as eq u al to nadu (op. cit., p. 46-7).
8 N a d u v ii-n a d u is referred to by o th e r n am es: N a d u -n a d u , N a d u v u n ila i-n a d u ,
an d N a d u v u -n a d u ; the tract is situ ated betw een T o n d a im a n d a la m an d C h o la­
m an d alam , betw een th e so u th e rn P enner a n d the n o rth e rn V ellaru rivers, ibid.,
p. 31; also see, K .S. V a id y an ath an , ‘A ncient G eo g rap h ical D ivisions o f T am il
 * T a b le III-I
Areal Variation of Selected N a d u s: a .d . 1300
A rea C h o lam an d a la m N ad u v il-n ad u
Sq. m i. S-- 68 S -- 15

8-12 5 8
15-20 24 4
25-30 23 1
35-40 10 0
50 I 1
65 I 0
70-80 2 1
180 1 0
300 1 0

o f these locality tracts. Subbarayalu’s maps show that most nadus

lie athwart water-courses, the principal discriminating physical
feature of much of the plain. Only the m ajor perennial rivers,
Kaveri and Penner, constituted boundaries. The flow of these
latter rivers was large and reliable enough for agriculturists on
both banks to use without the necessity of joint management.
N ot even the Kollidam (modern Coleroon) river, the most im por­
tant branch of the Kaveri, constituted a reliable demarcation of
nadus; several o f the localities on this river are found on both
sides of it (e.g. kaldra-kurram , vilanadu, vennaiyur-nadu )9 and
even the southern Vellaru, which was an ancient dividing m arker
between C hola and Pandya countries, appeared in the Chola
period to have some nadus lying on both of banks (e.g. olliyiir-
kiirram, kudalur-nadu, and peruvayil-nadu ) . 10
The spatial and tem poral dem arcation of nadus depends upon the
location o f villages said in the existing inscriptions to belong to a par­
ticular one. There are num erous detailed descriptions o f village
boundaries, but not a single inscription which defines the bounda­
ries o f a nadu. C ore settlements can be located with reasonable cer­
tainty from their names as well as from locational factors such as
streams, canals, tanks, and hills which are mentioned in inscriptions;
such places may often be found on m odem , large-scale (one inch to
N a d u ’, H andbook (K a iyed u ), I I International C onference-Sem inar o f Tam il Studies,
M a d ras, 1968, p. 234.
9 S u b b aray alu , op. cit., 22.
10 Loc. cit.
one mile) maps.. However, in other tracts, features such as forests,
jungle, scrub, long converted to cultivation {and in some cases
successively reclaimed as we know from eighteenth century records),
can rarely be found. In general, the further removed from the
principal water sources of a locality, the greater the difficulty in
delineating the boundaries of a particular locality. Many of the
villages referred to in inscriptions cannot be located at all. This
is hardly surprising; rather, it is striking how many settlements
of the medieval period still exist with names but little altered from
ancient times.
Another, quite minor, problem in m apping the nadus of the
Chola period is that a small num ber o f villages can be shown
probably to have been transferred from one to another.
Subbarayalu estimates that he has encountered about fourteen
cases of such transfers in Cholam andalam , a num ber he considers
insignificant considering that there were some 200 in Chola country
and the period over which he was surveying the evidence was
five hundred years.11 Moreover, the inscriptional record does
not deal explicitly with such apparent transfers, hence it is im­
possible to be even somewhat clear about why such transfers took
place. Subbarayalu does point out that in all cases, the transferred
settlements were at the borders of their respective nadus, that is,
at a relatively great distance from their core settlements.
Finally, of course, there is the problem which exists for any
research based on inscriptions, that is, the incompleteness of the
record. N ot all villages can be presumed to haVe received inscrip­
tional mention. The stone and copperplate records which contain
references to villages are not all preserved and legible; and not
all of these records have been collected or even noticed in epi-
graphical surveys. N otw ithstanding all of these problems, the
survey and mapping which has been done reveals matters of the
greatest im portance for an understanding of early agrarian rela­
tionships in South India.
The distribution of peasant villages in the some 500 nadus of
the territory under Chola overlordship was most decisively shaped
by the availability of reliable irrigation sources. The overall density
of villages in parts of the Kaveri delta reflects this relationship.
Subbarayalu has estimated that there was one village per two
square miles in places of reliable water supply. In Cholam andalam ,
ii Ib id ., pp. 22 (T.
most (i.e. two-thirds) o f the nadus with reasonably ascertainable
boundaries fell within the range o f fifteen to thirty square miles
and in N aduvil-nadu most (i.e., four-fifths) fell within the range of
eight to twenty square miles.12 These quite restricted localities
exist in the best agricultural tracts where definiteness of boundaries
is related to high population and thus to the existence of many
villages capable o f being assigned to one or another nadu. O f
the several in Cholam andalam with the largest num ber of villages,
the average size of the nadu , as suggested by Subbarayalu, was
twenty-four square miles.13 Ten nadus o f the 140 in C holam anda­
lam can definitely be assigned twenty villages or more, and all
but one of these can be regarded as having been very well sited
in relation to water.

III-2 T able

Selected Largea and Populous*3 C holam andalam Nadus

by Area, Villages and Relation to water supply
N am e M o d e rn L ocation Villages A re a c R elation to M ajor
(D istrict/T alu k ) (sq. m iles) W ater Source
R em ote P roxim ate

kilär- T an jav u r, P ap an asam 20 35 X

kürram
ärkä ttu - T an jav u r, T an jo re 24 36 X
kürram
pächchd- T iru ch irap alii, Lalgudi 20 45 X
kürram
pattana- T an jav u r, N a g a p a ttin a m 30 56 X
kürram
tiruvalandür T an jav u r, M a y u ram 22 56 X
nädu
lirunaraiyür- T an jav u r, K u m b a k o n a m 28 60 X
ncldu
purangaram - T an jav u r, M an n arg u d i a n d 27 75 X
bai-nädu P a ttu k k o tta i
uraiyur- T iru ch irap alii, K ulittalai 20 136 X
kürram
urrättür- T iru ch irap alii, K u littalai 24 180 X
künam an d K u la ttu r
vaUuvap- T iru ch irap alii, M usiri 40 200 X
padt-nadu

a. In excess o f 35 sq. m iles b. In excess o f 20 villages c. A p p ro x im ate

12 Ib id ., p. 22. 13 Ib id ., pp. 21-2.

 If the nadu territory and assembly were administrative institu­
tions of the Cholas, as the conventional historical view and Sub-
barayalu contend, then it is im portant to recognize that this is
an inference. No contem porary documents speak of the nadu
in terms of the Chola governmental structure or function. On
the contrary, these local units are referred to in donative inscrip­
tions for the apparent purpose of better identifying a village or
in recognition o f the special prominence of the leading people
o f the nadu — nàttàr — in m ost donative transactions. Therefore,
it would not be appropriate to suppose that when a locality —
nadu or kurram — is first mentioned in an inscription that it is
necessarily a newly created tract of settlement. There are, in fact,

III-3"
T a b le
New Nadus b(First References) in Cholam andalam and
Naduvil-AWw: a .d . 850-1300

P erio d C h o la m a n d a la m N ad u v il-n ad u

- 800 18 1
800- 850 0 3
850- 900 27 11
900- 950 27 6
950-1000 13 11
1000-10 50c ' 42 8
1050-1100 2 4
1100-1150 0 10
1150-1200 11 2
1200-1250 0 4
1250-1300 12 5

T o ta l 152 65

a. Based u p o n th e table in S u b b aray alu , op. cit., p. 20, w ith m odifications for
p resen tatio n .
b. Includes nadus an d kurram s.
c. T his is b ro k en dow n accordingly to differentiate the reigns o f R a ja ra ja I, (d.
a . d . 1014) a n d his successors:

Period C h o lam an d a la m N ad u v il-n ad u

1000-1014 36 - 4
1015-1050 6 4
few and questionable references14 to the creation of a nadu or
kurram, and the assumption that its first mention may be taken as
the approxim ate date of its creation is a doubtful one. Subbarayalu
makes this assumption in his discussion of the ‘evolutionary
character’ o f nadus and kurrams. He presents as evidence of
‘new units’, the first references to nadus and kurrams for C hola­
mandalam and Naduvil-nadu for the entire Chola period. On
the basis of these ‘first references’, the following distribution occurs.
Subbarayalu is led by this to suppose th at:
The density of population in the Choja-M andalam had reached, for
. the optimum conditions prevailing then, its saturation point . . . in the
first half of the 11 th century, after which the rate of growth was obviously
slow. In the Naduvil-nadu, however, new units were coming into existence
at a faster rate till the middle of the 12th century, after which the rate
was a bit slow.15
While it seems obvious that new agrarian tracts would soon open
around the rich ecological cores such as the Kaveri and the southern
Penner basins, the tem po of expansion suggested by Subbarayalu
cannot be accepted without corroborative evidence of very impres­
sive and sustained population growth or immigration. The large
num ber of new agrarian tracts of irrigated agriculture, utilizing
intensive labour, could not have been established without new
population or restructuring of the field labour force or a basically
new technology. There is no evidence o f these.
It is more likely that new agrarian tracts were opened on the less
fertile fringes o f both Cholam andalam and Naduvil-nadu. Some
of these were among the largest locality units of which we have
records. These include: vailuvappadi-nadu (Tiruchirapalli, Musiri
taluk) with forty villages and areas of about 200 square miles;
urrattur-kurram (Tiruchirapalli), with twenty-four villages and
an area of about 180 square miles. Others like kurunagan-nadu,
vembar-nadu, and kanakkiliyur-nadu were located on the western
extremity of Cholam andalam ; or, like idaikka-nadu and vaveliir-
nadu, they were located in the relatively dry zones between such
m ajor streams as the V arahanadi (South Arcot, Tirukkoyilur
taluk) or, like tunda-nadu and vilandaiyil-nadu between the western
portions of the northern Vellaru and the M urudaiyaru (Tiruchi-
14 A .R .E ., 1943-4, n o . 277, w here the ep ig rap h ist supposes a ‘re g ro u p in g ’ o f
these nadus. S u b b a ra y a lu ’s read in g o f this reco rd is different, o p. cit., p. 44.
15 S u b b aray alu , op. cit., p. 21.
rapalli, Udaiyarpalayam taluk), or finally, danava-nadu, around
the upper reaches of the Ambuliyar river (Tanjavur, Pattukkottai
taluk). M ost o f these tracts were isolated, large, and not very
populous; these areas of later occupation may, as Subbarayalu
has suggested, be the result of saturation densities in propinquit-
ous and better endowed neighbouring territories.
R ather than viewed as an ‘evolution’ or as an unfolding of
locality units as a result of ‘saturation’, it is proposed here that
m ost of the 500 or so localities referred to in Chola times were in
existence during the time o f R ajaraja I, and the references to them
indicate not a new existence, but a new recognition of Chola over­
lordship.
The assum ption th at locality units, variously called nadus,
kurrams, or, occasionally, by some other term , were in existence
at the time of the ‘Imperial Cholas’ is one of the three possible
assumptions which can be made. The prevailing conventional view
of the nadu is based upon the supposition that the centralized
state utilized, if it did not create, it as the prime unit o f local ad­
m inistration. In this view, the term ‘nadu’ is understood primarily
in its meaning as a territorial assembly, not as a territory; it is
the assembly which attended its governance as a territorial unit
of administrative convenience. The nadu, as an assembly, is seen
by N ilakanta Sastri as composed of some kind of regular repre­
sentation from each village or locality,16 or seen by M ahalingam
as composed of the more ‘influential’ persons of a locality.17
Both N ilakanta Sastri and M ahalingam tend to view the nadu
assembly as a link in the overall control of the Chola government.
This view rests on the unsubstantiated claim of the centralized,
bureaucratized character of the Chola state; this view also neglects
the agrarian and social basis of the locality territory as long an­
tecedent to the political order established by the Cholas.
Subbarayalu, in a second possible position, recognizes the
agrarian and social character of the nadu. However, in arguing
that those of Cholam andalam and the N aduvil-nadu in most
cases came into existence in the course of the Chola period, Sub­
barayalu is burdened by the undem onstrable assum ption of
impressive population increase such as to explain the esta­
blishment of ‘new units’ of agrarian organization leading to the
16 The Colas, p. 503-4.
17 South Indian P olity, p. 369.
eventual ‘saturation’ of the Kaveri and the Penner basins. He
also presumes a causal relationship between these developments
and the Chola state. It is not possible to dem onstrate that there
was a ‘Chola peace’ which ended a period of disorder inimical to
population growth or agrarian development nor that significant
new forms o f productive organization or techniques could have
accounted for the development implied in Subbarayalu’s for­
mulation.
The third position, the one adopted here, is that m ost of the
several hundred locality units of the Kaveri and Penner basins
(Cholam andalam, Naduvil-nadu and Tondaim andalam ) were in
existence prior to Rajaraja I, and that these units as the prime
units of social and agrarian organization provided the basis for
the Chola political order. The nature of the Chola state is dis­
cussed separately in chapter VII but as that state was based upon
an extant agrarian order, of which the nadu was the keystone,
attention must be given to its essential character.
M uch of our understanding of the persistence of territorially
segmented units of culture and society in South India, of which
the nadu was the m ajor m anifestation during the Chola period,
must await deeper scholarship on the society depicted in Classical
poetry and on the evidence of archaeology which in South India,
as elsewhere on the Indian sub-continent, has barely begun.18
These are the most promising frontiers of South Indian scholar­
ship. Though tentative, an assessment o f the features of South
Indian territorial segmentation may be suggested.
The legacy of the ancient past o f South India upon the segmentary
territorial structure of medieval South Indian society and culture
is manifold. Evidence of very early, widespread farming com­
munities over most of the southern peninsula comes from neolithic
and post-neolithic archaeology. While yet incomplete, this evidence
strongly suggests that agricultural settlements dotted the various
landscapes of the region, the more favourable riverine sites as well
as less favoured upland ones. M oreover, it would appear that
these early sites continued to be the loci of peasant settlement
18 T w o recen t w o rk s deserve special m e n tio n h e re : S. S ingaravelu, Social L ife
o f the Tam ils, D e p a rtm e n t o f In d ian S tudies, U niversity o f M a lay a, K u a la L u m p u r,
1966, especially ch. 9, ‘T h e N a tu ’ ; a n d C larence M aloney, ‘A rch aeo lo g y in S outh
In d ia: A cco m p lish m en ts an d P rospects', in E ssays on South India, ed. B u rto n
Stein, U n iv ersity P ress o f H aw aii, H o n o lu lu , 1975, ch. 1.
during the historical period. Further, at the time when the first
states were being consolidated in the Gangetic plain, in the middle
of the first millenium B .C ., there were relatively stable though
somewhat primitively formed, societies in the southern peninsula
known to Gangetic peoples. Reference to the Cheras, Cholas,
and Pandyans as well as other peninsular peoples in the Asokan
inscriptions m ark this recognition around the fourth century
B .C . Finally, in the Classical literature o f the first several centuries
of the present era, we are presented with a complex and sophisticated
social order in which diverse peoples lived in trade settlements
and engaged in an extensive trade network, including the M editer­
ranean region. At the same time, these cosm opolitan people shared
a general culture with a variety of South India folk in quite different
settings. As reflected in the cultural-ecological categories of the
five regions, or poetic situations (tinais) o f Tamil speakers, one
sees many well-established variants of a single, general culture.
One other factor should be regarded as im portant with respect
to territorial segmentation in the southern peninsula. The history
of South India is quite clear on the m atter of conquests by peoples
of cultures different from those of the macro region. After the
ancient incursions into the southern peninsula by M editerranean
peoples who comprise the basic Dravidian stock, there appear
to have been no conquests of the southern peoples within the macro
region of this study. There were neither the politico-military
subjugations which might have obliterated ancient, local ethnic
territories as known in Classical poems, for example, nor was
there a ‘cultural conquest’ such as might achieve the same end.
The introduction and dissemination of Aryan cultural forms —
Sanskrit language and the Indo- Aryan culture of the Gangetic
plain — appear to have been primarily the work o f South Indian
men. In a process which has been suggested by Kosambi and
assumed, if not explicitly supported, in the work of most South
Indian historians,19 Indo-A ryan culture was assimilated by South
Indians gradually and selectively. There was no sudden trans­
form ation o f Dravidian culture — if one may use such a term —
by strangers, but rather slow, quite varied, and m odest variations
19 D .D . K o sam b i, ‘T h e Basis o f A ncient Indian H istory. I ’, Journal o f the Am erican
O riental S ociety, 75, n o. 1 (Ja n .-M a rc h , 1955), p. 43. A v aluable, recent co n sid era­
tio n o f this m a tte r com es from N ila k a n ta S astri, Cultural C ontacts Between A ryans
and D ravidians, M a n a k ta la , B om bay, 1967.
in an already rich Dravidian culture through interactions with
carriers of Indo-A ryan culture, many of whom may themselves
be called ‘D ravidian’ in language, culture, and social allegiance.
The result o f these factors was a South Indian cultural system
based upon highly localized, very durable varieties of a single,
extensive culture, a related set of languages, and substantially
shared cultural traditions. The robustness of the Dravidian com­
ponent is evident in modern times as it has been throughout history.
The impact of early indo-A ryan forms, Muslim influences during
the medieval period, and, finally, European influence since the
eighteenth century have not obliterated the Dravidian foundation
of South Indian life. And it may well be that the segmentary
character of that society and culture served to preserve and protect
that foundation.
The salient dimensions of segmentary organization in early
South India were restricted m arriage and kinship netw orks; narrow
territorial social coalitions beyond kinsmen; and locally-based
agrarian relationships, political and religious affiliations and
loyalties.
The distinctive cross-cousin and m aternal uncle-niece forms
of m arriage apparently have been preferred from the beginning
to the period under examination here. Unfortunately, this pro­
position is difficult to assert with complete confidence because
inform ation relating to m arriage arrangements is surprisingly
sparse in the early, didactic literature o f Tamils, and they are non­
existent in the inscriptions except as marriages facilitated political
alliances.20 However, according to the eleventh century com­
m entary of Vijnanesvara on the Yajnavalkya Smriti, cross-cousin
marriages were recognized as appropriate in South India even
though disapproved in the general body of medieval dharmasastra,21
20 N o te d by K .A . N ila k a n ta S astri, Culture and H istory o f the Tam ils, p. 83, w ho
has been a m o n g the few h isto rian s to have consistently den ig rated the literary
sources o f th is p erio d as to o fanciful a n d d id actic a n d difficult to d a te to be reliable.
See his, Sources o f Indian H istory, pp. 55-6.
21 D .K . K a ra n d ik a r, E x tra c ts fr o m the Yajnavalkya S m riti, B o m b ay V aibhav
Press, B om bay, 1913, p. ii. M arriage w ith o n e ’s m a te rn a l u n cle’s d a u g h te r is also
allow ed by the n o rth e rn c o m m e n ta to r on M a n u , G o v in d araja, o f a b o u t the sam e
tim e (K an e, op. cit., I, pp. 313-15). E arlier references to the la tte r m ay be found
in B a u d h ay an a (possibly a so u th e rn er o f c: 500-200 B .C . a cco rd in g to K ane, ibid.,
pp. 27-30); G . Biihler, The Sacred B ooks o f the A ryas as Taught in the Schools o f
A pastam ha, G autam a, Vaishtha. and Baudhayana, v. 14, pt II, Sacred B ooks o f the
East, M o tilal B an arsid ass, D elhi, 1965, p. 146.
 and, in the Jain Tamil classic, the Civakacintamani, of a possibly
earlier date, one of the marriages o f the hero, Civakan, was with
the daughter of his m aternal uncle, in which the latter recognized
the hero’s preferential claim to his daughter.22
The spatially compressing character of the m arriage system
existed among the dom inant, land-controlling peasantry as well
as among m ost other locality social groups. In part this reflected
a sharing o f m arriage norm s from which no group, including
Brahm ans, was exempt; but, m ore, it was a concom itant of
coalitions which exist in any society dominated by peasant agrarian
relationships. In the absence of effective, continuous, extra-local,
non-peasant authority over agrarian relationships and over the
distributions o f the products of the land, it was only through secure
relationships with those local groups which controlled the land
that benefits could be obtained by groups without such control.
The superior privileges and com m on interests of the dom inant
peasant groups — which m ay at the beginning have consisted of
diverse ethnic segments — would have produced a cohesive,
endogamous locality group.23 Similarly, among those dependent
upon the dom inant locality peasantry, endogamous links were
formed based upon com m on or related occupations, or they would
form coalitions with those with whom m arriage relations were
not feasible. Where, as in most Coromandel localities, direct
management of the land and effective political authority were
combined in the same group, other locality groups developed
closed and separate corporate identities. ,
Territorially segmented marriage and kinship systems were
supported by caste principles of hierarchy and purity as these
principles may be identified in medieval South India. References
to castes as they are known to m odem Indian sociology are rare.
Suresh, in his study of Chola inscriptions, noted the names of over
1,800 persons, not including officials and women of chieftains’
families, mentioned in temple inscriptions. These persons included
women connected with the temple and other women donors,
male donors, temple staff, and Brahmans. Of this large number,
22 T .E . G n a n a m u rth y , A C ritical S tu d y o f the C ivakacintam ani, K alai K a th ir,
C o im b ato re, 1966; p. 209.
23 L ocalized c o n n u b ial system s can a n d h ave been m ain tain ed in w ays o th er
th a n th o se fo u n d in S o u th In d ia, o f co u rse; the w idespread p referen ce rules found
th ere are th erefo re n o t ‘ex p lain ed ’ here, m erely noted.
only Brahmans and about twenty other individuals are identified
by ‘caste’ and, of these, eighteen are either Vellala or setti, m ean­
ing ‘agriculturists’ and ‘m erchants’ of respectable standing. Suresh
concluded that the near total absence o f caste designations for
those other than Brahmans named in these temple records indicates
that ‘the caste system had not yet set’.24 This is a quite unnecessary
conclusion.
F or most Brahmans mentioned in the inscriptions studied by
Suresh and others, gotra designations are provided. However in
all cases, the village in the locality of the temple from whose inscrip­
tions the names of persons come — whether Brahmans or others —
is provided. It would therefore appear that for purposes of the
public recording which temple inscriptions served, the two essential
elements o f inform ation necessary for the identification o f per­
sons were whether they were Brahm ans and where they lived.
It is reasonable to assume that the persons mentioned who were
not Brahmans belonged to respectable social groups, for, pre­
sumably only those persons could be suitable participants in the
canonical temples. Thus, the list of names provided by Suresh
would include the two m ost prestigious social groups in South
Indian society at the time: Brahmans and respectable non-
Brahmans along with the inform ation of where they lived. To
say that the caste system was not set, as Suresh does, may not
therefore be altogether true, for what is suggested from his data
and th at of others is that the tripartite division of territorially
segmented social groups in South Indian society — the dom inant
structural element in modern South Indian caste — was already
well-established and recognized. In temple inscriptions of the
age it was sufficient to give the proper name of non-Brahm ans
along with his village to say all that was necessary about caste.
While the division o f social groups according to the varna schema
of vedic authority has had little sociological reality anywhere in
the Indian cultural sphere, its categories (the four varnas plus un­
touchables) appear to have provided clearer status categories
in northern India, particularly the Gangetic plain, than in South
India. In the tripartite division of peoples in South India, the most
ambiguous division was the middle one. Certainty about the
appropriateness of a prospective marriage would have been
greatest to attain within the middle caste ranks of dom inant
24 S uresh, H C G E S I, p. 315.
peasants where concerns about relative status were perhaps most
keenly felt. M arriage among near kin sought to assure this ap­
propriateness. Beyond that, participation with Brahm ans in the
support and operation of the canonical temple, as representatives
of the locally dom inant people, further supported their status
pretentions.
Peasant controllers of localities sought in their marriage system
and in their conspicuous association with Brahmans and canonical
religion to fix for themselves a status second only to the indis­
putably high rank of Brahmans. Other social groups, identified
primarily by their occupations (and the relation of these occupa­
tions to agrarian production) held ranked places below that of
the dom inant peasant castes, thus fortifying the latter’s political
and economic position. The periodically activated coalitions of
the right- and left-hand castes am ong occupational/social groups
below the dom inant land controllers which came into prominence
by the late eleventh century, served as a means for some lower
peasant groups and more mobile artisan-traders to combine across
locality lines, against the dominance enjoyed by the powerful
land-controlling peasantry, their dependents, and their allies,
the Brahmans.
Requirements of the agrarian organization, given the technology
of the age, made for territorially segmented units of production.
The nadu was thus an economic as well as ethnic territory. Bound­
aries of each territorial unit were defined by interactions between
the dom inant landed folk and those dependent upon them —
artisans, m erchants, and labourers — as well as with Brahmans
with whom the peasant groups were allied.
The earliest peasant settlements of the nadu usually gave their
names to the locality. The other major source of names, besides
the village toponym , stressed natural elements directly related
to agriculture. Thus, manni-nddu in K um bakonam taluk, modern
Thanjavur, is nam ed after the M anniyar River, and idaiyarru-
nadu, also Cholam andalam , is the same as the term ‘doab ’ in N orth
Indian usage, the land between two rivers; pulalerikki-nadu and
ambattur-erT-nadu, in Tondaim andalam , are nam ed after the
prominence of a great lake or tank as melmalai-velur-nddu refers
to a prom inent hill. Those well-watered, fertile tracts which at­
tracted agriculturists to particular localities in the first instance
remained the inner cores of each nadu. Here were the chief settle­
ments of those whose efforts had won the place from previous
occupants — forest people or pastoralists, who settled and cul­
tivated the land, and who recruited groups of service providers
— including Brahmans. In time, subordinate settlements were
created in other parts of the tract by these peasant settlers and
those associated with them, extending the arrangements evolved
in the earliest settlements with which newer ones were bound by
marriage ties, cult practices, defensive arrangements, allegiances
to supra-local chieftains, and common interests in the m anage­
ment of land relations.
An interesting aspect of the relationship of ‘founding’ villages
and their localities is that the earliest settlements after which locali­
ties were often nam ed assumed no apparent primacy am ong locality
villages, either as administrative centres or as fortified places.
It was rather the locality as a whole and the dom inant peasant
groups within them — referred to collectively and acting cor­
porately as the nattar — and dispersed throughout the locality
that was considered as im portant. In public terms, as reflected
in the inscriptional records, large Brahman settlements and
im portant trade centres enjoyed more prominence than even the
largest, essentially peasant, settlement of a locality.
The network of rural relationships which emerged in a nadu
territory was conditioned by many factors. Land was one such
factor. The proliferation and viability of colonies of the parent
peasant settlement depended critically upon the availability of
arable land. But nature could be moulded to the needs of peasant
cultivation; irrigation works could be developed to reduce the
hazards of aridity and wells could tap rising water tables resulting
from the careful husbanding o f water in the vicinity. The relative
availability of men and animals to work the land also operated
as a determ inant of expansion of peasant settlements and both
of these factors would be influenced by such short-term phenomena
as war and disease, as well as long term ones like endemic health
conditions hazardous to both man and beast. Able leadership
and warriors among a peasant folk could facilitate rapid expansion
just as proximity to well-organized tribal folk in the forests — pre­
dators upon wealthy peasant settlements — could retard, even
reverse, the development of a peasant micro region.
It is recognized that the designation of the nadu , an ethnic and
ecological micro region, as the basic unit of Chola society
diverges from the emphasis which is usually given to the ‘village’
by most scholars. In the judgem ent o f m ost historians, the nadu
is vaguely seen as an adm inistrative unit with a certain degree of
responsibility over self-governing villages. Thus N ilakanta Sastri
states: ‘The self governing village was the unit o f government.
A num ber o f them constituted a kürram or nadu or kdttam as it
is called in different parts of the country.’25
The word, ‘constituted’, which may or may not have been used
deliberately to denote the nadu as an arbitrarily defined space of
administrative convenience, well expresses the propensity of
scholars to disregard the sociological significance o f the micro
region.26
Emphasis upon the primacy of the ‘self-governing village’ and
the ‘constituted’ character of the nadu is questionable on at least
two grounds. First, the kind of ‘self governing villages’ to which
reference is made, and for which alone there is evidence, is the
Brahman village, and such villages were, in alm ost all cases, al­
ready thriving peasant settlements before being conferred as
brahmadeyas. indeed, if they were not prosperous villages, then
they could hardly have supported the non-productive Brahman
population and could hardly have been considered gifts.27 To the
extent that it is possible to speak o f village self-government, it is
widely recognized that the non-Brahm an assembly, the ür, was
probably as com petent as sabha o f the brahmadeya, though less
is known o f it, unfortunately. In general there is little inform ation
on the functioning o f any village assemblies prior to the ninth
century when inscriptional evidence becomes vastly richer and
riveted attention upon the very conspicuous B rahm an villages.28
25 The Colas, p. 465.
26 K .A .N . S astri, th e m o st scholarly, prolific, an d a u th o rita tiv e w riter on C hola
h isto ry h as p ersistently held to the co n stitu tiv e view o f th e nadu; in his recent w ork.
Sources o f Indian H isto ry: with special R eference to South India (H eras M em orial
L ectures (1961), A sia P u b lish in g H o u se, M a d ras, 1964, p. 68), he w rites, ‘they
[R ajaraja I an d R a je n d ra I] perfected a highly organized ad m in istrativ e system
w hich ad m irab ly h it th e m ean betw een cen tralizatio n an d local a u to n o m y ...’.
27 In o th e r tim es a n d places, B ra h m an -an d -te m p le -sp o n so red schem es o f agricul­
tu ral dev elo p m en t occu rred , b u t there is no evidence o f this in the p eriod u n d er dis­
cussion. See. fo r Bengal, L a n d S y ste m and Feudalism in A ncient India, ed. D .C .
S ircar, U n iv ersity o f C a lc u tta, C a lc u tta , 1966, p. 15; for S outh In d ia, B u rto n Stein,
'T em p les an d A g ricu ltu re D ev elo p m en t in M edieval S outh In d ia ’, E conom ic W eekly
A nnual, B om bay, 1960.
28 A m o n g th e b est o f w hich is K .A . N ila k a n ta S astri, Studies in Cola H istory
Secondly, the nadu, or micro region, far from being an adm inistra­
tive device of the Cholas, existed long before as an im portant
unit o f society and culture as well as politics.
Some o f the m acro regions o f the Chola period existed during
the Classical period.29 The Classical poetic canon, Tolkappiyam ,
distinguishes twelve dialectical sub-regions among Tamils in the
southernm ost part o f what was then formally regarded as Tamil
country ( tamilaham).*0 These Tam il dialects are differentiated
from the Tamil spoken north of the Kaveri — the ‘northern
language’ ( vada-cof) — which were regarded as unfit to be included
as ‘correct’ (sentamil) on phonetic grounds. Centreing on M adurai,
where the language was considered m ost correct, the formal
linguistic region, Tam ilaham , extended from K anya K um ari north­
ward on the M alabar Coast to about twelve degrees o f latitude
and on the Corom andel coast to the thirteenth degree o f north
latitude. The Vengadam Hills m arked the northern boundary of
Tamil speakers according to m ost Classical sources.31
Several o f the nadus o f the Classical period persisted as dis­
tinctive localities until quite recent times. Two on the west coast,
pdli-nddu and kuda-nadu are known until the conquest by Mysore
in the late eighteenth century when their names were slightly
altered to Polanad and K utnad as which they continue to constitute
sub-divisions o f m odern divisions.32 On the Corom andel coast, the

and A dm inistration, U niversity o f M a d ras, M a d ras, 1932; b u t also in such survey

w o rk s as R .C . M a ju m d a r, Corporate L ife in A ncient India, S .N . Sen, C a lc u tta,
1922, an d R .K . M u k h erjee, Local Governm ent in A ncient India, C laren d o n Press,
O x fo rd , 1919.
29 N o te th e ex ten d ed en try u n d e r the term nadu in S u b ra h am an ian , Pre-Pallavan
In d ex, p. 488-9.
30 S. Illa k k u v a n a r, T holkappiyam in E nglish; with C ritical Studies, K u ral N eri
P u b lish in g H o u se, M a d u ra , 1963, p. 142.
31 R egions by d ialect are listed in S rinivasa Iyengar, op. cit., p p . 150-1; in this
p erio d , T am il d ialect territo ries included th o se on th e west coast, in w hat is now
K erala. F ro m twelve degress o f n o rth la titu d e these included Polinadu, Kudanadu, and
Venadu, th e last o f w hich co n tin u ed to be an im p o rta n t T am il region, K a n ak a sab h ai,
op. cit., pp. 14-15. A cco rd in g to K .K . Pillay, this b ro ad d elin eatio n o f the T am il
region o f an tiq u ity is referred to in several o th e r classical w orks as well as the T olkap­
piyam (w here it is fo u n d in the payiram , o r la u d a to ry preface): the Purunanuru,
Ahananuru, a n d C ilapa dikaram ; A Social H isto ry o f the Tamils, pt I, U niversity
o f M a d ras, M a d ras, 1969, pp. 14-15.
32 W illiam L o g an , M alabar: The M alabar M anual, S u p erin te n d en t o f G o v e rn ­
m en t P ress, M a d ras [reprinted], 1951, v. 1, pp. 645 an d 662.
following ancient nadus continued to the Chola p erio d : aruvdnadu ,33
and oyma-nadu , 34 in Tondaim andalam (Chola: jayangondachdla-
mandalam ), and konadu in C holam andalam .35 Mala-nadu (malai-
ttddu) between C holam andalam and Kongum andalam , dates from
the Classical age and during the Chola period was designated as a
special link region called rajasraya-valanadu ,36 Malddu (or miladu),
in m odern Tirukkoyilur taluk of South Arcot, continued as a
nam ed locality until at least the late years o f the reign o f Rajendra
Chola I, and it is one o f the few Coromandel localities to which a
numerical designation was appended in the m anner of medieval
K arnataka (miladu, ‘2,000’).37
The nadus o f Gangavadi in southern K arnataka predated the
Hoysala state by perhaps two centuries. These may date from the
first kingdom to be centred in this region, the Gangas, whose
territory encompassed m ost o f what was the princely state of
M ysore during the nineteenth century. Gangavadi was distin­
guished in medieval inscriptions from northern K arnataka. The
latter, being called kuntala or rattavadi , included the extensive
tracts o f nolambavadi, banavasi, and palasige. As w ith all territories
in medieval K arnataka, from the seventh to the fourteenth centu­
ries, a numerical designation was affixed 10 these: Gangavadi-
96,000, nolambavddi- 32,000, banavasi- 12,000, and paiasige-12,000.38
G angavadi was clearly linked to the interior upland o f Tamil
country as evidenced by the large num ber o f Tamil inscriptions
there and by the movement o f peasant folk between Gangavadi
and neighbouring extensions of,the Tamil plain as recalled in the
Kdrmandala Satakam o f the eighteenth century.39
33 Subrahm anian, Pre-Pallavan Index, p. 56. Ptolemy in the second century
a .d. noted the territory between the north and south Penner as ‘A rvrernoi'; K.V.
Raman, op. cit., pp. 4-5.
34 Subrahm anian, Pre-Pallavan Index, p. 187.
35 Ibid., p. 346; N ilakanta Sastri, The Colas, p. 109; V. Venkayya, ‘Territorial
Divisions o f R ajaraja Chola’, S .L I., v. 2, pt 5, pp. 21-9.
36 K.S. Vaidyanathan, ‘M aia-N adu’, Quarterly Journal o f the M ythic Society,
‘Culture and Heritage N um ber’ (1956), p. 225.
37 N ilakanta Sastri, The Colas, 265-6. A nother is vadugavali-12,000 in m odern
N orth Arcot, S.I.I., v. 3, no. 42, p. 90, late ninth century.
38 G.S. Dikshit, Local Self-Government in Medieval Karnataka, K arnatak U ni­
versity, Dharwar, 1964, pp. 18-19.
39 Karmandala-Satakam (in Tam il) with a comm entary by P.A. M uthuthan-
davaraya Pillay, M adras, 1930.
 It may therefore be argued th at if there was a ‘basic unit of
government’, it was the nadu, not the ‘self-governing village’. The
nadu was the basic peasant unit o f the age, it was also an ethnic
region to which the later, prestigious B rahm an villages and the
great overlords adapted themselves. The essential governmental
significance o f the nadu was its ethnic coherence. All persons
and groups directly involved in the peasant agrarian system of
a locality and jointly dependent upon the successful exploitation
o f the land tended to constitute a discrete social universe. Where
land capable o f being turned to the plough ceased, where slope,
aridity, hazards to hum an or animal welfare, or the presence of
a hostile people — peasant or non-peasant — who could not be
displaced occurred, the locality ended. W ithin that spatial uni­
verse, in most parts o f the m acro region, those with sufficient
authority to compel it forced the acceptance of social rules based
upon hereditary hierarchy and segmentation. Those whose military
power and agricultural skills had originally converted a tract of
land to peasant cultivation m aintained authority through control
over cultivable land and through connections with supra-local
chieftains.
In some parts o f the macro region — its backwaters, late to
emerge as full-fledged peasant tracts — the relationships between
a particular ethnic group and the territory was perhaps stronger
than in the older peasant core areas, for example, southern Pandi-
m andalam as com pared with the riverine portion of Cholam an­
dalam or Tondaim andalam . W here such emergence was late, clan
organization could assume great im portance. This is well ex­
emplified in the cases o f the Kallars and M aravars in the southern
peninsula.
The m odern districts o f Ram nad and M adurai became marginal
extensions o f ancient Pandya country as Kallars and M aravars
progressively converted their lands to peasant agriculture, possibly
during later Chola times. Subsequently, they expanded their con­
trol into portions of modern Tanjavur and Tirunelveli, becoming
the territorially dom inant people in some portions o f Tanjavur
which they ruled as their ‘eighteen nadus ' . In both places, however,
the original Ram and and M adurai ‘hom eland’ and the places
into which they later migrated, Kallars and M aravars permitted
strangers, including higher status ones like Brahmans and Vellalas,
to m aintain their corporate character while superior locality
control was retained by K allars and M aravars.40 According to
the Census of 1901, Kallars were divided into ten m ajor endogamous
divisions by territory (nâdu). D uring the nineteenth century,
within each o f the K ailar localities, one K ailar sub-division was
accorded the title of nattar and assumed a dom inant place in the
affairs o f the place involving not only all Kallars, but problems
involving other castes as well.41 M aravars, it appears, followed
the same pattern.
Locality governance varied in the Chola m acro region, in the
central portion o f the Kaveri basin and Tondaim andalam , the
nattar seemed to operate as an assembly w ithout a single, local,
executive chief, whereas is southern K arnataka and Kongu42—
the interior, upland tracts o f the macro region — nadu chieftain­
ships prevailed, in K arnatak inscriptions, local political relations
revolved around locality chiefs. Here, men with titles of gâunda,
prabbu, and pergade to which are prefixed the word nâd were
recognized holders of locality authority with specified prerequisites
o f dues and land. These were hereditary offices, and those who
held the offices were ‘ranked chiefs’.43
The dispersed settlement of peasants in the relatively dry interior
upland of South India would have required greater reliance upon
highly localized chiefs for protection. M oreover, in both Kongu
and Southern K arnataka, there was evidence of conquest by
peasant groups from the Tamil plain which established a locality-
conquering élite unlike any that can be identified in the lowland
during Chola times. Arokiaswami, writing of K ongu country
during the twelfth and thirteenth centuries, has pointed to the
40 William Taylor, ‘M arava-Jathi-V ernanam ’ from U npublished Mackenzie
M anuscripts in Possession of the Asiatic D epartm ent of the M adras Literary
Society, Madras Journal o f Literature and Science, v. 4 (1836), p. 357.
4> Thurston, op. cit., v. 3, pp. 71ff. citing J.H . Nelson, The Madura Country:
A Manual. Asylum Press, M adras, 1868. More recent corroboration of this comes
from Louis D um ont’s study of a branch of the Kallars, Une Sous Caste de l'Inde
du Sud (M outon, Paris, 1957, pp. 143ff) where he describes the conventional struc­
ture of four chiefs (tëvar), eight districts (nadus.) and twenty-four subordinate villages
in which one of the four chiefs is recognized as superior (raja) over others who are
regarded as his ‘m inisters’.
42 Arokiaswami, Kongu Country, pp. 251-2, 261-2, 271.
43 D ikshit, op. cit., pp. 38-40. ‘R anked’ chiefs are distinguished from ‘cerem onial’
chiefs by their possession of preferential access to or control over hum an and non­
hum an resources, implying a stratified difference from the essentially honourary
office of chief.
heavy influx o f plains groups into the region which resulted in
segmentary relationships between the ‘foreign’ lowland intruders
and those who exercised local authority previously. These latter
linked caste groups, according to Arokiaswam i’s argument,
identified themselves in and with the place by the prefix ‘K ongu-’
which also emphasized their indigenous status.44
In the riverine plains o f the lowland, however, the pattern of
local authority and control continued to be corporate until the
Vijayanagara period. Collective terms such as, nadayisainda
nattom , which N ilakanta Sastri translates as, ‘. . . residents of the
nadu m et [formed] as nadu. . ,’45 refer to those who manage the
locality through specialists.
Two kinds of administrative functions are prom inent in records
which refer to the m anagement of more densely settled, lowland
localities. These are m aintenance of local accounts (nadu-kankdni),
revenue registers and assessments (nddu-vagai and nadu kuru).46
General attention was also increasingly given by the nattar to
temple management where the official nadu-kankani supervised
temple accounts and income derived from lands which had been
granted by persons of the locality as devadana (‘gift to the god’).47
W hereas locality leadership in m ore isolated parts o f the macro
region, as K arnataka, vested hereditary chiefs with considerable
authority and power based upon the dom inant peasantry to which
they belonged, leadership in the more densely settled, ethnically
diverse, and wealthy parts o f the plain is m ore difficult to identify.
Here, the public actions of the nattar spokesmen for the dom inant
peasantry of the locality are undertaken corporately. Revenue
and registerial functions within each locality were the work of
agents of the nattar. If one is to try to locate a kind of executive
authority in the more populous, diverse, and wealthy localities
o f the m acro region, the m ost prom ising possibilities appear to
be persons designated by such titles as muvenda-velar and mum-
madi.

44 Arokiaswami, Kongu Country, pp. 270-1. W hat Arokiaswami discussed as

'social stratification’ (p. 271) is better labelled ‘social segm entation’ since distinctions
am ong later ‘invasions’ o f Kongu do not appear to be o f a ranked sort; these were
merely different groups o f equal rank.
45 The Colas, p. 503.
46 Subbarayalu, op. cit., p. 42.
^ Ibid., pp. 43-4.
 Little attention has been given to the precise meanings of titles
in Chola inscriptions and such ideas which are held relating to
functions carried out by persons using titles are wholly con­
jectural. Arokiaswami published a brief note on the term mu-
vendavelcir, which he called, 'a curious title’.48 ‘C urious’ it is,
as it has been used by historians, and Arokiaswami was referring
to the usage by N ilakanta Sastri. The titles adigdri and muvenda-
veldr are examined by the latter in a lengthy paragraph dealing
with ‘officials in the service o f the king’. In this discussion, such
ancient honorific titles as enddi and mdrdyan and the titles of
chiefs and leading men (araiyan and peraraiyan) are included.49
The examples provided by Nilakanta Sastri fail to convince that
these are ‘Officials’ in the sense implied by his interpretation of the
Chola state as centralized and bureaucratized. Thus, examples
o f ‘civic occupations’ o f officials are kadigaimdrdyan and vacciya-
mdrayan, which mean, respectively, ‘palace-bard’ or ‘time-keeper’
and ‘palace musician’.50 As for the title adigdri, 51 which Nilakanta
Sastri and others regard as ‘high officers of the army and in general
adm inistration’, again, in the presumed framework o f a centralized
political structure, it is noted that the title is often used without
personal names, but very often with a place reference. Of these
‘officials’, N ilakanta Sastri cites a ‘quaint account’ o f a later
medieval com m entator on a religious text who w ro te:
M e m b ers o f these [adigdri] fam ilies o nly a ccep ted a p p o in tm e n ts as
m antris a n d did n o o th e r w o rk . It is an im p ro p e r th in g th a t they are
fo u n d h o ld in g p o sitio n s o f a c c o u n ta n ts in these d a y s; except th a t th ey
c o u ld n o t w e ar a cro w n , th ey a re e n title d to all th e o th e r insignia o f ro y alty ,
a n d it is im p ro p e r fo r th e m to a ccep t a n y p o sitio n s o th e r th a n th o se o f
m a n tris.-2

48 ‘The Inscriptional Term , ‘‘M uvendavelan” ’ J.I.H . (April 1956), p. 191. See
also the recent prelim inary work by N. Karashim a and Y. Subbarayalu, ‘A Statistical
Study of Personal Names in Tam il Inscriptions; Interim Report II', Computational
Analysis o f Asian and African Languages, ed. M.J. H ashim oto, no. 3 (M arch, 1976),
N ational Research Institute of Asian and African Languages and Culture, Tokyo,
pp. 9-20. Definitions are not considered in this work.
49 The Colas, pp. 462-4. Suresh, H CG ESI, pp. 173-4, notes eighty references to
these terms which he equates with raja or ‘king’.
50 The Colas, p. 483, no. 61 and Tamil Lexicon, p. 668, on katikai. Tamil Lexicon,
p. 3576 on vdcciya-mardyan.
51 Tamil Lexicon, p. 73.
52 The Colas, p. 463. N ilakanta Sastri notes that the gloss on this work, Tak-
kayapparani, was written some centuries after the period of the Cholas.
The term mantri is significant here because it relates to the role of
the king’s advisor or counsellor, not a subordinate office-holder
in the bureaucratic sense.53 Similarly, while adigari is used to
refer to a ‘'director’ or ‘head person’, it is in the sense of a person
of personal merit, not official precedence.54
Muvendavelar is a terra frequently used in Chola inscriptions.
It is often joined with the term adigari to designate an im portant
personage, often without his personal name. Arokiaswami’s un­
derstanding of the term is like that of Nilakanta S astri: it connotes
people with experience in ‘high administrative jobs’, where ‘jo b ’
is m eant to convey a bureaucratic office.55 M uvendavelar is a
phrase which appears to m e a n : Vellalas (velar) o f three (mu)
kings (vendan).56 Arokiaswam i has provided some instances of
Vellalas serving the Cholas and their predecessors which support
the association suggested by the term velar in the title.57 Again,
as in adigari, persons with this title appear less as occupying func­
tional political roles than as possessing an esteemed status, the
im portance which attaches to men who serve kings as full agents
of the ruler’s authority.
Persons with the titles m entioned above are frequently not
identified by their personal names, yet their locality or village is
often given. N ilakanta Sastri’s understanding o f place, in pre­
ference to personal, identifications is consistent with his general
notions about the Chola state: lie sees the reference as indicating
the place from which a royal officer derives his salary; he sees it
as a kind of fief, or livelihood (j'Tvita), which might include a village
or a ‘district’ (nadu).5% There is absolutely no evidence of such
prebends, or assignments o f village or nadu income, in Chola times,
and an arrangem ent o f this sort would seriously have weakened
the putative central control ascribed to Chola rulers. Another,
and more convincing explanation of the place designation, is that
the personage bearing the title was, in fact, a notable o f the place
m entioned in his title, that such persons were not royal officials
or central officers as purported, but locality chiefs.
53 Monier-W illiams, A Sanskrit-English Dictionary, p. 785.
54 Tamil Lexicon, p. 73.
55 Arokiaswami, ‘The Inscriptional Term “ M uvendavelar” ’, p. 191.
56 Tamil Lexicon, p. 3834. ventan and p. 3843, velan.
57 Arokiaswami, ‘The Inscriptional Term “ M uvendavelar” ’, p. 192.
58 The Colas, p. 464.
 This suggestion is supported by the data which Suresh has
collected. Though his understanding of muvendavelar conforms
closely to the conventional one, Suresh undertakes a closer study
of the term .59 Forty-nine references to muvendavelar are cited
by him from a list of 449 ‘officers’ culled from the corpus of Chola
inscriptions he examined. F or almost every one of these 449 men,
the place of the person is clearly recorded in the inscription along
with a title like muvendavelar 60 or brahmardyan, the latter indicat­
ing a Brahm an notable.61 In the cases of non-Brahm an notables,
a personage with the attributes of a chief was reliably indicated
by the suffixes: udaiyar, ajvdr, or araiyan. Suresh also notes that
another standard term affixed to the names of notables, mummadi
was equivalent to muvendavelar ,62 In the title mummadi-chdla
ilango vel reference is to a ‘chief’ or ‘king’ o f the Ilango lineage
which flourished in the Kaveri basin for a long time.63 M ummadi,
or more properly mummudi, thus, appears to be another way of
referring to the ‘three kings’,64 either in the sense that the Cholas
had become dom inant over the territories o f the three ancient
peoples — Cholas, Cheras, and Pandyas — but more likely in the
sense that the holder of the title was a local notable recognizing
the sovereignty of kings.
These titles were held by chieftains’ families o f long standing,
not by persons whose status derived from the officialdom o f the
Chola state. True, some great families of chiefs continued to be
referred to in inscriptions, thus, preserving an ancient identity.
They are described as an ‘aristocracy’ by Suresh and included
great families of chiefs (peraraiyar) whose existence is attested
during the Classical period. Tondaim an chiefs probably originated
in Tondaim andalam ,65 but their territorial power in the Kaveri
basin was well-established in villages (urs in which they were the
principal people (M ar). M aravar chiefs, similarly carrying the
title of draiyar, remained im portant at least until the time of
59 Suresh, H CG ESI, pp. 230-1.
60 Explained, ibid., p. 227.
61 Explained, ibid., p. 182.
62 Ibid., p. 227.
63 Ibid., pp. 194-5.
64 Ibid., p. 227; Tamil Lexicon, p. 3272.
65 The Perumpananuppatai of perhaps the second or third century celebrates
the chieftain T ondaim an Ilantiraiyan o f K anchi: K. Zvelebil, The Smile o f Murugan;
On Tamil Literature o f South India, E.J. Brill, Leiden, 1973, pp. 56-7.
 Kulottunga III; they appear less as a kin-linked ruling group in
the arid country south of the Kaveri than as an assortm ent of ruling
families originating from various parts of the macro region who
found opportunities on the southern Chola frontier. Among the
M aravar chiefs were the Ilangovels, referred to above, who enjoyed
a somewhat special relationship with the Cholas through marriage.
Kadavarayans, a chieftain’s family o f m odern Tanjavur district,
were also identified as M aravars though they also appear to have
been linked to another ancient chief’s family, the M uttaraiyars of
the relatively arid tracts north o f the Kaveri, in m odern South
Arcot. Other families of chiefs o f antiquity who are prom inent
during the Chola period include: Banas, Irukkuvels, Kadup-
pattigals, Sengenis, and Tamilavels. Additional great chieftains’
families were apparently Valavadaraiyans, Tennavans, and those
called Pallavans or Pallavaraiyans who may also have been warriors
from Tondaim andalam originally, though this is unclear.66
The prominence of these ancient chieftains’ families is marked
by the frequency with which they were mentioned in Chola ins­
criptions, the titles which they bore, the villages and localities
with which they were associated as principal people, and finally,
the practice of including Chola royal titles (e.g., ediriliehola or
vikramachola ) or the names of the Chola rulers in their own titles.
Suresh interprets the last practice as a sign o f m atrim onial relations
between ‘aristocratic’ families and the royal family;67 he also
supposes that there were m atrim onial relations am ong these
‘aristocratic’ families.68
A good example of the use of Chola royal names by quite in­
dependent chiefs comes from the interior upland of K ongu (m odem
Salem and Coim batore districts). A line of rulers of this territory,
called the ‘Kongu-Cholas’ is regarded as a ‘vice-royalty’ of the
Kaveri Cholas by Arokiaswam i.69 The Kongu-Cholas ruled from
the beginning of the eleventh to the late thirteenth century. Most
of the rulers in this line bore Chola regnal names, such as U ttam a-
chola ( a . d . 1100-1119), Virachola ( a .d . 1119-1134), and Vira
66 Suresh, H C G E Sf, pp. 301-8. Some o f these families were prom inent in Pallava
tim es; see M.S. Govindasam y, The Role o f Feudatories in Pallava History, Annam alai
University, A nnam alainagar, 1965.
f” Suresh, H CG ESI, p. 302.
68 ibid., p. 309.
69 Kongu Country, pp. 229-48.
Rajendra ( a .d . 1207-1249). While claiming the great territory of
kohgudesa, m ost of the 300 inscriptions of this line of chiefs were
concentrated in two places: D harapuram in the southern portion
of m odem /Coim batore and Avanashi in the central portion. Both
places seemed to be on an ancient road from the Kaveri basin to
M alabar over the Nilgiri Hills, and the progenitors o f this line of
chiefs may have been Irukkuvels who migrated from the southern
bank of the Kaveri during the Chola period.70 It may also have
been true that the Irukkuvels were linked in m arriage to the great
Cholas, but the independent career of the Kongu-Chola chiefs
is too clear to regard them as actual viceregal deputies of the great
Cholas.
Political alliances involving marriages am ong families of chiefs
and between them and the Chola ruling family certainly did occur
judging from the pedigree of some of the Chola queens. However,
the argument that the use o f a Chola royal title or name of a Chola
ruler in the title of a chief is a sign of m atrim onial relations is
questionable. Such titles seem to have been used frequently among
even quite m inor chieftains’ families noted am ong the 449 names
of ‘officers’ by Suresh.71 Unless it is to be argued that every local
notable — a person claiming the title of chief, and claiming kilvar
status with respect to a locality and using the title or name of
ruler — was actually tied to the Chola ruling family by marriage
relations, the practice m ust be explained some other way. W hat
is suggested in the evidence available is that the royal Chola family
exercised a ritual sovereignty over the numerous locality notables
in the plains and neighbouring upland; one expression o f this
sovereignty was the adoption of Chola royal names and titles by
local chiefs.
As in the case of the great families of chiefs referred to above,
it must have been the case that most m inor families of chiefs existed
before Rajaraja Fs time. There is no suggestion in the evidence
of the Chola period or in the interpretations of m odern historians
that either the great families of chiefs or m inor chieftains were
eliminated by the C holas; thus, the only reasonable position which
can be taken is that they continued to exist. In relatively isolated
portions of the plain — tracts less favourably situated for dense
70 Ibid, pp. 241-2. Also see, K.V. Subrahm anya Aiyer, ‘Seven Vattefuttu Inscrip­
tions from the Kongu C ountry', E.I., v. 30, no. 19, p. 100.
Suresh, H CG ESI, pp. 427-6.
settlements — older families o f chiefs as m ost o f those noted above
continued to hold sway, preserving their ancient family identities
even as they assumed additional titles linking them to the Cholas.
in those more populous tracts, well favoured by irrigation and
consisting o f quite diverse ethnic groups, the dom inant peasantry
continued as before to exercise control through their chiefs. How­
ever, as in the case of the more ancient families of chiefs, the identity
of these local chiefs — spokesmen for the nattar of the locality —
also changed to reflect a powerful new and politically able over­
lordship, that o f the Cholas.
The titles which appear to m ark off older locality chiefs in more
densely settled, pluralistic localities were muvendavelar and mum-
rnadi. References to the Chola hegemony in the titles of these
notables are not to political subordination, for the power of the
nattar and their chiefs seems to have continued undiminished.
The titles, rather, point to a formal allegiance to, and a ritual
linkage with, the Cholas by locality authorities whose power was
based upon their ties with local, dom inant peasantry from which
they themselves originated.
Shared sacral allegiance was another significant dimension of
the territorially segmented society of South India during medieval
times. The responsibility of the locally dom inant nattar for the
m aintenance and supervision o f shrines was vested in officials
(ndyagam) of the nadu assembly. The nattar of a locality in which
a Brahm an settlement was established were clearly the m ost im­
portant of the several groups which participated in the ceremonial
events m arking such foundations. But what is most interesting
about the relationship o f religion and the territorially and socially
segmented society of this age is the interaction of localized forms
of folk religion and canonical religion, and these with the Chola
state. This is considered in chapter VII.
Seen as micro regions of overlapping distributions of m arriage
and kinship, networks, social coalitions beyond kinsmen, agrarian
relations, political and religious affiliations, the nadus of the South
Indian m acro region varied greatly. In most parts of the Kaveri
and Palar basins — Cholam andalam and Tondaim andalam —
nadus were complex societies consisting o f relatively large popula­
tions, considerable wealth, and diverse social groupings living in
numerous, pluralistic settlements. In more remote upland portions
of the m acro region or in the less-favoured ecological settings of
parts of the Tamil plain, a simpler model may be considered, such
as th at presented by the anthropologist F .G , Bailey.
In his work on a part of the m odem peasantry of Orissa, Bailey
has used the concept of ‘clan’ to analyse relations between peasant
lineages and the territory they jointly occupy even though there
is no actual descent group which can be called a ‘clan’ in the classical
sense o f that term. This putative ‘clan’ structure appears as a
means for associating a diverse set of lineage groups with a parti­
cular portion of land. It provides an im portant element in Bailey’s
idea of the ‘bloc’, an integrated structure of discrete locality social
groups with w'hich other, similar, units have historically been ag­
gregated to form larger units, even states.72 Bailey’s arguments on
the character and functions of the ‘clan’ have been attacked on
theoretical grounds.73 On empirical grounds, it appears doubtful
that such ‘clan’ units as he has found in highland Orissa, to which
could be added such groups as the peasantry of Kongu or the
Kallars of M adurai and R am nad, ever constituted more than a
m inor variant o f peasant organization within the Indian cultural
sphere. The m aintenance of such sequestered social units has
been rare and probably restricted to relatively isolated territories,
cul-de-sacs , or shatter-zones.74 In perennial peasant territories,
involving the greatest numbers o f people, the problem for the
peasantry was not the m aintenance of some putative or real kinship
organization, but the form ation of viable coalitions based upon
the m utual interests of various kinship and occupational groupings
and their dependents. It would appear that the ‘clan’ model is
appropriate for some cases, but by no means of most.
The complex agrarian operations o f m ost nadu micro regions
are m ost clearly illustrated with respect to the redistribution
of locality income in the form of eleemosynary grants. Ironically,
this is the same evidence upon which those who argue the primacy
of the village depend. From Pallava times, changes in the status
o f villages and portions o f villages, in which dom inant cultivating
72 Political and Social Change in Orissa. Also, ‘Closed Stratification in India’,
Archives Européennes de Sociologie, v. 4, no. 1 (1963), pp. 107-24.
73 See, Louis D um ont, ‘A N ote on Locality in Relation to Descent’, Contribu­
tions to Indian Sociology, v. 7 (M arch 1964), pp. 71-6.
74 Bailey, ‘Closed Stratification', seems quite aware of this and suggested in his
postscript, pp. 122 and 124, that changes in scale of the territorial system would
effect caste and caste systems.
groups relinquished a portion of the wealth they controlled to
Brahman or iain a reli.gi.euse, involved transfer arrangements
executed through the dom inant locality peasantry acting cor­
porately as the nadu assembly. Form al actions solemnizing these
changes involved a purported royal order ( tirumugam ) or com ­
m unication (koriolai) presented to the nattar. An example of
this procedure is contained in one of the earliest Tamil inscriptions,
of the middle sixth century, from Pallankoyil, Tirutturaipundi
taluk, Tanjavur:
Let the assembly of the nadu (i.e ., the n a tta r ) of the Perunagara-nadu,
a sub-division of Venkunra-k-kdttam observe. We have granted the
village of Amanserkkai in your nadu as paU ichchandam to the teacher
(kuravar) Vajra-Nandi of Paruttikkunril. (Accordingly) you also walk
(around) the hamlet [padagai], plant [boundary] stones and milk-bush
(k a lli) and issue the order for proclamation to the assembly of the nadu
(n a tta r). And the members of the assembly of the nadu (n a tta r) having
seen the royal order, made obeisance and placed it on their heads, walked
(around the boundaries of) the padagai, planted stones and milk-bush
and issued the order for proclamation (araiyolai) according to which the
boundaries are. . . ,75

There are other instances o f this in Pallava inscriptions.76 Chola

inscriptions carry out the same arrangem ents with some variations,
perhaps reflecting the development, by that time, of various cor­
porate entities who shared in the changes in distribution of locality
resources. In the famous Larger Leiden Plates o f R ajaraja I,77
the village o f Anaimangalam was granted to support the Buddhist
shrine (p a lli) being built at N agapattinam by the M alaysian ruler
Chulam anivarm an of K adaram .78 Those addressed in the Tamil
portion of the plates included: the nattar o f Pattana-kurram , the

75 T .N . Subram aniam . ‘Pallava Jain C opper-Plate', Transactions o f the Archaeo­

logical Society o f South India, p. 82, hereafter: T.A .S.S.I., 1958-59; reported in
A .R .E . 1958-59 as C.P. no. 10 and comm ented upon in the Report, p. 3; also see,
Zvelebil, Tamil in 550 A.D ., pp. 13-14.
16 F or example, see ‘K asakkudi Plates o f N andivarm an II', dated c. a . d . 753,
S.I.I.. v. 2, no. 73, pp. 342-61.
77 E.I. v. 22, no. 34, pp. 213-66, dated a . d . 1005 and edited by K.V. Subrahm anya
Aiyer.
78 A recent record from N agapattinam is of interest in this connection, A.R.E.
1956-7, no. 166, dated in the reign of R ajendra I, a . d . 1019, and refers to what
appears to be a reciprocal gift by a Sri Vijaya ruler of Chinese Gold ( clna kanagam)
to support a temple in N agapattinam .
spokesmen (kilavar )79 for brahmadeyas, those in charge o f assemb­
lies (urgalildr)so o f H indu temple villages (devadana ), Jaina or
Buddhist shrine (paUichehanda) villages, those settlements exempt
from all dues (kanimurruttu ),8i villages whose income may have
supported public works or vedic sacrifices,82 and trade centres
(nagara). The nattar appear, from this record, to have been the
primary recipients of the order,83 and they appear also to have been
responsible for incising the record as suggested by the following
lines:
A roya! order ( tirum ugam ) embodying the above [specific income from
Anaimarigalam which were henceforth to be paid to the shrine] and with
the words, ‘it behoves you [the nattar] to be with those persons [named
as witnesses from many parts of the Coromandel plain] to point out the
boundaries, to go around the hamlet accompanied by a female elephant,
to set up [boundary] stones and milk-bush [markers], and to draw up
and give the deed of the gift’ . . . [This order] having been sent to us,
the n a tto m , . . . we the n a tto m respectfully received and carried on our
heads and accompanying the female elephant, walked round the hamlet,
set up stones and milk-bush, drew up and gave the deed of gift.84

79 The word 'kilavar' is translated in slightly different ways by those who have
comm ented on its usage. K.V. Subrahm anya Aiyer in his edition of the record,
£'./. v. 22, p. 258, renders it as ‘headm an’; T.N . Subram aniam (in the Pallankoyil
epigraph, p. 88, discussed above) translates it as ‘owner’ ; the Tamil Lexicon (p. 936)
gives another m eaning in addition, ‘aged m an’, or elder, which, with ‘spokesm an’,
seems to come closer to the m eaning in the context.
80 U nder urgalilar, the collectivity or ‘body of the u r\ Subrahm anya Aiyer appears
to put all village assemblies, including nagaras. except Brahm an villages with sabhas,
E.I. v. 22, p. 231.
si The epigraphist, K.G. Krishnan, in a personal com m unication, has suggested
the kanimurruttu grant was not for Brahmans, as usually understood, but for non-
Brahman teachers.
S2 Vettaperu is sometimes read as vcttiperu by Subrahm anya Aiyer (pp. 231 and
247, line 107) or by H. K rishnasastri, in his reading of the Tiruvalangadu Plates,
where the m eaning is as a service tenure of some sort, perhaps like that of sluice-
keeper as is suggested by Subrahm anya Aiyer in his reading of the Larger Leiden
Plates. However, T.N . Subram aniam reads She term as 'vei.ta(i)peru signifying
"the perform ance of (Vedic) sacrifices; Chola Jaina C opper Plate G rant', T.A .S.S.I.,
1958-9, pp. 91-2.
83 The Tamil portion of the T iruvalangadu Plates of Rajendra I is addressed to
the nattar and other locality groups in melmalai palaiyanur-nadu and directs that
the village of Palaiyanur, for which place the locality was named and which was a
Brahm an village (brahmadeya), was now to become a village subject to regular
dues from cultivators (velldn-vagai) and these were lo be granted to the temple ol
Tiruvalangadu as devadana. The opening Tamil portion follows the form cited
in the ‘Larger Leiden Plates' above, v. 3. p. 427.
»4 E.I., v. 22, p. 259, line 49fT.
After a detailed description of the lands, the income from which
comprised the gift, the signatures o f twenty-seven persons o f the
nadu, representing twenty-six m ajor settlements o f the locality,
are appended.85 Ten o f the twenty-six people o f the locality were
members o f B rahm an village assemblies; the balance o f the signa­
tories represented mixed Brahman and non-Brahm an assemblies
( urdr).
An analysis o f these nadu signatories is o f some interest.86
Lines 210 to 300 o f the Larger Leiden Plates provide the names,
titles, and villages o f each. O f the twenty-seven signatories, the
titles o f six was madhyastan and ten were called madhyastan-
karanattan. The first term means, ‘neutral person’ and sometimes,
‘headm an’87 while karanattdn m eans ‘accountant’; the latter is
given as the title o f two o f the nadu signatories. Five others were
entitled, karanattdn-vetkovan , the latter term meaning, ‘potter’.88
O f the balance, one of the signatories is identified as a vaikhdnasa
or Brahm an who followed the ancient and exclusive vaishnavdgama
o f th at name and, interestingly, represented not a brahm an village
assembly, sab ha, but an u r; as to the rest, the record provided no
other inform ation than the village which they represented. The
titles, madhyastan and karanattdn suggest persons whose functions
are nadu-wide, not simply powerful men o f individual villages.
Karanattdn especially, points to the fact that nadu management
was in the hands o f members o f the locality population, and it
anticipates the im portance o f the office o f the ‘karnam ’ in the
eighteenth century when the British first encountered it.89 Then,
and earlier, the office o f locality accountant was filled by im portant
local people. ‘M adhyastas’, people o f the middle, suggest the
position of the dom inant landed groups who sought to remain
aloof from divisions within the locality. They were not impartial
85 Ibid., List 'B \ pp. 237-8.
S6 Ibid., plates 11-15, pp. 263-6.
87 U nder this term, Subram aniara, S .l.T .l. Glossary, xxxii. gives ‘headm an’;
Subrahm anya Aiyer, throughout, uses the term ‘a rb itrato r’, cf. lines 2! Off; for
karanattdn, see Subram aniam , S J .T J . Glossary, p. xxv, ‘accountant'.
88 Subram aniam , S .l.T .l. Glossary, p. xci; Tamil Lexicon, p. 3821.
89 For example, the Tanjavur karnam continued to represent local interests long
after the appointm ent of government accountants in 1807; the form er were dis­
tinguished by the title kudi karnam in contrast to the government, or sarkar karnam
according to F.R, Hemingway, Tanjore (M adras District G azetteers); G overn­
m ent Press, M adras, 1915, p. 193.
m ediators, but arbitrators whose decisions involving locality pro­
blems was based upon their powerful positions within local society.
Again, one m ust look to the eighteenth century record for the
right- and left-hand divisions and the position o f the dom inant
peasantry which transcended both.90
This prominence given to the locally dom inant peasantry is
further supported by those parts o f the Larger Leiden Plates
which record signatures. The first signature on the list was a
mem ber of the dom inant peasant group in the village of Anaim an-
galam :
When the n a tta r were accompanying the [female] elephant and circumam-
buiating the hamlet of Anaimarigaiarn, I Kon Puttan, a Vellala, residing
at this Anaimarigaiam, mounted the elephant, was present with them
[the other witnesses in the procession], and showed the boundaries dearly,
and this is my signature.91
Other inscriptions also refer to the circum am bulation of a newly
constituted settlement or portion o f agricultural land granted to
persons or to a temple, and in these cases too there are num erous
local witnesses. In the Tiruvalangadu Plates o f Rajendra Chola
I, there is detailed reference to the circum am bulation procedure
and, again, a mem ber o f the dom inant cultivating group o f the
locality, 'one bom o f the fourth caste (chaturthanvaya)', led the
procession and concluded the other activities involved in dem ar­
cating the boundaries of the village.92
Dikshit has provided yet other examples o f corporate nadu
action primarily from southern K arnataka during the eleventh,
twelfth, and thirteenth centuries. In K adur district, a locality
assembly (nadu) of the Devalige '70’ territory, consisting of seventy
gaundas, m et in a . d . 1074 to reward the public work o f an indi­
vidual; at Hemm anahalli, in M ysore district, an assembly formed
o f gaundas or prabhugdvundas of a number of villages met with
a w arrior chief (dandanayaka ) in a .d . 1175; in another Mysore
district village, T ondanur, in a .d . 1158, prabhugavundas and the
representatives o f m ajor peasant households, okkalu, o f thirty-
villages met to arrange support for the Vishnu deity o f the place
(vittirunda perumala ).93
90 Cited in C.S. Srinivasachari, ‘Origin of the Right and Left H and Castes’, p. 85.
E.I., v. 22, lines 207-10. p. 262.
92 S.1.1., v. 3, verse 132, p. 426. Also see, C.P. no. 10 and no. 14, 1958-9 and
A .R .E ., 1958-9, pp. 4-5.
93 Dikshit, op. cit., pp. 44-5.
 It is necessary to underscore the central role taken by the do­
m inant peasantry in reallocations of locality wealth with w'hich
records like the Leiden Plates of R ajaraja 1 deal. Scholars have
long recognized the special functions of the nattar. We have
K.V. Subrahm anya Aiyer asserting that the prestigious representa­
tives o f brahmadeya villages, devadana villages, trade centres
(nagara) as well as other eleemosynary beneficiaries were under
the adm inistrative control o f the nattar.94 A dditional new light
has been provided in the recent, excellent work o f Subbarayalu.
He has clarified the connection between such well-recognized
bodies as the sabhas o f brahmadeyas, representatives of other
religious institutions, the well-recognized m ercantile interest of
nagaratiar o f a locality and the very dimly perceived velldn-vagai
villagers.
O f 1,300 villages noted by Subbarayalu in Cholam andalam ,
about 250 were brahmadeyas, about 50 were devadana villages,
part o f whose income was designated for particular temples and
over which temple authorities had some adm inistrative control,
and 26 were nagaras, trade settlements and subject to the special
influence o f m erchant assemblies. Villages identified as consti­
tuting kanimurruttu, pallichandan (supporting Jaina savants or
shrines), and vettaperu grants are few. This leaves the vast num ber
o f villages in the residual category o f velldn-vagai, or peasant
share villages, which m ost scholars recognize as the m ost num erous
o f all types o f villages.95
Subbarayalu notes that among the various representative groups,
one, the m ost num erous category, peasant villagers (uraa), is
missing. In a com petent review o f a series of inscriptions in which
the nattar figure prominently, Subbarayalu concludes that the
nattar can only be identified as the representatives o f this largest
category o f village units o f a locality—those called vellan-vagai
villages. H e goes on to argue as follows:96
If the N adu were a VeUanvagai group, the question may be raised why
they are confined to only their respective Nadu units, when similar groups

w E L , v. 22, p. 231.
95 A. A ppadorai. Economic Conditions in Southern India (A .D . 1000-1500), Uni­
versity of M adras, M adras, 1936, v. 1, p. 152 and implied by N ilakanta Sastri,
The Colas, where he speaks of ‘ryotw ari’ tenure (p. 531) and ‘peasant proprietor­
ship’ (p. 570).
96 S u b b aray alu , op. cit., p. 36. H is usage o f ‘n a d u ’ an d ‘nadu' is m eant to d is­
tin g u ish betw een th e territo ry and the assem bly.
 existed in the neighbouring units. It is really a fact that the N adus [i.e.,
nadu assemblies] functioned only within the limits of their Nadus [terri­
tories] at least till the end of the eleventh century a . d . The answer to the
question is closely related to the basis of the N adu region already con­
sidered. It was suggested that the Nadus were agricultural regions
formed of groupings of agricultural settlements. Since the N adus covered
only small areas, it is possible that each group of agricultural settlements
consisted mostly of kinsfolk. That is. each Nadu was basically a cohesive
group of agricultural people tied together by marriage and blood relation­
ships, the so-called chief villages forming the core of each group at the
beginning. Incidentally, it may be pointed out that only on this hypothesis
can a satisfactory explanation be given for the . . . many caste and com­
munal sub-divisions of today which are mostly territorial in origin.
Thus, then, because of the cohesive character of the segment o f the society
which each N adu contained, the N adu [assembly] functioned only within
the limits of the respective territorial units [. . . . and] the N a tta r were
the Vellanvagai group.

This view is very close to that o f another who has dem onstrated
a mastery o f Chola inscriptional evidence as well as a very acute
historical sense, T.N. Subram aniam. He has emphasized the pro­
minence of the nattar in Pallava as well as Chola times, noting th a t:
. . . devandana, palU chanda, and brahm adeya villages were, at the time
of their being granted . . . separated from the jurisdiction o f . . . the
nadu in which they were situated and constituted as autonomous villages.97

Peasant leadership in a locality, even if deeply rooted in networks

o f kinsmen and their dependents, could be threatened militarily.
Successful nattar control dem anded the capability o f defence
against both non-peasant m arauders as well as other peasant folk.
The exigencies o f defence and the opportunities open to the more
powerful for aggression against weaker neighbours provide one o f
the m ost im portant explanations o f the valangai velaikkara,
soldiers o f the right-hand division, who were the best-organized
military organization o f the Corom andel peasantry during the
Chola period. The conception o f the valangai velaikkara as a
standing, royal army is rejected here in favour of a conception of
these essentially local armed units constituting a potential force
which could be mobilized as required. Themselves the dom inant
peasantry in contiguous localities who periodically banded together
97 T.A .S.S.I., 1958-9, p. 88. He goes on to say that an analogous situation exists
in m odern Indian tow ns with panchayat unions within, but administratively separate
from , the m odem revenue district adm inistration.
for purposes o f mutual defence and occasional predatory raids
under the leadership of an im portant overlord, the valangai leaders
appear occasionally to have sought new agrarian territories.
Tracts which were already under the plough but poorly defended,
provided the most attractive targets.
The nattar o f Tondaim andalam looked to the settled lands to
the north-west for such expansion, and they were to remain the
dom inant peasantry there for centuries. The northern portion of
the Tamil plain, orT ondaim andalam , was called Jayarigondachola-
m andalam from the late tenth century to the twelfth. According
to a Tamil inscription, found in M ulbagal taluk, K olar district
o f m odern K arnataka, the conquering peasantry o f Jayahgonda-
chdlamandalam , ‘a 48,000 country’, recorded a revenue arrange­
ment in a . d . 1072.98 This arrangement apparently sought to
maintain revenue practices extant in Tondaim andalam in this
portion o f K arnataka which had been conquered during the early
years o f Rajaraja I " and perhaps reconquered during the time
o f Rajendra Chola I. Following an introductory section referring
to the conquest during R ajendra’s time, the inscription reads:
We of the assembly of the eighteen countries [or nadus] :1()0 [sn rdjendra-
chdla-padinen-pum i-periya-vishaya], the great army of the right-hand
armed with great weapons [perum badi-valangai-m ahdsenai ] . . . have
caused an order [sdsanam] to be engraved on stone to the effect that there
was no tax on cows and she-buffaloes since occupation of Nigarili-
choia-mandalam seventy-eight nadus [Nulambapadi or central Karnataka]
by the sacred family of the Cholas as well as in Jayangondachola-manda-
lam 48,000 country, in both o f which . . . the cultivators of the whole
country came and settled. Therefore, the tax upon cows and she-buffaloes,
ordered by [the officer] Adigarigal-Chola Muvendavelar should not be
paid. For dry lands with dry crops, there should be paid a m elvaram
of one in five. For lands under tanks, there shall be paid a m elvaram
of one in three. For every 1500 k u li of land in which k u m m a ri [crop by
shifting cultivation] is raised by the forest tribes [veclar], one cloth [pudavai]
shall be received. Additional dues such as k u m m a ra kk a ch c h a n a m [temple
dues], washermen’s fees, dues on good buffaloes and good cows, and the
like shall be two ka su per head . . . [damaged portion], for petty dues,
the d su v im a k -k a l [Ajivakas] shall pay one k d su per head, one-quarter

98 The text and translation are published in E.C., v. 10, no. 49a, pp. 86-7; other
partial translations, N ilakanta Sastri, The Colas, 538-9; K.V. Subrahm anya Aiyer,
Historical sketches . . ., v. 1, pp. 351-2; Dikshit, op. cit., p. 49.
99 N ilakanta Sastri, The Colas, p. 174.
100 D ikshit, op. cit., also notices this usage in K annada inscriptions: hadinentu-
vishaya or hadinentu-nadu (pp. 34 and 193).
 k a su shall be received from each house of teachers [uvatti], o f the men in
charge of the temple and talarar [watchmen or temple (tali) committee].
The houses set apart for the minor toils [siru-sungam] are exempt. The
lands shall be measured by rods 18 san in length, a san being equal to . . .
[damaged]. Thus, [we of] the eighteen countries and the great army
of the valangai possessing large weapons had this document engraved
on stone.

South-central and south-eastern Mysore, Gangavadi ‘96,000’

during the twelfth century, included the territory between the
K olar and Hassan districts o f m odern K arnataka and the northern
portion o f m odern Coim batore in Tamil N adu.105 That this
territory was under the dom ination of Tamils is clear from the
large num ber o f Tamil inscriptions there.102 Reference to nigarili-
chola-mandalam. confirms the fact that this part o f K arnataka
was under the control o f warriors associated with the Cholas;
it was also called Nulam bavadi and extended north beyond the
T ungabhadra river in what is now Bellary and A nantapur districts
o f A ndhra Pradesh and a portion of Chitaldrug district,
K arn atak a.103
The dom inant peasants who referred to themselves as ‘we of
the Sri- Rdjéndrñ-chola-mandalam-padinen-pümi- vishaya’, or the
assembly o f the eighteen nadus of Sñ-Rajendrá-chola-mandalam,
were possibly a part o f the arm y o f R ajendra I which conquered the
area in a . d . 1015, as stated in the introduction of the inscription,
or were migrants from the Coromandel plain after that conquest.
From inscriptions of the time of Kulottunga I, a . d . 1080, and
R ajaraja II, a . d . 1160, which refer to ganga-nadu in Nigarili-chdla-
mandalam, it appears that a deep thrust of the Chola overlordship
into K arnataka had brought eastern parts of that country into a
101 A detailed geographical description is contained in B.R. O opai, ‘The Later
W estern C hálukyas’. U npublished Ph. D. Thesis, University o f Mysore, 1961,
App. IV.
102 See M ysore Castes and Tribes, v. 1, p. 114, fig. 8, for a m ap showing the distri­
bution of Tamil inscriptions in the districts o f Mysore, Bangalore, and K olar with
a scattering in T um kur as well. O f such records in Mysore, there are over 100 along
this northern part of the Kaveri, and those of Bangalore and K olar cluster heavily
in the fertile basins of the Ponnaiyar and Palar with a secondary distribution along
the Penner.
103 This is defined in G opal, op. cit., with the designation, ‘a 32,000 country’;
an inscription o f a . d . 1219 (303/1913) equates Nigarili-chSla-mandalam with Nulum-
bapádi, see N ilakanta Sastri, The Colas, v. 2, pp. 344 and 723-4.
state of prospective settlement by Corom andel peasant settlers.104
The nucleus of th at settlement may have been military colonists
similar to those which, during the reign of K ulottunga I, were
established in other outlying areas of the Chola overlordship.
Colonies called nilaippadai were found on the principal lines of
com m unication in Pandya country and in T ravancore;105 these
colonies of warrior-peasants were reported to have been established
in the northern hill tracts around Kalahasti (modern Chittoor
district, A ndhra Pradesh) and may have existed in still other parts
o f the m acro region where C hola conquests occurred.106 The
references to Nulam bavadi and its association with peasant settlers
from Tondaim andalam also points to the vast territory which
had come to be included under the influence of cultivating groups
from the Corom andel plain by the eleventh century. Though this
great territory was not continuous, but broken by ranges of hills
and, at th at time, dense forests, those tracts capable o f settled
peasant agriculture based upon tank and riverine irrigation ap­
peared to be in close contact with each other.
A nother interesting aspect of the inscription discussed above,
in addition to the references to the valangai which is commented
upon further below (ch. V), is the designation o f Tondaim anda­
lam as a ‘48,000 country’. This occurs in other inscriptions of the
period, and it is am ong the few instances in which a territory is
given a numerical designation in Tamil inscriptions.107 As noted
above, such designations are com m on in K annada inscriptions
or Sanskrit inscriptions from medieval K arnataka; they also occur
in inscriptions from A ndhra.108 This M ulbagal record of the time
104 676/1905 and 18/1900; N ilakanta Sastri, The Colas, v. 2, pp. 564 and 653.
105 N ilakanta Sastri, The Colas, p. 313; Tamil Lexicon, p. 2281, glosses the word
as ‘standing arm y stationed in the capital of a country’.
106 v . V enkatasubba Ayyer, ‘Kalahasti and its Inscriptions’, Q JM S, v. 16 (1925-
6), p. 22.
107 O ther records o f the 48,000 (narpat ten-nayiravar) are; 238/1959-60, at Palghat,
10th C„ 273/1950-51 and A .R .E ., ¡950-1, p. 3 at Tanjavur and E.I., v. 27, no. 18,
p. 106, 12th Century, South Arcot. O ther num bered units include M iladu 2000
in Tirukkoyilur T aluk, South Arcot, A .R .E ., 1921-22, p. 8.
•°8 F or K arnataka: G opal, op. cit., and A.P. K arm arkar, Cultural History o f
Karnataka (Ancient & Medieval), K arnataka Vidyavaidaka Sangha, D harw ar
1947, pp. 64-5. F o r A ndhra Pradesh: Pala-N adu in Nellore is called a ‘21,000
country' in 22/1956-57; the text of this inscription was earlier published in A. Butter-
worth and V. Venugopaul Chetty, A Collection o f t h e Inscriptions on Copper Plates
and S t O ies in the Nellore District (3 Vols.), Governm ent Press, M adras, 1905, Ongole
of Kulottunga I being found in what is now K arnataka, but which
then contained a basic population of Telugu speakers over whom
Tamil peasant-warriors had extended their control,109 may account
for the rare usage in a Tamil inscription. Further, this usage as
encountered in a Tamil inscription, apart from what it reveals
about nadus, m ay offer a clue to the still vexing problem o f the
numerical designations in other parts of medieval K arnataka as
well as Andhra.
Numerical designations for territories in K arnataka appear to
date from the seventh century and continue to be used until the
early Vijayanagara period. The epigraphist, J.F. Fleet, was among
the first to speculate about the meaning of numerical suffixes
for territorial units, numbers ranging from 750,000 (seven and
one half lakhs) for R attavadi, comprising m ost o f northern K ar­
nataka, to a locality as small as Vavulatalla ‘twelve’.110 He argued
that smaller numbers — tens and hundreds — may have referred
to actual settlements within a particular territory whereas larger
numbers were conventional or ‘traditional’. Lewis Rice, the early
K arnataka epigraphist, drew Fleet’s criticism for his quite realistic
disbelief about such very large numbers of villages assumed in
even some of the smaller numerical suffixes. As to the larger
numbers, Rice suggested that ‘thousands’ referred to nads; thus,
Gangavadi ‘96,000’ would refer to a large tract comprised of
96 nadus or localities and Banavasi ‘12,000’ to a territory o f twelve
divisions.111 M ore recent students of medieval K arnataka have
striven to show that the .smaller numerical designations refer to
villages and have adduced some supporting evidence. However,

Taluk, no. 139, dated a . d . 1218, h i , pp. 1 129-30 showing that the segment referring
to the num bers for the territory were not clear. However, there is no ambiguity
in the recent epigraphical report; a KLannada inscription at W arangal of the Chaluk-
yas o f Kalyana, dated a . d . 1118 refers to the place as A nm akonda-7,000. Other
references to num beied territories in medieval A ndhra may be found in: A ndhra
Pradesh Governm ent Archaeological Series, Kannada Inscriptions o f Andhra
Pradesh, eds. P. Sreenivasachar and P.B. Desai, no. 3, Governm ent o f A ndhra
Pradesh, H yderabad, 1961, no. 73, p. 29; no. 88, p. 34; no. 94, p. 35.
109 Avani, where this K ulottunga inscription was found, is said to be in Andhra-
mandala in an inscription of the 4th C. a . d . : B. Lewis Rice, ‘M udyanur Plates of
Saka 261 of the Bana King M alladeva-N andivarm an’, I.A., v. 15 (June, 1886),
p. 172. It was then and later a well-known sacred site associated with Rama.
no ‘Ancient T erritorial Divisions o f India’, J R A S (1912), pp. 707-10.
m Discussed in Dikshit, op. cit-., pp. 26-7. In his opinion, this explanation was
plausible.
they also propose that the idea of ‘village’ may have been different
from the m odem one, though different in what respects is
unclear.112
A nother possible explanation of the smaller numerical suffixes
attached to place-names in medieval K arnataka and Andhra is
that they refer to the peasant household units, okkalu , of those
who originally conquered or colonized a locality. These conquerors
or colonizers might then have continued to refer to themselves,
as a corporate unit, by some conventional num ber based upon
the historical colonization event. This suggestion would accord
with a parenthetical statement of Dikshit who sought to explain
village assemblies (e.g., ‘Seventy of Kaginele’) as the num ber of
families which originally established the village and corporately
preserved their ascendant rights by the use of a numerical title
even after many more households had come into existence.113
Here, groups of peasant households, okkalu , are given a primacy
which conforms with the importance of the nadu to which Dikshit
gives full recognition. Numerical designations in their lower
ranges would thus be understood as referring to corporate groups
of peasant households within localities, or nadus. The larger
numbers, like Gangavadi ‘96,000’ would refer, as Rice long ago
suggested, to conventional ways of expressing the clustering of
such localities.
The term ‘48,000’ is im portant in the legends of Tondaim anda-
lam. Num erous local traditions refer to the w arrior Adonai
(or Ananda) C akravartin who conquered Tondaim andclam from
its pastoral occupants, the Kurum bars, and brought 48,000
‘selected and good’ families of Vellalas to settle the central and
northern portions of the Tamil plain.114 The inscriptional reference
112 D ikshit, op. cit., p. 28; S. Ritti, T h e Belavola D esa’, Jagadguru Tontadarya
College Miscellany, v. 4, pp. 2-3; B.R. G opal, ‘The L ater W estern Chalukyas (From
the Earliest Times to 1076 a .d . ‘, unpublished Ph. D. Thesis, Mysore University,
1961, especially App. IV, pp. 431-40 and a personal com m unication from Gopal,
19 Feb. 1968.
113 Dikshit, op. cit., p. 75.
114 See B. Ram asw am i N aidu, ‘Rem arks on the Revenue System and Landed
Tenures of the Provinces under the Presidency o f Fort St. G eorge’, J R A S [Com­
m unicated by John Hodgson] (1834), pp. 295ff; a similar version from the Mackenzie
Collection of the late eighteenth century is found in Castes and Tribes o f Southern
India, v. 7, pp. 382-5. Variants of this legend, involving the m igration of Vellalas
from Tondaim andalam to Pandya country may be found in M. Arokiaswami,
 to jayangonda-choia-mandalam 48,000 bhumi (Tamil: pumi ) is
thus significant in relationship to the legendary m igration of
Vellala families who divided the Tondaim andalam territory
(called jayangorida-chdlamandalam during the Chola period) into
48,000 sections.115 Though details of this Tondaim andalam legend
may be a late invention to justify Vellala local authority as the
nattar, o r leading people, o f Tondaim andalam , it is obviously
based upon earlier usage. It was almost certainly Vellalas, nattar
of Tondaim andalam , who controlled the area called Rajendra-
chola ‘eighteen countries’ and who comprised the valangai army
to which the a . d . 1072 inscription makes reference. This asser­
tion is further supported by evidence from the Kdrmaridala-
Satakam .116 F rom several verses o f the Satakam , we learn
th at the K aralar people were related to the ‘G anga people.117
Thus, Gangavadi ‘96,000’, as the tract including rajendra-
chola-padinen-vishaya was usually called during the medieval
period, appears to be the same as karmandala, ‘96 nadus’ of the
Satakam . The com m entator on the Karmandala-Satakam ex­
plained th at the references in verse seven to thirty-two nadus, and
sixty-four nadus m eant that there were 32,000 and 64,000 K aralar
families in these places, or 96,000 for the entire territory.
F rom extant inscriptions o f Tamil country and K arnataka, the
evidence is clear that the nattar represented the dom inant peasant
group or groups of the locality. In num erous records, the nattar
are differentiated from other im portant local groups having interests
in the land and in all cases the nattar have a primacy over these
other groups as the recipients o f the purportedly royal order and
The Early History o f the Vellar Basin, A m udha Nilayan Press, M adras, 1954, p. 36,
and involving the m igration of Vellalas from Tulu country (N orth K anara) to
Tondaim andalam in D.B. R am achandra M udaliar, ‘M udaliar’, Q JM S, v. 10
(1919-20), pp. 289ff. Suresh, H.C .G .E .S., p. 236, notes th at persons, including three
Brahmans use this numerical designation (narpattem ayira) in their names but
fails to recognize this as a reference to their origin, i.e. Tondaim andalam .
115 M onier-W illiam s, A Sanskrit-English Dictionary, ‘bhumi’, p. 763. The division
o f Tondaim andalam into kottam and nadu is attributed to Adonai C akravartin;
see Ramaswam i N aidu, op. cit., pp. 295-6.
116 The Satakam was written by one A raikilar of Avinasi and the com m entary
was prepared by P.A. M uthuthandavaroya Pillay, M adras, 1930, verses 1-7.
117 N ilakanta Sastri, The Colas, p. 507, where he cites in evidence: the Tiruvalan-
gadu Plates, S .I.I., v. 3, the Anbil Plates in E.I., v. 15, S .I.I., v. 3, no. 142, and the
'L arger Leiden Plates’ of E.I., v. 22.
were it executors. In w hat appears to have been a well-established
procedure, the nattar classified and registered lands within the
locality and took prime responsibility for effecting changes in the
distribution o f income from cultivated land.118 Only peasant
groups o f a locality who actually controlled the cultivable land
could have carried out the functions described in these numerous
records. They alone possessed the valuable land whose income
could constitute a gift to the pious and the learned; they alone
possessed the means for m aintaining the full productivity o f these
lands dependent as m ost were upon irrigation works which served
the entire locality; and they alone through their control over
dependent labourers — both skilled artisans and unskilled field
hands who actually carried out field operations — could have
assured th at once granted, the specified income from villages and
lands granted would sustain a flow o f income ‘in perpetuity’.
T hrough these references to the nattar, we perceive a corporate
entity consisting o f those prestigious, satvik peasant families linked
together by their com m on dom inion over the land and reinforced
by m arriage alliances, close social relations, shared religious and
ritual affiliations, and com m on allegiance to locality chiefs who,
in this period, were o f the nattar. These locality social systems of
peasant folk were m aintained by Vellalas throughout the Tamil
region and by the Vokkaligas and K apus in K arnataka and A ndhra
at this time. Locality interests and dominance would invariably,
in the South Indian context, come to reflect themselves as a multipli­
city o f territorially specific subdivisions am ong the nattar and would
be im itated by those lower status groups linked to those enjoying
locality dominance.
In the light o f the obvious im portance o f the nattar, it is interesting
to speculate about the elaborate solemnizing processes described
in m any o f the copperplate inscriptions. Ostensibly these docu­
ments recorded a gift and sought to assure its longevity just as
inscribing its provisions upon a temple wall did. The copperplate
inscriptions record gifts to individual priests or teachers — Hindu,
Buddhist, or Jaina — o r to groups of such persons as recipients;
attention is focused upon the receiver or receivers and that which
is received, and bo th are very elaborately described. The copper­
plates, thus, become instrum ents o f new rights created in land
or some other value. M ost stone inscriptions differ in that they
u s N ilakanta Sastri, The Colas, pp. 467 and 504-5.
record the beneficence o f a donor or donors to the god or priests
of the temple, and the m ajor focus is upon the giver.
The lengthy Sanskrit prasastis of copperplate inscriptions would
seem to belie this distinction between copper and stone epigraphs.
In m ost o f the bilingual — Sanskrit and Tamil — copperplates,
as in most stone inscriptions o f the Chola period, the reigning
overlord and, often, subordinate chiefs o f the neighbourhood are
praised. Special praise is lavished upon the petitioner ( vijñápti)
for the grant.119
The two languages employed in the Chola copperplate inscrip­
tions emphasize two purposes served by these records. The opera­
tive portions o f these inscriptions and others are in Tam il; they
are exact and detailed in describing the lands from which a portion
o f income is being transferred and in identifying the recipients
o f these grants. Sanskrit portions name the Chola overlord and
provide his genealogical and w arrior bona fidés. The Tamil portions
describe in detail the grant and the grantee, nam ing the most
im portant social groups in the locality where the grant is m ade and
providing for their solemn association with and assent in the grant.
While the Sanskrit portions o f the plates constitute the m ost im­
p o rtan t and comprehensive public records o f the day, the Tamil
portions provide a view of the key groups o f the locality in action,
and this, plus the solemn procession, suggest that the grant was a
way o f conferring honour upon such groups.
There is some evidence to suggest that the Sanskrit plates were
executed separately from the Tamil plates and at different times.
R ajaraja I introduced the technique of a standard prasasti which
was affixed to many o f the inscriptions of his time, a practice
followed by his descendents.120 The Sanskrit plates of the Tiru­
valangadu Plates o f R ajendra I were clearly added a decade after
119 See the ‘Larger Leiden’ Plates, Sanskrit and Tam il, 21 plates, 443 lines of
writing, £ ./., v. 22, no. 34; the T iruvalangadu Plates, both languages, 31 plates,
816 lines, S.I.I., v. 3, no. 205; the K arandai Plates, both languages, 55 plates, 2,500
lines of writing but, as yet, not properly edited, see N. Lakshm inarayan Rao, ‘Some
New Facts About Chola H istory’, Journal o f Oriental Research, v. 19 (1950), pp.
209ÍT.
120 N ilakanta Sastri, Sources o f Indian History, p. 69. He discusses some of
the differences between copperplate and stone inscriptions, noting that the latter
began to increase after the sixth century, 'but copper plate records continue to be
the m ainstay of the historian for some centuries m ore, and at no time can they be
left out of the reckoning (ibid., p. 61).
the donative Tamil plates,121 and, in one recently discovered set
o f plates o f the tenth century, only the Tamil language portion,
consisting o f six plates, inscribed on both sides, remains. In this
latter case, the Sanskrit plates were never added to the Tamil ones
which num ber from ‘13’, indicating that this addition was in­
tended.122 This is perhaps the clearest evidence yet discovered
relating to the differences between the two sections o f copperplate
inscriptions.
When, during later times, temples became the principal re­
positories o f public records, the relatively balanced and realistic
association o f peasant groups with the operation o f locality life
reflected in the C hola period became obscure. The prominence
o f stone inscriptions from temples accurately reflect an im portant
shift in institutional power, but, owing to their alm ost exclusive
concern with the affairs of the brahm anical temple, they inaccurate­
ly reflect other aspects o f social and economic life at a later time.
M ore than 550 nadus are m entioned in the inscriptions o f the
Chola period until about a . d . 1300. These cannot o f course be
considered as equivalent units, or segments, anym ore than the 333
taluks o f the m acro region during the twentieth century can be so
regarded.123 There were obvious and great variations among these
basic units o f agrarian organization: in size, population, quality
o f their resource base, isolation, and antiquity. To say a great
deal m ore about these hundreds o f nadus, or as they might otherwise
be called ‘nuclear localities’ or ‘segments’ is to venture well beyond
the evidence o f the Chola period, but such an excursion, so long
as it is understood as tentative, may be useful.
D uring alm ost the entire time o f the great Cholas — from
around a .d . 1000 to 1200— the 550 nadus o f the m acro region were
prim ary loci o f agrarian society. While each of these constituted
an alm ost self-sufficient ethno-agrarian micro region and while
121 S.1.1., V. 3, p. 384.
122 A .R .E ., 1961-2, C.P. no. 29, pp. 4-5; also published by T.N . Subram aniam ,
T.A .S .S .L , 1958-9, pp. 84-110.
123 Based on Subbarayalu, op. cit., App., pp. 98-109. Taluk data for the twentieth
century is as follows: Mysore districts in the divisions of Bangalore (48) and Mysore
(51) are 99 and M adras territory within the m acro region (excluding M alabar and
a few other places) had 234 taluks according to Governm ent of Mysore, D epart­
ment o f Statistics, Mysore State in Maps (1958), p. 3 and G overnm ent of M adras,
Alphabetical List o f Villages in the Taluks and Districts o f the Madras Presidency,
M adras, 1933, pp. iii-v.
most pre-dated the establishment o f the imperial Chola state,
all were linked together as parts o f a great kingdom. W hat linked
them were their common recognition o f the Chola kings; the
imitation by local chiefs of some of the royal styles of these kings,
especially their support of brahm anical institutions — brahmadeyas
and temples devoted to the worship of Vedic gods; and by the
occasional massing o f military resources o f a num ber of nadus
for predatory or defensive warfare. Each nadu was a segmental
part o f a single, unified conception o f Hindu kingship; each con­
stituted a basic bloc o f which the edifice of the realm was composed.
But, there were significant differences among these hundreds of
localities with respect to their internal organization and with respect
to their connections with other localities.
Some nadu localities were elaborately hierarchical. Here there
were ancient ruling lineages, some were vestiges o f an ancient
tradition o f kingship; here there were great establishments of
Brahm ans where the highest forms of brahmanical learning was
cultivated and transm itted and where the highest forms o f ritual
o f this hhakti age were practised; here too were large and wealthy
agricultural and m ercantile settlements. Political, religious, and
economic links am ong such nadus were strong; in the Kaveri, these
linkages am ong its populous localities were strengthened by the
proxim ity o f the royal power and authority o f the Cholas. As a
condition o f existence o f such wealthy and populous nadus was
the reliability o f water for irrigated agriculture. Protection and
regulation of water courses was a m ajor political concern for such
localities, and the certainty and relative abundance o f water meant
that they represented the greatest concentrations o f agricultural
labourers, then as now stigmatized as impure and sequestered in
a paraiyar-ceri, separated from the m ain settlements o f any locality.
Nadus of such strong vertical segmentation, or hierarchy, may be
designated ‘central nadus'.
There were central nadus in alm ost every part o f the macro
region, even in some of the m ore remote peripheral areas. In form,
these resembled the central localities o f the Kaveri core o f the Chola
segmentary state. Here, local society was most hierarchical in its
organization, C hola influences and political authority were m ost
evident, Brahm ans were m ost num erous and accorded high status
and the settlements controlled by their mahasabhas were often
linked together as satellites o f a central brahmadeya, chieftainships
were ritually assimilated to the C hola kingship (e.g. muvendavelar),
and nadus were smaller in size and larger in population as well
as m ore closely linked together as in the case o f the niyayams
o f C hola country or periyanadus elsewhere. Paradigm atically, the
central nadu consisted o f the following ranked elem ents:

Central Nadu

Cholas

Muvendavelar chiefs

_Brahmans _ ^C en tra l place

Brahmadeyas

Niyayam or^_ .D om inant Peasantry (N attar) _^Niyayam or

Periyanadu Periyanadu

Lower Peasantry
and Dependents

A rtisan-Traders

Landless Labourers

Yet another type of nadu was to be found in those tracts o f the

lowlands or in the large interior upland above the plain which
lacked reliable sources o f moisture for regular or extensive irrigated
agriculture. These may be designated as ‘interm ediate nadus'.
In these tracts, agriculture was a m atter of considerable risk and
uncertainty, requiring the utm ost skill and effort by agriculturists.
Successful occupation o f the land here was m ore a m atter o f such
skill and effort than military prowess; and here wealth for the
support o f elaborate royal or religious institutions was seldom
available, resulting in fewer o f the hierarchical elements found
in central nadus. Instead, various agricultural groups, usually
mobile, often migrants from elsewhere, joined older, established
peasant groups in localities which offered the prospect of simple
livelihood as a return on their cultivating skills and their m odest
and mobile agricultural capital o f animals, tools, and seeds. Dif­
ferences am ong such groups would not take the form of rank or
status distinctions, but o f subcaste and clan affiliation among all
o f whom a rough equality was maintained under the rule of a chief­
tain (o r occasionally, a m inor king) of those belonging to the
oldest, established peasant groups of the locality. This horizontal
segmentation am ong subcastes and cians along with the relatively
sparse populations supportable by mixed agricultural and pastoral
utilization of these dry lands made for fewer durable linkages
am ong neighbouring nadus as com pared with central ones. In­
termediate nadus, with their more diverse peasant groupings,
tended to be somewhat more independent, more isolated from
the others; often however, their constituent peasant and pastoral
peoples maintained cult and marriage contacts with their subcaste
and clan brethren in neighbouring nadus.
Nadus or segments of an intermediate character differed from
the hierarchically organized, central nadus in several crucial res­
pects. Each of the central nadus was closely linked to a ruling
dynasty; for most of the Chola period, this was the Chola dynasty
in the Kaveri core region. But, in the intermediate nadus the
sovereignty of other dynasties was recognized, in the Pandya
country, a part of the Chola segmentary state, the Pandyan dynasty
was barely suppressed by Chola power. It reasserted itself vigor­
ously in the thirteenth century. The Pallavas of Tondaim andalam
were remembered long after the demise of the dynasty by the
widely used chief’s title, ‘Pallavarayan’. Interm ediate nadus on the
edges of the Kaveri — the centre of Chola power — were capable
of being detached from allegiance to the Kaveri overlord by other
dynastic powers such as the Hoysalas and Pandyas.
A nother feature o f the interm ediate nadu segments was that
chieftains were less fully incorporated by the C hola kingship.
Here, the basis of authority of chiefs resided in their leadership
over the segmentarily organized, dom inant peasant folk of their
localities. In contrast to central nadus, where political m anage­
ment involved the exalted and separated, dharmic, king-like con­
trol over num erous groupings in complex interaction under power­
ful village representatives of the nattar, in intermediate nadus
more direct political control appeared to vest in local chiefs.
These hereditary chieftains of dom inant, local agricultural groups
exercised not lofty dharmic rule, but the direct, proprietory rule
o f k$atra.124 It was from am ong such chiefs th at the Pandya
dynasty, ‘the Five Pandyas’ m aintained a hovering presence
through much of the eleventh and twelfth centuries and re-emerged
in the thirteenth century; it was similar chiefs, called pattakkarar,
who controlled much of Kongu at the same time.
The num ber of Brahmans in interm ediate nadus was possibly
not insignificant, but the institutions with which Brahmans were
associated in the central ones were far fewer. Evidence of the
Chola period suggests that not all Brahmans, nor even most,
were members o f villages controlled by Mahasabhas, n o r were
they members of Vedic temple establishments protected and
patronized by the mighty. The territorialization of Brahman sub­
castes reported during the eighteenth and nineteenth centuries
and the status differences between the m ajor divisions of sastnc
Brahman castes and the vaidika or temple Brahmans may be
taken as evidence that most of them lived among peasant villagers.
Here they enjoyed special privileges deriving from their priestly
functions, but they lacked the visible institutional and sacral
prestige of sastn c Brahmans in the central nadu segments. Brah­
mans in the intermediate segments would thus have been closely
linked to ranked chiefs who, with their peasant clientele, they
served ritually.
Still another feature of the intermediate nadu segment was the
shallow hierarchy of relationships as com pared to that of the
central nadu; characteristically, horizontal rather than vertical
segmentation was characteristic of the former. W ithin each inter­
mediate nadu segment, villages might be dom inated by one or
another o f the several blocs o f peasant peoples of the m acro region
(e.g., choliyar, kongu); each nadu segment would be dominated
by one of these blocs. Within Pandim andalam and T ondaim an­
dalam, there appear to have been several blocs of castes whose
sub-units were dom inant in some o f the localities o f the zone
while in other localities they m ight be subordinate to another
zonal bloc sub-unit. Those dom inant in a nadu segment consisted
o f th at peasant group with numerical strength, with possession
o f share (parigu) rights over num erous parcels of land in the locality,
i24 See the discussion of Robert Lingat on the concepts of dharma and ksatra
in his The Classical Law o f India, University o f California Press, Berkeley, 1973,
pp. 211-13.
 and with close status and social relationships with basic service
groups, including artisans and merchants involved in the agricul­
tural economy. The lpahja j â t î (five castes) of Kongu in present
day Kongu exhibits the characteristics of caste dominance referred
to here.125
The tendency for the intermediate nadu segments to be dom inated
by peasants and service groups sharing a nearly common status,
identity, and a common zonal caste name produced two apparent
structural features which further distinguished the intermediate
from the central nadu. While the evidence on this is sketchy,
mobile artisan-traders — kam m alars or panchayattars — appear
usually to have been separated from this core population of the
locality and enjoying higher status than the same groups in the
central nadu segments. In the twelfth century, artisan-traders in
Kongu and elsewhere appeared to have striven for and attained
status parity with the local peasant folk among whom they lived.
Prior to th at time, and later, m obile artisan-traders maintained
linkages with others like themselves in other places and constituted
zonal alliances o f the idangai. Notwithstanding the greater diffi­
culty of intermediate nadus achieving the massing capability of
central ones as a consequence of their greater horizontal segmenta­
tion and the absense o f the proxim ate mobilizing activities of
kings, some of this did occur as we have seen. It was precisely
such a massing of dom inant peasant groups which took the epithet
of ‘right-hand castes’ as well as designations like the ‘48,000’ of
jayankondacholamandalam (Tondaim andalam ) in the M ulbagal
epigraphy o f a .d . 1072.
Diagramatically, the intermediate nadu segment may be
represented as on the next page.
A third type o f nadu may be distinguished. In those parts of
the m acro region least hospitable to sedentary agriculature or
even to mixed agricultural and pastoral activities were scattered
nadu localities in which neither vertical segmentation, or hierarchy,
nor elaborate horizontal segmentation prevailed. These may be
designated as peripheral nadus.
Nadus of the peripheral type in the Chola kingdom displayed
the strongest ‘tribal’ characteristics. That is, most of the people
125 Arokiaswami, Kongu Country, p. 267 and Thurston, op. cit., v. 3, p. 419. The
five castes comprise Vellalas, Chettis, barbers, washermen, and Paraiyar.
 Peasant M icro Regions: The N adu 139

Intermediate Nadu

Cholas
or
e.g. Pandya Zonal dynasty

Chief -> Zonal allies

Brahmans

Peasantry unit
Valangai o f zonal bloc A rtisan-traders 1 —>!dangai

Low Castes Low Castes

of these localities shared a single common ethnic identity, usually

a section of some larger regional group such as M aravar or Reddi.
Isolation and the hazardous basis of livelihood in the peripheral
tracts were the m ajor reasons for this tribal feature. It is doubtful
that the peoples o f the peripheral tracts were racially distinctive
in any sense whatever during the medieval period, if they had
ever been. In any case, they were not on this or any other account
excluded from contem porary dharmic society. In fact, as already
pointed out, M aravars and Kallars o f the southern peninsula
appear to have been acceptable sections of the Tamil-speaking
peoples in the early centuries o f the Christian era as this society
is revealed in Classical poetry. Only in British times did the pejora­
tive designation ‘criminal tribes’ come into existence.
While evidence of peripheral nadus is for the m ost part inferential
for the medieval period, it would appear that there was regular
contact between those of the peripheral nadus and more prosperous
peasant peoples contiguous to them. Relations with itinerant
m erchants and with colonies of plains m igrants were im portant
forms of contact ; participation in military expeditions with plains
people may have been another im portant form of contact. Such
interactions, however interm ittent, secured a place for the isolated
folk of the peripherai areas within the larger society of the macro
region; contact also provided a model of a way of life followed
by the peasant folk in neighbouring areas for at least the more
prosperous of the peripheral people to adopt. The latter, along
with genuine colonists from the intermediate nadus , created a
two-part society in the peripheral areas. D om ination was that of
the local section of the major ethnic group (e.g. M aravar), but
there would be a subordinate part approxim ating a simplified
organic model. In contrast to the pattern of intermediate nadu
segments, however, localities with the characteristics of central
ones were rare and dynastic affiliations were very weak, though
in common with most intermediate nadus the peripheral tracts
were ruled by strong chiefs and lateral linkages among locality
groups were im portant.
 The Coromandel Brahmadeya
Village

The sacerdotal élite of India has been a part of rural society from
an early time. Even after the emergence of ancient and medieval
urban places, Brahmans retained a ‘rustic’ character to which
both the diversity and flexibility of Indian civilization may, in part,
be attributed.1 However, the quality of rural life which South
Indian Brahmans m aintained over centuries was exceptional in
one crucial respect. That is, the character of the villages in which
many o f them resided was quite special, and the degree of secular
authority which they were able to exercise was very considerable,
in their own settlements, and in many other settlements dependent
upon Brahm an villages through their powerful spokesmen, the
‘great men’ of the sabha. In no other part of the sub-continent was
such a measure of influence achieved and m aintained by Brahmans
as in the villages and the localities of the Corom andel plain.2 The
Corom andel brahmadeya villages were unique centres of civilization
whose culture moulded that o f the peasantry around them ; they
were also thriving centres of agrarian activity.
There can be little doubt that the kinds of Brahman villages
1 M ore than any other historian o f India, D .D . Kosam bi stressed this fact in
his various works: see especially; "The Basis of Ancient Indian H istory’, Journal
o f the American Oriental Society, pt 2, v. 75, no.4 (1955), pp. 235-6; The Culture
and Civilization o f Ancient India in Historical Perspective, Routledge & Kegan Paul,
London, 1965, pp. 175-6. In general, however, the nature of urban places in early
India is poorly understood and deserves more serious attention than is provided
in works such as: B.N. Puri, C ities o f Ancient India, Meenakshi Prakashan, Meerut,
1966.
2 See A.S. Altekar, A History o f Village Communities in Western India, O.U .P.,
M adras, 1927, pp. 26, 123-4. A ltekar not only explicitly com pared the village types
of western and southern India but also criticizes some of the scholarship on this
issue which combines evidence from widely different times and places, especially
that of R.C. M ajum dar, Corporate Life in Ancient India, Calcutta, 1922 and R.K.
Mukerji, Local Government in Ancient India, Motilal Banarsidass, Delhi. 1958.
which dotted the Corom andel plain during the C hola period were
relatively scarce elsewhere in the macro region. From extant eviden­
ce, there were few Brahm an villages, governed by a sabha of the
learned and capable of the full management of their own affairs,
beyond the fertile, well-watered plain. W anting in most parts of
the m acro region beyond the plain were requisite peace and security,
and ecotypic conditions for agriculture prosperous enough to
support large populations o f Brahmans. W ithout such conditions
the Coromandel type o f Brahman village could not exist. Even
where such conditions could and did exist in the interior uplands
of the m acro region, there was not a dom inant peasantry which
regarded this institution as desirable or necessary as did the peasan­
try of the plain. However, if the Brahman villages themselves
could not be established and m aintained in the macro region beyond
the plain, the culture of these centres of civilization certainly did
reach these more remote areas. The mathas and other centres of
learning in the Corom andel brahmadeyas served as disseminating
points of a high and distinctive culture not only to young Brahmans
of various parts of the plain, but to those who came, often from
great distances, to study and to carry back to their home territories
ritual and theological elements which shaped the high culture in
the temples and mathas of the interior cultural hinterland. It must
be supposed th at other ideas were also carried in the same way.
Hence, while the great brahmadeya . villages, the subject of this
section, are appropriately identified with the plain, they must be
treated as distinctive institutions of the macro region as a whole.
How Brahmans were able to establish and m aintain these unique
rural centres of civilization has been suggested above in the dis­
cussion of Brahmans and peasants during and after the Pallava
period. The complementarity of social, political, and ideological
objectives of the Corom andel Brahmans and the dom inant peasan­
try of that region was responsible for this accomplishment; the
persistence of reciprocal advantages in the relationship between
Brahmans and peasants assured the durability of these settlements.
Basic changes, when they did occur after the twelfth century, were
only partly a consequence of the development of divergent aims
between peasants and Brahmans and from pressures in society
beyond this relationship. Forces were actually generated within
the Brahman villages and the Brahman world which contributed
significantly to the decline of these villages.
 However unique and im portant Brahm an villages as centres of
civilization and agriculture may be shown to be, their place and
functions can be, and have been, exaggerated, and the role of the
village as a unit of society, politics, and economy correspondingly
distorted. M any of the inscriptions which we possess deal with
brahmadeya villages; a few refer to the locality with its numerous
settlements of non-Brahman cultivators. From the point of view
of agrarian relations, the latter were obviously more im portant.
Because Brahm an villages were well-defined, well-organized, and
highly visible in the historical record of the time, the assumption
has been that peasant villages were but slightly different versions
o f the brahmadeya villages. This view has been strengthened by
the m anner in which the British dealt with the village unit for
revenue purposes on the assumption, or with the justification, that
they were adapting to the usage of ‘time immemorial’, when
actually it suited British political requirements and their agrarian
preconceptions.3 Historians have not adequately recognized that
the atomic village of recent times reflected im portant agrarian
changes after the Chola period; too much emphasis has been
placed upon village units in a period when this was not a primary
unit of social and agrarian organization, which it was to some
extent later to become.
Seen in the context of the nadu, or the peasant micro region,
Brahmans were one of several kinds of social groups dist inguishable
by the fact of their significantly higher status than even the most
powerful peasant people and their ritual and learned functions
which set them apart from others. As noted in the Chola copper­
plate inscriptions, particularly, Brahmans in brahmadeya villages
shared a common position in relation to the nattar along with
others engaged in sacral functions such as certain Vedic, Jaina,
and even Buddhist functionaries and teachers. M erchants and
artisans involved in commodity production and exchange outside
of the immediate local peasant economy were also accorded a
special, autonom ous position with respect to the dom inant peasant
people of a locality with whom substantial political authority and
economic power vested. However, Brahmans were distinctive in
W alter C. Neale, ‘Land Is To Rule’, in R obert E. Frykenberg (ed.). Land Con­
trol and Social Structure in Indian History, The University o f W isconsin Press,
Madison, 1969, pp. 3-17; Louis D um ont, ‘The Village Com munity from M unro
to M aine’, Contributions to Indian Sociology, v. 9 (Decem ber 1966), pp. 67 ff.
that they alone maintained village establishments governed, by-
assemblies of their most learned members, and these settlements
were further distinctive in being the most im portant religious and
educational centres of the time. Though it is not to be presumed
that all — or indeed most — Brahmans lived in these sabha-
governed villages but, rather, that many Brahmans lived in villages
attached to temples (devadána ) or in ordinary peasant villages,
and though it is also clear that there were many non-Brahmans
who lived within the large brahmadeya villages, still the latter were
among the most im portant cultural institutions of the time.
in the Corom andel lowland by the ninth century, and in other
parts o f the m acro region later, control over cultivable land and its
wealth was held either by peasants comprising the nattar or by
Brahmans. Distinctions between the two were carefully maintained
as suggested by the terms ürar and veUan-vagai, for the former, and
sabha and brahmadeya, for the latter. Urars and sabhas were
assemblies composed of representatives of the two basic types
of agrarian settlements within a nadu. The terms vellan-vagai
and brahmadeya refer to categories of major recipients of agricul­
tural incom e: veUan vagai, a sharing of agricultural income among
cultivating peoples, peasant families comprising the various urar
of the nadu ;4 brahmadeya, a sharing of income among Brahman
families under the authority of the sabha. Appropriately, the
control of the urar, or urgahlar as it was also called, comprehended
all lands not under the authority of the sabha, including Hindu
temple land (devadana ), those lands from which income supported
Jaina shrines (paUic-candan ), and Vedic sacrifices conducted
outside of temples or ‘public works’ ( vettapperu ).5 In all of these
cases, as well as that involving arable lands around trade centres
under some degree of m erchant nagaratiar control, the dom inant
peasantry of the locality distributed the proceeds of cultivation
according to the deeds of gift specified in inscriptions, or according
to custom ary payments and dues to groups associated with agrarian
production from the lowliest labourer to the most skilled craftsman
and the locality overlord.
4 N ilakanta Sastri, The Colas, pp. 570-80; Appadorai, op. cit., v. 1. pp. J52-4;
vellan-vakai, according to the Tamil Lexicon, v. 6, is ‘that which belongs to the
vélala class, as lands in a village’, p. 3793, citing a South Indian inscription.
5 E.I., v. 22, no.34, ‘Larger Leiden Plates’, p. 231; the m eaning of vettapperu
is doubtful.
 Carefully and ceremonially dem arcated from the most fertile
and productive lands were the fields under the authority of the
Brahm an sabha. Within a locality, or nadu, there may have been
several Brahm an villages, some very large, others smaller. These
were easily identified by their names and references to the sabha
and often to the fact that the village was called a brahmadeya.
Brahm an villages of this period were referred to by a variety of
terms beside brahm adeya: chaturvedimahgalam, mahgalam, agaram,
agraharam. agra-brahmadeya, agra-brahmadesa , brahmadesam,
brahmapuri, and brahma-mahgalam.6 O f these, only agraharam
seems to have conveyed a substantially different m eaning within the
general meaning o f brahmadeya, ‘a donation to Brahm ans’. A g­
raharam appears to have been m ore often used in inscriptions to
designate a set of privileges held by Brahmans living in villages
over which they did not enjoy the same dominance as the Brahmans
of brahmadeyas.1 The usage is m ore common in the inscriptions
o f K arnataka and, in general, during the period after the thirteenth
century when the im portance o f brahmadeyas was much reduced
and therefore may reflect only a change in the circumstances of
landed Brahmans under the altered agrarian conditions of a later
period.8 An alternative means o f identifying Brahm an villages
was by reference to assemblies through which these great villages
were governed. Thus, inscriptions refer to the sabha and maha-
sabha, or use the equivalent Tamil terms, kuri and perunguri, or
perungurimakkal , ‘the great men o f the assembly’.9
Brahm an assemblies with their functional committees, or
offices ( variyams ), through which control and supervision were
6 A ppadorai, op. cit., v. 1, p. 140; S.I.T .I., v. 3, pt 2, ‘Glossary’.
7 Monier-W illiams, Sanskrit-English Dictionary, p. 6; A ppadorai, op. cit., v. 1,
p. 158n, cites F.W . Ellis (Replies to Seventeen Questions Proposed by the Government
o f Fort George Relative to M irasi Right; with Two Appendices, G overnm ent Gazette
Office, M adras, 1818, p. xxv, n.8, to the effect th at agra m eans ‘before’ and harati
‘it is taken’ which is understood to be the grant o f village income to Brahm ans which
would otherwise constitute a revenue payment. However, Ellis used agrahara to
refer to villages which paid a full revenue as well as those which paid part or no
'inam Relative to melvaram, where m el — before any other or first, see n.75 below.
8 On agraharam of K arnataka see G.S. D ikshit, Local Self-Government in Medieval
Karnataka. K arnatak University, D harw ar, 1964, chs. 5 and 6, where the contrast
with Corom andel Brahm an villages, though not explicit, is striking; also, K.R.
Basava Raja, ‘A graharas in Medieval K arn atak a’, Journal o f the Karnatak Uni­
versity (H um anities), v. 2 (1970), pp. 106-14.
9 N ilakanta Sastri, Cola Studies, p. 82n; A ppadorai, op. cit., v. 1, p. 138.
exercised, do not appear to have existed before the Pallava period,
though Brahmans living am ong peasants in their villages and upon
a share of peasant production dates from an early tim e.10 Not
until the eighth century are there records which refer to the means
o f selection and scope o f functions of the Brahman assemblies.
A Pandyan inscription o f a . d . 782 from M ananilanallur (m odem
M anur near Tirunelveli), a brahmadeya in Kalakkudi-nadu, may
be among the very first o f those records which anticipate the
elaborate organization described more fully in the U ttaram em r
inscriptions of Chingleput district.11 The purpose of the brahmadeya
was to provide a reliable source of support to Brahmans for the
pursuit of their sacral responsibilities, and the gift (Sanskrit:
dana or deya) of arable land, part of the proceeds from which
constituted a stream of income to learned Brahmans, was one of the
m ajor sources of merit to pious H indus.12 In order to achieve this
objective, reliable and substantial wealth had to exist to be placed
at the disposal of Brahm an assemblies. Merely providing for the
subsistence of the Brahman populations o f some of the great
Brahm an villages assumed very advanced productive agriculture.
The population o f a place like U ttaram erur during the tenth
century, according to the Parantaka I inscriptions of a . d . 919 and
922, m ust have been very large if the elaborate rules for selecting
holders o f the village offices ( vdriyam ) were even partially followed.
These rules provided for the selection of forty-two members of the
mahasabha to serve on five committees, twelve on the annual and
garden committees and six on the tank, assessment,13 and gold
committees. Each of the mahajanas , or members of the great
10 M inakshi, op. cit., p. 121; Sathianathaier, op. cit., pp. 34-5; and K.M . G upta,
The Land System in South India Between c. 800 A.D. and 1200 A .D ., Motilal Banar-
sidass, Lahore, 1933, pp. 30-1; N. Subrahm aniam . Sahgam Polity, pp. 255 ff.
11 N ilakanta Sastri, Cola Studies, pp. 81-2.
12 P.V. Kane, History o f Dharmasdstra (Ancient and Medieval Religious and
Civil Law ), B handarkar Oriental Research Institute, Poona, 1941, v. 2., pt 2, ch.
25; A ppadorai, op. cit., v. 1, p. 156; and note the term dana khanda, in E.I.. v. 14,
p. 97, a . d . 1369, referring to the merit of constructing tanks.
13 Some of the term s used in connection with committees are unclear; for example,
pahcavara-variyam which is interpreted by N ilakanta Sastri to be related to the
function of assessment (Cola Studies, p. 142-3) and by Sathianathaier as ‘Standing
com m ittee’ (op. cit.. p. 33; S .I.T .L . v. 3, pi 2) ‘G lossary’, takes both positions it
appears (pp. viii-xlii) while K.V. Subrahm anya Aiyer (E.I., v. 23, p. 22) favours
the first.
assembly, selected for membership o f the committees had to meet
rigorous qualifications and be free of specified disabilities. The
qualifications were those of age (between 37 and 70 years), Vedic
learning and teaching experience, and m inimum property ex­
pressed as a share of the lands possessed by the Brahmans of the
place; disabilities involved having served on the sabha in the
previous three years, being shown derelict in the execution of
previous offices (a disability which extended to other agnatic and
affinal kin as well), having com m itted the sin of incest or other
similarly serious transgressions as theft, consuming polluted food
without having undergone purificatory rites, and having been
adjudged an ‘enemy o f the village (grdm akantakd)\ Considering
these restrictions and the rotational rules on balloting, a village
like U ttaram erur would have had a large population in order to
provide anything approaching an adequate panel o f candidates.
Yet, this village was not great in size; it m easured less than five-
eighths of a mile from east-to-west and one-half mile from north-
to-south during the tenth century.14 In 1932, N ilakanta Sastri
reported a population o f 11,000; its population in 1961 was 13,
879.15
Distinctions am ong Brahm an villages m ust be recognized
notwithstanding the fragm entary knowledge which exists about
them. It is certainly clear that all Brahm an villages in South India
were not like U ttaram erur, by which Brahm an villages have been
measured, not only in South India, but elsewhere in India.16 There
were, first, other settlements like U ttaram erur which m aintained
a set of central place functions, in which authority was most ela­
borately organized among Brahm an residents, and in which the
most persistent relations with the great overlords of the region
— the Cholas and their allies — obtained. However, the extant
evidence suggests that Brahm an villages o f the Chola period ranged
in size from those which may have been the largest settlements on
the plain to quite small villages, just as they ranged in social com ­
14 Based upon the locations o f the m ajor temples referred to by contem porary
inscriptions in the m odem village, see N ilakanta Sastri, Cola Studies, pp. 98-9. These
are m entioned in S.I.I., v. 6, pp. 273-325, 238-335, 336-8, 339, 340, 341-3, 376-7;
Subrahm anya Aiyer, Historical Sketches . . . . v. 2, p. 215.
15 Census o f 1961, IX, Madras District Census H andbook; Chingleput, v. 2, p. 305;
the U ttaram erur Panchayat U nion population was 92, 773, p. 307.
16 Note particularly this assum ption in Dikshit, op. cit., p. 98.
position from those with a large proportion of Brahman families
to those with a few. The residence site of a Brahman village might
comprise over two square miles, as in the case o f U ttaram erur,
which was subdivided into twelve sections (cm's) and thirty smaller
divisions called kudumbus.11 A nother internal division of this
and other im portant Brahm an settlements may have been ac­
cording to sect or to occupational grouping; it was called sanka-
rcippâdi, and may have included only non-Brahm an residents of
these villages.18
Size reflected the location and the circumstances under which
the gift o f village income was conferred upon a group o f Brahman
families. Those extensive tracts with relatively well-developed,
reliable irrigation facilities, as in deltaic portions of the plain, or
similarly large areas whose surface configuration and drainage
characteristics made possible large, perennial tanks, provided the
condition and the requirement for large peasant villages. Isolated
parts o f the rest o f the macro region provided the same possibilities,
17 This estim ate of the area is based upon the location of temple m arkers (n. n o .14).
Cëri names in U ttaram erur appear to be based upon twelve names o f Vishnu thus
suggesting the clustering of Brahm an families according to their ritual functions
(Subrahm anya Aiyer, Historical Sketches . . . , v. 2, p. 273; Nilakanta Sastri. Cola
Studies, p. 103n). In another case of the eleventh century, A lam bakkam , alias
M adhurantakam -chaturvedim arigalam , about twelve miles south-east of U ttara­
m erur, cëris seem to be named for other Brahm an villages, suggesting colonization
from places, or they were nam ed for previous Chola rulers using their surnames
as many Brahm an villages did (A .R .E ., 1910, para. 25. referring to 726/1909). An
inscription of K ulottunga I seems to support the latter theory of cëri names based
upon royal titles (N ilakanta Sastri, The Colas, v. 2, pt 2, p. 560). The m eaning of
kudurnhu is unclear. Usually translated as 'w ard’ as by N ilakanta Sastri (The Colas,
pp. 496, 502, 517n.), the term is obviously better understood as a reference to
lineage’ or, as used by Subrahm anya Aiyer, ‘family groups’ in Historical Sketches . . .
(v. 2, pp. 273-86 and passim) ; also see kutumpi in Tamil Lexicon, v. 2, p. 974. Kudumbu
was also used am ong peasants to constitute what appears to be work units; see
below, ch. 6, n. n o .24.
18 At Tiruvalangadu, in a . d . 1072, an inscription refers to a portion of that large
temple centre called the sankarappadi and describes the shifting of twenty-five
families of oil-m ongers from there to another part o f the settlement and the require­
ment that they provide oil for fifteen temple lamps (14/1896, in N ilakanta Sastri,
The Colas, v. 2, p. 272n.). Still other references to this residential grouping come
from U daiyargudi (C hidam baram taluk) which refers to the sankarappadi of Kadam -
bur (550/1920, S .L I., v. 13, no.58, pp. 26-7) and Sivapuram (Sriperum budur taluk.
Chingleput) (A .R .E ., 1960, no. 284, a . d . 1030; also see François G ros and R. Naga-
swamy, Uttaramerur : Legendes, Histoire, Monuments. Institut Français d ’Indoiogie,
Pondicherry, 1970, p. 98.
but it was in the coastal plain portion of the macro region, the areas
of the m ost ancient peasant settlement, that m ost Brahm an villages,
large and small, were established. O f the smaller Brahm an settle­
ments less is known o f course, for it is from inscriptions alone that
anything is known o f any Brahm an villages, and the num ber of
inscriptions is clearly correlated with the size and importance of a
place. Thus, most o f the Brahm an villages o f the macro region have
left an insufficient record upon which to estimate even relative size.
Whereas in a place like U ttaram erur, procedures for the selection
of members for service oil one o f the functional committees o f the
mahasabha are explicit, the presum ption o f a relatively large
Brahman population is unavoidable. Similarly, where a village
was granted as brahmadeya to a large num ber o f Brahm an families
— even if, as is often the case, the num ber was a conventional one
like ‘108’ families — it was only a m atter o f time before the Brahm an
population would have become substantial.19 M any extant records
refer to brahmadeya grants to smaller groups of Brahmans, including
those to single persons (ékabhógam ).20 in the latter case, the des­
cendants of the original donee would have constituted a m inor
fraction of the population o f a village, the m ajor share of whose
income they enjoyed, and probably lived on a single street much as
Brahmans o f the agraharam section o f m odern South Indian
villages do.
The proportion o f Brahm ans o f the macro region who did not
live in these villages is among the im portant issues in the social
history of early South India which have never been raised. It is
certainly significant to know — or at least to speculate about —
the extent to which Brahmans resided and were part o f the society
outside o f B rahm an villages. T hat m ost did seems beyond question.
19 Srivillipputtur, a Brahm an village in the northern part o f m odern Tirunelveli
was established as a brahmadeya for 108 Brahmans in the late eighth or early ninth
century (T.K.T. Viraraghavacharya, ‘The Srivilliputtur Temple of Sudikkodutta
N achchiyar’, Tirupati-Tirumaiai Devasthanam Bulletin, v. 6, no.3, pp. 1-2). This
num ber is very comm on, though there are other multiples o f twelve which are often
encountered as well; e.g. the village of K orraparru in the K istna delta, established
as a brahmadeya in the middle of the eighth century, certainly one of the earliest,
was granted to twenty-four Brahmans (S.L I., v. 1, n o .35, pp. 31-6). See ch. 6, n.
no.9.
20 S.L T .L , ‘G lossary’, p. xiv; grants to individual Brahm ans who might later
redistribute shares o f income among other Brahm ans seems to have occurred more
frequently, indeed characteristically, in K arnataka than in the Corom andel plain
(Dikshit, op. cit., pp., 100-1).
Villages under the control o f a Brahman assembly were bein g created
throughout the C hola period; small or large groups o f Brahm an
families came into new residential quarters o f established and
prosperous settlements from other peasant settlements near and
far, constituted themselves as a sabha, and corporately managed
the landed wealth which had been bestowed. It appears reasonable
to suggest that many more Brahmans continued to live in non-
Brahman villages as indeed they had in the period before Pallava
times when the establishment of such settlements gained great
prominence. If this supposition is correct, it cannot be argued
that life in the great Brahm an villages, such as U ttaram erur,
constituted the ordinary way of life of the sacerdotal élite of the
macro region in this age. On the contrary, it must be supposed
that, as m ost Brahm ans lived in prosperous peasant villages in the
way they had in the past, perhaps in a special quarter in the vicinity
o f a Vedic temple, their life-styles would have been close to those
o f the peasants and artisans am ong whom they lived. Life-styles
o f B rahm an and peasant, under the circumstances o f long-termed
residential propinquity as well as the tendency in this period of the
latter to emulate the ways of the Brahmans, cannot have varied
substantially. The implication of this line o f speculation is that
it would be false to exaggerate the gap between Brahm ans and
those peasant groups with whom they shared a common rural
social context and culture.
M ap IV-1 shows a distribution o f some 300 Brahm an villages
o f the C hola period. It cannot be claimed to be a complete re­
presentation o f B rahm an villages of the period for new ones come
to light with new publications of South Indian inscriptions, and all
of them m ay never be known. Beyond that, many o f the known
Brahm an villages o f the period cannot be located with enough
certainty to m ake their notice useful, and others cannot be located
at all. Still others which could have been plotted were not because
they were parts o f clusters o f such villages, discernable only on
large-scale m aps.21 In M ap IV-1, it will be seen that the distribution
o f Brahm an villages conforms to the ecotypic conditions most
21 As with all other distributions of medieval South India, the accidents of pre­
servation and discovery o f brahmadeya village inscriptions determines the extent
of possible identification and location. Lacking either local or supra local listings
of Brahm an villages at particular times and places, the universe of such places which
can be positively identified and dated must be considered as partial. In the famous
IV-1 B rahm adeya Distribution c. a .D . 1300
favourable to Corom andel agriculture. The most significant con­
centrations occur in the riverine tracts o f the Kaveri and its tri­
butaries, the Ponnaiyar, the western Tam braparni, and the Palar.
Most o f the Brahman villages lie below the 250 foot contour, the
exceptions being only those which follow the major river basins
through the graded descent from higher elevations to the sea-
level plain below. Very few Brahm an villages are found at higher
elevations where reliable irrigation sources, whether riverine or
tank, could rarely be developed. Even considering the partial
nature o f the distribution shown, the m ap points to significant
and predictable aspects o f the relationship o f Brahm an villages
to the agrarian structure. One of these is that the Brahm an village
is a reliable and accurate ‘m arker’ o f the most m ature agrarian
localities in the m acro region. The high consum ption requirements
o f these villages were supportable only under the m ost advanced
conditions o f agriculture. M oreover, these same localities, it must
be presumed, bore the highest burden o f tribute payments to local
and great overlords. The second fact dem onstrated by M ap IV-1
is that m ost of the nadus, which can be located had within their
boundaries, and usually at their core, one or more B rahm an villages.
This is evident from records o f brahmadeya s of the period in which
the nam e of the nadu in which the Brahman village existed is almost
invariably given. N. Karashim a, in his work on the Chola ins­
criptions, has suggested that during the period o f R ajaraja I,
m ost o f the nadus o f Cholam andalam contained two or three
Brahm an villages; in some there were four or m ore.22 The relation­
ship between the nadu, the prim ary unit of agrarian organization,
and these prestigious Brahman settlements will be discussed below.
Previous research on the Brahman villages o f the Pailava and
Chola periods has treated each as an independent, self-contained
unit. The effect o f this has been to homogenize ail villages in which
the sabha existed. It has been suggested above that a more realistic
perception o f Brahm an villages is that they varied from the most
modest, hazardously supported, village settlements, under the

T anjavur inscriptions of Rajarajesvara temple, of the tw enty-ninth year of Rajaraja

i, a great m any villages are listed to supply functionaries for the temple (e.g.,
v, 2, no. 69, pp. 312 ff); this is exceptional in the extent of the m any villages m en­
tioned, but the total num ber of locality villages are not.
22 ‘The Power Structure of the Cola R ule', Paper presented to the ii international
Conference-Sem inar of Tamil Studies, M adras, 5 January 1968, p. 3 and app.
control of a sabha, to a class of great, populous, and wealthy Brah­
m an-dom inated enclaves within a micro region, of which U ttara­
m erur was a conspicuous member. O f this latter class, it would
appear that perhaps thirty-five constituted premier settlements
by reason o f their size and the impressive array o f central place
functions.
‘Central place’ Brahman settlements were sometimes designated
as taniyür, literally ‘separate settlem ent’, or, as scholars have at
times suggested, ‘free-' or ‘independent-settlement’,23 Among such
places are the following: M adhurantakam , Tirum ukkudal and
U ttaram erur in modern Chingleput district; Tribhuvani in Pondi­
cherry ; and Chidam baram and Ennayiran in modern South Arcot.
Taniyür settlements are often referred to as constituting a m inor
and separate division o f the nadu in which they were: tanküru.24
A more reliable measure o f the special status o f the central-
place Brahman settlement than such designations, however, was
the various functions which many of them carried out. Among
these, the most conspicuous and, perhaps, im portant were religious
and ritual functions. The privileges enjoyed by Brahman families
in these villages were derived from sacral activities for -which they
alone could be responsible. From the earliest brahmadeya ins­
criptions, there appear to have been a regular set of ritual functions
for which learned Brahm ans were responsible and for which sup­
port was given. These included: adhyayanam or archanâ, the
recitation o f Vedas ; bharati or bhciratavritti, recitation o f the Maha-
bharata; panchangam, calendrical activities to provide the aus­
picious times for marriages, festivals, and ploughing; purânam ,
recitation o f purdna ; in addition there were paym ents for teaching
the Vedas to other Brahmans, vëdavritti; and payments to sup­
port learned Brahmans, bhattavritti.25 Most of these activities
were carried o u t in temples, and all such villages had at least one
temple to provide the locus for these and other ritual activities.
During the tenth century, most temples were identified in the
inscriptions simply as ‘the mahddëva temple’ of some particular
25 Appadorai, op. cit., v. 1, pp. 150-1; 'G lossary’, p. ixiii.
24 U ttaram erur was called a taniyür and tan-küru in the famous a . d . 921 inscrip­
tion of Parantaka I (N ilakanta Sastri, Cola Studies, pp. 99, 171). Tirukkalukunram
{Chingleput taluk) is also designated as tan-küru: E.I., v. 3, no.38B, dated a . d . 919.
25 Ibid., pp. 118, 125-7; S. Sundararaja Iyengar, Land Tenures in the Madras
Presidency. Commercial Press, M adras, 1921, p. 213.
village, reflecting ihe greater prom inence o f Saivite religious
activities over Vaishnavite at the time. After the tenth century,
however, many Brahm an villages had several temples, most of
which continued to be Saivite shrines and all under the management
o f the s a b h a which supervised and supported the educational and
ritual activities o f each. By the twelfth century, most of the temples
had come to be managed by special bodies, separate from the
s a b h a , and fully capable of receiving endowments of money and
land for the m aintenance of the shrine. This independence was to
weaken the Brahm an assembly after the twelfth century.
Special hagiographical im portance is attached to m ost of the
great Brahman settlements as ancient, im portant Saivite places,
celebrated in the lines of the devotional hymns of the náyanárs
and often prom inent in the lives o f the sect’s saints. Examples of
these are: Takkolam , Tiruvallam, Alangudi, Tillaisthanam , Tiruk-
kadaiyur, Tiruvidaim arudur, Kiipaiavur, Tiruppalatturai, N allur.26
Others were am ong the most sacred places o f the Vaishnavas, the
‘108 sacred places (nutteitu tirupati)’ : Tirrukkoyiiur, Kum ba-
konam , Tirunaraiyur, Tiruchchirai, Tirum ayam, Shiyali, Kovil-
adi, Kandiyür, Adanur, T irukkurungudi, Srivaikuntam, Sri-
villiputtür, Kilanbil, Uraiyur, Tiruvellarai.27 Some o f these places,
like Tirukkoyilur and K um bakonam , were equally sacred to both
of the m ajor sects, and in all of them were one or m ore temples
which were rated am ong the m ost im portant in the m acro region.28
In fact, it was only in such great Brahm an settlements and in the
few towns like Kanchi, Tanjavur, and G angaikondacholapuram
— im portant centres o f Chola authority — that personnel and
resources for the increasingly elaborate rituals o f the times could
be supported. Thus, U ttaram erur had two Vishnu and five Siva
temples, N allur had two Vishnu and four Siva temples, and Ennayi-
ram is said in an inscription o f a .d . 1037 to have had twelve Vedic
temples and village shrines.29
26 These are most conveniently found in V. Rangacharya. A Topographical List
o f the Inscriptions o f the Madras Presidency (Collected till 1915), 3 vols., G overn­
m ent Press, M adras, 1919. The volume and page are provided here: v. 1: 512,
37, 73; v. 2: 1359, 1413, 1308, 1265; v. 3: 1614, 1617, 1580 [Cited hereafter as IMP}.
27 Ibid., v. 1, p. 512; P.V. Jagadisa Ayyar, South Indian Shrines, The M adras
Times Printing and Publishing Co., M adras, 1920. pp. 221-30.
28 IM P , v. 1, p. 226; v. 2. p. 1235.
29 N ilakanta Sastri, Cola Studies, pp. 101, 85; A .R .E ., 1917, no.335, dated a . d .
 Central brahmadeyas appeared to maintain an elaborate and
enduring network of relationships with other institutions in South
India. They were places o f record in a society in which there were
few other means for recording and preserving im portant public
facts. Among such facts would be those related to the great w arrior
houses o f the time including, but not limited to, royal families.
Im portant officers o f the Chola overlords are panegyrized in long
inscriptions, the purport o f which was to provide a gift to a temple
or to some Brahmans. These great kavya inscriptions in Sanskrit
as well as other languages must be considered as a form o f public
notice of an essentially non-religious nature notw ithstanding their
ostensible purpose.30 The efficacy o f incising public notices in
these settlements derived from their attraction for large numbers of
people participating in periodic religious ceremonies and other
activities, including educational, economic, and political.
M ost Brahm an villages, however modest, carried out at least
a limited set o f educational functions; the central Brahm an villages
were the premier educational centres of the time. M at has, or
seminaries, of considerable size as well as individual teachers were
provided with regular income by the sabhas. While such grants
to individual teachers are am ong the most commonly recorded in
the inscriptions o f the macro region, evidence relating to centres
of advanced learning involving m any people are m ore rare. An
inscription of a .d . 1048 refers to 260 students and twelve teachers
at Tribhuvani; a record at Ennayiram , somewhat earlier, mentions
370 students o f which m ost were junior in status, the rest, perhaps
seventy, being regarded as senior scholars o f Vedas and Prabhan-
das; in addition there were fourteen teachers.31 Both of these
inscriptions specify the allotments o f food-grains to each of the
several classes o f persons associated with the schools. At Tiru-
vadutturai, a school for medical and grammatical training was
provided with lands by a military officer (sendpati ) from which the
sabha exempted all dues.32 These activities — individual scholars
1037: idem, The Colas, v. 1, p. 563; Madras District Gazetteers, South Arcot, ed.
B.S. Baliga. Governm ent o f M adras, M adras, 1962, p. 483.
30 John F. Fleet, one of the early epigraphists in India (governm ent epigraphist,
1883-6) recognized the public notice function of these records, a fact insufficiently
recognized by m odern scholars (cited in S ./.T .I., v. 3, pt 2, p. 164).
31 N ilakanta Sastri, The Colas, pp. 630-1; A.R.E.. 1917, no.333 and A.R.E.,
1919, no.176.
« Ibid., p. 632; A .R .E ., ¡925, n o .159.
expounding particular portions o f brahm anical knowledge as the
M imamsa of Prabharkara or larger groups of scholars pursuing
m ore varied subjects — were at the core of the purpose of the
Brahman village. The clientele o f this tuition were Brahman
youths just as the instructors were Brahmans, and toward the
objective o f transmission of this sacred and semi-sacred lore, the
sabhas o f Brahm an settlements m aintained a broad network of
relationships with other, minor, Brahm an villages, as well as non-
Brahman settlements.
Mercantile and artisan groups comprised an im portant segment
o f the population of the central Brahm an settlements. As populous
places which attracted pilgrims from near and far, there were
profitable opportunities open to both groups. A unique temple
like the Rajarajesvara shrine at Tanjavur cannot, of course, be the
measure o f the scale o f operations extant in other, smaller places,
for it was a special m onum ent o f one o f the greatest o f the Chola
warriors, R ajaraja I. Thus, its employment o f over 600 temple
servants, according to an a .d . 1011 record, included musicians,
accountants, and various kinds of artisans as well as ritual func­
tionaries and teachers. Tanjavur must be regarded as unusual.33
The m agnitude of mercantile and artisan activity at m ost central
Brahm an settlements was certainly m ore modest, yet integrally
related to the pilgrimage custom o f these centres. Temples of
the central Brahman settlements were consumers of large quantities
o f goods, many of which could not be provided locally. Various
kinds of condiments, oils, and the ephemera of ritual offerings had
often to be acquired from distant places.34
Principally, central Brahm an settlements are to be identified
by the num ber o f inscriptions found in them which pertain to the
affairs o f other places. Anticipating the great temple centres of
the later medieval period in South India, when im portant events
o f the locality, territory, and kingdom were registered in the
lengthy prasastis o f inscriptions, the central Brahm an settlements
were repositories o f local public notice. In the course of a half
century of epigraphical collection by archaeologists of the Govern­
ment of India and several of the larger Indian States, such as
33 A ppadorai, op. cit., v. 1, pp. 275-6; S.1.1., v. 2, no.66; and K.A. Nilakanta
Sastri, T h e Economy of a South Indian Temple in the Cola Period-, Malaviya Com­
memoration Volume, ed. A.B. Dhruva, Banaras H indu University, Banaras, 1932.
34 C am phor is one such material, A ppadorai, op. cit.. v. 1-, p. 288.
 Mysore, Travancore, and H yderabad,35 most o f the macro region
has been surveyed, thousands of inscriptions copied, and many of
them published. Though the resultant record is incomplete, it is
possible to identify some o f the places which occupied a central
recording function in relation to other locality institutions.
One turns inevitably to U ttaram erur, the place with perhaps
the largest num ber o f sabha inscriptions (over ninety) o f any
Brahm an settlement — indeed, any viilage — in South A sia;
it is m oreover one o f the most studied of ancient Indian settlements,
the most recent and excellent example being the work o f F. Gros
and R. Nagaswam y.36 Here, two early, undated records of the
ninth century refer to endowments in support of two temples,
characteristically identified simply as mahudeva , or Siva, temples
in villages o f the locality. One was in the village o f Tiruppulivalam,
3 .5 miles north o f U ttaram erur on the main road to the Cheyyar
river;37 the other was Tittatur, m odern Tittalam , about 5 .5 miles
south-east of U ttaram erur.38 The endowment for the temple at
Tiruppulivalam was m ade by the residents of the non-Brahman
quarter (sahkarappadi ) o f the north-east bazaar ( vadakilangddi}
o f U ttaram erur with the consent o f the mahasabha. whereas the
grant to T ittatur was m ade directly by the latter. These neigh­
bouring villages were clearly not Brahman villages, but Tirup­
pulivalam may have been a village whose m ajor share of income
had been granted to a temple in U ttaram erur (i.e., devaddna).
Similar to these is the complex and interesting record o f U ttara­
merur, dated a . d . 1 1 3 3 , in which a temple o f the village, or one
nearby, having demanded repayment of a loan which had been
made to the U ttaram erur assembly, and the latter, being without
funds to meet the payment, including accumulated interest, trans­
ferred the share o f income enjoyed by the mahasabha from a
35 B. Lewis Rice (ed.), Epigraphia Carnatica, 12 vols., M angalore and Bangalore.
1886-1905; Travancore Archaeological Series, v. 1, M adras, 1910-13; v. 2, ff.,
T rivandrum ; Hyderabad Archaeological Reports, Calcutta, 1915; now superseded
by the Andhra Pradesh Government Archaeological Series, H yderabad.
36 Op. cit.
37 A .R .E ., 1898, no.79, N ilakanta Sastri, Cola Studies, p. 120; Subrahm anya
Aiyer, Historical Sketches . . . . v. 2, p. 214.
38 A .R .E ., 1898, no.4; N ilakanta Sastri. Cola Studies, p. 120; Subrahm anya
Aiyer, Historical Sketches . . . , v. 2, p. 231. Others m entioned in the neighbourhood
of U ttaram erur are; Adalam pundi (Anaram pundi), 3.1 miles SE. of U ttaram erur,
ibid., p. 235; Puliyur, 2.1 m iles'SE. (ibid., p. 237), and M arudattur (M arudam ),
3.5. miles NE. (ibid., p. 254). Also, G ros and Nagaswamy, op. cit.
neighbouring village, Verm akuttanallur, to the temple in order
to discharge the debt.39
Several points deserve notice in this record. One is that this is
an example o f a Brahm an village assembly — one o f the greatest
in the macro region — having become so indebted to temple
authorities, in their own and in other villages, that they were com­
pelled to divest themselves o f a portion o f their own income sources.
This phenom enon is general during the twelfth century, and it
constitutes irrefutable evidence of the decline of the great Brahman
villages by that time. The other point o f interest in this a .d . 1133
inscription is th at there was a provision th at the nam e of the village
transferred from the Brahm an village to the temple, Vennakutta-
nallur, a name derived from one o f the appellations o f the god
Krishna, a Vishnu deity, should henceforth be known by the
Saivite name, Tiruvekam banallur.40 The change was not a lasting
one; the village of V ennakuttan still exists in M adhurantakam
taluk o f Chingleput district. Still another U ttaram erur inscrip­
tion, o f a .d . 1018, relates a transaction o f the mahasabha with four
Vaishnava priests o f the Vaikanasa school in which they were
assured o f income in U ttaram erur by a grant o f income shares for
temple worship (archanavritti) to compensate them for the loss
o f their income in the distant Brahm an village o f Arasanim angalam
(m odem C uddalore taluk, South Arcot district).51 This record
dem onstrates, as m any o f the eleventh century do, the widespread
network of a great settlement like U ttaram erur, in this case about
fifty miles, and also the spatial mobility of priestly families during
this and subsequent ages. That it was not only Brahm ans who
were capable o f cooperation in the affairs o f the great Brahm an
villages is clear from another eleventh century inscription from
U ttaram erur in which one o f its merchants, a resident o f what
appears to be a Vaishnava quarter of the village, Govindacheri,
together with a m erchant of the central bazar (nadiivilahgadi) of
U ttaram erur, jointly endowed a temple lamp in U ttaram erur.42
The same pattern o f relationship between other Brahm an and
39 A .R .E ., 1898, no.68; N ilakanta Sastri, Cola Studies, p. 128.
4« S.I.T .I., ‘G lossary’, p. xli.
41 A .R .E ., 1923, n o .171; N ilakanta Sastri, The Colas, v. 1, p. 544.
42 This is probably m istaken for the m odern Valarigadi village, A .R .E ., 1923,
no. 187; N ilakanta Sastri, Cola Studies, p. 124. See, Gros and Nagaswamy, op.
cit., p. 91.
non-Brahman peasant settlements can be shown for other central
Brahm an villages. From Tribhuvani, in m odem South Arcot,
a mahasabha order o f a . d . 1028 altered the status of a village named
V arakkur, a devadana village, into one in which forty-eight culti­
vators were assigned full rights over the production o f the village
save only a continued and specified payment to a particular temple
and the responsibility for maintaining the tank of the village.43
An a . d . 1113 inscription by order of the Tribhuvani mahasabha
directed th at a group of its village-servants — priests, carpenters,
accountants, and others — m ust desist in the practice o f plying
their crafts in other locality settlements.44 This restriction is
interesting in the light o f an earlier Tribhuvani mahasabha order
which granted an income from agricultural land to a goldsmith,
presumably to assure his services to the Brahm an village.45
in Tanjavur, the central Brahman village of Tiruvadutturai
(M ayavaram taluk) m aintained a close relationship with several
villages from which income in support of one of the Tiruvadutturai
temples was obtained or from which consumables for ritual
offerings, such as coconut and flowers, were obtained. The
Brahm an village o f Sattanur, about three miles from Tiruvadut­
turai in m odern K um bakonam taluk, existed as a satellite for these
purposes for alm ost two centuries, it would appear. Several ins­
criptions o f the ninth and tenth centuries suggest the close interac­
tion which m ust have obtained between the large and ancient
Brahm an village and its small neighbouring settlement.4^ By
a . d . 1018, Sattanur, already a Brahm an village, underwent a name
change to abhayasraya chaturvedimahgalam , suggesting that the
com m unity o f Brahm ans o f the place were in high esteem as scho­
lars ; yet, the sabha continued to register some o f its inscriptions,
as o f old, in the village o f Tiruvadutturai.47 O ther satellite B rah­
m an villages associated with Tiruvadutturai in some o f the same
ways were: Sirupuliyur (Nannilam taluk) and Sirranaichchur
(M ayavaram taluk).48 The num erous inscriptions o f Tiruvaduttu-
43 189/19 19 in N ilakanta Sastri, The Colas, v. 1, p. 557.
44 205/1919 in The Colas, v. 2, p. 595.
« 210/1919 in The Colas, v. 1, p. 585, undated, but probably of the period from
A.D. 1018-54.
4® 135/1925 and 127/1925 in The Cotas, v. 1, pp. 355, 396; and 120/1925, 102/1925,
150/1925 in The Cotas, v. 1, pp. 485, 543, 592 respectively.
47 102/1925 in The Colas, v. 1. p. 543 and 147/1925 in ibid.. v. 2., p. 567.
48 For the form er, 107/1925 in The Colas, v. 2, p. 391 and 111/1925 and others
ra i suggest an impressive netw ork o f relationships with other
settlements in this densely populated portion o f the delta. From
the non-Brahm an village o f K aran ur, some ten miles from Tiru-
vadutturai, income from land was given to m aintain a temple lamp
in a .d . 1017; a year later a Brahm an village neighbouring K aranur,
Peravur, undertook to provide regular income to a temple in
T iruvadutturai in return for a lump-sum paym ent.49 O ther brah­
madeya villages in regular relationship with Tiruvadutturai were:
Tiruvilimilalai, five miles to the south (Nannilam taluk); Tiraim ur;
Tiruvidaim ardur, five miles south-west whose m erchant assembly
(nagarattar ) participated in the purchase and gift o f a substantial
portion o f land to the T iruvadutturai temple in a . d . 942; and Palai-
vanavanmadevi-chaturvedim angalam about sixteen miles west
of Tiruvadutturai from which the income from some land was
purchased in A.D. 1016 to support physicians in the latter p la c e d
Arrangem ents similar to these could easily be evinced for many
other Brahm an villages.51
Beyond these transactions between central B rahm an villages
and other settlements involving transfers o f rights over agrarian
wealth, loans and gifts, specialists recruited and in some cases
restricted, these great institutions served as the m ost prom inent
locality site for the registration o f events quite beyond the locality.
in The Colas, v. 1, pp. 434, 437; for the latter, 139/1925, 126/1925, 125/1925 in The
Colas, v. 1, pp. 414, 423-4, 433; also IM P , v. 2, Tanjore, nos.600, 604.
49 102/1925 in N ilakanta Sastri, The Colas, v. 1, p. 543.
50 K a ra n u r was said to be in P erav u r n a d u , a territo ry nam ed for its im p o rta n t
B ia h m a n village, PeravuT (109/1925 a n d 102/1925 in The Colas, v. 1, p. 543; also
149/1895 an d 113/1925 in ibid., v. 1, pp. 433, 537).
51 Examples: a T iruvalangadu inscription of a . d . 1124 records the sale of land
by a brahmadeya called Valaikulam a alias N ittavinoda-chaturvedim angalam for
a lamp in T iruvalangadu (485/1905 in ibid., v. 2, p. 618). At Ennayiram in a . d .
1061 a village, part of whose income was previously used to m aintain temple ritual,
i.e. a devadana, called N annaderpakkam , of the temple at Eydar (m odern Eydanur,
C uddalore taluk. South Arcot) was* now to pay some portion o f its income to a
person as a life gift while the rem ainder was to be included in local revenue accounts
(336/1917 in ibid., v. 2, p. 652). T iruvadatturai served this function for a num ber
of settlements such as Tirunallam , 5 miles south, in which endowm ents were arranged
to support malhas in T iruvadatturai (T irunilvaithankan-m atha); (144/1925, dated
a . d . 1101 in ibid., v. 2, p. 581 and 155/1925, dated a . d . 1110 in ibid., v. 2, p. 590).

Others connected to T iruvadatturai include K ottur, 30 miles south in M annargudi

taluk (152/1925, a . d . 116 in ibid., v. 2, p. 599) and Siruppuliyur (62/1926, 69/1926,
71/1926, 107/1925 in ibid., v. 2, p. 391).
U ttaram erur was called Rajendrachola-chaturvedim angalam du­
ring the reign o f R ajendra I, and in the middle o f the thirteenth
century it was called Gandagopala-chaturvedim angalam for the
Telugu-Chola overlord.52 So genera! is the change of names
o f these great settlements to those based upon the personal name
or surname o f a ruler or his favoured deity as to suggest that bet­
ween these villages and the m ost powerful w arrior overlords of the
m acro region, a special relationship existed.
The presence o f C hola authority in peasant micro regions is
nowhere clearer than in the central Brahm an villages, and the
‘central’ functions of such places are perhaps m ost dram atic when
considered with reference to that authority. N ot all central Brah­
m an villages were equally influenced by kings; location appears to
have affected the extent, to which a settlement might be involved
in the activities o f the C hola overlords and their military officials.
The im portant B rahm an centres o f Tribhuvani, Ennayiram ,
and T iruvadutturai all contain inscriptions indicating that warriors
serving the Cholas took part in the decisions o f the respective
mahasabhas as they related to agrarian matters. An a .d . 1053
record from Tribhuvani reported an order from a military official
to the sabha directing that they alter the classification o f dues upon
some o f the lands under sabha control, part o f whose income was
alienated to a tem ple.53 Again, in a .d . 1099, the sabha was directed
to provide lands for the cultivation o f areca trees from which no
dues were to be collected, and the sabha was instructed to replace
the lost dues from other lands.54 External intervention is again
evidenced in an a .d . 1093 order to the Tribhuvani sabha that land-
holding arrangem ents involving w hat appears to be a military
tenure (parasavak-kani) be protected from the encroachment of
non-military cultivators, in this case, potters.55 Several inscrip­
tions from Ennayiram of the eleventh and twelfth centuries also
attest a significant degree o f external authority over the affairs of
the sabha , with soldiers instructing the assembly to provide income
for temple rituals o r being present when lands o f the Brahman
village were sold o r the arrangements pertaining to the allocation

52 N ilakanta Sastri, Cola Studies, pp. 99-100, 126.

53 188/1919 in The Colas, v. 1, p. 583.
S'1 201/1919 in ibid., v. 2, p. 577.
55 206/19 19 in ibid., v. 2, p. 572; S .I.T .I., ‘G lossary’, p. xliv.
o f proceeds from land altered in some way.56 An im portant record
o f a . d . 1001 from Tiruvadutturai relates that some money owed
t o temple weavers by the sabha was appropriated by ‘the king’,
and the assembly, fearing that the weavers would migrate else­
where if not paid, borrowed the sum from a temple in return for
which it freed some dëvadâna lands o f th at temple from certain
dues.57 A part from the evidence o f arbitrary exactions to which
a mahasabha might be exposed by those with power, this record
indicates that the mahasabha could be placed in a position of
jeopardy between these powerful outsiders and such village ins­
titutions as temples and artisan groups, all o f whom m aintained a
degree o f corporate strength with which the Brahm ans o f the
sabha had to cope.
The temple at Sucindram, in Travancore, provides similar
evidence o f the displacement o f the mahasabha which had managed
the affairs o f the temple from the middle of the tenth century.
Under the supervision o f a committee o f the mahasabha, called
the mülaparuda as at other Saivite shrines, the following m atters
were attended : general financial management, the managem ent of
lands apportioned to the m aintenance o f the temple, the super­
vision of ritual and endowments, and the prom ulgation of orders
relating to the temple. In the middle o f the thirteenth century,
the mülaparuda committee o f the mahasabha was displaced by a
group of Malayali Brahm ans called ydgakkar or pôrris, a priestly
group which may have originated in the northern coastal area of
Tulu-country, m igrated to Travancore, and developed m arital
links with N am budri Brahm ans.58
Vulnerability to these pressures upon and against the maha-
sabhas seems to have been greater in the southern portions of the
Corom andel plain than in its central and northern portions. In
Tondaim andalam , judging from the inscriptions o f U ttaram erur
and those at the temple centre o f Tiruvalangadu, which referred
to brahmadeya affairs, these Brahm an centres appeared to have
preserved a greater degree of freedom from the intrusion o f power­
ful warriors as well as less pressure from the internal power position
of temple organizations. Distance from the Chola heartland in the
56 330 and 335/1917 in The Colas, v. 1, pp. 276-7, 562; 348 and 351/1917 in ibid.,
v. 2, pp. 588, 596.
57 105/1925 in ibid., v. 1, p. 498.
58 Pillay. Sucindram Temple, pp. 143-50.
Kaveri m ay have been an im portant fa cto r; another appears to be
the greater diversity and corporate strength o f groups in the central
and northern plain.
Again, it is to the great Brahm an village o f U ttaram erur that
one turns for evidence o f well-organized corporate elements besides
the Brahm ans o f the assembly comprising the settlement. A variety
o f such groups m aintained their identity from the earliest period
o f the inscriptions to the latest. Among the m ost early are residen­
tial groups such as the people o f sankarappadi and perhaps those
o f the twelve c e ris ; others include: mahesvaras, sthanattars,
perilam aiyar, sraddhamantas, srtvaishnavas, virganattar, kaliga-
nattar, and srikrishnaganapperumakkl.59 All o f these groups are
associated with temples, and they figure prom inently in the affairs
o f the village as inscriptions reflect this. The com position o f these
groups is difficult to ascertain. In some cases, they suggest groups
not capable o f serving on or being represented in the mahasabha.
F or example, those o f the sankarappadi were non-Brahm ans as
were the same group in Tiruvalangadu; the perilam aiyar consisted
o f middle-aged women attached to tem ples; and the viraganattar
referred to Jainas.60 The mahasabha of U ttaram erur appears to
have been m ore successful than Brahm an villages in the Kaveri
basin in m aintaining their authority and control over resources
of the settlement from the expanding rights o f temple organiza­
tions. In the Kaveri region, as noted, inscriptions o f the twelfth
century indicate that sabhas were often unable to resist the shift
o f greater economic power to temples which had in the past been
entirely dependent on the assemblies o f these villages. Thus,
the Brahm an village of Sattanur, although closely linked to the
larger village of Tiruvadutturai, became for a time at least, a
tiruvidaiyattam ; that is, its income was endowed to a Vishnu temple
of Tiruvadutturai. This village was therefore under some influence
from the temple.61 A neighbouring Brahm an village, Ilachchikudi,
alias Vikramasinga-chaturvedimangalam , appears also to have
made payments to the same temple, at least once on behalf of
Sattanur.62 U ttaram erur was not, of course, completely proof
59 All of these were corporate groups involved in temple adm inistration (N ilakanta
Sastri, Cola Studies, p. 102).
60 Perilamaiyar may have been middle-aged women attached to temples (S .I.T .I.,
‘Glossary’, p. lii); vTraganatta may refer to Jainas of the village (ibid., p. xciv).
61 101/1925, dated a . d . 1009 in N ilakanta Sastri, The Colas, v . 1, p . 513.
62 See ibid.
 against this encroachment o f temple influence in sabha affairs
judging from the need which this great assembly had to transfer
its interests in the village of V ennakuttanallur to the Saivite shrine
of Ekam bam -udaiyar in repayment of a loan.63 However, a
century later, when the temple functionaries (mahesvaras and
sthanattars o f the Siva temple in Tiruppulivalam , nearby) demanded
that the U ttaram erur mahasabha resume its obligation to maintain
eight temple lamps, according to various endowments o f the past
for this purpose, the assembly was able to settle with the temple
managers for a lower schedule o f paym ents.64 Thus, even U ttara­
m erur felt the growing pressure of temple dem ands by the thirteenth
century.
This discussion o f the central B rahm an villages o f the Chola
period inevitably exaggerates their importance. In all, these great
rural centres m ight have comprised about one-tenth o f all Brahman
settlements and only a m inor fraction o f the totality of rural settle­
ments of all sorts. The cultural and religious significance of these
relatively few places obscures the fact that most Brahm an settle­
ments were smaller and more modest in function; many were
shadows — satellites — o f those centres o f civilization referred to
here as central Brahm an villages; and m ost were more closely
integrated with culture and society o f the nadu o f which they were
part. O f m ost Brahm an villages, only their names and references to
their Brahm an assemblies and Vedic temples distinguish them from
the m ore im portant peasant villages o f the localities in which they
were. Yet, every Brahm an village, however modest and remote, was
a centre o f those cultural and social forms which gave to a locality
an essential status element which its peasant people supported to
the fullest extent th at their resources permitted. If the Brahman
village was a m arker o f the prosperous, peasant agrarian localities
o f the m acro region, as has been suggested, then it was no less
a social and cultural m arker o f the m aturity and propriety o f the
collection o f peasant groups comprising the nattar.
Relations between Brahm an villages and the locality peasantry
— nattar — have been touched upon in a variety o f ways in this
discussion. These.relations are elusive, for they are not the subject
of many inscriptions. N or should it be expected that these relations
would be elaborated in great detail in stone and copper records
63 68/1898, dated a.d. 1133 in N ilakanta Sastri, Cola Studies, p. 128.
64 67/1898, dated A..D. 1215 in N ilakanta Sastri, Cola Studies, p. 129-30.
o f the time. Occasionally, o f course, detailed data relating to
grants are provided, but these are exceptional and, even then,
m uch is left unspecified. Clearly, relationships involving the
distribution of proceeds from arable land was a m atter of the
highest im portance, and the absence of such details means that
these relationships were either well enough known and fixed by
usage or they were subject to a custom ary procedure or determ ina­
tion th at perm itted allocation and reallocation of income from the
land without the need o f a perm anent record. Both o f these sup­
positions are probably correct in some degree. In every peasant
locality, the division o f produce among those with rights to a
share would have been well-established am ong locality folk.
Wide variations in the absolute am ounts which might be realized
from shares had to be anticipated owing to several factors: the
variations in cropping patterns and practices, the relationships
between those who tilled and those who exercised dominance
over the land, the shifting relationships between those directly
involved in agricultural production and service groups, and finally,
the relations between those with local dominance over the land
and overlords — great and small — to whom some tribute was
transm itted. The scope for arbitration and adjustm ent of share
divisions would have had to be considerable. Agricultural output,
then as now, would have been subject to seasonal variations, and
tributary demands would vary according to the exigent require­
ments o f local defence or perhaps the military, or to the enthusiastic
piety of the ruler. U nder these circumstances, the conclusion is
inescapable that locality custom and established modes o f adjudica­
tion to deal with the exigent would have obviated the need for
detailed descriptions o f many aspects o f the agrarian relations
between Brahm an villages and the nattar of a locality.
Such references as are occasionally found in inscriptions seem
to bear this out. M ost inscriptions record that a donor, usually
well-identified, provided land o f an unspecified quality and quantity
in a nam ed village for the benefit of the god of a temple. Where
other kinds of grants are made, there was a similar lack o f specificity
about precisely how the grant was to be realized. F or example,
many o f the grants to temples during the tenth century, and even
later, were to m aintain lamps. George Spencer’s analysis of such
procedures is most valuable.65 G rants usually specified the num ber
65 George W. Spencer, 'Tem ple M oney-Lending and Livestock Redistribution
of sheep or, at times, cattle, whose milk was to yield ghee as the
fuel. At times, grants state that these animals were to be entrusted
with a named village or person in a village whose responsibility
was to supply the temple with the fuel; for the m ost part, not even
this inform ation is recorded. Lack o f specificity about how, in
w hat am ounts, and by whom the gift o f land, animals, or money
were to be converted into the ritual service cannot be inferred to
mean that there was indifference about these things. Certainly
the donor cannot have been indifferent, nor should we assume,
were those o f the temple in charge of such m atters. One can only
conclude th at in m ost cases the procedures for fulfilling such solemn
obligations were well enough known and understood by all that
there was no need to state them in the record o f the grant.66 It
was probably sufficient for m ost pious donors to have the permanent
record of their charity inscribed in stone, leaving to those res­
ponsible for temple affairs the proper execution o f the grant in
accordance with accepted procedures. It is only in the light o f such
unwritten and custom ary arrangem ents, understood by all o f the
locality folk involved in the support o f the religious and educational
activities of the Brahm an village and temples, that the records of
the Chola period can be understood.
In the larger sense, these unwritten, essentially custom ary ways
of doing things indicate an im portant aspect o f the relationship
between the Brahm an village and the peasantry am ong whom they
existed. There was m utual confidence and respect between the
two. N ilakanta Sastri, in discussing the relationship between the
sabha and ur o f U ttaram erur, has stated that ‘. . . a vague transla­
tion of Vr and Ürórn into “village” and “we, the inhabitants o f the
village” is hardly satisfactory’. Pointing out that the ur was an
ancient locality organization, he goes on to argue that this body
o f peasants had a com m anding voice within the locality in respect
to many issues.67 In several U ttaram erur mahasabha inscriptions,
in Early T anjore', The Indian Economic and Social History Review, v. 5, pp. 277-93.
66 Alternatively, there is some evidence that such m atters were also recorded
on m ore ephem eral m aterials, such as palm leaf (e.g. A .R .E ., 1961-62, no.449,
a . d . 1106) where it is stated that white ants destroyed the palm leaf record and

thus lost; this kind of procedure may have been followed often. However, to assume
that m uch was left unwritten appears to be m ore plausible and conform s more
realistically to w hat would be expected in an essentially noil-literate society which
m ost of those concerned would have comprised.
<>7 N ilakanta Sastri, Cola Studies, pp. 103-4.
the ur of the locality undertook special responsibilities for temple
endowments which involved the sale of rights over village land by
the ur and the subsequent supervision of the endowment, under the
general control o f the sabka.6S in another case, the ur o f U ttara­
m erur assigned all the dues to which it was entitled from the southern
hamlet, called Ulliyur, to a Saivite temple in that village.69
Where one must differ with N ilakanta Sastri about the relation­
ship between such great Brahm an villages as U ttaram erur and the
locality peasantry is his assum ption o f the involuntary subordina­
tion of the peasantry as a consequence of external pressure:
. . . the simplest explanation of the existence side by side, as in U ttara­
m erur, of both the organizations is to suppose that the Or was the more
ancient form and that the Sabhá came on top of it when, at the will of
some king or chieftain, a considerable number of new Brahman residents
. . . were settled in the village, and endowed with perpetual rights of
property in a part of the village lands.70
There is no evidence to support the argument that the rights con­
ferred upon B rahm ans were extracted by force or the threat of
force from the non-Brahm an, peasant people o f a locality. On
the contrary, as even N ilakanta Sastri has pointed out, there is
every indication that Brahm ans were granted such rights as they
possessed on a wholly voluntary basis by the dom inant peasantry.71
Further, the idea o f conflict between the Brahm an sabha and
the peasant ur, speaking for the locality peasantry, is based upon
the obviously diminished role o f ur within the precincts of the
brahmadeya village. This reduced scope o f ur functions within a
Brahm an village, and presumably the complete elimination of the
ur in some cases, did certainly occur, providing scope for the super­
visory functions o f the committees ( variyam ) o f the sabha. However,
the ur was not simply a village institution in the same sense as the
sabha. The urar were members o f a locality social system in which the
village which was granted as brahmadeya was located. In ceasing
to enjoy certain rights in these villages, they did not forfeit their
68 Ibid., pp. 105-6.
69 41/1898, 9th C. in ibid., pp. 122-3.
70 Ibid., p. 104.
71 T h ey [Brahmans] lived on voluntary gifts from all classes . . . and devoted
themselves exclusively to learning, teaching, and writing. They showed themselves
capable of detached thinking on social questions and provided patterns of ethical
and religious conduct [and were] active helpers and disinterested arbitrators. . . .’
(Culture and History o f the Tamil, pp. 95-6).
prerogatives as members o f locality society. The decline of the
ur in Brahm an villages did not m ean a diminished capacity of its
peasant constituents to exert authority in those settlements and
over those lands beyond the jurisdiction o f the sabha. The fallacy
arises from viewing the relations between Brahm ans and peasants
within the framework o f the individual village, when it was not the
village, but the locality — the nadu — which was the prim ary
unit o f peasant organization. So long as the nattar were capable
o f acting together with effectiveness, as they were long beyond the
time of the great Brahm an villages, the village unit was a means for
providing requisite self-government to non-peasant groups. The
locality was dom inated by peasant folk under the authority of
locality chieftains, themselves drawn from peasant ranks and
dependent upon peasant support. Brahm ans were accordingly
granted rights to be exercised under the sabha, and to assume that
such village units were m odal in character is to ignore the capability
o f the locality peasantry which, in the period after the twelfth
century, showed itself to be increasingly able to exercise authority
separate from such village bodies, on the one hand, and the great
C hola overlords, on the other.
The extent o f support for Brahm ans and their ritual activities
by the locality peasantry cannot be assessed with any reliability from
existing records. There were variations from place to place in
accordance with the num ber o f Brahm an families constituting
beneficiaries o f the brahmadeya, the relative fertility and agricultural
potential o f the lands granted, the elaborateness o f ritual facilities,
and support from other than local people. U nder the term s o f the
inscriptional charters, Brahm an villages realized their basic support
from a grant of the m ajor portion of income, or ‘m élváram ’ as it was
called, produced from the lands around the newly constituted
Brahm an settlement. The melváram, ‘higher-share’, was that
which had been enjoyed by those with control over the locality of
prosperous settlements from which the Brahm an villages were
necessarily created. The actual proportion o f output represented by
the mélvaram is not specified in inscriptions, and there is no basis
whatever for suggesting some regular proportion o f output.72
72 To suppose that the sastric injunction of 1/6th as the appropriate share to the
king was followed, as some suggest, is baseless. Equally difficult to accept are such
references as 1/5th for forest land and l/3rd for rice land, as specified in the a . d .
1072 inscription o f K ulottunga I, referred to above, as constituting a regular basis
A nother portion o f income to which reference is made is the
kTlvaram, ‘lower-share’, or the kudiváram , ‘cultivator’s-share’; this
appears to have gone to those who actually supervised cultivation
and perhaps took part in it. These income shares, varcrm, were net
proceeds since there were in addition numerous, other small dues
to which reference is m ade in the inscriptions of the period which
seemed to be in excess o f these shares.73 In most Brahman villages,
it must be supposed, it was the ‘higher-share’ o f the net proceeds
from cultivation that went to the Brahman beneficiaries, and the
former cultivators carried out their supervisory activities as before.
This may be inferred from the references to an arrangem ent some­
times mentioned in which the previous cultivators are said to have
been removed from the land and the village at the time of being
granted.74
Displacement o f those who supervised cultivation and received
the k ih a ra m , or iow er-share’, appears to have been unusual.
However, its occurence helps to explain the mention in many of
the mahasabha inscriptions o f specialized committees or offices
o f the assembly charged with the supervision o f various aspects
o f cultivation.75 Two o f the im portant committees (variyam) of
the U ttaram erur assembly were the tank committee (eri-variyam)
and the garden committee (tótta-váriyam ) ; service on these com ­
mittees was considered a necessary qualification to service on the
for calculating the income o f those who could extract a portion o f peasant production.
73 I am grateful to T.V. MahaHngam for perm itting me to use a glossary of
dues of this period which runs to 43 typed pages and presents a bewildering array
of charges against peasant production in excess of such regular paym ents o f m élváram
and kllváram (‘U nit for the Preparation of Topographical List of Inscriptions in the
M adras and Kerala States: Epigraphical Glossary of Terms Relating to Taxes and
other Dues, Custom ary and Feudal’, D epartm ent of Ancient History and Archaeol­
ogy, University o f M adras, M adras, 1967). An im portant analysis o f some of these
terms is found in N. K arashim a and B. Sitaram an, ‘Revenue Term s in Chola Inscrip­
tions', A jia A fu rk a Gengo B unka K en kyu (Tokyo), no.5 (August 1972), pp. 87-117.
74 This is expressed by the term kudi-ningáya and refers to lands which have been
granted to a temple for its support, S .I.T .I., ‘Glossary’, pp. xxix-xxx.
75 The term variyam is usually translated as com m ittee; however, as N ilakanta
Sastri has pointed out, it is m ore to be understood as an office for carrying out some
im portant function of the Brahm an settlement. As a committee formally con­
stituted by the sabha , the variyam could be seen more as a forma! than functional
institution. In some cases this may have been true, but in others the váriyams did
provide supervision to operations with which they were charged; Cola S tu d ies ,
p. 133, and S.1.1., v. 14, no.78, p. 50, a . d . 860, for regulating irrigation channels,
and S .I.I.. v. 3, no .5, pp. 8-9, a .d . 883, for removing silt from a tank.
annual or standing com m ute o f the assembly {samvatsara-va-
riyam ).76 The brahmadeya Tirupparkadal contains a record from
another Brahm an village, Kaviri-Pakkam, alias Am aniarayana-
chaturvedim angalam , which refers to the following offices o f its
mahasabha'. kalani-variyam for the general supervision of cultiva­
ted lands, éri-váriyam for tank supervision; kaliñgu-váriyam
for the supervision and m aintenance of sluices, and tadivali-
variyam for the supervision o f paths and roads around cultivated
fields. These offices are enum erated in an inscription of about
a . d . 960 which records a grant by the brahmadeya Kaviri-Pakkan

to a temple in the larger village of Tiruparkadal o f some o f its

as that of U ttaram erur of about a . d . 924. In one, the tdtta- variyam,
cultivable waste land, m a n jikka m , for reclamation in order to
support a temple service.77 T hat the mahasabha offices actually
supervised appears clear from a few o f the existing records such
or garden officials, were authorized by the assembly to acquire
sufficient land along an im portant irrigation channel in order to
carry out dredging and repairs; the record states that lands were
purchased and the work com pleted.78 Where inscriptions speak
o f the displacement o f the form er cultivators, as some do, it may
be presumed th at supervision of cultivation was undertaken by
Brahm an members o f the sabha and, under these circumstances,
kllvaram , as well as the melvciram, would have gone to the assembly.
Supervision o f cultivation in m ost cases remained the preroga­
tive o f those who had held this right prior to the grant of the brahma­
deya, and these peasants continued to retain a portion of the net
proceeds o f cultivation while passing to the sabha that portion
called m elva ra m jy Continuity with the older pattern of agrarian
relations was m aintained by the persistence o f the ur as a corporate
76 N ilakanta Sastri, Cola Studies, p. i 74.
77 S .I.I., v. 3, no. 156, pp. 32-3.
78 S .I .Í . , v. 6, no.292, p. 147; also see Subrahm anya Aiyer’s interesting reading
of this inscription in which he takes the param esvara-vadi to be a road not a channel
as it is usually taken in inscriptions (cf. T am il L exicon, v. 6, p. 3489; vati, as channel
according to usage in S .I.I. (H istorical Sketch es . . . , v. 2. pp. 249-51).
79 The term m élváram and niiyatchi both seem to refer to the income which was
taken by those with the superior claim. The particle m el m eans not only ‘higher’,
‘upper’, or ‘m ajor’, but also ‘superior’ in the sense of priority of entitlem ent and
‘first’ in the sense of that which is taken first ( T am il L exicon, v. 6, pp. 3354 and 3356
and S .I .T .I .; ‘G lossary’, pp. xxvii and xxviii; also G upta, op. cit., p. 192 and
Minakshi, op. cit., p. 138).
entity within the Brahman village, even if its functions were reduced
by those of the sabha. The relationship between the Brahman
sabha and the peasant ur has properly been called ‘hazy’ by Nila­
kanta Sastri with reference to the best documented Brahm an village
which exists — U ttaram erur.80 This relationship appears to be
a one-sided one, favouring the Brahmans, which surely reflected
the fact that, as N ilakanta Sastri has stated, all o f the evidence
o f the relationship in U ttaram erur date from a time after the village
had become a chaturvédimañgalam and all o f the transactions
recorded were those o f the sabha. M any o f the rights which the ur
or its executive, the áluñganattár, exercised from the time when
there was no sabha came to be executed by the sabha. Yet, even
in this great Brahm an settlement, a taniyür in its own small territory
(tan-küru), the non-Brahraan essentially peasant, assembly con­
tinued to have an existence which can be glimpsed through the
exclusively Brahm an records o f the place.
N ilakanta Sastri concludes his discussion o f the ur and the
sabha with the assertion th a t: ‘the ancient Or by the side o f the new
Sabha was secured as a part o f the new order’.81 T hat ‘new order’
centred upon the existence o f Brahm an settlements throughout the
Corom andel plain, in every peasant locality, enjoying the support
o f the dom inant peasantry with whom Brahmans had formed
close relations during the Pallava period. Brahm an settlements
were linked to each other by common ritual ties and, judging
from the territorialization o f Brahm an sub-castes, by kinship ties
as well within the localities in which they were; they were also
linked across nadu lines by similar bonds and through the central
place functions which involved not only the multifarious activities
o f the sacredotal élite, but those o f merchants and artisans as well.
At this time when the fundam ental organization of South Indian
society was segmented into many isolated peasant localities, the
central Brahm an villages played a significant integrative role.
Possessing a high degree o f spatial mobility, Brahm an families
were able to move from peasant villages to Brahm an villages and
to temple centres; they were invited from places in one part of
the plain to others hundreds of miles away for their particular
ritual or sastric knowledge. In these centuries Brahm ans COntri-
so N ilakanta Sastri, Cola Studies, p. 105.
Ibid.
buted much to a m acro regional culture which bridged the numerous
social and cultural differences among territorially segmented
peasant peoples.
The Brahm an village functioned as a vital hinge in this macro
regional culture. Standing between the great temple centres —
populous towns o f often great antiquity as Kanchi and M adurai —
and the m ajority of Brahmans who lived in small, separate parts
o f prosperous peasant villages, B rahm an villages of the Corom andel
plain m aintained institutions essential for the preservation and
transmission o f Indo-A ryan culture. In these settlements, with
control over substantial agrarian resources — a token o f the
support o f the peasantry am ong whom they lived — and with the
patronage o f locality chieftains and great overlords in Tanjavur,
temples, schools, and mathas of various kinds elaborated and
disseminated the ritual and philosophical doctrines o f the age.
Here too, the basic facilities for the pursuit o f proper (satvik )
m ode o f brahm anical life was m ost readily realizable. Ritual
and social purity according to agamic and sastric proscription
could be m aintained in the m any peasant villages and the few towns
in which Brahm ans lived, of course; but in these places there was
the proximity o f m any people o f low ritual status and the lack of
ritual specialists required for many life-cycle and other ceremonials.
Brahm an villages, to be sure, were also pluralistic settlements.
In the original charter of Srivilliputtur, for example, six shares of
the m ajor portion o f income from the village were allocated as
penisai vritti , or shares for skilled artisans.82 Vellalas are often
referred to as residents of Brahm an villages in Tondaim andalam
and elsewhere.83 M erchants too lived in the great Brahm an villages
and participated in the life o f the place as distinctive, corporate
elements.84 The largest portion o f the non-Brahm an population
in B rahm an settlements as in others m ay well have been those who
laboured in the fields o f Brahm ans and Vellalas as dependents
o f these two powerful groups. Such people would have lived in
82 V iraraghavachari, ‘Srivilliputtur Tem ple’, p. 1.
83 A .R .E ., 1909, no.729, 11th C., M adhurantakam -chaturvedim angalam ;
A .R .E ., 1922, no.20, 11th C., Arinjiyam angalam .
84 A .R .E ., 1919, no.399, early 11th C., refers to a m erchant o f Nellur (Nellore)
as a resident o f M uranottam angalam , S.1.1., v. 14, nos.72, 46 of early 10th C. refer
to a m erchant and carpenter; also, K.V. R am an, The E arly H isto ry o f the M adras
Region, A m udha Nilayam Private Ltd., M adras, 1959, p. 174.
hamlets apart from the ceris o f the principal settlement — as they
do today — but well within the precincts o f the locality dominated
by the Brahm an assembly. Finally, in Brahm an villages at the
edges o f the plain, or deep in the upland plateau, as in K arnataka,
even tribal people frequented, if they did not at times dwell in,
some Brahm an villages.85 N otw ithstanding its pluralistic charac­
ter, the Brahm an village, better than other possible residential
situations, afforded a unique context for the requirements of the
pious and learned Brahm ans o f Coromandel, and to the extent
that it is accurate to say that Brahmanical culture and life-styles
were an im portant element in social models of the macro region,
the Brahm an village was a keystone of Corom andel culture because
it was the Brahm an place, p ar excellence.
Life-styles in Brahm an villages of the Corom andel plain influen­
ced the behaviour of those dom inant peasants who sought to
separate themselves from others in the countryside. Dietary rules,
language, ritual activities, and the general modes of social interac­
tion o f the dom inant peasants o f the plain were modelled upon
those of Brahm ans am ong whom they lived. The extent to which
this* influence was effective in m oulding a set of broad, subregional
variants of respectable peasant culture becomes somewhat clearer
after the twelfth century when dom inant peasant groups began to
establish their own ritual and educational centres. These followed
those of the Brahmans but w ithout the participation of the latter.
Peasants also shifted their allegiance in ritual m atters from the
sabhas of the Brahmans to temples within and outside the Brahman
villages.

85 This somewhat late reference of the 14 C. or 15 C. is from an inscription from

the brahm adeya o f Bellur alias V ishnuvardhana-chaturvedim angalam in which
a K arnataka tribal folk (Sahara) were m entioned as disciples of Brahm an teachers
(M y s o re Archaeological R eport, ¡913-14, p. 6; cited in B.A. Saletore, The W ild
Tribes in Indian H istory, M otilal Banarsidass, Lahore, 1935, pp. 58-9).
 Right and Left Hand Castes
( valangai and idangai)

The nadu, the basic territorial unit in the South Indian macro
region, gave to the agrarian system o f the Chola period a highly
fragm ented character, elements o f this remain to this day. Culti­
vated land and the nexus of relationships involving land exercised
a strong centripetal influence upon the structure o f social relation­
ships in South India as in other pre-industrial agrarian contexts.
Added to this, however, are the distinctively regional characteristics
of spatially compressed marriage, kinship, and political relation­
ships resulting in cores o f peasant settlements which were dis­
continuous and relatively small. These settlement units remained
small and isolated until the thirteenth century in m ost parts of
the Corom andel plain and even longer in the western uplands,
assuring to the «adw-locality its primacy as a structural unit.
Two factors tended to offset the isolation of the nadu without
diminishing its integrity. One was the network o f brahmadeyas
from whence, during the tenth to the twelfth century, em anated
a general, highly aryanized culture spreading from the Coromandel
plain over the entire m acro region. These were powerful, corporate
institutions which exercised continuous influence for several
centuries. The other was the emergence, by the eleventh century
at least, o f dual social divisions rooted in the num erous nadu
societies but capable of transcending the isolation of these localities.
These were potential social formations which could be activated
for a variety of purposes, but which were not corporate or con­
tinuous in character. W hat the Brahm an settlements o f the region
did to foster integrative cultural bonds among dom inant peasant
folk within the m acro region, the divisions o f the ‘right-hand’
and ‘left-hand’ peoples or castes appeared to do in forging signi­
ficant social links am ong a variety of dependent peoples o f diverse
localities. In both cases, cultural and social integration beyond the
level of the nadu was the consequence although it was not until
after the thirteenth century that the nadu began to lose some of
its early primacy as the focus of society and culture in the macro
region.
Labels for the dual social divisions have persisted for almost a
millenium. Valangai, the Tamil word for ‘right-hand’ or ‘right-
side’, as a social designation dates from the tenth century when
contingencies of R ajaraja Fs armies, valahgai-velaikkara-padaigal,
are m entioned.1 D uring the early eleventh century, persons calling
themselves valangai, m ade endowments to temples as in the case
of the temple at V em barrur, alias Sri-Cholam attanda-chatur-
vedimangalam in Tanjavur.2 References to groups o f the ‘left-hand’
or ‘left-side’, idangai, appear somewhat later; one o f the earliest
recorded an affray between people of the right and left hand in
a . d . 1072. This record reads in part:

. . . in the second regnal year of the king (Kulottunga I) there was a

clash between the right-hand and left-hand communities in which the
village was burnt down, the sacred places destroyed, and the images
of deities and the treasure of the temple (Mummudi-Choja-Vinnagar-
Alvar temple) looted.3

Thus, by the late eleventh century, there is evidence of two broad

and at times hostile divisions o f the population in at least some
parts of the Corom andel plain; shortly it was to cover almost
the whole of Tamil country. These divisions appear also to have
existed in other parts o f the macro region at about the same time
though there is less convincing inscriptional evidence. In southern
K arnataka, the equivalent K annada terms for right and left-
hand, balagey and edagey, were used as designations for the divi­
sion.4 O ther designations later used among K annada speakers
were: desa, for right-hand people and nadu for those of the left-
hand division.5 Am ong Telugu-speaking people of the macro
1 V. Venkayya, S .L I., v. 2, ‘Introduction’, p. 10: C.S. Srinivasachari, ‘The Origin
of the Right and Left H and Caste Divisions', J .A .H .R .S ., v. 4 (1929), p. 80.
2 1.M .P ., v. 2, p. 1287, 341/1907, dated a . d . 1014.
3 A .R .E ., 1936-7, para. 27; also summarized in N ilakanta Sastri, The Colas,
p. 551.
4 C. H ayavadana Rao (ed.), M ysore G azetteer, v. 1, G overnm ent Press, Ban­
galore, 1927, p. 178; L.K. A nanthakrishna Iyer, The M ysore Tribes and C astes, v. 2,
Mysore University, M ysore, 1935, p. 114; L. Rice, M ysore, v. 1, p. 222.
5 Rice, M ysore, v. 1, p. 224. He reports that the term p ete was used in place of
the term nadu.
region, slightly different designations were used. One, kam pulu
(literally ‘protector’ but in com m on usage, ‘agriculturist’), appears
to have had the same meaning as the terms used in Tamil country
and K arnataka for the right-hand designation while the terms
pahchahanam varu and pahchdnulu were the same as the left-hand
division elsewhere.6 The latter term in Telugu inscriptions refers
to five artisan-trader groups usually consisting of goldsmiths,
blacksmiths, braziers, and stone-and-wood sculptors, hence, pahcha,
or five.7 in later centuries especially, but apparently even in the
twelfth and thirteenth centuries, another way o f referring to the
dual division am ong Telugu speakers was by their sectarian affi­
liations. Adherents to Vishnu worship. srTvaisnavas, being the
counterpart of the right division and Siva adherents allegedly
corresponding to the left division.8
Analysis of the origins and functions of the dual divisions of
peoples of the macro region have posed difficult problems. Though
the subject of serious scholarly speculation for almost a century,9
the origins o f the divisions remain obscure. Classical poetry
contains no references to the divisions, and the terms have little,
if any, contem porary currency.10 A note on origins is appended
to this chapter. It has proven just as difficult to understand the
functions of the right and left divisions, for the dual divisions
resist analysis according to such Chola cultural categories as
caste, sect, and territorial (nadu) affiliation.
6 R. N arasim ha Rao, Corporate L ife in M edieval Andhradesa, Secunderabad,
1967. pp. 110-23; S. Chandrasekhara Sastri, ‘Economic Conditions U nder the
Hoysalas’, The H a lf-Y e a rly M ysore University Journal, v. 2 (1928), pp. 215-16.
7 N arasim ha Rao, op.cit., pp. 56-63.
8 N. Subha Reddi, ‘C o m m u n ity Conflict am ong the Depressed Castes of A n d h ra’,
M an in India, v. 30, no.4 (Oct.-Dec. 1950), p. 4; N arasim ha Rao notes the use of
‘Srtvaisnulu’ in groupings o f what appears a right-hand character in medieval
Telugu records (op.cit., p. 116). However, see contra, Sidney Nicholson, ‘Social
Organization of the M alas — An Outcaste Indian People’, Journal o f the R oyal
Anthropological In stitu te o f Great B ritain and Ireland, v. 56, (Jan.-June, 1926),
p. 91: ‘m ost Right-hand castes are Saivite in faith . . .’ This discrepancy cannot
be explained by regional variation since the references appear to be to the Rayala-
seema portion of south-western Andhra.
9 Cf. the work of O ppert which was published in 1893; notice two somewhat
earlier, brief, and less analytic discussions in the Indian A ntiquary: J.S.F. Mackenzie,
‘Caste Insignia’, v. 4, (Nov., 1876), pp. 344-6 and James F. Kearns, ‘The Right-
H and and Left-H and Castes', v. 5 (Dec., 1876), pp. 353-4.
10 T hough she reports that the dual divisions are not known to m ore than a few
 While each of these categories may be found at times to have
been related to the dual divisions, the divisions are essentially
different. Thus, ranking seems present at times, as in an a .d . 1227
inscription in which newly adm itted groups to the division in a
part o f m odern South A rcot are declared ‘the eyes and the hands
of the idahgaV, body images suggesting the perform ance of service
for other members o f the division.11 Generally, however, the
divisions give the appearance o f being non-ranked groupings of
local social groups. Also, while certain elements o f sect organi­
zation may be seen at times in the references to an idangai perceptor
or a mandapam, the divisions are not essentially religious groupings.
Finally, while the divisions have territorial focus— there being no
m acro region-wide divisions as such — that territory appears
always to have been greater than the nadu.
The categories o f caste, sect, and territory fail to help in an
understanding o f the dual divisions in South India because the
divisions are different from each and all, and because, at least in
the early period under consideration, the scope of these three
social categories was very highly localized whereas the dual divisions
appear to be essentially supra-local in character. It is therefore
little wonder that where the divisions have been considered as
something to explain by historians, these divisions are often treated
together with other, so-called, ‘corporate’ institutions in a modest
genre o f historical literature dealing with what is called, ‘corporate
But, two persistent features of the early right-left divisions
militate against their dismissal as ‘corporate groups’ in presumed
ensemble with other like groups o f a caste, sectarian, or territorial
in kind. One is the im portance o f references to the assimilation
o f previously outside people to the left division and their alliance
with other generally similar groups; the other is the military and
life’ or ‘local government’.12
of her inform ants, Brenda E.F. Beck m ade the conception a central organizing
principle of her recent, P easant S o ciety in K ohku, University o f British Columbia,
Vancouver, 1972. Com m enting on her usage, D um ont in H om o H ierarchicus,
p. 288, called it 'sui generis'.
11 A .R .E ., 1940-1, n o .184; discussed below more fully.
12 The best known o f these for early India generally are: R.C. M ajum dar, Cor­
p o ra te L ife in A ncient India , C alcutta, 1918 and R.K. M ookerji, L ocal Government
in A ncient India , Oxford, 1919; for South India: R. N arasim ha Rao, Corporate
L ife in M edieval Andhradesa, Secunderabad, 1967 and F.S. Dikshit, L ocal Self-
G overnm ent in M edieval K arnataka, D harw ar, 1964.
colonization ventures with which both divisions o f the Chola
period were associated. Together, these two features o f assimila­
tion/alliance and m ilitary/colonization convey the sense of a social
order which is not fixed in term s o f its structural constituents
nor in space, but a social order which is in flux, one expanding
from its relatively isolated local forms o f organization to ever
wider forms o f societal and cultural integration. The corporate
imagery o f the existing historiography with its presumption of
fixedness and stability around bureaucratic kingship, caste, and
guild or its conception o f conflict resolution through factious
groupings fails to appreciate the dynamism o f Chola society.
Viewed as a ‘corporate institution’, the dual division is looked
upon as guild-like, or as a srenl, i.e. a multicaste body of traders,
artisans, and agriculturists. This guild conception is based upon
the well-recognized association o f the right division with agriculture
and related activities, including trade and some processing of
agricultural commodities, as well as the equally consistent asso­
ciation o f left division groups with artisan-trader activities. The
guild or srenl notion also fits well with the general Indian institution
usually called ‘theja/m aw system’ 13 — localized exchanges o f goods
and services centred on the ritual and economic dominance of
agricultural patrons (jajman , S anskrit: yajam ana) and their clients.
However, any essentially cooperative and interdependent model,
whether guild/ire«T or jajm a n i , fails to deal with the often con­
flictful relations between the divisions which are attested in his­
torical records from at least the eleventh century, as noted above.
Thus, some scholars, cued by conflict between the right and left
divisions, have applied the term ‘faction’ following the usage of
some British adm inistrators.14
‘F action’ denotes an alignment o f persons for the purpose of
attaining some objective in com petition with others.15 Conflict
13 See the excellent sum m ary discussion by D um ont, H om o H ierarchicus , pp. 97-
108.
14 M cKim M arriott, C aste R anking and C om m unity Structure in Five R egions o f
India and P akistan, Deccan College M onograph Series, no.23, Deccan College,
Poona, 1960, pp. 46-51, where he speaks of ‘factional lines of ritual patronage and
alliance’. J.H . H utton speaks of ‘factious rivalry’, Caste in India, Oxford, 1963,
p. 67. The term is also used in the following o f the M adras G azetteers: Salem ,
F.J. Richards, pp. 125-6; Trichinopoly, F.R. Hemingway, pp. 92-3.
15 S ee Alan R. Beals and Bernard J. Siegel, Divisiveness and So cia l C onflict:
is the business o f factions, and the term fits, most aptly, certain
o f the activities with which the right and left divisions in South
India have been associated during recent centuries certainly, and
possibly from a much earlier time.
As sociological elements, factions have been viewed in many
ways by m odem scholars, though all might agree with the hum orous
observation of Nicholas that ‘the faction is a troublesome form
o f social organization’.16 Factional alignments can and have been
relatively persistent in some societies, particularly at times of
special internal strain and external stress.17 And, however un­
stable they may be, factional systems can achieve some im portant
objectives through means not usually considered appropriate and
often in contravention to norm s regarding conflict resolution.
This would seem especially true in cultures which emphasize
‘harm ony and unanim ity’ or where ‘cooperation’ among social
groups is given high value as it is in caste culture according to
many scholars.18 Finally, even if factions may be evanescent,
‘their com ponent cliques and families may be stable groups’.19
Accordingly, the dual division o f social groups in the South Indian
m acro region may plausibly be seen to lend itself to analysis as
factional systems even as it is recognized that there were im portant
changes in the composition, purposes, and context in which the
divisions operated in the course o f perhaps eight centuries.
Neither ‘faction’ nor ‘guild’ appear fully satisfactory terms for
discussion of the early phase o f the dual division of social groups
in South India. If one were to adopt Nicholas’ definition of faction
— ‘a noncorporate political conflict group, the members o f which
are recruited by a leader on the basis of diverse ties’20— it would
A n A nthropological A pproach , O .U .P., Bombay, 1967, pp. 21ff; Lewis Coser, The
Functions o f Socia l Conflict, Free Press, Glencoe, Illinois, 1956; R aym ond Firth,
‘Factions in Indian and Overseas Societies; Introduction', British Journal o f S o ­
ciology, 8, no.4 (Dec., 1957), p. 292.
16 Ralph W. Nicholas, ‘Factions: A Com parative Analysis’, in M. Banton (ed.),
P olitical S ystem s and the D istribution o f Power, London, 1969, p. 21.
17 These positions are examined in Beals and Siegel, op.cit., p .166.
Ibid., p. 158.
« Ibid., p. 166.
20 Ralph Nicholas, ‘Structures o f Politics in the Villages of Southern Asia', in
Stru ctu re and Change in Indian Society, eds. Milton Singer and Bernard S. Cohn,
Aldine Publishing, Co., Chicago, 1968, p. 278. N ote the difference here from
A ppadurai who appears to reserve the term ‘faction’ for m anifestations of right-
be necessary to reject the term during the Chola period or at any
time prior to the seventeenth century. The dual social divisions
in South India were not solely nor primarily conflict groups at
this earlier time, and, while ties am ong each o f the divisions may
have been varied, the core o f the interests defining each are per­
sistent and clear. M oreover, to the extent that factions may be
viewed as ‘ego-centred’, essentially individual-participant ‘quasi­
groups’, as M ayer has called them, the South Indian dual divisions
would not qualify. The constituent units of the divisions are always
localized caste groups.21
As neither the concept of ‘faction’ nor ‘guild’ which have been
used to describe the right and left divisions appear to fit certain
of the im portant characteristics o f the divisions, some other way
o f speaking about them is necessary. The recent essay about the
right and left division by A rjun A ppadurai postulates a cultural
model to deal with the conflict of right and left castes, especially
in M adras city during the seventeenth century where he also uses
the concept of faction. However, for the Chola period, the stress
upon conflict is misconceived. Conflict appears a m inor aspect
of the divisions during this early period however im portant it
becomes later. To emphasize conflict between the divisions at
this early period is to impose later characteristics upon the divisions
of the Chola period and thus to distort an understanding of the
institution in Chola times.
M ost Chola inscriptions pertaining to the valangai and idangai
do not refer to conflict, but to the typical subject m atter of in­
scriptions: gifts to Brahmans and temples. From these references,
we learn of various kinds of groups cooperating beyond their local
bases. Among the most prom inent were the valañgai-másénai22
and idahgai-másénai :23 the great armies o f the right-hand and
left-hand. Trade activities and especially relations with im portant
left conflict in M adras city (Arjun A ppadurai, ‘Right and Left-Hand Castes in South
India', l.E .S .H .R ., v. 11 [June-Sept. 1974]).
21 A drian Mayer, ‘Quasi-Groups in the Study of Complex Societies', The Social
A nthropology o f C o m p lex Societies, ed. M. Banton, Association of Social A nthro­
pologists M onograph, Edinburgh, 1966, p .116.
22 A .R .E . 1933, no.232 and 233 discussed by K.S. Vaidyanathan, ‘The M embers
o f the South India Arm y (sénai): Their Assembly and Its Functions’, Q J M S . XXXII
(1941-2), pp. 301-3.
23 A .R .E . 1961-2, no. 478 in the script of the eleventh century, from C hannapatna
taluk. Bangalore district.
itinerant trade organizations were other reasons for extra-local
cooperation among locality peoples. Thus there are the numerous
inscriptions referring to nanadesi merchants meeting together with
local merchants of the nadu and nagara, that is merchants of
ordinary agricultural villages of a locality as well as special trade
settlements, including valangai weavers.24 And, finally, there are
rare Chola records of agreements by lower caste people of the
valangai and idangai divisions, in some places at least, for resistance
against ‘the Brahm anas Vellalans who hold the proprietory rights
(.kani ) over the lands o f the district’.25
The divisions are thus seen not as ‘absolute’ social entities,
for example, as ‘super castes’ as suggested by the terms ‘right-hand
castes’ and ‘left-hand castes’, but as ‘relative’ or ‘potential’ groupings
of established local groups. Such aggregate groupings were capable
of dealing with extra local problems beyond the scope and capability
of existing locality institutions of the time and capable of being
called into existence in response to a variety of problems, including
conflicts, requiring extra-local cooperation.26 At any time and
place, the com position o f right and left divisions would vary ac­
cording to the exigent condition which brought them into being,
and they would lapse into latency with the passing of that condition.
Viewed as relative or potential groupings rather than as enduring
corporate ones (e.g. guilds) or as a d hoc conflict groupings (i.e.
factions), the dual divisions of Chola times assume an anachronisti-
cally m odern appearance. T hat is, the valangai and idangai divi­
sions of the Chola period appear as broad ethnic coalitions which
are neither internally ranked in the m anner of castes into subordi­
nate sub-castes nor externally ranked with respect to the other
bloc or division. Rather, in the m anner of horizontally integrated
South Indian caste associations of the recent past — the N adars
and Vanniyakula Kshatriyas27— an absolute quality is claimed
on the basis of birth into a named group and the ascriptive right
24 A .R .E . 1912, no.342 and para. 25/1913.
25 A .R .E . 1913, no. 34 at A duturai, Tiruchchirapalli, probably in the time of
K ulottunga III ( a . d . 1178-1218).
26 As noted by M arshall Sahlins, ‘The Segmentary L ineage: An Organization
o f Predatory Expansion’, in C om parative P olitical System s, ed. Ronald Cohen
and John M iddleton, New York, 1967, pp. 105-6.
27 Sum m arized by Lloyd I. and Susanne H. R udolph, The M o d ern ity o f Tradition-.
P olitical D evelopm ent in India, The University of Chicago Press, Chicago, 1967,
pp. 36-64.
to certain emblems and insignia. Hierarchical bases of status or
m oral condition are ignored. The whole or part of such large
groupings are capable of acting together for certain purposes,
but they do not lose their localized bases of organization and,
typically, interm arriage among the constituent groups does not
occur.
The need for such supra-local coalitions was particularly great
until well into the thirteenth century by which time urbanization
provided a reliable supra-local focus at least for leading artisan-
trader groups, usually designated ‘left-hand’ people. By the same
time, the widespread merging of nadu -localities into the periyanadu,
or great nadu, provided for collaboration among leading agrarian
peoples — those of the right-hand — on a supra-local level. Prior
to the thirteenth century, however, the dual divisions, with their
varied constituencies from place to place, represented perhaps
the sole means by which groups other than Brahmans and some
military chiefs could on occasion transcend the borders of the
nadu- locality.
One of the m ajor reasons for seeing the dual divisions as ‘relative’
or ‘potential’ structural entities, rather than ones which had an
absolute (i.e. ‘corporate’) existence in particular places is that
neither division finds mention among those groups named in the
detailed Chola inscriptions dealing with m atters requiring the
assent of or the cooperation from im portant local groups. In
brahmadeya inscriptions, these local bodies inevitably include the
following: the nattar; assemblies of neighbouring brahmadeyas
represented by their spokesmen (or headmen), the brahm.ad.eya-
ki[avar\ assemblies of villages (ur ), part of whose income was
previously granted to support Hindu temples (devadana ), Buddhist
and Jaina shrines ( paUiccanda ); assemblies of villages which were
trade centres and under m erchants’ control, nagarattar or naga-
ragalitar; and assemblies of villages (ur), some portion o f whose
income was diverted to other forms o f special purposes (kanimur-
uttu and vettiperu ).28 Other Chola inscriptions refer to other
bodies, including k il kalanaigal, who are described as including
carpenters (taccan), blacksmiths (kollar), goldsmiths (tatt&x), and
koliyar (weavers?29).30 Seen in these references are the various
28 These terms are taken from ‘The Larger Leiden Plates o f R ajaraja I,’ £ ./.,
v. 22, p. 258.
29 k a u lika appears to be the Sanskrit word upon which the Tamil term koliyar
caste groups comprising the agrarian-centred division o f the
valangai in Tamil country and the mobile artisans of the idangai ,
but there is no mention of these divisions themselves.
Hence, rather than use terms such as ‘faction’ or ‘guild’ to
speak of the dual division, the term ‘division’ will be used. The
meaning attached to the term ‘division’ is that o f the occasional
com bination o f agrarian-centred groups, on the one hand, and
artisan-traders on the other at levels beyond the localities in which
both kinds of groups lived.
One of the m ost im portant functions of the idangai division
was the assimilation of groups to the expanding order o f the
Chola period. From the tenth to the thirteenth century new
tracts of land not previously committed to sedentary agriculture
were being brought into the expanding ambience of the Chola
agrarian order. W hether by conquest or by the peaceful extension
of the Chola agrarian system, people of these new tracts were
brought into the dual divisions, and the groups thus included in
the dual divisions might be agriculturists who had previously
practiced shifting cultivation or they might be artisans or they
might be any one of the various kinds of occupational groups
which were not already aligned with one or the other dual divisions.
In either case, the newly recruited groups could henceforth make
alliance claims upon others in their division and even cause the
division in any place to change from potential to actual groupings
for a variety of purposes.
This process of assimilation is well exemplified in two early
thirteenth century idangai inscriptions. The first, from the Ut-
tam acholan temple of U rrattur, fifteen miles north of the Kaveri
River (Lalgudi taluk, Tiruchirapalli), is dated a .d . 1218 and reads
in part:
In order to k'ill the demons (that disturbed) the sacrifices of Kasyapa
[the priest o f Visvakarma, patron god of artisans] we were made to appear
from the agni-kunda (sacrificial fire pit) and while we were thus protecting
the said sacrifice, Chakravartin Arindama honoured the officiating sage
priests by carrying them in a car and led them to the Brahmana colony
(newly founded by himself). On this occasion we were made to take our
seats on the back of the car and to carry the slippers and umbrellas of
is based (Sircar, Epigraphical G lossary , p. 159 and Monier-W illiams, Sanskrit
D ictionary, p. 317, where another meaning is ‘left-hand’ sa k ta worshippers).
id A .R .E . 1888, n o .118; a Chidam bram inscription of c. a .d . 1036. (Rajendra
I, 24th year); see N ilakanta Sastri, The Colas, v. 1, p. 562.
 these sages. Eventually with these Brahmana sages we were made to
settle down in the [same] villages. . . . We received the clan name idangai
because the sages (while they got down from their cars) were supported
by us on their left side. The ancestors of this our sect having lost their
credentials and insignia in the jungles and bushes, we were ignorant of
our origins. Having now once learnt it, we the members of the 98 subsects
enter into a compact, in the fortieth year of the king [Kulottunga III]
that we shall hereafter behave like the sons of the same parents and what
good and evil may befall any one of us, will be shared by all. If anything
derogatory happens to the idangai class, we shall jointly assert our rights
until we establish them. It is also understood that only those who, during
their congregational meetings to settle communal disputes, display the
insignia horn, bugle, and parasol shall belong to our class. Those who
have to recognize us now and hereafter, in public, must do so from our
distinguishing symbols — the feather of the crane and the loose hanging
hair. The horn and the conch shell shall also be sounded in front of us
and the bugle blown according to the fashion obtaining among the
idangai people. Those who act in contravention to these rules shall be
treated as the enemies o f our class. Those who behave differently from
the rules (thus) prescribed for the conduct of the idangai classes shall
be excommunicated and shall not be recognized as srutiman (members
o f the community). They will be considered slaves of the classes opposed
to us.31

The second record is from V aranjuram (Vriddhachalam taluk,

South Arcot) and is dated a . d . 1227. It reads:
We, the nadus [assemblies of eleven localities] having assembled at the
village of Tiruvalanjuram . . . got the following resolution engraved on
the Tiruvalanjuram-udaiyar temple: ‘the malaiyamakkal and the natta-
makkal of these nadus shall henceforth be admitted into the idangai-
talam [left hand class of m en]; they shall be considered the eyes and hands
of the idangai; if we violate this resolution, we shall be considered as
wrong-doers to the caste. 32

The resolution was endorsed by Brahmans, and other leaders of

the locality as well as by those calling themselves of the idahgai-
31 A .R .E . 1913, para. 39; summarized by N ilakanta Sastri, The Colas, pp. 551-2.
Also note the interesting comment on this record by D.C. Sircar, The Guhilas o f
Kiskindha, Sanskrit College, Calcutta, pp. 22-3.
32 The nadus m entioned are: irungelappddi, ku n nra-kurram , tunda-nadu, iirumu-
naippddi, cengunra-nadu, Vanakappudi adaiy-nadu, pangala-nadu melkarai-nadu,
gangappddi and paranurru-nadu. N ilakanta Sastri also refers to this record —
184th/1940-1 — which he locates in the Kallakkurichchi taluk of South Arcot
(T h e C olas , p. 552); the text cited above was from a translation by Y. Subbarayalu,
research scholar, D epartm ent o f Archaeology, University of Madras.
talam including kaikkolars and sdliyar (weavers), vanigars (mer­
chants) and others.33
These, idangai records of the thirteenth century presume the
existence o f an established supra-local social entity into which new
groups could be initiated. In the first inscription it is not clear
who those of the lost credentials and insignia were, though the
‘ignorance’ about origins and the references to the jungle suggest
persons who, in other circumstances, would be low in ritual status.
Those mentioned in the second inscription are more readily identi­
fiable and interesting cases to which reference shall be made shortly.
A point which m ust be taken up first pertains to the apparent
lack of emphasis upon stratified relations within either the idangai
or the valangai division. It is as if the divisions were homogeneous,
pluralistic aggregates in which all constituent groups shared a
common status and com m on symbols of rank. There are two ways
in which this apparent homogeneity among the constituent units
of the divisions is expressed. One way is in shared natural sub­
stance, that is in attributes ascribed to the divisions as living things
which possess unique endowed qualities arising from how they
came into existence. Thus, according to a later source, the Idangai-
valangai Puranam o f a . d . 1692-3,34 both divisions were created or
brought into being by the actions of gods. In one context, Siva and
Indra are made responsible for the left-hand division and Brahma
and a rishi (bhrigu , Tam il: p iru ka ) for the right-hand division;
in another context, the divisions are seen as the result of a disagree­
ment between Siva, as Paramesvara, and his consort Parvati.
M ore specifically, the Idangai-valangai Puranam assigns to each
division different somatic markers. To the left division, the most
im portant are blood, skin, and eye-balls; to the right division,
bones, nerves, and brain. The other symbolic way in which the
constituent units of the divisions appear to have been accorded equal
status and thus to constitute a pluralistic aggregate of equivalent
units, was in the common emblems each division possessed even
though each constituent unit had its own emblems. In the U rratur
inscription above, several insignia were given prominence. Ac­
cording to the Idangai-valangai Jatiyar Varalaru of the Mackenzie
33 In place of the word, idangai-ta[am , N ilakanta Sastri has the word, idahgait-
tanattom or 'other Idangai people o f the area’ (T he C olas , p.552).
34 D ated in 1615 of the saka era or 4794 of the kaliyuga\ Oriental M anuscripts
Library, University of M adras, Mss. no. D. 2793.
collection, the ninety-eight castes o f the right division had common
emblems of the ‘Brahm ani’ kite (Garuda?), the half-human, half­
animal form (‘purusha-m irukam ’), the elephant, ass, and eagle, while
the left division had the tiger, fox, the male bird (p ô tu ), swôrd,
crow, ‘Brahm ani’ kite, horse, lion, and a mythological animal with
a face bearing features o f the lion and elephant (ya/z).35
Such shared insignia and symbols o f common ‘natural’ attributes
am ong units o f the divisions may appear to imply ‘corporateness’
in the sense that castes are corporate. That is, the dual divisions
may be supposed to have been something like ‘super-castes’ with
the same quality o f durable and diffused solidarity which charac­
terized a caste. However, this would be incorrect. Rather; the
divisions were groupings with quite specific elements of solidarity,
such as possessed by sectarian groups. The religious sect, with
certain exceptions like the Lingayats, was comprised of persons
o f many castes (though excluding the very lowest castes), but strati­
fied interactions were irrelevant when sect votaries acted in religious
contexts: Thus, in the confines of the sect centre, all sect members,
regardless of their caste affiliations, interacted as equals in ritual
activities.
Différences between the dual divisions and sects are im portant,
however. The religious sect was an absolute, not relative social
form. Its enduring character was sampradaya, a tradition passed
from sectarian leader, âcharyâ to disciples; its institutional base
was the m atha or sect centre. In relation to the sampradaya and
the matha all laic members suspended their caste identities though
such identities obviously continued. Similarly, the dual divisions
were composed of localized caste/occupational groups who interac­
ted according to caste norm s in their own localities. These norms
could be-altered to enable joint action with others in a broad,
essentially occupational, alignment on a variety of matters. Co­
operation in military ventures and in support of religious institu­
tions is how the divisions are usually seen. These other activities
m ust be seen as ancillary to the maintenance of occupational
35 Idangai-valangai Ja tiya r Varalâru [History o f the Left H and and Right H and
Castes], the Mackenzie Mss., Oriental Mss. Library, University of M adras, no.R.
1572. This is undated; it may be of the eighteenth century. In the enum eration
of insignia for the divisions provided to the Chinglep.ut m agistrate Colem an in 1809,
some twenty were described as appropriate for all or any right division caste and
slightly fewer for castes of the left division (J.S.F. Mackenzie, op.cit., p. 345).
interests at the supralocal level and, of course, at the local level
where it m ost counted. In disregarding caste distinctions among
their constituent units, the dual divisions were not denying caste
in the sense that sects did in obedience to bhakti principles. Reli­
gious bodies o f the medieval period often affirmed the supernatural
order prevailing at the sect temple centre by suspending the ‘natural’
order of caste relationships. F or the right and left divisions it is
rather that caste, whether viewed as localized ethnic groups or
as ritually ranked parts of a m oral order, was not salient for the
supra-local, occupational functions of the divisions, at least at this
early time. Thus, caste groups are m entioned, but appear to have
little to do with the way in which the divisions were transform ed
from latency to deal with the issues they did.
The endorsement o f Brahmans and other prestigious members
of local society in the Varanjuram record cited above does not
clarify the m atter of internal stratification o f the divisions. Along
with Brahmans, referred to in that inscription as andanar, there
were ekayar, i.e. ascetics, and niyayattars , i.e. local persons of
prominence.36 Weavers and merchants who endorsed the resolution
have the appearance of being persons of wealth, but there is no
definite attribution of their superior status in the record. Since
Brahmans are outside the divisions in most accounts, the association
o f Brahm ans in both the inscriptions tends to confirm the equality
of status that existed among the constituent castes of the divisions.
It is as attendants of ‘Brahm ana sages’ or through the endorsing
function o f Brahm ans that the claim o f respectability and mem ber­
ship in one o f the divisions is m ade and justified.37
The absence o f references to stratified divisions within the dual
divisions does not eliminate the possibility of internal strata.
F o r some scholars such strata appear at times to exist. References
to groups like the ka ikkd la r and other weavers and merchants,
who are mentioned along with Brahmans in the endorsement,
suggest this to C.S. Srinivasachari. In his pioneering work on
the dual divisions he states that there were indeed strata and that
the divisions reveal a process of ‘low castes striving for higher social
36 ‘G lossary’, p. xliv.
37 A ttention is again drawn to A ppadurai’s recent discussion, especially his
iteration and analysis o f various accounts of the origins of the divisions (op. cit.,
p. 233 ff).
positions’.38 F urther support for this view is provided by the
nattam d kka l folk discussed in the V aranjuram inscription.
Thirty years after the date of the Varanjuram inscription cited
above, the n a tta m d kka l claimed for themselves the status of
pum iputtirar, ‘sons o f the soil’, in two inscriptions from Vengur and
Tirukkoyilur, near the site of the Varanjuram record.39 The title
pum iputtirar is significant since it is claimed by Vellalas, the do­
m inant peasants of the right division, the valangai. For the natta­
m dkkal, their membership in the left division but lately attained
(i.e. a . d . 1227) this was an ambitious claim indeed! But such claims
become m ore com m on later.40 T hat the n a tta m d kka l o f modern
South A rcot m ade this claim is supported by usage in Jaffna,
northern Sri Lanka, where n a tta m d kka l are called ‘kings of Vellalas
(orusar v e la la f indicating that those migrating from what is now
South Arcot to Sri Lanka during the medieval period may have
carried this relatively exalted designation.41 However, during the
nineteenth century, the nattam dkkal as well as the m alaiyam akkal
m entioned in the a . d . 1227 inscription were still closely linked
territorially and in m arriage.42 This suggests that while the plains­
men (nattam dn ) o f this part o f South A rcot might have sought to
differentiate themselves from the hill folk (malaim an ) by arrogating
to themselves titles such as pum iputtirar, and by association.
38 Srinivasachari, op. cit., p. 84.
39 Vengur inscription is 206/1936-7, comm ented upon in para. 43 of A .R .E . 1936-7;
the T irukkoyilur record is 117/1900 and is discussed in the same place as well as
I .M .P ., v. 1, no.845, p. 226.
40 Buchanan cites evidence of this in recording the disputes between the right-
hand groups o f K om ati and Panoham Banijiga merchants, the form er claiming
leadership of the division on the basis of being Vaisyas whereas the Banijiga were
but Sudras (op. cit., v. 1, p. 80).
41 T am il L exico n , v. 4, p. 2149, ‘n a tta m d k k a l ’■ The association of these people
with Sri Lanka is further supported by an inscription o f a . d . 1518 and one o f a . d .
1425 in which Vanniyar warriors are referred to as conquerors of Sri L anka (in
T.N . Subram aniam , ‘Fresh Light from a T iruvottiyur Inscription’, Sem inar on
Inscriptions [Kal-vettu Karuttarangu], ed. R. Nagaswamy, Books India Private
Ltd., M adras, 1968, p. 197). Further confirm ation o f the connection of Vanniyars
with Sri L anka comes in a purdna allegedly composed in a . d . 1615 in which a warrior,
Devakirtti, ‘em peror in all countries where Vanniyars lived, seized Lanka and Vengi'
idangai-valangaipuranam [Ancient Story of Left and Right Castes] (Oriental Mss.
Library, M adras, doc. no. D. 2793). N a tta m d k k a l are associated with the caste
of Vanniyar through their connection with pal/is.
42 Thurston, op. cit., v. 7, p. 206.
Vellala-like status, this claim did not hold. The nattam akkal
remained a peasant people below the status o f the Vellala and were
often identified as part of the Palli caste, a peasant group incon­
gruously of the left division.43 Hence, if status differences were
at times stressed, these were not always successful.
M ilitary actions by the dual divisions occupy a conspicuous
place in the early records of the divisions and pose most sharply
the question of the potential or relative character of the divisions.
There are references to the ‘great army' (m asenai ) of valangai and
idangai, and to fighting men called velaikkara which comprised
part of the Chola army in Sri Lanka during the late eleventh ce n tu ry
according to a Tamil inscription at Polonnaruva.44 Also, inscrip­
tions of the middle of the eleventh century from Tiruvenkadu in
Tanjavur and Tiruvallam in N orth Arcot refer to grants by m em­
bers of both divisions to temples, notwithstanding the fact that in
other places, notably Kanchi, though perhaps not at this early
time, the two divisions used different temples, halls, and dancing
girls.45
The term valangai is first encountered m connection with military
contingents under the first of the great Cholas, R ajaraja I. Analysis
of Chola military organization, granted the meagreness of the
evidence, leads to the conclusion that the designation valangai
could only have referred to armed contingents raised and com ­
m anded by the dom inant peasantry of the Chola heartland of
Cholam andalam and Tondaim andalam . D uring the time o f R aja­
raja I (985-1014) and R ajendra I (1014-44), inscriptions from
Tanjavur enum erate regiments o f the army of which almost one-
half (thirteen o f thirty-one) were entitled valangai.46 Other forces
43 Ibid., 212. However, the Mackenzie Collection ‘History o f Left Hand and
Right Hand Castes' (fn. 34), refers to n a tta m a kka l and m a la ym a kka l as right-hand
groups.
44 K.S. Vaidyanathan, ‘The M em bers o f the Ancient South Indian Army (Senai):
Their Assembly and its F unction’, Q J M S xxxil (1941-2), pp. 301-3; E .I., v. 18,
no.38, ‘Polonnaruva Inscription of Vijayabahu I’, pp. 330-8; A .R .E . 1961-2, no.478,
11th C.
45 S .I.T .I., v. 3, pt 2, 63. O ppert (op. cit.) refers to the special place of Kanchi
(p. 62) and also m entions that valangai and idangai worshippers at the SelvapiUai
Temple o f Melkote were given different hours for worship; also mentioned in A .R .E .,
1921, para. 47.
46 The following are the term s m entioned in the T anjavur inscriptions related
to m ilitary units of Rajaraja 1; each is followed by the words, terinda-valahgai-
included household troops and troops draw n from territories on
the m argins o f Chola country, ‘vadugan ’ from the northern, Telugu,
tracts and malaiyâlar from the hill borders o f m odern K erala in
the west.47 These forces serving the Cholas bore the designation
vëlaikkârar, and the valangai were further identified by the addition
to their titles of one of the many pseudonymns taken by R ajaraja
and a connective word, terinda, or ‘selected’.48 Thus, there was the
unit called: nittavinoda-terinda-valahgai-vëlaikkarappadaigal, ‘Nit-
tavinoda’s [R ajaraja’s] select right-hand w arrior regiment’.49
The meaning of vëlaikkârar has vexed historians for half a
century, and it is still not clear. That it refers to warriors is unam ­
biguous from the contexts in which it appears in inscriptions and
literary sources o f the tenth to the twelfth centuries. But ‘whose’
soldiers they were is still at issue. The central problem involving
this term is whether the vëlaikkârar were special and permanent
troops o f the Chola overlords or whether they were enlisted for
extraordinary or occasional military service as is suggested by the
vêlai, one o f the meanings of which is ‘occasional’.50 If they
were the perm anent troops o f the Chola overlords, the valangai
units cannot simultaneously be considered local peasant militia
units. Alternatively, if the valangai vëlaikkârar were recruited to the
military adventures of the Cholas from existing military units among
the peasantry — controlled and led by the peasantry — then the
association of the valangai with the peasantry would appear as
strong in the early period as it is in the lists of the eighteenth century.
The prevailing view of the vëlaikkârar is stated most clearly by
N ilakanta Sastri. He says that they ‘were the m ost permanent
and dependable troops in the royal service . . . they were ever
ready to defend the king and his cause with their lives when occasion
(vêlai) arose’.51 This view is supported by the editor of the Tam il

vefaikkarar except th e first tw o in w hich th e term s valahgai-vëlaikkârappadigal

o nly ap p ear. T hese a re: perundanattu-, sirudanattu - ; a n d aragiya-sora-, chanda-
p ara kra m a -, ilaiyarâja-, kshatriyasikham ani-, m uratavikram âbharana-, rajaraja-,
nittavinôda-, raja ka n th irava-, râjavinôda-, ranam ukha-bhim a-, vikramàbharana-,
V. V enkayya, v. 2, ‘In tro d u c tio n ’, p. 10.
47 E .I., v. 18, n o .38, lines 39-44, p. 338.
48 N o te d by V enkayya, S .I.I., v. 2, ‘In tro d u c tio n ’, pp. 10-11; see also Tam il
L exico n , iv. p. 2034, 'ten".
4» V enkayya, S .L I.. v. 2, ‘In tro d u c tio n ’, pp. 10-11.
50 T am il L exico n , v. 6, p. 3844.
51 The Colas, p. 454; T am il L e x ic o n , v. 6, p. 3844, v. 2, p. 895.
Lexicon who writes, under the entry, velaikkarar: ‘devoted servants
who held themselves responsible for a particular service to their
king at stated hours and vow to stab themselves to death if they fail
in th at’.52 As evidence for this view, N ilakanta Sastri refers,
with uncharacteristic vagueness, to later literary sources while
the Lexicon cites the commentary of Periyavachchapillai on
N am m alvar’s Tirumoli, to the effect that these soldiers committed
suicide for their king.53
This view of the devotion of the velaikkarar to their king, whom
they served presumably on a permanent basis, replaced an older,
less heroic, view held by the epigraphists Hultzsch, Venkayya,
and K rishna Sastri. They spoke o f ‘troops of servants’, ‘volunteers’,
or simply, soldiers of lower status (‘working classes’) who fought
in Chola armies.54 Other scholars have suggested that the velaikka­
rar were mercenary troops as were others in the Chola forces.55
It may also be noted that the word velaikkarar differs from the
word velaikkarar, servant or workman, only in the retroflex ‘1’.
At issue here is whether these warriors, representing half of the
known regiments enumerated in inscriptions of the great Cholas,
could be considered a permanent force, supported from the re­
sources of the Chola overlords or whether the valangai velaikkarar
were mobilized from among existing peasant military units for
some limited purpose, and were thus an extension of the valangai
(or idangai) as a potential social formation. The former view is
congenial to that of N ilakanta Sastri and others who have tended
to place heavy reliance upon the comprehensiveness and effective­
ness of the Chola state. However, there is no evidence of the basic
means of supporting a large army any more than there is for m ain­
tenance o f an elaborate bureaucracy. Neither, of course, was
necessary. Just as locality institutions provided m ost of the ad-
52 T am il L exicon, v. 6, p. 3844.
53 Cited in E .I., v. 18, no.38, p. 334. F or a more complete version of this argu­
ment see: K.A. N ilakanta Sastri, ‘A N ote on V elaikkarar', Journal o f the R oyal
A sia tic S o ciety (C eylon), v. 4, (N.S., 1954), pp. 67-71.
E .I., v. 18, p. 334.
55 K.K. Pillay, South India and Ceylon, University o f M adras, 1963, p. 80, fn.,
in which he speaks o f mercenary soldiers; however, on p. 144, Pillay says they were
a perm anent element of the Chola armies. The view that velaikkarar were m er­
cenaries finds recent support in a doctoral thesis by Kenneth R. Hall, ‘The N agaram
as a M arketing Center in Early Medieval South India’, unpublished, Dept, of
History, University o f M ichigan, 1975, pp. 191-2.
ministrative functions required at the time, so, too, it must be
supposed that the m ajor forces involved in the wars of the Cholas
were supplied from the existing organizations of the locality-based
society of the time. To the core of household troops m aintained
by the Cholas, who may indeed have held a special loyalty to their
overlord, and some mercenary troops from the western and northern
forests, those under the control of peasant locality leaders alone,
during the tenth and eleventh centuries at least, could have pro­
vided the military units under Chola command. By the twelfth
century, idangai forces were added to this pool o f militarily organi­
zed folk within the macro region who could be mobilized to join
the Chola kings in defensive and predatory campaigns.
The association of the Corom andel peasantry with valangai
military forces is supported by an im portant record of the time
of Kulottunga I, a .d . 1072, at Avani (M ulbagal taluk, K olar
district) in Gangavadi. As discussed above,56 the claim of the
dom inant locality folk in this area to membership in ‘48,000
bhumi’ of Tondaim andalam makes their identity as Vellalas from
that adjoining territory a relatively firm one. The central purpose
of the record bears out the unm istakable peasant interests of these
locally dom inant folk. Furtherm ore, this inscription is somewhat
unusual in being one of that small class of stone inscriptions which
do not relate to a temple endowment.
This K olar inscription records how various local agrarian
groups were to be taxed in a locality, called the ‘eighteen vishqya
of Rajendra-C hola’, under the control of persons identified as
valangai of Tondaim andalam . Commenting on this im portant
record, N ilakanta Sastri sees in it the capacity of local people to
thw art the efforts of a ‘self-willed and autocratic ruler’ and the
expression of a ‘popular consciousness [that] there was a clear
limit to the taxing power o f the governm ent. . . .’ 57 According
to Professor Sastri, this epigraph records a unilateral modification
of revenue arrangements imposed upon the local peasantry by a
Chola revenue adm instrator.
There are several reasons for suggesting a different interpretation
of this inscription. It is, first of all, extremely unlikely that the
Chola overlords presumed to establish a system of detailed rates
on all the specific sources o f revenue m entioned in the record:
56 Ch. III.
57 The Cokis, p. 539.
the m achinery for such control was simply not there. It is further
questionable whether the rates set for land tax (melvdram ) in this
record — one-fifth for forest and dry crop tracts and one-third
for tank-irrigated paddy land — was for the benefit of the Chola
state at all.
N ilakanta Sastri and other historians of medieval South India
have taken the term melvdram to mean ‘government’s share’,
when it m eans: ‘m ajor’, ‘higher’, or ‘first’ share.58 M elvdram
is characteristically used in relationship to the division of produce
from the la n d ; it designates the m ajor share claimed by those who
held dom inant land rights. Kudi- (cultivator) or kil- (inferior)
varam was the lesser share. In his discussion of revenue terms,
N ilakanta Sastri does not include melvdram among a somewhat
doubtful list o f term s,59 but, on the contrary, in his discussion of
relations between those who cultivate and those who control the
land, m elvdram figures very prom inently.60 The equation of
‘m ajor share’ or m elvdram with ‘government share’ from cultivated
land is based upon nom enclature of the British ryotwari system
rather than upon early South Indian practice. The British adopted
this well-established term relating to divisions of produce between
what they regarded as ‘landlord and tenant’, then, assuming the
politic fiction that the government was landlord, the British claimed
the right to a substantial portion of produce.61 However, in its
historical context, there is no connection between the melvdram
and the share which may have gone to the ‘state’ during the Chola
period.62
The order executed by those in control of the Avani locality,
calling themselves valangai of Tondaim andalam , was addressed
to the local, ruling groups over whom the control of Tondai­
m andalam soldiery had been extended by conquest at some earlier
time. In N ilakanta Sastri’s discussion o f this record, it is treated
58 T am il L exicon, v. 6, p. 3356.
59 The Colas, pp. 521-3; am ong the m ore improbable taxes enum erated is one
upon begging (iraru) !
60 Ibid., pp. 585-6.
61 Examples o f this usage: H.R. Pate, T innevelly G azetteer, Governm ent Press,
M adras 1917, pp. 283, 321; A .F. Cox, N orth A rco t M anual, G overnm ent Press,
M adras, 1895, p. 282.
62 This view is supported further by the use of the term as understood by others
who have worked on the Chola records; see S .I .T .I ., G lossary, p. xxxvii; K.M .
G upta, op. cit., p. 192; A ppadorai, v. 1, op. cit., pp. 172-3, 176-7, 325.
as a protest ‘against unusual levies’ o f a ‘self-willed and autocratic
ruler or chieftain’.63 But, the inscription is cast in quite usual
terms with a laudatory preamble dedicated to the Chola king;
it is not an obvious record of protest though it does declare that
an order o f the adigargal-sqla-muvendaveldr would not be followed.
The muvendavelar referred to in the inscription as having pro­
m ulgated this new and inappropriate revenue regulation could
of course have been an agent of the Chola overlords or perhaps a
well-placed military officer acting on his own behalf.64 But,
it is most likely that this person was a leader of the conquering
Tondaim andalam valangai forces claiming to exercise the superior
prerogatives of a chief. In any case, as N ilakanta Sastri has noted,
defiance of his orders are clear: ‘ . . . there being no tax on cows
and she-buffaloes since the rise of the sacred family of Chojas in the
Solamandalam nadu [or] in the Jayarigonda-Sola-m andalam . . .
no such tax should be paid in accordance with the order o f the
. . . Soja-muvendavelar . . . .’ 65 If it is supposed that, like the
w arrior spokesmen o f the valangai of Avani, this muvendavelar
personage was part o f a conquering Chola army, then the conflict
was between him and others o f the conquering force which had
taken up residence in and was claiming control over this tract.
Claimed by the latter were the same rights as they enjoyed in
Tondaim andalam . The dem and for an enhancement of payments
from this tract by a tax on milch animals was therefore rejected by
those in control of the area on the basis o f custom followed in the
neighbouring area from which they had come.
While expressly rejecting the special and, to those Tondaim anda-
63 The Colas, p. 539.
64 As with m any other ‘officials' of the Chola state, the m uvendavelan is difficult
to relate to adm inistrative functions. An officer with this title adjudicated a dispute
involving a temple in Ennayiram , alias rdjaraja-chaturvedim ahgalam , in which
it was granted land taxes as low as other temples o f the place; 330/1917, dated a . d .
1048. A nother record o f an audit of accounts of a temple in T iruverkadu (Sri-
perem bur taluk, Chingleput) names an officer with this title; 386/1958-9, dated
a . d . 1054. See M. Arokaiswami, ‘The Inscriptional Term “ M uvendavelan” ', Journal

o f Indian H istory, v. 34 (April 1956) in which he concludes that the term refers to
Vellalas (p. 192), but little else; also see the discussion infra , ch. III.
65 E .C ., v. 10, pt 2, p. 87. K.A. N ilakanta Sastri in The Colas, p. 538, mistakenly
speaks o f ‘the 78 nadus of Nigarilli-sd.la-mandalam’; ‘Nigarili’ does not appear
in this record; cf. English translation and the Tamil text provided in E .C ., v. 10,
pt 2, p. 54, line 12; also cf. translation o f K.V. Subrahm anya Aiyer, H istorical
S ketch es, v. 1, p. 351.
lam colonists in control of the locality, inappropriate demand for
this tax, the rates of revenue to be paid to them as the land controllers
of the locality were explicitly stated. This could only have been
m eant for those who cultivated the land around Avani, or the
‘eighteen vishaya o f R ajendra-Choia and K andam adam ’, as the
place was called. An interesting postscript to the inscription
throws light on who the latter might have been.
Shortly after the a . d . 1072 inscription was engraved, it was
defaced.66 A new record, dated in the same year, was accordingly
incised stating:
We [the inhabitants] of the eighteen great vishaya and the great army of
the right hand [valangai], armed with great weapons, have also caused
it to be engraved in stone that those who [violate] this order shall incur
the heinous sin of having destroyed Brahmanas, herds of tawny cows,
and Varanasi and shall become hereditary enemies of the great vishaya
and the great army of the right hand armed with great weapons while
those who maintain this order shall acquire the merit of having performed
many horse sacrifices . . . 67

The strongly worded imprecation of this postscript to the Avani

record o f a .d . 1072 appeals to well-recognized m oral sanctions
(e.g. the sin of destroying Brahmans and so on), to retribution
from the Tondaim andalam valangai, and, interestingly, to kingly
honour in the form of the, horse sacrifice. The last suggests an
effort to influence local chiefs. It is moreover reasonable to assume
that the defacement of the original inscription was done under the
orders of or with the complicity of the ancient ruling groups of the
Avani locality in protest against ,the revenue demands made upon
them and their cultivators specified in the record of the Tondai­
m andalam valangai. These demands, even with the tax on live­
stock excluded, were probably greater than those which had existed
before; in any case, after all, it was not for the benefit of local
cultivators that the general tax on kine was rejected, but for the
benefit of those who had invested themselves with superior rights
through conquest, the Tondaim andalam valangai colonists. It
would be they who would enforce this regulation.
As mentioned, the a . d . 1072 records o f Avani are am ong those
66 E .C ., v . 10, pt 2, M ulbagal,' no.l 19. a . d . 1072. The epigraphist noted that
this record was presum ably m eant to substitute for gaps in the original inscription,
no.49a, which had been ‘defaced’.
67 Ibid.
rare documents which do not relate to a temple endowment or to
temple business, but to the resolution of some problem. It is even
more rare in recording the explicit rejection and modification of
an order of a m uvendavelar or some similar personage of authority.
However, the reasons for the rejection are not difficult to under­
stand.
The Avani region of K olar was an ancient, settled, apparently
prosperous area of Telugu agriculturists when it came under Chola
dominance through the military enterprise of Tondaim andalam
valangai soldiers. Inscriptions from the neighbourhood of Avani
from as early as the. fourth century a .d . speak o f the region as part
of andhra-mandala , or ‘Telugu country’, and it was a celebrated
place of antiquity and sanctity — avantikakshetra — associated
with Valmiki, composer of the Ramayana. Its sacred places during
the tenth century included several devoted to figures in the R am a
legend as well as a Smartha m atha.6% The Chola conquest therefore
involved the displacement of quite ancient, local ruling families,
or at least their subordination to the recently intrusive T ondai­
m andalam valangai. It is scarcely surprising therefore to discover
evidence o f altered revenue arrangem ents under the new overlords
of Avani and its general neighbourhood. It is not surprising
either to find resistence to these arrangements by the local Telugu
cultivators of the place and their chiefs who were only recently
subjugated by a people with a different language and customs.
The defacement of the original a .d . 1072 inscription and the post­
script added later may be attributed to the fact of this conquest.
..Still puzzling, though, is the repudiation of the bovine tax order
in the Avani inscription. As there is no further evidence on the
m atter, it is assumed that the revenue rates established in a .d .
1072 were put into effect. On that assumption, it cannot be serious­
ly considered that the defacement of the order which prom pted
the strong imprecation of the postscript was the work of the m u­
vendavelar adigdirgai Such a feeble response to the rebuff o f the
valangai colonists of Avani would appear unworthy o f one with
the title of m uvendavelar unless it were to be thought that the title
could be claimed and recognized as appropriate by persons of
very m odest coercive capacities. This is, o f course, a possibility
and one consistent with the view of muvendavelar which is pro-
68 Lewis B. Rice, ‘M udayanun Plates of Saka 261 of the Bana King M alladeva-
N andivarm an’, Indian A ntiquary, v. 15 (June 1886), pp. 174-5.
posed below. However, the more prudent reading of this aspect
of the Avani record is that, whatever the power of the adigdrigal,
the rebuff was accepted by him. This is to agree with N ilakanta
Sastri’s interpretation of the inscription as evidence of the ability
of local power holders to thw art unacceptable impositions, even
by one who may himself have been a valangai chief o f the forces of
conquering Tondaim andalam agriculturists.
The left hand, idangai , division of lower social groups in the macro
region during the Chola period was as certainly associated with
mercantile and craft occupations as the right-hand division was
with agrarian activities. The core idangai groups in all parts of the
macro region were certain merchants and craftsmen conventionally
expressed by the num eral ‘five’ as in the terms panchalar (or pah-
chalattar or panch-kam m alar) and ahjuvannam ,69 These usually
included goldsmiths, silversmiths, blacksmiths, and skilled car­
penters and stone cutters. Others characteristically associated
with the left division, according to evidence of the eighteenth
century, were oil processers using presses operated by two or more
bullocks, implying supplies o f raw m aterials and m arkets which
might be found in urban places. Certain weavers were also of the
division according to later evidence, though most were of the right
division. In the case o f weavers, there appears to be no particular
reason for the association with the left division unless scale of
operation and production for the m arket (rather than for a fixed
clientele) was a factor for weavers as it appeared to be for oil
producers.
It is possible to project the seventeenth and eighteenth century
occupational alignment of the left division backward in time.
One im portant link with the past is found in rathakara inscriptions
o f the twelfth and thirteenth centuries. Craftsmen identified as
rathakara had enjoyed an ancient honourable status according
to Vedic and later Vedic sources,70 but by the early centuries o f the
Christian era they had come to be regarded as Sudras according
69 S .I.T .I., ‘Glossary’, p. xlviii; also p. vi where reference is made to the redundant
term ahju-panchalattar.
10 S e e A.A. M acdonnell and A.B. K eith, Vedic In d ex o f N am es and Subjects,
v. 2, London, 1912, p. 265 and F. M ax Muller, The Sacred B ooks o f the E ast, 14,
‘The Sacred Books of the Aryas as T aught by the Schools of A pastam ba. . . .’,
p. xxxviii.
to the A m arakosa ,71 Reflecting the early high status o f rathakara
in South India, a late fourth century Pallava copperplate inscrip­
tion found in the K rishna district o f m odern A ndhra, dated in the
fourteenth year o f Nandivarm an I, records the grant o f an agra-
hara to one of the rathakara caste who was called a chatuvejja,
that is, one who has studied the four Vedas.72
Craftsmen of a later period occasionally used the rathakara
designation in what must be considered an attem pt to strengthen
their claims to high status. A well known inscription from Uyya-
kondan-U daiyar (Tiruchirapalli taluk) of a . d . I l l 873 records a
gathering of learned Brahmans (bhatta ) at Rajasraya-chaturvedi-
mangalam to consider the status o f a group of craftsmen, including
goldsmiths and silversmiths, carpenters, stone cutters and masons
calling themselves rathakarar.14 Having examined sastric autho­
rities,75 the Brahmans concluded that since rathakarar were of
high and correct birth (mahishya and anuldm a),16 they were entitled
to the sacred thread investiture and access to other im portant
rituals. A nother rathakara inscription from Alangudi, alias Jana-
natha-chaturvedi-m angalam (Nannilam taluk, Tanjavur) of a . d .
126477 records an agreement among craftsmen calling themselves
rathakarar, to raise a fund from am ong their num ber in specified
localities for the construction o f a pavilion in th at brahmadeya .78
71 D .D . Kosam bi, ‘The W orking Class in the A m arakosa , Journal o f O riental
R esearch, v. 24 (1955), p. 60.
72 E .I., v. 31, n o .l, ‘Two Salarikayana Charters from K anukollu (G udivada
T aluk, K rishna D istrict)’, p. 3.
73 A .R .E . 1909, para. 45, 479/1908.
74 The activities with which these craftsm en were associated were: architecture,
building coaches and chariots, constructing temple towers (gopuram ) with sculptures
on them , preparation of instruments used by Brahm ans in worship (i.e. ladle or
srik ), constructing pavilions (m andapam ), and making jewels for kings (ibid.).
75 Am ong those m entioned a re : Y ajnavalkya, G autam a, Kautilya, and B odhayana
(ibid.).
76 The term s used to define their ‘good b irth’ were, mahishyasa, karani, and
anuldm a. M ahishyasa. sons of K shatriya fathers and Vaisya m others, were the
fathers of the ra th a k a ra s; karani , daughters of Vaisya fathers and Sudra m others,
were their m others; thus, the father in each case was o f the higher varna, making
it a superior, anuldm a, m ixed-varna birth (ibid.).
77 Discussed by K.V. Subrahm anya Aiyer, ‘Largest Provincial Organizations
in Ancient India’, Q J M S . v. 45, no.2 (1954), pp. 77 ff, and Sundaram , op. cit., pp.
31 ff.
78 This record provides an excellent source of ancient locality names in several
m odern taluks o f m odern Tanjavur. These are listed in Subrahm anya Aiyer, ‘Largest
The fund was to be created from a special cess, inavari, upon
craftsmen, and it was to be collected by Saivite temple functiona­
ries,79 in the named localities. Am ong the signatories of the order
were carpenters and goldsmiths. It may safely be assumed, with
the epigraphist K.V. Subrahm anya Aiyer, that, the four classes
of artisans referred to in the Alangudi inscription were.pañchalar
or Kam m alar, the core group of the left division of that eariy time
and later.80 They had simply appropriated rathakárar myths
involving the god Visvakarma and his priest, K asyapa.81
A nother element of evidence linking those using the ancient
rathakárar title with left division artisans of the medieval period
is to be found in the identification of rathakárar as kil-kalanai,
subordinate professional people,82 who seemed to have lived in
separate residential quarters ( ceri ) in larger villages.83 An inscrip­
tion of a . d . 1036 from Chidam baram distinguishes between non-
Brahman inhabitants of superior status, kudigal, and those of
inferior status, kll-kalanai ,84 The kudigal included two merchant
groups, sañkarappadiyar and vyápárin,85 plus three groups usually
associated with the right division: Vellala, Saliyar (cloth merchants)
and Pattinavar (fishermen). The subordinate workmen, kTl-kalanai,
were taccar (carpenters), kollar (blacksmiths), tattar (goldsmiths),
Provincia! Organizations . .’ Q J M S , v. 45, no.2, p. 79.
79 Called vTra-máhesvaras in line 9 o f the record (ibid.).
so Ibid.
81 In the U rrattur record cited above, the idangai claim to be created to kill demons
who disturbed the sacrifices of Kasyapa (A .R .E . 1913, para. 39). According to
the B rahm ana texts, Kasyapa was the priest o f the god Visvakarm a: P.V. Kane,
H isto ry o f D harm asastra. v. 2, pt II (B handarkar Oriental Research Institute,
Poona, 1930-62, p. 840). Srinivasachari (op. cit., p. 80), cites references in the
Tondaim andalam S a ta k a m associating artisans with the god V isvakarm a; T hurston
(op. cit., vh, p. 362) also records m yths of the K am m alar in which their descent
is traced from V isvakarm a; according to a Nellore inscription of the eleventh
century, a grant to a temple was made by members o f the Visvakarma caste (A .R .E .
1954-5, no.35 or Inscriptions o f the Nellore D istrict, p. 818, no.50).
82 S .I.T .I., G lossary, p. xxii, kalanai and xxvii, kil-kalanai.
83 M entioned in Subrahm anya Aiyer, ‘Largest Provincial O rganizations . .
op. cit., p. 35.
84 E .I., v. 22, no.24, p. 146.
85 sañkarappadiyar are discussed by Subrahm anya Aiyer; the relative status of
vyápñrin is touched on in a fifteenth century Sanskrit work, Vaisyavam sasudhakara
(in V. Raghavan, ‘The ‘V aisyavam sasudhakara’ of Kolacala M alinatha', New
Indian A ntiq u a ry, v. 2 (1939-40), p. 444).
and kdliyar (weavers).86 The epigraphist Hultzsch, in discussing
some inscriptions of a slightly earlier period, a . d . 1013, noted the
term ‘kam m ânachëri ’ and hazarded that this was the residential
quarter of the kam m alar, or artisans; the propinquity of the arti­
sans’ quarters to those of the paraiyan , ‘p a ra ic c ë rf , suggested the
low status of the artisans.87 Thus, whatever the high status of the
rathakarar in ancient times, and notw ithstanding the use which
craftsmen of the Chola period sought to m ake of this ancient status
in assuming the title o f rathakarar, they had come to be identified
with middling and even poor rank in the eleventh century.
During the twelfth century, however, the status o f artisans and
merchants associated with the left division began to change. The
Polonnaruva (Sri Lanka) inscription o f the first quarter of the
twelfth century speaks of the idangai vëlaikkârar for the first time
and m erchant groups later to be m entioned prom inently.88 This
record suggests strongly th at the idangai vëlaikkârar were the
military arm of the merchants m arking the beginning of the rise to
prominence of the great itinerant guilds whose military power
was so conspicuous for the next two centuries.89 At the same time,
artisans o f the left division began to dem and and to receive privileges
which m arked an enhancement of their status. A series of ins­
criptions from the Kongu country during the late twelfth century
refer to the Kam m alar of vengâlanâdu (modern K arur taluk) who
claimed for themselves the right to use the double conch and drums
at times of marriages and funerals, to use footwear (ciruppu ), and
to cover their houses with plaster as a m ark o f their respectabi­
lity.90 The interpretation by D r M. Arokiaswami of these Kongu
inscriptions, is that valangai colonists o f the region, including
Vellalas and Kaikkolars, had oppressed those of the left-hand
faction until the intervention of the Chola ruler, K ulottunga III
was brought by the left division leaders there, the K am m alar.91
86 T am il L exico n , v. 2, p. 1197, kôlakan.
87 v. 2, 47n, 63n.
88 E .I., v. 18, no.38, lines 25-39 in which the valanjiyar are referred to as the leaders
of those making the grant to the temple of the T ooth Relic in Sri L anka; nagarattàr,
local merchants in contrast to the valanjiyar who appear to be. itinerant merchants
(nànâdësi), are also mentioned in this inscription. See the discussion by T.N.
Subram aniam , S .l.T .L , v. 3, pt 2, pp. 63-4.
8» S .I.T .I., v. 3, pt 2, p. 65.
90 S.1.1., v. 3, no.46; also see M. Arokiaswami, ‘Social Developm ents U nder
the Im perial Colas’, T .A .S .S .I., 1956-7, p. 5.
«> Ibid.
 Developments similar to these were taking place in A ndhra and
K arnataka as well as in Tamil country. During the eleventh and
twelfth centuries, m erchants and artisans of A ndhra attained
strikingly public presence for the first time, particularly in the rela­
tively densely settled parts o f the A ndhra plain called Vengi, com ­
prising m odem East Godavari, Krishna, and G untur districts. In
these places there are num erous temple inscriptions which record
gifts o f m erchants calling themselves, ‘the lords o f Punugonda’ and
often citing gotra nam es.92 itinerant merchants plying extensive
trade networks between K arnataka and A ndhra endowed temples
in these regions as well as in Tamil country. These endowments
are recorded in inscriptions which extoll the virtue, bravery, and
dharm ic pursuits of their members.93 Artisans o f K arnataka,
calling themselves Vira Panchala, had formed special relationships
with certain temples and seminaries (mathas) such as the Airiya-
kula-m atha in the Hoysala capital of D orosam udra (m odem
Halebid, Hassan district);94 artisans of Andhra, with the name
pahchanamuvaru, were also associated with particular temples of
the time and even referred to themselves as a corporate
group.95 Among the most self-consciously striving groups of the
time were the oil-mongers (teliki) of Bezwada and its vicinity.
They called themselves, ‘the one-thousand’, and in their records
of the eleventh and twelfth centuries boast of being the hereditary
servants of the Eastern Chalukya rulers of the area. According to
a copperplate inscription o f a . d . 1084, Teliki bridal couples were
given the special right to visit the king on horseback and to obtain
offerings of betel from his hand.96
Changes in the status of the left division people during the
twelfth century were dependent upon changes in South Indian
society. New im portance was accorded to urban artisans and m er­
chants as a result of the temple urbanization of the period. W ith
92 Sundaram , op. cit., pp. 62-4.
93 Ibid.; also see Burton Stein, ‘Corom andel Trade in Medieval India’, pp. 50-1.
94 S. C handrasekhara Sastri, ‘Economic Conditions U nder Hoysalas’, The
H a lf-Y ea rly M ysore University Journal , v. 2 (1928), pp. 216-17.
95 Sundaram , op. cit., p. 26, in which these artisans also refer to themselves as
‘the Seventy-four com m unity’ and ibid., p. 30, in which they are associated with
the Kam atesvara temple. Also, N arasim ha Rao, op. cit., pp. 64-75.
96 The ‘Teki Plates of the 17th year of K ulottunga F, according to S .I-T .L , v. 3,
pt 2, p. 68, which should g : /e the year as a . d . 1087; however Sundaram , op. cit.,
p.37, gives the date as a .d . 1084.
that came a fundam ental modification o f the position o f mercantile
and craft groups from that o f the previous centuries when they
were n o t only constrained to accept a subordinate place in relation
to the dom inant peasantry, but to suffer the indignity o f a corporate
stat us of pollution which is ineluctably associated with the left-hand.
The terms valangai and idangai and balagey and edagey literally
mean ‘right hand and left hand’ in Tamil and K annada. While
it is possible to attach excessive im portance to the simple positional
distinction of left and right, there is ever-present the taint of pollu­
tion owing to the use of the left hand in bodily functions. It is
also well to recognize the distinction of left and right at the level
of ritual. The designation of left-hand has been attached to corrupt
or perverse forms of worship called vamTs (from vama: ‘left-side’
or ‘reverse’) which are secret Tantric ritual forms o f Siva worship.97
This distinction is illustrated by the early sixteenth century story
of how the Saivite teacher Appaya D ikshit o f Kalahasti sought
to discredit a Vaishnava teacher of the Tatacharya family by
accusing him of having given a blessing to the Vijayanagara king,
Achyutadevaraya, with his left hand as he did with people o f the
lower castes.98
Various explanations have been offered for the left-and right-
hand designations. M acleane’s suggested that the ‘hands’ imagery
arose from the fact of five artisan groups o f the left division, the
panchalar or Kam m alar, as opposed to five non-artisan castes —
i.e. as fingers on a hand. This is clearly unacceptable, for the
num ber here, ‘five’, is as conventional as the num ber ‘ninety-eight’
which is used for each division in m any o f the sources.99 The
more usual explanations about the right and left hand are posi­
tional: people of the valangai being on the right-hand side of
gods, sages, Brahmans, or kings in some legendary context
in which status was determ ined.100 G. Oppert appears to have
been the first to notice the implication of ritual pollution in the
91 H.H. Wilson, E ssays and L ectures in the Religion o f the H indus, L ondon,
1862, pp. 250-61 and M. M onier-W illiams, H induism , London, 1906, p. 126. M onier-
Williams, A S a n skrit-E nglish D ictionary, p. 941.
1,8 R. Sham a Shastry, ‘Sivananda’s Life of Appaya D ikshit’, Q .J .M .S ., v. 11
(1920-1), pp. 116-17.
99 M a nual o f M adras A dm inistration, v. 1, p. 69.
10° These positions are mentioned in various sources; see Srinivasachari, op.
cit., p. 85 and A .R .E . 1913. para. 39.
right-and left-hand division. He attributed the dual division with
its pollution implications to the conflict between Jainas and Hindus
during the pre-Pallavan period. In this connection, Oppert stated :
‘The influence of Jains was perhaps strongest in the towns where
artisan classes form an im portant portion o f the population,
while the Brahm ans appealed to the land owning and agricultural
classes.’101
While Brahmans remained neutral with respect to the divisions,
Jainas were apparently associated with the left-hand division,
edagey, in K arnataka until a . d . 1368 when the Vijayanagar ruler,
Bukka Raya, intervened in a dispute involving Vaishnavas and
Jainas over sect emblems and decreed that Jainas were to be con­
sidered members of the right-hand division.1«2 As suggested above
in the discussion o f the Pallava period, the nominally religious
conflict which was bitterly carried out during that early period
was based upon im portant ideological factors; U nder the Pallavas
and their peasant and Brahm an supporters, Jainism was treated
as a dangerous error, and association with Jaina teachers and
institutions polluting. O ppert pointed to a similar orientation of
the Chola rulers tow ard the Jainas and the Jaina-supporting
H oysalas.103
Considering the stigma o f pollution which attached to the left-
hand, it is to be wondered that those of the idangai would have
acquiesced in the title. T hat they did is clear from the U rrattur
inscriptions cited above and num erous other inscriptional and
literary documents of the idangai which are to be found in all
parts of the m acro region. There are num erous references to
regular local dues collected from the left-hand people as idahgai-
vari as well as subscriptions collected from and on behalf o f the
division, the idahgdi-magamai .104 During the earliest period for
which there are records, it appears that the idangai occupied an
inferior and perhaps despised position am ong people o f the region.
Later, in the twelfth century and after, when idangai groups under­
took to alter their positions with respect to the dom inant peasant
population, they continued to identify themselves by the idangai
title notw ithstanding that this title might have originally been a
101 Oppert, op. cit., p. 62.
102 H ayavadana R ao, M ysore G azetteer, v. 2, pt 3, p. 1483.
103 Cited in Srinivasachari, op. cit., p. 84.
104 A ppadorai, op. cit., v. 2, p. 729; S .I.T .I., ‘Glossary’, p. xxxii.
sign o f their degraded status in the society o f the m acro region.
The title was retained until the nineteenth century in most places.
A part from Oppert and, recently, Arjun A ppadurai, few m odem
com m entators on the right-left divisions in South Indian society
recognize the signification of impurity or pollution which accom­
panies left sym bolism ;105 none— including these tw o— addresses
the question o f why those of the left division accepted its derogatory
designation. A ppadurai notes: ‘As in other cultural systems, the
left-hand in South India has connotations o f impurity whereas
the right-hand has powerful and positive norm ative associa­
tions. . . .’106 Where Oppert explains the designation as origi­
nating in the success o f the Brahman-led Hinduism o f agricul­
turists over the ‘heretical’ Jaina artisans and merchants o f pre-
Pallavan trade centres, he does not ask why the presumed deni­
grating ‘left’ title persisted and, especially, why those o f the left
division continued to use it in their own records later. N or does
A ppadurai’s thoughtful and bold explanation of the division as
a ‘root paradigm ’ of conflict107 permit us to understand finally
why, if, as he says, the left-hand connotes impurity, those o f the
idangai use th at title.
A possible explanation of this puzzling phenom enon is that the
utility o f the idangai title as a well-established symbol o f identity
outweighed for its users o f the C hola period and later any stigma
which might have attached to the title from an earlier time. It
is after all not only in the labels which are affixed upon or chosen
by a group that basic significance inheres, for new myths can be
m ade to offset older meanings. The proud adoption o f the label
‘Slav’ (from ‘slave’) in nineteenth century Europe and the more
recent use of the label ‘Black’ in American society remind us of
this property o f ethnic labels and labelling. The capacity o f ethnic
labels to serve as symbols of identity and m obilization — whatever
the origin o f the labels and their possibly once derogatory
105 Arjun A ppadurai, 'R ight and Left H and Castes', p. 221.
11,6 Loc. c i t . .
107 The concept of 'root paradigm ’ is taken by A ppadurai from Victor Turner:
‘ . . . these are consciously recognized cultural models which emerge during the
life-crises of individuals or groups, and have reference to the social relationships
of those involved, as well as to the cultural, ideological or cognative patterns which
incline them to alliance on divisiveness. . . . As a root paradigm in South Indian
history, the function of this particular m etaphor is to give expression to a wide variety
of empirical conflicts, anomalies and antagonism s’ (loc. cit.).
connotations — explains as well as any reason why the title ‘left’
or ‘left-hand’ continued to be used by a substantial num ber of
South Indians even after the twelfth century when those using the
title found impressive new opportunities and importance.

Sum m ary and C o n c l u s io n

The span of the dual division o f lower castes in South India

extended over eight centuries. Only a brief part o f that span
has been examined here, and this early phase o f its development
may not have been its most im portant phase. Evidence from the
late eighteenth and early nineteenth centuries suggests that the dual
divisions provided an essential means by which many lower castes
in South India im proved their means of coping with the extra­
ordinary changes attending the development o f British institutions,
particularly those o f towns and cities.
However, urban contexts are o f little consequence during most
o f the Chola period, and, to the extent that urbanization became
im portant in the late C hola period, it contributed to the weakening
o f the state and to changes in the society o f the m acro region
previously within its authority. It is especially during the Vijaya-
nagara period and, later, during the British period, that towns
assume an im portance not known before, except, perhaps, during
the Classical age o f the Cilapadikaram. O f this most early period,
only speculation is possible and that is the subject of an appending
note on origins o f the dual divisions.
During the Chola period, several characteristics distinguish the
divisions from their later history and from other elements o f
Chola society apart from the rural context in which they operated.
Among the most im portant are the military and colonization
activities o f each o f the divisions. The eleventh century Avani
inscription o f the Tondaim andalam valangai from m odern Kolar
district and several idangai records from modern South Arcot
pertain to these matters. Each portrays conquest groups who have
successfully installed themselves in new territories and exercise
dominance over them. In these inscriptions and others, another
function o f the divisions was the assimilation o f persons previously
beyond the expanding frontier of the sedentary agricultural order.
This is attested in the two thirteenth century inscriptions from
U rrattu r in m odern Tiruchirapalli and V aranjuram in modern
South Arcot. Both places would appear to have been zones of
recent agricultural expansion in the thirteenth century; both were
located on the edges of older riverine settlement areas. Such
newly assimilated folk did not gain new identities, but, having
been granted appropriate insignia, were associated with others in
the loose alliance structure o f a division — usually the left division
— and thus eligible for the support o f others o f the division when
that was required. They also became eligible for participation in
the occasional military forays which expanded the land under
regular cultivation and settlem ent; they finally supported temples
by occasional cesses and subscriptions (vari and magam ai) and
thus became eligible for participation in forms of temple worship
and perhaps temple honours from which they would otherwise
be excluded.
These functions were occasional; the divisions were not absolute,
corporate, or continuous, but potential groupings. This conception
is adm ittedly an unexpected one and jars the presuppositions
which any student o f H indu society brings to the study o f the
subject. For, the divisions o f the right and the left are not only
occasional in their organization, and functions, but they also appear
to be without the internally ranked segmentation characteristic of
Hindu institutions. They resemble sodalities or sectarian groupings
in these respects, except that there is no core continuity as might
be provided to cults by joint worship at specific tutelary or even
canonical shrines, nor is there the evidence, except in post-Chola
times, o f an acharya or a guru who might serve as leader of the
divisions in particular places. "Hie divisions resemble more the
m odern socio-political caste associations in linking spatially dis­
persed groups possessing some shared ethnic identifications or
interests into active associations for occasional cooperation. Like
the m odern movements, myths were prom ulgated by the divisions
to justify their existence, and, like the modern movements, joint
efforts may have involved conflict with others over rights to certain
symbols (e.g. thirteenth century artisans o f the left division de­
m anded and received the right to the use o f the conch and to
sandals just as the nineteenth century N adars claimed the breast
cloth). These joint efforts might also involve cooperation in the
support o f certain activities (e.g. medieval temple endowments
and m odern ‘temple entry’ rights, modern school hostels and
cooperative housing societies). The analogue o f the dual divisions
and m odern caste associations may be pressed further. In both
cases, public conflict with adversaries was resolved by the inter­
vention of legitimate authority. Chiefs, kings and Brahm ans play
an increasingly public part in the records o f the dual divisions
during the later Chola period ju st as British officials do in the
conflicts between the divisions from the seventeenth to the early
nineteenth century. In the latter cases, conflicts at times give the
impression o f being staged or prom oted by members of one or the
other division in order to produce intervention and legitimate
adjudication. This would seem to have been the strategy of caste
associations in more recent times, and ‘agitational politics’ is the
term that has come to be applied to the execution o f that strategy.
During the Chola period, these strategies can hardly be identified
from the relatively few and dispersed inscriptions, but the forms
do appear to be present, and they are to become clearer during
the Vijayanagara period under the intensified processes o f ur­
banization and political change.
In Chola times, the divisions o f the right and left present them ­
selves as linking form ations concerned with military activities,
expansion o f agrarian forms o f the age, and the assimilation of
new peoples to the society o f the age. Linkages were essentially
horizontal, that is the divisions brought into potential alliance
structures those groups o f neighbouring nadu localities whose y
interests were not represented in the dom inant nattar of the locality.
Lower agricultural groups, various artisans and traders of a g ri-,
cultural commodities, as well as field labourers could rely less I
and less upon the protection of their interests by the dom inant j
nattar o f their isolated localities. Even m ore hazardous was the
position of these artisan and m erchant groups whose economic
activities articulated poorly with the agricultural economy, whose
products were not exchanged in the nexus o f agricultural patron-
client relations, but exchanged m ore widely and impersonally;
whose economic fortunes, therefore, were often tied closely to
itinerant m erchants seen as little better than bandits by settled
agriculturists, and whose pasts were tainted by urban and heretical
connections which, added to everything else, made them objects
of suspicion and disdain.
To both such disadvantaged groups — lower agricultural groups
and mobile artisan-traders — alliances across ancient nadu lines
provided a measure o f security and political leverage with respect
to the notables o f each nadu , the nattar. These horizontal linkages
across nadu boundaries complemented and were probably the con­
sequence o f the vertical integration o f nadus during the time of
the great Cholas, R ajaraja I and Rajendra I, when the Chola
segmentary state was being perfected. The local beneficiaries of
this vertical integration were the m ost powerful of the nattar
who, in accepting the ritual sovereignty o f these great kings and
im itating their royal style, separated themselves as an increasingly
powerful, local ruling stratum . This political development is treated
below in ch.VII. However, even as the horizontal linkages among
the less powerful occupational groupings was occurring as a means
of fortifying their interests against the increased prestige and power
o f the nattar , the latter were themselves establishing horizontal
linkages with dom inant agricultural groups like themselves in
neighbouring localities, creating the periyanadu, or ‘greater nadu’’
which comes into historical view during the twelfth century. To
this and to related changes during the twelfth century attention
must now be given.

A N ote on O r ig in s

Discussion of the right and left divisions during the Chola period
has been shaped by the necessity of examining the fragmentary
evidence of this early period in the light of the more complete
inform ation of the seventeenth to nineteenth centuries. However
great the gaps in evidence and understanding o f the dual divisions
between the tenth and thirteenth centuries, these are as nothing
compared to the difficulties of considering the origins of the divi­
sions. The vagueness with which the dual social divisions may
be seen in the period from the tenth to the thirteenth centuries
may appear to make suggestions about origins particularly hazar­
dous. However, it is precisely because knowledge about the divi­
sions is so fragm entary at this time and even later that any analysis
m ust imply a set of notions about their origins. Such implicit
notions are best made explicit.
The development of the dual division of lower castes appears
as the consequence of two significant structural features of the
society o f the m acro region during its early history. These are the
territorial segmentation of society and culture, and the ambiguous
social status o f the non-Brahm an population given the commitment
to Brahmanicai ideology by the Pallava period. Early attem pts
by peoples of the macro region to cope with these structural features
produced a status system pivoting upon the power of the peasantry
of the numerous localities into which the southern peninsula was
divided. For a substantial portion of the population — those who
were not of the dom inant peasant groups — the status system of
the Pallava and m ost of the Chola periods was unsatisfactory,
and by the eleventh century, merchants and artisans along with
their urban and rural dependents — began to function indepen­
dently of and to some extent in opposition to peasant leaders o f
local society. This modification occurred without altering the
basic territorial segmentation of society or the ambiguous status
positions of non-Brahmans in South India, and the dual social
divisions of the eleventh century were to remain im portant for
centuries. M embership in, or at least association or alliance with,
established status groupings in an increasingly stratified society
was the principal m otivation for the division. The process of an
expanding agrarian system — partly by military means, partly by
peaceful means — created the need to assimilate new peoples;
this was an im portant feature of the Pallava and early Chola
periods. In this process, occupational and residential groupings
were the prime organizational loci. The expanding requirements
for the services of essentially urban-based merchants and artisans
provided the opportunity for the development.
The emergence o f groups identifying themselves as the left
division in the eleventh century does not preclude the possibility
th at there might have been an earlier foundation for such a division.
One such hypothesis is that the divisions represent peoples of
different racial origin within South India dating from ancient
times.108 Racial adm ixtures there were, but, apart from the
‘Aryan m igration’ which introduced Brahmanicai as well as Jaina
and Buddhist institutions, there is no convincing historical evidence
o f a significant, racially distinct stratification which could account
for the divisions as known after the tenth century. N or does the
view here deny a variant o f the ancient racial argument which
distinguished between ‘indigenous’ people of a culture area and
‘strangers’ who took up residence there. The latter hypothesis
is supported by occasional references‘in K arnataka where nadu
108 H utton, op. cit., pp. 166-8.
and dësa (for strangers) are co-equal with edagey and halagey. IC>9
Valangai vëlaikkârar are considered by Srinivasa Iyengar to be
‘Tam ilians’, whereas the idangai vëlaikkârar, according to him,
consisted o f warriors from A ndhra ( vadugan ), K erala (malaiyalar),
and others not o f the Chola heartland o f Cholam andalam and
Tondaim andalam .110 Another m anifestation of this ‘stranger’
versus ‘indigenous’ classification is in the epigram o f the low caste
o f Paraiyan, invariably associated with the right-hand division
in recent centuries : ‘the paraiyar are not o f the left hand, they are
Tam ilians’.111 While the conception o f ‘stranger peoples’ may
have entered into the dual division, the core elements of the divi­
sions, on both sides, m ust be considered as ‘indigenous’.
A nother view o f how earlier divisions in the society o f the macro
region m ay have served as the basis for the subsequent development
o f the right- and left-hand divisions appears to be emerging from
recent scholarship on the Classical or Sangam era. In the im portant
work o f N. Subrahm anian, Sangam P o lity,112 there are depicted
two, possibly simultaneous social orders. One was urban, cos­
m opolitan, trade oriented ; the other was ‘tribal’, rural, and relatively
simple in economic organization. The interactions between these
two disparate social orders are very unclear. Poets o f the one order
may not have been poets o f the other, and while certain cultural
continuities existed between the orders — in language and beliefs
at least — it is as yet difficult to see them as constituting a single
civilization. However, these two orders might have comprised a
single society as proposed by Subrahm anian, and this could
have established the basic fram ework for the later, historical
divisions.113
Such questions o f possibly earlier fissures which might have
formed the basis o f the later divisions cannot be considered in
greater detail at this point ; the state o f our knowledge will not permit
that. The need for m ore archaeological and systematic literary
109 Rice, M ysore, v. 1 p. 224. An alternative form of nadu in this context was given
as pete.
no Srinivasa Iyengar, T am il Studies, pp. 106-7; other groups m entioned included
Pallan and M allar from Pandya country, Bedar from K arnataka, and Madigas
from Andhra.
111 O ppert, op. cit., 64-5 and 65n.
112 Op. cit.
i*3 See this w riter’s review of the Sangam P olity in The Indian Econom ic and
S o cia l H istory Review , v. 5, no.l (M arch 1968), pp. 109-15.
research is essential for carrying these speculations beyond the
present state. F or the purposes o f exploring the .origins o f the
divisions o u r starting place m ust be restricted to the period when
the evidence is superior to that afforded by racial drift theories
and the inform ation gained from the undated poems o f hundreds
of authors. It is to certain of the general characteristics of society
which were manifest during the Pallava period and after that
attention m ust be given. Am ong these is the high degree o f territo­
rial segmentation o f society and some o f the social structural
consequences o f the m ature Brahmanicai culture o f that time.
Territorial segmentation has referred principally to the isolation
of the many locality cores o f peasant population, nadus , scattered
over the Tamil plain. The degree o f isolation conformed with
physiographical factors to a significant extent; thus, in the central
Kaveri basin, there was less isolation, greater continuity o f settle­
m ent, than in the western parts o f the basin (parts o f m odem
Tiruchirapalli, South Arcot, and Salem) and in the tracts South
o f the Kaveri. In most o f Tondaim andalam , the central Tamil
plain, the nodes o f peasant settlement conform ed with physical
features, especially the drainage patterns of the principal streams
upon which tank storage o f irrigation water depended would have
been especially im portant. Judging from the distribution of ancient
inscriptional and modern records referring to the right and left
divisions, the greater the degree o f isolation among the peasant
cores of settlement, the m ore im portant and conspicuous was the
dual division. The central Kaveri basin appears never to have
developed the intensive divisional alignments found in the central
portion o f the Tamil plain where the dual divisions were both an
early and persistent phenom enon with Kanchi serving as the centre
for both.
By the late Pallava period, and certainly in the Chola period,
Brahm an groups were virtually closed, priestly corporations dom i­
nated by Smartas o f various divisions. M ost prestigious, perhaps,
were the Vadam a Sm arta Brahmans. Their names suggest a north­
ern origin ( vada means ‘north’), but the title, vadama could also
refer to the proficiency in Sanskrit and vedic ritual, which then and
earlier were associated with the north. O ther Sm arta Brahmans
were divided into territorial sub-divisions.114 There was also a
114 These divisions are listed in various places (sec Thurston, op. cit., v. 1, p.
333 ff.).
smaller group o f Brahmans o f Vaishnava persuasion, the Vaikana-
sas; it was not until some centuries later that Vaishnava Brahmans
began to constitute a somewhat larger proportion o f the Brahman
population, and at that later time, the num bers were primarily the
result o f a shift of Smartas to the fold o f the dynamic and expanding
Sri Vaishnava sampradayas. Increasingly secure and possessing
great secular authority in rural settlements over some o f which they
enjoyed complete control, the Brahm an m onopoly over higher
sacral functions was firm and was to remain so until the thirteenth
century. Brahm ans were unchallenged in this sacred or secular
authority either by rival religions o r by warriors determined or
disposed to diminish their role. Brahmans thus constituted a strong
and impenetrable strata o f the contem porary social system. No
less clearly defined in this social system were those at the bottom.
Enjoying neither the high ritual status nor the status provided
by holding land or possessing skills essential to the maintenance
o f peasant agrarian operations, agrestic labourers occupied the
unambiguously lowest strata o f contem porary society. With the
expansion o f wet rice cultivation based upon secure irrigation sour­
ces, the num bers o f such persons increased. Partly, these increases
represented the assimilation of agricultural groups who had
previously cultivated lands deficient in reliable irrigation sources.
Such peasants were situated at the margins o f productive peasant
agriculture. W hen these m arginal tracts were brought under
irrigation by the expansion of prosperous and powerful peasant
groups with requisite organization, capital (mostly in the form of
superior skills but also livestock), adequate manpower, and superior
military power, the former occupants of these lands were reduced
to labour dependents of the expanding peasantry, or they fled
to still more m arginal tracts only to be incorporated later or forced
once again to flee. A nother source o f this lowest tier o f social
groups were those o f the forests to whom regular sedentary agri­
cultural pursuits were unknown. W hen forests were felled by
expanding peasant agriculture, as they were throughout the Pallava
period, the fate o f the forest dwellers was the same as that to whom
agricultural labour was already a fulltime though hazardous
basis of livelihood.
A part from the degraded status which attached to those who
laboured on the fields of others and were therefore w ithout sub­
stantial rights or means for am eliorating their conditions except
the threat o f absconding, there were those whose purchase in
peasant social organization was even worse, by a slight degree,
because they combined some polluting craft with their principal
agrestic labour. Such were the leather workers, Sakkiliyar and
Madiga, for example, whose low occupational status as field
workers was negatively reinforced by their work with leather and
the preparation o f hides. Others included musicians and dancers
who constituted part o f the corps of bards in the classical period.115
Between the poorest field workers, artisans and artists and the
highest strata o f priests were the m ajority o f the population. During
the recent, m odern past, ranking pressures have been the m ost
severe at this level o f society consisting o f powerful land-controlling
peasant groups and wealthy merchants, bankers, and artisans.
Beneath these has been a second order of peasants who lacked the
means to support a claim to being dom inant peasants , pumipputirar,
‘sons o f the soil’. Finally, there have been many kinds o f village
artisans and service groups (as washermen, barbers, and potters)
whose work was tainted by a not always logical set o f pollution
norms. Among such varied groups, the terminology dictated by
Brahmanicai usage, such as ‘Sudra’, serves no analytical purpose,
and, in the recent past, that label has been rejected by those upon
whom it was placed by other Indians or by British adm inistrators.
The m odern term ‘non-B rahm an’, comes closer than any other
to encompassing the m iddle groups of the early period in the sense
o f being ‘respectable’, socially mobile, and yet clearly neither o f the
highest nor lowest strata. To the m odern ear, however, ‘non-
B rahm an’, is difficult to disassociate from the twentieth century
context when the term was claimed by educated and politically
mobilized groups of Tam il N adu and A ndhra bent on displacing
Brahmans from w hat they regarded as places o f disproportionate
advantage. Still, the term ‘non-B rahm an’ is m ore appropriate
than the varna terms Vaisya and Sudra in the South Indian context
though the latter terms occur in ritual manuals (agama ) and ins­
criptions occasionally. In fact, there appears to be no generic
term for those beneath the status o f Brahmans in the medieval
period or earlier. Sectarian term s such as cattatavan and saiva,
denoting votaries of Vishnu and Siva who were not Brahmans,
115 S ee George L. H art II!, ‘Ancient Tamil L iterature: Its Scholarly Past and
Future’, in E ssays on South India, ed. Burton Stein, University Press of Hawaii,
Honolulu, 1975.
are too narrowly circumscribed in their reference. In the corpus
o f South Indian inscriptions, there occur num erous specific ethno-
occupational groups below the status o f Brahm ans such as:
adavimar, ayogavar, kaikkdlar, saligar (weavers), alavdr, parampar,
vellan (cultivators), ahjuvannam, kam m alar (artisans), pattinavar,
bharatavar (fishermen), davana-chetti, teliki. (merchants), Tlavar
(toddy tappers), kannakkanan (brazier), mannan , vannan (washer­
men), manradi (shepherd), navisdn (barber), taiyan (tailor); and
vetkdvan (potter).116 The terms idangai and valangai are themselves
references to general classes o f non-Brahmans, except that the
dual divisions included among their m ost active members those
of very low status.
It seems evident that there was little o f the rank striving and
conflict am ong local ethnic groupings that one finds in contem porary
South India. The nineteenth and twentieth century claims o f some
middle groups to the status o f Brahm ans may have occurred
earlier; Velialas and Reddis, dom inant peasants of the m odern Tamil
and A ndhra plains, occasionally equated their control over land
with the rights o f Kshatriyas. Such claims were as irrelevant
in an earlier age as they were during the nineteenth and twentieth
centuries. R anking postulates are not based upon unilateral
claims, but upon interactions am ong claimants with each other
and with those considered the lowest and m ost polluted, on the
other side.
From the earliest references to the dual divisions, their com ­
positions appear to have included a wide range of groups which in
m ore recent times m aintain punctilious differentiation and relative
ranking. Yet, the diction o f the idangai and valangai inscriptions
and the administrative and judicial records of the seventeenth to
the twentieth centuries give no im portance to relative rank within
the divisions. On the contrary, the divisions represent themselves
or are represented as pluralistic collectivities enjoying an apparent
equality as in various kinds o f sodalities. The critical factor in the
divisions o f this early period and later was not relative ranking
am ong the constituents, but shared substance and interests. It
was not ra n k , but membership within a division which was im­
portant. Only the very powerful or very marginal could claim
116 These are listed in S .I.T .I., Glossary. On p. iv, he suggests the term agambacji-
udaiyan as ‘citizen’ or ‘subject’ with which were associated other terms. Also see
Tam il L exicon, v. 1, p. 12; v. 2, pp. 970, 968, 974.
or afford a position o f neutrality (e.g. the status o f m adhyasta in the
Baram ahal Records) with respect to the dual-alliance formations.
The powerful nattar were in the best position to separate them ­
selves from others o f the ambiguous middle strata. Their economic
control and military authority within the framework o f the Chola
overlordship, their claims to ancient respectability and political
primacy, and their close relations with Brahm ans all combined
to effect this separation. To those less favourably placed, but still
im portant in agrarian relationships — peasant groups of lesser
status and certain merchants and artisans — the right division pro­
vided alliance support o f importance. Craftsmen and merchants less
directly involved in agrarian relations sought alliance refuge in
the left division, for in the balancing o f status and locality solidarity
which hinged upon links to the land, they had but poor claims.
Moreover, to the extent that such craft and mercantile activities
could be identified with the urban social order, they bore an addi­
tional status handicap of having been associated with heterodoxy.
A core of military forces stood ready to support the latter coalition
in the form of armed contingents of itinerant traders.
The supra-local system o f dual divisions became evident first
in the valangai military units o f the first o f the great Cholas, R aja­
raja I. At that time there appears to have been no contending
idangai. The designation valangai at this earlier time may have been
the way of referring to the armed peasantry of Cholam andalam
and Tondaim andalam who had overcome a considerable measure
of isolation and had begun to cooperate militarily, first under
various chiefs and later under the great Cholas. They had become
the weapon, the ‘right-hand’, of ambitious warrior leaders. These
were potential or relative structures, not absolute and continuous
ones, as suggested in the term vêlai, in velaikkarar, meaning
‘occasional’. It was not until two centuries after the first references
to essentially localized peasant groups collaborating on wider
regional lines for military purposes that the left division appears
to have achieved the same degree of supra-local potential organiza­
tion. Venkayya's view th at the existence of the valangai soldiers
in the time o f R ajaraja I implied the existence of the idangai would,
under this interpretation, be questionable. It is more probable
that the prior existence of the peasant-dom inated right division
led to the development of an opposing division at a later time,
that is during the eleventh century. ■
 Removed from the great brahmadeya centres of orthodoxy of
the tenth and eleventh centuries, which afford the most im portant
sources o f inform ation of this period, the rise of the left-hand
division and the establishment of a true dual division was slow to
be registered. Mobile artisan/trader groups, even after they had
shifted their allegiance from the heterodox to the Saivite orthodox
faith of the era, probably continued to be held in suspicion. How­
ever, such groups could claim an ancient and honourable past
during which they enjoyed respectable status if evidence such as
the Classical epics, Ciiappadikaram and M anim ekalai are to be
credited even partially and if the rathakara connection was accepted
widely. Neither these artisan/trader groups nor the itinerant
traders with whom they were linked would long have willingly
accepted the low status which had befallen them in many agrarian
settlements. They would, accordingly, seize upon the new oppor­
tunities of the twelfth century and after to alter that status. Among
the most im portant o f the opportunities were those associated
with temple development in that age.
 The Transition to Supra-local
Integration in the Twelfth and
Thirteenth Centuries

Significant structural changes occurred during the late twelfth

and thirteenth centuries causing instability to the agrarian order
consolidated during the reign o f R ajaraja I. These changes include :
(a) the emergence o f assemblies called the periyanadu acting over
an enlarged locality and signalling diminished isolation among
previous nuclear regions, or nadus, and, at the same time, augment­
ing power in the hands o f supra-local leaders o f the enlarged
locality; (b) the integration o f portions of the m acro region which
had previously been marginal in settlement and im portance, but
which now emerged as m ature agrarian regions including much
o f the interior area o f m odern South Arcot which had earlier
separated Tondaim andalam from Cholam andalam , and the upland
tract of Kongu comprising m odern Salem and Coim batore; and
(c) the emergence o f a new tier o f centres o f civilization — towns
— which served to integrate the enlarged localities o f the period
and displaced the earlier civilization centres, brahmadeyas, in
both sacred and secular functions. These changes — essentially
a new ordering o f the elements o f the earliest period o f agrarian
integration — provide the clearest explanation o f the decline of
the Chola state and the establishment o f the political culture o f the
Vijayanagara period. However, supra-local integration o f the
twelfth and thirteenth centuries was transitional in the sense that
earlier forms continue to be im portant for a considerable time
longer.

Supra-loca! assemblies
Beginning in the late twelfth century, there is evidence o f impressive
integration of local, nadu institutions at higher spatial levels than
ever before achieved. These supra-local institutions were first
commented upon by the senior epigraphist and historian, K.V.
Subrahm anya Aiyer, in a series o f articles entitled, T h e Largest
Provincial Organizations in India’, published in 1954-5.1 Most
recently, Subbarayalu has stressed the importance o f these supra-
local assemblies:
. . . from the beginning o f the 12th century, the Nadus . . . at times
transcended their territorial limits through assemblies of the Chittrameli-
Periya-Natfar. The Periya-Nattar were none other than the Nattar
of the whole [of any section o f the] country assembled together. . . ,2
It is well to note th at the inscriptions upon which the discussions
o f Subrahm anya Aiyer and Subbarayalu are based were known at
the same time as those o f the mahasabha institutions, having been
collected during the early years o f systematic epigraphical activity
in South India, around 1890. Yet, the supra-locai assemblies never
attained the same significance am ong South Indian historians as
did the brahmadeyas.
The term, ‘supra-local assembly’, is used here for periyanadu
in preference to th at used by Subrahm anya Aiyer, ‘provincial’,
because o f the implication in the latter usage o f an intermediate
l e v e l o f the state between the m onarchy and the locality. Subrah­
m anya Aiyer explicitly associates concrete, centrally-directed,
adm inistrative functions with these assemblies. Referring to a
periyanadu which convened at M annargudi (Tanjavur district)
in a . d . 1288, he states that they are seen ‘. . . exercising t h e powers
o f the State in three [ways]: the levying o f fees on articles o f mer­
chandise, the assignment o f revenues so obtained to the temple,
and causing the engraving o f the transaction on the temple w all’. 3
While a short time later in his discussion he seems t o deny t h e
above assertion, stating that it was ‘a popular assembly . . . not
one appointed by the State’,4 there are other places subsequently
1 Q .J .M .S ., v. 45 (New Series), n o .l, pp. 29-47; no.2, 70-98; no.4, 270-86; v. 46,
n o .l, pp. 8-22.
2 Subbarayalu, op.cit., p. 36.
3 Q .J .M .S ., v. 45, n o .l, p. 35.
4 Ibid., p. 35; the padinen-vishayattom ‘was . . . a popular assembly . . . not one
appointed by the State’. He later states that the ‘provincial assem bly' ‘functioned
in place o f the king in the country’, ibid., no.4, pp. 283-4, 286. Subrahm anya Aiyer
cites in support of the latter contention the phrase rdjyapari-pdlakdndm , but this
term means ‘protector of the country’ as well as king or ruler ( T am il L exico n , v. 4,
p. 2515).
where the same idea o f the assemblies as instruments o f a central
government is again expressed.
The assum ption that various kinds o f supra-local assemblies
o f the late twelfth and thirteenth centuries were interm ediate levels
of a centralized state structure is rejected here. In arguing that
these assemblies were the instrum ents of the centralized m onarchi­
cal state, Subrahm anya Aiyer ignored the overwhelming evidence
o f a weakened Chola overlordship which m ost historians recognize.
N or is his assum ption supported by any evidence from contem po­
rary records, for nowhere do these records specify explicit delegated
functions. Rather, his position is based, somewhat inconstantly
as it has been observed, upon the identification of all administrative
and political functions with a putative centralized state structure.
South Indian inscriptions afford too many examples of payments
assessed by local groups upon their members for specific purposes
and the recording of such arrangements on stone to accept this
imputation.
The region com prehended by the term ‘supra-local', is vague.
It might have varied considerably in different places and at dif­
ferent times. In the late thirteenth century M annargudi inscription
mentioned above,5 and in two others from the same place, one of a
slightly earlier time and the other later,6 reference is made to
supra-local assemblies o f vishayattar (‘people o f the country’)
o f eighteen vishayas (padinen-vishayattom ). A series o f records
from the same place earlier in the thirteenth century, in the twenty-
second regnal year o f R ajaraja III, a . d . 1239, register the decision
o f an assembly which included the nattavar (nattar) o f five localities
(nadus) which are nam ed7 and can be located in the vicinity of
C hidam baram .8 W hether the vishaya and nadu territories are the
same is not clear; both term s may refer to localities, and they
often do.9 M oreover, whether the reference to ‘eighteen vishayas 1,
can be taken literally is highly doubtful, for ‘eighteen’, like ‘one
5 A .R .E ., 1897, no.89; S.1.1., v. 6, no.40.
s A .R .E ., 1897, no.95; S .I .L , v. 6, no.47, dated a . d . 1268; A .R .E . , 1897, no.90;
S .l .I . , v. 6, no.41, dated a . d . 1313.
i A .R .E ., 1897, no. 98; S .l .I . , v. 6. nos.50, 58; A .R .E .. 1897. no.96; S .I.I., v. 6,
no.48.
8 See ibid.
9 T am il L exicon, v. 6, p. 3744, visayam is given as ‘territory, region o f country’;
see ibid., v. 4, p. 2210, natu is given as ‘country, district, province' or ‘locality’ (v. 4).
hundred and eight’ or ‘one thousand and eighty’,10 appears to have
been a conventional number, one o f the several sets which serve
to cast doubt upon m uch detail in Indian epigraphy. Dikshit has
noted the appelation ‘eighteen districts’ (K an n ad a: hadinentu
vishaya or hadinentu nadu) in G angavadi; he and D. Desai have
also drawn attention to the convention in K arnataka o f referring
to a complete village as one in which the ‘eighteen social groups’
or ‘castes’ (K annada: hadinentu sam aya) were present.11 One
cannot be absolutely certain, therefore, that at M annargudi the
scope of the supra-local assembly grew from the five localities
m entioned in a . d . 1239 to eighteen in a . d . 1288; however, it is
certainly possible th at a relatively small grouping o f localities of
the early thirteenth century expanded to include others and came
to be regarded as a relatively stable circle o f cooperating localities
as seems suggested in the term , ‘eighteen vishayas ’ used in the
M annargudi inscription and in those o f other places.12
From extant epigraphical evidence alone, supra-local assemblies
dealt with only a few functions, though there is the assum ption of
stable forms o f cooperation am ong constituent locality institutions
involving other functions. M ost o f the records of supra-local
assemblies relate to temple endowments, of course. In the M annar­
gudi inscriptions referred to above, agreements are recorded among
local merchants (nagarattdr ), presumably in the areca nut trade,
to present a sum of money to two temples o f the place in order to
support ritual offerings and repairs. These grants were to be raised
by a duty on the areca nuts as they passed through octroi posts
around the to w n .13
10 Kanchi is supposed to have had 108 Siva temples and eighteen Vishnu shrines
according to P.V. Jagdisa Ayyar (South Indian Shrines, M adras, 1920, p. 21);
there are supposed to be 108 Vaishnava pahcaratra samhitcis, Farquhar, op.cit.,
p. 182; 108 is the usual num ber of Brahmans upon whom brahm adeyas were con­
ferred, and 108 pots o f water were appropriate for the bathing o f temple images.
K.V. Subrahm anya Aiyer, H istorical Sketch es . . . , v. 1, p. 339.
11 G.S. Dikshit, L o ca l Self-G overnm ent in M edieval K arnataka, K arnataka
University, D harw ar 1964, pp. 34, 72-3; D inkar Desai, The M ahdm andalcs varas
under the C alukyas o f Kalyani, Indian Historical Research Institute, Bombay,
1951, pp. 342-3.
12 The a . d . 1234 record from N arttam alai in Pudukkottai contains the term
padinen bhum i according to A .R .E ., 1904, no.364 in I.M .P ., v. 3, p. 1640. The Avani
record of K ulottunga I, a . d . 1072 refers to the inhabitants of the eighteen vishayas
(E .C .. v. 10, pp.86-7 and K.V. Subrahm anya Aiyer, H istorical Sketch es . . . . v. 1,
pp. 351-2).
13 Two o f these posts are mentioned in the inscriptions.
 In what way the vishayattar o f the eighteen vishayas are involved
in this grant is not clear; they appear neither to be donors nor
donees. Instead, the vishayattar seem to assume the position of
general responsibility for the final agreement to establish the new
duty in support of the gift to the temples. This view is based upon
references to the temples whose deities bore the name o f the supra-
local assemblies, thus indicating the supervisory or protective
function which the assemblies exercised in relation to the temples.14
In a classic inscription o f a supra-local assembly from Tirukkoyilur
(modern South Arcot), the deity, Bhumidevi, (‘Goddess o f the
Earth, or Soil’) or, ‘Chittiram elivinnagar of the eighteen lands
(.bhumi ) and seventy-nine countries (nadus)’, was the beneficiary
of a grant from the citram eli-periyandttdr :15
Lord Hari. H ail! Prosperity! This is the order [sdsanam] of the chitra-
raela, just and good to all people, being the prosperous sons of the soil
and those subsisting on cow’s milk. Let this order, which is for the
protection and strengthening o f the sons of the land, who are bom of the
four castes, endure in this world. We the Chitrameli-Periyanattar,
who are the sons of the earth goddess [Bhumidevi] who have studied the
pure language [centamil] and the northern knowledge [i.e. Sanskrit
knowledge], who have become adepts at heeding the laws, who are sons
of the goddess of wealth whose flowers never wither, who are the lights
to all quarters, who deal with sweet words in the case of the good and
exorcise evil with harsh words, who prosper in this wide world bounded
by the four seas, with the lord of the winds blowing gently, the lord of
waters [Yaruna] showering the water, the lord of the heavens illuminating
all, and the people of all the quarters seated in peace (with the lands filled
with coconut, jackfruit, mango groves, plantain, areca nut, sweetly
scented flowers on the creepers, and the birds flocking enchantingly)
increasing in numbers without waning, with justice prospering and in­
justice diminishing, with their fame spreading and their enemies capitulat­
ing, with their mace of authority [cetigoi] preceding them in all quarters,
having the ploughshare as their god [citrameU\, with golden fertility of the
full purse as their goal, conducting the affairs of their organization with
tolerance and sympathy, having high and true justice as the source of
their towering fame, having fully learned the Muttamil m alai,1^ make this
benefaction. As even previously the gate decorated with the plough

14 See below.
15 S .l.I-, v. 7, n o .129; A .R .E ., 1900, n o .1)7. On paleographical grounds, this
inscription has been dated in the first quarter of the thirteenth century (personal
com m unication from K.G. Krishnan, Office of the Epigraphist for India, 16 Oct.
1968).
16 T am il L exicon, v. 3, p. 1757; probably m uttam il m a ra i : Kural and bhakti
hymns.
 [mejfl and the Goddess Bhumidevi had been installed and as such the
temple of the god in Chittiramelivinnagar o f the eighteen lands [bhumi]
o f the seventy-nine countries [nadus] alias the PerumaJ who is pleased
to stand in Tiruvidaikali in Tirukkovalur [Tirukkdyilur] in Kurukkai-
kurram in Miladu alias Jananatha-valanadu, had become the respon­
sibility o f our organization and as the endowments we had already made
were lost during the calamities, we again endow a padakku of paddy per
plough and a kuruni o f paddy per person including those from outside
our organization.

The opening lines o f this Tirukkoyilur are in Sanskrit in contrast

to the rest o f the record which is in Tamil. The Sanskrit portion
constitutes w hat Subrahm anya Aiyer has called, ‘. . . the charac­
teristic feature o f [the] Provincial Assembly in charge o f A dm inistra­
tion and A griculture’.17 While the purported delegation o f adm i­
nistrative functions of these assemblies as seen by this learned
epigraphist may be questioned, their association with agriculture
is beyond challenge. The names m ost often taken by such supra-
local bodies involved the words for ‘plough’ (Tamil and K annada:
m eli\ Telugu: m edi), as in the title, citram eli-periyanattom in the
inscription above, or in two twelfth century northern K annada
records where the term ‘m e ji- 1000’ (me/i sasirvaru) occurs.18
Such records also frequently display a decorative element which
further supports the association with agriculture; this was an
ornam ental arch with the representation o f a plough incised in
the slab bearing the inscription.19 The titles o f the groups under
whose auspices the inscriptions are prepared — periyanattar ,
m eli- 1000, and the assembly o f the eighteen districts of the four
quarters (nangu-tisai padinen vishayattdm ) further support the
close association of the assemblies with agricultural groups, the
nattars, who dom inated land relations.20 In some o f the inscrip­

17 ‘Largest Provincial Organizations . . .’, Q .J .M .S ., v. 45, n o .l, pp. 44-5; the

verse is given in ibid, p. 70 and in A .R .E ., 1953-4, p. 6.
18 A .R .E ., 1953-4, p. 6, provides a series of such records. Also see ‘Largest
Provincial O rg a n iz a tio n s ...’, Q .J .M .S ., v. 46, no.2, p. 71 (text) 1. 7. Further
references to these organizations in K arnataka are contained in A .R .E ., 1958-9,
nos.519, 665, 666; A .R .E ., 1960-1, p. 25, relating to a record of the time of
Vikram aditya VI (a.d. 1076-1126).
19 'Largest Provincial Organizations . . Q .J .M .S ., v. 45, no 1, p. 45 and A .R .E .,
1953-4, p. 6. The epigraphist, T.N. Subram aniam , reports a copperplate inscrip­
tion of the late thirteenth or fourteenth century dealing with the citram eli and
shaped as a m iniature plough ( S .l .T .l ., v. 3, pt 2, p. 217).
20‘Largest Provincial O rganizations. . .’, Q .J .M .S ., v. 45, n o .l, p. 47; Tam il
L exicon, v. 3, p. 1867 ( tisai).
tions, the constituent locality representatives, nattar, and their
assemblies, nadu , are referred to explicitly. An inscription from
Nellore, dated a . d . 1197, reads21 as follows after the standard
introductory verse:
In the Saka year 1119, the inhabitants of [the following] nadus: Pedai-
nadu, Peratti-nadu, M ungalarattai-nadu, Kadaiyasinga-nadu, Purigai-
nadu, Tongai-Punul-nadu, Chhicaia-nadu, Pottappi-nadu of Jayankonda-
cholamandalam having all assembled at Chittirameli-mandapa [pillared
hall] at Tiruparkadar-Chittiramejivinnagar, gave to the god at Chittira-
melivinnagar, free of tax, lands to the extent of 2,200 kuli.
The record from M annargudi dated a . d . 1239 mentions land
controlling groups from five nadus who met as a body to settle
complaints by some of their num ber against the demands of the
Brahman sabha of M annargudi (Rajadhiraja-chaturvedim anga-
lam).22 An a . d . 1234 record from T iruppattur (Tirupidavur, in
Musiri taluk, Tiruchirapalli district), refers to an assembly of
twelve nadus ,23 and from Varanjuram (Vriddhachalam taluk,
South Arcot) an epigraph dated a .d . 1227, already noted in
connection with the admission of peoples called nattam akkal
and m alaiyam akkal into the idangai, was issued on behalf of
eleven nadus.24
N o t all grants o f the periyanattar were to temples over which they
had supervisory responsibilities. From the shrine of Sri Govinda-
raja at Tirupati, there is the following Tamil record of a . d . 1235.25
Hail, Prosperity! This is the edict of Bhudevipiitra-Chitramela issued
for the maintenance of the dharma observed by the four varnas. As
per the oral order of the king issued previously,. . . we, the Periya-Naita-
var, having assembled in the council-chamber (attached) to the Tiruvi-
¡arikoyil [situated] in Tiruchchukanur in full numbers without omission
of the necessary adjuncts, and resolved upon the representation of Sri-
sathakopadasarpiUai, Kollikavalidasarpijjai, Aruvaraiyanaiyakoyilpillai
and Kalikarridasarpillai in respect of the provision (to be made) for the
amudupadi and sattuppadi (offerings) for (the image of) Tirumarigaiyalvar,
the bestower of blessings (on the devotees), who was installed in Sri
Govindapperumal’s temple which is a plastered sanctum of Vishnu,
through the charity of the Periya-Nattar, have witnessed that the Tiruk
21 A. Butterw orth and V. Venugopal Chetty, A Collection o f Inscriptions on
Copper-pkv.es and Stones in the N ellore D istrict, M adras, 1905, v. 2, p. 824 (at Nellore
town).
22 A .R .E ., 1897; S .l .I ., v. 6, no.50.
23 I.M .P ., v. 3, no. 185, p. 1532.
24 A .R .E ., 1940-1, no. 184.
25 T .T .D .I., v. 1, no.40, pp. 64-6.
 kudavürâr have granted with libations of water Kottakâlvày situated
in Tirukkudavür-nàdu to . . . yielding paddy . . . and direct that (its
produce) be amalgamated with and collected along with the income of
Sri Gôvindapperumàl by the supervisors of the treasury of the temple of
Tiruvênkatamudaiyân and that this charity be conducted as long as the
moon and the sun (last).
He, who obstructs [the conduct of] this charity, (will surrender) his body
to the N âttavar and will beget the sin of killing a tawny (coloured) cow
on the bank of the Ganges.
We, the Periya-Nâttavar (hereby witness this transaction). This is the
signature of the Periya-Nâttu-Vëlân.
This evidence of the continuity of the nattar, their assemblies,
and their cooperation within supra-local assemblies of the late
twelfth and thirteenth centuries is significant. It is im portant that
the latter, as new integrative institutions, did not replace the more
ancient locality bodies, but rather incorporated the latter, thus
achieving improved means for com binations across locality lines.
As such, the supra-local assemblies of the nattar, now periyanattar,
may be seen as an im portant further development of South Indian
segmentary society. But, there was a cost borne by the older local
assemblies.
The emergence o f a ruling strata increasingly divorced from the
locality peasantry accompanied this supra-local cooperation.
This is seen in several ways. One was the differentiation of the
nattar, now collaborating across nadu lines, from the ordinary
peasantry; another was the close cooperation between the supra-
local élite and itinerant m erchant groups; finally, there appeared
to be the development of a distinctive sub-culture associated with
what appears as a new ruling class.
This increased status and power differentiation of those who
comprised the supra-local bodies from the local peasantry is
suggested by the set of inscriptions from M annargudi. During the
twenty-second regnal year o f the C hola ruler Rajaraja III, a .d .
1239, a complicated, joint regulation ( vyavastha ) o f the sabha and
mahasabha of Rajadhiraja-chaturvedim angalam , as M annargudi
was called, and nâttavar of five nadus of the locality was recorded.26
26 A .R .E ., 1897, no.98, 104; S .I.I., v. 6, nos.50. 58. Also A .R .E ., 1897, no.96;
v. 6, no.48. There is the suggestion in these records, and in a few others, that
in some places the two assemblies — sabha and mahasabha — may have been
different bodies, the latter presumably including non-Brahm ans as well as Brahmans.
It is stated that members of the locality peasantry, Vellalas, had
complained that they could no longer m aintain themselves since
their income — cash and paddy crop — was already committed
to too many purposes. Therefore, it was resolved that obligations
of the Vellalas for the repair of the river bunds should be in pro­
portion to their taxable land; that for various other obligations,27
payments were to be based upon a strict levy per unit (m a ) of
cultivated land; and that no additional payments were to be
demanded on various pretexts. It was further provided that new
kudum bus, a term which in this context may be understood as work
groups,28 should comprise only those eligible to cultivate land and
that Brahmans should not interfere in the constitution of the
kudum bus or in m atters involving the land taxes29 from the hamlets
of the locality. Finally, the additional dem and on Vellala house­
holds o f a special tax to support Brahmans (brahm anar-perkkada-
m at) was prohibited. The regulation concluded with the injunction
that if any Brahmans or Vellalas made false allegations against
other Brahmans or Vellalas of the locality, or against those in
charge of the kudum bus or in charge of the collection of taxes
(puravu ) and accounts (urkanakkar) to powerful people beyond

The m eeting recorded in 96 o f 1897 was at the temple o f Sri K ailasam udaiyar which
is one o f the m ost ancient shrines at M annargudi where one version of this inscrip­
tion is found.
21 Tirum ugan tevai, wages for those carrying out royal orders (?); chennlr-vetti,
cannot be identified; n etta l and kuraivaruppu, connected with dredging the river;
and m akkat-chevagapperu, emoluments for village servants.
28 This is to follow the m eaning of kudum bu, groupings of families, as Subrah-
m anya Aiyer does in his discussion o f the U ttaram erur mahasabha inscriptions
(H isto rica l S ketch es o f A ncient D ekhan, v. 2, ed., K.S. Vidyanathan, Coim batore
Co-Operative Printing W orks, C oim batore, \967, pp. 273-4) where it has the same
aggregative m eaning as the term kula. This is consonant with the m eaning given
in the T am il L exico n (v. 2, p. 974), k u tu m p a m , and it deviates from the meaning
which N ilakanta Sastri used in his discussion of the same inscriptions in S tudies
in Cola H isto ry and A dm inistration, especially where he seems to confound the
terms kudum bu and ceri as sections of the settlement (p. 158). The idea that the
kudum bu m ight also have been a grouping o f agricultural families for purposes of
cooperation in agricultural activities is certainly suggested in this inscription where
the process o f periodic reconstitution can hardly be reconciled with fixed kinship
units and where kudum bu m em bership is specifically restricted to those who cultivate
land.
29 These are enum erated as follows: ur-viniyogam , village expenditures; kudum bu-
ka c u , lineage (or work unit) paym ents; and various temple dues.
the locality (mudalis), they would be considered as traitors to the
village and nadu.
The mudalis in this inscription can have been none other than
the nattavar o f the five nadus themselves, and the purpose of the
regulation would appear to have been to create a better balance
between Brahmans and the local peasantry with respect to divisions
of local resources with some favour being shown to the latter.
The injunction against carrying ‘false allegations’ beyond the
immediate vicinity, i.e. to the m udalis , suggests further the interest
of the supra-local assembly o f the nattavar in having local problems
disposed of by those of the immediate locality. The implication
would accordingly be that efforts were made on the part of those
more powerful elements of the peasant population who had attained
an ascendant place over several peasant localities to separate
themselves from the local peasantry. The periyanattar appeared
not to be acting as the patrons of the local peasantry, from which
they derived, but as m ediators in the relations between that peasan­
try and the Brahman sabha of the vicinity who appear to have
lost some of their former prestige.
A more reliable indicator of the separation of the leadership
segment of the peasantry with respect to local peasants as a whole
would be an increased dem and from peasant production. To a
degree that seems suggested in the M annargudi records mentioned
above, it might be supposed that the greater the proportion of
agrarian output taken by the Brahm an sabha o f Rajadhiraja-
chaturvedim angalam , the less was available tp those with supra-
local authority, who might, in turn, have passed some of this to
temples. Thus, the effort to limit and regularize payments for
strictly local purposes, including the support of local Brahmans,
may be understood as a step toward freeing resources for transfer
to the supra-local authority of the nattar for their distribution.
Because revenue details are seldom ever given in the records of the
period and because even such as exist are unclear,30 it is not possible
to strongly support the assertion that there was a shift in the level
of payments from the prime ancient recipients of shares of locality
income, such as Brahmans, to the new class of locality leaders
or that the dem and upon the lower peasantry actually rose in the
period.
30 See A ppadorai’s discussion, op. cit., v. 2, pp. 676-80.
 One scholar, L.B. Alayev, who has given special attention
to this problem, suggested that during the twelfth century in Chola
country there were indications of an increased demand from
holders o f the kllv dr am , i.e. the share of income to the peasantry
proper, judging from the flight of peasants from their lands. Alayev
attributes the increased dem and to pressures from the growing
temple institutions of the time which were, already in the twelfth
century, holders o f the m elvaram .31 An objection may be posed
to his suggestion about the temples as a source o f pressure upon
the peasantry since temples themselves possessed no coercive
powers. If the indications to which Alayev points are correct,
they signify the exercise o f enhanced power by those who constituted
the supra-local bodies and who were the chief supporters of temples
from this time on. A thirteenth century inscription from Srirangam
would seem to support the view that substantial if not an increasing
portion o f peasant output was taken by supra-local chiefs during
the twelfth century and later. This was a record of the Hoysala
ruler Vira Somesvaradeva in about a . d . 126432 which provided
that the locality authorities (nattavar ) of idaiyarni-nddu and
kurai-parru were to deposit in a granary (kottakaram ) - 10,000
kalam o f paddy out o f the dues (kadam ai ) taken by them which
am ounted to 20,000 ka la m ,33 This'was stated to be in accordance
with the proportion taken previously by the chiefs of the region,
the akalanka-naddhar, who were apparently displaced by the
Hoysalas.34 The dem and of half of the paddy dues, kadam ai ,
from those who m anaged local production suggests that the de­
mands of the latter upon those who cultivated was high. However,
31 Unpublished paper presented to a sem inar on ‘Problems o f Social and Cul­
tural Change in South India’,held in M adras, 14and 15 June 1968, under theauspices
of the M adras Centre o f the Am erican Institute of Indian Studies. The paper was
entitled ‘Land Rights in Medieval South India as a M easure of Social Status’.
32 S .l .I . , v. 4, no.435.
33 A ppadorai provides the inform ation for converting this quantity of paddy
(20,000 kalam ) into 120,000 bushels o f the grain (op. cit., v. 2, pp. 407-10).
34 A chief with this title was one o f two mentioned in an a . d , 1196 inscription
in this part of Tiruchirapalli recording an agreement (nilaim ai-tittu ) to cooperate
together against others (A .R .E ., 1908, no.483, discussed in K.A. N ilakanta Sastri,
The Colas , v. 2, pt 2, p. 693). A nother inscription which refers to the ascendant
position of the Hoysalas in this part of South India at the time is an A .D . 1245 one
from Tirum ayam , Tiruchirapalli district (K.R. V enkataram an, ‘An Interesting
A w ard’, in Nagaswam y, -Seminar, on Inscriptions, pp. 3-7).
without better data on the division of shares prior to the twelfth
century, it is not possible to accept Alayev’s contention of an
increased demand.
A nother dimension along which the powerful class of supra-
local authorities was separated from their peasant bases and origins
may be seen in the collaboration of the periyanattar with trade and
artisan groups. It is one of the distinctive characteristics of this
transitional period of the twelfth and thirteenth centuries that
groups which before had no separate corporative existence now
enjoyed increasingly independent identification: the powerful,
agricultural cittram eji-periyanattar; the itinerant, mercantile tisai
ayirattu ainnurruvar, or ncmadesi; and mobile artisans, sometimes
identified by an ancient title, rathakara, at other times by such
terms as K am m alar.35
Relationships among these groups are somewhat clarified in the
set of epigraphs collected by K.V. Subrahm anya Aiyer for his
discussion of the ‘provincial organization’. Six o f the nine records
examined by him for the period a . d . 1168 to a . d . 1314 involved
arrangements between mercantile groups and the periyanattar .36
From M annargudi, three inscriptions record provisions for the
payment of a duty on areca nut and the pepper trade to temples of the
place; from Piranm alai (Tiruppattur taluk, Ram nad) a wide range
of trade goods became leviable to support ritual services and repairs
to a temple, ‘under the protection of the vishayattar of the eighteen
vishayas" ,37 In these inscriptions, the supra-local authorities, called
vishayattar, which is the equivalent of periyanattar, assume no
liability for payments to the temples which are the recipients of the
grants. The vishayattar seem merely to be assenting in and associat­
ing themselves with an allocation of resources which they, the
supra-local authorities, would otherwise have received. Moreover,
in the M annargudi records, the mercantile groups (nagarattar ),
are local traders subject to the authority of the vishayattar whereas,
in the Piranm alai record, the mercantile groups include those from
35 S J .T .I ., v. 3. pt 2, p. 203.
Jt> S .I.!., v. 6, no.40 (89 o f 1897), 41 (90 of 1897), 47 (95 of 1897); v. 7, n o s.198
(701 of 1902), 291 (21 of 1903); v. 8, no.442 (154 of 1903).
37 This record (S.1.1., v. 8, no.442 [154 o f 1903]) is an extremely im portant one
not only for an understanding of the supra-local assemblies of the period but also
for its enumeration of trade goods (‘Largest Provincial Organizations . . Q .J .M .S .,
v. 46, p. 17).
specific trade centres in pandya- and kohgu-m andalam who are
also called nagarattdr ,38 But, in addition, there are several groups
of itinerant traders with the title, tisai ayircittainhurruvar, generally
understood as itinerant traders.39 Thus, while the M annargudi
inscriptions are those of a supra-local assembly alone registering
their agreement to the arrangement involving duty to be paid by
local merchants, in the Piranm alai grant there appears to be some­
what greater parity between the vishayattar and the mercantile
groups since both seem to be issuing the grant as suggested by the
terms: citrameli-sasasanam-tribhuvanasraya-pahchasata-vlra-sasa-
nam in the opening lines of the inscription.40
On the whole, however, the relationship between the supra-local
assembly o f dom inant land controllers and the mercantile bodies
referred to in these records assumes the subordination of the latter
to the former. In the a . d . 1235 inscription from Anbil in m odern
Tiruchirapalli,41 several groups are mentioned in addition to the
citram eli periyanattom . These are: itinerant merchants (tisaia-
yirattaihhurruvcir ), local merchants (ndttu chettigal), other m er­
chants (devala-chettigal and jayapalar), as well as artisans (kalanai )
and soldiers (munai and munai virakkodiyar). These are collectively
referred to by the supra-local group as ‘our people’ (n a m -m akkal ).42

38 These include the nagarattar of several places including: Aruvim angalam

alias K ulasekhara-pattinan, Eripadainallur-vadam attai, Pudutteru alias Seranara-
yana-puram , T irukottiyur-m aniyam balam , Alagapurara alias Seliyanarayanapuram ,
Sundarapandiyapuram .
39 There is disagreement am ong scholars on this term. They are spoken o f as
the ‘1,500’ by Subrahm anya Aiyer (‘Largest Provincial Organizations. . . .’
v. 46, no.2, p. 73) and as ‘500’ of the thousand districts by N ilakanta Sastri (The
Cotas, p. 319); simply as ‘m erchants and traders in m any lands', by T.N. Subra-
maniam (S .I .T .I ., v. 3. pt 2, p. 203).
40 ‘Largest Provincial O rg an izatio n s. . . . , Q .J .M .S ., v. 46, n o .l, p. 8. It is note­
worthy th a t contem porary inscriptions from K arnataka point to developments
similar to those in Tamil country. A damaged record from M udanur (Shorapur
taluk, G ulbarga) recently noticed contains a resolution of a supra-local assembly
(mahanadu) involving several m erchant groups and the regulations they framed
(A .R .E ., 1960-1, no.519). A nother record from G abbur in R aichur of the twelfth
century refers to an organization called m eli sdsiruvaru (‘1000 ploughs’), equated
by the epigraphist with the citram eli, as a guild of agriculturists (ibid., no.665 of
1958-9, p. 25) dating from the time of the KaVyani Chalukya ruler Vikram aditya VI.
41 A .R .E ., 1902, no.601, the text o f which is found in ‘Largest Provincial Organi­
zations . . .’ (Q .J .M .S ., v. 45, no.2, p. 74).
42 Ibid., p. 73; reading is ‘our sons’.
In the judgement of K.V. Subrahm anya Aiyer, the supra-local
assemblies of the period were ‘all powerful’ with respect to other
corporate groups, and though there is no reason to accept his
argument that these supra-local groups were instruments of the
kings and ‘exercised State powers’,43 his recognition of their
plenary local power must be accepted.

The Sub-Culture of the Periyanattar

Concom itant with and perhaps underlying the impressive authority
of the supra-local assemblies in the economic and political m anage­
ment of the agrarian order was a cultural development modelled
on existing Brahmanicai institutions, but decidedly reflective of
the self-confidence of the ruling members of these assemblies.
The late twelfth and thirteenth centuries may have witnessed the
eclipse o f Chola authority, such as to prom pt N ilakanta Sastri’s
observation that ‘the period was m arked by no striking develop­
ments in polity or society. . . .’,44 however, the emergence of
powerful locality authorities, boasting o f special relations with
temples and seminaries im athai) and the increasing importance
of what were essentially new urban places m ust be considered as
most ‘striking’.
The m anner in which influential, non-Brahman, agrarian groups
were involved in the affairs of the sacred centres of South India
during the period of the great Cholas (tenth to twelfth centuries)
has been touched upon in several ways. The nattar played a vital
role in the support of brahmadeyas. It was peasant-agrarian
resources collected by superior locality groups and voluntarily
granted to Brahmans, individually and corporately, that sustained
these m ost significant agents of ritual and education. Royal
gifts and the support o f those close to the great overlords o f the
43 See ibid., no.4, 284-5. Here Subrahm anya Aiyer attributes neglect of these
bodies to their confusion as primarily religious in character: ‘The reason for the
non-recognition of this all powerful body [the periyanattar] which wielded the
greatest influence in the land is due to the m istaken notion that the word sam aya
o r sam aya-dharm a m eant the dharm as o f the comm unity in all religions.' Sam aya-
dharm a is regarded by him as the same as vyavastha\ 'regulations’ or ‘recorded
transactions' which are equivalent to raja dharm a to be obeyed as if they were the
orders o f the king (ibid., p. 281).
44 H istory o f South India , p. 202. He goes on to state that industry, trade, and the
arts continued to flourish.
Corom andel plain were of im portance principally in the magnificent
Brihadisvara temple at Tanjavur, the temple, at Garigaikondachola-
puram , and in a few other places. These were as much m onuments
to the power o f their Chola builders as m arks o f their piety. Direct
‘royal’ grants were few, relative to the overall pattern o f support
to Brahm an institutions, though such royal gifts must be considered
as models for the overwhelming support by local dom inant groups
to such institutions. To both the local and great overlords, such
support brought honour and fortified legitimation in addition
to serving the motive of piety. The sustaining support for all
but the few institutions which enjoyed direct royal favour came
from the spokesmen and leaders of the peasantry o f the region.
To what extent these supporters of Brahmanicai institutions
o f learning and ritual were direct beneficiaries of these institutions
is, however, questionable. Too little is known about the modes of
temple worship from the tenth to the twelfth centuries to assess
the degree of non-Brahm an ritual participation. The probability
is th at there was little direct participation. Certain kinds of know­
ledge related to ritual were transm itted within the context o f the
Brahm an family or, for higher forms of such knowledge, in a
relationship with a guru or acharya. The place of guru as trans­
m itter of high sacred knowledge became increasingly im portant
during later Chola times as differences between Saivite and Vaishna-
vite doctrine became firmer,45 and as these led to religious orders,
‘sects’ (Tamil: cam piratdyam , from the Sanskrit sam pradaya).
‘Sectarian’ education conducted in a seminary (m ath a or gh atika)
provided the comprehensive studies judged necessary for the
maintenance of Brahmanicai traditions of the age. Here, careful
screening of students served to exclude all but Brahmans with
appropriate genealogical and sastric credentials. The most com­
plete educational facilities were the gh atikas o f Kanchi, Tiruvadut-
turai, Tiruvorriyur, Ennayiram , and Tribhuvani where curricula
extended beyond sacred knowledge p e r se to include medicine,
poetics, and other arts. Inscriptions from these places provide
an impression o f heavy B rahm an control. Most teachers were
Brahmans, the language o f instruction was Sanskrit, and m ost of
the stipendary students were Brahmans.
45 K .A . N ila k a n ta S astri, D e v e lo p m e n t o f R eligion in Sou th India, pp. 61-5; Far-
q u h a r. O u tlin e o f R elig io u s L ite r a tu r e , p. 257.
 Exclusion from these educational contexts to which Brahmans
controlled access does not mean that Sanskrit and the sacred
and sastric lore in that language could not be obtained by non-
Brahmans. Some o f the great figures of the age who were not
Brahmans commanded Sanskrit and at least a part of its knowledge.
The thirteenth century form ulator of Saiva Siddhanta, M eykandar,
is recalled as a pious Vellala who translated twelve Sanskrit sutras
o f the raurava-agam a into Tam il.46 indeed, the classification
of satvik-sudra applied to Vellalas in medieval South India, if
it means anything, means that some o f them were deemed worthy
of some Brahmanicai education.
The general character of non-Brahman culture in South India,
its pretentions to ritual purity, was based precisely upon such
Brahman norm s and knowledge47 as noticed in the Tirukkoyilur
inscription quoted above. How non-Brahmans acquired Brah­
manicai learning is obscured by the dominance of Brahmans over
the most im portant cultural institutions of which we have record.
However, there is a suggestion that the term kanim urruttu may
refer to support of higher forms of non-Brahman learning during
the tenth to twelfth centuries. This term appears in Chola copper­
plate grants to indicate a right to income from land enjoyed by
persons outside o f the brahm adeyas; other rights o f a similar sort
were those of devaddna , paU iccandan , and vettiperu with which the
kanim urruttu right is usually included. The other terms relate
to income rights o f persons attached to H indu temples, Jaina
shrines, and those who clear forest tracts around settled areas and
possibly engage in other public works.48 Kanim urruttu was ex­
plained by K.V. Subrahm anya Aiyer as an income right to those
called kadigaiyar, sometimes identified as announcers of the time,49
but m ore plausibly as court poets.50 Recently, a learned epi-
46 Ib id ., p. 94 an d F a rq u h a r, O u tlin e o f R elig io u s L itera tu re , p. 257.
47 In th e recent p ast, th e ideological w ritings o f leaders o f the n o n -B rah m an
m ov em en t in T am il co u n try argue th a t alleged B rahm an n o rm s are really ancient
V ellala ones a d o p te d by B rah m an s an d used ag ain st th e V ellalas. S e e Swam i Veda-
ch alam , V elalar N a g a rig a m [The C iv iliza tio n o f A n cien t V elalas], T .M . Press, Pal-
lav aram , 1927.
4!i ‘L arg er L eiden Plates o f R a ja ra ja I ’, E .I., v. 22, pp. 213fF; ‘T iru v alan g ad u
P lates o f R a je n d ra I ’, E .I., v. 3, p. 383ff; S .I .T .L , 'G lo ssa ry '; an d T a m il L ex ico n ,
v. 6, p. 3296. m urruttu.
4» E .L , v. 22, p. 231.
50 T a m il L e x ic o n , v. 2, p. 668.
graphist, K.G. Krishnan, has suggested that kan irnu m ittu may
refer to payments to the learned (perhaps poets?) who were not
Brahmans and who were thus excluded from participation and
benefits from brahm adeyas, gh atikas, and vidyasthân as .51 The
evidence of achievements by non-Brahmans in Sanskrit and higher
forms of knowledge prior to the late twelfth century and the high
probability that non-Brahm ans were substantially excluded from
other sources of support, strongly favour K rishnan’s suggestion.
Beyond limited access to the sacred knowledge o f Brahmanically
oriented institutions, non-Brahm ans would presumably also have
had been denied access to many ritual performances involving
Brahman functionaries. F or one reason, the shrines of canonical
deities during the early Chola period and most of the period of the
great Cholas were located in the residential quarters of Brahman
villages, thus non-Brahmans would have suffered restrictions of
movement into these quarters, it would also appear to have been
the case that from the ninth century, most shrines of Brahmanical
gods were simple, even austere, reflecting the aniconic propensities
of the Sm artha tradition of the age. This is said to have been a
reversal o f the trend to greater iconic worship before a .d . 800 and
thus may reflect a deliberate effort to exclude all but the sastric-
oriented Sivabrahm ans.52 The remains of early Chola temples,
stripped o f their later elaborate additions — m inor shrines, great
walls providing impressive am bulatory corridors (p ra k à ra ), great
gate-ways (gôpura ) — consist o f small shelters for a lingam. Here,
during m ost o f the C hola period, the most elaborate ritual would
have consisted of lamp offerings and recitations of Sanskrit m antras
and, increasingly, Tamil hymns in praise of Siva.53 Given the rules
of purity governing sacred places and the modest scale of ritual
51 P erso n al c o m m u n icatio n from K .G . K rish n an , Office o f the E p ig rap h ist for
India.
52 K .A . N ila k a n ta S astri, D e ve lo p m e n t o f R eligion in S ou th India, p. 111. K .R .
S rinivasan (su p erin ten d en t o f th e D e p a rtm e n t o f A rchaeology, so u th e rn circle
[M adras], G o v ern m en t o f India), has suggested th at a tran sitio n from iconic to
an ico n ic fo rm s occurred a ro u n d a . d . 800, the sivalin ga replacing iconic form s:
'. . . a reversal o f w hat had o b tain ed earlier an d w hat has been persisting in low er
su b -strata o f society th ro u g h o u t’ ( S o m e A sp e c ts o f R eligion as R ev e a le d b y E a rly
M o n u m en ts a n d L ite r a tu r e in Sou th In d ia , U niversity o f M ad ras, M adras, 1960,
p. 62).
53 T his d ates from the late P allava p eriod according to N ila k a n ta S astri ( D evelo p ­
m ent o f R elig io n in S o u th In d ia , pp. 120-1).
performances, the scope for non-Brahm an ritual participation
would have been small. It was only with b h a k ti ritual in its more
elaborate and popular form based upon the inclusion of folk
elements of iconic worship of the twelfth century and after — a
development associated with the development of Vaishnavism
after Ram anuja — that the non-Brahm an place in Brahmanicai
temple activities became significant. These developments form
an im portant episode in South Indian religion.
Direct involvement of the dom inant landed folk of the C oro­
mandel plain and portions o f the contiguous upland with Brah­
manicai (Vedic) temples became greater during the twelfth century
and later. Considerable significance m ust be attributed to reforms
in Vaishnava temples o f this period with which the name o f R am a­
nuja has persistently been linked. According to the chronicle
(.sthalapurana ) o f the Srirangam temple at Tiruchirapalli, K o il
Olugu, the great Vaishnava acharya spent an extended period at
the temple, succeeding the Vaishnava teacher Alavandar as superin­
tendent.54 Among the ten divisions o f temple functionaries estab­
lished by Ram anuja was one consisting of Sudras called sa tta d a
Vaishnavas, a term which means Vaishnavas without ‘threads’,
or Vellalas.55 The roles which such non-Brahmans carried out
were those of storehouse managers, guards, accountants, as well
as some ritual posts including the offering of coconuts to the
deity, R anganatha.5f> It appears that the position o f the non-
Brahm an functionaries at Srirangam was not to endure beyond
the late fourteenth century when, along with non-Brahm an func­
tionaries at the other im portant Vaishnava shrine at Tirupati,
they were replaced in their duties by Brahmans or by ascetic orders
comprising Brahm ans and non-B rahm ans.57
However, at lesser Vaishnava shrines, and probably at most
temples of the time, non-Brahm an leaders were most consistently
54 K o il O lugu, The C hronicle o f the S rira n g a m T em p le w ith H is to r ic a l N o te s,
ed. V .N . H ari R a o , R o ch o u se an d S ons, M a d ras, 1961, p. 41ff. A lso see H ari R a o ’s
d o cto ral thesis, ‘A H isto ry o f T rich in o p o ly an d S rira n g a m ’, U niversity o f M ad ras,
M a d ras, 1945, co n su lte d a t th e U niversity o f C hicago L ibrary.
55 Ib id ., p. 90.
56 Ib id ., p. 94.
57 B u rto n Stein, ‘Social M obility an d M edieval S o u th In d ian Sects’ in J. Silverberg
(ed.), S o c ia l M o b ility in the C a ste S y s te m o f India, M o u to n , T h e H ague, 1968, pp. 78-
95. T h ere are suggestions o f B rahm an o p p o sitio n to som e o f the reform s involving
S u d ras in the tim e o f R a m a n u ja (H ari R a o , K o il O lugu, . . .).
influential. Num erous records of the twelfth and thirteenth cen­
turies refer to citram eji-vinnagar temples, that is Vaishnava shrines
under the protection o f the citram eU -periyandttar. The ancient
shrine o f Trivikram a Perumai in Tirukkoyilur (South Arcot),
also called tiruvidai-kqli-ndyancir , celebrated by the early hymnists
(a lva r s), was under the protection o f a supra-local assembly accord­
ing to an early thirteenth century inscription.58 A portion of this
record reads:
To the tem ple o f E lupattonbadu-nattu-padinen-bhum i-cittram eji-vinna-
gar [the Vishnu deity o f the supra-local assembly o f the 79 countries
and 18 districts] alias T iruvidaikaU 'ninraruliya-pem m al at T irukovalur
in K u rak k ai-k u rram o f M aladu alias [its C hola territorial designation]
Jananatha-valan ad u , w hich has received the [symbol o f the] plough
[meh toranam] and has been left under our protection [and] bhum ideva
[the earth god] having been consecrated by us, we m ade the following
endow m ent: as the gifts m ade by us to this tem ple in form er years had
been lost along with the edicts relating to them during invasions, we have
now ordered th a t one pad a kku o f paddy on each plough and one kuruni
o f paddy per individual shall be paid as had been form erly decreed by us.59

in modern Nellore town (Nellore district, A ndhra Pradesh) a

Vishnu deity was granted income from 2,200 kuli of land (about
7 acres)60 in a . d . 1197 by a supra-local assembly meeting in the
pillared hall of the temple (m andapam ) called the citram eli-m an-
dapam . Introducing this record is the standard invocation of the
citram eli-periyan attar, followed by:
In the Saka year 1119, the inhabitants o f the n ad u s [eight are enum erated]
o f Jayankondacholoa-m andalam , having all assembled at the chittra-
m eli-m andapa at T iruparkadar-chittiram eli-vinnagar, gave to the god. . .
free of tax, lands to the extent o f 2,200 kuji.

In the im portant citram eli-periyan attdr inscription from Piran-

malai (T iruppattur taluk, Ram nad), already discussed in part,
it is stated:
. . . the sacred tem ple o f M argavagaittirta-m udaliyar-nayanar-udaiyar
. . . at the fo o t o f T iru k k o d u n k o n ram hill in T irum alai-nadu, w ith its
sacred tank, m ath a and T irum adaivilagam had been left under the p ro ­
tection o f (assembly o f the) Eighteen D istricts (padinen-vishaya). . . . 61

58 A .R .E ., 1922, p ara. 8, pp. 8-9.

59 A .R .E ., 1900, n o .117, text an d tra n sla tio n in ‘L argest P rovincial O rg an iza­
tio n s . . Q .J .M .S ., v. 45, n o .2, pp. 70ff.
■ 60 N e llo re In scrip tio n s, op. cit., v. 2, p. 824. T he conversion o f k u li follow s A p-
p ad o rai, op. cit., v. 1, 262n, 406.
61 ‘L argest P rovincial O rg an izatio n s . . .’, O .J .M .S ., v. 45, no.4, pp. 270 fit. I .M .P ..
in a . d . 1217 a n inscription registered a gift to the temple Tirumer-

koyil-citram eli-vinnagaralvdr at Vijayamangalam (Erode taluk,

Coimbatore) said to be under the protection of the left division
of castes (idangai).62 A nother similar record from Paruthipalli
(Salem taluk, Salem) refers to a temple as citram eli-vinnagaram
and to endowments of the citram elincittar ,63 Other places whose
Vishnu temples and m at has were under the protection of the
supra-local assembly and accordingly called citram eli-vinnagaram
were at V alarpuram ,64 Sendam angalam ,65 and Anbil.66
The reform impulse in contem porary Saivism also involved
participation by non-Brahmans. Saiva Siddhanta philosophy with
its intense monotheism and devotional emphasis was transform ed
into a popular sect during the thirteenth century. Its famous
early teachers, using Tamil as the vehicle for their teaching, included
two Vellalas and two Brahmans. M eykandar, the first of the
Tamil expounders of Siddhanta and a Vellala, translated twelve
Sanskrit works into the Tamil Siva-jnana-boclham.61 His disciple
was a Brahm an, Arulnandi. Marai, the next great teacher of the
sect, was a Vellala ;6^ his disciple Um apathi, at the end of the
thirteenth century, was perhaps the most accomplished theologian
and is alleged to have suffered excommunication from his Brahman
community at Chidam baram for his association with Vellalas.69
Other reforms in. Saiva worship as this involved active non-Brah-
man participation were the Vira-Saiva m ovement in K arnataka
and the sect led by Aradhya Brahmans in parts of medieval A ndhra
which m aintained close relations with the Vira-Saivas.70 Appro-
v. 2, p. 1122, nos. 534, 535, 539, 543, 544. are o th er exam ples o f these assem blies of
the th irte e n th cen tu ry . T he q u o ta tio n is fro m A .R .E ., 1903, n o .154; S.1.1., v. 8,
no.442.
62 A .R .E ., 1905. n o .5 6 4 ; I .M .P ., v. 1, p. 540. T his record refers in an u n u su al
way to the ‘n in ety -n in e’ o f the left h and.
63 A ro k iasw am i. K ongu C o u n try, p. 289, citing E .C ., v. 4, H g 17, v. 7, Sk 118.
(>4 A .R .E ., 1911, n o .28.
65 P u d u k k o tta i S ta te In scrip tio n s, no. 171, p. 100.
6« A .R .E ., 1900, n o .22.
67 V iolet P a ra n jo ti, S a iv a S id d h a n ta in the M e y k a n d a r S a s tr a , L o n d o n , 1938;
L uzac co n tain s a list o f im p o rta n t w orks on th e sect, (p. 9).
68 FC.R. S rinivasa Iyengar, ‘D rav id ian L anguage an d L ite ra tu re ; Saiva S id d h an ta
L ite ra tu re ’, in R .C . M a ju m d a r (ed.), H is to r y a n d C ultu re o f the Indian P eople, v. 5,
B h a ratiy a V id y a B h av an , B om bay, 1966, p. 366.
69 F a rq u h a r, O u tlin es o f R elig io u s L ite ra tu re , p. 257. A lso, K .A . N ila k a n ta
S astri, 'D evelo p m en t o f R elig io n . . .', p. 94.
70 Ib id ., pp. 64-5.
priate to the intellectual thrust of these developments, seminaries
(imatha s) played an im portant role, and within these institutions
Saivite, non-Brahman agriculturists were conspicuous. Many Sai-
vite m athas were under the control o f non-Brahm an gurus or
m ath apatis during the thirteenth century. As centres of Saiva
Siddhanta, these m athas were the institutionalized aspect of the
m ajor role taken by the sect’s Vellala progenitors. A study of
Saivite institutions to the fourteenth century by M. Rajamani-
ckam attributes central significance to these seminary institutions.71
Saiva m athas involving non-Brahm an participation and leader­
ship arose in the first half of the thirteenth century.72 Among the
most im portant of these were: the Tirunana-sam bandham matha
at Tiruchchattim urram (Rajarajapuram in Tanjavur) with branches
in several places;73 the M aiigai-m atha at Tiruvidaim arudur also
in Tanjavur with branches as distant as Chidam baram (Tillai) ;74
the Senbaikkudi m a th a ; the Acharam alagiyan-matha at Tiruvarur
in R am nad;75 the Tiruitondaitogaiyan-m atha at G ovindaputtur;76
the Siruttondan-m atha at Sengattan-gudi (Tiruchchengattangudi),
T anjavur;77 the Tirutturaiyur-m atha in South A rcot; the Tiru-
vadutturai-m atha in T anjavur;78 the D harm apuram m atha in
M ayavaram taluk, Tanjavur.79 These seminaries and their branches
were under the control and supervision of non-Brahm an teachers,
sivayogins or m ahesvaras, and the mathas themselves were often
named for the non-Brahm an Saivite saint (ndyandr ) Tirunavuk-
karasar. Seminaries with the latter designation existed at Tiruvili-
milalai in Tanjavur and Tiruppalatturai in Tiruchchirapalli.80
Headships of the respective matha organizations constituted a line
called a santdna and usually bore the title m udaliydr, presumably
71 M .R a ja m a n ic k a m , The D e ve lo p m e n t o f S a ivism in S ou th India (A .D . 300-
1 3 0 0 ). D h a rm a p u ra m A d h in am , D h a rm a p u ra m , 1964.
72 ib id ., p. 379; th ere is a m a jo r discussion o f these institu tio n s in A .R .E ., 1909,
p ara. 53, pp. 103-5.
73 A n aik k a, A .R .E .. ¡908, n o .586; U s attan am , ibid., n o .218; V ilim ilalai, ibid.,
n o .392; V aliv alam , A .R .E ., 1911, nos. 108, 109.
74 R a ja m a n ic k am , op. cit., p. 232.
73 Ib id .. pp. 233-4; A .R .E ., 1909, p a ra . 53.
76 A .R .E ., 1909, para. 53.
77 ibid .
7« Ibid.
7,) Ibid.
so Ibid.
to distinguish them from Brahman m ath a leaders.81 Still another
distinctive designation, which occurs in the case of the m atha
at Tiruvanaikkaval, in Tiruchirapalli, a branch of the Tiruchchatti-
m urrattu m atha at Rajarajapuram in Tanjavur, is närpattennäyi-
ravan-m atha, or the ‘48,000’ m a th a ß 2 The ‘48,000’ designation,
as already noted, is one of those applied to Vellalas of Tondai-
mandalam.
There are some indications that the flourishing non-Brahm an
m athas were opposed by Brahmans, in one case, that o f the
‘48,000’ m atha at Tiruvanaikkaval, the institution appears to have
been taken over by Brahmans who changed its name to the Sam-
karacharyaswam i-m atha.83 In another case, a m atha or guhai
established for non-Brahm an ascetics at Tirutturaipundi in Tanja­
vur was the scene of some disturbances (kalagam ) in the year
a . d . 1200 and resulted in the appointm ent of a Brahm an teacher.84

Such attem pts by Brahmans to overturn the control o f non-Brah­

man leaders at these centres of learning were not, in the main,
successful. As Rajamanickam pointed out in discussing the
disturbances at T irutturaipundi in a . d . 1200: ‘. . . in the time of
Rajaraja III and subsequently . . . [non-Brahman guhais] flourished
under the patronage o f ruling chiefs and private individuals.’85
An im portant, yet barely noticed, development in temple worship
during this period when non-Brahmans began to have an influential
place was the full emergence o f goddess — A m m an or D evi —
worship. From the thirteenth century, temples were constructed
to include a shrine for a female deity whose identity was linked
to the m ajor deity of the temple; moreover, these shrines were
added to many of the temples which had been constructed before
that time.86 The significance o f this development is that it represen­
ted an assimilation of folk conceptions of deity.
81 R a ja m a n ic k am . op. cit., p. 231.
82 A .R .E ., 1909, p ara. 53, reg ard in g n o s.585, 586. K .V . V e n k a ta ra m a n associated
th e ‘48,000’ g ro u p o f Saiva S id d h an ta teachers w ith a line (sa n ta n a m ) called g o la k i
o r la k sh a d h y a y i as well as n ä rp a tte n -n ä yira tta n ä r, o r ‘th e 48,000’ ; ‘A n Interesting
A w ard ’, op. cit., p. 4.
S3 A .R .E ., 1909, p a ra . 53.
M A .R .E ., 1912, no.471, discussed by N ila k a n ta S astri, The C olas, v. 2, pt 2.
p. 722. A lso se e R a ja m a n ic k am . op. cit., pp. 239-41; T a m il L ex ico n , v. 2, p. 775
( k a la k k a m ).
85 R a ja m a n ic k am , op. cit., p. 241.
80 K .R . S rinivasan, T ir u k a m a k o tta m ’, op. cit., p. 53; also his ‘A spects o f Re-
 The female deity, then as now the m ajor focus of village, clan,
and locality devotion — in part as protectress, in part as fertility
deity — attained a central place in Vedic temple worship. Generally
called tiru kam akdttam , the am m an shrine within the Siva temple
complex became a full-blown architectural element during the
thirteenth century after a long, slow evolution beginning in about
the eighth century. According to dgam a texts and inscriptional
evidence of that early time, goddess images were at times installed
in existing Vedic shrines.87 Among the m ost im portant of these
early female deities were Durga, the ‘Seven M others’ (sapta-
m atrikas), and the somewhat fearsome sister of the benign Lakshmi,
Jyeshtha. D urga is m entioned in one of the two principal Vishnu-
oriented vaikhanasdgam as, and among the iconographic attributes
of this goddess were the conch and discus of Vishnu. 88 In the period
after the eleventh century, and particularly from the thirteenth
century, temples of both Siva and Vishnu deities came to include
a shrine for a goddess appropriately nam ed for her association
with the principal male deity, e.g. Brihadesvara and the goddess
Brihannayaki, Peruvudaiyar and Periyanayaki in Tanjavur, Ranga-
natha and Ranganayaki at Srirangam, Sundaresvara and Minakshi
at M adurai, Ekam resvara and Kamakshi at K anchi.89 Within the
Vaishnava tradition, the more conservative vaikhánasa ritual form
gave way to that of the páñ carátra, following the reform activities
of Ram anuja, and shrines for Vaishnava consorts (ndcciyar )
became im portant.90 The female hymnist (dlydr), Andal and the
goddess Lakshmi were among the im portant deities in such shrines.
The m ajor shrine of Andal (also called k o d a i and su dikkodutta
ndcciyar) at Srivilliputtur in R am nad is one o f the m ost famous of
D evi shrines among Vaishnavas; it was constructed in about a . d .
1160.91 O ther Vaishnava nachchiyar which strikingly bear out this
ligion R evealed by E ariy M o n u m e n ts . . op. cit., p. 22 for agam ic evidence.
In scrip tio n al evidence is fou n d , curiously, in Ja in a inscriptions o f the eighth century
in w hich th e im p re catio n includes a reference to th o se w ho d estroy k á m a k ó tta m s
(discussed in M .D . S a m p a th ‘Ja in a In scrip tio n s o f S a tta m a n g a la m ’, in N agasw am y,
S em in a r on In scrip tio n s, p. 158).
87 S rinivasan, A sp e c ts o f R elig io n R e v e a le d b y E a rly M on u m en ts, p. 22.
88 Ib id ., p. 30.
8« Ib id ., pp. 32-3.
90 K .A . N ila k a n ta S astri, ‘D e ve lo p m e n t o f R eligion . . .’, p. 67; F a rq u h a r, ‘O utlin es
o f R elig io u s L ite r a tu r e . . . , p. 244.
91 A n o th er d ate suggested for th e installation o f this deity is a . d . 973 by the
argument o f folk cult assimilation a r e : the Chenchu Nachchiyar at
the A hobilam Narasim haswam i temple, the M alayalam Nachchiyar
at Kanchi, and the U raiyur Valli N achchiyar at Srirangam.92 The
attainm ent of the status of m ajor deity by goddesses, especially
those with ancient tribal cult associations, even though subordinate
to Vedic male deities, was one of the clearest signs of religious
changes in the thirteenth century and m arked the deepened connec­
tion between the peasant culture and high culture of the age.
Religious changes from the late twelfth century strongly suggest
the efflorescence of a significant cultural variant not before visible
in medieval South India. It is impossible to separate this cultural
development from the leaders o f peasant society in the macro
region, for in all cases they are not only intimately involved in the
changes, but they were its prom oters. Individuals such as Mey-
kandar and M arai within the Saiva Siddhanta movem ent and the
num erous santana-m udaliyars were Vellalas; the offices reserved
for non-Brahm ans within the leading Vaishnava shrine at Srirangam
and Tirupati were occupied by Vellalas; temples, m athas, and
even Brahman settlements during the twelfth and thirteenth
centuries came under the protection of the supra-local assemblies
composed o f Vellala n attars o f the citram eli-periyan attar 93 The
emergence o f D evi shrines in both Vaishnava and Saivite temples
is further evidence of this transform ation.
Considering the significant association of leading peasant
groups with Brahmans from the pre-Pallavan period, considering
too that peasant support of Brahm anical ritual and learning was
entirely voluntary during the m ature era of the Pallavas as well as
during the period of the great Cholas, it would be incorrect to speak
of the developments of the twelfth and thirteenth centuries as
new. They, of course, were not. In an essential way, these later
developments may certainly be understood as the evolution of
devotional forms o f religion in which the assimilation o f folk
elements was present from the beginning in the hymns of the
a u th o r T .K . V ira ra g h a v a c h a ry a (‘T h e S riv illip u ttu r T em p le o f S u d ik k o d u tta
N ach c h iy a r’, T iru m a la i-T iru p a ti D eva sth a n a m B u lle tin , v. 6, no.4, pp. 1-3).
92 Ib id ., v. 6, n o .5, pp. 5-6. A lso discussed by V. R a g h av an , ‘V ariety an d In ­
teg ratio n in th e P a tte rn o f Indian C u ltu re ’, J o u rn a l o f A sia n S tu d ie s, v. 15, no.4
(1956), p. 500.
93 K..R. V e n k a ta ra m a A yyar, ‘M edieval T ra d e , C ra ft a n d M e rc h a n t G u ild s in
S o u th In d ia ', v. 25 (1947), pp. 274-6.
alvdrs and nayanars. it is further to be recognized that the folk-
devotional variant in South Indian Hinduism had always had a
reflective aspect m aintained by learned non-Brahmans, proficient
in Sanskrit and informed by sastric knowledge which was trans­
mitted in Sanskrit. If the interpretation of the term kanim urruttu
as a form of agrarian income meant to provide support to non-Brah­
man savants is correct, it would appear that the maintenance of
such learning was considered appropriate by Brahmans and local
chiefs as well as great Chola overlords.
However, even if this term is incorrectly interpreted, it is obvious
that this highest learning by non-Brahmans was supported in
some m anner and that learned non-Brahmans, along with the
ritual functionaries involved with non-Vedic deities and Tamil
siddharacharyas (T am il: cit tar) who were yogic, medical practi­
tioners,94 enjoyed consistent patronage. Here, as in many other
aspects of peasant life, it m ust be recognized that such activities
need not have become a part of the extant inscriptional record
since it fell outside the Brahm an-centred culture, the records of
which were engraved on the surfaces of thousands o f Vedic shrines.
Support by the peasantry of non-Brahm an, non-Vedic ritual and
learning was obviously well-established and did not require the
special notice that transfers of resources to Brahm ans did. It is
not possible to understand the careers o f num erous non-Brahmans
in the Siddhanta movement and in Vira Saivism without supposing
an ancient and established tradition of erudition, in Sanskrit as
well as other South Indian languages, among non-Brahmans.
The conspicuous participation of non-Brahm ans in the most
im portant religious and learned activities of the twelfth and thirteen­
th centuries does not, therefore, reflect new competences so much
as new opportunities for using quite ancient ones.95 Further,
94 In these tra d itio n s from a b o u t the tenth to the sixteenth an d seventeenth
cen tu ries, n o n -B ra h m a n teachers c o n d em n ed idolatry an d tau g h t a n intense
m o n o th eistic faith (N ila k a n ta S astri, ‘ D evelo p m en t o f R eligion . . p. 95, an d the
fo rth co m in g essay o f N . S u b ra h m an iam , provisionally en titled ‘B rah m an s in S o u th
In d ia ’, by th e U niversity o f M a d ra s based u p o n lectures at the U niversity o f M ad ras
in 1967. A lso, Z velebil, S m ile o f M uru gan, ch. 14).
95 O ne n o tes the c o n te m p o ra ry litera ry figure K am b a n , w ho w as a n o n -B rah m an
uvallan, a caste o f tem ple d ru m m e rs an d priests o f K ali shrines an d the poet K u tta n
w ho w as a w eaver (N ila k a n ta S astri, H is to r y o f S ou th In dia , pp. 359-60 an d T am il
L e x ic o n , v. 1. p. 462). K a m b a n ’s p a tro n was also a n o n -B rah m an nam ed S ad aiy ap p a
V ellala (C .P. V en k a ta ra m a A iyar, K am ban a n d H is A rt, M ad ras, 1913, p. 107).
the manifest weakness of the Chola overlordship would preclude
even the most enthusiastic modern scholarly proponents of the
view of bureaucratic kingship in medieval South India from attribut­
ing this support to the Chola rulers.96 On the contrary, this develop­
ment registers the power and self-confidence o f a rural class which
was progressively detaching itself from its peasant base and estab­
lishing deep relationships with the growing urban segment of
South Indian society and culture.

The Urban Milieu of the Twelfth and Thirteenth Centuries

T ow n life’ and its relationship to the society and culture of the
countryside continues to be a vexing problem for South Asian
social scientists. Accustomed to the convenient, and on the whole
valid, dichotom ization of many contem porary and historical
societies and cultures into ‘urban’ and ‘rural’ spheres, social
scientists working in South Asia have often encountered serious
problems in the application o f these concepts to South Asian
evidence. This is perhaps more true for historians than others,
and historians of the medieval period have been very heavily
dependent for information about towns and town life upon the
reports of those with questionable knowledge of India, foreigners,
beginning with Chinese Buddhist pilgrims and ending with Euro­
pean traders. Such places as Kanchi, which from Pallava times
at least, was an im portant pilgrimage centre and locus o f regional
sect and social life, and Tanjavur, which during the period of the
great Cholas appears to have been a m ajor regional centre of
pilgrimage, trade, and adm inistration, are but rarely seen. And,
there were few places with the stature of these in medieval South
India.
Of course, the extent to which it is possible to be informed about
medieval society and culture from conditions of the present is
limited and hazardous. Certainly, however, administrative and
m arket functions of urban places during more recent times cannot
1,6 T he o p p o site relatio n sh ip — th at o f kings dep en d en t u p o n p easan t g ro u p s —
is suggested in K..A. N ilak an ta Sastri an d T .N . S u b ra m an iam (‘H em avati P illar
in sc rip tio n o f K u lo ttu n g a C h ela III, Y ear 2’, E .I.. v. 31, n o .37, pp. 274-5). H ere
it is stated th at the p e riy a n a tta v a r "paved the way for the u ltim ate success o f K u lo t­
tu n g a in g etting th e C h o la th ro n e '. T his conjecture is deprecated by the E .I. ed.,
D .C . Sircar.
be assumed for an earlier period when neither the m odem bureau­
cratized political system nor the m odern international economy
impinged upon the peasant agrarian context, it may be proposed"
that there were no fundam ental differences am ong settlements
in medieval India, South and North, such as to permit the delinea­
tion o f ‘urban places’ of however primitive a character.97 Terms
such as nagaram , puram , pcittinam may be seen to be attached to
certain settlements in recognition of their m arket functions and
possibly their somewhat more diverse (though not necessarily
greater) populations. However, such settlements may be seen
as but m inor variations within the complex structure o f essentially
rural localities in South India before a . d . 1100. In most settlements
of th at time, peasants, m erchants and artisans along with others
lived in close association, sharing not only interdependent econo­
mic relationships, but also a common involvement in the cultural
life of the locality. It is thus difficult to assume that settlements
designated variously as n agaram , puram , or p a ttin a m — and
usually taken to mean towns or cities by many modern researchers
— were ‘centres’ in the sense of being qualitatively different from
most other large settlements of a locality of the macro region.
97 It has been a rg u ed recently th a t the In d ian ‘u rb a n p la c e ’ is n o t d isc o n tin u o u s
w ith its ru ral su rro u n d in g s an d th a t the ‘u r b a n - r u r a l’ d ich o to m y is in ap p ro p ria te
in In d ia. D av id F. P ocock has presen ted this issue m o st cogently. S peaking o f
c u rre n t sociological research in the In d ia n cu ltu ral sphere, P o co ck states:
W here it is recognized the city an d village are elem ents o f the sam e civilization,
how does th e q u estio n of th eir co n tin u ity arise? Is it n o t because the sociologist
has assum ed (alm ost unconsciously) a division w hich his later ob serv atio n s
w o uld lead him to m en d ? If we have posited the village from the outset we
have au to m atically o pposed it in o u r m inds to the city. W hen we com e to knit
u p w h at we have b ro k e n we can only do it by way o f a d escrip tio n betw een the
tw o en tities . . . T h a t the p ro b lem is a false one becom es clear w hen, m oving
to the level o f relations, the d ich o to m y disappears.

‘S ociologies: U rb a n a n d R u ra l’, C o n trib u tio n s to Indian S o c io lo g y , v. 4 (A p ril 1960),

p. 81. A n ap p licatio n o f P o c o c k ’s distin ctio n is m ade by O w en M. L ynch (‘R ural
C ities in In d ia : C o n tin u ities an d D isc o n tin u ities', India a n d C e y lo n : U n ity an d
D iv e r s ity , ed. P. M a so n , O .U .P ., L o n d o n , 1967, pp. 142-58). Involved in this dis­
cu ssion is a n o th e r dim en sio n w hich m ay be viewed as m eth o d o lo g ical, o r perhaps
epistem ological, in w hich the cen tral q u estio n is the co n d itio n u n d e r w hich any
‘pre-estab lish ed c o n c e p ts' from on e cu ltu ral context can be applied to an o th er.
S e e F .G . Bailey, ‘F o r a S ociology o f In d ia ? ' C o n tribu tion s to Indian S o c io lo g y ,
v. 3., pp. 80-101 an d the editorial rejo in d er (P ocock an d L. D u m o n t) in ibid., v. 4,
pp. 82ff.
It cannot even be said of such places that ties of trade and exchange
predom inated over those involving the land and its management.
However, by the twelfth and thirteenth centuries, there were
a few settlements which did exhibit more clearly the characteristics
of an urban place in most m odern usage. Here were found a
greater array of specialists than were to be found in most large,
essentially agricultural settlements — that is, a significant difference
in degree. Another urban characteristic which would be en­
countered in some places was a major, regional temple (with
its constellation of ancillary institutions such as seminaries, choul­
tries) and a m ajor regional market, i.e. a difference of a qualitative
kind.
Settlements which differed qualitatively from even the most
populous agrarian settlements in the greater diversity of the resident
groups, might have included many o f the older brahm adeyas,
those dating from the tenth to the twelfth centuries. As settlements
with a substantial population of Brahm an teachers and ritual
specialists, a wider array of service groups were to be expected
and appear to have been present according to the brahm adeya
records of that earlier period when such settlements were the most
pluralistic of any. Though there was a decline in the establishment
o f new brahm adeyas after the time of K ulottunga I and though
the formerly substantial self-government appears also to have
diminished, these large, pluralistic settlements do not appear to
have changed rapidly. Some seem to have reverted to peasant-
managed settlements, others were placed under the management
of temple authorities as devadana villages; in still other cases these
Brahman settlements became the cores of larger and more diverse
settlements in which different kinds o f sacral functions displaced
those of the earlier Brahm an settlement and in which m arket
functions — always an aspect o f the largest brahm adeyas — became
even more im portant.98
Thus it was that while many, perhaps most, Brahman settlements
of the earlier age continued for some time to exist as im portant
98 T h is ap p ears to have h ap p en ed in the case o f the b ra h m a d e ya T iru k k u d a-
m ukkil, by w hich n am e K u m b a k o n a m tow n w as referred until a b o u t a . d . 1018
(T.V. M a h aim g am , ‘T h e N agesvarasvam i T em p le’, p t 1 (A pril 1967), pp. 73-4).
In a general way, this p o int is even, m ore strongly p resented in the sum m ary tables
p rep ared by S a th ia n a th a ie r (S tu d ies in the A n cien t H is to r y o f T on dam andalam ,
ap p . ‘B ’, pp. 64ff).
places within a locality and to m aintain an impressive array of
religious specialists, the significance of these settlements appeared
to have diminished in relation to other large and diverse settlements
during the twelfth and thirteenth centuries. The latter may be
called ‘towns’ in the sense of being sociologically distinctive from,
if not discontinuous with, the countryside and in the sense that
these settlements were the loci o f functions previously dispersed
over the locality. These centres tended increasingly to become
the dom inant political, ritual, and trade places in the enlarged
localities of the twelfth and thirteenth centuries; they constituted
a new tier of' settlements that contributed to the integration of the
periyanadu.
The Vedic temple of the later Chola period is the m arker o f the
urbanization of the age. Where before the hrahmadeya village
had been a m ajor repository of lithic records, the site of the great
temples provides a m ajor source of records of the twelfth century
and la te r." Temples of this age were undergoing basic architectural
changes which have long been recognized by art historians. Accord­
ingly, the census officer in charge of the census temple survey of
1961 divided Chola temples into those of the early period ( a . d .
900-85), the middle period ( a . d . 985-1070),. and the later period
( a . d . 1070-1250) after which time there was a general mergence
of the Chola style temple with what has been called the ‘Pandyan
style’.100 Architectural developments of the late eleventh century
99 T he reliability o f this assertion is n ot high, n o r will such p ro p o sitio n s be until
tens o f th o u sa n d s o f in scriptions are o rganized better than they are. H ow ever,
if one takes the lists o f reco rd s for tw o C h o la reigns o f ap p ro x im ately equal d u ra tio n
and with ab o u t the sam e n u m b er o f ex tan t records, the d im inished role o f the brah-
m a d e y a . is clear. T h u s, using the list o f reco rd s in N ila k a n ta S a stri’s The C o la ,v,
v. 1 an d v. 2, pt 2, for R a je n d ra ! ( a .d . 1012-45) in v. !. pp. 530-71. involving 221
inscrip tio n s, an d R a ja ra ja l l i ( a .d . 1216-56) in the hitler place, pp. 721-60, involving
284 inscrip tio n s, it was found th at in the earlier reign 79 in scriptions (36%) recorded
th e affairs o f th e h ra h m a d eya settlem ents an d in the la tte r reign 40 inscriptions
(14% ) involved the affairs o f these settlem ents. A n analysis o f the qualitativ e as­
pects o f th e reco rd s from these tw o reigns m akes this im pression m uch stronger.
100 G o v ern m en t o f India, C ensus o f India, 1961. v. 9, ‘M a d ra s’, pt 11-D , ‘T em ples
of M a d ra s S tate; C h in g lep u t D istrict an d M a d ra s C ity ', com p, by P .K . N a m b ia r
an d N. K rish n a m u rth y (1965), p. 7. A lso, S .R . B a la su b rah m an y am , E a rly C hola
A n , pt 1, A sia P ublishing H o u se, B o m bay, 1966, espec. pp. 253ff for his discussion
o f p erio d s o f,e a rly tem ple styles a n d Jam es C. H arle, T em ple G a te w a y s in Sou th
India ', the A rc h itec tu re a n d Ico n o g ra p h y o f the C id a m b a ra m G opuras. B runo C assirer,
O xford, 1963, pp. xii-xiii.
and after include generally larger and more complex, structures
than known before. Ornately carved pillared halls (m andapam )
are characteristic of this period; long stairways to the sanctorum
of shrines set on hills was another; tall gateways (gopuram ) which
became the massive and distinctive element of the Vijayanagara
period made their regular appearance; and, as already mentioned,
the erection o f separate shrines for A m m an or D evi worship were
parts of the new temples constructed in this period and were added
to existing temples in many cases.’05
These temple centres, on the side of whose buildings and walls
most of the records of this later period and the following period
are to be found, were dependent upon a large and varied population
in order to function. Im portant temple centres had always been
so in India. This is noted by the historian of D h arm asdstra , P.V.
Kane, who pointed to an ancient recognition that a village, even
a large one, was distinguished from a ‘town’ (nagaram ) on the
basis that in the latter only, all castes were to be fo u n d ;102 another
scholar, commenting on early South India, stated similarly:
The difference between a village and a tow n ( N a g a r a ) was generally th at
the latter had a tem ple o f high reputation. A ttached to it were the priests
versed in the A gam as, B rahm ins learned in the Vedas, musicians and
others. The aggregation o f a large population due to the shrine or due
to the protection afforded by the fort or tem ple walls gave an industrial
bias to tow n life. . . ,103
To these general characterizations of the temple centre may be
added K.A. N ilakanta Sastri’s reference to Chola temples:
. . . Every tem ple, great or small, held in relation to its neighbourhood
exactly the same position th at the G reat Tem ple [B rihadisvara tem ple
at Tanjavur] had in the capita!. The difference was only one o f degree.
As landholder, em ployer, and consum er o f goods and services, as bank,
school, and m useum , as hospital and theatre, in short, as a nucleus which
gathered round itself all th at was best in the arts o f civilized existence. . .
the medieval Indian temple has few parallels in the annals o f m a n k in d .104
That institutions of the scale of the Brihadisvara temple were
rare during the early eleventh century when it was completed,
101 Sriruvasan, ‘T iru k a m a k o tta m ’, op. cit.
102 H is to r y o f D h a rm a sd stra (A n c ie n t a n d M e d ie v a l R eligiou s an d C ivil L a w ),
v. 3, B h a n d a rk a r O rien tal R esearch In stitu te, P o o n a, 1946, p. 183.
103 K .R . S u b ra raa n ia n , ‘E conom ic C o n d itio n s o f the T h ev ara m P erio d '. Q .J .M .S .,
v. 18 (1927-8), p. 271.
104 The C o la s , p. 654.
but widespread during the twelfth and thirteenth centuries, is well
documented. Harle’s work on the gateways (go pu r am ) of C hidam ­
baram helps to clarify the relationship between one of the distinctive
architectural elements of the age and the consequences in the scale
of the temple complex.
Gopuras, due to the way in which the S outh Indian tem ple evolved after
the twelfth century, cam e to supersede the central shrine as the largest
and architecturally the m ost im p o rtan t buildings in the temple. The
same evolution called for additional enclosures and consequently m ore
and m ore of these gateways in all the large temples, with the result th at
the South Indian tem ple builders henceforth devoted the greater p a rt o f
their talents and energy to building gopuras.105
Along with Chidam baram , the m ajor temple construction which
was completed during the time of Kulottunga III ( a .d . 1178-1217),
m any other im portant temple centres have been identified by art
historians as dating from this period: D arasuram (K um bakonam
taluk, Tanjavur), Tribhuvanam (Pondicherry), Tiruvanaikkaval
(or Jam bukesvara on Srirangam Island, Tiruchchirapalli).106
There is further evidence that stylistic elements of this later Chola
temple architecture influenced construction as far away as K onarak
in Orissa where the famed chariot-like (ratha) appearance of this
temple of the thirteenth century was apparently an adaptation
of the same motifs on lesser scale at D arasuram and Chidam­
b aram .107
The enlarged temple precincts of the twelfth and thirteenth
centuries, the addition of independent shrines with their ritual
functionaries, servants, and workmen, the heightened tempo
of participation by increasing numbers of pilgrims from all res­
pectable groups — not only Brahmans — direct attention to the
basic change in venue of the prime religious activities of the age.
Sizeable urban settlements became an adjunct of great temples;
where they did not exist, they were created. It is said of the town
of Tirupati that the Vaishnava teacher Ram anuja, while sojourning
in the Vengadam Hills and at its sacred shrine on Tirumalai,
ordered that temple functionaries and others serving in the new
105 H arle, op. cit., p. viii.
106 T h is iden tification was m ade by B ala su b rah m an y am . op. cit.. an d con cu rred
in by H arle (op. cit., pp. 40-1, 70) for D a ra su ra m an d T rib h u v an i. T iru v an ak k av a l
is discussed by H arle (op. cit., pp. 76-7), citing the opinion o f Percy Brow n.
107 C. S iv aram am u rti, R o y a l C o n q u ests a n d C u ltu ra l M ig ra tio n s in Sou th India
a n d the D ecca n , In d ian M useum . C a lc u tta, 1955, pp. 18-19, 23.
temple o f Sri Govindarajaswarai at the base of Tirum alai, must
take up residence n ear the new temple. The tow n was thus estab­
lished i n a b o u t a . d . 1130 to serve the requirements o f the shrines
of Venkatesvara an d G ovindarajaswam i.108 In a som ew hat sim ilar
way, the Bhaktavatsalam temple at the Saivite centre o f T iruk-
kalukkunram in Chingleput district was built at the base of the hill
shrine during the thirteenth century, forming the nucleus of the
medieval to w n .!()C) The w estern part of m odern Kanchi came to
have as its most im portant shrine that of the god V aradaraja who,
to the eleventh century, was housed in a small shrine in the village
o f Atiiyur, a suburb of Kanchi. This small shrine was enlarged
to its grandiose medieval proportions during the twelfth century
to form the core of a Vishnu centre — involving eighteen temples
according to traditions of the place — now called ‘Vishnu-
K anchi'.110 S ubstantial enlargement of the ancient Saivite shrine
at Tiruvorriyur (Chingleput district) including the outer walled
enclosure (prakaram ) o f th at temple was built by the chief Adi-
mangalam during the thirty-first year o f K ulottunga III, a . d .
1209.111 A t Suchindram in the far sou th (Travancore), the temple
began to become the ‘spiritual m etropolis of N ancinad’, as K.K.
Pillay has p u t it, after the beginning of the thirteenth century when
control shifted from the m ahasabha o f the place to a group of
temple functionaries called yd g a k k a r. Pillay observed that:
. . . several prosperous villages exercised a considerable influence on
the pagoda, all located within a radius o f four miles o f Sucindram . . . .
Benefactions and endow m ents have freely flowed from them . Crowds
o f visitors and devotees from these places used to flock to the shrine
everyday, and particularly on festive occasions. . . . The people o f these
villages have contributed to the prosperity and fam e o f the pagoda, while
they themselves were influenced by the sacred tem ple and its in stitu tio n s.112

Other m odem towns around M adras city apparently date from

the same time, judging from architectural remains.
>08 v ira ra g h a v a c h a ry a , The T iru p a ti T em p le, v. I, pp. 357-9.
109 B ala su b rah m an y am , op. cit., p. 222. A lso see Jag d isa A yyar, Sou th Indian
Sh rin es, op. cit., p. 25.
110 K .V . S o u n d ara R a ja n , ‘K a u s tu b h a P ra sa d a — N ew L ight on Ja y a k h a T a n tra ’,
Jou rn a l o f the O rie n ta l In stitu te , v. ¡7. no. I (1967), pp. 73-5. H e notes th at at the end
o f th e eleventh cen tu ry the god w as called A ttiy u r A ivar after the place.
111 K .V . R a m a n , The E a rly H is to r y o f the M a d ra s R egion , A m u d h a N ilayam
P rivate L td., M ad ras, 1959, pp. 242, 244-58.
112 Pillay, S u cin dram T em p le, op. cit., p. 12.
 The very process of construction must have lasted for years as
suggested by H arle11} and would have brought into the neighbour­
hood of construction workmen, skilled and unskilled, from quite
distant places. Workmen would remain as an urban population
for an extended period and would thus have required the services
of still others. The total effect of such activities would thus have
been to create settlements with modes of organization considerably
different from the peasant villages of the locality and from those
established brahm adeyas which were not themselves the core
settlements of the new towns.
Changes in social organization would have been most evident
among trading groups and artisans. Reference has already been
made to the presumed impact of temple construction and the
generally increased significance of these new centres upon traders
and artisans. Skilled artisans comprising the Kam m alar are seen
to have achieved new prerogatives and social privileges during the
late eleventh and twelfth centuries as their services in connection
with temple building appreciated. This enhanced position was
the springboard for launching the effective left-hand grouping
o f castes ( idangai) in Tamil country. And, north of Tamil country
— in what is now Andhra Pradesh — the work of K. Sundaram
describes the rise to new prominence, during the twelfth and thir­
teenth centuries, of groups like the Telikis (oil mongers) of Bezwada,
the merchants (Vaisyas) of Penugonda, and the Balanja merchants
of Ayyavali who ‘imitated the chiefs and potentates and assumed. . .
titles likewise’.114
The relationship between new temple centres and social, and
economic change permit the use of the term ‘urbanization’ during
the twelfth and thirteenth centuries; this is perhaps most clearly
documented in A ppadorai’s study of economic conditions. In his
chapter entitled, ‘Towns and Internal Trade', he states that ‘a
temple has often been made the nucleus around which a town in
course of time grew’.115 As is often the case in his discussion,
there is no effort to discern changes within the five centuries covered
by his study. However, it is inscriptional evidence of the twelfth
and thirteenth centuries along with foreign travellers’ reports for
the Vijayanagara period upon which his discussion of towns is
113 O p. cit., p. 41.
114 S u n d aram , 'S tu d ies in E co n o m ic a n d S o c ia l C o n d ition s . . .', p. 30.
A p p ad o rai, op. cit., v. 1, ch. 3, pp. 338 ff.
based. The only exception to this is in his references to the ancient
towns of Tanjavur and K anchi.116 O f considerable importance
in his discussion of the 'economic organization of the town’, 117
is his attention to the nature of guilds and their relationship to
ja ti s.
Appadorai contrasts the medieval European guild with what he
calls 'the Indian caste-gild’, 118 on the basis of recruitment. In the
former, membership was restricted to those of the same profession
in a relatively confined area whereas in the latter, ‘in practice’ and
7« the m ain birth-determined m em bership.119 He further observes
that according to sm riti — the M itaksh ara of Vijnanesvara of the
twelfth century — guilds (srenl) comprised those who earned
their livelihoods by the same kind of work though belonging to
different castes (ja tis ).i2<) That trade and artisan groups were
often members of different castes receives support from inscrip­
tional usage. The term kalanai is often used to designate artisans;
this has been m entioned in connection with left-hand castes121
where the same term is used to refer to persons of ‘mixed caste'.122
Related to this is the suggestion by the epigraphist K.G. Krishnan
that the term sahkarappadiyar, which often appears in inscriptions
from the tenth century on, means those who reside in the quarter
reserved for mixed or confused (sah kara ) castes, n o t, as earlier
suggested by K rishna Sastri, those residing in a quarter for Siva
worshippers.123 Groupings such as the sah karappadiyar and k i[
kalan aiydr or pu ra-kalanaiydr are collective terms for merchants
and artisans living in populous centres; such terms occur at an
early time. An inscription of a . d . 1036 in the reign of Rajendra I124
116 Ibid., v. !, p. 350, n.64.
I'? Ibid ., 356 ff.
u s ib id ., p. 357.
1|l-) Italics in the o riginal (ibid., p. 357).
Ib id .. pp. 358-60.
121 ‘G lo ssa ry ’, p. xxii, k a la n a i, an d xxvii, kJl-kalanai. Also, 'L arges!
Provincial O rg an izatio n s . . .’, O .S .M .S ., v. 45. n o .l, p. 35.
122 A .R .E ., 1909 (para. 45) refers to k a r a n j in connection w ith the artisan group,
the rathakciras.
I2i A .R .E ., 1964-5 (p. 15) regarding n o .305 o f th at year but also applies to the
follow ing reco rd s: 293, 300, 309 from V alik an d ap u ram (P e ra m b a lu r taluk, T iru-
ch irap alli). T his is in c o n tra st to the older view expressed in S .!.!.. v. 3, p. 275n.,
as cited by M a h aiin g am (S ou th Indian P o lity , p. 389n).
>24 S .I .i.. v. 4, n o .223, line 25ff.
refers to the founding of a trade centre (nagaram ) in the eastern
hamlet of the great brahm adeya o f Chidam baram. This hamlet
came to be called Gunam enagaipuram and was settled by mer­
chants ( vydpdrins ) and cultivators of high status (vellans ), lesser
m erchants (sah kappadiydr ), cloth m erchants (sdliyar) and fisher­
men (pattin avar), as well as lesser artisans (ktj-kalanai) including
carpenters ( taccar ), goldsmiths {ta ttd r ), blacksmiths (hollar), and
coarse d o th weavers (k o liy a r ).125
These early references to trading and artisan groups whose
economic activities were only tangentially related to agrarian
production, whose consumers were not restricted to specific
landed patrons but necessarily to a wider market, and whose
activities made them mobile in ways different from groups attached
to landed patronage are essentially references to what has here
been caiied the ‘outer core’ o f nuclear agrarian territories, nadus .,26
However, it is clear that such mobile occupational groups existed
as part o f the population of the inner core of nadus as well, in many
brahm adeyas during the tenth to twelfth centuries, and as part of
the general economic order outside these Brahman settlements;
it is also apparent that these mobile, mercantile and craft groups
constituted the prestigious core membership of the left-hand castes.
With the rise of towns during the twelfth century, these groups
attained a greater public presence and im portance as supporters
of Vedic temples. Artisans and m erchants are referred to by the
collective title pu ra-kalanai in an a .d . 1343 inscription from Pulip-
parakoyil (M adhurantakam taluk, Chingleput). This term speci­
fically included cloth merchants (sdliyar and ka ik k d ld r), general
m erchants (vanigar), and military as well as temple or palace
shopkeepers (senaiahgadiyar and koyil-ah gadiya r). 121 U rban places
o f the twelfth and thirteenth centuries, provided the basis for vastly
enhanced social power than was achievable by such groups in the
125 D iscussed in detail in ‘L argest P rovincial O rg an izatio n s . . Q .S .M .S .. v. 45,
n o .l, p. 35. S u b ra h m a n y a A iyer also discusses this m a tte r in E .I.. v. 22, n o .24,
‘U tta ra m a llu r In scrip tio n o f P a ra n ta k a 1', pp. 146-7.' Also note, Tam'd L e x ic o n ,
v. 2, p. 1197, k o lik a r , an d k o liy a n , loc. c i t .: an d T a m il L e x ic o n , v. 6, p. 3586, vanikam .
126 A n im p o rta n t th eoretical sta te m e n t o f these two classes o f castes is fo u n d in
D . P ocock. ‘N o tes on ja jm a n i re la tio n sh ip s’, C o n trib u tion s to Indian S ociology..
v. 6 (1962), pp. 78if.
127 A .R .E ., 1910, n o .298. R elevant p o rtio n s o f this in scrip tio n are tran slated
in E .I., v. 22, n o .24, p. 146. n. no.4.
rural context where the inner-core o f land-linked relations con­
tinued to dominate.
Having said all of this and having indicated the processes at
work during the twelfth and thirteenth centuries over substantial
parts of the South Indian macro region, it is still a fact that little
is known about any particular place. Ironically, impressions about
particular urban centres are much more vivid during the Sangam
Age than in any period thereafter until Vijayanagara times when
foreign accounts are a major source of information. One foreign
source for the twelfth century is the Chinese Chu-fan-chi by Chan
Ju-kua.128 Here, there is a reference to thirty-two trade centres
(‘pu-lo’ or pu ra) twelve o f which are on the west coast, eight in
the south, and twelve in the north. O f these, few can be identified
with certainty.129 A nother set of vague references to urban places
is found in inscriptions of the citram eli-periyan d tta r and other
records in which m erchant groups are mentioned, in a Piramalai
inscription of the early thirteenth century discussed by K. V. Sub­
rahm anya Aiyer,130 reference is made to eighteen established towns
with palatial buildings (m ddavJdu-padinettu-pattinam ), thirty-two
growing towns (valarpuram ), and sixty-four other urban places
(,kadigai-talvalam or sth dna).u l These numbers have the suspicious
m ark o f the conventional about them, and reliance can only be
placed on the list of towns in the last part of the inscription referring
to n agarattars from particular places in Pandyam andalam around
Piranm alai (including portions of Kulittalai taluk in Tiruchirapalli)
and some places in K ongum andalam which can be located.132
These latter places are called nagaram , p attin am , and puram-,
they are also called ur and perunteru (‘great ro ad ’).133
From the twelfth century on, there is considerable evidence
suggesting that trade and urban life were becoming much more
significant than these had been during the period of the great
Cholas when merchants (n agarattar) were almost always linked
subordinately to brahm adeyas and other large agrarian settlements.
128 T ran slated an d a n n o ta te d by F rie d rich H irth an d W .W . R o ckhill, St. P eters­
burg, 1911, pp. 93-102.
129 Ib id ., pp. 94-5, 99, n. n o .5.
130 ‘L arg est P rovincial O rg an izatio n s . . Q .J .M .S ., v. 45, n o .4, p. 270, Is. 1-4.
131 Ib id ., p. 280.
132 Ib id ., p. 272.
133 T a m il L e x ic o n , v. 4, pp. 2036-7, teru.
T hat many Brahman villages were in fact settlements of considerable
size, with diverse populations and functions, is quite clear. How­
ever, during the twelfth century merchants and artisan groups
began to separate themselves from the constraints of Brahman
and peasant rural control. This process of separation, in the
Corom andel plain, at least, was accompanied by new status claims
by merchants and artisans, as noticed in the discussion of the
left division of castes and as further noticed in temple honours
sought and gained by these groups. Essentially, the growing
importance of urban places was dependent upon the increasingly
vigorous trade throughout the macro region resulting from the
wealth and stability of the agrarian integration of the previous
two centuries. Settlements with substantial numbers of merchants
and artisans rose to new importance. An im portant additional
factor in the rise of such settlements was the increasing tempo
of temple building in which places of perhaps ancient sacred
importance, neglected during the period of the ascendancy of
brahm adeyas as sacred centres, were recognized anew. The require­
ments o f the various new or enlarged pilgrimage centres both
prom oted and facilitated the development of urban trade centres.
Simultaneously, an increasingly differentiated ruling class in the
countryside — the p eriya n a tta r — found it advantageous to co­
operate with those townsmen and the powerful, itinerant mercantile
groups with whom urban merchants were affiliated. This is evident
in the many inscriptions from most parts of the macro region
speaking of the citram eH -periyanáttavar and the tisai aiyirattu-
aiññürruvar to which K.V. Subrahmanya Aiyer and others have
drawn attention.134 Supra-locality rulers dramatically extended
the basis of their power beyond the locality peasantry from which
they had emerged and beyond an earlier alliance with rural Brah­
mans. Towns provided new sources of wealth and new sources of
prestige from association with temples and m athas. While still
essentially a rurally based ruling stratum where, as members of
supra-local bodies, they dom inated agrarian affairs, these erstwhile
local peasant leaders found in the weakness of the Chola overlord­
ship — for which they bore some responsibility — and in the
wealth and prestige of towns the means of enhancing their authority
154 K.V.' S u b ra h m an y a A iyer, 'L arg est Provincial O rganizations. . .’, O .J .M .S .
and K .A . N ila k a n ta Sastri and T .N . S u b ram an iam , 'H tm a v a ti Pillar Inscription
o f K u lo ttu n g ach o la (H i), Y ear 2 ', E .I.. v. 3 i, n o .37, pp. 274-5.
as it was not before possible to do. The m ajor shift from an agrarian
base o f power in which locality organization was founded upon a
balanced and close relationship between communities of Brahmans
and peasants (n a tta r) to one in which a part of the latter, as the
p eriya n a tta r , had attained hegemonic authority over several,
contiguous localities and their constituent communities appears
to have been accomplished by the fourteenth century. This change
produced an altered relationship with the Chola overlords of the
macro region.
The earliest manifestations of an altered power structure dates
from the twelfth century succession struggles am ong later Chola
rulers as noticed by K.A. Nilakanta Sastri and T.N. Subramaniam,
who state:
The organization o f the Sittiram eli-Periyanattavar which came into
existence about this tim e [the fo u rth regnal year of R ajadhirajachola 11,
c. a . d . 1172] very soon obtained a firm footing in the Tamil country and
was very influential th roughout the reign o f K u lottunga III. We may
not therefore be w rong in surm ising th at this organization paved the way
for the ultim ate success o f K ulottunga in getting the C hola th ro n e.135

During the next three centuries, this stratum o f powerful men was
to become merged in the new and highly m artial power system of
the Vijayanagara period.

135 K .A . N ila k a n ta S astri an d T .N . S u b ra m an iam , op, cit., pp. 274-5. It should be

noted th at this surm ise is q uestioned by the learned e d ito r o f E .I.. D .C . S ircar (op.
cit., 275n.).
 The Chola State and the
Agrarian Order

i
Among the greatest states in medieval India in its durability and
the scope of its authority was the Chola state. Considering only
the period of its dominance over m ost of Tamil country, the Chola
state lasted for three centuries, and, during its great days, from
about a . d . 950 to a . d . 1100 (the reigns of Sundara Chola to Kulot-
tunga I), Chola authority covered most o f the southern half of
peninsular India. It is therefore quite understandable that this
state should have received the very substantial attention which
it has from three generations of historians. W hat is singular,
however, is that notw ithstanding the deep study and appreciation
of the Chola state there is profound difficulty in determining the
relation of the state to the agrarian order of that time.
Why this condition should exist is more appropriately the subject
of a separate historiographical essay.1 Here, it may suffice to
consider briefly the way in which the Chola state is conceived in the
conventional historiography and how this conception has prevented
an adequate understanding o f either the political system or the
agrarian order of the age. It may also be noted that the problem is
not limited to the C hola state, but includes most South Indian
states prior to the establishm ent o f British rule in South India.
At the outset it should be noted that ideas about the relationship
of the medieval state and the agrarian system m ost often err with
reference to the state. This is unexpected since South Indian
historiography has been as deficient in economic-historical analysis
as most other regions o f the sub-continent. But, and perhaps this

1 See th e a u th o r ’s ‘T he S tate an d the A g raria n O rd er in M edieval S o u th In d ia:

A H isto rio g rap h ical C ritiq u e ’, in E ssa ys on South India, ed. B u rto n Stein, A sian
Studies in H aw aii, n o .15, 1975, T he U niversity Press, H aw aii, H o n o lu lu .
is true of pre-m odern historical research in other parts o f the sub­
continent, the evidence pertaining to the agrarian economy is the
same as th at pertaining to political arrangements. It is the same
limited corpus o f stone and metal inscriptions. South Indian
historians, with very few exceptions, have taken the data on
m ost economic arrangements as given in the inscriptions. These
records describe quite localized systems of agrarian production
and exchange.
W hen dealing with the state, however, the preconceptions of
the historian have been permitted to intrude upon the basic evidence
with the result that the Chola state as well as other medieval South
Indian states have tended to be conceived as great unitary states
under powerful kings whose will was worked through an elaborate
bureaucratic apparatus assisted by a powerful military establish­
ment. The m ore prudent and accurate depiction o f agrarian
arrangements results, to some extent, from a neglect of, or at
least a diminished interest in, such material m atters whereas the
the deep interest in and attention to p o litica l arran gem en ts have
served principally to distort notions about the state.
This contrast is m ost evident in the treatm ent of what is almost
always called ‘local government’. M ost discussions of the Chola
state speak of several levels of state organization, from the ‘central
royal adm inistration’ to the ‘local government’; m ost have also
presumed ‘self-sufficient’ villages. Actually, only two levels of
government have been dealt with. These are the king, with ‘his’
m ilitary and ‘his’ bureaucratic organizations, and the local level.
Linkages between these two levels are poorly defined; for the
m ost part, they are inferred. In some cases, linkages are altogether
ignored. M ost discussions of local, rural society, display a keen
appreciation for the well-developed, self-governing institutions
o f locality society; the village community, castes, and guilds.
Local taxation and the local m anagem ent of institutions are ex­
tensively docum ented since the vernacular portions o f inscriptions
recording grants pertain to these issues primarily. At times,
however, this appreciation becomes excessive as when N ilakanta
Sastri likens South Indian villages to the Rom an cities of Gaul
and cites Fustel de Coulanges to the effect that if the Gallic city
(and the South Indian villages to which it is compared) ‘was not
a free state; it was at any rate a state ’.2 In yet another place Nila-
2 The C o la s , p. 515.
 kanta Sastri speaks of Chola society as 'a federation of [‘hereditary
and voluntary’] groups’ appearing to have no territorial basis.3
As com pared to the richly documented vigour of locality in­
stitutions within the territory claimed by the Cholas and in which
their inscriptions are found, the ‘central government’ has a poorly
documented and often tortuously argued existence. N ilakanta
Sastri speaks of ‘the almost Byzantine royalty of Râjarâja’.4
Explicitly, he intends to contrast the kingship o f the earlier Classical
period with that o f the Chola period. He and others have certainly
produced convincing evidence of a conception of kingship during
the Chola period which is very different from that of the Classical
age. However, the contrast denigrates the changes in South Indian
kingship of the intervening Pallava age when a full-blown, Brah-
manical royal institution was brought into being and shaped the
Chola kingship. In the inscriptions of both kingships, a m ajor
expressive element consisted of royal support of grants to Hindu
temples and to Brahmans. It is noteworthy that ancient royal
sacrifices (y a g a s ) such as the asvam ëdha, prom inent in the early
Pallava period (i.e. before the eighth century), are rarely mentioned
in Chola inscriptions. Instead, the Chola kings devoted substantial
wealth to the construction of great temples such as the Brihadisvara
temple of Rajaraja I at Tanjavur and the Gangaikondacholapuram
temple of R ajendra I. These temples are considered as sepulchral
m onum ents by some scholars, including N ilakanta Sastri, who
are led to compare this feature to that of the God-king (dëvarâja )
conception o f South-east Asia.5 Still other temples possessed
portrait images of some Chola rulers who might have been
w orshipped.6 Some o f these issues will be considered further
below.
However, another kind of meaning m ust be attributed to the
phrase, ‘Byzantine m onarchy’. That is, the Chola king is seen as
the executive over a vast and powerful bureaucracy and military
organization. N ilakanta Sastri writes: ‘W hat distinguished [the
Cola government’] . . . was the superior executive strength it was
able to develop by bringing into existence a highly organized and
thoroughly efficient bureaucracy’.7 A great army and navy are
3 Ibid., p. 462. 6 Ibid., p. 453.
4 Ibid., p. 451. 7 Ibid., p. 462.
s Ib id ., p. 452-3.
presumed to have existed under the control and maintenance
of Chola kings. The existence of a ‘Chola navy’ is based upon a
few questionable references to ‘numberless ships’ in Chola inscrip­
tional preambles and upon the still mysterious adventurers of
Rajendra I in M alayan waters. M ore cogently, the presumed
m aritim e activity of the Cholas rests upon the old view still clung
to by m any Indian historians of a ‘G reater India’ and Hindu
colonization in South-east Asia, a view which has been brought
into serious question by recent work in the history of that region.8
Chola armies have a more substantial reality. The conquests of
Chola rulers from Sundara C hola ( a . d . 956-73) to K ulottunga I
( a .d . 1070-1120) are real enough, but these conquests neither
presuppose nor required a vast, standing army. M ilitary power
was possessed by numerous chiefs, by itinerant m erchant groups,
and others throughout the Chola macro region. These were locally
recruited and controlled forces. It is one of the m ajor flaws in
the reading of Chola evidence to see armed force as the monopoly
o f the Chola kingship, and it is im portant to recognize that upon
the specious notion of a vast, central and standing army much of
the justification for the Chola bureaucracy rests.
There is, in fact, no direct evidence of a central bureaucratic
organization with competence over the C hola macro region.
Such an organization is called into existence by the presumption
of a need for it. It is not the case that evidence of the existence
of a centralized, bureaucratized administrative structure prom pted
the reasonable question of what this adm inistration was for in
medieval South India. The idea of such an administrative structure
is based on the unsubstantiated supposition o f a central military
establishment, and the maintenance of this establishment and
other presumed ‘central’ functions was seen to require a centralized
bureaucratic state structure. It can be shown that there existed
neither a centrally controlled military organization nor centrally
coordinated redistribution of what were regarded as state re­
sources. M ilitary needs and the support of religious institutions
and Brahmans were met from resources allocated by those under
whose the control such resources w ere: locality chieftains, not the
8 E g.. O .W . W o lters, E a rly Indonesian C o m m e rc e ; A S tu d y o f the O rigin s o f
S n v ija y a , C ornell U niversity Press, Ithaca, 1967, pp. 64-5; a n d P aul W heatley,
The G olden K h e rs o n e se ; S tu d ies in the H is to r ic a l G e o g ra p h y o f the M a la y Peninsula
before A .D . 1500, U niv ersity o f M alay a Press, K u a la L u m p u r, 1961, pp. 201-3.
Chola rulers, except as the latter were themselves locality leaders
in the central parts o f the Kaveri basin.
N ilakanta Sastri has been the most eloquent spokesman for the
presumed centralized state structure o f the Cholas, but he is not
alone in this conception. In considering the agrarian system of
the Chola macro region, there is some advantage in examining the
position of A. A ppadorai, one of Nilakanta Sastri’s best students.
W ith his m entor, A ppadorai assumes that the royal government
permeated all aspects of life, and yet, A ppadorai’s discussion of
the state and the economy is one o f the briefest chapters o f his
large, two volume work, E conom ic Conditions in Southern India
(1000-1500 A .D .) ?
The key to A ppadorai’s understanding o f the relationship
between the state and the agrarian order is his understanding of
the term kadam ai. He, along with m ost Chola scholars construe
the term to refer to land revenue paid in kind to the ‘central govern­
m ent’. This is confirmed by his distinction between ‘land revenue’
(kadam ai) and a group o f other taxes which ‘were not likely to
reach the central government’.i° In attem pting to estimate the rate
of this payment to the ‘central government’, Appadorai presented
a table based on inscriptional references to payments due in kind
per unit of land.11 These references range from a . d . 1011 to
a . d . 1504. The assessments are given in or converted to kalam s

(a dry measure equal to six cubic feet) per veli or m d (equal to 6.6
and 1.65 acres respectively).12 The crops include paddy, sugar­
cane, sesamum, millets, and gingelly. As A ppadorai observes,
these data drawn from several parts of the Tamil plain at different
times yield great variation in rates. F or paddy alone in two eleventh
century inscriptions from Tanjavur, the difference is a thousand
per ce n t! Dissatisfied with the variation produced from these data,
A ppadorai cited an a .d . 1325 Pandyan inscription from Vada-
vanpatti (modern Sivaganga taluk, R am nad) where the kadam ai
is given as three kalam s o f paddy per m d yielding forty k a la m s .13
This rate of three-fortieths is rejected by Appadorai as ‘too favour­
able to the peasant’ and ‘exceptional’ ! The rate which is finally
9 A p p a d o ra i, op. tit., C h. V I, ‘T h e S phere o f the S ta te ', v. 2, pp. 661-732.
10 Ib id ., p. 681.
11 Ib id ., pp. 677-8.
12 Ib id ., ‘A p p en d ix ’, pp. 783-5.
13 A .R .E .. 1924, p ara. 38; 39/1924.
settled upon is the classical one-sixth, the conventional fraction
o f production due to the king according to sm riti texts from the
time of M anu.14 In support o f this rate, A ppadorai cited a text
o f the Vijayanagara period and statem ents o f the early British
adm inistrator, Thom as M unro.15
T.V. M ahalingam, in his South Indian P o lity ,16 repeated Ap-
padorai's argum ent without, however, m entioning his work. Like
A ppadorai, M ahalingam does not question to whom the ka dam ai
was paid, it being assumed that the land revenue went to the
‘central government’.17 Also like A ppadorai, and all other writers
who hold the same position, M ahalingam does not ask how the
k a d a m ai in kind was converted into income for the ‘central govern­
m ent’. Considering that the kad a m a i is not expressly stated in
the inscriptions o f the period to be a payment to the ‘central govern­
m ent’, crucial im portance attaches to the question of how the
proceeds from agricultural production were in fact realized by the
Chola rulers in the Kaveri. The same question does not arise with
the Vijayanagara state on the Tungabhadra with which both
M ahalingam and Appadorai deal because both accept foreign
accounts which describe tribute paym ents from subordinate
Vijayanagara warriors as the m ajor source o f state revenue.
Several factors operate to explain the persistence of the belief
that the Chola state exercised m anagem ent over the agrarian
economy of the time. One is ideological. Three generations of
South Indian historians have taken pride in w hat they perceived
to be elaborate, centralized governments during the medieval
period. This sentiment was reinforced by ancient and medieval
political treatises which spoke o f powerful m onarchical states and
by later British writers with their own commitment to centralized
bureaucratic government and with their, often self-deluded (and
self-serving), notion that the government of the East India Company
simply restored proper government to South India after a century
or m ore of ‘chaos’. The robustness o f the belief of historians
14 The O rdin an ces o f M anu, ed. an d tran sl. A .C . B urnell a n d E .W . H o p k in s,
L o n d o n , 1884, ch. V II, verses 130-1. A ctually, three rates are given h ere: one-sixth,
one-eig h th , a n d o ne-tw elfth.
15 A p p a d o ra i, op. cit., v. 2, pp. 679-80.
16 O riginally p u b lish ed in 1954 by the U niversity o f M a d ra s ; revised edition,
1967 used here, pp. 158-72.
17 M a h alin g am does q u estio n the one-sixth rate, h o w ev er; ibid., p. 166.
and others in centralized polities of the medieval period apparently
discouraged attem pts to form ulate a theory of government which
conformed better with the facts known of the medieval period.
This was true despite the suspicion with which literary sources
of the period were held by N ilakanta Sastri especially.18 Even
M ahalingam, though less troubled by the often didactic m ti texts,
appears cautious in his use of a ‘m anual of adm inistration’ of the
Vijayanagara period attributed to M adhvacharya. First, he cites
the following statem ent from the Parasara M a d h a v iy a :
A s the florist in the garden plucks blossom s successively p u t forth and
does not eradicate the flowering shrub, so should the king draw ing revenue
from his subjects, take the sixth p a rt o f the actual pro d u ce; like the m aker
o f charcoal extirpating the tree burns the whole plant, let no t the king
so treat his subjects.19

M ahalingam then cautions t h a t:

. . . w orks (ike the ParaiaramadhavTya deal m ore w ith the theoretical
side o f taxation than the practical side o f it; and unless such indirect
and less reliable evidence are co rroberated by the evidence o f contem porary
inscriptions it is difficult to accept their value.20

Notwithstanding this prudent reservation, M ahalingam as­

sumes that the central government — the king — received a portion
o f agrarian production from even distant territories. This is as­
sumed w ithout questioning how these royal claims to revenue
were realized. Medieval n fti formulas are little more reliable than
the speculations o f later British adm inistrators with respect to
the function o f the state during the medieval period. Both assume
a uniformity o f usage which is not substantiated by more reliable
evidence; both also exaggerate the legitimacy and capability of
South Indian medieval rulers to income from tracts over which
they did not exercise direct dominance.
The m atter of uniform ity has received recent attention from
N. Karashim a and B. Sitaraman. Based on one of the most exhaus­
tive surveys o f Chola records,21 these two scholars analysed Chola
19 M a h alin g am , S o u th Indian P o lity , p. 167.
20 Ib id ., p. 169.
21 T hey estim ate th a t o f ap p ro x im ately 9,000 C h o la in scrip tio n s th at h ave been
noticed, som e 3,500 have been published. O f these, fo rty per cent refer io revenue
term s, lan d g ran ts, an d lan d transfers. N o b o ru K arash im a a n d B. S itaram an ,
'R ev en u e T erm s in C hola In scrip tio n s’, J o u rn a l o f A sia n a n d A frican S tu dies. T okyo,
n o .5 (1972), p. 88.
inscriptions from a . d . 846 to a . d . 1279 (the reigns of Vijayalaya
to Rajendra III) in the two great territories of the Chola macro
region: Cholam andalam and Tondaim andalam (or, to use the
Chola period designation for the latter, Jayankondacholam an-
dalam). Taking terms which occur ten times or more, twenty-
seven revenue term s are analysed chronologically and topographi­
cally. It was found that eleven of these frequently used terms
(forty per cent) are found in Tondaim andalam only. The authors
conclude that ‘there existed a different administrative procedure
or economic set-up in one m andalam as distinguished from the
other’.22 While this analysis on a m andalam basis minimized
differences within each of the great territories — differences that
would be very great — these findings should have a sobering
effect upon those who propound, or more often simply imply,
a unitary state theory as do m ost writers on the Cholas.
M oreover, the analysis o f Karashim a and Sitaram an casts deep
doubt upon the view that some portion of agrarian production
from even the central core of the territory claimed by the Cholas
found its way to a central treasury. From K arashim a’s previous
writing, this view o f a centralized fiscal structure may have been
expected.23 Few o f the twenty-seven most frequently mentioned
term s refer explicitly to cash payments, e.g. kasu ayam and veli
kasu. The former term, used in Tondaim andalam , means quite
literally a tax or paym ent in m oney; the latter term which occurs in
both of the m andalam s means a charge o f one kasu (probably
a copper coin) per veli, 6.6 acres of cultivated land.24 Others among
the twenty-seven terms may have been cash payments, of course.
K urrat-tendam was a fine levied in Tondaim andalam and ta tta r-

22 Ib id ., p. 90.
23 A m o n g the w ritings to w hich reference can be m ade are: ‘A llur an d Isana-
m a n g a la m : T w o S o u th Indian Villages o f th e C o la T im es’, P ro ceed in g s o f the F irst
In tern a tio n a l C onferen ce S em in a r o f T a m il S tu d ie s, K u a la L u m p u r, 1966, pp. 426-
37; ‘T h e P ow er S tru c tu re o f the C ola R u le ’, P resented to the S econd C onference
S em in ar on T am il Studies, Ja n u ary , 1967; ‘H isto rical D ev elo p m en t o f S outh In d ian
S ociety’ (‘M in am i in d o shakai no re kishi teki h a tte n ’), in Indo shi o k e ru toch i seido
to k e n ryo k u R o zo , ed. T o ru M atsui a n d T o sh io Y am azu k i, T o k y o U niversity Press,
T o k y o , pp. 73-105.
24 ‘E p ig rap h ical G lo ssary o f T erm s R e la tin g to T axes and o th er D ues C u sto m ary
a n d F eu d al (P rep ared fo r a T o p o g ra p h ic a l L ist o f Inscriptions in the M a d ra s and
K erala States), D e p a rtm e n t o f A ncient H isto ry a n d A rchaeology, U niversity o f
M a d ras, 1967 (typescript).
p a tta m was a tax on goldsmiths applied in both territories.25
Both could have been payable in cash. However, most o f the
terms enum erated by Karashim a and Sitaraman refer to payments
in labour or in kind. The point previously raised about the m anner
in which payments in kind were converted to money or otherwise
made available to the ‘central government’ is underscored by
these findings.
But, it is the kadam ai from all parts o f the Chola territory which
is regarded as the prim ary source o f revenue for the Cholas as it
was for the British. This term occurs with six others which K ara­
shima and Sitaram an found were used in both of the m andalam s
and which occur with at least twice the frequency of the other
twenty terms. Glossaries o f inscriptional terms o f the Chola period
identify kad a m a i as a land tax payable in kind to governm ent;26
most would agree with Appadorai that ‘government’ m eant ‘central
government’. M ore generally, however, the words k a dam ai or
ka4am mean ‘duty’ or ‘obligation’ then as now.27
There is no question that k adam ai was used in Chola times to
designate a tax, but there are questions of whether it was a tax
paid to the ‘central government’ as implied in most writing and
whether it was a tax on land only. A ppadorai’s view, as already
noticed, is that kad a m a i was ‘land revenue’ paid to the ‘central
government’ and was distinguishable from other taxes on agrarian
production which were ‘utilized for maintaining irrigation works,
for paym ent of village officers, and for the m aintenance of
temples’.28 The last three categories of payments are seen as
local cesses. Similarly, in the analysis of Karashim a and Sitara­
man, the seven m ost frequent revenue terms found in both Chola-
m andalam and Tondaim andalam include kadam ai. These writers
do not consider the meanings of the terms they have classified
chronologically and topographically, but the six other terms would
appear to be local dues exclusively. Thus, the terms echchoru,
25 K ara sh im a a n d S itara m a n , op. cit., pp. 90-1.
26 See th e glossary o f M a h alin g am , Sou th Indian P o lity , p. 426 an d S .I .T .I .,
‘G lo ssary ', p. xx. A lso T a m il L e x ic o n , p. 659.
27 T h e term s figure p ro m in e n tly , if dubiously at tim es, in th e d o cto ral thesis o f
Stanley J. H eg in b o th am , ‘P a tte rn s a n d S ources o f Indian B u reau cratic B ehaviour;
O rg an izatio n al P ressures an d the E th ic o f D u ty in a T am il N a d u D ev elo p m en t
P ro g ra m ', M a ssach u setts In stitu te o f T echnology, 1970. H ere, k a d a m a i is u n ­
acco u n tab ly tra n slite ra te d 'k a rd e m a i', an d is eq u ated w ith the term dh arm a, p. 321.
28 A p p a d o ra i, op. cit., n, pp. 680-1.
m u ttaiy-al and ve tti refer to obligations by local groups to perform
labour; the term antarayan refers to taxes levied by local bodies.
K u dim ai ,29 another of the frequently m entioned term s is generally
understood to refer to all taxes and dues except those from the
land; ku dim ai is thus contrasted with kadam ai. Finally, the term
ta tta r-p a tta m is a professional tax levied on goldsmiths and pos­
sibly realized in money. Excluding the last term and the generic
terms ku dim ai and kadam ai, the other four m ost frequently m en­
tioned terms appear to be local dues. In the absence of any known
process for the conversion o f payments in kind to money, the
only revenue source which could practicably have been realized
by the ‘central government’ was the professional tax. There is
no evidence to support the view that kad a m a i was a payment
from localities to the ‘central government’. The generic revenue
terms kadam ai and kudim ai appear simply to have been terms
for taxes; they imply no reference to the source (i.e. whether from
the land or not) nor the recipient (i.e. whether local or extralocal).
N ilakanta Sastri notes that the meanings o f kad a m a i and kudim ai
are ‘quite close in m eaning’; both refer to ‘duties of the k u dis ’
or people, and he assumes that both terms refer to payments to
the ‘K ing’s government’.30 The latter notion is based on the
extremely slender evidence of phrases in inscriptions such as the
following of R ajaraja I from Tiruvaduturai in C holam andalam :
‘Any kind of ku dim ai [is] due at the Sacred Victorious Gate [tiruk -
korra-vasal including] the taxation (varippadu ) levied by the ur
(town or village), and any other type of ku dim a i .’’ 31 The phrase
tiru k-korru -vdsal, upon which much hangs in this interpretation,
can also mean ‘the masonry gateway before the chief god o f the
temple’, and since the record cited pertains to an arrangement
for the provision o f dues to a temple, the latter meaning of the
phrase may be what was intended. In several places, N ilakanta
Sastri provides evidence against his own definition. He cites
inscriptions in which the k adam ai was possessed not by the ‘king’s
government’, but by various kinds o f settlements including those
under the control o f temples (devadana and tiruvidyaitam ), Jains
( paU iccandam ), learned persons (agarapparru, m adapuram ), village
29 Tam il L exicon, p. 971.
30 N ila k a n ta S astri, The Colas, pp. 522-3.
31 Loc. c it.; a su m m ary o f this in sc rip tio n is fo u n d in The Colas, i, p. 505, referring
A .R .E . 1925, n o .121. A lso, Tamil Lexicon, p. 1917 (tirvdcal) a n d p. 1167 ( k o r r u ) .
servants and officials (jTvataparru ), and soldiers (padiaparru and
vanniyapparru).i2 Moreover, he provides examples of the use of
the term kadam ai to refer to dues on productive capacity other
than regular field agriculture.33 This included several references
to kadam ai for oil-mongers and one to kadam ai on areca trees.34
This discussion of Chola revenue has intended to draw atten­
tion to the largely conjectural and imputed nature of the Chola
state in the existing historiography. It is the purpose of this chapter
to propose a model of the Chola state which is fundamentally
different from the existing model, which also fits the known
evidence of the Chola period better.

II
In order to consider a better model of political arrangements
in medieval South Indian states than the existing one, it is neces­
sary to recognize that there are many kinds of states, of which the
unitary, centralized state is but one. It is, however, precisely
that one which m ost South Indian historians assume to have
been normative during the medieval period, and, except when
the king was weak or the kingdoms troubled by natural disasters
or invasions, these historians claim actually to have existed. In
fact, this type of unitary state did not exist nor could it have
existed in medieval South India any more than in medieval India
as whole, with the possible exception of the M ughal state of
the seventeenth century; nor did such a state exist in most of the
world prior to the industrial revolution which provided the tech­
nology and mobile force required to sustain unitary states as we
know them.
In this section, another type of state formation is proposed as
m ore in conformity with extant evidence. This is the pyramidally
segmented type of state, so-called because the smallest unit of
political organization — for example, a section of a peasant
village — was linked to ever more comprehensive units o f political
organization of an ascending order (e.g., village, locality, supra-
locality, and kingdom) for various purposes, but that each unit
stood in opposition to other, similar units (e.g., one section o f a
32 N ila k a n ta S astri, The C o la s, p p . 505, 585.
-'3 ib id ., pp. 526, 532.
34 Ibid., p. 532.
village as against another) for other purposes. The most fully
elaborated discussion of this type of political system has been
provided by Aidan Southall for an East African society, the Alur.
In his study,35 Southall is concerned with a political order which
he contends is different from, and c o n tra sts w ith, the unitary
state, but is yet a ‘state’. The ‘state’ in m ost m odern usage is seen
to be comprised of four elem ents: territorial sovereignty, centra­
lized government, specialized ruling or administrative classes, and
monopoly of legitimate force or political control by the centre.
To the contention that state forms other than unitary ones are
‘transitory’, Southall responds with the description of a stable,
and, he contends, widespread form which he calls ‘segmentary’.
The characteristics of the segmentary state are the following:

(1) T erritorial sovereignty is recognized but limited and essentially

relative, form ing a series o f zones in w hich auth o rity is m ost absolute
near the centre and increasingly restricted tow ards the periphery, often
shading off into a ritual hegemony.
(2) T here is centralized governm ent, yet there are also num erous
peripheral focuses o f adm inistration over which the centre exercises
only a limited control.
(3) T here is a specialized adm inistrative staff at the centre, but it
is repeated on a reduced scale at all the peripheral focuses o f ad m inistra­
tion.
(4) M onopoly o f the use o f force is successfully claimed to a limited
extent and w ithin a lim ited range by a central authority, b u t legitimate
force on a m ore restricted order inheres at all the peripheral focuses.
(5) Several levels o f subordinate focuses may be distinguishable,
organized pyram idally in relation to the central authority. The central
and peripheral authorities reflect the sam e m odel, the latter being reduced
images o f the form er. Sim ilar pow ers are repeated at each level w ith a
decreasing range; every au thority has certain recognized powers over the
subordinate authorities articulated to it, and form ally sim ilar offences
differ in significance according to the o rd er to authorities involved in them.
(6) The m ore peripheral a subordinate au th o rity is the m ore chance
it has to change its allegiance from one pow er to another. Segm entary
states are thus flexible and fluctuating, even com prising peripheral units
which have political standing in several adjacent pow er pyram ids which
thus become interlocked.36

A lu r S o c ie ty : A S tu d y in P ro cess a n d T y p e s o f D o m in a tio n , W . Heffer an d Sons

L td ., C am b rid g e, 1956.
36 Ib id ., pp. 248-9.
 Professor Southall’s explicitly comparative approach, his efforts
to consider political systems other than the Alur in contem porary
Africa and in other times, invites further consideration. Special
attention may be given to three o f the elements of Southall’s
definition: sovereignty, centralized government, and administrative
specialization.
Territorial sovereignty in the unitary state and in the segmentary
state is different. In the unitary state, effective political control
and adm inistration defines the territory of the state; it also estab­
lishes, or constitutes, the quality of sovereignty, that is, the m anner
in which political authority and political power coincide. While
in the unitary state there is likely to be no substantial gap between
political control and authority, or what Southall speaks of as
‘ritual hegemony’, in a segmentary state these two aspects of rule
are markedly divergent. Thus Southall, speaking of a segmentary
political situation, states; ‘ritual supremacy is often accepted where
political control is not, and segmentary states may characteristically
be more highly centralized ritually than politically.’ 37 This affects
the nature of central authority as well as of territory, of course,
but the implications for the latter are im portant. If one accepts
the validity o f the distinction between ritual and political aspects
of rule, as one m ust in order to deal with the concept of the seg­
mentary state as Southall presents it, then it would appear necessary
to consider two different notions of ‘territory’ at a given time for
a particular segmentary state. One sense o f territory would be the
scope of ritual supremacy, the other of political control. This is
the distinction upon which Southall appears to base his notion of
a ‘series of zones’. It is necessary to recognize that in a segmentary
state there is a dual or divided idea of territoriality. Whereas in a
unitary state this dual sense of territory occurs only when some­
thing has gone wrong, in the segmentary state, the fact and legitimacy
of this duality is of crucial theoretical and empirical significance.
The dual sense o f territorial sovereignty as, on the one hand,
an essentially ritual sort exercised by a king in a segmentary state
37 Ib id ., p. 261. In a later sta te m e n t, S o u th all seem s to m odify w h at is here an
a p p a re n t equivalence o f ‘ritu a l h eg em o n y ’ an d ‘a u th o rity ’ by sta tin g t h a t : ‘A u th o rity
is th e legitim ate exercise o f im perative c o n tro l (i.e. \ . . the p ro b ab ility th at a co m ­
m an d will b e o b eyed’) ; T y p o lo g y o f S tates a n d P olitical S ystem s’ in M. B an to n
(ed), P olitical S ystem s and the Distribution o f Power, A .S .A . M o n o g rap h s, n o .2
(1969), T av isto ck P u b licatio n s, L o n d o n , p. 120.
and, on the other hand, as an essentially political or controlling
sort which the king exercises in his own domain, but which is
appropriately exercised by subordinate rulers in their domains, is
seen as appropriate in the medieval Indian situation. The terms
k satra and rajadharm a in dharm asdstra texts denote the distinction.
In the recent work of Lingat,38 these terms are contrasted in
two different ways. First, the terms contrast with respect to the
scope of kingly ru le: ksatra perceives the king as a fully competent,
independent political actor charged, as is the father in the family,
with the protection and control of his subjects; rajadharm a perceives
the king as an actor of limited power within an interdependent
system of hierarchical relationships, who bears the full moral
consequences o f his own actions and those o f his subjects.39 The
second contrast is more relevant.
[Rajadharma]40 is a universal rule in this sense, th a t every king is subject
to it and suffers its sanctions, w hatever the extent o r situation o f his
kingdom . M oreover it is a duty and an obligation o f a personal ch aracter
which is incum bent on the king’s conscience and obtains stability only
through his will.41
Ksatra . . . is a pow er o f a territorial character, exercised within a given
territory and stopping at the frontier o f the realm. . . . O f the same
nature as property, it implies a direct pow er over the soil. T h at is why
the king is also called svdmin, a w ord which can be applied equally to a
pro p rieto r as to a husband or a chief, and w hich denotes an im m ediate
pow er over a thing o r over a person.42

A similar contrast is suggested by Louis D um ont in his discus­

sion of danda and artha as elements of what he calls ‘ “ conventional” ’
or ‘ “ rational” ’ kingship in India. D um ont contrasts danda,
or ‘legitimate force’, and artha, or ‘interest. . . at once economic
and p o litica l. . and notes that danda, m ost abstractly conceived
is ‘a kind of immanent power of justice . . . m ore or less identical
38 R o b e rt L in g at, The C la ssic a l L a w o f India, tra n s. w ith ad d itio n s by J. D u n can
M . D erre tt, U niv ersity o f C alifo rn ia Press, Berkeley, 1973. O riginally published
in 1967 u n d er th e title, L e s sources du d ro it dans le s y s te m e tra d ito n n el de I’ln de,
M o u to n an d C o ., P aris.
3« Ib id ., pp. 211-12.
40 ‘D h a r m a ’ is used in th e original b u t L ingat n o ted th a t ‘it is perfectly correct
fo r th e M a h a b h a ra ta to declare th a t all d harm as are com prised in the raja-dharm a
an d th a t all have ra ja -d h a rm a at th eir h e a d ' (ibid., p. 208).
“i Ib id ., p. 212.
42 Loc. cit.
with dharma. . . ,’43 D harm a is thus ‘action in conformity with
the universal norm , and hence disinterested, while artha is action
in conform ity with interest without regard to the universal
norm. . . ,’44
Southall’s second characteristic of the segmentary state, ‘centra­
lized government’, requires clarification. The segmentary state
has two kinds of ‘centres’, or rather, centres exist in two conceptual
and empirical senses. As to the first sense, the segmentary state
exists as a state (and not a congeries of independent political
entities) only insofar as the segmentary units comprising it recognize
a single ritual authority — the king. It must be this recognition
which provides some of the legitimacy for constituent segmentary
units — those ‘num erous peripheral foci of adm inistration’ in the
state — which are themselves ‘centres’ in the second sense. In
this connection, it is difficult to understand Southall’s suggestion
that one difference between segmentary and unitary states (and
the factor chiefly involved in the transform ation from the former
condition to the latter) is the concept of legitimacy.45 Southall
supposes that legitimacy is lacking in the political relationships
of the segmentary state. Yet, what else but a ‘belief in ‘ “ legitimacy” ’
(to follow Southall in the use of W eber’s definition of legitimacy)
could explain the efficacy and durability of the Alur state and
other systems to which he applied the label, ‘segmentary state’,
it would seem to be ‘legitimacy’ alone that can be understood in
Southall’s statement that ‘in a segmentary system tribute is received
in direct return for ritual and jural services rather than in recognition
of any regular fiscal obligation’.46 It is a similar sense of legitimacy
which K rader presumably has in mind when he states:47
. . . in any segm entary polity, the local chiefs or o th er authorities do
not give their political pow er io the central ruler; the political pow er
rem ains in th eir hands. N evertheless, a central office m ay be established
in these kingdom s with a sym bolic function o f representing the unity
o f the people and the land.

43 ‘T h e C o n c ep tio n o f K ingship in A ncient In d ia' in L. D u m o n t, R e lig io n /P o litic s

a n d H is to r y in In d ia : C o lle c te d P a p e rs in Indian S o c io lo g y , M o u to n an d C o., T he
H ag u e, 1970, p. 76. H e c o n tra sts ‘co n v en tio n al' w ith 'd iv in e', p. 71.
44 Ib id ., p. 78.
45 S o u th all, op. cit., p. 252.
46 Ib id ., pp. 261-2.
47 L aw rence K ra d e r, F o rm a tio n o f the S ta le (F o u n d a tio n o f M o d e rn A n th ro ­
pology Series), P rentice H all, E nglew ood Cliffs, N ew Jersey, 1968, p. 35.
 If legitimacy does not attach to ritual centralization in the segmen­
tary state, Southall’s second defining characteristic on centralized
government becomes puzzling. For, it appears contradictory to
assert that ‘there is centralized government’ at the same time that
there are ‘peripheral foci of adm inistration over which the centre
exercises limited control’. It m ay be granted that there are limits
to central control in even the most m odern, unitary state. However,
the distinction which is emphasized in considering the segmentary
state is th at overarching central p o litic a l control may not be very
im portant at all. This is a very different conception from that of
‘centralized government’. In a segmentary state, while political
control is appropriately distributed among many throughout
the system, ritual supremacy is legitimately conceded to a single
centre.
A similar and related question arises in the ‘m onopoly in the use
of force’, Southall’s fourth characteristic of the segmentary state.
Involved here is the same kind of apparent contradiction as that
alluded to in the discussion o f ‘centralized government’. It is
difficult to accept that the ‘m onopoly of the use of force’ has been
‘successfully claim ed’ by the centre when the legitimate, even
if restricted, use of legitimate force also ‘inheres at all peripheral
foci’. This is the same as accepting that a centralized government
could exist which had almost no political control over parts of
the system. These apparent paradoxes are mitigated, if not resolved,
in the distinction between political and ritual authority which has
already been discussed and in the relationships between centres
and their internal subordinate parts, to which attention can now
be directed.
Southall’s fifth characteristic states in part:
Several levels o f subordinate foci m ay be distinguished, organized pyram i­
dally in relation to the central authority. The central and peripheral
authorities reflect the sam e m odel, the latter being a reduced image of
the form er.

The concept of pyramidal segmentation has the greatest saliency

in Southall’s form ulation; it may be the most distinctive feature
of his form ulation of the segmentary state, for the conception of
pyramidal segmentary organization assumes an upward reaching
set of localized units of society, such as peasant families, to the
m oral centre of society, the king. As used by Durkheim, the principle
of segmentary structure denoted the horizon tal repetition of like
social units at the base of society, units such as clans and lineages,
for example. Later anthropological analysis, especially that of
Evans-Pritchard on the Nuer of Africa have noted that there was
also a form o f vertical segmentation. Here, the highest segment
might be a tribe, or m ajor section of a tribe, beneath which nested
what Evans-Pritchard called secondary and perhaps even tertiary
sections of the same tribal unit. The lowest level of this segmentary
structure might include several villages,48 and at this level are to be
seen many of the characteristics of the tribe itself: a distinctive
name, shared bonds of sentiment, and a territory of its own. How­
ever, each segment is also internally segmented with oppositions
among its parts, Thus, Evans-Pritchard writes, \ . . a tribal segment
is crystallized around a lineage of the dom inant clan of the
tribe. . .’ 49 in opposition to which stand other lineages. Further:
The political system is an expanding series o f opposed segments from the
relations w ithin the smallest tribal section to inter-tribal and foreign
relations, for opposition between the segments o f the smallest section
seems to us to be o f the same structural character as the opposition between
the tribe and its neighbours. . . ,50

The im portance of this conception vests in its simultaneous

attention to the linkage of social units at various horizontal levels
— even very local levels — and the integration of diverse social
units at increasingly more comprehensive spatial and societal
levels, it is of course recognized that there are profound differences
between African societies and those of India, including the far
weaker (but not absent) clan and lineage structures at the lowest
levels o f society in India, and the overarching ideological and ritual
integration achieved under Indian conceptions of kingship. But
such differences do not alter the essential structural similarities
of certain fundamental relationships in these two very different
kinds o f social orders.
The basic segments of the South Indian medieval segmentary
political system were nadus under the leadership of chiefs. In the
Chola period, such personages held titles such as udaiyar, arasar,
48 E .E. E v an s-P ritch ard , The N u er, O xford U niversity Press, N ew Y o rk , rep rin t
1971, p. 139; cited by S outhall, op. cit., p. 249.
49 E v a n s-P ritc h a rd , op. cit., p. 142; also pp. 211-14.
so Ib id ., p. 148.
mummudi, or m üvendavélár. The dom inant basis of opposition
o f these localized units o f society was not that o f ethnically or
culturally differentiated peoples, as in the case of the Alur des­
cribed by Southall. In medieval South India opposing elements
within the nadu units of society were of a different nature, and
often asymmetrical. These would include the opposition between
families of chiefs and the dom inant castes from which they had
emerged, between locally dom inant landed groups and subordinate
ones; between agricultural and non-agricultural groups, between
established castes o f a locality and newcomers or outsiders, and
among sect and cult groups. Many of these oppositions took
concrete form in the right and left caste groupings.
W ithin each nadu segment, social units in balanced or com­
plementary opposition may have conceded to a chief a degree o f
executive authority — political and ritual power — sufficient to
satisfy the tasks of governance within a territory, but this was a
limited concession. It was limited by the demands of constituent
groups within the segment that their separate identities and internal
regulation be respected and preserved from usurpation by the
power of the chief on the one hand, and from assimilation by other
groups, on the other hand.51 M aintenance of the internal relations
among groups within the segmentary unit accomplished more-
than the limitation of the chiefs power; it also m ilitated against
pressures from beyond the segmentary unit which might threaten
this internal constitution. Thus, preservation of the oppositional
structure of groups within a nadu, while limiting the chiefs authority
over the unit, simultaneously strengthened the office of the chief
by assuring support to him in protecting the nadu from external
aggression. Chieftainships of this sort may additionally, though
only temporarily, be supported by intra-segmentary cooperation
in acts of aggression against others. As long as the disparate
interests of groups within the segmentary unit were maintained,
the segmentary political order had stability. A segmentary unit
under gifted leadership or special circumstances, such a natural
catastrophe, may have shifted to the status of a more powerful
chiefdom or, on the other hand, it may have yielded to superior
force and become integrated into a supralocal, centralized political
system. But, neither of these transform ations are likely to be
51 S e e L in g a t’s d iscussion o f D h a rm a sa stra on th is p o in t; op. cit., pp. 246-8.
either easy or enduring. As long as the subordinate social units
preserved the tension of opposition to each other and to the proxi­
mate central authority, i.e. the local chieftain, each segmentary
unit contributed to the segmentary character of the whole.
Each segment replicates the overall pattern of central and
peripheral relationships and at every level there are centres. South­
all states:
The distribution o f pow er in the segm entary state is characterized by the
fact that, w ithin any one segm ent, at any level o f the pyram idal structure,
there is at any one m om ent a certain degree o f m onopoly o f political
pow er, developm ent o f adm inistrative staff an d definition o f territorial
limits, w hereas, w ithin the system as a whole, the political relations of
the various segm ents are determ ined by m uch the sam e factors as in the
case o f segm entary societies w hich have no political specialization at all.
Such specialized pow er as exists at the top o f the pyram id becomes p ro ­
gressively w eaker in p ro p o rtio n to distance from the center, b u t the
degree o f pow er centralization w ithin segments m ay vary very little from
one segm ent to an o th er w ithin each pyram idal system .52

All subordinate centres of a segmentary order are bound together

by their joint recognition of the ritual sovereignty of the highest
central office or personage considered as the authoritative source
o f ritual cohesiveness for the state as a whole. W hat distinguishes
centres from each other at various levels of the system is the vertical
discontinuity of power relations. Hence, the scope of political
control falls upon diminished numbers of persons as one descends
from the apex o f the segmentary state (the raja) to the base, the
local chief (e.g., udaiyar). But the nature of authority of each
centre is the same.
It is essentially this point of many centres in relationship to
superior and subordinate levels with which Southall is concerned
in his distinction between pyramidal social structures and hierarchi­
cal power structures.53 In the former every legitimate kind of
political authority and control can be found at all levels, but these
operate upon a diminishing constituency as one moves from level
to level. In hierarchical power structures, different kinds of
authority and control are appropriate to centres at different levels,
and what m arks the passage from one to another level is a particular
bundle of executive authority and power. That which is ‘reduced’
52 S o u th all, op. cit., pp. 251-2.
53 Ib id ., pp. 250-1.
in the ‘reduced image’ of pyramidal levels is the range of com ­
petence not the kind o f political control.
‘Specialized administrative staff, the third basic element of
Southall’s definition refers to agents of a centralized authority
who are the means by which centralized authority and power are
articulated. However, in his usage and that of most students of
pre-industrialized societies, apart from China, perhaps, the reference
is not to specialists of the sort associated with modern, unitary,
bureaucratized states. The latter are functional specialists res­
ponsible for attending a fractional aspect of central political
control such as when a m agistrate adjudicates disputed claims
according to codified, statutory law. W hat appears to be meant
by ‘specialized administrative staff in Southall’s conception o f the
segmentary state is the existence within a society of persons with
distinct political roles in contrast to the fused character of kinship
and politics in stateless societies. M oreover, in Southall’s form ula­
tion, the reference is not to administrative specialists in the sense
that a m agistrate or a tax collector is one, but, in fact, to a general
agent of central authority, a m an tri in the strict sense of the
term.
Considering the very im portant distinctions in the concept
of the segmentary state which are made between ritual and
political authority and power on the one hand, and central
and segmentary focuses on the other, it would be expected
that the adm inistrative personnel — those with clear political
roles — are of two kinds. First, there are those who are agents
of the central authority’s actual political control within the
limited sphere o f th at competence; secondly, there are those
who are agents o f the central authority’s ritual function in
binding subordinate authorities together into either subordinate
pyramidal segments or state systems comprised of such units.
Perceived in these terms, Southall’s third characteristic may
perhaps be stated as follows:
Specialized adm inistrative personnel are o f two kinds. O ne is essentially
involved in political control, and the scope o f their activities m ay n o t
extend beyond the segm entary unit (recognizing that the segm entary
unit o f the great chief may be m uch larger than others). The second
kind o f adm inistrative personnel is involved in essentially ritual activities
and tends to be inter-segm entary in scope.
A partial summary of the clarification and elaboration of South-
all’s conception o f the ‘segmentary state’ which have been discussed
above may now be offered.54
1) In a segmentary state sovereignty is dual. It consists of
actual political sovereignty, or control, and what Southall terms
‘ritual hegemony’ or ‘ritual sovereignty’. These correspond in
Indian usage to k sa tra and rajadharm a respectively.
2) In a segmentary state there may be num erous ‘centres’ of
which one has primacy as a source o f ritual sovereignty, but all
exercise actual political control over a part, or segment, o f the
political system encompassed by the state.
3) The ‘specialized adm instrative staff’ — what in some unitary
states would be called ‘the bureaucracy’ — is not an exclusive
feature o f the prim ary centre, but is found operating at and within
the segments of which the state consists.
4) Subordinate levels, or ‘zones’ of the segmentary state may
be distinguished and the organization o f these is ‘pyram idal’.
T hat is, the relationship between the centre and the peripheral units
o f any single segment is the same — in reduced form — as the
relationship between the prime centre and all peripheral focuses
o f power. There is a contrast here with hierarchical forms of
political organization in which, at successively subordinate levels
o f a system there are different kinds o f executive authority whereas
in the segmentary state, executive authority is the same at the prime
centre and at any subordinate segmental centre except that it is
exercised over fewer people. In the Indian context this principle
is expressed in the term s Tittle kingdom s’ and ‘little kings’ to
describe a local ruler whose ‘kingly’ authority is that of any great
king, but more limited in scope.
The principle o f pyramidal segmentation — o f complementary
opposition am ong segments within an expanding framework
o f vertical integration — is as salient in the concept of segmentary
states as the conceptual bifurcation o f the ritual and political
aspects o f rule. A unitary political order in the process of form ation
or decay m ay exhibit these characteristics. However, when the
fundam ental structure o f political relationships is predicated upon
the legitimacy o f pyramidal segmentation and o f the kind o f separa­
tion o f authority and power delineated here, the system is extremely
34 W hile th ere have been discussions betw een the a u th o r a n d S outhall on m ost
o f th e issues raised here, it is n o t suggested th a t the concept o f the segm entary
state as used h ere is 'a u th o riz e d ' o r ‘e n d o rse d ’ by S outhall.
difficult to bring under unitary rule from above or to alter from
below because political authority is inextricably tied to opposed,
localized segments. Given the pyramidal segmentary form of
organization, the only possible supra-local, extra-segmentary
integration which could occur would be of a ritual sort. Seg­
m entary social organization may, and often does, occur in stateless
societies or in tribal societies; segmentary social organization
among peasant peoples, whose dom inant form o f political organiza­
tion is the chiefdom, would require some kind of segmentary state
with characteristics conforming to those outlined by Southall.
The conception outlined above pertains to the ways in which
relatively self-sufficient, enduring, and often quite ancient localized
societies can be linked together to form a state. Such a state is not
an am algam ation or absorption of localized units into an organic
greater unit such as implied in the unitary state, but is an arrange­
ment in which the local units — segments — retain their essential
being as segmental parts of a whole. One reason why each of
the segmental units remains autonom ous is that each is pyramidal,
that is, each consists o f balanced and opposed internal groupings
which zealously cling to their independent identities, privileges and
internal governance, and demand that these units be protected by
their local rulers.
A nother quite different reason for the durability o f these arrange­
ments is th at rulership — whether that of chieftainship or kingship
— is sacral: it is dependent upon ritual, not administrative, in­
corporation. The very term raja alerts us to the ritually incor-
porative character o fln d ian kingship since one meaning of the root,
raj, is ‘stretching o u t’, ‘the king being one who “ stretched himself
out and protected (other men) under his powerful arm s” ’ as do
divine powers (the Vedic Savitar and Agni).55 Segmentary forms
of organization cannot exist except under such ritually incorporative
leadership.
Rulership in medieval South India was based upon ancient
canons of A ryan kingship. This was true from the Pallava period,
when brahm anicai kingship was established, modifying the earlier
form of kingship described in the puram poems of the Classical,
or Sangam, tradition. While both the earlier and later forms of
55 J. G o n d a , A n cien t Indian K in gsh ip fro m the R elig io u s P o in t o f View, E .J. Brill,
L eiden, 1966, p. 122; he draw s here u p o n the w ork o f T. B urrow .
rulership — of chief or king — were sacral in character, the form
introduced in the Pallava age provided the means for attaining
encompassing, incorporative universal kingship. The pre-Palla-
van, or Classical, period was one in which three kingships and a
great num ber of chieftainships existed among Tam ils; from the
Pallava period, the Tamils could have but one great king, one who,
by means of ritual, incorporated all lesser rulers. Ritual incorpora­
tion was the essence of medieval South Indian kingship, and the
introduction of this conception into South India was as politically
significant as was the introduction, later, of bureaucracy, by the
British, for modern South India.
Ritually incorporative kingship has been neglected by all but
a few writers on Indian kingship and the Hindu state. Among the
scholars of the Chola period, sacral kingship is entirely ignored.
Nilakanta Sastri’s discussion of Chola kingship, in The Colas
and other writings, is concerned with the king as warrior, as the
putative director of an elaborate administrative structure, as the
benefactor of Brahmans and temples, and as the follower of con­
ventional rajadharm a in other ways.56 T.V. M ahalingam ’s treat­
ment of kingship is more detailed, but not essentially different
from that of his m entor, N ilakanta Sastri.57 Though Mahalingam
gives more serious attention to literary and poetic sources than
N ilakanta Sastri, he shares the latter’s distrust of them. Hence,
he considers ‘overdrawn’ poetic references to the king as m ediator
between the cosmos and earthly society, such as the following lines
from the M an im ekalai: ‘If the king swerved from his righteous
path, ‘The planets would all change their course. . . . " 58
Among a few scholars, however, the sacral character of Indian
kingship is absolutely central. Two elaborate treatm ents of the
subject are found in J. G onda’s A ncient Indian Kingship from the
Religious P oint o f View and A.M. H ocart’s K ings and Councillors ,59
G onda’s title is somewhat misleading since it is clear that he regards
Indian kingship as intelligible only from the religious point of view.
Thus, while allowing that the view of Kautilya of the ‘practical
56 O p. cit., pp. 460 ff.
57 S ou th Indian P o lity , pp. 11 if.
58 Ibid., p. 15.
59 G o n d a , op. cit., an d A .M . H o c a rt, K ings an d C o u n c illo rs; A n E ssa y in the
C o m p a ra tive A n a to m y o f H um an S o c ie ty , U niversity o f C hicago Press, 1970, p. 197;
originally p u b lish ed in C airo , 1936.
side of Indian public life as opposed to the religious’ was im portant
in recognizing the king to be the ‘central figure of the state’, he con­
cludes his disquisition with the observation that the social order
which the king was bound to nourish and sustain was prior to the
state.60 As protector of this social order and its people, the king
possessed ‘a morality of his ow n’.61 The m ajor thrust of G onda’s
essay is the sacrificially attained divinity of the Indian king. Created
from ‘eternal and essential particles of Indra and the seven other
great devas the king is one of the lokapalas (guardians of the world)
according to M anu and later writers.62 The Vedic ra tn im , seen
by many historians as an ‘administrative council’ of the ancient
Indian king, are better seen, according to G onda and his student
Heesterman, as a group o f persons endowed with ‘sacral qualities’
which are incorporated by the king in a marriage-like ritual.63
The king as m arried to his realm or to the goddesses of fortune,
victory, and the earth are figures of speech frequently found in
royal inscriptions o f medieval India.64 As gods are honoured by
man, so are kings because of the latters’ anointm ent in the maha-
bhisekha which consecrates (i.e., ritually transform s) each hum an
king just as it did the prim ordial divine king, Indra.65 Sacrificially
created power is thereafter passed through the king as in the asva-
m edha sacrifices. G onda points out that the royal horse sacrifice
does not merely assert ownership over the territory circum am bu­
lated by the king’s sacrificial horse; the horse transm its to that
territory its divine power acquired as a result of ritual.66 The
horse sacrifice is repeated each year by the royal sacrificer and
hence becomes a cycle of sacrificial regeneration of the land and
its people.
H ocart provides the most general conception of incorporative
60 G o n d a , op. cit., p p . 137-8. E lsew here (p. 103): ‘It is th erefo re no hap p y idea
sh arp ly to d istin g u ish betw een the religious and secular aspects o f kingship, the
fo rm er req u irin g fro m th e m o n arch certain acts o f p ro p itia tin g gods and unseen
p ow ers . . . w ith the help o f th e p u ro h ita and sacrificial priests, the la tte r including
acts th a t lead to p ro sp erity o f realm an d subjects.’
«> Ib id ., p. 138.
« Ib id ., pp. 24-30.
63 G o n d a , op. cit., p. 44; J.C . H eesterm an , The A n cien t Indian R o y a l C o n secra ­
tio n ’s, M o u to n an d C o ., s-G ravenhage, 1957, p. 52.
64 L in g at, op. cit., p. 212 a n d Sircar, Indian E p ig ra p h y, p. 35 In.
65 G o n d a , op. cit., pp. 56, 82-3.
66 Ib id ., pp. 110-14.
kingship in which the ancient Indian monarchical tradition is but
one example. Sacrifice is the central means by which the king
is made sacred and is the means by which all other elements of
the realm are incorporated and controlled: lesser kings and chiefs,
and their territories; other gods and their priests. E.W. H opkins’
view of Indian sacral kingship deserves special mention, for he
introduces a dynamic element missing from the conceptions outlined
above. H opkins notes that the concept of divine kingship is at the
foundation o f all ancient Indian texts dealing with governance.67
His contrast of the Brahm an and the king in later Vedic thought
is instructive: the Brahm an is ‘born divine. . . a king becomes divine
only by virtue of a religious ritual’ regardless of his birth as a
K shatriya.68 This conception, H opkins states, is found in many
later law texts (dharm asastras). Thus, Kautilya speaks o f the
anointed king as one to be honoured and obeyed as a god, and
M anu goes even further to equate the anointed king not only with
Indra — king o f the Vedic gods, but with seven other great gods
who protect the world (gods o f the sun, m oon, fire, wind, wealth,
judgem ent or Varuna, and death).69 Later, epic poets speak of
five rather than eight divine protectors of the world, including the
king, and epic poets ‘never say that the king is like a great divinity. . .
the king is a great divinity; he is like this or that god’.70
Epic poets also introduced a new element into the notion of
sacred kingship, that of ‘divine incarnation’. Thus, of the first
king, Prthu, Hopkins states: ‘Visnu entered his body, and so the
world bows to this king as to a god among hum an gods.’71 The
avatar form of incarnation provided the means of assimilating
Ram a and Krishna to Vishnu as hum an forms of the god, and
provides a conceptual model for the sacral transform ation of hum an
kings to divinity in medieval South Indian kingship.72
D harm asastra texts extend the temporal range of sources on
sacral kingship to the period of the Cholas.73 Lingat has noted that
67 ‘T he D iv in ity o f K in g s’, J A O S , v. 50 (1931), p. 309.
68 Loc. cit.
69 Ib id ., p. 311.
™ Ib id ., p. 313.
« Ib id ., p. 313.
72 Ib id ., p. 314; a n d S ircar, Indian E p ig ra p h y, p. 337.
73 J. D u n c a n M . D e rre tt h a s d raw n atte n tio n to this fact in h is ‘T w o In scrip tio n s
C o n cern in g th e S tatu s o f K am m alas a n d the A p p licatio n o f th e D h a rm a sa stra ’,
while many of the sastric texts on dharma speak of the divinity of
kings, the idea is rarely encountered in the earlier d h arm asu tras ;
he also notes a difference in the way this divine aspect of kingship
is perceived in early, Vedic, texts and later, dharm asastra , texts.
Vedic texts attribute divinity to kings because of their participation
in ritual which identified the king with a god; this is well illustrated
in G o n d a’s discussion, of course. The same sense o f sacred kingship
is conveyed in the meaning attached to the terms raja and rajyam
in Vedic usage, that is the idea o f the ‘splendor, magnificence and
prestige’ of the king ‘rather than his power to com m and and
the idea of power and force’.74 But, Lingat argues, by the time
o f the dh arm asastra s of M anu and N arada, the divinity of kings
is posited on other grounds, namely as a reason for obedience
to royal orders. It is, therefore, the institution, not the royal
person, which is deified.
The dom inant idea o f the dharmasastra w riters seems to have been th at
it was not the king who had a divine nature, but the royal function it­
self. . . ,15 The exercise o f the royal function is equivalent to the celebra­
tion of a sacrifice o f long du ratio n (sattra), and th at is why the king
rem ains pure, w hatever acts he is led to com m it. . . ,76

An interesting and corrolary conception which Lingat finds

in his study of dharm asastra is that the social origin o f kings is
not considered im portant; it m atters only that the king has been
anointed (abh isikta ) and that he is an able warrior. Kshatriya
birth is insisted upon by M anu,77 but m ost other dharam asastras
do not. Kulluka, a Bengal sa strik a of the late twelfth century,
attributes the title o f raja to any m an duly anointed and affording
royal protection, regardless of his varna ,78
Notwithstanding the various aspects of sacral kingship which
Lingat has evinced, his summary judgement of the conception of
kingship in the medieval dharm asastra tradition is puzzling in its
emphasis upon the secular:
in P rof. K .A . N ila k a n ta S a s tr i 80 th B irth d a y F elicita tio n V olum e, M a d ras, 1971:
a b rie f o n this ap p ears in L ingat, op. cit.. p p . 273-4.
74 L in g at, op. cit., p. 211 n.
75 Ibid., p. 208. H e n o tes th at this co n cep tio n h as been the subject o f heated
co n tro v ersy a m o n g scholars.
7« Ib id ., p. 215.
77 Ib id ., p. 209.
78 ib id ., p. 210; also see K an e, op. cit., v. 1, pt 2, pp. 756-9.
 . . . th e k in g a p p e a r s to o w e h is a u t h o r i ty n e ith e r to d iv in e w ill, n o r to
h is b ir th , n o r to a n y s o c ia l c o m p a c t, b u t so le ly to th e fo rc e a t h is d is p o s a l.
. . . . H is a u t h o r i ty is e n tir e ly te m p o r a l a n d s e c u la r. P u n is h m e n t is th e
so le in s tr u m e n t o f h is p o lic ie s. B u t lik e a n y m o r ta l, h e h a s h is o w n
d h a r m a . . . [an d ] t h o u g h h e is a n a b s o lu te so v e re ig n , h e is s u b je c t to th e
law o f k a r m a . . . .7‘'

W hat Lingat appears to mean here is that for the dharm asastra
writers, the possession of force, which alone provides the eligibility
for kingship, is made perfect as danda: pure force which becomes
a divine institution of punishment for the welfare of all.80 Perfec­
tion of that force can only occur with proper anointm ent, and this
creates the royal function; secular force is danda only as it is em­
bedded in the extended sacrifice (sa ttra ) which kingship is construed
to be. During this time, the king is freed from the consequences
of sin or error, though in his after-life he is as fully accountable for
both as any person. This dh arm asastra conception of kingship is
certainly different from th e conception conveyed in Vedic texts
upon which Gonda, Hocart, and Hopkins primarily depended, but
the effect of this dh arm asastra view is not to diminish, but rather
to amplify and elaborate th e notion of sacral kingship. It is to b e
excepted that the concept of royal authority in these sm riti texts,
based as they are upon Vedic sru ti sources as revealed in the com­
mentaries of learned Brahmans, would tend to yield little more to
kings than the responsibility and the attendant force {danda) to
compel obedience by the king to the instructions of the Brahman
lawgivers. One consequence o f this was the depersonalization
of kingship by emphasizing the royal function and making o f it
an extended sacrifice81, thus conferring contingent divinity upon
the king as agent. But this can be hardly said to diminish the
sacral attributes o f medieval kings; it seems rather to augment
them.
Two quite different kinds of formulations — the pyramidally
segmentary state and sacral kingship — have been presented here
to provide the basis for a better understanding of the medieval
political system of South India and for a better explanation of
79 Loc. c it..
8» Ibid., p. 214.
81 G o n d a to o (op. cit., p. 15) rem arks on Manu’s conception o f the extended
sacrifice o f w hich kingship is sup p o sed to consist as a necessary co n d itio n freeing
the king from asctnca, o r im p u rity an d the consequent restriction from the p e rfo rm ­
ance o f religious acts.
political evidence from that period. These apparently disparate
formulations are held to comprise a complex, unified conception of
the South Indian states o f the time. Neither form ulation alone can
be considered adequate even though many scholars might agree that
all medieval Indian states were segmentary and many might also
recognize that the Hindu conception of m onarchy was essentially
sacred in the sense that kings were created by ritual and m aintained
through the moral authority engendered by ritual. The incom­
pleteness of the former conception stems from obvious differences
between African societies, analyses of which produced the concept
of pyramidal segmentation, and those of India. Crucial here is
the differential im portance o f clan/lineage organization in African
and Indian varieties as well as differences in the degree o f elabora­
tion, complexity, and pervasiveness o f ideas pertaining to kingship
and society in India and Africa. The sacral kingly conception
is not less incomplete, for the means by which India’s sacred kings
actually ruled territories as vast as that o f the Chola state can
hardly be fitted to the modest scale implied in the discussions of
those who have, quite rightly, spoken of sacral kingship in India
(i.e., G onda, Hocart, Heesterman, and Hopkins). It is only as
these formulations of pyramidal segmentation and sacral kingship
are brought together that a single cogent model of state formation
and statecraft can be realized.
The concepts o f pyramidal segmentation and sacra! kingship both
imply a political system of fluidity and indeterminacy which may
offend our m odern notions of states as fixed and certain entities.
Yet, it is precisely the fluid and indeterminate — or what Southall
called the ‘fragile’ — character of medieval Indian states which
most clearly strikes all who study them. Boundaries are often
vague and shifting as are capitals. Gaps and bends in the royal
line frequently occur. Great states are reduced to m inor ones
kaleidoscopically. The reach of royal intervention is at times
astonishing as when a great king personally sets aright the ritual
arrangements at some small temple very distant from his capital.
For some scholars, this is evidence not only of a centralized state,
but of one in which the king is its m ost active agent! M ost of the
time, however, the king is not only a distant figure, but one of
such god-like majesty or fierce war-like mien as to make him an
utterly improbable agent for the resolution of minor, local prob­
lems. A sacred king does possess all of these attributes: he is a
god; he is a great w arrior; he is responsible for order and prosperity
throughout his realm o f however great an extent that may be;
he is the personification of divine energy (s a k ti) which is without
limit. But it is in being all of this that makes for difficulty in grasping
the political system as something other than a stage for the enact­
ment of a sacred drama.
There are elements of a more fixed and certain nature in the
medieval South Indian state, and they are combined as a state
only in that they are incorporated under a king’s energetic protection
of the world, his realm. In Chola times these elements existed
concretely at the level of the nadu. Here there were villages under
the control of assemblies (urs) of the dom inant cultivating groups
o f the locality corporately known as nattars; here there were trade
settlements, nagaram s under the control o f local m erchant groups;
here, finally, were chiefs who, from the time of the Pallava king
N andivarm an II, Pallavamalla ( a . d . 731-96), were honoured in
royal inscriptions as the protectors o f a locality, ‘little kings’ who
were ‘m arried’ to the nadu realm just as kings were ‘m arried’ to
their royal realm.
The nadu was not a sealed world, nor was it a seamless one. M ost
of the constituent elements within a nadu had connections beyond
its borders which were activated on occasion: the n a tta r to other
n attar groups with whom, by the twelfth century, they cooperated
on an almost continuous basis as the periya n a ttd r; merchants of
the nagaram with other locality merchant groups and with the
itinerant m erchant groups of the southern peninsula; Brahmans
with other brahm adeyas and with sacred centres; and chiefs with
others like themselves serving in the armies of the kings to defend
the kingdom or to seize booty or to add to the lustre of their king
by wars with other kingdoms. These relationships between nadu
groups and other like groups in other nadus are one significant
aspect of the pyramidal articulation o f segments in Chola society,
a massing of an ascending order that reached the royal centre of the
state.
But, equally significant and characteristic of pyramidal segmen­
tation were the segmental divisions within each nadu. All of the
elements mentioned above plus the dual division o f right and left
lower castes constituted integral units in their own right, and ulti­
mately, maintaining the integrity of each and all, was the responsi­
bility o f local chiefs and, finally, the king.
 Above all else, the two formulations joined together in this
discussion serve to focus attention on the m atter of chieftainship.
In Chola historiography, the institution of chieftainship is virtually
ignored, a victim o f the centralist-bureaucratic bias of N ilakanta
Sastri and others. Thus, the institution does not appear in the
sections of The Colas devoted to the Chola system of government.
However, in every one of the chapters of this im portant work,
there are num erous references to houses o f chiefs with whom the
Chola kings interacted, at times as enemies, but often as supporters,
in which case chiefs are called ‘officials’ or ‘feudatories’. The same
lapse appears in M ahalingam ’s South Indian P olity.
Chieftainship of various kinds existed during the Chola period
and later. Some o f these were great houses o f chiefs which on
occasion produced a warrior leader of such ability as to break
the surface of obscurity to which m ost such personages are con­
signed in the inscriptions of the time. In the time of Parantaka I
( a . d . 907-55) the Paluvettaraiyan chiefs" o f m odern Tiruchirapalli
district attained sufficient prominence to give a bride to the royal
family,82 and during the time of Rajaraja I, descendants of this
line of chiefs built temples in im itation of the Chola king.83 When
Rajendra I faced a serious threat in the conquered regions of
Pandya and Chera (Kerala) countries in the late years of his reign,
one of his opponents was a chief of Iram kudam .84 Among Rajen-
dra's most dependable supporters was the Vallavaraiyar chief of
modern Salem district; a nadu there was nam ed after this family.85
And, in the northern portions of the Chola overlordship, in m odem
Coorg and K arnataka, two distinguished families of chiefs held sway
as Chola supporters: the Cangavalvas and the Kongavalvas.86
Again, in the time of Kulottunga I, m ajor support to the royal
fortunes came from the Kadava chieftains who claimed connection
with the ancient Pallavas and were centred in m odem South
A rcot.87 Finally, in the troubled reigns of R ajadhiraja II and
Kulottunga III, of the late twelfth and early thirteenth centuries,
numerous lines of chiefs are m entioned; indeed, it is the prominence
s2 N ila k a n ta S astri, The C olas, p. 187.
w Ib id ., p. 187.
*4 Ib id .. p. 221.
S5 Ibid.. p. 226
»6 Ib id .. p. 227.
87 Ib id ., pp. 349-50.
o f these cheiftains at the time, which is seen as having contributed
decisively to the decline o f the dynasty.88
Such great chieftains are most often assimilated to Chola polity
as ‘officials’, a term suggesting that these powerful persons owed
their political existences to positions in the Chola government.
But, the evidence presented by N ilakanta Sastri and others
quite belies this. Great chiefs and lesser chiefs, the latter often
designated by the title adigari (and thus identified by Nilakanta
Sastri as ‘officials of the government’)89 or the title m uvendavelar
(a chief whose family served the three kings: Chola, Chera, and
Pandya) were fully independent local rulers. N ilakanta Sastri
cites a commentary on the twelfth century poem of O ttakuttan,
the T akkayagapparan i , regarding the pretentions of chiefs to
entitlement to all of the honours of kingship except the crown
itself.90 He calls this a ‘quaint account’, but it is quite m istaken
to trivialize claims of chiefs of this sort. For such claims to authority
have the same nature as fully royal ones; they derive from the same
ideology of rule.
Chieftains, just as kings during the medieval period of South
Indian history, possessed the attributes of k sa tra and dharma.
K satra , the power to com mand and possess resources, inhered in
the chiefs by virtue of their connection with dom inant agricultural
groups whose leaders they were. Along with this went political
control within the territory of chiefs and the ability to amass
resources from all within the locality, whether these were in money,
in kind, or in service. K satra for chiefs as well as kings was a claim
to resources based upon force; in this case, as with kings, k sa tra
implied a set of reciprocal relationships— one might say contractual
relationships — between the chief and the various groups within
his realm: in return for protection from the chief, resources were
given. As with kings, so with chiefs, this force became an entitle­
ment, or was purified, by dharm a which meant the pooling of
resources and their redistribution. D harm a was appropriately
claimed and realized through the redistributional nexus within the
territory of the chief; at the centre of this stood the chief. The
clearest and most persistent forms of dharmic activities of great
and m inor chiefs were religious. The alienation of income from
Ibid., pp. 388, 400-7.
sg Ib id .. p. 513.
9(1 ibid.. p. 463.
 established agrarian settlements to Brahm ans and to temples was
an expression of the dharmic, yajam ana, function, and the cere­
monies which accompanied such grants symbolically enacted
the moral and dharm ic unity of the chiefs, territory just as similar
grants and ceremonies involving kings enacted the unity of the
kingdom as a whole.

HI
Three levels of structure must be identified in the South Indian
segmentary state system. Taking the macro region as a whole during
and shortly after the period of Chola ascendency, the three levels
k of structure, or ‘zones’, to use Southall’s term, are: the central
zone of the Chola segmentary state, C holam andalam ; three inter­
mediate zones of which Tondaim andalam and Pandim andalam
were most im portant, and Naduvil-nadu, less im portant;91 and
the peripheral zones of Kongu and Gangavadi. Each of these
zones, in turn, can be further differentiated into central, inter­
mediate, and peripheral areas according to the types of nadus which
are found, as already suggested in C hapter III. (See M ap VII-1.)
Chola country in the Kaveri basin, the central zone of the Chola
segmentary state, was called, C hoianadu, until the early eleventh
century when Rajaraja I adopted the designation m andalam for
this territory as well as for other m ajor zones of the Chola state.92
At about the same time, Tondainadu became Jayankondachola-
mandalam and Pandinadu became R ajarajam andalam . Bounded
by the northern and southern Vellaru rivers according to references
in Classical works as well as those of the Chola period, the core of
Chola country was the Kaveri river along its east-west traverse
to the sea. Where the Kaveri changes from its north-south course
to its east-west graded movement to the sea, to the west of what
was called mala-nddu , Kongu country began.93
Prior to about a.d. 1000, inscriptions refer to ‘eighteen’ establish­
ed localities (nadus) comprising Choianadu. These were concentra­
ted in the riverine core area of the territory, the northern, central

1.1 O th er nam es for this territo ry are nadit-nadit and nailuvu-iuiclu.

1.2 T he earliest reference to vnlanmmlalum cam e in a . d . 1009; A .R .E ., 1922, n o .22.
1.3 K. V aid y an ath an , ‘A ncient G eo g rap h ical D ivisions o f T am il N adu", op. cit.,
p. 234; aiso his, ‘M a ia -N a d u ’, Q J M S {C ulture and H eritage N um ber), pp. 225-60.
and western portions o f m odem Tanjavur district, and are referred
to in inscriptions o f the ninth and tenth centuries with the indication
that they were either on the north bank of the Kaveri or the south
bank ( vadakarai or tenkarai). To this perhaps conventionally
designated ancient core were added about 120 more nadus be­
longing to Chola country by a . d . 1300. As argued above, these
additional localities represent not newly established areas of
settlement in m ost cases, but more probably the inclusion of
previously settled places under the expanding overlordship of the
Cholas.94 By the fourteenth century, there were well over a thou­
sand villages in this overall territory of some 8500 square miles
whose deltaic core comprised about 3000 square miles.95
The deltaic portion o f Cholam andalam can be designated the
central area of the central dom ain o f the Chola segmentary state.
It was here that the power of the kings from Rajaraja I to Kulot-
tunga I was concentrated; this was their n adu ; they were the leaders
o f its n attar. Curiously, however, it is difficult to locate the seat
o f Chola power within this deltaic core. Vijayalaya, the founder
o f the great Chola line, established Tanjavur (then, Tanjapuri)
as his capital in the middle of the ninth century, and this continued
to be the fortified centre of the Chola overlords through the time
of Rajaraja I. Rajendra I shifted the capital to G angaikondachola-
puram on the northern bank of the Kollidam, about thirty-five
miles north-west of R ajaraja’s capital. In both cases, the selection
of the m ajor centre o f government may have been in part dictated
by strategic considerations related to the protection o f the riverine
source of wealth to all of the central zone. Two other places also
appear to have been especially im portant for the Cholas. One was
near the western fringe of the old delta, U raiyur (near m odem

94 S u b b aray alu lists the earliest nadus as th e follow ing: A rv a la k u rra m (M an -

n arg u d i ta lu k , T an jav u r), A vur-k. (K u m b a k o n a m /P a p a n a sa m t.), In g an -n a d u
(N a n n ila m /T a n ja v u r t.), In n a m b a r-n . (K u m b a k o n a m /P a p a n a sa m t.), K ilar-k.
(P a p a n a sa m /T a n ja v u r t.), K u la m a n g a la m -n . (A la n g u d i/K a la ttu r t.), M a n n i-n .
(K u m b a k o n a m t.), P am b u n i-k . (M a n n a rg u d i t.), P a n a iy u r-n . (N an n ilam t.),
Poygai-n. (T an jav u r a n d L alg u d i/U d a iy a rp alam t., T iru ch irap alli), Poyyil-k.
(T an jav u r t.), P u ra n g a re m b a i-n . (M a n n a rg u d i t.). T iru n a ra iy u r-n . (K u m b a k o n a m
t.), T iru v ali-n . (Shiyali t.), T iru v a la n d u r-n . (M a y u ra m .t.), T iru a ru r-k . (N ag ap at-
ta n a m t.), T iru v in d alu r-n . (M ay u ram t.), U raiy u r-k (T iru ch irap a lli/K u littalai
t.): p. 18, (A p p en dix 1).
m Ib id ., p. 31.
VII-1 The Chola Macro Region c. a .d . 1300
Tiruchirapalli), from whence the Cholas apparently originated.96
The other was the trade centre of N agapatanam on the sea.
These capitals and major centres at the fringes of the core of the
river basin suggest that the core region was well developed by the
tenth century and th at the perceived task of the Cholas was to
protect the basin from its m ost dangerous enemies, such as the
M uttaraiyar chiefs, who had previously held the central core,
and the allies o f the latter, the Pandyas, to the south. From the
locations of the Chola capitals, protection was more im portant
than the fostering of settlement and development of the region.
From such fringe positions, it is difficult to see the Cholas as
managers of a ‘Byzantine m onarchy’! W hat is more in evidence
in this central area of the Chola state is a kingship established upon
firm bonds with existing locality leaders o f the rich lands of the
inner core — the deltaic fan — and, m aintaining a network of
ritual allegiances to fortify their military power. It is im portant
to note that no apparent ancient sacredness attached to the Tanja-
puri site o f R ajaraja’s great shrine, Rajarajesvaram , nor that of
R ajendra’s royal shrine at Gangaikondacholapuram . The shrines
were not for the shelter and honour of ancient gods of these places,
but were symbols of royal power. The key to Chola authority
within their central domain being military power, another factor
in the location of the Chola capitals was to facilitate better control
over the interm ediate area adjoining the central area of the central
domain.
The largest part of the Kaveri basin, about 5000 square miles,
comprised the intermediate and peripheral areas of the central
dom ain of the Cholas. The intermediate area included the large,
relatively sparsely populated, nadus of the western part o f the
Kaveri where the river bends tow ard the sea and close to where
Kongu country begins. This interm ediate area extended eastward,
toward the sea along the north bank of the Kaveri to where rough
terrain begins (the Kollimalai and Panchaim alai hills) and along
the southern bank to a depth of about fifteen miles where rugged
country is again encountered. The southern peripheral area of the
central zone was known as K onadu, Jayansingakula-valanadu,
and R ajaraja-valanadu which formed part of the m odem Puduk-
kottai state whose core was the southern Vellaru river, the tradi­
tional southern boundary of Chola country. To the north of the
96 N ila k a n ta S astri, The C o la s, p. 110.
Kaveri, the area between the Kollidam and the northern Vellaru
rivers constituted another peripheral area of Chola country.
Naduvil-nadu, beyond the northern Vellaru and west of the narrow
coastal plain at this point, may thus be considered an independent
zone between Chola and Tondai countries.
Chola control over the Kaveri core began with the founder’s
— Vijayalaya — displacement o f the M uttaraiyar chiefs from
Tanjapuri and his successful defence against the Pandyan leader
Varagunavarm an. Aditya Chola, who succeeded Vijayalaya and
ruled from a .d . 871-907, was similarly successful against the Pallava
overlord A parajita who sought to re-establish an earlier Pallava
overlordship in the Kaveri. It is to be recognized that the Pallavas
were the first South Indian dynasty to establish a kingship incor­
porating the m any locality chieftainships of the m acro region and
to use ritual authority as the basis of their authority. But, outside
the Pallava heartland of Tondaim andalam , this accomplishment
was short-lived, soon to be displaced by the Kaveri-based Cholas.
The Pallavas having been repulsed from the Kaveri, Aditya
even succeeded in securing a Chola foothold in Tondaim andalam
by a . d . 890 as well as in Kongu country by the end of his reign.97
P arantaka I, in the first half of the tenth century, pressed the Chola
hegemony in several directions. In a series of raids into Pandya
country, the Cholas established what was to be a long and difficult
presence in the southern peninsula and Sri Lanka. Chola warriors
also raided and subordinated some of the great and ancient chiefs
on the northern fringe o f Tondaim andalam . In these military
activities, P arantaka had the collaboration of other chiefs formerly
under Pandyan overlordship, especially the Irukkuvels of Kodum-
balur in Konadu, whose territory on the southern Vellaru sat
uneasily between Pandya and Chola countries. Such chieftains
allied to the Cholas gained in their association with Chola expansion.
The K onadu chiefs, for example, established themselves as almost
independent kings in Kongu for several centuries.98 Other ancient
families of chiefs — Banas and Vaidambas in northern T ondai­
m andalam — who had recognized the overlordship of the Rashtra-
kutas of the Bom bay-K arnatak, were sharply reduced in influence.
R ashtrakuta power, under K rishna III, brought an end to this
9 i Ibid.. pp. 1(2-13.
98 E .I., v. 30, n o .19, ‘Seven v atte lu ttu Inscriptions from th e K o n g u C o u n try ’, ed.
K .V . S u b ra h m a n y a A iyer, pp. 95-112.
phase of Chola expansion. K rishna’s inscriptions in Tondai-
m andaiam for almost thirty-five years — a . d . 945-70 — are elo­
quent testimony of the resistence to the Chola bid to replace Pallava
influence in the central Tamil p la in ." U nder R ajaraja I, the issue
of whether the central and northern Tamil plain was to be a peri­
pheral zone of R ashtrakuta (or Deccan) influence, through the
intermediate zone of Gangavadi under the agency of Ganga chiefs
like Butaga, or an intermediate zone of Kaveri influence was
finally settled in favour of the C holas.100 N otw ithstanding the
setback of Chola power of Tondaim andalam by Krishna ill, the
decisive sign of Chola influence in the central and northern Tamil
plain — the first of such influence by them — came in the time
of Parantaka I. His inscriptions relating to the autonom y of
brahm adeyas in Tondaim andalam both date and help to define
this im portant part of the Chola segmentary state.
Tondaim andalam and Pandim andaiam were the two inter­
mediate zones o f the Chola state. Each possessed its own central
area: the Palar basin with Kanchi, the Vaigai basin with M adurai;
in each there was an established tradition of kingship, and, in the
case of the Pandyas, one as ancient as that of the Chola. Each of
these intermediate zones of the Chola state also had its own inter­
mediate and peripheral areas with distinctive qualities of their own.
Except for Sathianathaier,101 none who have written on the
Chola period have adequately recognized the extent of the dif­
ferences between Cholam andalam and Tondaim andalam in the
time of the great Cholas. It is not necessary to accept his argument
that the distinctiveness of Tondaim andalam derived from its having
been included in the empire of Ashoka, for he does not. and cannot,
suggest why that, if true, might have been im portant. However,
the significance that he attributes to the fact that Buddhism had a
longer life in Tondaim andalam than elsewhere in the Tamil plain
appears more cogent.102 These pre-Pallavan developments are
difficult to assess. But, with the establishment o f Pallava power
at Kanchi, the distinctiveness of Tondaim andalam emerges
clearly. According to Sathianathaier, the Pallavas of Kanchi were
99 K .V . S u b ra h m an y a A iyer, S k e tc h e s o f the A n cien t D ekh an , I, pp. 55, 226-31.
>«“ Ib id ., p. 54.
101 R . S ath ian ath aier, S tu d ie s in the A n cien t H is to r y o f T on dam andalam .
102 Ib id ., pp. 1-26.
responsible for the establishment of brahm adeyas, those large,
populous settlements under the regulation o f an assembly of
Brahmans (m ahasabha ). It is his contention that this institution
— to which he attributes great im portance in the determ ination
of the quality of Pallavan culture — was first developed in Tondai-
m andalam and that during Pallava times, the eighth to the tenth
centuries, m ost o f the records o f mahasabhas come from villages
in Tondaim andalam .103 The late eighth century records of Danti-
varm an Pallava, both Sathianathaier and N ilakanta Sastri agree,
show the first development of the committee ( variyam ) system
which was characteristic of the governance system of these villages
in the Chola period.
W hether, as Sathianathaier asserts, the classical brahm adeya
was first developed in Tondaim andalam , or whether the develop­
ment o f brahm adeyas occurred at the same time in both the central
and southern parts of the Tamil plain, perhaps even in Venadu
(modern Kerala), it is clear that Tondaim andalam was a m ajor
region for the concentration of these institutions. In contrast to
both C holam andalam and Tondaim andalam in which there are a
large num ber of brahm adeya inscriptions dating from the ninth
to the fourteenth centuries, the other parts o f the Tamil region
provided many fewer.104 This analysis of Sathianathaier is based
upon fifty brahm adeyas which yield four or more mahasabha
inscriptions for the five centuries as o f about 1940. M ap IV-1 locates
brahm adeyas m entioned in one or m ore inscriptions and covers
the period of epigraphic activity to the middle 1960s. This map
shows a markedly different pattern from that discovered by Sathia­
nathaier; more than three times as many brahm adeyas are located
in Chola country than in Tondaim andalam .105 Since the problems
of locating with certainty those m odern villages which were pre­
viously Brahm an centres are no greater in one o f the plains regions
103 Ibid., p. 27. tie notes seventeen places where mahasabha inscriptions of the
Pallavas have been discovered. O f these seven were in Chingleput, six in N orth
Arcot and C hittoor, one from T anjavur, one from South Arcot, and two from Tiru-
chirapalli.
104 Sathianathaier’s totals for the period are: T ondaim andalam , 307; C hola­
m andalam , 300; Pandyam andalam , 25; K ongum andalam , 14; total, 646.
105 C holam andalam (Tanjavur, Tiruchirapalli. and the C hidam baram taluk of
South A rcot): 223; Tondaim andalam (Chittoor, Chingleput, N orth Arcot, and
most of South A rcot): 83; Pandyam andalam (M adurai and Tirunelvelli): 42;
Kongum andalam : 9.
than the other, it is necessary to conclude that Sathianathaier’s
notion about the primacy of Tondaim andalam in relation to these
settlements is incorrect, or, at least, questionable.
Other issues raised by him to account for the alleged primacy
of Tondaim andalam in the m atter of the distinctive Brahm an
village are less easily evaluated. The seventh century Chinese
Buddhist pilgrim Hsiian Tsang is read by Sathianathaier to indicate
that Buddhist institutions continued to survive, if not to thrive,
in the central Tamil plain after they had declined in the Kistna-
Godavari and Kaveri deltas and elsewhere.106 Also, Hsiian Tsang
is understood to suggest that there was at Kanchi as great a centre
of Buddhist learning as at Nalanda. Jaina learned men are also
reported to have been im portant here. Noting that the Pallava
records of the middle of the ninth century from Bahur (m odem
Pondicherry) refer to centres of higher Brahmanical learning
( vidyasthanas ), Sathianathaier is led to conclude that the factors
responsible for Buddhist survivals in Tondaim andalam also
account for later ‘local self-government and . . . higher educa­
tion’.107 Thus, he argues that the significance of Brahm an centres
in Tondaim andalam , committed to learning as well as to the
propagation o f resurgent Vedic religion, were a direct outgrowth
o f the previous im portance o f Buddhist and Jaina intellectual
and religious activities against which the subsequent enthusiastic
sponsorship o f Brahmanical learning and religion by the Pallavas
of the Simhavishnu line were directed.
Difficult though it may be to assess these factors presumed to be
im portant in explaining the alleged primacy of Tondaim andalam
in the provenance and the num ber o f brahm adeyas, one must agree
with Sathianathaier — and all historians do — that this institution
was of the greatest im portance in the preservation and dissemination
of Brahmanical knowledge and religion. One must, moreover,
agree that Tondaim andalam saw the early, if not the earliest,
development of this institution, though it is not the case that the
num ber of brahm adeyas in Tondaim andalam and Cholam andalam
was so nearly equal as Sathianathaier argues. The concentration
of Brahm ans and those settlements which they controlled appears
to have been much greater in the Kaveri core than in the central
plain. M oreover, if it is granted, as an explanation of the signifi-
106 Ibid., pp. 23-6.
107 Ibid., pp. 26, 50-1.
 cance of brahm adeyas there, that the persistence of Buddhist and
Jaina influence in Tondaim andalam was greater than elsewhere,
we have no way of ascertaining why this should have been so. Why
did Buddhism and Jainism flourish in Tondaim andalam for a
longer time?
The answers suggested are not final. One factor in the persistence
of ‘heterodox’ faiths was the city o f Kanchi. As a major trade
centre strategically sited on the principal route between the two
deltaic regions of the Kistna-G odavari and Kaveri population
centres, with good access to the interior upland behind the plain
as well as an opening to sea-born coastal and long-distance trade,
Kanchi is recorded in early Chinese sources. Trade centres near
Kanchi (e.g. M arkanam ) are also mentioned in the Periplus .108
This evidence of the ancient economic significance of Kanchi is
bolstered by what may have been an im portant association of
artisan-traders with the place. In later times, Kanchi was a prime
centre o f artisans o f the left-hand castes ( idan gai ).109 Oppert and
Mclean, writing in the late nineteenth century believed that the
left-hand division of castes at Kanchi harked back to an ancient
association o f these merchants and craftsmen with Jainism in
a m anner similar to the association o f merchants with Buddhism .110
Another factor which must be considered is that even beyond the
support of Buddhism and Jainism by m erchants and artisans —
a pan-Indian phenom enon of antiquity — there is reason to
suppose that these two faiths may have enjoyed wide, general accep­
tance among peasant folk of the Palar-Cheyaru basin core of the
Kanchi region. This is suggested by the facts that relatively few
Saivite bh a k ti hymnists (ndyanars ) are identified with Tondaim an­
dalam and that there are relatively few sites sacred to these hymnists
in this area as com pared to the Kaveri. O f the sixty-three Saivite
saints traditionally honoured (assimilating the Jaina conception
of sixty-three ‘great souls’), only six were of Tondaim andalam as
108 K..A. N ilakanta Sastri, F oreign N o tic e s . . . . pp. 4, 45 from Pan Kuo of the
first century a . d . ; ‘Periplus M aris E rythraei’ (in R.C. M ajum dar, The C la ssic a l
A cco u n ts o f India, C alcutta, 1960, p. 307) where three ports on this section of the
C orom andel coast are n am ed : Cam ara, Poduca and Sopatma.
1(19 Maclean, M a n u a l o f M a d ra s A d m in istra tio n , v. 3, p. 1037.
11(1 Maclean, ibid., and G. O ppert (O r ig in a l In h a bitan ts o f B h aratavarsa
p. 62), who notes that at the temple centre of Meikote, there was a special hall and
other facilities for castes of the left-hand.
against most of the rest who were of the Kaveri area.1*1 Moreover,
of the 279 sacred places referred to in the Saivite hymns, D evdram ,
260 can be located in South India and Sri Lanka, and of these the
largest number, over 200, were in Cholam andalam . The distribu­
tion of Vaishnava hymnists is less strikingly biased in favour of
the Kaveri. Five of the ten alvars appear to have come from T ondai­
mandalam, though only twenty-two of the 108 sacred places of the
Yaishnavas ( tirupatis ) are in Tondaim andalam .112
The disparity in the sacredness associated with Tondaim andalam ,
in contrast to Cholam andalam in the hagiography of South Indian
b h a k ti Hinduism, may perhaps be explained in many ways. Among
the explanations, however, is that Buddhism and Jainism had
established a position o f primacy among peasant folk as well as
townsmen in the central Tamil plain.
In the Kaveri, by contrast, the form of Aryan religion which
had come to be most valued by peasant folk was that devoted
to Vedic gods, and this provided a more receptive context for the
devotional hymns o f the devotees of the Puranic gods Siva and
Vishnu. This may account, too, for the choice o f the great teacher
Sankaracharya o f Kanchi as the site for one of his centres ( m a t has )
for the dissemination of Advaita Vedanta. If the tradition of the
origin of the Sankaracharya m atha is accepted, then one explanation
of this eighth century decision was to m ount an attack upon Bud­
dhism and Jainism in the very centre of their strength, Kanchi.
If that was its purpose, it was surely successful, for all overt vestiges
o f these religions have long been expunged from Kanchi. In
addition, according to some Saivite traditions, the powerful
female deity, Kamakshi, was transform ed by the piety of Sankara
111 K.V. S u b ra h m an y a Aiyer, H isto ric a l S k e tc h e s o f the A n cien t D ekhan, v. 1,
p. 9, note no.3.
112 George W. Spencer, 'The Sacred Geography of the Tamil Shaivite Hym ns’,
N U M E N ; In tern a tio n a l R eview o f the H is to r y o f R elig io n , v. 17 (December 1970),
pp. 236-7. Spencer lists thirteen sites in South Arcot district of which most, if not
all, can be included in Chola country. In addition Tanjavur had 160 and Tiruchira-
palli district had eighteen. Also see: K.V. Subrahm anya Ayyar, H isto rica l S k e tc h e s
o f the A n cien t D ekh an , v. 1, p. 9, note 4 and P.V. Jagdisa Ayyar, op. cit., pp. 221-30.
N ilakanta Sastri has argued that the first four Vaishnava saints, all of T ondai­
m andalam (viz., Poygai, Puclam, Pey and Tirum alisai), were 'waifs and strays',
that is, of low birth, thus strengthening the difference between the Kaveri. where
Saivite devotees came from all castes, and T ondaim andalam , where the first four
were of apparently very low status: ('Vaishnavism and the early Alvars', T .A .S .S .l
Silver Jubilee Volume, 1962, pp. 123-4).
to a benign, bh akti deity providing Kanchi, and Tondaim andalam
in general, with what has continued to be its most popular devotional
deity.113
Existing evidence scarcely permits a final choice o f explanations
for the distinctions between Tondaim andalam and Cholam andalam
in the m atter of religion, but the differences in the two regions with
respect to religion appear quite clear. The durability of Buddhism
and Jainism in the central plain may indeed have been an im portant
factor in the proliferation of prestigious Brahm an settlements as
suggested by Sathianathaier.
But the most significant reason for the emergence of T ondai­
m andalam as an intermediate zone of the Chola segmentary state
was the impressive kingship and state established by the Pallavas.
From Simhavishnu in the late sixth century to Nandivarm an II,
Pallavamalla, in about the middle o f the eighth century, a line of
powerful warriors secured an effective overlordship in T ondai­
m andalam and southward into the Kaveri and Vaigai basins.
Cholas and Pandyas were compelled to recognize the Pallava hege­
mony, as were those chiefs north and west tow ard the Deccan
strength of the Chalukyas and R ashtrakutas. After the middle of
the eighth century, Pallava control was weakened, and for the
next 150 years the Pallavas were linked, perhaps subordinately,
to warriors of southern K arnataka, the Gangas. This vestigal
Pallava overlordship was finally ended in the early years of the
tenth century when the Pallava A parajita was conquered by Aditya
Chola I.
The broad outlines of the Pallava state are too well-known to
require reiteration here. It is sufficient merely to speak of the
evidence of a powerful kingship and a network of political loyalities
in most of Tondaim andalam and, for a time, much of the C oro­
mandel plain. The Pallava state was a more impressive state than
any previous one in the m acro region. Beyond politics, one notes
its participation in the revival of H indu institutions to a place of
dom inance after the long ascendancy of Buddhism and Jainism;
its monum ental architectural accomplishments in Kanchi, Maha-
balipuram and elsewhere which permanently influenced the forms
of temple structures in South India and beyond; its participation in
the agricultural development of Tondaim andalam which made of
n i K.V. S u b ra h m an y a Aiyer. H istorical Sketch es o f A ncient D ekhan, v. 1, p. 14.
the central Tamil plain the significant secondary core of peasant
agrarian organization in the m acro region from that time. The
primary question which arises from these accomplishments between
the sixth and ninth centuries is why Tondaim andalam did not be­
come the central dom ain o f a state encompassing the entire plain
from the K istna-G odavari delta to the tip of the peninsula, K anya
Kum ari. The essential answer to this question would appear to
involve three factors: (1) the fact of an ancient, indigenous Chola
kingship in the Kaveri basin; (2) the superior population and
wealth o f the Kaveri and (3) the special im portance of that area
for the hymnists o f devotional Saivism. These were surely signifi­
cant in the long run, but still leave unclear more proximate reasons.
F or a brief time in the early seventh century, as already noticed,
the Pallavas did in fact achieve widespread dom inion in the m acro
region. Simhavishnu (c . a .d . 574-600) having first cleared the
plain of the ‘K alabhras’, thus ending that perceived menace,
successfully campaigned against the Cholas, then a power of
little consequence, and against the more formidable Pandyas of
M adurai. He and his successor M ahendravarm an I (c. a .d . 600-
630) also pressed their authority northw ard, perhaps to the Kistna-
Godavari. However, by the end of M ahendravarm an’s reign,
Pallava power had reached its peak. M ahendra was a king whom
history has remembered with much adm iration for his building
at M ahaballipuram and Kanchi, his conversion from Jainism by
the hymnist saint A ppar (himself a convert to Saivism from Jainism)
and his repute as a poet and musician. M ahendra’s successors
could do little more than to protect the wealth o f the central plain
from the raids of their long-standing enemies of the Deccan, the
Chalukyas of Badami. Later Pallavas were only partially successful
in this effort, for the Chalukya Vikram aditya II ( a .d . 733-744)
is said to have taken and looted Kanchi three times, a testimony to
the wealth which Tondaim andalam had come to possess and
its attraction for military adventurers of the age. To this may be
added the succession problems of the Pallavas which produced a
break in the main line with N andivarm an II, or Pallavamalla.
N andivarm an II was almost certainly a usurper of the Pallava
throne,114 one who may have come from a distant place, perhaps
even from South-east Asia as some scholars have suggested.115
114 Ibid., v. 1, p. 135 and N ilakanta Sastri, H S I, p. 148.
115 T.N . Subram aniam , The Pallavas o f K anchi and South-east A sia, The Swadesa-
But, more im portant than his usurpation — though perhaps a
result o f that act — the Pallava kingship altered its expression of
authority and sovereignty in fundamental ways. No longer were
Pallava kings referred to as the descendants of kings whose rule
was legitimated by reference to founders who performed Vedic
sacrifices, but rather on the basis of being the source of dana,
gifts, to Brahmans and to gods. Though this is a shift in the language
of royal authority, Pallava kingship was certainly no less sacral
than it had been. The king was still the anointed ruler whose
m ilitary prowess was overwhelming and was purified by the maha-
bhiseka as depicted in the panels sculpted upon the walls of the
Vaik unthaperum al temple in K anchi.116 But, thence-forward,
Pallava kings present themselves as munificent yaja m d n a s , centres
of a redistributional system of resources which encompassed the
world, their realm.
As significant as this change in the expression of royal authority
was a change in the conception of the king’s sovereignty in which
great local chieftains attained a status of generous dom inion only
slightly less than the king’s. This was achieved in the references
in inscriptional p rasastis to the vijhapti , or petitioner of grants
from the kings. In Pallava inscriptions earlier than Nandivarm an
II, a dom inant figure in inscriptions recording grants to Brahmans
and to gods was the ajnapti, or executor o f royal gifts, usually a
w arrior com panion of the king. Records of the time of Nandi-
varman and his successors make no reference to this djhapti,
but give great attention to the petitioner of a grant who, in all
cases, is presented as a powerful local chieftain. These records
portray this local personage as one who not only confers wealth
which is his own, within the scope of his k sa tra , but following
an established form of subordination to regal power seeks the
assent of the king for his grant. M oreover, the portion o f the
record praising the liberality o f the local petitioning chieftain
was not much less elaborate in praise of him than verses that
praised the reigning king.
An unpublished paper of Nicholas D irks117 provides a partial
m itran Ltd.. M adras, 1967, pp. 81-95.
116 C. Minakshi, ‘The Historical Sculptures of the Vaikunthaperum al Temple.
Kanci', M em o ires o f the Archaeological Survey o f India, no.63, M anager of Publica­
tions, Delhi, 1941.
117‘Political Systems in Ancient South India', February. 1974, D epartm ent of
basis for the preceding discussion o f changes in the expression
of Pallava authority and sovereignty. It is argued there that the
broadened conception o f sovereignty represented by the more
conspicuous place of local chiefs as king-like prestators was forced
upon the Pallava kings by the exigent pressure o f its enemies from
beyond Tondaim andalam . Such pressures cannot be discounted
of course, but what is more persuasively shown by D irk’s analysis
and the work of others is the broad acceptance of the Pallava
kingship as the focus of loyalty and participation by the great
chiefs of T ondaim andalam ; that is, as a proper regional kingship
co-equal with any of the macro region. From the late eighth
century onwards, Pallava kings using established ritual formulas—
of which the inscriptional p ra sa sti is one element — succeeded in
incorporating the established prestige of regional chieftainships;
it is upon such incorporation that kingship depended.
But their kingship was substantially restricted to this area
owing to the almost constant pressure of predatory raids of the
Chalukyas and their allies, the Gangas, from the north, and the
Pandyas and their allies, the Sinhalese, from the south. In all
of this, Chola country was reduced to the political ignominy of a
shatter zone over which Pallava and Pandyan soldiers moved
even as the sacred character of the Chola country was being etched
by the hymnist devotees of Siva and Vishnu. All that we see of
Cholas at this time is a line of warriors in the dry country north
of Tondaim andalam (modern Cuddapah, K urnool, and Bellary)
who called themselves Chola (or Choda) in inscriptions; they are
reported by Hsiian Tsang as ruling in these northern parts. Of
the U raiyur Cholas nothing is know n.118
Pandyas, not Cholas, were the great contenders with the Pallavas
for power in the southern peninsula prior to Vijayalaya Chola;
later, during the period of C hola ascendency, Pandyan territory
in the m odern districts of M adurai and Tirunelveli, along with
Tondaim andalam , constituted the second, m ajor intermediate
zone of the Chola segmentary state. Pandya country shared with
Chola country a regal tradition in classical antiquity which Tondai­
m andalam and the Pallavas did not. Beginning with the obscure
Pandya K adungon (c. a . d . 590-620), a line of warriors began their
History, University of Chicago.
118 K.V. Subrahm anya Aiyer, H istorical Sketch es o f Ancient D ekhan, v. i, pp.
112-13. Epigraphists refer to Telugu Chodas.
 kingly rise at the same time as the line of Pallava Simhavishnu.
As in the case o f m ost Indian dynasties — south and north — the
origin and early successes of the early rulers are known from a
few inscriptions. The Pandyan line is reconstructed from four
sets of copperplates of the late eighth and tenth centuries.119
These plates set forth mythic and real genealogical details and a
record of the peoples against whom the ruling lineage first contended
to establish their authority in the territory which became their
central dom ain — the Vaigai basin for the Pandyas — and then
those peoples and territories over which an overlordship was
extended.
The earliest Pandyan warriors mentioned in these inscriptions
are the same as those named in some of the Classical poems (e.g.,
the Purananuru ) and refer to kings whom the poems show to have
been great warriors and supporters of Brahmans and Vedic sacri­
fices (yagas) including the rajasuya and asvam edha. From the
account o f the ubiquitous Hsiian Tsang, the inference can be
drawn that the central dom ain o f the Pandyas, in the seventh
century at least, was probably a land of poor agriculture owing
to the salinity of its soil, though its people were avid traders.120
This conforms reasonably well with references to the Vaigai region
in Classical works and suggests that the interest o f Pandyan warriors
in the Kaveri to the north, which they came to dom inate in the
early ninth century, displacing the short-lived Pallava overlordship
.there, may have been prom pted by the rich food resources of the
western Kaveri delta.121 As in the case o f the Pallavas, one o f the
early rulers was spectacularly converted from Jainism. The
Pandyan Arikesari M aravarm an o f the late seventh century is
said to have been won to the Saivite faith by the saint, Jnanasam-
ban d ar.122 Finally, the Pandyan connection with Lanka from the
ninth century provides yet another sign of the energetic and ag­
gressive propensities of these rulers to strengthen the base of their
political operations. Beginning with Srimara Srivallabha in the
1,9 Discussed conveniently in ibid., pp. 99 ff: the late eighth century ‘M adras
Museum Plates o f Jatilav arm an '; the ‘Velvikudi Plates of N edunjadayan’; the
‘Smaller Sinnatnanur Plates’ ; and the tenth century ‘Larger Sinnam anur Plates’.
120 Ibid., p. 120 where Hsiian T sang’s ‘M alakuta’ is identified as m ilai-kurram .
121 Ibid., p. 115.
122 The dates of this ruler are given by Subrahm anya Ayyar, ibid., p. 127, as a . d .
650-80 and by N ilakanta Sastri, H isto ry o f South India, p. 165, as a . d . 670-700.
early ninth century and continuing through the time of Varaguna-
varman ( a . d . 862-8 80),123 Pandyans were actively involved in
Lanka politics not simply as predators, but seemingly seeking
a durable connection with it. Here again, the sea is to be seen not
as a barrier, but as a route o f opportunity no more hazardous
than the tracts of desolate country which bordered other im portant
population cores of the peninsula.
Unlike the Tondaim andalam intermediate zone of the Chola
segmentary state, Pandya country proved a difficult place for the
Chola overlords. Again and again the country lying south of the
Kaveri required punitive expeditions against would-be restorers
of Pandyan kingship. Indeed, so implacable was the resistance
to Chola rule that the policy o f Rajendra I to directly rule modern
M adurai and Tirunelveli through a viceroy could not be sustained
later. Parantaka I invaded Pandya country three times between
a . d . 907 and 953, and though he claimed the title m adiraikonda

(‘who took M adurai’) and, after his third incursion, the title
madiraiyum-Tlamum-konda (‘who took M adurai and Lanka’),
assertions of a separate kingship by warriors of the southern country
continued. Throughout the Chola period, Pandyan warriors,
often in alliance with warriors o f Lanka, challenged the Chola
claims to hegemony in the relatively well-settled core regions of
the Vaigai and T am braparni (Ambasamudram). In the more
desolate parts of the southern country the Chola overlordship was
often non-existent. R ajaraja I in a .d . 995 made the reconquest of
Pandya country one his first great ventures outside of Chola country,
renaming it, rájarája-pándinádu. U nder R ajendra I, resistance
to the Chola overlords resulted in the short-lived viceregal arrange­
ment, but this was to no avail.124
During the twelfth and thirteenth centuries, when Chola power
waned, a series of Pandyan warriors, notably one Kulasekhara,
had free rein over Lanka and constituted a m ajor source of m olesta­
tion there according to the Lanka chronicle, the M ah a va m sa .125
He and others also pressed claims to overlordship of the northern
parts of the m acro region as well, and under M aravarm an Sundara
Pandya I ( a .d . 1216-27) and Jatavarm an Sundara Pandya II ( a . d .
123 N ilakanta Sastri, H isto ry o f South India, p. 165.
124 Subrahm anya Aiyer, H istorical Sketch es o f A ncient D ekhan, v. 1, pp. 143-
55; N ilakanta Sastri, The P andyan K ingdom , pp. 100 ff.
125 Subrahm anya Aiyer, H istorical Sketch es o f A ncient D ekhan, v. 1, pp. 155-60.
1227-51), Pandyan power became supreme in the southern penin­
sula, looting and extracting hom age from warriors in southern
Chola country, Tondaim andalam , and the Hoysalas of Gangavadi
and K ongu.126 This century o f Pandyan expansion was brought
to an end by the invasion o f Muslims in a . d . 1310.
Pandya country differed from Tondaim andalam in the continuous
resistance o f its chiefs to an overlordship o f the Cholas. One factor
in this resistance was that Pandyans shared with the Cholas an
ancient tradition of kingship. Yet, the development of institutions
in this southern portion of Tamil country were similar to those
o f the central plain. In Pandya country, as in Tondaim andalam ,
there came to be large tracts of fertile land under the control of a
peasantry which called itself ‘Vellala’ and which shared a mythic
origin as well as the pretension to ritual parity with the dom inant
peasantry of the Kaveri and the central plain. Here, too, there
were brahmadeyas established with the ceremonial collaboration
o f such peasant groups and their chiefs and dedicated to the same
Brahmanical learning as elsewhere. These Brahm an villages,
as shown on M ap IV -1, are concentrated in the river basins of the
southern peninsula at Am basam udram , where ten can be located,
and at Tiruchchendur, Tirunelveli, Nanguneri, Srivaikuntam,
Sankaranainarkoil, K ovilpatti, Tenkasi, and Srivilliputtur. Around
M adurai, there were brahmadeyas at Periyakulam, M adurai itself,
N ilakkottai (in Dindigal taluk of the nineteenth century), and
Palni. These latter were in the western portions o f m odern M adurai
district which constituted a fertile catchm ent area for run-off
from the hill boundaries of the district. M any of these villages
date from the Chola period, some are earlier, and many were
formed during the heroic days o f Pandyan expansion in the
thirteenth century when a large num ber o f peasant villages were
brought together to form great settlements under the supervision of
a mahasabha. Finally, in Pandya country, as in Tondaim andalam ,
and in Chola country itself, there were substantial tracts of inferior
land, unimproved by irrigation and generally hazardous to man
and beast alike where there lived people whose culture and society
and predatory ways were such as to make their territories marginal
to each zone as a whole.
In Pandya country, as in Chola country, the names o f the inhabi-
126 Ibid., pp. 164-71 and N ilakanta Sastri, The Pandyan K ingdom , pp. 122ff.
tants o f such m arginal or peripheral tracts were M aravar and
K allar; in Tondaim andalam , such groups were more diverse,
including Telugu and K annada speakers in the northern parts
o f T ondaim andalam and, in the southern parts, Tamil-speaking
Pallis (who also were im portant in the western parts of Naduvil-
nadu). These folk were but interm ittently linked to the normative
order over which the Brahmans occupied an im portant place and
where the landed people, closely tied to Brahmans, exercised major
dominance under chiefs who recognized the overlordship of the
Chola kings. Some accom m odation of these often fierce, clan-like
folk was m ade by the peasantry and chiefs of the central domain
of the Cholas and those of the intermediate zones o f Pandim anda-
lam and Tondaim andalam , but in the main, peripheral areas can
be so designated because they were physically removed from the
core cultures of the zones proper. Thus, while these isolated tracts
or peripheral areas were considered a part of the central and inter­
mediate zones and though they were in regular contact with the
m ajor tracts of established peasant society and culture, these isolated
places developed distinctive characteristics of their own.
M uch o f m odern R am nad comprised a peripheral area within
the intermediate Pandya zone of the Chola segmentary state.
M ost of this area was p a la i according to the tinai classification of
the Classical poetry. It was a land so poorly endowed by nature
as to preclude large populations and reliable agriculture. During
medieval times, three kinds o f folk can be identified there, and the
interaction among the three provided for a distinctive social,
political, and economic configuration. The three peoples were:
M aravars, a militarily capable, clan-organized people who did
not fully acquire peasant culture until the nineteenth century when
irrigation provided a margin o f security to the hazardous agriculture
o f the region; a sizable community of fisher-folk along the some­
what sheltered coast o f the Palk Strait; and a vigorous trading
people, the N attukkottai Chettis, who date their immigration into
the T iruppattur and Devakottai divisions of m odem Sivaganga
and Ram nad from Chola tim es.127 D uring the late Chola period
and the time o f Pandyan expansion o f the thirteenth century, this
part o f the southern peninsula was rarely graced by temples and
127 Thurston, Castes and Tribes, v. 5, pp. 250-2. This claim to ancient residence
is reportedly denied by other trading groups of this region.
inscriptions, and obviously stood apart from the dom inant culture
zones to the west, though tenuously connected by trade and some
interaction. The N attukkottai Chettis, in recent times a prosperous
trading and money-lending community in the southern peninsula
and in South-east Asia, apparently began their trade activities
as carriers and traders of salt. An early name for these traders
was U ppu (meaning ‘salt’ in Tamil) Chettis.128
M aravars had an early ethnic identity in Classical works where
they are depicted as personifications of the harsh and forbidding
p alai. T hroughout the medieval period, these were a people apart
though they were never reduced to the opprobrium o f the British
period when they were called, with others, ‘criminal castes’.129
Indeed, M aravar chiefs achieved considerable eminence and
became m ajor political actors in the southern peninsula during the
time o f the Nayaks o f M adurai, in the seventeenth century. Later,
even after the British had branded M aravars as a ‘criminal caste’
they conferred recognition upon some of them who had not chal­
lenged, but at times assisted, the establishment of British control
during the eighteenth century. The latter were granted zamindari
estates and the greatest of them was recognized as the Sethupati
of R am nad.130 The progenitors o f many o f the M aravar chiefs
who played im portant roles in the seventeenth century and later,
may have been immigrants from the Kaveri or from Pandya
country displaced by the earlier consolidations of these zones rather
than descendents of the people described in Classical poems.
F or the adventurous warrior, or simply one forced to flee his
ancestral lands, this wild, dry country at the margins of increasingly
well settled lands dom inated by zonal ruling houses would have
provided opportunities, albeit dangerous ones.
W hat modern R am nad was for the Pandyan intermediate zone
— a land of poor agricultural capability, sparsely peopled, and
dom inated politically by clan-organized peoples — modern Puduk-
kottai was for C hola country. The northern edge of this Chola­
m andalam peripheral area, m odern K ulattur taluk, is directly south
of m odern Tiruchirapalli. In Chola times, this area consisted
of a large, sparsely peopled territory called K onadu or Trattapadi-
12» Ibid.. p. 262.
129 F.S. Mullaly, N o tes on the Crim inal Tribes o f the M adras P residency, M adras,
1892.
iso Ibid., 111-12.
kondachola-valanadu which extended to the southern Vellaru
river. F or the m ost part, then, as more recently, this tract had few
reliable water sources and poor soil of high salinity (kalar soil).131
The area constituted a substantial barrier between the population
o f the Kaveri around Uraiyur (modern Tiruchirapalli) and the
Vellaru basin in its upper and middle course near K odam balur,
about twenty-five miles from the Kaveri. The southern Vellaru
m arked the edge o f the peripheral area below the K averi; it formed
the basis for a large, if scattered, population concentrated in a few
large settlements with sufficient wealth to support some brahm adeya
settlements and temples. Almost all of these are within a few miles of
the river.132 The inscriptions of the southern Vellaru heartland
o f m odern Pudukkottai State refer to sixteen brahm adeyas of which
five existed in the Chola period, the balance having been established
during the Pandyan expansion in the thirteenth century.133 On this
somewhat narrow ecological base, a m inor power centre developed
which was uneasily poised between the Pandyans and the Cholas.
Here, the Irukkuvel chiefs exercised local dominance and some
degree o f larger influence on Chola politics.
These chiefs at K odam balur m aintained themselves in the
hazardous interplay o f Chola and Pandya conflict during the time
of Chola consolidation under Parantaka I and Aditya I. An
ancient line of chiefs, the Irukkuvels appear to have been in close
relations with the dreaded Kalabhras, and it is the speculation of
some scholars that they were part o f the K alabhras.134 During
the ninth century, the Irukkuvels prudently cast their lot with the
Cholas and became the mainstay of Chola power, serving in Chola
armies and intermarrying with the ruling family.135 During the
twelfth century these warriors suffered the fate o f many other
chiefs at any earlier time — they ceased to be visible as im portant
131 K .R. Venkataram a, A M a n u a l o f the P u d u k k o tta i S ta te , Pudukkottai, 1944,
v. 2, pt 2, p. 1012.
132 A C h ro n o lo g ica l L is t o f Inscriptions o f P u d u k k o tta i S ta te , Pudukkottai, 1929,
pp. 8 ff, for Chola inscriptions.
133 V enkataram a, op. cit., v. 2, pt 2, p. 652.
134 M. Arokiaswami, The E a rly H is to r y o f the V ellar B a sin ; w ith S p e c ia l R eferen ce
to the Iru k k u ve ls o f K o d u m b a lu r, A m udha Nilayam Ltd., M adras, 1954.
135 Ibid., pp. 100-1. Arokiaswami claimed that the title of m uvendavefar was
the special prerogative of the Irukkuvels, thus vastly exaggerating the im portance
of these chiefs. His later discussion of the term in the I H Q , op. cit., w ithout referring
to this earlier discussion, corrects that view.
political actors; unlike other chiefs whose independent identities
were incorporated by the Chola overlordship, however, the Iruk-
kuvels may have lost their dominance over the southern Vellaru
basin completely, and shifted their base of power to Kongu whose
chiefs, the Kongu Cholas, identified themselves as k o n a tta r (of
konadu).
In Arokiaswam i’s study o f this line of chiefs, the m ajor thesis
about the Irukkuvels and their origins is that they were the progeni­
tors of the Vellala people of the peninsula! It is a view which grants
to the mythology o f origin a plausibility seldom accorded, and it is
a view which is unacceptable. This dry tract, Kodam balur, in
the narrow , fertile land formed by the Vellaru its core, had chiefs
not drawn from Vellala peasants primarily but from the tough
hunters and pastoralists who peopled the Ram nad region in this
early period. Kallars o f Pudukkottai, like those o f Ram nad, and
M aravars o f both dry regions, were the prim ary sections of the
population. Here, land in only a few places was capable of support­
ing the full orchestration of elements which m arked the society
and culture of the plains proper. Only slowly — in m any places
not until the last two centuries — did peasant peoples from the
Kaveri seek fields and villages in these ecological and culturally
marginal areas. And, it was only under extreme pressure — much
o f it caused by threats from those tightly organized, rugged people
of the dry country — that peasant people found their way into
these areas. Such pressure is suggested by the ensconced presence
of ka lla r-m a k k a l in the lush kulam ahgala-nadu around Tirup-
puvanam in m odern K um bakonam taluk of Tanjavur according
to a thirteenth century Pandyan inscription there.136 Simultane­
ously, o f course, as tank-supported irrigated agriculture was very
slowly established in these dry places of the southern Tamil plain,
people there modified their lives to take advantage of these changes.
Some groups of Kallars and M aravars were transform ed to the
status of Vellalas by im itating the modes of life appropriate to the
peasantry, accounting for the Tamil proverb descriptive of these
places: Kallars in time become M aravars who, in turn, changed
to Vellalas.137
Having surveyed some of the general features of the com ponent
136 Thurston, op. cit., v. 3, p. 63.
w I.M .P . v. 2, p. 1245: A .R .E . 1911, n o .159.
parts of the Chola segmentary state, closer attention can now be
given to certain variations. Am ong indicators of differences
am ong the broad divisions o f the South Indian m acro region as
well as differences within each such division, the distribution of
brahmadeyas and temple inscriptions are perhaps the m ost sensitive
and im portant.
The central Tamil plain contained two groupings of segmentary
units: the m ajor grouping comprising the intermediate zone of
Tondaim andalam and a m inor grouping comprising the discrete
sub-region o f Naduvil-nadu. Since each o f these can be divided
into a large num ber of constituent parts, one is confronted with a
very complex region as a whole.
Partly, this complexity arises from the dispersed pattern o f
settlement over the central Tamil plain, a pattern determined by
the fact that at fairly regular intervals the plain receives drainage
from the interior upland which, given an ability to preserve water
through tank storage, made it possible to sustain clusters of ag­
rarian population. Through num erous breaks in the upland catch­
ment area at around the 1000 foot contour, some sixty-five miles
from the coast, these streams nourish the soil upon which are
dependent the two large populations of peasant peoples in T ondai­
m andalam and Naduvil-nadu, and many m inor ones from Lake
Pulicat in the north to the Ponnaiyar basin in the south. The two
m ajor population concentrations are the prim ary one of the Palar
Cheyyar basins around Kanchi and the secondary one o f the
Ponnaiyar-M allattar basins around Tirukkoyilur. Num erous
streams draining the K alarayan and Javadi hills traverse the plain
between the two m ajor settlement areas.
The distribution of brahmadeyas and temple inscriptions of the
Chola period are strongly correlated with these physical characteris­
tics of the plain, given tank irrigation. Brahmadeyas are found
in the central, Kanchi, portion of Tondaim andalam , in the middle
course o f the Cheyyar, and along the 250 foot contour between the
Cheyyar and K ortalaiyar river (13:15 degrees north). In the m odem
district of Chingleput, this would include the taluks o f Chingleput,
Conjeevaram, and Saidapet. Only a few Brahman settlements are
found in M adhurantakam taluk and fewer in Ponneri and Tiru-
vallur taluks. In N orth A rcot district, m ost o f such m arker
settlements are found in Cheyyar taluk (Arcot o f the late nine­
teenth century and earlier), Arkonam , G udiyattam , with a few
in the western parts o f W andiwash where the Cheyyar flows.
As to the southern cluster o f the central Tamil plain, the locus
of brahm adeyas and inscriptions is the Ponnaiyar basin. This
riverine tract and its surrounding territory, comprised Naduvil-
nadu. This zone seems never to have been long included in either
o f the m ajor concentrations o f population in Tondaim andalam or
Cholam andalam , though, at times, it may have been considered
a part of two broad valanadu territories which lay in the peripheral
area of Chola country north of the Kollidam : rajendrasim ha-
vcdanadu and virurajabhayahkara-valanddu. A substantial popu­
lation was concentrated in the Ponnaiyar basin, but beyond this
core, population was sparse and the basis of peasant agriculture
too slender to support much beyond the hardy Palli peasantry
(Vanniyars of later times), the only m ajor cultivating people to be
consistently associated with left castes (idahgai ).
The coastal plain supporting these m ajor peasant regions is from
forty to sixty miles deep; between the two clusters centred on
Kanchi and Tirukkoyilur the plain shrinks to about half of that
depth. Almost no Brahm an settlements are found above the 250
foot contour, and none are above the 500 foot contour which
runs at a depth o f fifty miles from the sea. W ithin this narrow,
well-watered corridor, the distribution of brahm adeyas and ins­
criptions of the age can be used to define the central, intermediate,
and peripheral areas of the two clusters.
Table VII-1 displays some highly selective Chola data com paring
the central Tamil plain with the core o f the Kaveri basin. The table
is based upon inscriptions collected between 1888 and 1915, and
published by the D epartm ent of Epigraphy, Governm ent of India.
V. Rangacharya organized these inscriptions arranged by taluk,
and district for the M adras Presidency,138 and the table is partially
based upon that topographical com pilation.139 In addition, data
upon which M ap IV-1 is based have been used for the taluk-wise
138 A Topographical L ist o f the Inscriptions o f the M adras Presidency, 3 vols.,
M adras, 1919.
139 As in any aggregative, as opposed to anecdotal, use o f inscriptions, it is im­
portant to consider that all of the extant inscriptions have not been noticed and
copied and that a large num ber of inscriptions have been destroyed and damaged.
O f the inscriptions used here it is also im portant to note that included are only those
collected to 1915 which means that they are very incomplete as well as biased toward
the more prom inent of the Chola sites, thus exaggerating some of the differences
drawn here.
 distribution of brahm adeyas. While a num ber of serious reserva­
tions must inevitably be raised about all of these data and their
use, Table VII-1 is intended to do no m ore than indicate certain
gross differences in the distribution of these m arker elements in
the central Tamil plain during Chola times. These data represent
a first approxim ation o f m ajor kinds o f segmentary units o f the
time.
Chola temple inscriptions and brahm adeyas are unevenly dis­
tributed in ways which would be anticipated from a knowledge of
factors related to agricultural production, hence to population
and wealth. Taluks with the greatest num ber of Chola inscriptions
or those which have the largest proportion of their total inscriptions
from Chola times also tend to have the largest num ber of brahm a­
deyas. In Tondaim andalam , these taluks are: Arkonam and
G udiyattam in N orth Arcot and Saidapet and Conjeevaram in
Chingleput; in Naduvil-nadu, the core localities are Tirukkoyilur
and parts o f Tindivanam, Cuddalore, and Villupuram (Chidam ­
baram taluk being grouped with Cholam andalam ). Certain anom a­
lies in these data should not be overlooked, however. Tiru-
vannamalai taluk with about seventeen per cent of the inscriptions
for N orth Arcot district, o f which about the same proportion
dates from the Chola period, shows only one brahm adeya, Puli-
yurnadi-chaturvedimangalam . There are also distortions which
must inevitably occur in aggregating data according to the taluk
and district territories of British India, for these administrative
units were constituted regions bearing no considered relationship
to modern social realities (except as the creation of these units
established such realities), much less historical ones. N otw ith­
standing such anomalies and distortions, the core portions of
Tondaim andalam and Naduvil-nadu are seen to have a lower
density o f both Chola inscriptions and brah m adeyas though the
absolute num ber of both inscriptions and Brahm an settlements
for the whole of T ondaim andalam (as shown in the taluks of
Table VII-1) slightly exceed those o f the core portions o f C hola­
m andalam (as shown in the data from Tanjavur district).
The same data are presented in Table VII-2 in a form which
presents areal clusters of the central Tamil plain according to
whether they were central, intermediate, or peripheral within the
two zones. The effect generally is to emphasize the graded quality
of the dispersal pattern over the entire central plain. Bounding the
308 P easan t S ta te and S o cie ty in M edieval South India

T able V II-1
D istribution of C hola Inscriptions and B rahm adeyas by
Selected M adras Taluks to 1915

Total Chola Percentage of

D istrict/Taluk Inscriptions Inscriptions C hola B rahm adeyas
Inscriptions

N orth A rcot 727 284 38 28

Arkonam 151 136 90 8
Arni 70 4 5 0
Cheyyar 91 34 27 6
G udiyattam 60 35 58 6
Polur 30 8 26 0
T iruppattur 6 1 15 1
Tiruvannam alai 124 22 18 1
Vellore 61 18 30 1
W alajapet 49 19 39 3
W andiwash 85 37 43 3

Chingleput 1221 596 49 47

Chingleput 230 93 40 6
Conjeevaram 219 118 55 16
M adurantakam 214 91 42 5
Ponneri 89 38 42 2
Saidapet 378 228 60 14
Tiruvallur 91 28 30 4

C hittoor
Chandragiri 22 8 36 1
C hittoor 15 4 26 1
Kalahasti 110 75 68 0
Punganur 62 19 30 3
Puttur 30 16 53 0
Tiruttani 80 48 60 4
Nellore
G udur 133 57 43 0
Polur 45 9 20 1

S o u th A rcot 1094 316 35 32

Chidam baram 195 52 38 6
Cuddalore 160 97 60 3
Kallakurichchi 2 0 0 0
Tindivanam 85 50 59 4
Tirukkoyilur 480 271 56 11
Villupuram 93 61 65 4
Vriddhachalam 46 6 13 4

Tanjavur 1639 1300 80 93

Arantangi 7 2 0 0
K um bakonam 478 324 68 25
M annargudi 116 85 73 6
M ayavaram 131 93 70 24
Nannilam 169 124 73 6
N agapatnam 77 58 75 4
Papanasam 121 99 82 9
Pattukkottai 49 30 60 4
Shiyali 34 32 95 5
Tanjavur 336 325 98 10
Tirutturaipundi 130 118 90 2

central clusters o f Tondaim andalam and Naduvil-nadu were

localities of diminishing populations sited with increasing jeopardy
as the rising, drier lands of the bordering hill country is approached
and phasing into the peripheral zones of Kongu on the west and
Gangavadi to the north. Zonal divisions of Tondaim andalam are
shown in M ap VII-2. The central (C), intermediate (I), and
peripheral (P) zones are represented along with some o f the more
im portant sacred centres and settlements o f the m andalam .
Kongu country and Gangavadi were peripheral zones of the
Chola segmentary state system. Their links with the Chola system
are less related to conquests by Chola armies from the plains
than to the movement o f cultivators to these previously isolated
upland extensions o f the Tamil and Mysore plains. Prior to the
time of the Cholas, the isolation of Kongu, a relatively sparsely
settled region,140 was only occasionally disturbed by m arauding
warriors and peoples moving between the Corom andel plain and
K arnataka. At such times, Kongu was a shatter zone and, like
other such regions in India, showed complex overlappings o f peoples
and cultures. Gangavadi, on the other hand, had long been the
centre o f a m inor peninsular dynasty — the Gangas — after whom
the area was named. It was conquered by R ajaraja I in a . d . 1004
and was thereafter subjected to heavy Tamil in-migration consisting
not only of Tamil peasant groups but of Tamil Brahmans as well.141
From the eleventh to the fifteenth century the fundamental influence
upon both peripheral zones came from the Kaveri, and the result
was an overlay o f Chola culture and institutions which did not
expunge the earlier diversity of these places but only added to their
variegated character.
Kongu was a troubled frontier region during the early period
of the great Cholas. Called malai-nadu in the Tiruvalangadu
Plates, or vTrachdlamandalam in m ost other Chola inscriptions, it
was a place where Chola and Pandyan warriors sought dominance
and where distinctions were not always made between Kongu,
southern K arnataka (Gangavadi), and m odem C oorg.142 An
a . d . 1012 Tamil inscription143 o f R ajaraja I from Balmuri in
Gangavadi refers to his conquests in Gangavadi, M alenad (Coorg
and the western hill portion o f m odern K arnataka), and Kongu
as well as Nolam bavadi (modern south-central K arnataka),
A ndhra, Kalinga, and Pandya countries. However, after Rajaraja
I’s time there are few Chola records. Chola inscriptions of
P arantaka I, Parakesarivarm an U ttam a-Chola, R ajaraja I,
Rajendra I, R ajendra II, Virarajendra, and K ulottunga II and III
are found at Erode, Valappanadu, Tiruchchengodu, and K arur
140 Arokiaswam i, K o n g u C o u n try, p. 257.
141 Rice, M y s o r e ; A G a ze tte e r, p. 316.
142 Arokiaswami, K o n g u C o n try, p. 202; also Subrahm anya Aiyer, ‘The History
of K ongu C ountry’, in H is to r ic a l S k e tc h e s o f A n cien t D ekh an , v. 2, where other
names for K ongu are listed: A dhrrajarajam andalam in the eleventh century and
Cholakeralam andalam in the twelfth century, p. 55.
143 Cited in Arokiaswami, K o n g u C o u n try, p. 205; E .C ., v. 3, sr. n o .140.,
VTí-2 T o n daim andalam Z onal Divisions c. a .d . 1300
in the northern and eastern parts o f Kongu. The southern and
western portions of the territory appear to have been less directly
affected by the Cholas. Though the inscriptional link to the Cholas
and the Tamil plain were tenuous, there was certainly a movement
of peasant peoples from the plains to the Kongu upland, and this
m igration was little affected by the replacement of Chola by Hoysala
ascendency in the early twelfth century.144 Hoysala Ballala I
in about a . d . 1120 extended his dom inance over Kongu from
Talakad on the upper Kaveri in Gangavadi (and formerly the
G anga capital), thus joining for a time the two peripheral zones
of the Chola segmentary state to provide a base for an assault
upon the lower Kaveri itself by his Hoysala successors.
Though Kongu was an isolated area, there was continuous
contact with the Kaveri region by an ancient road, the kohgu-
peruvali, connecting the Corom andel and M alabar coasts along
which have been discovered hoards o f R om an coins.145 Moreover,
the devotional m ovement o f Tamil Saivites touched Kongu where
there are seven sacred sites according to the D evara m ,146 N o t­
withstanding these contacts, the isolated, or at best the route
character of Kongu is underscored by an analysis of inscriptions
found there.
Arokiaswami has estimated that about 900 Kongu inscriptions
had been noticed by epigraphists by the mid 1950s. O f these about
630 can be attributed with fair accuracy to particular dynastic
periods, the remainder being miscellaneous grants to temples
by local notables recognizing no overlordship.147 Well over half
of the inscriptions which can be associated with an overlordship
in K ongu belong to three lineages of chiefs, each o f which took as
their surnames, titles, and epithets those o f one o f the great ruling
lineages of the macro region.148 Thus, there were Kongu-G angas
144 Subrahm anya Aiyer, in H is to r ic a l S k e tc h e s o f A n c ie n t D ek h a n , v. 2, p. 55.
1« Ibid., p. 44.
1« Ibid., p. 49.
147 Arokiaswami, K o n g u C o u n try, p. 22.
148 K.V. Subrahm anya Aiyer (‘History of the Kongu C ountry’, p. 45 H is to r ic a l
S k e tc h e s o f A n cien t D ek h a n , v. 2) points out th at the K ongu Cholas even used the
rotational terms p a r a k e s a r i and ra ja k esa ri. In the same a uthor’s edition of several
Kongu inscriptions of the tenth and eleventh centuries, he argues that this line of
chiefs should be called ‘K ongu K o n a tta r' (from K onadu in m odern Pudukkottai:
E l v. 30, no. 19, ‘Seven V atteluttu Inscriptions from the Kongu C ountry’, p. 95).
from the eighth century with ten inscriptions, Kongu-Cholas during
the period from a . d . 1004 to 1275 with about 300 inscriptions, and
Kongu-Pandyas of the thirteenth century with some 120 records.149
In each case, titles and personal names are taken from rulers of
dom inant regional dynasties and used by great Kongu chieftains
who bore no relationship — kinship or otherwise — to the royal
Cholas, Pandyas, or G angas.150 It was a rudim entary form of
kingship appropriate to the peripheral zone of a segmentary state
in which the royal style was quite specifically linked to that of the
great kingships of the macro region because there were lacking
traditions of legitimacy to sustain a fully autonom ous kingship
in these places. Borrowed, or vicarious, legitimacy was necessary
and this apparently sufficed to provide for the Kongu region
a ritual centre around which locality chiefs could order their
political relationships and attem pt to cope with increasing pressures
of external spoilers in the form of invading armies from the south
and north as well as with the influx of new people.

T able V I1-2
Segmentary Areal Groupings in the Central Tamil Plain
during Chola Times a

Classification K otta m /N a d u D istrict/Taluk

T O N D A IM A N D A LA M

Peripheral Areas
N orth and West Tiruverigadu-k. C hittoor Kalahasti
C h itto o r P u ttu r
Pavattiri-k. C hittoor Chandragiri
Panganuru
Paiyyür-Ilan-k. C hittoor
Nellore G udur
K u n ra v a tta n a -k .
Nellore Pôlür

a. Based on Subbarayalu, op. cit., Appendices and Maps.

149 Arokiaswami, Kongu C ountry, p. 242.
150 Ibid., p. 22 and K.V. Subrahm anya Aiyer, ‘History of the Kongu C ountry’,
H istorical S ketch es o f A ncient D ekhan, v. 2, p. 60. The latter (p.44) argues that
K ongu was not a kingdom, but merely a territory, since its rulers lacked titles of
their own. However, the functions o f these rulers were royal in the then prevailing
sense of it on a primitive level and w ithout benefit of clergy (e.g., the m ahâbisekha).
 Chingleput Tiruttani
Chingleput Tiruvaliur
lkkâttu-k. N. Arcot
Manayil-k. W alajapet
Mëlür-k. N. Arcot Vellore
Interm ediate Areas
N orth and West Paduvür-k. Chingleput Ponneri
Paiyyür-k. Chingleput Tiruvaliur
Palakunra-k. Chingleput T iruttani
Pulal-k.
Central Area Àmür-k. N. Arcot Arkonam
Chengattu-k. N. Arcot G udiyattam
Eyii-k. N. Arcot Cheyyar
K a|attür-k. Chingleput Chingleput
K.aliyür-k. Chingleput Conjeevaram
Puliyür-k. Chingleput Saidapet
Tamar-k.
Interm ediate Areas U rrukattu-k.
South and West Chembür-k. N. Arcot Arni
IndOr-k. N. Arcot Polur
Venkunra-k. N. Arcot Tiruvannam alai
N. Arcot W andiwash
N. Arcot T iruppattur
Chingleput M adhurantakam

Oyma-n. S. Arcot Tindivanam

Peripheral Areas Singapura-n.
South and West
Peru-Mukkil-n.

N A D U V IL-N A D U

Peripheral Area Àdaiyür-n. N. A rcot T iruvannam alai

N orth and West
Interm ediate Area Vavalür-n. S. Arcot Villupuram
N orth and West
Central Areas Kllkondai-n. S. Arcot T irukkoyilur
U daikkattu-n.
K urukkai-kürram
Idaiyârru-n.
K udal-Ilâdaippâdi-n.
Interm ediate Areas Kllanm ur-n. S. Arcot Cuddalore
South and West K ayappâkkai-n. S. Arcot Vriddhachalam

Vagür-n.
M êrkür-n.

b. T iruttani has been part o f C hittoor district at times.

314 P easan t S ta te and S o cie ty in M edieva l South India

Kudal-n.
Peripheral Area M elam ir-kurram S. A rcot K allakurchi
South and West

In his work on Kongu — one o f the few serious attem pts at

local history during the pre-m odem period — Arokiaswami out­
lines the way in which new peoples were assimilated into the
society o f the isolated upland region from about a .d . 1000. He
sees the horizontal segmentation of caste groupings as a means
by which newcomers found places in the emergent society of this
peripheral zone. The pafija ja ti, or five Kongu castes of modern
C oim batore,151 are an example o f that process which dates from
the first Chola incursions in the tenth century. The peoples displaced
by these newcomers from the Tamil plain were pastoral and hunting
peoples referred to as K urum bar and Vedar in the chronicle
literature used by Arokiaswam i.152 Among the migrants to Kongu
o f the early imperial Chola period who took the appelation of
Kongu as a caste name, there came to exist special, close and
interdependent relationships reflected in the kinship terminology
and in ritual interactions.
These multiplex bonds served to differentiate the immigrants
of the tenth and eleventh centuries from those who came in the
thirteenth century according to Arokiaswam i.153 The latter are
distinguished by different caste appelations and by a different
order o f interrelationships such as to constitute a parrallel ordering
o f castes ranging from respectable Sendalai Vellalas ( tentisai-
choliya-velldlar ) o f the Kaveri and Padaiyatchi-Vellalas of T ondai­
m andalam to untouchables.
Arokiaswam i’s view of this process may exaggerate the extent
to which Kongu was opened to peasant occupation and agriculture
in a time as late as the Chola Period. The inscriptions o f the Kongu-
Gangas and others o f the pre-Chola period suggest a well-estab­
lished, if sparsely settled, agrarian society, and there are very
early references to two of the bulwarks of K ongu agriculture: well
irrigation and a superior strain o f draft bullock from around
151 K ongu C o u n try, p. 267; these include cultivators (Vellalas or K avundars),
m erchants (Chettis), barbers, washermen, and untouchable field labourers.
152 Notably, the K o n g u d e s a ra ja k k a l , English trans. by the Rev. William Taylor,
The M a d ra s Jou rn a l o f L ite r a tu r e a n d S cien ce, v. 14, pt 1 (1847), pp. 1-66.
153 K ongu C o u n try, pp. 270-1.
Kangayam. It is more likely that the Chola period saw an intensifi­
cation of migration to the Kongu upland, possibly the result of
military activities here and in K arnataka by the Cholas, and an
adaptation by newcomers o f the existing agricultural techniques
which have proved to be am ong the most adaptive in South India.
The crystalization o f these diverse populations of Kongu into
a single, complex local society may have come, as suggested by
Arokiaswami, in the thirteenth century. M igrants of that time
arrived in Kongu carrying new elements of culture from the plains;
there is also evidence to suggest that these most recent migrants
may have been encouraged and assisted by the threatened Chola,
K ulottunga III.154 These later m igrants brought m ore of the
established elements of plains culture, especially Vedic temples
and brahm adeyas, institutions which were at that very time under
considerable pressure in the coastal plains; in addition they brought
notions about and experience with such cultural and social forms
as temple towns and the division o f castes into those of the right
and left.155
The focus of much o f the society and culture o f Kongu Vellala
castes was upon four chieftainships which may have risen by the
later Chola period. These four families of chiefs, called p a tta k k a -
ran, bore titles allegedly conferred at that time, and in two cases
at least early evidence in support of these claims are found. The
four houses of chiefs are: the Palaiakottai p a tta k k a r a r of the
p a yira clan of Kongu Vellalas; the K ataiyur p a tta k k a ra r o f the
p o ru lan tai clan; the Putur p a tta k k a r a r of the sehkannan clan;
and the Sankarantam palaiyam p a tta k k a r a r o f the p eriya clan.
The first element in their titles (e.g. p a la iy a k k o tta i) refers to the
village centre of the chieftainship; all are located in the central
part of Kongu. Clan (kulam ) designations o f these chiefs families
appear in the earliest docum ents in which they find m ention,
and the m odem functions o f the clan organization upon which
these chiefs’ houses are based is given considerable attention by
Brenda Beck in her recent work on Kongu peasant society.156
154 S e e A rokiasw am i’s argum ent on this in ‘Social Developm ents under the
ChSlas’, T .A .S .S .I., 1956-7, pp. 3-4.
155 Kongu C ountry, p. 278.
iss Beck, op. cit., pp. 41-2. She refers to these chiefs as ‘titled families’ and as
an ‘aristocracy’ whereas the Tam il L exicon, v. 4, p. 2419, gives as one m eaning ‘title
holders’ m eaning by that the possessor of a reven u e p a tta in British usage; the Lexicon
Her reports from contem porary inform ants as well as her references
to vernacular histories of the Kongu Vellala (or Kavundar). com­
munity on the antiquity of the four families of chiefs are supported
by inscriptional and literary evidence going back to the sixteenth
century. Two inscriptions o f the time o f M allikarjuna of Vijaya-
nagara refer to the Payira clan chiefs as kohgaveladaraiyar (chiefs
o f the Kongu Vellalas),157 and the eighteenth century Kongu-
m an dala-Satakam of the Jaina-Brahm an Karm egakkavinar refers
to the k d ta iyu r chiefs o f the poru lan tai clan and the p u tiir chiefs
of the sehkannan-kulam renown for having received the title of
m um m udi-Pallavan from a Chola king.158 These several chieftain­
ships of the dom inant cultivating castes of Kongu not only pro­
vided political focuses, an expression o f their political primacy
among cultivating groups in this area, but they linked even these
remote Tamil cultivators to a kingly centre in the Kaveri.
Gangavadi, like Kongu, is a relatively dry tableland girded by
m ountains which served to preserve its isolation until Chola times.
However, this territory achieved considerably greater political
prominence than Kongu because, around the fourth century, there
emerged a m inor kingly lineage calling itself Ganga. This line of
rulers persisted in southern K arnataka for almost six centuries.
Though the Ganga kings were not unique in claiming a connection
with Aryavarta, their persistent support of and possibly their
personal commitment to Jainism sets them apart from most other
ruling families in South India.159
Gangavadi was the oldest, largest, and best developed of the
three ancient territories o f central and southern K arnataka. The
other two were Nolam bavadi and Banavasi. Nolam bavadi formed
the north-eastern boundary o f Gangavadi, north of the north
Pennar river and extending to the Tungabhadra. It was settled
by sedentary cultivators long after the Ganga country when, around
the eighth century, Tamil warriors called N olam bas associated with
the Pallavas o f Tondaim andalam established their control there.
also gives the m eaning of headm an of T ottiyar and K ongu Vellala castes. The
notion o f chief would seem to fall between these meanings.
1s 7 A .R .E ., 1920, nos. 235 and 239.
158 korigum an4aIa-satakam o f ka rm eg a kka vin a r, ed. T.A . M uttusam ikkonar,
M adras, 1923, verses 51, 53, 72.
159 On the question of Jainism see S. Srikantha Sastri, E arly Gahgas o f Talaka4,
published by the author, Bangalore, 1952, pp. 44-9.
Banavasi, to the north-west o f Gangavadi, was even later to emerge
as an im portant area o f peasant agriculture; it was a small territory
and of lesser political significance.160 Though m ost of the period
to the eleventh century when Chola power was established in
Gangavadi, G anga rulers dom inated their two near neighbours in
the m anner of medieval Indian states, that is, little was done to
diminish the local authority of its subordinate families o f chiefs.
The prime orientation of the Ganga chiefs was not to northern
K arnataka, where political dom inance was exercised by the Chaluk-
yas and R ashtrakutas and whose subordinates the greatest of the
G anga chiefs were, but to the south and east. Early in the history
of the family, G anga chiefs, as allies of the Chalukyas, raided the
Pallava country o f Tondaim andalam on its eastern border. At
that time, G anga power was greatest in the east, on the edge of
Tondaim andalam , where the m ountain fortress o f Nandidrug,
at 5000 feet, provided a base for predatory operations.
Later, the centre o f Ganga power shifted westward to Talakad
on the Kaveri, and in the late eighth century, under the chief Sri
Purusha, the G anga capital was M anne or M anyapura, some thirty
miles from m odem Bangalore.161 From here, later Gangas pressed
upon their southern neighbours in Kongu. Butuga and his son
M arasim ha provided the m odel of kingship for K ongu chiefs who
adopted G anga symbols of kingship. But, beyond that influence,
later Ganga rulers, soon to be displaced by Chola influence,162
laid the basis for a new line of rulers, the Hoysalas. The latter
came to control Gangavadi in a .d . 1116 and changed its name to
to hoy salar aja .163
D errett’s judgem ent of the six centuries o f Ganga rule is some­
what harsh. He states that after the impressive reigns of Butuga
and M arasimha,
G anga affairs . . . sank back into the state o f com parative insignificance
which had been norm al before B utu g a’s time, when the central and

160 B. Lewis Rice, ‘G angavadi’, C o m m e m o ra tive E ss a y s P r e se n te d to S ir R a m -

krish n a G o p a l B h a n d a rk a r, Poona, 1917, pp. 237-8; J.D .M . D errett, The H o y sa la s:
A M e d ie v a l Indian R o y a l F a m ily, O .U .P., M adras, 1957, pp. 12-13.
161 S e e S. Srikantha Sastri, E a rly G angas o f T a la k a d , Rice, ‘G angavadi’, p. 239.
162 D errett, op. cit., p. 14. Parantaka I, U ttam a Choja, and R ajaraja I raided
eastern G anga country in the late tenth century; R ajendra I took T alakad in a . d .
1004.
163 Rice, ‘G angavadi’, p. 248.
 southern parts o f K arn atak a cou n try were rem arkable for n o distinctions
in the fields political, literary, architectural, o r religious. M oreover, from
the sm all num ber and often illiterate character o f the records th at survive
it m ay be judged th a t th e general level o f prosperity w as very low com pared
w ith the succeeding centuries, and it may be safely assum ed that the
historical obscurity th a t su rrounds m uch o f the G an g a period is due very
largely to the backw ard co ndition o f the m ajority o f its subjects.!®4

In becoming the central dom ain of the Hoysala state of the

twelfth and thirteenth centuries, Gangavadi ceased to be a peri­
pheral zone of the C hola segmentary state. But, in order to be the
Hoysala central domain, Gangavadi had to possess population
and wealth upon which to base such a powerful state. Thus, one
m ust alter D errett’s assessment of the Gangas and Gangavadi
prior to Hoysala times to the extent of recognizing that while the
Gangas did not match the Hoysalas in the quality and power of the
state or in literary, architectural, or religious accomplishments,
theirs was a territory which could sustain the excellence o f the
Hoysalas. In any case there is no evidence that the Hoysalas did
anything dram atic to enrich the land that they gained from the
Gangas and that sustained them. D errett makes no mention of
intensive irrigation development under the sponsorship of the
Hoysala state, nor was there even the m ore modest irrigation
development which can be dem onstrated for the Cholas and the
Vijayanagara rulers later.165 The m any localized irrigation schemes
developed during the age were the work of locality folk in the
nads under their headmen, the nadgdimdas, just as it was elsewhere
in the m acro region.166 Indeed, the very lacklustre of the Ganga
rulers who preceded the Cholas and Hoysalas suggests that they
were essentially peasant chiefs who neither sought nor m anaged to
break their ties with the dom inant peasant folk of the territory. That
peasantry still identifies itself with the ancient G anga designation;
they are called, garigadikaras who in 1891 comprised forty-four
per cent o f the total population o f the land-controlling peasantry of
M ysore State (i.e. Vokkaligas).167 G angadikara is a slight contrac­
tion of the term gangavadikara, ‘men of the Ganga country’.168
164 D errett, op. cit., p. 14.
165 See his chapter seven on adm inistration and his analysis o f revenue which
indicates no irrigation charges.
166 Dikshit, op. cit., pp. 41, 46, 52.
161 A nanthakrishna Iyer, M ysore Tribes and Castes, v. 1, p. 125.
168 Rice, ‘G angavadi’, p. 237.
 The G angadikara peasantry of Gangavadi appears to have been
m ore significantly linked to the K ongu peasantry to the south
than to peasant peoples in the central and northern parts of medieval
K arnataka. Similarly, the M arasu Vokkaligas of eastern Bangalore
and central and southern K olar districts appear to have been
linked to Tondaim andalam . Chola inscriptions are heavily dis­
tributed over the three modern districts o f Mysore, Bangalore,
and K olar, the core of G angavadi.169 One of these Tamil inscrip­
tions, th at o f M ulbagal taluk, Kolar, dated a . d . 1072, relating to
the interaction of Tondaim andalam peasants of the valangai
and those o f m odern K olar (then ‘Sri Rajendrachola Eighteen
Countries’) was discussed at length above.170 It may have been
just such ancient connections as those between the peasantry of
Ganga country and of Kongu which persuaded the Hoysala Ballala
II (Vira Ballala) to fish in the troubled waters of Chola country in
a . d . 1218. His son, Narasim ha II (Vira Narasim ha) established

Hoysala power in the lower Kaveri at K annanur (near Srirangam,

Tiruchirapalli district). D errett calls this action an ‘aberration’
because it prevented the establishment of a ‘K arnataka national
empire’ that D errett believed to be a ‘proper conclusion to the
period of upheaval which had preceded the rise of the Hoysala’.171
Such an assessment of the Hoysala incursion into the lower Kaveri
must be judged a modern historian’s ‘aberration’ which sees
nationalism as a factor, a ‘natural force’, in an age to which it
was wholly inappropriate. It was rather the rich Kaveri delta,
exposed by the weakened and divided Cholas, which offered the
best potential for Hoysala power. The Gangavadi ‘hom eland’172
of the Hoysalas, after all, opened not to the northern territories,
bleak and troubled, but to the rich delta of the Kaveri which
inevitably tied southern K arnataka to the Tamil plain.
The K arm andala S atakam , possibly o f the seventeenth century,
fortifies the presumption o f linkage between the peasant peoples
of southern K arnataka and those of Kongu and the Tamil plain.173
The work o f a Kongu Vellala, Araikilar, this sa ta k a m describes the

170 Supra, chap. 3.

171 D errett, op. cit., p. 105.
172 A nanthakrishna Iyer, M ysore Tribes and Castes, v. 1, p. 106 passim uses this
expression.
173 M adras, 1930, with a comm entary by P.A. M uthuthandavaraya Pillay.
country of the karalar peoples as composed o f two parts: rich
plains (polil-nadu ) in the east and hills ( varai-nadu ) in the west.174
K arm andalam , or garigapadi was bounded by these western hills
and the ocean on the west and ‘Pallavam ’ (i.e. Nolambavadi),
‘K adam ba’ (i.e. Banavasi), Tondai, and K ongu.175 The territory
included the Coorg Hills (Kudaku-m alai), the Pushpagiri-Subrah-
m anyam Hill (of U ppinangadi taluk, South Kanara), the Pennai
(N. Pennar) and Kaveri rivers, and it consisted of thirty-two
nadus in the western hill country, and sixty-four nadu s in the plains
(verse 7); the com m entator on this sataka m suggested that the
ninety-six nadus of Gangavadi account for its being called a ‘96,000’
country.
Verse forty-five of the satakam states that gahgapadi was known
as m um m adichdla-m andalam as well as mahilasura-nddu. The
latter name is the slightly altered m ahisasura-nadu (to become
(‘M ysore-nadu’). M udikondachola-m andalam was the common
name by which the Cholas referred to Gangavadi whereas mum-
m adichola-m andalam usually referred to n o rth ern . L anka.176
K aralar means ‘ruler of the clouds’ and is thus the same as
k a ra k k a ta r as in the sub-caste title, K arkatta Vellala.177 The
goddess Cham undi, whose shrine is in m odern Mysore City, is
said in the satakam to have provided the K aralar cultivators of this
country with rice seeds for their fields (verse eight), and another
famous town, Tagadur, or Dharm apuri, in m odern Salem, north
of the Toppur divide, was also included in the Ganga country
(verse eleven). Verses thirty-one to thirty-three state that the
K aralar are o f the Ganga people who include the vTrattiyar (Jainas)
in the western hill country and srd ttiya r (worshippers of Siva
and Vishnu) in the eastern part. They are progeny of the powerful
G anga chiefs, and (verse forty-six) the satakam ridicules the notion
that R ajaraja the Great conquered karm an dalam ; rather, his rule
was sought because of his famous devotion to the Lord Siva.
174 P o r il m eans forested or park land; v a ra i m eans m ountainous. D E D , nos.
3723 and 4315.
175 Because the N olam bas were understood to be a branch of the Pallava family,
N olam bavadi was referred to by that title (A nanthakrishna Iyer, M y s o re T ribes
a n d C a ste s, v. I, p. 99 n); similarly, Banavasi was also designated by its great w arrior
family, the K adam bas (Dikshit op. cit., p. 16 and G. M oraes, The K a d a m b a K u la ;
A H is to r y o f A n cien t a n d M e d ie v a l K a rn a ta k a , Bombay, 1931).
176 N ilakanta Sastri, The C o la s, pp. 472, 173.
t77 Thurston, C a ste s a n d T rib es, v. 3, pp. 249-50.
Finally, verse forty-eight states that the K aralar rule the people
of Ganga, Kongu, and Kalinga; the last-mentioned reference
is to the Eastern Gangas of Orissa.
Setting aside the bravura of the sata k a m genre, the K arm andala
S a ta k am helps to define the southern K arnataka country as one
linked to the segmentary state o f the Cholas. R ajaraja I’s conquest
of southern K arnataka — including Gangavadi, Nolam bavadi,
and Taigaivadi (modern Mysore district) — came by the sixth year
of his reign, a . d . 991. The Chola conquerors came through Kongu,
and this was clearly a part of the consolidation of the Chola over­
lordship in the western peripheral regions. In early tenth century
records from Gangavadi, Ganga chiefs are said to recognize
R ajaraja’s rule.178

IV
The P o litica l Culture o f the Chola S egm en tary S ta te.

The structure of the Chola segmentary state — its central, inter­

mediate, and peripheral zones with their differentially segmented
internal divisions — was not explicit in the numerous inscriptional
records nor in the various genre of literature of the age. It is an
inferred structure. Lacking an explicit, relevant, contem porary
theory of the state, it has been necessary to suggest another formula­
tion to organize the existing political facts of the age. The m ost
appropriate of alternatives is the theory of the segmentary state
proposed by Southall.
As argued above in the discussion of the structure of the Chola
segmentary state, effective territorial sovereignty of the Cholas
was confined to the rich, populous core of the Kaveri delta. Beyond
this sub-region, Chola sovereignty was an increasingly ritual
hegemony as the peripheral zones o f the state were approached.
The core territory of the Cholas included the following six deltaic
territories: pdndi-kulasam i-valanddu, including Tanjavur; nitta-
vindda-valanddw, arum olideva-valanadu, also called ten-kaduvay
or the territory south of the Kaduvay River; kshatriyasikham cm i-
valanddu, also called vada-kaduvdy or the territory north of the
Kaduvay; u yya-kon dar-valan adu ; and rajadhirdja-valanddu ,179

178 N ilakanta Sastri, The Cotas, pp. 174-5.

179 Venkayya, ‘Territorial Divisions of R ajaraja-C hola’, S .I .L , v. 3, pt 5, pp.
21-9.
The names o f these territories were epithets of R ajaraja I, fixing
upon the entire core region an identity with the ruler which was
as clear in purpose as the great shrine he built at Tanjavur. The
special relationship between the Chola rulers and their central
domain, the Kaveri core, and the role of the Rajarajesvara temple
together distinguish Chola country from the other m ajor segments
o f ‘zones’ comprising the state. The two features are symbolically
linked in a variety of ways.
In the origin m yth of the Chola family, prim ary emphasis is
placed upon the Kaveri region and symbolic connections established
am ong a num ber o f elements which were to remain im portant
throughout the history o f the great Cholas. This m yth occurs
in a Sanskrit inscription o f a .d . 1070 recording a grant in the
time of V irarajendra C hola to the shrine of the goddess Kanya-
pidariyar at K anya K um ari.180 After a lengthy iteration of the
solar dynasty with which the Cholas linked themselves (verses
4-27), a series o f verses (28-35) relate an episode modelled upon the
R am ayana. Here, an eponymous Chola w arrior who roamed the
Gangetic forests in carefree hunt was, like Prince R am a of Ayodhya,
deceived by a demon in the guise o f a deer and led southward from
Aryavarta. The w arrior slew the deer on the Kaveri and while
there looked for Brahm ans upon whom to confer gifts. Finding
none, he settled some Brahmans from Aryavarta who recognized
the Kaveri as a river of even greater sacredness than the G anga.181
Im portant in this m ythic account are three elem ents: the putative
Aryan connection of the progenitor of the line, the dedication to
Vedic, or at least Brahmanical rites leading to the introduction of
A ryan Brahm ans, and the superior sacredness of the Kaveri.
As the learned editor of the K anya Kum ari inscription, K.V. Sub­
rahm anya Aiyer, notes, the Cholas were not a northern people.
N o r is it any longer possible to accept the idea that the Brahmans
of medieval South India were primarily descendents of early
migrants from the north even though there was a continuous move­
m ent of small groups of N orth Indian Brahmans into the south
during ancient times.
180 K.V. Subrahm anya Aiyer, ‘K anya K um ari Inscription o f V irarajendra’,
T .A .S ., v. 3, pt t, no.34, pp. 87-158. The grant portion of this long inscription
consists o f barely twenty lines of Tamil.
isi Ibid., p. 96.
 The m yth of Gangetic origin is rather to be understood as an
effort to place the Chola kingship upon a basis o f legitimacy superior
to th at which could be claimed with equal validity by Pandyans and
by warriors of the Chera country. This was no easy task considering
the continued currency o f the titles m uvendavelar and mummudi,
held by chiefs recognizing all o f the three ancient kingships o f the
m acro region. In addition, such a claim o f Gangetic provenance
was appropriately antecedent to, or supportive of, the royal cult
to Siva which came into existence during the time o f Rajaraja.
The notion of territorial sovereignty fortified by religious asso­
ciations is scarcely unique with the C holas; similar associations
involving the conception of A ryavarta were made by kings of the
Gangetic plain at the tim e182 and are a basic element of sacral
kingship of time. In addition, however, the conception derived
from two processes in South Indian medieval Hinduism. These
processes came to be joined during the Chola period and provided
the foundation for the royal Siva cult centred upon the Rajara-
jesvara temple. Both processes may be identified during the Pallava
period, and it is to the credit o f R ajaraja’s political acumen that
they were incorporated into the state system which he created.
The first process was the revival, or more accurately, the
establishment o f worship of the Vedic Siva as the m ajor form of
high ritual in Tondaim andalam under the sponsorship of the Palla-
vas. Shrines in the central plain which had previously housed Jaina
and Buddhist images were reconsecrated to the m ahadeva (Siva)
alm ost always represented by a linga, called m ahesvara, and under
the custodianship o f Brahmans. Later Pallava inscriptions and
temple remains in Tondaim andalam verify that this was a non-
puranic, aniconic form of w orship.183 The conception of deity
and ritual here is one congenial to a continuing accom m odation
to Jaina and Buddhist forms (e.g. the sixty-three Jaina great
souls and Saiva nayanars and perhaps accounts for the ease with
which some o f the early Saivite and Vaishnavite saints o f the bhalcti
movement and their illustrious converts could switch allegiance
from the ‘heretical faiths’ to Vedic deities.184
182 K .M . M unshi’s ‘F orew ord’ to T he A g e o f Im perial Kanauj ( a .d . 750-1000),
v. 4 o f T he H isto ry a nd Culture o f the Indian People, ed. R.C. M ajum dar, Bharatiya
Vidya Bhavan, Bom bay, 1964.
183 K.A. N ilakanta Sastri, D evelopm ent o f R eligion in S outh India, p. 60.
184 K.V. R am an, The E arly H isto ry o f the M adras R egion, A m udha Nilayam
 The second religious process was of another kind. This was the
devotional (bh akti ) worship of the puranic gods Siva and Vishnu,
primarily a movement in the Kaveri region. Bhakti hymnists in
the Kaveri did not so much contend against the Aryan heresies of
Jainism and Buddhism — though their judgements of these sects
was often harsh; it was more the vigour of devotion to female
tutelary deities that moved them. The m ajor male deities of the
ancient religion of Tamils proved to be easily transform ed into
appropriate Vedic, male guises.185 But, certain of the great goddes­
ses and the many village goddesses (gram a devata or uramman )
o f ancient worship were m ore difficult to assimilate. Nambi
A rurar, one of the three early Saivite saints whose works form an
im portant part o f the D evaram , treats the goddess theme with
respect and appreciation but more in aesthetic, than devotional
terms and, in any case, he viewed female divinities quite clearly
subordinate to, dependent upon, and inseparable from Siva.186
In time, as has been shown, goddesses were also assimilated to the
Hinduism of the age, as the consort o f one o f the puranic gods,
and worthy of separate worship in her k d m a k o tta m . But this
did not occur until the thirteenth century as a general feature of
the religion of the m acro region.
Crowning these broad developments o f the Chola age and
contributing to the sacral significance of the Chola domain in the
Kaveri was the Siva cult of the Chola rulers. Here, the work of
Suresh is germane. In his analysis of the inscriptions and icono­
graphy of the Rajarajesvara temple at Tanjavur, Suresh has made
a strong presumptive case for the existence of the deliberate ‘pro-
pogation of canonized religion in Tam iland’ by the Chola kings.187
While he cautions that m ore detailed study of this hypothesis is
necessary, the main outlines of this argument are persuasive.
Private Ltd., M adras, 1959, pp. 186-94.
185 N ilakanta Sastri, D e ve lo p m e n t o f R elig io n in Sou th India, p. 21. Also, in the
T o lk a p p iya m . a gramm atical w ork of the late Classical period, there was already
an association o f Vedic gods with the traditional five landscapes: m u lla i- M ayon-
V ishnu: Awnn/i-Seyon-Siva; m am fam -V endam -Indra; neydal-W 'dm na. ; p a la i-'D m g n
(Balasubrahm anyam , op. cit., pp. 8-9).
186 M.A. D orai Rangaswam y, The R eligion a n d P h ilo so p h y o f the T evaram , 2
vols, University of M adras, M adras, 1958, pp. 210-23, 244.
187 Suresh, ‘Raajaaraajeesvara . . .’, pp. 449. Also, G.W. Spencer, ‘Religious
N etw orks and Royal Influence in Eleventh Century South India', Jou rn al o f the
E co n o m ic a n d S o c ia l H is to r y o f the O rien t, v. 12, pt 1 (January 1969), pp. 42-56.
 At Tanjapuri, which had never before the time of Vijayalaya
enjoyed the status o f a sacred place188 and at which the founder
o f the great Chola line constructed a modest temple to the goddess
N isum bhasudani,189 R ajaraja chose to create what he m ust have
intended to be the greatest Siva shrine in South India. ‘The tem ple’,
as N ilakanta Sastri has stated, ‘was altogether a creation o f Raja­
raja’s policy’.190 From the architectural point of view, Rajaraja
certainly succeeded; but his success was as great in having created
a symbolic shelter for and the incorporation o f the puranic Siva.
In the Brihadisvara temple at Tanjavur were established a full
display of the manifestations o f not only the puranic Siva, but
representatives o f other Vedic and puranic deities: Surya, Vishnu,
and B rahm a.191 The female principle o f deity — the preponderat­
ing one in the continuing popular religion as well as in earlier
forms of high religion192 — was reduced to a single category of
u m aparam ésvarl s, consort deities o f the m ajor representations
of Siva.193 O f the few non-canonical male deities, only one, Aiyan,
is even m entioned in the inscriptions of the tem ple.194
It is Suresh’s contention that the Cholas instituted a deliberate
policy which he speaks o f as ‘aryanization’. He means by this that
an effort was made by Chola rulers, especially Rajaraja I, not to
incorporate the deities — mostly goddesses — worshipped by
Tamils, but to displace these cults. Suresh sees this effort as a
sharp break from the previous religious history of the Tamil country
which he briefly examined.195
188 N ilakanta Sastri, The Colas, pp. 184-5.
189 Balasubrahm anyam , Early Chola A rt, p. 43.
190 The Colas, p. 185. It is interesting to notice that if any ancient sacredness
could be assigned to T anjavur it was in association with Vishnu, not Siva. Accord­
ing to a purana cited by Jagdisa Ayyar, op. cit., p. 87, the nam e Tanjavur is said
to be for the demon (rákshasa) ‘T anjan’ who was overcome by the lord Vishnu
as Nllamégaperumál.
191 Suresh, ‘R aajaraajeesvara . . pp. 438-9. Also see J.M . Som asundaram ,
T he G reat Tem ple at Tanjore, M adras, 1935.
192 See the discussion of temples in various tenth century reigns of Cholas in
B alasubrahm anyam , op. cit., pp. 43-80, 86-229. Also, see Suresh, ‘R aajaraajees­
vara . . .’, p. 441, where the prominence o f such deities as Pidaris, K adukais as
well as D urga, Jyeshtha and others are mentioned.
193 Suresh, ‘R aajaraajeesvara . . .’, p. 439.
«4 Ibid., p. 442.
195 See, especially, his section entitled, ‘Tem ples’, pp. 124 ff and 328-32, as well
as his essay, ‘Raajaraajeevara . . .’ P roceedings o f the F irst International Conference
 Evidence of Classical times indicates to Suresh, as to others,
that, notw ithstanding the presence o f Brahmans and some Vedic
ritual performances, the ancient deities of the Tamils remained
domin an t.196 These deities, the most im portant of whom was
M urugan, the m ountain god, overshadowed the Aryan deities
Siva and Vishnu to whom reference is made but the worship o f
whom, if it occurred, was subordinate to that of the Tamil gods.
Suresh sees the period from the Classical period to the advent of
the Pallavas, c a .d . 200-500, as one in which Aryan religious forms
— Buddhism, Jainism and Brahmanical cults of Siva and Vishnu
— co-existed with localized Tamil cults. This view has already
been suggested above and is quite standard in South Indian historio­
graphy. Along with others, Suresh bases this view upon the two
contem porary epics, the M an im ekalai and C ilappadikaram , though
he appears to have reservations about the degree of harmony
am ong faiths depicted in these works. Some religious strife, he
believes, m arred relationships among the ‘aryan’ faiths, and this
conflict, according to Suresh, may have deepened owing to the
political disorders o f the time. It would appear that by the onset
of Pallava rule both Buddhism and Jainism had suffered serious
setbacks, not so much extirpated by as merged with Brahmanism
which trium phed during the Pallava period. W hat the Pallavas
began after the fifth century a . d . in utilizing state power to establish
Aryan cults under the control of Brahmans, the Cholas continued
and vigorously expanded after the ninth century. Sanskrit learning
was encouraged, and this provided the mythic sources o f elaborate
Aryan cults to the detrim ent of more ancient Tamil religious cults.
Suresh’s findings regarding the influence of the Aryan gods
under the Cholas are impressive. Inform ation on 174 temples
o f the Kaveri basin is analysed. O f these, many are shrines of the
m ajor deities of the Vedic pantheon, pre-eminently Siva, but several
are dedicated to Vishnu and, somewhat unusual, four to B rahm a.197
o f T a m il S tu d ie s, K uala Lum pur, 1966, pp. 437-50.
196 The degree o f dom inance conceded varies among scholars, however; A ryan
and non-A ryan elements are seen as fused by N ilakanta Sastri (D e v elo p m en t o f
R elig io n in S o u th In d ia , pp. 32-4), whereas George H art sees Tamil elements as
predom inant at this time (‘Ancient Tamil L iterature; Its Scholarly Past and F uture’,
in B urton Stein (ed.), E ss a y s on S o u th India, The University Press o f Hawaii,
H onolulu, 1975, pp. 41-64).
197 This analysis is less than perfectly clear since there are no tables nor clear
categories for classification. The Brahm a temples include: the b ra h m a -k ü (ta m
Siva shrines existed in a variety o f forms. Some were devoted to
Siva in his principal agamic m anifestations: e.g., C handrasekara,
K ailasam udaiyar, or Karanisvaram . But m ost Siva shrines were
simply identified by the suffix -Isvaram just as m ost Vishnu shrines
were identified by the suffix -vinnagaram . Isvaram temples cons­
tituted the largest single category among the 174 temples examined
by Suresh (49 o f 174 temples).198 A substantial num ber o f these,
alm ost half, were temples named for a Chola king or queen (22
of 49); the rem ainder of these isvaram temples were named for
Siva directly or for a goddess associated with Siva. The next largest
category of shrines (28 o f 174) were nam ed after locality deities
such as Allur n akkan or M ullur nakkan. N akkan , according to
Suresh, was a way of indicating a Siva deity in the time before the
great Cholas; the term ceased to be im portant after R ajaraja I.199
Suresh noticed th at temples named after places constituted a sizable
portion of the total, 24 of 174 shrines. The term ur, in this usage,
Suresh regards as a way of designating village deities. But since
most nadus were nam ed after villages (urs) and since in some cases
the term nadu is also used, the assumption here is that such deities
were locality, rather than simply village, deities.200
The various ways of naming temples during the Chola period are
significant. Relatively few o f the temple inscriptions examined
by Suresh refer to a Vedic deity — some agamic name for Siva
or Vishnu or their consorts. The overwhelmingly large num ber of
shrines are identified either with Chola royal titles or with place
names, and it is these two categories which provide an im portant
basis for Suresh’s ‘aryanization’ interpretation. In considering the
first category, Suresh, quite properly, attaches special importance
to the prototypic royal shrine, that o f Rajarajesvaram at Tanjavur.
This central temple structure bears inscriptions of the great ruler’s
twenty-ninth year ( a .d . 1014), but refers to endowments of the
period from a .d . 1009 to 1014 to the main Vedic deities, Brihadis-
a t T anjavur and brahrmsvaram temples at Tanjavur, Perunagar, and T irukkaluk-
kunram . The first three o f these are o f the time o f Rajaraja I, the fourth is o f the
time of K ulottunga I, H C G E S I, p. 25.
'«a ibid., pp. 124-30.
199 This is supported by the findings of K arashim a and Subbarayalu (op. cit.,
p. 13) where 85 per cent o f those bearing the title o f na kka n were found in Chola-
m andalam and, o f these, around the same proportion occur in inscriptions prior
to R ajaraja I.
200 Suresh, H CG ESI, pp. 124-5, 130.
vara and Chandesvara.201 M ost images, however, are those of
Siva as D akshinam urti, Lingapurana, Atavallár, Rishabavána
and A rdhanári. Consorts of Siva are represented by the general
image, UinJtparamesvari. There are a few images representing the
Chola royal family and some of the n a ya m rs (Saivite hymnist-
saints).
W hat is suggested to Suresh by the Tanjavur temple is the
foliowing:
W hen a person takes the R aajaraajeesvaram inscriptions. . . into con­
sideration and analyses the images kept for w o rs h ip .. .the n atu ral conclu­
sion he arrives at is th a t the religion presented . . . is purely canonical
and highly aryanized and is the representative picture o f the religion
existing in the country at the tim e.202

Or, again,
R ájarájésvaram m arks the stage, when the canonical pan th eo n was
ju st nearing com pletion, when the non-canonica] gods, goddesses and
tem ples found it difficult to survive the new p an th eo n an d when im p o rtan t
ancient religious centres like A rür, A iyaru, and V enkádu began to survive
as T iruvárür, T iruvaiyáru, an d T iruvenkádu after canonizing themselves.
In short, R ájarájésvaram m arks the stage of, an d is the concrete presenta­
tion o f the canonical religion only, w ith Siva as its head. H aving con­
solidated his em pire, and with [re-] sources an d tim e a t his disposal,
R ajaraja I could m ethodically present the new C ola pantheon, to be
developed by his successors.203

Why R ajaraja I pursued this course is not explored by Suresh

except as he notes that a ‘reason for the success of the Cola pantheon
during and after R ajarájesvaram ’, was that canonical temples
like it, ‘. . . handled the economics of the country. . .’204 Such
a factor is m ade much of by C hola historians,2°5 but none appear
to come as close as Suresh to the supposition that the economic
factor may have been upperm ost in the motives o f the Chola rulers.
T hat this great temple and others o f the Chola period had im portant
economic functions is, o f course, undeniable; they were large,
201 Texts and translations in S .I.Í., v. 2.
202 ‘Raajaraajeesvaram . . op. cit., p. 439.
2»3 Suresh, H C G E S I, p. 299.
Ibid., p. 300.
205 E.g., N ilakanta Sastri, The Colas, p. 654, or even m ore extremely, in N ilakanta
Sastri’s essay, ‘The Econom y of a South Indian Temple of the Cola Period’, M alaviya
C om m em oration Volume, pp. 305-19.
complex institutions required to care for a large staff of temple
servants and pilgrims. But, it is doubtful that these institutions —
which were after all not num erous in Chola country — did more in
economic term s than to manage their own complex operations.
They cannot have been created to serve or displace economic
functions which existed prior to their creation.
Suresh’s studies and many of his arguments are about canonical
gods favoured and supported by the Chola rulers and, as he would
have it, thrust upon those whom they ruled in opposition to those
deities whom the latter had long venerated. He observes, quite
correctly, that what we are permitted to know of religion during
the Chola period comes from documents (inscriptions) from canoni­
cal temples alone.206 It is primarily through a variety of indirect
references th at other forms of religion of Chola times can be seen.
Using these indirect references of the Chola period and the few,
scattered inscriptions which date from the pre-Chola period,
Suresh discusses what he calls, ‘non-canonicaP religions, in terms
similar to W hitehead and others: deities are most often female;
male deities, when they rarely occur, are essentially tutelary deities.
Contrastively, canonical deities are male, Siva or Vishnu or con­
sorts, and they have no territorial attribute — th ey are universal.
In the pre-R ajaraja I religion of the Tamils, temple nomenclature
is seen to correspond with these differences. Two terms for shrines
occur in the pre-Chola and early Chola inscriptions: k o ii and tali.
According to Suresh, the former (from the Tamil root, k o , or
‘king’) is used for shrines of goddesses or for shrines bearing the
name o f a village or locality. Tali, from the Sanskrit, 'sthalT or
sacred place, is almost always attached to toponymic shrines of
either gods or goddesses. The suffix -isvaram predominates in
references to temples established in the post-R ajaraja I period
and signify, for Suresh, the full ‘canonization’ o f the deity. As
Suresh states it, these terms
tell the history o f the temples in a very simple way. -Kooil represents
the earliest stage, -ta li the second stage represents the introduction o f
B rahm anic elem ents converted to suit native usage, and -iisvaram the
final stage speaks o f the dom ination o f canonized tem ples.207

Suresh’s inventory and classification o f temple inscriptions of

206 Suresh, H C G E S I, pp. 328 and p a ssim ; ‘Raajaraajeesvara . . op. cit., p. 439.
2«7 H C G E S I, p. 441.
the Chola period has clarified much about the religious system of
the age. Some o f this bears upon the question o f territorial seg­
m entation, for his analysis reveals the widespread vitality o f essen­
tially localized shrines devoted to female deities — pid a ris and
k a d u k a h . These were tutelary cults enjoying the devotion and
patronage of people of the locality. We know little of the ritual of
these places, but many o f such shrines must have been under the
ritual supervision of Brahmans as well as non-Brahm an specialists.
Among the goddesses worshipped at some of these shrines were
such ‘canonized’ deities as D urga and Jyeshtha which points to
some participation of Brahm ans learned in the puranic-agamic
tradition.
It is Suresh’s view that these localized cults underwent changes
during the early Chola period. Some shrines in the vicinity of
Tanjavur, he believes, were forced to send many of their male and
female staff to serve in the great temple as a result of which these
shrines apparently ceased to function.208 At other shrines, goddes­
ses which were the object of local cult worship were transform ed
into Um aparam esvaris and thus linked to Siva. This ended their
careers as independent, place deities. Suresh sees this incorporation
o f tutelaries, rather than their complete neglect, as a necessary
compromise o f the canonicalists with the prevalence of female
deity worship; he notes that ultimately the place o f goddesses
was m aintained through their continued popularity as consorts
of the m ahadevd as exemplified in the continued popularity o f the
goddess Brihadnayaki, the consort of Brihadisvara, at Tanjavur.209
Thus, the drive to establish canonical forms of religion by R ajaraja I
and his successors was not wholly successful according to Suresh.
Female deities continued to be worshipped, possibly in ancient
ways, despite the changes in their mythic character; goddess shrines
were shortly after to be erected in the precincts o f temples devoted
to Siva and Vishnu; ancient male deities o f the Tamil pantheon —
Aiyan and M urugan (as Subrahmanyam) were also honoured
by shrines within the walls of the great temple at Tanjavur and
those elsewhere.
But, is it appropriate to consider, as Suresh does, that another
208 Suresh, ‘Raajaraajeesvara . . op. cit., p. 443: 1 . . . most probably these
tem ples died a death of starvation. .
209 ‘Raajaraajeesvara . . .', op. cit., p. 450.
purpose of R ajaraja I for imposing canonical religion was to
expunge ‘native usage’, as he puts it? Here, there is some confusion.
At one point he argues that by Pallava times ‘. . . due to some force,
the native and Aryan [Brahmanical] faiths combined and got
inseparably blended, and powerful enough . . . to fight the other
two faiths [Jainism and Buddhism].’ 210
Yet, he appears to see R ajaraja’s policies o f ‘propagating the
canonical religion’,211 and his ‘pressure o f canonization’212 directed
against ‘native usage’. The issue does not therefore appear to be
‘aryan’ or ‘canonical’ religion against non-aryan Tamil or non-
canonical religion, for, as he points out, the religion o f Tamil
country was significantly, perhaps completely ‘aryanized’ centuries
before Rajaraja. W hat appears to have been in process in the
policies o f the C hola overlords was the attem pted incorporation
o f localized cults into a religious order linked to the royal cult of
Siva. This attem pt seems to have been quite successful. Therefore,
Chola religious policies are not to be understood as arising from
economic purposes nor from the presumed zeal o f R ajaraja and his
successors, to expunge existing forms o f ritual affiliation, but from
a political design calculated to encompass independent and localized
cultic affinities within an expanding Chola hegemony. The incor­
poration o f local place and caste tutelary deities was one form of
ritual sovereignty in which the lesser gods o f local chiefs and places
honour the god o f the king and the god o f his realm.
The Tanjavur temple dedicated to Siva as Brihadisvara was
built during the late years of R ajaraja’s reign, and most o f the
records inscribed on the sides o f the m ain shrine are his; only
a few belong to the time o f R ajendra I and K ulottunga I. O f
R ajaraja’s inscriptions, all are o f the twenty-ninth year of his reign
which was possibly the last year o f his rule.213 The entire effort —
construction o f the temple, recruitm ent o f its staff, provision of
income and ritual implements, and, finally, preparation o f the
lengthy and elegant inscriptions — was carried out in a short time,
presumably under great pressure from the now old king. Each
210 H C G E S I, p. 331.
211 ‘Raajaraajeesvara . . op. cit., p. 450.
212 Ibid., p. 443.
213 N ilakanta Sastri, The Colas, p. 183. Suresh notes that an epigraph collected
in 1953 (370/1953-4) is dated in R ajaraja’s thirty-second year; (‘R aajaraajeesvara . . .’,
p. 450 n).
o f these factors contribute to the sense o f deliberateness, even
urgency, o f the undertaking.
The forced development o f the temple is evident from yet other
features. The shrine belongs to that small class o f temples which
had no pretension to ancient sacredness; it was not mentioned
am ong the hundreds o f places cherished and celebrated by Saiva
hymnists, and the chronicle o f the temple (sth ala puràna) does
not even bother to invent an ancient sanctity. Like other temples
possessing instant sanctity, the god is revealed in a dream. Accord­
ing to the B rihadisvara M ah atm ya, the chronicle o f the temple,
the Brahm an guru o f Rajaraja related a dream to the king in which
Isvara (Siva) appeared and advised that if the king wished to rid
himself o f a disease (black leprosy or krishna kustha) he should
build a temple with a great tower to shelter a linga from the N arbada
region called Brihadisvara.214 The design features o f the temple
have been extensively noted by art historians and add weight to
the argument. Unlike alm ost every other South Indian temple,
it was created and executed to a unitary design. Quite apart from
the striking effect produced by this, in contrast to the m élange
o f structures which clutter m ost temple precints, this feature
points to the builder’s purpose and determ ination. Also, this
effort to create a stunning m onum ent — and it was surely this
which inspired the towering 190 foot central shrine surmounted
by a huge granite sphere — was obviously not intended to supply
a deficiency to the Kaveri region. There were already numerous
impressive temples in the immediate vicinity, and at Chidam baram
(Tillai) at the northern edge o f the Kaveri core region was perhaps
the premier Siva temple o f the m acro region.215
A further indication of the m anner in which the temple was
pressed to completion by R ajaraja was in the means o f recruiting
temple staff. This imperious incorporation o f the servants of
other gods o f the region was a dem onstration o f dominance in the
central dom ain o f a ruler w ithout parallel in South Indian history.
Several lengthy inscriptions on the base o f the central tower of
the Rajarajesvari shrine inventory the personnel com m anded by
R ajaraja to fill the num erous and varied functions o f the place.
Three hundred celibate priests from two hundred villages; four
hundred women (presumably singers and danceis, though Nila-
2,4 Som asundaram , op. cit., p. 40; citing ch. 11 of the purâna.
215 Balasubrahm anyam , op. cit., p. 65.
kanta Sastri calls them ‘hetaerae’)216 drawn from every social
stratum and from a large num ber of goddess shrines; musicians,
male hymn singers, artists, accountants, watchmen, and armed
guards: all were com m anded to come to the temple from their
named villages in the Kaveri core. Special streets ( talicceri ) were
designated for their residence around the temple.217 Income for
all o f these servants and the various kinds o f ritual were elaborately
provided for and recorded in inscriptions.
W ealth came from two sources. First, there was the treasure
that R ajaraja had accumulated from his successful military cam ­
paigns in all parts o f the m acro region as well as Lanka. These
predatory activities and the prizes produced by them are enumerated
in the temple inscriptions; from this plunder came jewels to bedeck
the images o f the temple and gold and silver to be fashioned into
ritual implements and ornaments. The second source o f wealth
has the appearance of being tribute gathered by the king from
various territorial associations (n iyayam ) of the Kaveri core terri­
tories. These were cash endowments deposited with the m anager
o f the temple. Loan operations of these funds by the temple
manager yielded a stream o f income to support ritual.218 The
source of these cash ‘gifts’, the niyayam s, are called ‘regiments’
by N ilakanta Sastri and most other scholars, nam ed for the vala-
nadu from which they came.219 The impression conveyed in con­
sidering n iyayam as ‘regiments’ o f a centralized army, as is usually
done, is that such temple grants were a redeployment o f state funds.
However, if the term, n iyayam , is taken in its m ore inclusive m ean­
ing of a group or an association of persons having similar functions,
as other scholars have taken it, then the explicit military identifica­
tion of the groups is weakened, and the money payments may be
seen as tribute or even extortion com m anded by the king from
wealthy locality groups on the Kaveri core.220
216 The C o la s, p. 653.
217 S .I.I., v. 2. n o .69, pp. 312 ff. and no.70, pp. 328 ff.
2*s N ilakanta Sastri, The C olas, p. 654.
219 The C o la s, p. 454; K.V. Subrahm anya Aiyer, T .A .S ., v. 3, p. 74, and Suresh,
‘Raajaraajeesvara . . .’
220 Tamil Lexicon, p. 2258, ‘n iy a y a m ’; S .I .T .I ., ‘G lo s s a r y ’ a n d D C. Sircar, Indian
E p ig ra p h ic a l G lo ssa ry (Delhi: 1966), pp. 221-2 agree with the m ore inclusive de­
finition. K.V. Subrahm anya Aiyer, in T .A .S . , v. 3, p. 74, in discussing a tenth century
Succindram inscription notes that while the term n iy a y a m may often refer to soldiers,
in the particular record and in most others it has a m ore inclusive m eaning as an
 Though considerable urgency appears to have been taken to
complete the Tanjavur temple by it builder, construction was not
completed at the time o f R ajaraja’s death. Suresh draws attention
to this.221 He notes that many of the incomplete parts of the
structure had to await the attention o f the sixteenth century Naya-
kas of Tanjavur, and he wonders why R ajaraja’s successor Rajendra
did not complete his father’s work. One answer to this question
is that the latter soon began m ajor construction on his palace
and temple at the new capital of G angaikondacholapuram , but
this only raises another query: why did Rajendra abandon the
site which had served as the Chola capital from the time, of Vijaya-
laya and which had been magnificently enriched by the great temple
itself?
One answer to this question has already been proposed, that is,
strategic considerations might have dictated the locational decision.
G reater control over or defence o f the Kaveri basin from the nor­
thern interm ediate zones of Tondaim andalam and Naduvil-nadu
could have been im portant. Alternatively, or perhaps additionally,
it is possible that R ajendra’s creation of a new capital is to be seen
as an alteration o f Chola kingship o f a magnitude as significant
as that witnessed in the Pallava kingship during the time of N andi­
varm an II. This change, initiated by R ajaraja, was the depiction
o f the king as a m ore complete secular representative o f sacred
authority, the head of a royal cult. Such a suggestion is fully
consonant with the medieval Indian conception of kingship as
Hopkins long ago pointed out and as already m entioned above.222
This is the avata ra notion m etam orphosed to apply to the king as a
divine incarnation and dramatically set forth in the M ahabharata
and R am ayana. Support for the recitation o f the epics was as
popular among South Indian kings as am ong N orthern ones.
M oreover, as D.C. Sircar has pointed out, the Chalukyan kings of
Badami appeared to be claiming just such a status in their use of
the title sri-prithivT-vallabha (or s r i vallabha or simply vallabha )
along with their personal names, suggesting that they were incarna­
tions of Vishnu (i.e., husband of sri and p rith ivi ).223
association. S.1.1., v. 17, no.460 (429/1904) refers to niyayattar as a general associa­
tion.
221 Suresh, ‘Raajaraajeesvara . . pp. 449-50.
222 ‘The Divinity of Kings', J A O S , v. 50 (1931), p. 314.
223 Indian Epigraphy, p. 337.
 It is this aspect of sacral Hindu kingship which received its
m ost elaborate development among the kings of South-East Asia.
The Khmer king Jay avarm an II of the early ninth century is regarded
as the first to m ake use of the deva-raja (god-king) concept accord­
ing to D .G .E. Hall:
[Jayavarm an II] took into his service a B rahm an, Sivakaivalya, who
becam e the first priest o f the new cult which he established as the official
religion. It was that o f the D eva-raja, the G od-king, a form o f Saivism
which centred on the w orship o f the ¡inga as the king’s personality tran s­
m itted to him by Siva through the m edium o f his B rahm an chaplain. The
prosperity o f the kingdom was considered to be bound up w ith the royal
linga. Its sanctuary was the sum m it o f a tem ple m ountain . . . which
w as the centre o f the capital . . . . the axis o f the universe. . . . F rom his
tim e onw ards for several centuries it was the duty o f every K hm er king
to raise his tem ple m ountain for the preservation o f the royal linga, which
enshrined his sacred ego. T hus arose the great tem ples which were the
glory o f the A ngkor region.224

The god-king conception is one o f the most perplexing elements in

the still unclear relationship between ancient India and South-east
Asia. It is the one element o f ‘H induized’ kingship in early South­
east Asia which cannot be found fully developed in contem porary
India, though some features of the m ature institution can be identi­
fied. While it would be inappropriate to attem pt to deal conclusively
with the m atter here, it is possible to indicate a line of developmen t in
medieval South India which resulted in something close to the deva-
raja institution of South-east Asia monarchy. This line of argument
stems from the emergence of royal tom bs or funerary temples in the
m acro region. As to the probable origin o f this institution, the
following statem ent by Balasubrahm anyam appears prudent:
. . . we do not know w hether (the ‘p ersonality o r ‘devaraja’ cults) had an
earlier com m on origin or was an outflow from In d ia o r had an independent
parallel developm ent in the tw o regions (o f S outh-east A sia and India).
We have to suspend judgem ent till such time as we get m ore light on the
question, and, m eanw hile, rest satisfied w ith the know ledge th at sim ilar
cults were in existence m ore or less in the sam e period in the lands bordering
on both sides o f the Bay o f Bengal, though the devaraja cult o f S outh­
east A sia had its own peculiar developm ent, w ith individualistic features
o f its ow n.225
A m ajor element in the god-king conception of ancient Cambodia,
224 A H isto ry o f S o u th -E a st A sia, New Y ork, 1965, p. 93.
225 Balasubrahm anyam , op. cit., p. 20.
Java, and southern Annam (Champa) was the royal funerary
temple. Coedes used the terms ‘funerary temple’ and ‘m ausoleum ’
for temples in Java, Bali, and A nkor,226 and long before him, a
thirteenth century Chinese, Chan Ta-kuan, spoke of A nkor Wat
as the tom b o f the Chinese god of architecture, Lou Pan.227 As
tombs, these South-east Asian structures were believed to have
received the ashes or body of the dead king for whose honour and
in whose nam e the edifices were known. As temples, they sheltered
a sivalinga, m ost commonly called Bhadresvara with which the
king was connected.228 Bhadresvara appears to have been the
tutelary god of the kings of Cham pa229 from at least the fourth
century when the w arrior Bhadravarm an constructed a wooden
temple in honour of the sivalinga Bhadresvara as ‘the national
shrine o f C ham pa’ according to Briggs.230 Thus commenced
‘the Cham custom of forming the name of the national tutelary
deity by combining the name of the founder or reigning king with
-esvara ( = Siva), a custom which does not seem to have taken root
in Cam bodia until several centuries later",231 Briggs also follows
Coedes in attributing the first use of kingly suffix varman to this
fourth century Cham pa king.232
The royal sivalinga tutelary Bhadresvara may have come into
existence as Briggs assumed, that is,’ as a deification of the king
Bhadra signified in the suffix Isvara. However, a Bhadresvara
sivalinga existed contem poraneously in India. Religious seals
bearing this linga are seen in the fourth or fifth century a . d .233
There is no suggestion that this Indian sivalinga was connected to
a royal cult, as in Cam bodia, nor is there other evidence of such a
connection during the G upta period or later, until, perhaps, Chola
times. The identification o f kings with Siva appears thus to have
22<>Cited in Lawrence P. Briggs, The A n cien t K hm er E m pire, Philadelphia, 1957,
p. 204.
227 Ibid., p. 203; B.R. Chatterji, Indian C u ltu ra l Influence in C a m b o d ia , Calcutta,
1964, pp. 206-14.
228 Chatterji, op. cit., p. 72; other references to Bhadresvara: pp. 37, 99, 147,
172, 215, covering a period from the late ninth to the middle o f the fourteenth
centuries.
229 Briggs, op. cit., p. 15: ‘patron saint of the Chams’.
23<>Ibid., p. 25.
«1 Ibid., p. 24.
232 Loc. c it..
233 J.N. Banerjea, D evelo p m en t o f H indu Icon ography, Calcutta, 1956, pp. 182-3.
been a South-east Asian phenom enon, but the sivalinga as an
element of South-east Asian kingly tradition may have been trans­
ferred from India to South-east Asia during the G upta period.
The sivalinga at Tanjavur, Brihadisvara, is not of the classic
and well-known ones. The name does not appear among the many
listed in the standard iconographic works of T.A. G opinatha Rao
or Banerjea.234 Moreover, the form of this sivalinga is distinctive.
It is a very large, wide column, and the name brihad- probably is a
reference to its massiveness.235 Massive lingas exist at relatively
few South Indian temples, and the Brihadisvara sivalinga is found
at only two temples: R ajaraja’s temple at Tanjavur and R ajendra’s
temple at G angaikondacholapuram .236
Popularization of public linga worship and thus the appro­
priateness o f the sivalinga as an element o f a royal Siva cult in
South India was a relatively late development. It does not appear
much before the sixth century.237 Sankaracharya, the great Siva
teacher, may have contributed to its increased im portance pursuant
to the assimilation of popular sex cults into H indu orthopraxy as
P. Thom as and others have suggested.238 A nother probable
factor in the popularization o f public linga worship may have
been the prevailing use of columns or upright stones in the ancestor
worship of South India.239 Both o f these factors — sex cults and
phallus worship as well as the established use of the stone pillar
in ancestor rites — help in an understanding o f the m ajor royal
shrines at Tanjavur and Gangaikondacholapuram . Both shrines
were religious centres which emphasized canonical forms of Siva
worship, but they combined with this the powerful elements of
ancestor worship o f the massive linga and the focus upon the royal
house as cult centre through the worship o f the dead king at his
palH ppadai or sam adhi shrine. The massive sivalinga incorporated
both developments symbolically.
The Brihadisvara temples at Tanjavur and G angaikondachola-
234 E le m en ts o f H in du Ico n o g ra p h y , Paragon Book R eprint C orporation, New
Y ork, 1968; originally M adras, 1914, v. 2, pt 1; Banerjea, op. cit., pp. 454 ff.
235 M onier-W illiams, A S a n sk rit-E n g lish D ic tio n a r y , p. 735. The brih d tsa m ih ita
of V araham ihara, a sixth century w ork on images and lingas, discusses the brih at
form ; a translation of a portion of which is found in Banerjea, op. cit., pp. 579-89.
236 Banerjea, op., cit., p. 463; Jagdisa Ayyar, op. cit., pp. 63, 87.
237 S e e , ‘D ravidians’, E n cyclo p ed ia o f R elig io n a n d E th ics, v. 5, p. 12.
238 P. Thom as, E p ics, M y th s a n d L e g e n d s o f India, 8th ed., Bombay, n.d., p. 69.
239 Banerjea, op. cit., p. 463.
puram were funerary temples, that is, shrines dedicated to the dead.
Though images o f the dead were expressly proscribed in many of
the texts governing ritual practices, the installation of memorial
p ortrait statues was resorted to by the Chola family prior to R aja­
raja’s time. Balasubrahm anyam noted a Siva temple at Solapuram
(Vellore taluk, N orth Arcot) of the late ninth century which is
called a ‘Siva temple’ (jsv a ra -a la ya m ) as well as a ‘tom b’ (a tlta
g rih a m ).240 This shrine was apparently erected over some royal
personage. Later, the Chola queen Sembiyan-Mahadeviyar, wife
o f G andaradityadeva and m other o f U ttam a Chola who preceded
R ajaraja I, installed a portrait statue of her husband at the Koneri-
rajapuram temple which she erected.241 The installation of portrait
images by R ajaraja and his sister in the Tanjavur temple had
precedent therefore, and R ajaraja is known to have established
palU ppadais at M elpadi in Tondaim andalam for the Chola Arinjaya,
son of P arantaka I242 and at Tondaim an-A rrur, near Kalahasti
as a memorial to Aditya I.243 But the personal cult characteristics
at Tanjavur were deepened by the inauguration there o f recitations
and dram atic presentations based upon R ajaraja’s life. The drama
rajarajesvara-natakam and the eulogistic poem (k a vy a ) rajaraja-
vijayam were performed at times o f festivals, according to the
inscriptions o f the temple. U nfortunately, no extant versions of
these works have survived so it is impossible to say more about
them .244 M oreover, as George Spencer has pointed out,245 the
frescoes of the Rajarajesvara central shrine — notable art works
in their own right — contain a painting of Siva in the warrior pose
of Tripurantaka. Sivaram am urti commented on this depiction
of the god Siva, as conqueror, and Spencer underscores the sugges­
tive relationship to the conquering king, R ajaraja.246
The im portance o f funerary shrines of the tenth century, prior
to the time of R ajaraja I, may have stemmed as much from the need
240 Op. cit., p. 20.
241 A .R .E ., 1927, para. 13; N ilakanta Sastri, The C o la s, pp. 142, 152, 452-3.
242 A .R .E ., 1927, para. 14; Arinjisvara temple.
243 T .A .S ., v. 3, pt 1, p. I l l ; Asityesvara temple. Discussing this shrine in a
recent work on C hola temples, M athuram Bhootalingam has referred to the funerary
function as a ‘Buddhist practice’, M o v e m e n t in S to n e : A S tu d y o f S o m e C hola
T em p les, Somani Publications, 1969, p. 16.
244 N ilakanta Sastri, The C o la s, pp. 663-4.
245 Spencer, ‘Religious Netw orks and Royal Influence . . p. 50.
246 Sivaram am urti, R o ya l C o n q u ests a n d C u ltu ra l M ig ra tio n s, p. 6.
to clarify the line o f succession as to honour a dead king. The
period of G andaraditya and Arinjaya (c . a .d . 956) was one of
considerable political confusion when different kings were men­
tioned for the same period.247 A t times none were mentioned
at all. Thus, there were a series o f contem porary inscriptions from
Tondaim andalam , complete in all ways, including Saka dates,
but failing to mention the ruling king. The epigraphist who edited
these records surmised that for at least a decade there was much
uncertainty in the central plain about which o f various members
of the Chola family was ruling, or whether any was.248
No ambiguity surrounded R ajaraja’s tenure. It would therefore
seem that the purpose o f the Tanjavur temple as a funerary shrine,
a place where Chola rulers were given the status o f ‘divine beings.’249
as well as the centre o f the royal cult, was related to other efforts
by him to use canonical religious forms to extend the scope of Chola
royal authority within and beyond the Kaveri domain.
Given the obvious and effective power of Rajaraja and his
successors in their Kaveri central domain, it cannot be supposed
that the royal Siva cult was simply for the purpose o f securing the
Kaveri basin under Chola control. The royal cult, the prominence
o f brahm anical forms, and the network o f Brahmanical institutions
in the intermediate and peripheral zones o f the Chola state is best
viewed as a means by Chola rulers to affect ritual hegemony over
the num erous locality chieftains o f the m acro region. Each of
the latter exercised effective and legitimate rule (k sa tra ) over a
greater or lesser territory at the beginning of R ajaraja’s reign and
m ost continued to enjoy that power throughout the Chola period.
W hat R ajaraja and his successors perfected was a conception of
dharmic incorporation which did two things : it strengthened Chola
authority and it provided locality chiefs with stronger ritual or
symbolic bases for their own local rule. Both objectives depended
upon sharing in the brahmanically-centred canonical system
fostered by the Chola rulers and centred in Tanjavur. The particular
means adopted by Rajaraja and his successors to convert older
forms of local legitimacy of chiefs, first seen during Pallava times,
into a consistent and regular system based upon the ritual supremecy
247 N ilakanta Sastri, The Colas, pp. 142-3.
2 « A .R .E ., 1927, para. 14.
2*9 Ibid. .
o f the Chola rulers were very successful. Each local chief, his
power based upon the internal organization of his locality, came
to acquire a ritual competence related to the ritual supremecy of
the Chola rulers.
The means of articulating the ritual sovereignty em anating from
Chola kings and extending over local chieftains must be examined.
This examination is guided by the propositions advanced by
Professor Southall dealing with adm inistration and adm instrative
staffs which linked parts o f the segmentary state to the whole
through the m anipulation o f ritual elements.
In reviewing the adm inistration o f the C hola segmentary state,
it is necessary to recognize the im portance o f sacred symbols.
R ajaraja I brought two processes o f religious change together in a
royal cult with the purpose of providing an enhanced, if not a new,
dynamic element o f ideological integration to the political order
of South India. In this, he m ade astute use of individual as well as
communities of Brahmans who, operating from their unique
settlements, as part o f an overarching ideological framework,
knit together far-flung and diverse peoples o f the southern peninsula.
Here, rather than in any putative centralized and bureaucratized
state, is the genius o f R ajaraja seen and appreciated. In his con­
tinuous, at times massive, support of canonical deities, R ajaraja
raised Brahmans o f the broad territory in which his influence
prevailed to a new peak of esteem and social power. Perhaps not
since the age o f the Brahm ana texts, when the great srautci sacrifices
dom inated high religious ritual, did Brahmans enjoy the patronage
they had under the Cholas. It is true o f course that the religion
of the Chola age was no longer based upon the sacrificial rites of x
the Later Vedic age, but was now a religion o f temple worship
and personal devotion in which the intercession of Brahmans was
weakened. But, in return for this diminished sacral power, Brah- >
mans, at least in South India, attained a degree of secular authority
not previously held.
T hat secular power was not just a result o f royal design and
support. It was based upon the patronage of local chiefs, some of
ancient lineages of chiefs, but m ost of ordinary peasant stock, long
the source of locality leadership in South Indian society. During
the Pallava age, Brahmans and peasants established close inter­
dependencies based upon m utual benefits; the Chola period saw
this alliance flower to the full extent of its possibilities. Most
 nadus came to possess at least one brahm adeya where some forms
o f sastric learning and the devotional worship of puranic deities
were carried on. The n attars, as the sponsors of these activities,
probably never sought and certainly did not achieve the exalted
status o f yajam an a warriors o f the Later Vedic age. But these
peasant locality leaders did enjoy prominence, precedence, and a
degree o f interaction with Brahmanical learning and ritual —
adopting much of this culture as their own in their public and their
domestic behaviour — which separated them from the lower
status groups of local society. W ith Brahmans, these dom inant
peasant groups maintained a tradition of Tamil learning which is
solidly evidenced in the great writers o f the age who were Vellalas,
such as Sekkilar, author of the P eriyapu ran am ; with Brahmans,
also, as the Saiva Siddhanta m ovement of the thirteenth century
made clear, some among these locality peasant groups cultivated
the highest competence in Sanskritic knowledge.
The ideological framework of Chola ritual hegemony is analys-
able in terms o f concrete institutions and the essentially adm inistra­
tive processes revolving around them.
The keystone of the system of ritual hegemony was the royal
> Siva cult. Chola kings, or kingly centres, were the most obvious
source of the canonical emphasis upon which was based the culture
of the Chola segmentary state and, to a large extent, the vast
territorial sphere o f its influence. Temples dedicated to vedic
gods — primarily Siva — greatly multiplied during the eleventh
and twelfth centuries, and these institutions were focal points
in the network of Chola hegemony. M any o f these temples
sheltered gods bearing the nam e o f Chola kings, and m ost bore one
or more of the long, standardized inscriptions of the reigning
C hola king. The Brihadisvara temples o f Rajaraja at Tanjavur
and Rajendra at G angaikondacholapuram were unique in their
size and artistic qualities, fitting symbols o f a powerful, sacred
kingship. M oreover, in their canonical character and their pro­
cedures of worship, they were models for the hundreds of temples
erected in the two centuries o f Chola rule. F or the many older
temples which were enlarged and renovated, these premier temples
were also models. Consider the impact of just one architectural
element established at the G angaikondacholapuram temple. There,
an impressive symbol was created to connect the temple with
the sacred Ganga. Rajendra constructed a tank which was
called, ‘a liquid pillar of victory (jalam ayasta m b h a f and named
Cholaganga to com m em orate his ‘conquest’ of the Ganga. This
tank was created of water brought from the sacred river which had
been carried in pots from the north. These pots of G anga water
were supposedly stored tem porarily at places of rest as they were
carried to Tamil country, and at each such place, a gahgam andapam
or hall for the water, was constructed.250 At the same temple,
images were installed which were quite different from those usually
seen in South India at the time'. Brahma and Agni were depicted
as old men, bearded and pot-bellied, as they were iconographically
represented in contem porary northern India and in contrast to the
youthful forms these deities usually took in South India.251 Siva
was depicted in the form o f Lakulisa, a form most popular in
G ujarat and Kalinga, but also im portant at the im portant Siva
shrine of Tiruvorriyur in northern Tondaim andalam .252
Rajendra seems to have been drawn to the im portant Siva centre
at Tiruvorriyur. This shrine was celebrated by the nayanar
A ppar who visited there in the seventh century,253 and Sankara­
charya is said to have perform ed a miracle upon the goddess of
Tiruvorriyur similar to that which he performed with the presiding
goddess at Kanchi, Kamakshi. Hearing th at the Tiruvorriyur
goddess was fierce and that she demanded animal and human sacri­
fices, Sankara is said to have converted her to b h a k ti forms.254
While purged of some unacceptable ritual elements, Tiruvorriyur
continued as a centre of Pasupata Saivism against which there was
much contem porary inveighing.255 This temple was the centre
of an order (sam pradayam ) devoted to the Soma Siddhanta or
Pasupata school o f Saivism under a line of teachers with the title
of C haturanana Pundita. One o f the gurus o f this line, born of a
family of chiefs in K erala, rose to a high post of advisor to Rajaditya,
2?o Sivaram am urti, op. cit., pp. 22-3.
251 The distinctions between northern and southern iconographic forms are
discussed in the b rih a tsa m h ita in Banerjea, op. cit., p. 583 and in G opinatha Rao,
op. cit., v. 2, pt 1, p. 92.
252 Sivaram am urti, op. cit., pp. 23-4, 27, T he Raman, E a rly H is to r y o f the M a d ra s
R egion, p. 195.
253 Raman, The E a rly H is to r y o j the M a d ra s R egion, p. 13.
254 Ibid., p. 196, where the suggestion is m ade that the offending goddess was
destroyed, though it is still worshipped.
255 N ilakanta Sastri, The C o la s, pp. 648-9.
son of Parantaka, before becoming a member of the ascetic order.256
Rajendra m aintained close relations with this Pasupata m a t ha
as evidenced by two inscriptions of a .d . 1012 and 1043257 which
record the celebration of the birthday of the king and refer to
temple construction he sponsored at Tiruvorriyur.258 Later Cholas
also continued this association with the Pasupata teachers as well
as following Rajendra in his grants to Brahmans who recited Vedas
there.259
Close association of the Cholas with distinguished Brahmans
was made a prom inent issue in the inscriptions of the time of
R ajendra and K ulottunga I. In a num ber of records, Chola kings
are said to have conferred grants upon Brahman supplicants
while dining. According to the K arandai Plates o f Rajendra i,
the king was dining in a hall at Chidam baram when he was asked
to make a grant by a high Brahm an official.260 In another of
R ajendra’s grants this is also m entioned,261 and K ulottunga I is also
described as dining when he made a gift to Brahmans while in
K anchi.262 These references to dining are peculiar since inscriptions
are usually free of such m undane factual inform ation. Occasionally,
a king may be reported to have issued orders for a certain grant while
‘in camp' at some place, but this reference was usually in the context
of praising recent military victories o f the king. The appearance
o f the king being in a situation of intimacy with religious leaders,
brahm adeya leaders, or some Brahm an in his service is contrived.
Considering the im portance of food taking as a context for register­
ing status relations, these references cannot simply be dismissed
as a rare bit o f verisimilitude in the inscriptions.
Relations o f the Chola rulers with individual savants of high
status and accomplishment such as the C haturanana Punditas
of Tiruvorriyur and the illustrious brahm arayans in their service
were less im portant for the propagation o f a standard royal ideology
256 Ibid., p. 649, based on an inscription of a .d . 960; R am an, The E arly H isto ry
o f the M a d ra s R eg io n , p. 197.
257 Ram an, The E a rly H is to r y o f the M a d ra s R eg io n , p. 197, referring to 104/1912
and 126/1912.
258 ibid., p. 198, citing a twelfth century inscription of Rajadhiraja II.
259 Ibid., p. 228.
260 A .R .E ., 1961-2. p. 12; the grant was m ade on the request o f a brah m ardyan .
2M Ibid., no.429, dated in the third year of the king and found at U yyakondan-
Tirum alai, Tiruchirapalli taluk.
2« S .I.I., v. 17, no.207, p. 74.
than the association o f the Cholas with brahm adeyas and other
corporately organized communities of the learned. In this, as in
other things, the Cholas were following the precedent of the Palla­
vas who appeared to have maintained an interest and perhaps
some control over centres of brahmanical learning in T ondai­
m andalam .263 Such Pallava and, later, Chola policies have most
often been understood by historians to be manifestations of the
appreciation o f Indian rulers for brahmanical learning for its
own sake; it is otherwise seen as a means of demonstrating proper
regal (rajadharm a ) behaviour in accordance with their claim to
being of the solar line of Kshatriyas and thereby entitled to perform
the Vedic Kshatriya ritual of the asvam edha as allegedly done by
Rajadhiraja, R ajaraja Fs grandson.264 However, these attributed
reasons for C hola royal munificence to and association with the
learned Brahmans and non-Brahmans undervalues the importance
o f the latter in m aintaining the system of ritual hegemony o f Chola
kings.
Brahm adeyas were adm irably suited to this purpose. Beyond the
multifarious secular functions of these settlements and their
complex internal and external relationships, brahm adeyas were
the m ajor points in the spatial network of Chola ritual hegemony.
In these hundreds of settlements throughout the macro region,
Sanskrit and Sanskrit learning were preserved and transm itted;
here temples dedicated to canonical deities set the style of appro­
priateness in the worship of devotional H induism ; here, finally, was
the last asylum for cognoscenti and practitioners of the ancient Vedic
religion which had all but been displaced by temple-centred devo­
tional religion.265 And, it m ust be remembered that brahm adeyas
were not exclusively Brahm an settlements. The only social group
which was explicitly forbidden residence in a brahm adeya was
the Ilava, those who tapped the coconut and palm yra palms for
a fermentable juice (‘toddy’ in modern South India).266 A part
from Ilavas, brahm adeyas were multi-ethnic settlements where
peasants, artisans, and merchants, as well as Brahmans resided,
2« S .I .T J ., V . 3, pt 2, pp. 11-12.
264 Ibid., pp. 10-11.
265 N ilakanta Sastri, D evelopm ent o f Religion in South India, pp. 26-7 and The
Colas, p. 451.
266 See for example, 'The Tiruvalangadu Copper-Plates o f the Sixth year of
R ajendra-C hola I’, S.1.1., v. 5, pt III, p. 437 and N ilakanta Sastri, The Colas, p. 578.
judging from the specification of and dues of these groups as
residents o f the Brahman settlements. Thus, while enjoying con­
siderable and conspicuous self-government, the brahm adeya was
not isolated from its rural context. On the contrary, it was in
every way woven into the texture of rural society and ideally suited
to its political function in the system of ritual hegemony m anaged
by the Chola rulers.
B rahm adeyas of the Chola period had come to supersede the
more ancient centres of learning, gh atikas and sa la is ; the functions
of preserving and transm itting higher learning formerly associated
with the relatively few such centres had for the most part come to
be encompassed within, or better, diffused among brahm adeyas
from the tenth century onwards. Even then, however, a few
brahm adeyas continued to be distinctive for their institutions of
higher learning: Ennayiram, Tribhuvani, Tirum ukkudal, and
Tiruvaduturai. It is hardly surprising that the pluralistic settle­
ments governed by Brahm an assemblies, m ahasabhas, should
have become the premier centres ofhigher learning during the Chola
period. A part from a few well-endowed temples, there were few
other places which could have supported great learning. But,
the proliferation of these centres of Brahm an specialists of know­
ledge and ritual give an appearance o f design which should be noted.
A ttention is again drawn to R ajaraja I as one who recognized the
potential contribution which sastric and ritual scholars could
make to the m aintenance o f the Chola state.
Very early in R ajaraja’s reign, his third year,267 a military
campaign was launched into southern Kerala from which he
emerged with the first of his long list o f conquest titles — m um m adi
cola devar (the Chola ruling the three kingdoms). He also claimed
another achievement which was repeated by his successors, and
this claim remains one of the oldest historical puzzles o f the Chola
period. The claim is contained in the phrase: kandalursdlaik-
k a la m a ru tta ; it is found in the inscriptions of Rajaraja I, Rajendra
R ajadhiraja I, and Kulottunga I. A reconsideration of this ins-
criptional phrase by T.N. Subram aniam suggests the possibility
that the actions o f Rajaraja and his successors at K andalur, may
not have been as im portant militarily as ideologically.268
267 N ilakanta Sastri, The Colas, p. 189, note 5.
268 S .J.T .I.. v. 3, pt 2, pp. 1-13.
 The phrase has been read in two quite different ways. It has
come to be interpreted by historians as, ‘the destruction of the
ships at the port (or roadstead, in the navigational sense) of
Kandaiur in Venddu’.269 However, another interpretation of this
phrase given by E. Hultzsch in the first volume of South Indian
Inscriptions (1890) and by T.A. G opinatha Rao in an early volume
of the Travancore inscriptional collection.270 The latter interpreta­
tion takes the word, salai, which is in the now conventional inter­
pretation understood as ‘p o rt’ or ‘roadstead’, in its m ore common
meaning as a ‘hall’, ‘institution of learning' (c.f. patha said) or
‘charitable institution’.271 This latter meaning is more acceptable
on two grounds. First o f all, salai, glossed as ‘roadstead’ or ‘p o rt’
occurs in this phrase only within the corpus of Chola inscriptions;
secondly, the meaning adopted by Huhzsch, G opinatha Rao, and,
more recently, Subram aniam is more sound linguistically.272
Salai is the tadbhava form of the Sanskrit word said, meaning:
‘m ansion’ or ‘building’. It is this gloss of salai found in an im portant
earlier inscription, one of a . d . 865, describing the establishment of a
facility (salai) for ninety-five learned celibates (sattars) said in
the inscription to be like the salai at K andaiur.273
The establishment of a salai modelled after the one at K andaiur
was at a place called Parthvasekharapuram near modern T rivand­
rum city; K andaiur is a part o f modern T riv and rum .274 Beneficia­
ries of this grant were advanced Vedic students, those seeking
proficiency in three branches of Vedic study, and it was stipulated
as a condition of admission to the salai that each scholar have an
adequate foundation in vyakarana, m im d m s a , and paurohitya.275
2b9 N ilakanta Sastri, The Colas, pp. 169. ¡89-90 and 222: K.V. Subrahm anya
Aiyer, H istorical Sketch es o f A ncient D ekhan. v. i, pp. 246, 258: S ./.f. v. 2. pp. 35,
47, 241. 250.
270 S.1.1., v. 1, p. 65, where Hultzsch provides the reading: ‘ [the king] was pleased
to build a jewel-like hall at K andaiur; G opinatha Rao, T .A .S ., v. 1. pp. 2, lOn;
S. Desikavinayakam Pillai, Kerala S ta te Papers, Srs. 2, pp. lOOff., cited in S .I.T .I.,
v. 3, pt 2, p. 2n, and discussed in The Colas, p. 190.
271 Proposed by S.D. Pillai as cited in The Colas, p. 190, and S .I.T .I., v. 3 pt 2,
p. 2n.
272 T .A .S .. v. 1, pp. 2, 10, note 22; Monier-Willians, A Sanskrit-E nglish Dictionary,
p. 1067.
273 ‘The Huzur Office Plates’. T .A .S .. v. 1, pp. 1-14.
274 N ilakanta Sastri, The Colas, p. 190.
275 T .A .S . v. 1, p. 13; these terms refer to gram m ar, logic, and domestic ritual.
J.F. Staal has noted that Veda recitations continued to be performed in Trivandrum .
Other provisions stipulated in the plates make it clear that this
institution was a type of Vedic and Sanskrit school (pathasala)
in which the ninety-five scholars were not only to become proficient
in Vedic studies and logic, but were also to gain competence in
secular subjects. Thus, this institution was very like the gh atika
at Kanchi to which the Brahm an king-to-be, K adam ba Mayur-
sarmam, went in the fifth century in order to acquire proficiency
in pravachana, or recitation,276 and, possibly also like the Kanchi
school, the salai at Parthivasekharapuram was not simply an
institution of higher Vedic studies, but a school in which secular
training was also im portant. This is suggested by several references
in the H uzur Office Plates o f a .d . 865. Among the requirements
for appointm ent to the salai was, that the sa tta rs possess competence
in learning necessary for the affairs of the three kingdoms (trairajya
vyavahdra). The secular learning may have included martial
arts as there are references to fines for infractions of members of
the salai which included physical assault with or without weapons
or wearing weapons to meetings.277 A final provision in the grant
strengthens the secular character, or at least the less than rigorous
asceticism, o f its nominally celibate members is the explicit exclu­
sion of ‘maid servants’ or ‘concubines’ ( veUdttigal) from their
quarters.278
The Tamil language H uzur Office Plates of a . d . 865 state at the
outset that the new salai endowed by the king Sri K arunandadakkan
was established in conformity with the rules (m aryada ) of Kandalur.
The opening portion of the first plate reads in part as follows.
The fifteenth day o f the ninth year o f (the reign of) the king K aru n an d a­
d akkan (being current) on this day, having acquired gradually from the
Sabha o f M inchirai, by granting oth er lands in exchange for the (plot
of) land know n as U iakkudivilai which belonged to them : letting (loose)
an elephant round the land (for m arking its bou n d ary ); raising on it a
beautiful tem ple; setting in the tem ple (the image of) V ishnubhattaraka
and calling (the village) P arthivasekharapuram , the king Sri K arunan-
dadekkan m ade (established), in conform ity w ith the rules o f K andalur,
a salai for ninety [sic] five s a tta r s ,219

at the p alace: N am budiri Veda R ecitation (V G ra v e n h a g e , 1961). p. 34.

276 G eo rg e M oraes, The Kadamba Kula, pp. 10, 15; E .C ., v. 7, S k: 176, p. ! 13
o f tran slatio n s.
277 T .A .S ., v. 1. p. 13.
27s T .A .S ., v. 1. p. 13; S .I.T .I.. v. 3, p t 2, p. 5.
■■1<} T .A .S ., v. 1. p. 10. N o te 22 here stales th at salai ‘m eans a public in stitu tio n
The clarity with which this is stated and the obvious meaning of
salai as a learned institution in this record argues against the
continued acceptability o f the conventional interpretation about
destruction o f ships at K andaiur. One reason for the conventional
interpretation is that the phrases kan daiur-salaik-kalam arutta
occurs in Chola inscriptions am ong conquests claimed by the
kings in the eulogistic introductions of whose inscriptions it is
found.
In his review of the problem presented by the phrase and the
H uzur Office plates of Trivandrum , T.N. Subramaniam offers
an explanation which addresses the m atter in a more convincing
way. Considering the H uzur Office plates and other evidence of
the im portance o f similar institutions o f higher learning, Subra­
m aniam concludes that the idiom o f conflict and conquest connec­
ted with K andaiur was appropriate in the records of R ajaraja
and his successors since it referred to a polemical contest between
the king and learned men of the salai resulting in the defeat of the
latter and their submission and subsequent cooperation with the
Cholas.280 This is, of course, a classic kind of encounter in ancient
India in which the defeated accept the view o f the victor.281 This
idiom of conflict and conquest figures prom inently in the hagio-
graphical sectarian literature o f the medieval period in which great
saints and teachers subdue rivals in argument and earn their
o f a ch aritab le n a tu re ’ an d gives textual evidence in su p p o rt; also n oted is tadbhava
fo rm o f w o rd fo r the S an sk rit, .idla.
28« S .I .T .I ., v. 3, pt 2, p. I t .
281 K in g J a n a k a o f V ideha is an early exam ple o f such a victor. A ccording to
a Satapatha B raham ana tale, J a n a k a m et several B rahm ans, including the learned
Y ajn av alk y a. an d asked how each perfo rm ed the agnihotrd rite. H aving heard
each o f the B rah m an s, th e king said th a t he considered Y a jn av alk y a's account
the b est a n d rew ard ed the latter w ith 100 cow s, b ut only after saying th a t n o t even
th a t m o st learned o f B rah m an s could tell w h at becam e o f the tw o lib atio n s o f the
ritu al. W ith th at, Ja n a k a drove off leaving several o f the B rah m an s chag rin ed th a t
this rajanya h ad m ad e fools o f them an d resolving to ch allange th e king to a fo rm al
d isp u ta tio n . Y ajn av alk y a dissuaded the oth ers from th a t course w ith the cau tio n
th a t if th e B rah m an s w ere to win, n o n e w ould be im pressed, for th a t is w hat w as
to be ex p ected ; b u t, suppose the king w ere to win! Y ajn av alk y a th en ra n to o v er­
tak e the king an d to o b ta in from him th e co rrect answ er. A fter receiving th a t answ er,
th e B rah m an sage resolved th a t h en cefo rw ard he w ould ask the qu estio n s an d
listen respectfully to J a n a k a ’s answ ers. J a n a k a th ereafte r w as a B rah m an . The
Satapatha B rahm ana according to the T e x t o f the M adhyandina School, trans.
J. Eggeling, The Sacred B ooks o f the E ast, ed. M . M uller, v. 44, p t 5, M otilal
B an arsid ass, D elhi, 1963. p. 115.
support thereby. This sort of homage victory is what also occurred
in the political sphere as well. W arriors defeated by others, like
the Cholas, were perm itted to continue their rule, but on condition
that they recognized the supremacy o f the Cholas. There is, thus,
some reason for including K andalur, which served as a model of
learned institutions, among the conquests o f several Chola kings.
But, it is possible to speculate beyond Subram aniam and suggest
that the Cholas, like the Pallavas before them, took a strong instru­
mental interest in learned institutions such as the one at K andalur
in southern Kerala. Both dynastic lines sought to make these
institutions useful to the states they were creating. It may indeed
have been considered a ‘conquest’ to win to the services of the state
the skills and prestige of previously independent learned men and
institutions. In what m anner the learned served the state can be
little more than suggested in the present state of our knowledge,
but there are two ways which may be proposed. One is in relation
to the organizational forms and functions o f brahm adeyas o f the
Chola period; the other is in the articulation of ritual hegemony
in inscriptions and other public contexts.
Similarities between the forms followed in the establishment
of the salai at Parthivasekharapuram , according to the K andalur
rules, and the forms followed in the establishment of brahm adeyas
in the Chola period are striking. The land involved in the new
salai was acquired gradually and through exchange with the
Brahm an settlement M inchirai (the m odern village of M unjire)282
in a m anner followed in later times. The reference to the sabha
of the latter village indicates that the sa tta r of the new settlement
were different from those of the older village and the grant is
emphatic in assuring the autonom y o f the new settlement. In
later times in the Corom andel plain, Brahman-governed settlements
could and did include such specialized institutions o f learning,
such as the one at Ennayiram. Also described in the grant was the
circum am bulation o f the land o f the new settlement by an elephant ;
the establishment of a new Vishnu temple to the deity Vishnu­
bhattaraka; and the provision o f cash and kind income to the
temple and the salai from peasants who worked the lands of each.283
The grant concludes with the statem ent that the arrangements

282 Som e tw o m iles d ista n t fro m P a rth iv a p u ra m (T .A .S ., v. 1. p. 5).

283 jr.a.s..
v. l, P. 10.
 stipulated were m ade according to the com m and (ajnapti)2*4
of the headm an (kilavan ) o f Tenga-nada, the locality in which the
new settlement was established.285 The headman, one Sattan
M urugan, was called a vennlr Vellala, that is Vellala by birth.286
Considering the secular concerns of the salais of K andaiur and
Parthivasekharapuram , and the similarities of these with gh atikas
elsewhere in the Corom andel region, it is fair to conclude they may
have been a model for the kind o f learned institution which would
have had an appeal for Rajaraja and his successors.
But, there is m ore than a formal, organizational similarity
between the salais at K andaiur and Parthivasekhara and the
later brahm anical forms — including the brahm adeya — of the
Chola period. An im portant further connection has been suggested
between the m e yk k ir tti, or inscriptional preamble of Chola records
and a particular literary form which appeared to have developed
in the vicinity of Kandaiur.
The literary form referred to here is the poetic epilogue (p a tik a m ) ;
these were later additions to one o f the early Classical collections
of Tamil poetry, the P atirruppattu. W ith other poems of the Classi­
cal period, those o f the P atirru ppattu were collected in anthology
form in perhaps the ninth century.287 The title of the work means
‘the ten-fold ten’, a reference to its ten poems, each consisting o ften
odes to particular kings of the ancient Chera, modern southern
K erala.288 Courtly themes predom inate in the odes; they celebrate
the victories and the careers o f heroic kings.289 Long after the Clas­
sical period, ea ch 'o f these poems acquired an elaborate epilogue;
these have been preserved in m anuscript versions o f the poems
which also contain commentaries and explanatory notes.290
284 Monier-W illiams, A Sanskrit-E nglish D ictionary, p. 133.
285 T .A .S ., v. 1, p. 5, notes that there is a village called Tengapattanam near
Parthivasekharapuram .
286 Ibid., p. 13; Tam il L exicon, p. 3779.
287 Zvelebil, The Sm ile o f M urugan, p. 25. Perundevanar who is given credit
for the collection o f these poems and others is considered by N ilakanta Sastri to
have lived in the thirteenth century or after, not the ninth century ( H istory o f South
India, pp. I l l and 381).
288 J.M . Som asundaram Pillai, A H istory o f Tam il Literature, Annam alainagar,
1967, pp. 170-1.
28 9 t.E . G hanam urthy, A C ritical S tu d y o f the CTvakaantam ani, Coimbatore,

1966, pp. 2-3.

290 Som asundaram Pillai, A H istory o f Tam il L iterature, p. 170.
These epilogues have been com pared to the Chola m eykkT rtti
by scholars o f Tamil literature. K. Kailaspathy in his recent
T am il H eroic P o etry has written:

W hile the precise relationship o f the ‘epilogues’ [of the Patirruppattu]

to the inscriptions o f the C ola period can n o t be determ ined w ith any
certainty, it is plausible that they were fairly close to each other in po in t
o f time [i.e., the ninth century]. O bserving the rem arkable resemblance
betw een the M e y k k ir tti o f the C oja inscriptions and the ‘epilogues’ . . .
[one noted scholar o f the Tam il Classical w orks]291 suggested with some
reason th at R aja R aja I, who was the first to introduce the M eykkT rtti
as preferatory m aterial in his inscriptions, m ight have com m enced the
practice after his invasion o f the C eral [Chera] country, where he came to
know7 o f the ‘epilogues’.292

There were, thus, two features connecting that part o f ancient

Chera country which the Cholas called Venadu with im portant
aspects of Chola cu ltu re: the salai form as a possible model for the
m ature Chola brahm adeya and the colophonic epilogue as a possible
model for the inscriptional m e y k k irtti. The connections are sug­
gestive, rather than conclusive. The brahm adeya institution itself
cannot, o f course, be traced to the deep southern portion of the
peninsula, and gh atikas were known elsewhere and particularly
in the central Tamil plain near Kanchi where brah m adeya s also
arose early. Similarly, the eulogistic preamble of the inscriptions
may have developed independently o f the epilogues later added
to the Classical Tamil poetry o f the Chera country. However,
the coincidence of the assimilation of these two forms and their
detailed similarity with those of R ajaraja’s period and that king’s
invasion of southern Kerala is not easily dismissed. It is difficult
not to see these factors, coming when they did, as vitally related
to other developments which strengthened the brahmanical forms
of the Chola period and the ideological evolution of the Chola
segmentary state.
In support o f this contention is the evidence of the increasing
favour shown by the Cholas and those who modelled their public,
political behaviour on the C hola rulers to Brahm an learned men
over non-Brahmans. The learned who were supported at the
salais and gh atikas in the period before R ajaraja may or may not

291 M entioned was T.V. Sadasiva Pandarattar in his work, ila kkiya aracciyam
ka lvettu ka lu m in K . Kailasapathy, T am il H eroic P oetry, O.U .P., Oxford, 1968,
p. 223n.
292 Kailasapathy, op. cit., p. 223.
have been Brahmans. It is not clear, however, whether those who
received the greatest favour during the period o f the great Cholas
were conspicuously, if not exclusively, Brahmans. That non-
Brahm an men with a profound knowledge o f Sanskrit continued to
exist we know from the thirteenth century Saiva movement, but it is
difficult to ascertain their sponsorship. It has thus been necessary
to follow the speculation of the epigraphist, K. G. Krishnan, that
the term kanim urrüttu in Chola inscriptions, may have referred
to the support o f non-Brahm an savants by the non-Brahman
nattars. However, m ost Chola inscriptions are explicit in recording
grants to Brahmans. This is the meaning conveyed in the suffix
chaturvédim añgalam attached to m ost brahm adeya s and in countless
specific references in inscriptions. The Karandai Plates from what
is now a suburb of Tanjavur are a rather rem arkable instance of this
feature. The Brahm an recipients listed in these plates of Rajendra I
num bered 1083 in all. They were for the most part Yajurvedins
o f the A pastham ba sutra. O f the total, 775 were Brahmans bearing
the names of villages in m odern A ndhra (of whom 615 were Apas-
tham bas); the rem ainder of the grantees were Tamil Brahmans,
some following the A pastham ba and m ost of other ritual traditions.
This was a massive infusion o f Vedic scholars into the heart of
Chola country.293
Brahmans, brah m adeya s, and great temples were the nodal
points in the Chola ideological system of rajadharm a; all were
knit together by tens o f thousands o f stone and copperplate ins­
criptions. W hile unevenly distributed, the network of inscriptions
carried to every part of the large territory of the Chola segmentary
state evidence that their claim to ritual hegemony was recognized
by locality chiefs, even very remote ones.
M ost inscriptions m ust be viewed as a com bination of ritual
and control evidence. T hat is, stone and metal inscriptions were
records which intended to present a particular understanding of
relations am ong persons and institutions in the contem porary
society as well as to provide a record of changes in the control over
resources am ong persons and institutions. Inscriptions are rather
easily analysed according to these two functions o f ritual and
control.
293 K G- K rishnan, personal com m unication, 1968 o f draft m anuscript on
K arandai Plates.
 The Y ajn avalkyasm riti states that a charter or order for a gift
iddna sasana), which was the m ost common category o f inscrip­
tion,294 should include the genealogy and personal eulogy of the
donor as well as a description of the gift and date.295 The great
majority o f Chola stone inscriptions faithfully follow this form.
M ost provide a genealogical and eulogistical introduction (prasasti
in Sanskrit, m eykkT rtti in Tamil) a date given as the regnal year of
the king, and a description, usually brief, o f the gift. There are
exceptions to this, of course. A stone inscription of R ajaraja at
Tanjavur recording his gift of jewels for the deity devotes some
120 lines o f Tamil to the jewels and their ritual uses in a text contain­
ing about 190 lines 296 By and large, however, stone inscriptions
refer m ost tersely to the gifts and dwell m ore upon eulogy for the
donor. There is little question in the case o f m ost dana sasanas
of Chola times that each locality, and possibly each m ajor settle­
ment, m aintained detailed records o f the lands and personal services
which had been granted as devadana to support temple worship,
just as it is almost certain that temples m aintained records of
endowed lands, money, and animals, such as goats and cows,
whose milk was used in ritual performances. The same would
apply to Brahm an settlements. M ost gift records contain only
sketchy descriptions of the gifts which occasioned their creation.
Thus, the greatest saliency must be accorded to the ritual content
of stone inscriptions which constitute the largest class of extant
historical records possessed for the Chola period.
Copperplates (tam ra sasana) differ from stone inscriptions in
several ways. D.C. Sircar lays prim ary stress upon copperplate
inscriptions in his discussion of Indian epigraphy even though,
in South India at least, there are only a few hundred of these records
extant while there are several tens of thousands o f stone records.297
Long before stone temples existed, Indian kings issued grants
incised on copperplates.298 However, in post-Pallava times, the
num ber of stone inscriptions which have survived came greatly
294 Op. cit., pp. 400-3 and Sircar, Indian Epigraphy, p. 103.
295 S .I.T .I., v. 3, pt 2, p. 190 and Sircar, Indian Epigraphy, p. 126 ff.
296 S .I.I., v. 2, no.93, pp. 428 ff.
297 N ilakanta Sastri, Sources o f Indian H istory, p. 69.
298 Sircar, Indian Epigraphy, notes that the Buddhist pilgrim Fa Hien reported
that m onasteries possessed plates allegedly dating from the time of the Lord Buddha.
Sircar rejects this claim but acknowledges the antiquity o f the use of copperplates
(p-74).
to exceed those on copper. The view of some scholars that in every
case a stone record was engraved, there was a copperplate version
given to the grantee is difficult to accept from the sheer quantum
o f metal involved. Thus, it m ust be supposed that stone inscriptions
came to displace copper inscriptions with the emergence of stone
temples.
Still, copperplate inscriptions continued to be produced and
for the South Indian historian these records have had fundamental
importance. The following great plates o f the Chola period form
the basis for the firm chronology of the time: R ajaraja I’s ‘Larger
Leiden Plates’, 443 lines on twenty-one plates; Rajendra I’s ‘Tiru-
valangadu Plates’, 816 lines on thirty-one plates, and most specta­
cular, the latter’s ‘K arandai Plates’, 2500 lines on fifty-five large
plates (about sixteen by nine inches and weighing a total of 246
pounds without the seal-ring).299 Most of twenty-two of the
fifty-five plates of the K arandai grant are devoted to a eulogy of
R ajendra I (1041 lines), one o f the longest p ra sa stis in existence.
A substantial num ber of lines in the K arandai plates are given to
the enum eration of the 1080 Brahman grantees of the new village
which consisted of over 20,000 acres contributed by fifty-seven
different villages. The new settlement was named tribhuvana-
m ahadevi-chaturvedim ahgalam in honour of the m other of the
king. 300
Detailed attention to the beneficiaries of the grant is an im portant
characteristic of copperplate inscriptions. Whereas stone inscrip­
tions are often perfunctory in the description o f the gift and its
beneficiaries, stressing rather the eulogistic introduction, copper­
plates characteristically emphasize inform ation pertaining to the
gift and its recipients. It is little wonder, therefore, that some
scholars have presumed that stone and metal inscriptions were
com plementary in function and assumed that for every one of
the former a copperplate grant was retained by the grantee. Finally,
as copperplates were in the possession of persons who were the
299 Sircar, Indian E pigraphy, pp. 123-4. The plates are published respectively
in: E .I., v. 22, pp. 213 ff; S .I.I., v. 3, pp. 383 ff; and E .I. in press, edited by K.G.
K rishnan, D epartm ent o f Epigraphy, G overnm ent o f India. The K arandai Plates
received brief m ention by the epigraphist N . Lakshtninarayan R ao, ‘Some New
Facts about C hola H istory’, Journal o f O riental Research, M adras, v. 19 (1950),
pp. 148-57.
300 Based on the typescript generously provided by K .G . Krishna, 1968.
beneficiaries of the grants recorded, these were more susceptible
to alteration than the public stone records; they could also be
forged more easily. For some historical periods, these disadvanta­
ges diminish the usefulness of metal inscriptions for the historian;
but that is not true of the Chola period.
Copperplate grants provide the best examples of the standard
eulogistic pream ble of Chola inscriptions. That is one reason
why the great plates of the Cholas have been im portant for historical
purposes. However, as T.N. Subram aniam has noted, Chola
inscriptions differ from o th ers:
Dharmasastras prescribe th a t the nam es o f at least three generations of
the donor, i.e., great grandfather, grandfather, and father o f the donor.
This rule is scrupulously observed in the earlier grants generally and the
Sanskrit charters o f the Pallavas are the best examples for this. Som e­
times the m ilitary exploits and other eulogistic m atters regarding them
are also given. In som e grants this genealogy is not limited to three
generations, but is given for the entire ruling family from its founder.
The copperplate grants of the C halukyas, b o th the E astern and W estern
branches, give the prasasti o f their family in set verses which are added
to w ith the accession o f each new king. But in the T am il inscriptions
o f the Im perial C holas from the tim e o f R ajaraja I, and also of their
successors, the Pandyas, the prasasti o f only the reigning m onarch is
given in m etrical form . This euiogy know n as meykkirtti in Tamil is
m ore precise and historical, and consequently m ore reliable than the
exaggerated poetical com positions o f the Sanskrit form s. This meyk-
klrtti is different for each king and w ould increase in length as the reign
progressed, new victories being added as they are gained. Even gram m ati­
cal w orks on T am il prosody, know n as pattiyal, define the form o f meyk-
kirtti. T he nam e of the reigning m onarch is given at the end o f the meyk-
klrtti ,301

Because there are relatively few copperplate inscriptions, only a

few exist which record the same m eykkTrtti as are found on stone
inscriptions. One o f these is the V irarajendra record at Kanya
Kum ari, discussed above in relation to the Chola origin myth.
The same m e y k k ir tti is found, verbatim, in the C harala Plates of
the same ruler in Vengi, at the opposite end of the m acro region !302
The mobility and completeness of the copperplates suggest that
the m e yk k lrttis recorded on them might have been prepared in a
few places and disseminated to the localities around these places
S .I.T .I., v. 3, pt 2, p. 231.
302 N ilakanta Sastri, Sources o f Indian H istory, p. 69 and N. V snkataram anayya,
The E astern C alukyas o f VengJ, M adras, 1950, p. 221.
for preparation as preambles of inscriptions on temple structures.
The most likely disseminating points for these inscriptional
preambles with their genealogical, eulogistical, and mythical
content were brahmadeya^ though a temple centre like Kanchi
might also serve as in the case of the Sanskrit portion of the ‘Larger
Leiden’ plates which were prepared by five artisans at K anchi303
In these settlements were the literate men — Brahm an and non-
Brahm an — to superintend the transcription of what may originally
have come to them on more ephemeral material such as cloth and
palm leaf. At such places there were also the skilled artisans to
prepare the metal and incise the letters. And in the precincts of
these settlements were the largest num ber of temples upon whose
walls inscriptions could be placed and over the priestly custodians
of which the brahm adeya residents would have had influence.
In several of the copperplate grants — notably the ‘Larger Leiden’
plates and the K arandai plates — the procedure for preparing the
plates from other, ephemeral documents is outlined.304 Consider­
able time was necessary to complete the final record; in the case of
the K arandai plates, the grant was made in October, a .d . 1019,
and the plates were completed in August, a . d . 1 0 2 1 .j 05
If the surmise that the copperplates served as prototypes for the
m eykkT rtti portions of stone inscriptions is correct, the role of
scattered communities of literate Brahmans and non-Brahmans
in the technical m aintenance of the system of ritual hegemony of
the Chola segmentary state was very im portant. While the pro­
cedure of using copperplates as models for the ubiquitous stone
records appears to have been followed by the Chalukyas and the
Pallavas, as Subram aniam has suggested, the standardization of
the m eykkT rtti form after the time of Rajaraja represents a perfec­
tion of the procedure and the general adm inistration of ritual
sovereignty under the Cholas.
Adm inistration of ritual hegemony under the Chola segmentary
state was centralized in the ruling house. In the time of Rajaraja I
and Rajendra I, particularly, but also during the time of their
successors into the twelfth century, a deliberate and successful
effort was made to superimpose over much of the southern peninsula
E .L, v. 22, p. 222.
■W4 Three documents are referred to: tlttu, tirumugan, and aravnlai according to
the T irum ukkudal inscription (N ilakanta Sastri, The Colas, p. 468).
505 Krishnan, m anuscript cited.
a new, or at least altered, ideological framework. The following
ideological elements were salient: a royai cult o f Siva created
in which the shrine of first the Tanjavur capital and later that of
Gangaikondacholapuram enjoyed special prominence and set
the canonical style of temple worship for the entire macro region;
hundreds o f brahm adeyas were established throughout the central
and intermediate zones of the segmentary state accompanied by
impressive ceremonials at times jointly participated in by the
Chola ruling house and locally dom inant personages; and use was
made of sastric Brahmans and other learned men to sustain a
standardized symbolic system for which copper and stone inscrip­
tions were the chief instruments.
Num erous inscriptions refer to ritual specialists attached to the
royal household whose titles indicated that they were royal scribes.
Documents prepared by these writers have often been construed
as political directives through which the centralized political
control of the Cholas was effected. This attribution by historians
is hardly feasible since the only examples we have of the work of
these scribal officials are ritual documents, not bureaucratic orders.
It is thus more appropriate to see the work of such scribal specialists
as part of a system of transmission of docum ents intended to in­
corporate the hundreds of locally powerful chiefs under the sacred
umbrella of Chola kingship. The submission of the former to
ritual dominance of the Cholas enhanced their own authority over
increasingly pluralistic local societies.
Brahman and non-Brahman savants, scribes, and metal and
stone artisans were all part of the system for m aintaining and
spreading the ritual hegemony of Chola rulers. Within the centra,!
domain of the Cholas, in the intermediate zones of Tondaim anda­
lam and even the rebellious Pandya country and the more remote
peripheral zones, the most m inor, local chieftains supported and
fostered this system. For, these local personages, along with
Brahmans and other learned men, were m ajor beneficiaries of the
Chola ritual system of hegemony. By establishing brahm adeyas,
centres of sastric learning and by erecting canonical temples,
the learned and the locally dom inant brought increased stability
and prestige to their already well-established secular power as
local chiefs. Possessing no impressive credentials for exercising
dharmic rule, being themselves o f peasant stock and thus not
easily distinguished from other groups of dom inant landed folk,
or n attar , the measure of legitimacy thus gained was im portant.
It is necessary to distinguish between two groups of persons
who together constituted the m ajor agents in the system of ritual
hegemony and were m ajor beneficiaries of that system. Brahmans
and other learned persons while sometimes local men, were often
brought to a locality and provided the means to cultivate their
learned pursuits in well-supported settlements. To a life of learning
and ritual activities, they added a somewhat m inor function,
administrative specialization. For the most part, their adm inistra­
tive tasks were limited to participation in public events at brahma-
deyas and temples in which some gift was conferred upon those
like themselves and in which the presence of the reigning Cholas
was invoked to solemnize the gift. Brahman and non-Brahm an
learned persons might be involved in the preparation of the dana
sasanam , deed of gift, to be incised on stone or metal for the reci­
pients of the gift. W ith rare exceptions, ceremonies were arranged
when gifts were conferred by local persons of wealth and status.
Such grants were ostensibly for some future religious merit, but
an immediate benefit of esteem accrued to the donor.306 In all
such transactions, the presence of the Chola house was felt, for
almost every endowment carried a eulogistic reference to the royal
house. It has already been noted, too, that there were reguiar
recitations of eulogistical poems (kdvya) at the temples at Tanjavur
and elsewhere.307 Finally, there are n um erous references to the
elaborate ceremonials conducted in conjunction with grants to
Brahmans and to temples, including an impressive procession, led
by an elephant, m arking off the lands granted. Here were contexts
for a public recitation of the appropriate royal eulogy which was
then recorded as the Tamil m e y k k ir tti on the inscription comme­
morating the grant.
Others whose participation was vital and whose wealth, directly
or indirectly, made pious, public gifts possible were local men of
power. The basis of their power and much of the legitimacy of
their secular authority was membership in and leadership over
the peasant groups of their localities. In most cases, management
of a peasant locality was dem anding and complex, for such local
societies and economies could not be considered as patrim onial
30(5 According to early and medieval dharm asastras, the king acquired one-sixth
of the spiritual merit of his subjects (Lingat, op. cit., pp. 211-12).
3117 T iruppundurulti temple was another place where this was done (N ilakanta
Sasiri, The Cd/as, p. 663).
properties to be operated by a lineage or ‘brotherhood’ as in some
parts of India. These were complex, pluralistic units of social and
political organization the control of which required impressive
political skills.
To construe such locality leaders as ‘adm inistrators’ serving
the Chola segmentary state can only be true in a very limited sense.
The right of such men to rule derived fundamentally from the inter­
nal constitution o f each segmentary locality and the recognized
dominance enjoyed by those called the nattar. It is of course true
that such localized chieftainships was part of a larger political
system. It is this consistent element which was established under
the Cholas.
There were two ways in which men of standing and power in
their localities served as ‘administrative specialists’. One was in
the context of ceremonies solemnizing arrangements pertaining
to brahmadeyas and other eleemosynary activities. In these,,they
participated with Brahmans and presumably for the same purpose:
to witness prestations, to enrich the solemnity o f those occasions
in which grants of land and other values were commemorated.
At such times, they are called witness and might simply be identi­
fied by the title ‘muvendavelar with their locality specified. It
frequently occurred that both Brahmans and these locality rulers
would come from quite distant places in order to participate in
ceremonies. In the celebrated P arantaka I inscription at U ttara-
merur, in which the complex voting procedure was stipulated, it
is m entioned that one tattamlr muvendavelar was in attendance at
the meeting of the mahasabha which prom ulgated the rules.308
T attanur is a village in modern Tiruchirapalli, almost 200 miles
away from U lta ra m eru r.309 Two years later, in perhaps a . d .
922, after the prom ulgation o f the procedures m entioned above,
further provisions were made in another inscription. This rambai-
nadu in Cholam andalam , almost the same great distance from
U ttaram erur in T ondaim andalam .510 In the ‘Larger Leiden’
plates o f R ajaraja I, about forty persons are m entioned as being
involved in the execution of the grant at its various stages from
being an oral order to the engraving of the final arrangements.
Of these forty, five bore the title of m uvendavelar and another
-,os Subrahm anya Aiyer, Sketch es o f the A ncient D ekhan, v. 2. pp. 262-4.
309 Cf. the M eiappaluvur inscription o f the tenth century, A .R .E ., 1924, no.367.
31o Subrahm anya Aiyer, S ketch es o f the A ncient D ekhan, v. 2, p. 279.
fifteen persons are identifiable as non-Brahman notables, the
remainder being Brahm ans according to their designations.311
Most of the Brahman and non-Brahman penicipants came from
villages within a fifty mile radius of the newly established settlement
of Anaimangalam (N agapattanam ). The titles of these forty
persons from the various valanadus around kshatriyasikham ani-
vajandu, in which the new village was located, indicate that their
functions were scribal.312 Designations such as dlai-eludum , man-
dira-olai-nayakan, and karu m am arayam imply scribal functions.313
The same designations and functions were specified in the Tiruva-
langadu plates of Rajendra I which was almost contem porary with
the ‘Larger Leiden' grant o f R ajaraja I,314 the p ra sa stis o f both
being drawn by the same person, A nantanarayana, R ajendra’s
court poet.315 Thus, together with Brahmans, locality leaders
from nearby, and at times from distant places, participated jointly,
and it would appear, in interchangeable roles, in the ceremonies
which conspicuously emphasized protection over royal and other
centres which played a crucial function in the articulation of Chola
ritual sovereignty.
The other, and truer, sense in which local, non-Brahm an per­
sonages performed as ‘administrative specialists’ under the Chola
segmentary state was within their localities. It was here that their
proprietory k sa tra was exercised. They were the principal executors
and trustees of grants made for the maintenance of brahmanical
institutions which gave structure to the ritual network of Chola
hegemony and their own sub-regal authority. This is again verified
in great Chola plates. In the Tiruvalangadu Plates of Rajendra
I, one Avaneri, ‘of the fourth caste (chaturthdnvaya), pure on both
sides’ is m ade responsible for carrying out the elaborate provisions
of the grant which included leading the circum am bulation of the
new settlement by a female elephant (karini-bhram ana ) ; further,
311 E .I.. v. 22, pp. 235-6, ‘List ‘A ’.
312 Ibid.; these include; N iitavinoda, Uyyakondar, and Arum olideva valanadus
and the more distant one o f kar-nadu in rajendrasimhavalanadu.
313 Ibid.; the editor o f this inscription, Subrahm anya Aiyet, translates there
terms as: ‘Superintendent o f Royal W rits', ‘Secretary', and ‘various tax registrars’.
314 The ‘Larger Leiden’ plates of R ajaraja I being posthum ous (£ ./., v. 22,
p. 222). -
115 Ibid., and ‘The Tiruvalangadu Copper-Plates of the Sixth Y ear of Rajendra-
Chola F, v. 3, p t III, no.205, pp. 383 ff; and especially the Tamil portion o f
the grant, pp. 428 ff.
it is stated that, having received the written order from various
scribai functionaries, \ . . we the (chief) men of the district [nadu]
went out (respectfully), received and placed (it) on (our) heads
and accom panying the female elephant, walked around the ham ­
lets.’ 316 Or, according to the ‘Larger Leiden’ grant:
. . . we the nattom (i.e., the assembly o f the district), seeing it (i.e., the
order) being brought, respectfully advanced (tow ards) received and
carried (it) on our heads and accom panying the female elephant, walked
round the ham lets, set u p stones and m ilk bush [as b o u n d ary m arkers]
and drew up and gave the deed o f g ift.3’7

These and other grants of the time also include attesting signatures
to the elaborate arrangements described in the grant records.
Most signatories were residents of villages neighbouring upon the
newly established Brahman settlement and contributors to the
lands possessed by the new village. These witnesses are said to have
accompanied the circum am bulation procession and to have
assisted in the final drawing of the document. In the ‘Larger
Leiden’ plates from the end of R ajaraja I’s reign, there were twenty-
seven such attesting statements and signatures contained in the
final part of the Tamil portion of the plates.318 In later Chola
times, the internal adm inistration, or exercise of ksa tra , by these
locality leaders became even more im portant when, as the p eriya -
n a tta r , they spoke of themselves in the dharmic terms of protectors
and managers o f temples.

P yra m id a l segm en tation an d Chola p o litic a l culture.

According to theories of the state in medieval India, whether derived

from puranic or sastric sources, society was prior to polity; social
collectivities defined by kinship (ku la), by shared and cooperative
interests (sreni ), and by common residence (gana ) are considered
competent and responsible for the governance of their affairs, and
the state, or more particularly kingship, was an institution whose
function it was to maintain a social order comprised of such col­
lectivities.319 Kingly maintenance o f the social order was achieved
316 ‘T iruvalangadu Plates', S .U ., v. 3, pt 3, p. 389 and verse 132, p. 426; qu o ta­
tion, line 143, p. 430.
311 E .I.. v. 22, p. 259.
318 E.J., v. 22, ‘List B’, pp. 237-8 and lines 207-330, pp. 263-6; also ‘Tiruvalan­
gadu Plates’, S .I .L , v. 3, pt 3 pp. 437-9.
31<) K ane, op, cit., v. 3, pp. 280-1; Lingat, op. cit., pp. 246-8.
through the moral operation of royal justice, that is through a d hoc
adjudication of conflicting claims by groups in society, thereby
creating or protecting entitlements of groups through the dharmic
protection of danda.
The pyramidally segmented society of medieval South India
had developed a massing capability well before Chola times;
nadus possessed the attributes of janapadas characterized by cor­
porate control over a tract of land by agricultural groups sharing
a common ethnic identity, and artisan and trade groups maintained
wide-ranging networks. These were autochthonous structures
of social groupings in both the material and moral senses; they
existed before the Cholas and persisted after them. Territorial
chieftainships and primitive kingships facilitated some of the
horizontal and vertical massing which is referred to in the Classical
poetry of the Tamils. However, this political element was thin by
com parison with the elaborate kingship of the Pallavas and the
Cholas. K ailasapathy’s analysis of themes in the puram poems
discloses a fixation on m artial prowess and w arrior bravery by
such political personages.320 It is perhaps only after the time
o f N andivarm an II, Pallavamalla, that the chief, acting as agent
for massing resources and the incorporative strategy of the Pallava
kings, becomes significant.
Facilitating this massing and incorporative process o f these
South Indian kingships were inscriptions. Chola stone and copper­
plate inscriptions are normative documents. They present a
particular moral conception of society and polity, and are essential
expressions of both. In idiom, inscriptions are puranic and sastric.
Kings ‘capture’ and ‘m arry’ distinctly powerful agents in ways that
at once attest their royal might and magnify it. Rajendra I, accord­
ing to his Tiruvalangadu Plates, took the river Ganga and held it in
the Cholaganga tank constructed at his capital ‘to sanctify his own
country’ ; the name of his capital itself, G angaikondacholapuram ,
commemorates this puranic conquest.321 Rajaraja, Rajendra, and
other Cholas, again in puranic style, spoke of being m arried to the
goddesses o f fortune, victory, and the earth.322 Sastric ordinances
are heeded in the form followed in inscriptions, and the Cholas
refer to themselves as upholders of the dharm a codified by M anu.
320 Kailasapathy, op. cit., pp. 24-6.
321 S .I.I.. v. 3, pt 1.
322 Ibid., v. 3, pt 1, n o .4, pp. 6-8 and S .I.I., v. 1, no.67, p. 959.
 Beyond these expressive features of inscriptions, these documents
have an im portant sacrificial aspect; they record the processes
and the results o f sacred transform ations. The process, or set of
linked actions, which resulted in inscriptions of the Chola kings
always involved a massing or pooling of hum an and material
resources, often on a grand scale. These resources were drawn
from diverse, independent, and at times, locally opposed elements
in Chola society (e.g. from both idangai and valangai groups)
for the general welfare of all in that society. Typically, this was a
gift to a god or to Brahmans, and the presentation process was
homologous to, and was very likely considered equivalent to, a
sacrifice ( yajh a ).
The conception of dana, or gift, as ya jh a , or sacrifice, appears in
early dharam asastras. Dana is seen to convey the same range of
benefits to the donor as the sacrifice does to the yajam ana.
M anu and others equate the two as the principal aspects of religious
life in the dvapara-yuga and kali-yuga respectively.324 One of the
dharm asastra elements yoking these conceptions of dana and
yajha is ista: what is sacrificed; what is given. A nother linking
element is pratigraha. This is the idea elaborated by M edhatithi,
a m ajor medieval com m entator on M anu, to the effect that
some gifts are distinguished (as pratigrah a ) in being ‘offered
with a view to some transcendental result, and received with
due m an tras’.325
In Chola inscriptions, kings are often, mistakenly, assumed to
have been the actual yajam ana agents, or donors. This is quite
incorrect. However, the invocation of royal protection for a gift
transform ed that which was given to the status of pratigraha and
the donor to the status of a yajam ana. Inscriptions of the early
Chola period of the ninth and tenth centuries even go as far as to
call forth royal protection by the depersonalized titles of ra ja k esa ri
and p a ra k e sa ri ; presumably, this was as sanctifying as the elaborate­
ly eulogistic royal invocations of R ajaraja and his successors. And,
since the invocation of the reigning king (even abstractly by the
ra ja k esari title) conferred upon the ruler some part of the merit

323 Kane, op. cit., v. 2, pt 2, pp. 838-9.

324 Ibid., p. 837.
325 M a n u sm rti; The Law s o f M anu, with the Bhasya o f M edhatithi, trans. by
G anganatha Jha, University of C alcutta Press, Calcutta, 1922, v. 2, pt 2, 'Discourse
IV ’, p. 304. Also, Kane, op. cit., v. 2, pt 2, p. 842.
gained by the gift, every such gift increased the protective power
of the king.
In all of this, Chola kings may be seen as having perfected means
of incorporation under their royal aegis which the Pallavas had
begun perhaps a century or more before.326 The means used by
Chola kings were not essentially different from those of the Pallavas,
but they were pressed with greater energy and urgency. Thus,
the Cholas followed the Pallavas in the construction of temples
to house Vedic gods and in the lavish support of Brahmans, but with
an intensity not before seen. R ajaraja’s remote ancestor Aditya I
( a .d . 870-907) is said in the Anbil Plates of Sundara Chola ( a . d .
956-73) to have lined the banks of the Kaveri with Siva temples,
and m odern research confirms that the claim was a substantial
one.327 D uring the reign o f U ttam a Chola ( a . d . 969-85) and the
early part of R ajaraja’s reign, quantity may be said to have yielded
to quality as the Rajarajesvara temple in Tanjavur commanded
most of the King’s resources. And, with what remarkable effect!
It was m oreover during the time of R ajaraja that the rate at which
brahm adeyas were established attained a peak.
To these means of accentuating the dharmic character of Chola
kingship were added two others. One was the sacralization of the
royal lineage through the construction of funerary tomb-temples,
paU ippadai, by Parantaka I at Tondaim anad, by R ajaraja at
Melpadi, and by Rajendra at R am anathakoyil.328 The other was
the effort by R ajaraja particularly — though by Chola rulers
generally — to appropriate for themselves a special relationship
to the m ahadeva Siva as his chief devotees, if not as incarnations.
The Pallavas also introduced the elaborate technical process
of which inscriptions were the material result. As in other aspects
of dharmic kingship, the Cholas perfected this richly ceremonial
and expressive medium. Chola inscriptions do not portray a
unified social and political order so much as they constitute an
element which fashioned or created that order. If one considers
326 See the brief discussion by B alasubrahm anyam , E arly Chola A rt. ch. 3, ‘Temple-
Building in the Pallava Age’.
327 Discussed by Douglas Barrett. E arly Cola A rchitecture and Sculpture: 866-
1014 A .D .. Faber & Faber Ltd., 1974, p. 49.
328 N ilakanta Sastri, The Colas, p. 453. Balasubrahm anyam , op. cit., p. 18, notes
an earlier, he believes earliest, example of such sepulchral tom b-tem ples constructed
by a G anga chief at Solapuram (Vellore taluk. N orth Arcot) in a . d . 8 8 6 .
the 3500 or so published Chola epigraphs,329 a great variety of
discrete and disparate historical features of society can be cata­
logued. The inscriptional corpus depicts a society from the tenth
to the thirteenth century which is very diverse: hundreds of nadus
of differing internal structure; hundreds of chiefs, great and small;
thousands of gods each of which was the locus of cult affiliations
and practices; hundreds of term s for local taxes and dues; and
other particularistic traits. But, transcending this diversity is
a unified conception of a Chola society and a Chola kingdom,
and that conception is not the technical, bureaucratic one inferred
by most Chola historians. This is a false idiom. It is rather the
idiom of a dharmic universe realized through the sacral kingship
of the Cholas.

529 This estimate is given by N. K arashim a and Y. Subbarayalu, op. cit.

 CH A PTER VIII

Vijayanagara State and Society

Introduction.

The purpose of this final chapter is threefold. One is to examine

the political system of the South Indian macro region during the
Vijayanagara period as a segmentary state. This further elabora­
tion o f the pre-m odern, South Indian political system is intended
to provide both theoretical and empirical support for the idea
of the segmentary state described in detail in the previous chapter.
A second purpose is to consider new elements of South Indian
society, especially Tamil society, in the centuries after the ‘imperial
Cholas’. Im portant among these new elements in the macro
region were the expansion of Telugu and Muslim control over a
substantial portion o f the southern peninsula in the early post-
C hola period and the introduction o f Europeans from the middle
of the period to its end, when the English East India Company
established its rule in the late eighteenth century. Thirdly, this
chapter intends to provide a cogent background for the considera­
tion of changes in South Indian society during the nineteenth
century, a concern which inspired this investigation of early South
Indian society by the author many years ago and in which he con­
tinues to have a lively research interest.
Treating the several centuries after the Chola period in a single
chapter is to condense a rich and varied historical period of South
Indian history. This will appear cavalier to many. In justification,
several related considerations were determinative. The present
m onograph is already long by m odem publishing standards and
to treat the post-Chola period fully, as fully as it may be thought
to deserve, would require a volume once again as large as the present
one. While it would be absurd to argue that this extended his­
torical period is undeserving of that fullness of attention which
is eschewed here, it must also be pointed out that the period
cannot be said to have been neglected since the time that R obert
Sewell spoke o f ‘Vijayanagar, A Forgotten Em pire!’ The work
o f Sewell himself, of Krishnaswami Ayyangar, N ilakanta Sastri,
V enkataram anayya, Saletore and others have made o f Vijayanagara
history one o f the better researched fields — qualitatively and
quantitatively speaking — o f any in Indian history. Hence, the
urgent need appears less to add yet more pieces o f inscriptional
and literary evidence to illuminate the period than to propose
some perhaps new and different conceptions of that period and
its rich evidence with a view to stimulating and perhaps redirecting
further research. Such new, and it is hoped stimulating, con­
ceptions as are presented below spring from a variety o f sources,
including the possibly idiosyncratic interests of the author. How­
ever, it is supposed that by viewing the Vijayanagara period against
a detailed consideration o f the society and culture o f the Chola
period in the way that has been done here, a potentially useful
basis for precisely that kind o f revisionistic focus which appears
to be called for at this time will be laid.
In quite summary form, the following comparisons and contrasts
between the m ature Chola and Vijayanagara societies may be
outlined. To begin with, there was fundamental continuity between,
say, the eleventh and seventeenth centuries with respect to several
im portant aspects of society and culture within the peninsular
m acro region. The political system continued to be one which is
called ‘segmentary’ or ‘pyram idal’. Vijayanagara kingship, like
Chola kingship, was ‘ritual’ in respect to rule over peoples and
territories o f the macro region beyond the ‘hom e’ territories of
each kingship. Corresponding to the ritual centre of kingship
in both cases — and again outside of the riverine core regions of
the two kingships, i.e. the Kaveri o f the Cholas and the Tunga-
bhadra of Vijayanagara — locality units of the political system
were not merely self-governing — linked to imperial centres
neither by resource flows nor com m and — but were reduced
images of the two centres. This remains true even as it is recognized
that the conception, size, and complexity of ‘locality’ may be seen
to have changed considerably over the two periods.
Another related element of continuity between the two periods
of South Indian history is the central place o f religious institutions.
M ore is intended in this observation than that South Indians were
and continue to be religious. Religious institutions, particularly
the brahm anical institutions of the Chola brahm adeyas and the
temples o f puranic gods later, were core institutions in every sense.
Not only did these institutions directly fulfill the narrowly construed
religious needs o f the most respectable (sa tv ik ) groups in South
Indian society, but served as models for and models o f an appro­
priate order for even ordinary folk by the Vijayanagara period.
And, as in the earlier period, religious institutions of the latter
period continued to have m ajor economic importance, to be
centres o f learning, to be focuses o f local pride and identity, as well
as being theatres o f every form of artistic expression.
Finally, and notwithstanding the sturdy independence of locality
society and culture, both ages witnessed and reflected the importance
o f migration and conquest. The peripatetic ways o f many in
South Indian society are not recognized by most historians. In
both periods, however, new peoples were continuously integ­
rated into established locality societies. M ost conspicuously this
involved military conquerors such as Tamil peasant soldiery under
the Cholas in the K aranatak and Andhra extensions of the Tamil
plain or Telugu peasant warrior intrusions into Tamil country,
especially after a . d . 1450; it also involved the movement of Brahman
specialists (e.g. the thousand, learned Apastambhikas from A ndhra
to the Kaveri basin recorded in the. K arandai Plates to, Rajendra
Chola I) or ritual specialists from one to another temple to install
and maintain a particular b h a k ti ritual form. F or the movements
of persons and groups involved in temple ritual, inscriptions
constitute a record of considerable specificity.
Less well-recorded, but nevertheless evident, were movements
o f lower groups and their progressive inclusion in the expanding
agrarian and trade systems o f the macro region. The continued
process of integration and alliance represented by the dual division
of right and left castes often convey an impression of tension and
conflict attending the movement of persons and groups in spatial
as well as status terms. The appearance of the divisions as conflict
groups is exaggerated, however. Conflict was but one of the causes
for converting the divisions from potential or latent groupings
at any time and place into actual groupings; evidence on the
divisions from the colonial city of M adras also contributed to this
characterization o f the divisions as primarily conflict-ridden,
factious groups. During the Vijayanagara period, somewhat less
of the factious quality of the divisions is evident.
Given the degree of continuity outlined above, the notion that
there might have been significant discontinuity may appear difficult
 to sustain. Certainly the historiography on the question of historical
discontinuity is mixed and controversial. Scholars whose research
has focused on K arnataka and A ndhra during the Vijayanagara
period (e.g. Saletore and Venkataramanayya) have emphasized
continuity and the preservation o f ancient usage while those who
have worked on Tamil country (e.g. N ilakanta Sastri and Krishna-
swami Pillai) have drawn attention to basic changes incident upon
the extension o f Vijayanagara rule over the Tamil region. But,
whatever the historiographical disposition, all could presumably
agree that the following changes occured and were im portant.
The m artial character o f the Vijayanagara state and the seemingly
incessant warfare of the period from the middle of the fourteenth
to the late seventeenth century is implicitly contrasted to the
apparent civil order o f Chola times. Here o f course, the threat
o f Islam to H indu society is given im portance both as explaining
the violence o f the age and the adaptations o f H indu society to
that violence. Indeed, the Islamic threat is given a decidedly
m odern complexion as the ideological factor which brought the
Vijayanagara state into existence and, on the one hand, explained
that state’s steadfast maintenance o f custom , or, on the other hand,
for some scholars, justified departures from custom in order to
defend Hinduism and varnasram adharm a. F or scholars who give
m ajor attention to this ideological factor, the Vijayanagara state,
in its essence, would be considered very different from the Chola
state, particularly in its ‘feudal’ elements.
While not all scholars o f Vijayanagara speak o f ‘feudalism’,
all give prominence to a feature o f political organization which,
for those who do use the concept, is decisive. That is, the nayankara
system. W hether seen as part of a feudal estate similar in some
essential m anner to those of medieval Europe or .simply as a new
political feature o f South Indian history, the nay a k a of Vijaya­
nagara times is an im portant and dicontinuous fact of the age.
W arriors who used the title o f nay a ka or am aran aya k a , or to whom
that title is affixed, cannot be defined easily in terms of particular
office, ethnic identity, privileges and duties. Yet, it can be said
that during the Vijayanagara period there came into existence, or
( at least into sharp focus, a level of supralocal chieftainship which
appears to be different from anything which existed before. As
locally powerful personages, these chiefs may not have been different
in any particular from supralocal personages of the p a st: they were
first and foremost warriors with armed followings; they joined
with great kings in defensive and predatory warfare; they drew
resources from the territories under their rule and shared little
or none of these resources on a regular basis with those o f super­
ordinate authority ; they distributed, or, more correctly, redistribu­
ted the resources they com m anded to a variety o f religious and
social purposes. In these activities, there was perhaps not much
new or different. But, in the degree of power of these chieftainships,
in the m agnitude of local resources commanded and redistributed,
in their independence from local social and cultural constraints,
their ability to intrude into local society, and in their persistent
independence from and occasional opposition to superordinate
authority, this political category is unprecedented. These nayaka
warriors constitute a level o f power and authority not before seen
in South India.
A second, technological, factor o f discontinuity may be seen
to explain the n ayan kara system. N a ya k a power rested upon the
substantially increased military capability afforded by firearms,
fortifications, and superior cavalry m ounts. The ability to
com m and these improved military means was limited to the rich
and powerful o f whom the Vijayanagara kings were the first. It
was upon these new means o f warfare that the spectacular military
success of the Rayas was based, and their experience with and use
of Muslim soldiers from the fifteenth century on made possible
much o f this superiority. Telugu followers o f the Rayas employed
the same means to establish themselves as the independent nayan-
karas o f the age, and gradually these superior military elements
came into the possession o f other, non-Telugu, warriors as well.
The latter added to the complexity of the supralocal military
élite designated by the term nayankara.
Dissemination o f this assemblage of militarily superior means
not only made n ayakas the m ost powerful local and supralocal
chiefs South India had known, but almost assured conflict between
them and the Raya overlords o f the macro region. Such conflict
was an early feature of Vijayanagara rule in South India and is
ultimately responsible for the decline of the state in the early part
o f the seventeenth century when civil wars rent the southern
peninsula. Conflict between the Vijayanagara kings and the
powerful nayakas is another element o f discontinuity which is
significant. D uring the late Chola period — the thirteenth century
 — similar conflict between kings and powerful chiefs can be seen,
but it was neither as enduring nor as much a part o f the system of
political relationships o f that time as it was to be from almost the
beginning of the Vijayanagara period. The level o f these internal
conflicts heralded the end o f the Chola overlordship in South India;
it was a central part of the Vijayanagara overlordship from its onset
in the fourteenth century; and just the fact o f having to cope
with it, provides a final element of discontinuity with the Chola
state. This was the significant political role taken by Brahmans
during the Vijayanagara age.
It appears to have been a deliberate policy of the Rayas — cer­
tainly it was a policy o f Krishnaraya who was perhaps the greatest
o f all — that Brahm ans had been given m ajor political roles. In his
time, and to some extent before, Brahmans were the comm andants
/ o f m ajor fortresses and were considered territorial ‘governors’
by contem poraries, South Indian as well as foreign. A part from
the maintenance of royal forts in all parts of the macro region,
there appears to have been only one governing task for which
Brahmans were responsible: that is, checking the fissiparous designs
of the nayakas. There are few ,reigns in the dynasties of Vijaya­
nagara in which it is not possible to identify Brahmans as major
agents of Raya rule, and the central if not sole requirement for these
prestigious persons in their secular, political functions was to
defuse the explosive potentialities inherent in the Vijayanagara
segmentary state with its powerful, interm ediary level o f nayankaras.

V ijayanagara P o litica l S ystem

There are many im portant questions about Vijayanagara political

and social m atters which remain vexingly unclear after more than
a half century of scholarly attention.1 As com pared to the Chola
period o f South Indian history, a greater num ber o f historians have
contributed to Vijayanagara studies; there is nothing like the
dominance of a single view in the study of this latter period — as
that o f N ilakanta Sastri for the Chola period. One consequence
of this greater variety is a richer historiography, but one which is
m ore difficult to evaluate.
1 S. Krishnasw am i Aiyangar, A n cien t India and S o u th Indian H istory and Culture
II, Oriental Book Agency, Poona, 1941; first essay published in 1920; R. Sewell,
A F orgotten E m pire, L ondon, 1924; Rev. H enry Heras, The A ravidu D ynasty o f
Vijayanagara, M adras, 1927.
 Almost from the outset there has been agreement among Vijaya­
nagara students with respect to certain im portant interpretations.
These have tended to remain acceptable to more recent scholars.
The first of these durable interpretations pertains to the success
of the Vijayanagara state in limiting the expansion of Deccani
Muslim power. The fourteenth and fifteenth century Vijayanagara
state did stabilize the frontier between the Bahmani sultanate, its
successors, and what was to remain a dom inantly H indu social
and political order south of the Kistna-Tungabhadra. This leads
to what appears to be the second broadly agreed upon view of the
Vijayanagara state: that it created the conditions for a defence
of H indu culture and institutions. This defensive role is also
seen to have been self-conscious and ideological. Finally, and
again a consequence o f the encounter with the powerful Muslim
slates o f the Deccan, the Vijayanagara state is seen as an essentially
military state, in Nilakanta Sastri’s words: ‘the nearest approach
to a war-stale ever made by a Hindu kingdom’.2
The social and economic implications of these generally agreed
upon views o f political history stemming essentially from the
Muslim presence in the peninsula have only been partially explored.
Am ong the potentially relevant issues which might be queried are
the following. G ranted the support which powerful Vijayanagara
personages — from the kings of the several Vijayanagara dynasties
down to m inor, local notables — gave to H indu religious institu­
tions, what other differences between a general Hindu overlordship
and a Muslim one can be dem onstrated, or suggested? That
is, in terms m ore general than the support o f Hindu temples and
Hinduism generally, what difference did it make that this was a
‘H indu’ state? M ore fundamental, perhaps, what was the relation­
ship between the variety of religious institutions and other social,
political, and economic institutions o f medieval South India?
O r again, what were the conséquences of the ‘warlike’ character
of the Vijayanagara state within a general institutional and ideolo­
gical context? To those familiar with studies o f the period, these
queries will seem to have been answered in the num erous m ono­
graphs and essays dealing with the Vijayanagara state. And so
they have, but usually in terms so general and so simplistically
based upon the overwhelming saliency o f Islamic power, that
2 H istory o f S outh India, p. 295.
great reliance cannot be placed upon these answers. This is especial­
ly true when note is taken of considerable controversy among
scholars on the most im portant questions.
Some of the questions upon which there remain deep differences
am ong scholars have been noticed in the m ore recently published
research on Vijayanagara. V asundara Filliozat reminds us of
continuing differences among scholars about the origin of the
first or Sangama dynasty o f Vijayanagara.3 Older views about the
origin o f the dynasty as tribal peoples, K urum ba or K adam ba,4
have given way to two other m ain theories. N ilakanta Sastri
and N. V enkataram anayya have taken the position that the founders
o f the Vijayanagara state were Telugus; others, including R.
Narasim hachar, B.A. Saletore, H.Heras, and S. Krishnaswami
Ayyangar, have identified the founders as K arnataka warriors.
A recent study o f the second, or Saluva dynasty sees its founders
as both, that is, as having originally moved from K arnataka to
A ndhra country.5 If Filliozat has decided for himself which of the
theories is best supported by the inscriptional evidence he very
closely examines in his work, his conclusion is not clear. Yet,
the question is of some im portance if, as it is usually assumed, the
state structure created by this first dynasty was different from
other state forms which had existed in South India (if in nothing
else, its m artial emphasis).
There are, in fact, two m ajor questions about the nature of the
Vijayanagara state which may be considered linked to the probable
origin of the founding dynasty, but possessing intrinsic importance
separate from the question of origins. One of these has been widely
discussed as the question o f whether ‘local institutions’ continued
to flourish during the Vijayanagara period as they had, especially
in Tamil country, prior to that time. And, if they did not, was this
a deliberate policy of the new rulers of the Vijayanagara state?
According to Saletore, whose principal work centred upon
K arnataka during the Vijayanagara period, the integrity o f local
institutions, or as he puts it, 1pü rvam ariyâde (ancient constitutional

3 L 'E pigraphie de Vijayanagar du débout à 1377; P ublications de l'école Français

d e ’E xtrèm e-O rien t, v. 91, Paris, 1973, pp. ix-xv.
4 Sewell, op. cit., p. 33.
5 P. Sree Ram a Sarm a, ‘Sâluva Dynasty o f V ijayanagar’, unpublished Ph.D.
Thesis, Osmania University, Hyderabad, 1972, p. 38.
usage)’,6 continued under the Vijayanagara rulers much as it had
before. This is not the view of historians who have sought to see
the Vijayanagara state at the height of its territorial influence,
that is, including Tamil country. K.V. Subrahm anya Aiyer, Nila-
kanta Sastri, and T. V. M ahalingam, and, more recently, A. Krishna-
swami, see great changes in the functioning of various local institu­
tions (e.g., sabhâ, ür, nâdu) in Tamil country. These changes are
not, however, attributed to deliberate neglect or subversion by the
Vijayanagara rulers, but rather the consequences of new political
forms of the military state. This latter view is vigorously presented
by Krishnaswami who devotes two chapters to ‘feudalism’ and
‘feudatory relations’ in his work on Tamil country during the
Vijayanagara period.7
The issue o f ‘feudalism ’ has arisen with respect to the Vijaya­
nagara state as it has with few other pre-modern states in India.
W ith very few exceptions, ‘feudalism’ has been used by Indian
historians to cope with the specific features of pre-modern, Indian
political organization, rather than as a presumed genuine variant
o f those political forms of Europe or Japan to which the term
can legitimately be applied.8 In connection with the Vijayanagara
state, as perhaps with m ost other Indian states to which the label
‘feudal’ has been affixed, the prim ary question goes to the effective­
ness o f supralocal governmental functions, that is to the Vijaya­
nagara state. As with the Cholas, there is no question of whether
there was a state. R ather, the questions are — as with the Cholas
— what kind o f state, and particularly what kind o f governing
functions it carried out in contradistinction to those executed
at the level o f localities, and, moreover, by what means were locali­
ties linked to the Vijayanagara state. Feudalism, taken most
broadly, appears to answer these questions, and the concept has
been considered by many of the scholars who have written on
Vijayanagara.
D.C. Sircar, who has condemned the use o f the term ‘feudalism’
for all other medieval states of India, reservedly considers that the
6 B.A. Saletore, S o cia l and P olitical L ife in the Vijayanagar E m pire (A.D. 1346-
A.D. 1646), M adras, 1934, v. 1, p. 342.
7 A. K rishnasw am i, T he T am il C ountry under V ijayanagar, Annam alai University,
Annam alai, 1964, pp. 100-2.
8 R.S. Sharm a’s, Indian F eudalism : c. 300-1200 (University of Calcutta, Calcutta,
1965) is the m ajor exception for Sharm a here proposes a comprehensive system
of ‘feudal’ relationships.
term may appropriately be applied to the Vijayanagara state
because o f the central feature o f the am aram tenure, usually con­
strued as a military service tenure.9 N. Venkataramanayya,
whose work on am aram tenure is the most complete, denies the
appropriateness o f the concept ‘feudal’ on the basis o f the complete
absence of any idea of fealty and hom age.10 As utilized by scholars
less careful than these two, the concept of ‘feudalism’ in Vijaya­
nagara history merely obscures the failure to deal satisfactorily
with the fragmentary and contradictory evidence and interpreta­
tions of V ijayanagara inscriptions with respect to the state.
W hat is perhaps most striking about the use o f a feudal concep­
tion in Vijayanagara history is that it appears to stem from the
descriptions of European contem poraries of K rishnaraya and
Achyutaraya o f the sixteenth century. The Portuguese Domingo
Paes’ and Fernao N uniz’ reports on the reigns of these two kings
of the third dynasty, in a .d . 1520-2 and a .d . 1535-7 respectively,
provide the basic inform ation on the n ayaka system. These P ortu­
guese views have been among the m ost influential in forming an
understanding o f Vijayanagara polity precisely because it is as a
system that the polity is described. Paes’ account is the most
im p o rtan t:
Should any one ask w hat revenues the king possesses, and w hat his
treasure is th at enables him to pay so m any troops, since he has so m any
and such great lords in his kingdom , w ho, the greater p a rt o f them , have
themselves revenues, I answer th u s: These captains w hom he has over
these troops o f his are the nobies o f his kingdom ; they are-lords, and they
hold the city, and the tow ns and villages o f the k in g d o m ; ,|here are captains
am ongst them w ho have a revenue o f a m illion and a m illion and a h alf
o f p a rd a o s , others a h undred thousand pardaos, others tw o hundred,
three hundred o r five hundred th o u san d p a rd a o s , and as each one has a
revenue so the king fixes for him the num ber o f tro o p s he m ust m aintain,
in foot, horse, and elephants. These troops are always ready for duty,
whenever they m ay be called out and w herever they m ay have to g o ; and
in this way he has this million o f fighting m en always ready. E ach of
these captains labours to turn o u t the best troops he can get because he
pays them their salaries; and [in the review o f tro o p s by K rishnaraya]
. . . there were the finest young m en possible to be seen, for in all this array

9 L andlordism a nd Tenancy in A ncient and M edieval India as R evealed by Epi-

graphical R ecords (The D r R adhakum ud M ookerji Endowm ent Lectures of 1964),
University o f Lucknow, Lucknow, 1969.
10 N. V enkata Ram anayya, Studies in the H isto ry o f the T hird D ynasty o f Vija­
yanagara, University of M adras, M adras, 1935, pp. 171-2.
 I did n o t see a m an th a t w ould act the c o w ard . B esides m a in ta in in g these
tro o p s, each c a p ta in h a s to m ak e his a n n u a l p a y m e n ts to the king, a n d the
k ing h a s his ow n salaried tro o p s to w hom he gives p a y .11

Nuniz supports Paes’ description by enumerating some of the

great n ayakas, their income and military contribution to the king’s
armies and their contribution to the central treasury. He states:
. . all the land belongs to the King, and from his hand the captains
held it. They make it over to the husbandm en who pay nine-tenths
to their L ord; and they have no land of their own, for the Kingdom
belongs entirely to the King. . . ,’12
. Vijayanagara historians, especially those who urge the feudal
concept, have placed great reliance upon these brief descriptions.
Few have evinced much discom fort at the ease with which Euro­
pean forms of political organization could be transferred to medieval
South Indian society. V enkataram an ayya , am ong the deepest
and best scholars o f Vijayanagara history, expresses his reservation
about applying ‘feudal’ to the system outlined by Paes and Nuniz.
‘The n ayan kara system’, he writes, ‘has no doubt strong affinities
to feudalism, but it has also many differences.’13 Am ong the diffe­
rences are that the political bonding of European feudal relations
based upon fealty and hom age were absent as was the practice of
‘sub-infeudation’.14 In fact, Venkataram anayya is content to see
the ‘system’, rather vaguely, as one of military tenure under a
central authority: ‘. . . land was held immediately or mediately
o f the em peror on condition o f military service.’ 15 D.C. Sircar,
who has taken the strongest opposition to the use of the concept
‘feudalism’ in India, prefers instead to speak of ‘landlordism ’
u Sewell, op. cit., pp. 280-1. The pardao was a gold coin o f uncertain value;
apparently it was equal to 360 Portuguese reis of the sixteenth century and is reckoned
by Sewell to have had a value, ‘in more recent days’ of three-and-a-half-rupees
or seven English shillings (op. cit., pp. 270-1 n).
12 Ibid., p. 379. U nfortunately, neither Portuguese chronicler com pares the
situation observed in Vijayanagara with what they knew of Portugal, or perhaps
more particularly, Portuguese colonies. However, the term captainici was a royal
grant to ‘proprietory landlords’ (donatarios) who settled portions of Brazil at their
own expense in return for which they enjoyed adm inistrative, fiscal, and legal con­
trol over colonists (C.R. Boxer, Salvador de S a and the Struggle fo r B razil and A ngola,
University o f L ondon, London, 1952, p. 3).
13 V enkataram anayya. S tudies in the Third D ynasty, p. 171.
14 Ibid., p. 172.
is Ibid.
(which raises at least as many problems as the term it replaces).16
In his analysis o f Vijayanagara inscriptions, Sircar concedes a
special element — am aram tenure — which contrasts the Vijaya­
nagara state with all other medieval Indian states. But, correctly,
he dismisses the arguments of those who speak o f an ‘Indian
feudalism’.
. . . the landlordism o f ancient and medieval India. . . should n o t be
confused w ith feudalism . In India, the king was never the actual owner
o f the land under perm anent tenants. T he m ajority o f the num erous
charters, discovered all over India. . . records grants o f land w ithout
stipulating any obligation o f the B rahm ans and tem ple authorities to
the donors. Obviously, the priestly class was the m ost unsuitable for
rendering services o f the feudal type. . . . [on the contrary] it is generally
stated in clear term s in the grants th at the donees were exem pted from
all obligations including the supply o f unpaid labour and sometimes
also they were entitled to sell o r m ortgage the d onated property. The
object o f the grant is generally stated to have been the religious m erit and
fam e o f the d o n o r and their parents. T here are only a few charters re­
cording grants o f land to people o f the w arrior and other classes, som etimes
for services rendered to the king. But, excluding the amara tenure o f late
medieval South India, there is absolutely no m ention o f obligations having
resem blance w ith those o f the feudal type. . , 17

At another point in his discussion, Sircar attem pts to accom modate

am ara tenure to his terminology in the following w a y :

. . . the amara tenure was sim ilar to the allotm ent of land to the priest,
barber, w asherm an, carpenter an d others for the services to be received
from them regularly. . . . The Amaranayakas gave their lands to m inor
landlords on sim ilar term s o f m ilitary service ju st as the subordinate
rulers had various grades o f vassal chiefs u n d er th em .18

This form ulation, homogenizing as it does the privileges o f political

and ritual power with the rights of those providing m inor, village
service, appears to be based upon the conception o f inam tenures
of British India and leaves unclear, and certainly undem onstrated,
the continuities o f legal and bureaucratic forms which m ust be
16 L andlordism and Tenancy in A ncient and M edieval India as R evealed by
E pigraphical R ecords. The D r R adha K um ud M ookerji Endowm ent Lectures,
University o f Lucknow, 1964; University o f Lucknow, Lucknow, 1969. Sircar's
specific views in this publication are substantially different from his earlier views in
a work also cited here, L a n d S ystem and Feudalism in A ncient India (1966).
17 ibid., p. 33; emphasis mine
18 Ibid., p. 32.
 shown to have provided the foundations both o f medieval and
m odern arrangements.
Nevertheless, Sircar’s position, even with its serious faults, is
to be preferred over those o f other scholars who, like Sircar, have
relied primarily upon inscriptional evidence and still hold to a
conception of Vijayanagara feudalism. A. Krishnaswam i’s recent
study, The T am il C ountry under Vijayanagar is unreserved in its
view o f the nayaka system as feudal.19
Krishnaswami makes use o f the term ‘feudalism ’ in the titles
of several of the sections of his m onograph,20 and he makes
statem ents such as: ‘this nayankara system o f the feudal arrange­
ments in the Tamil country seems to have been in existence from
the time o f the conquest o f the region by K um ara K am pana.’21
He also uses phrases such as ‘feudal revenue’ to refer to various
kinds o f payments, to refer to many kinds o f subordinates and
‘feudal vassal’.22 Though one o f the strengths o f this work is the
careful consideration o f extant interpretations on m any difficult
points, Krishnaswam i does not query others’ views on feudalism.
This is especially peculiar in the case of Venkataram anayya upon
whom Krishnaswami relies in m any other ways. In fact, there is
much carelessness in Krishnasw am i’s usage, sometimes resulting
in distortion of the evidence. This is most tellingly seen in bis
quotation of the Paes passage cited above. In the second sentence
of th at much quoted paragraph, Krishnaswami inserts the word
‘feudal’ which occurs nowhere in the original chronicles. Thus,
that sentence in Krishnaswami reads: ‘These captains whom the
king has over the feudal troops. . . .’23
At the base of this question o f ‘Vijayanagara feudalism ’ is the
fragm entary evidence o f the perio d ; it is even m ore difficult than
Chola evidence. Therefore, systemic views or given paradigms of
the political system weigh very heavily on how this evidence
19 Annam alai University Historical Series, no.20, Annam alai University,
Annam alainagar, 1964.
20 Ibid., pp. 161, 176, 179.
21 Ibid., p. 181.
22 Ibid., p. 183.
23 Ibid., p. 177, citing the correct pages in Sewell, that is p. 280. Curiously,
V enkataram anayya in Studies in the T hird D ynasty, p. 170, also misquotes the same
sentence for he deletes the words ‘o f his’ after the word ‘troops’ thereby weakening
a point that is consistently pressed by the Portuguese chroniclers, i.e. that all troops
were those o f the king.
is ultimately evaluated. The great im portance o f the Portuguese
descriptions is precisely in providing a view of a systemic whole
to which the disparate inscriptional evidence o f the Vijayanagara
period can be related. A part from inscriptions, the other m ajor
source o f evidence comes from the Mackenzie Collection. M any
of the accounts of this eighteenth century collection of local records
deal with the m igration of Telugu warriors and their followers
to those parts of Tamil country which had been peripheral zones
of settlement during Chola times. In m ost accounts, a m igrant
w arrior takes service under a n ayaka chief, clearing forest to es­
tablish a settlement, and prospering with the military fortunes of
his n ayaka patron. Given the construct o f feudalism, whether it is
taken without qualification as Krishnaswami does or taken as
possible variant as Sircar and Venkataram anayya appear to take
it, it is not difficult to see these documents confirming ‘feudal’
arrangements. But, try as one might, it is impossible to find firm
evidence for the ‘feudal system’ in Tamil country as seen by K rishna­
swami.24
The weight o f historiographical judgem ent about South Indian
political history clearly opposes the idea o f a Vijayanagara ‘feudal
system’. M ost historians of the period, while they may use terms
such as ‘feudatory’ and ‘vassal’, do not seriously consider the
conception. F o r them, the new elements which come into the
political system from about the fourteenth century — as outlined
in the introduction o f this chapter — do no t alter the basic political
organization; they treat the Vijayanagara political system as an
elaboration of th at system of political relations which had existed
in the southern peninsula from at least the tenth century and dis­
cussed here as a segmentary polity.
This is m ost clearly seen in works whose scope extend beyond
dynastic periods. T.V. M ahalingam perceives no basic discontinui­
ties during the middle period o f South Indian history in his South
Indian P olity. But, as already noted,25 M ahalingam, remarkably,
treats the entire post-Classical age in South India as a vast, un­
differentiated period with evidence of political usages from widely
disparate times and places taken as elaborations upon some single
structure of power relationships. The m ore critical work of Appa-
24 A sum m ary paragraph of ‘the feudal arrangem ents in Tamil country’ is found
in Krishnaswam i, op. cit., pp. 180-1.
25 See. Stein, ‘The State and the A grarian O rder p. 68.
dorai on economic history assumes essential continuity though,
again as already mentioned, he does not give much attention to
formal political arrangem ents,26 tending rather to accept Nila-
kanta Sastri’s view o f Chola polity and assuming its continuation
during the period o f Hoysala and Vijayanagara ascendancy in the
southern peninsula.
Sastri himself in various works, however, places more emphasis
upon the disjunctive character of the Vijayanagara period. This is
evident in several ways. First, there is his perception of the martial
character o f the state in Vijayanagara times and the assumption
o f a new and dynamic ideological thrust underpinning this m artial
state. Then there is his assessment that the formerly robust local
institutions of Tamil country were decisively weakened. Finally
there is his awareness o f the m ore cosm opolitan nature of the
Vijayanagara period as seen in the historical sources of the period,
including Muslim and European reports as well as greater and
more reliable literary sources from within South India.27 Sastri’s
sensitivity to these issues sets him apart from other South Indian
historians. The latter, in general, whether they study the Vijaya­
nagara period as a whole or o f one o f the Vijayanagara dynasties,
do not give serious consideration to the question of continuity
or discontinuity with the Chola state; or, like M ahalingam, in his
work on Vijayanagara adm inistration, they take as the prime
reference point, ancient, formalistic, and often didactic sastras
on government.28
Considered as a continuation of the earlier segmentary state,
Vijayanagara polity certainly presents changes in the system which
existed under the Cholas, but the continuity is impressive. In
territorial terms, the scope o f the Vijayanagara state during the
first dynasty was as great as it had become under R ajaraja Chola I.
Thus, it was not only a political system of very substantial dimen­
sion, as the Chola state had been, but it encompassed the m ajor
population concentrations of Tamil-speakers as well as K annada-
26 E conom ic C onditions in Southern India, A .D . 1000-1500, v. 2, ch. 6.
27 Especially his discussion in Sources o f Indian H isto ry w ith Special R eference
to S o u th India, (Heras M em orial Lectures, 1961), Asia Publishing House, M adras,
1964, chapter 3.
28 Cf. V enkataram anayya, Studies in the H istory o f the T hird D ynasty o f Vija­
yanagara and M ahalingam , A dm inistration and Social L ife Under Vijayanagar,
M adras, 1940.
and Telugu-speakers m uch as the Chola state and the later M adras
Presidency did. Also like the Chola state, was the gradual unfolding
o f the new ritual sovereignty o f the Vijayanagara state during the
fifteenth century; so was the expansion o f the Telugu w arrior élite
into Tamil country similar to the earlier movement o f Tamils into
A ndhra and K arnataka.
W ithin thirty years of the establishm ent of the dynasty upon the
foundation of the failing H oysala house under Ballala III, the
early Vijayanagara warriors brilliantly extended their overlordship
to the southern part of the peninsula, ending Muslim rule in M adu­
rai in a . d . 1371. The dram atic reconquest of M adurai transform ed
the Sangama dynasty of Vijayanagara from a powerful, if hazar­
dously based and remote, kingdom into a worthy successor to the
Chola. In this sense, the youthful founders o f Vijayanagara were
able to accomplish what the Hoysalas even under sometimes extra­
ordinary leadership had failed in. And, it is perhaps ironic that
one reason for the success of the Sarigam w arriors was the establish­
ment in a . d . 1347, alm ost simultaneously with their own beginnings,
o f the M uslim Bahmani state on their northern frontier. While
posing a continuous hazard to the young Vijayanagara state, it
also forced the Vijayanagara rulers to establish a lateral defensive
system westward and eastward across the peninsula from their
principal locus o f power on the Tungabhadra. Being constrained
to such a policy, the Vijayanagara rulers avoided one of the salient
weaknesses o f their predecessors, the Hoysalas. The latter vacil­
lated between expansion northw ard into what came to be called
the ‘Bombay K arn atak ’ and southward into the Kaveri basin, and
they succeeded in neither. It may also be added that the expansion
of the first o f the Vijayanagara rulers laterally across the peninsula
was made necessary by the success of the first rulers o f the Bahmani
sultanate in establishing close collaboration with H indu warriors
of A ndhra country, notably the K apaya N ayaka o f W arangal.29
The expansion of Vijayanagara sovereignty to Tulu country
in the west and Penukonda in the east reflects this pressure from
the north and created a war frontier between the states in the doab
tract o f Raichur. Denied expansion to the north, the Vijayanagara
rulers were forced into what was to become the second element of
strength and durability of their state, that is the expansion into
Tamil country.
29 N ilakanta Sastri, H istory o f South India, p. 231.
 The movement of the Vijayanagara overlordship southw ard came
in slow stages, being punctuated by a series o f wars with the
Bahmanis from the 1350s to the 1470s.30 In some cases at least,
the victims o f Vijayanagara expansion southward were Hindu
chiefs. Thus, the defeat o f the Sambuvaraya chiefs o f Rajagam-
birarajyam by K um ara K am pana, the son o f the Vijayanagara
king Bukka I, around a . d . 1363, was as impressive a victory as his
conquest o f the M adurai Sultanate a few years later. Rajagambi-
rarajyam included a substantial part of Tondaim andalam just as
the Sultanate included a substantial part o f Pandim andalam .31
But, such conquests were an exceptional m anifestation o f the
expanding Vijayanagara overlordship. In the fourteenth and
fifteenth centuries what m ore characteristically ‘expanded’ was a
successful claim of overlordship, not direct Vijayanagara control.
G reat chiefs, like the Zam orin of Calicut,32 and small chiefs, like
the n atta r o f Ponpattai (in Cholam andalam ),33 recognized the
overlordship o f Vijayanagara in m ore appropriate ways: the
former by responding prom ptly to a command from Devaraya II
for the presence at the royal court of the Persian am bassador
A bdur Razzak (who recorded the event) and the latter by the
acknowledgement o f the right o f the Vijayanagara prince K um ara
K am pana to present a gift to a temple in his territory.
D uring the expansion o f the Vijayanagara overlordship from the
earliest rulers o f the first dynasty through the relatively short-lived
second, or Saluva, dynasty ( a . d . 1486-1505), the im portance of
dharm ic ideology for Vijayanagara rule is revealed: the mission
of the State in preserving H indu institutions against the depreda­
tions o f Muslims o f the Deccan. Venkataram anayya states: ‘the
history o f Vijayanagara may be said to be the history of a fierce
struggle between the Hindus o f the Deccan and tiie M uham m adan
rulers o f the D eccan’,34 N ilakanta Sastri similarly states:
. . . the basic n atu re of the historic role o f V ijayanagar . . . was to preserve
South India as the last refuge o f the traditional culture o f and institutions
o f the co u n try ,35 . . . th a t great em pire which, by resisting the onslaughts
M Ibid., pp. 232-45.
31 Krishnasw am i, op. cit., pp. 6-21. The territory included portions of modern
Chingleput and N orth and South A rcot districts. On other such conquests, ibid.;
pp. 18-21, 34, and 215-19; also N ilakanta Sastri, H istory o f S outh India, pp. 264-77.
32 N ilakanta Sastri, H istory o f South India, p. 260.
33 Krishnaswam i, op. cit., pp. 93-4.
34 T hird D yn a sty o f Vijayanagara. p. 145.
35 H isto ry o f South India, p. 305.
 o f Islam , cham pioned the case o f H in d u civilization an d culture in the
South for close to three centuries and thus preserved the ancient tradition
o f the country in its polity, its learning and its a rts.35
There can be no question that the existence of the highly militarized
power o f the Vijayanagara state south of the K istna-G odavari
had the effect o f stemming Muslim expansion. However, the
dharm ic posture o f the Vijayanagara rulers as protectors of Hindu
culture is, above all, ideologically significant. Vijayanagara king-
, ship and the Vijayanagara state were constituted upon, or soon
1 came to acquire, an ideological principle which distinguishes it
from previous South Indian states. It is this principle which most
decisively identifies the Vijayanagara overlordship, not presumed
differences in the basic structure of the State.
V ijayanagara kingship, like all medieval H indu kingships,
expressed appropriateness in term s of the m aintenance of dharma,
and especially varnasram adharm a. This is captured in Vijayanagara
inscriptional p ra sa tis in a very different way from Chola inscrip­
tions even though the dharm ic qualities of their kings were the
same. Both kingships claimed to be in the hands of conquerors
whose military exploits made them digvijayans; these kings were
the greatest of prestators whose gifts to gods and Brahmans assured
the welfare o f the world.
Among the things that distinguish the Vijayanagara from the
Chola overlordship is the personal character o f the form er or their
agents as dharm ic actors. Vijayanagara inscriptions of the four­
teenth century onwards depict the Vijayanagara king, his son,
or preceptorial agent m aking gifts to temples or to Brahmans,
adjudicating disputes among such personages, or re-establishing
temple worship long interrupted by Muslim depredations or
other disorders. There is in the Vijayanagara records an immediacy
o f the royal presence that is largely absent from m ost Chola ins­
criptions. In the latter, royal dharm a and prestations are realized
through the often im personal and remote m ediation of an unnam ed
ajndpati , or ‘executor’37 It is this remoteness and indirection
of the expression of Chola rajdharm a — the elaborate set of docu­
ments and orders that connect a gift to Brahmans or gods with the
Chola kings — that gave to earlier kingship its bureaucratic,
36 Ibid., p. 264.
17 I am indebted to discussions with A rjun A ppadurai and Nicholas Dirks for
clarification o f this point. Also see, Sircar, Indian Epigraphy, pp. 143, 374.
even ‘Byzantine’, tone. Vijayanagara inscriptions, by contrast,
place the king, his kinsmen or guru in the arena o f prestation in
a very direct way.
In nothing else is the ritual focus o f the Vijayanagara kings so
clear as in the m ahanavam i festival, an annual royal ceremony of
the fifteenth and sixteenth century occuring about 15 September
to 15 October. Since Vijayanagara times, the nine day festival,
followed by a tenth and final day — dasara — has been im portant
in many parts o f the m acro region. M ost famous in K arnataka,
as D asara, it has been continuously celebrated at least since Raja
W odeyar sponsored it in late September, a . d . 1610 at Seringapat-
am .38 The same festival, also called nava ra tr i , is celebrated by
the n ayaka successors o f Vijayanagara. The M ahanavam i is
described in vivid detail by the Portuguese sojourners of the six­
teenth century, Paes and Nuniz, and also mentioned by the Italian
Nicolo Conti whose report o f a visit to Vijayanagara about a . d . 1420
is the earliest extant and by A bdur Razzak, the Persian am bassador
ordered to Vijayanagara by the king Deva Raya II around 1442.39
The theoretical works o f H ocart and G onda40 prepare the way for
the acceptance o f the ritual actions o f ancient Indian kings, but it
is scarcely possible to find a better medieval example than the
nine days o f the M ahanavam i dedicated to protection and regenera­
tion at the capital city of Vijayanagara.
Puranic sources speak o f two im portant nine day festivals which
m ark the turning o f the three seasons o f the sub-continent: in
M arch-April, after the harvest o f the sam ba or rabi crop and the
38 C. H ayavadana Rao, H isto ry o f M ysore, ¡399-1799 A .D ., Governm ent Press,
Bangalore, 1943, v. 2, p. 68 and H istorical Sketch es o f the South o f India in an A ttem p t
to Trace the H isto ry o f M yso o r . . . to 1799 . . . . by Lieutenant C olonel M a rk W ilks,
P olitical R esident at the Court o f M ysore, ed. M urray Hammick, Governm ent Branch
Press, Mysore, 1930, v. 1, pp. 61-3. The M ahanavam i begins on the ninth day of
the increasing m oon o f the seventh lunar m onth, dsvina (Tam il m onth, purattaci).
See th e fo rth co m in g essay (Princeton University Press), on this festival by the present
author in a volume o f studies on the G upta age edited by Bardwell L. Smith, 1979.
39 Sewell, Forgotten Empire, Paes’ chronicle, pp. 253-64 and N uniz' chronicle,
pp. 357-60. C onti’s and Razzak's accounts are found in R.H. M ajor, India in the
F ifteenth C en tu ry; a Collection o f N arratives o f Voyages to India. H akluyt Society,
London, no.22, 1857.
40 A.M. H ocart, Kings and Councillors, University o f Chicago Press, Chicago,
1970 (originally published in Cairo, 1936) and J. G onda, A ncient Indian Kingship
fro m the Religious Point o f View. E.J. Brill, Leiden, 1966.
onset o f the hot season, and in September-October, after the harvest
o f the k a r or k h a r if crop and the onset o f the cold season. The
first of these is associated with the god Ram a and the second with
the goddess Devi, or Durga. Elements of both these festivals, as
described in puranic works, can be seen to exist in the M ahanavami
o f medieval Vijayanagara, and the Ram a m otif m ay have been
as im portant as the Devi m otif though it is the later (i.e. September-
October) nine days that was celebrated at Vijayanagara.
The m ost elaborate description of the festival was that of Paes,
from whose account the following features may be stressed.
T hroughout the nine days, festivities are centred in the ‘citadel’ area
o f Vijayanagara, before the palace and on two, large, permanent
structures: one o f which is called ‘The House o f Victory’ by Paes
(‘Throne Platform ’ by Longhurst, the archaeologist o f Vijayanagara.
and the mcthancivamJ dibba according to the m odern residents of
Ham pi) and the other ‘The K ing’s Audience H all’. The ruins of
both are massive granite slab platform s showing structural signs
o f having borne large wooden superstructures as described in the
medieval chronicles. These were constructed by Krishnadevaraya
around a .d . 1513 following his Orissan campaign and victory over
the Gajapatis.
A round and within these buildings were enacted the events of
the festival.41 Here the king observed the many processions,
displays, and games and here he accepted the hom age and the
gifts of throngs of notables as he sat upon his bejewelled throne.
The king sometimes shared this throne, or sat at its foot, while
it was occupied by a richly decorated processional m u rti o f a g o d ;
at other times he was alone. The god is not identified. W ithin the
‘House o f Victory’ was a special, enclosed, and, again, richly
decorated chamber in which the image was sheltered when it was
not on display before the public participants in the festival. At
several points in the proceedings, the King, sometimes with Brah­
m ans and sometimes alone, retired to this enclosed chamber of
the deity for worship. Both the Audience Hall and the Throne
Platform bear bas-relief sculpture along their granite sides depicting
many of the events described by Paes and Nuniz.
In front o f the two structures which were the centre of the festival
activities were constructed a num ber o f pavilions which contributed
41 A.H. Longhurst, H a m p i R uins D escribed and Illustrated, G overnm ent Press,
M adras, 1917, pp. 57-70.
to the aura o f wealth and sumptuousness o f the festival as a whole.
They were elaborately decorated, in among other ways, with ‘de­
vices’, apparently heraldic symbols, o f the grandee occupants for
whom the pavilions were tem porary housing during the festival.
Nuniz reported that there were nine m ajor pavilions (he called them
‘castles’) for the m ost illustrious o f the notables and that each
m ilitary com m ander also had to erect one in the broad space
before the palace.42 Razzak, as am bassador from Persia, was
ensconced in one of these.43
Access to the guarded, central arena o f festival activity was gained
by passage through several gates enclosing wells o f the temple
precincts; Paes’ description o f this suggests passage through a
series o f gateways as the medieval pilgrim moved toward the
sanctorum in the great temples o f South India.44 Once gained, the
spacious open area before the palace, the Audience Hall, and the
house o f Victory was ringed about with the pavilions referred to
and with shaded seating from which the great — soldiers, sectarian
leaders and others — viewed the proceedings immediately before
the House o f Victory on whose higher levels the K ing sat.
W hat was viewed was a com bination o f a great durbar with
its offerings o f homage and wealth to the King and the return
gifts from the King — exchanges o f honours; the sacrificial recon­
secration o f the K ing’s arms — his soldiers, horses, elephants —
in which hundreds or thousands o f animals were slaughtered;
darshana and p u ja o f the King’s tutelary — the goddess — as well
as his closest agnatic and affinal kinsm en; and a variety o f athletic
contests, dancing and singing processions involving the King’s
caparisoned women and temple dancers from throughout the
realm, and fireworks displays. The focus of these diverse and
magnificent entertainm ents was always the King as glorious and
conquering warrior, as the possessor o f vast riches lavishly dis­
played by him and his women (queens and their maids of honour)
and distributed to his followers. The King was fructifier and agent
of prosperity of the world. M ost succinctly assessed, the M aha­
navami appears as a com bination o f the asvam edha (the greatest
o f all royal sacrifices, with its celebration and consecration of
42 Sewell, F orgotten E m pire, p. 357.
43 M ajor, op. cit., ‘N arrative of the Voyage of A bd-er Razzak, Am bassador
from Shah Rukh, A.H. 845, A.D. 1442', p. 36.
44 Sewell, op. cit., pp. 253-4.
kingly military prowess as symbolized in the horse o f the king)
and the description of R am a’s return to Ayodhya in C anto 130
o f the final book o f Valmiki’s R am ayana.
Com paring the M ahanavam i o f Vijayanagara with the archaic
horse sacrifice may seem superficial or strained, for the Vijayanagara
period knows o f no such royal sacrifices; none are even alluded
to as during the Chola period. Certain com m on features o f the
archaic asvam edha and the medieval M ahanavam i may indeed be
superficial. Both are ten day rituals and at least one of the great
asvam edhas was celebrated on a m ahanavam i (the M arch-April
or C haitra one) by Yudhisthira, hero o f the M ah abharata.4^ Yet,
the anointing of royal arms by priests 46 and by royal women along
with animal sacrifices are prom inent features o f the Vijayanagara
festival; these elements figure conspicuously in the two most detail­
ed accounts of it. Paes wrote of the King’s women:
They com e in regular order one before the other, in all perhaps sixty
w om en fair and strong, from sixteen to tw enty years o f age. W ho is he
th a t could tell o f the costliness and value o f w hat each o f these women
carried on her person? So great is the w eight o f the bracelets and jewels
carried by them th a t m any o f them cannot su p p o rt them , and women
accom panying them assisting them by supporting th eir arm s. In this
m anner and in this array they proceed three tim es aro u n d the [K ing’s]
horses, and at the end retire into the palace.47

Com pare this with the Scuapatha Brdhmancr.

It is the wives th a t anoint (the horse), for they — to wit (m any) wives —
are a form o f prosperity. . . ,48 the wives w alk ro u n d (the horse) . . .
thrice they w alk round. . . 49

An even m ore strikingly parallel feature o f the M ahanavam i is

the symbolic significance of the K ing’s horse in the consecration
of his kingship. Paes describes a troop o f richly caparisoned horses
45 The M ahabharata o f K rishna-D w aipayana Vyasa, trans. by P.O. Roy, Oriental
Publishing Co., C alcutta n.d., v. 12, ‘Aswamedha Parva’, sec. 84, p. 161 and passim .
46 Priests offer prayers for and sprinkle water upon the K ing’s horses and ele­
phants: N uniz in Sewell, op. cit., p. 358.
47 Seweli, op. cit., p. 263.
48 J. Eggeling (T rans.), The Satapatha-B rahm ana A ccording to the T ext o f the
M adhyandina School in Sacred B ooks o f the E ast. Clarendon Press, Oxford, 1900,
v. 44, pt 5, p. 313. On the sprinkling anointm ent of the sacrificial horse, see pp.
278-9.
49 Ibid., pp. 322-3.
brought before the King at one point in the M ahanavam i, and
leading this troop was one bearing ‘two state umbrellas o f the
king and grander decorations than the others’.50 O f this one,
Paes writes:
Y ou m ust know th a t this horse th a t is conducted w ith all this state is a
horse th at the king keeps, on w hich they are sworn and received as kings,
and on it m ust be sw orn all those th a t shall com e after them ; and in case
such a horse dies they put an o th er in its place.51

This suggested com parison with the archaic, Vedic asvam edha
is more than anachronistic; it is also flawed in being an apparent
violation of Vedic prohibitions as well. The bright half o f the lunar
m onth o f dsvina (Tamil m onth o f p u ra ttd c i) when the M ahanavami
festival occurs is said to be inauspicious for Vedic learning and by
extension for other Vedic activities as well. These activities are
enjoined as an adhyaya according to the dharam asastra o f Apara-
ka.52 Brahmans do figure in the M ahanavami festival as ritual
performers along with the King in relation to the King’s tutelary;
they also are recipients of royal gifts. But, Brahm ans do not
dom inate the ritual arena, which is very much the King’s and in
one description o f the festival, that o f Conti around a . d . 1420,
Brahm ans appear to have been publicly reviled.53
The association o f Sri Ram a with the M ahanavam i is somewhat
more direct. The same m ood o f celebration by a people of their
king found in the final verses o f the Valmiki Rcimayana suffuses
the vivid descriptions o f the festival o f Vijayanagara. And, there
is the further shared conception o f regal deliverance from threaten­
ing evil. Saletore captures this quality in his statem ent on the
festival.
Religious in atm osphere, it is essentially political in its significance.
F o r it com m em orates the anniversary o f R a m a’s m arching against
R avana, and in its tw ofold aspect o f the w orship o f D u rg a and o f the
ayudhas o r arm s, culm inating in the V ijaya-dasam i [victorious tenth day],
was particularly suited to V ijayanagara times when fatal issues loom ed
om inously in the political h o rizo n .54
50 Sewell, op. cit., p. 262.
51 Loc. c it..
52 Kane, H isto ry o f D haram sastra, v. 2, pt I, p. 395.
53 'The Travels of Nicolo Conti in the East in the Early Part o f the Fifteenth
Century', in M ajor, op. cit., p. 28 and Sewell, op. cit.. p. 83.
54 Saletore. Socia l and P olitical L ife in the Vijayanagara E m pire, v. 2, p. 372.
 The god R am a appears im portant in yet another way. One of
the m ajor temples o f the capital city under the Tuluva dynasty of
Vijayanagara, and especially in Krishnadevaraya’s time was that
dedicated to R am achandra, the hero o f the R am ayana and the
seventh avatar o f Vishnu. A separate shrine within that temple
was dedicated to the Devi consort o f Rama. This temple was also
called the H azara R am a temple, and it was the only shrine within
the palace precincts and thus quite proxim ate to the varied activities
o f the festival. The unidentified image in the accounts o f the
M ahanavam i m ay have been the consort goddess of Sri Ram a if
the proximity o f the Ram a temple is taken as im portant and if the
judgem ent of Longhurst, the archaeologist of Hampi, or Vijaya­
nagara, is correct that the R am a shrine was the private place of
worship of the Tuluva kings.55
Though all historians o f Vijayanagara have mentioned the
M ahanavam i festival, often quoting long excerpts from the accounts
o f Paes and Nuniz, the festival has not received analysis as a single,
unified ritual event. Culled from the detailed reports have been
odd facts (e.g. ‘the rents’ of the n ayakas are paid at this time and the
King ‘owns’ all the land) and descriptions o f the King, his high
officials, and queens. However, this festival, like others, is perceived
as a unified system o f action and meaning, and should be interpreted
in that light.
Two aspects o f the m ahanavam i tlrtha immediately seize atten­
tion: its overwhelmingly royal character and its symbolically
incorporative character. These aspects confirm conceptions of
55 Longhurst, op. cit., p. 71. There is some uncertainty about the personal reli­
gious preferences of K rishnadevaraya. He appears to have been a Vaishnava bhakta
as were others of his dynasty, the preceding Saluva and succeeding Aravidu dynasties,
and there is his special relationship with the great M adhva scholar Vyasatirtha.
The latter was the K ing’s guru as well as the head of a school in Vijayanagara which
had been supported by the Saluvas as well. (S e e M. Ram a Rao, Krishnadeva R aya,
National Book Trust, New Delhi, N ational Biography Series, 1971, pp. 39-40).
This relationship m ay have been m ore personal than sectarian, and m ight not,
therefore, presum e a preference by the King for the god Krishna, the deity o f the
M adhva sect, over the god Ram a. It is interesting to note that the large K rishna
tem ple in Vijayanagara was built by K rishnadevaraya about the same time as the
R am a temple and the T hrone Platform ( ‘H ouse of Victory’). This Krishna temple
is about 1.5 miles from the palace area, and according to L onghurst (op. cit., p. 96),
sheltered an image taken from the G ajapati fortress o f Udayagiri around a . d . 1513.
This K rishna image is thus more a trophy of the King than his tutelary.
ritual kingship assumed in the concept o f the segmentary state
suggested some forty years ago by Hocart and since then largely
ignored by Indological scholars.56 H ocart’s conception o f the
king as ritual perform er and the prim ary agency for the prosperity
and welfare o f the realm, and his attention to the symbolically
integrative character o f the temple and the city are extraordinarily
well-realized in the M ahanavam i festival. Kingly ritual power
is expressed in num erous ways: in the m anifestation o f wealth
displayed and elaborately redistributed at many points o f the
nine day festival; in the various consecratory actions involving
the King’s arm s as the means o f his royal fame and protection;
and also in the King’s frequent and often solitary worship of
(and ultimately identity with) the deity who presides with him over
the festival, and in whose name and for whose propitiation the
festival occurs. Certain signs o f the Devi (Durga) worship are
clear in this festival, and they deserve notice. According to the
D evi B hagavatam Purana, females in procession before the goddess
form an essential element o f the N ava-ratri or M ahanavam i
festival.57 It is also well to recall that since the M ahanavam i is
considered a time o f danger, the protective power o f the Devi
and also the King are enhanced.
Incorporative elements to which H ocart drew attention in his
w ork included the subordination o f all gods and all chiefs to the
king. This incorporation is signified by various means in the M aha­
navami festival. The palace site o f the festival is reached through
two large gates over which towers are constructed.58 These massive
gateways were apparently destroyed by the Muslim invaders o f the
city in the sixteenth century; gates in other parts o f the Hampi
ruins, at some distance from the ‘citadel’ of K rishnadevaraya’s
time,59 confirm Paes’ description o f these structures which resemble
the gopuram leading to the sanctorums of H indu temples of the
time. King and god are at least homologized, if they are not
56 H ocart, K ings a n d Councillors. He is not mentioned by G onda, A ncient Indian
Kingship fr o m the R eligious P oint o f View. D um ont’s H om o H ierarchicus, is a recent
exception to the general neglect o f H ocart.
57 S ee B. R am akrishna Rao, ‘The D asara Celebrations in M ysore’, Q J M S , v. 11
(July, 1921), pp. 302-3 and The S r i M a d D evi Bhagavatam in Sacred B ooks o f the
H indus, v. 26. Trans. Swami Vijnanananda (H.P. Chatterji), Panani Office, Alla­
habad, ‘On N avaratra’, pt 1, ch. 26, pp. 225-9.
58 Sewell, op. cit., pp. 253-4.
59 Several are shown in L onghurst, op. cit., pp. 47-9.
equated. The pavilions erected in the spacious, interior courtyard
before the Throne Platform to house the notables o f the realm are
called ‘castles’, or dwellings o f great men placed within the pre­
cincts of the palace and thus under the protection o f the great king.
Gods of the K ing’s realm are also incorporated in his city.
Included here are permanent resident deities like the G ajapati
Krishna of Udayagiri and Vitthala (Vithoba) from distant M aha­
rashtra.60 It also appears that deities from elsewhere in the realm
are brought to the capital during the festival and presented to the
King for his adoration.6! And, servants of gods throughout the
K ing’s realm come to do obeisance. This included priests, but
m ost conspicuously it was the temple women (whom the Portuguese
called ‘courtesans’) of shrines everywhere. These temple dancers
and musicians performed before the King just as they did before
the god to whom they were dedicated.62
Following H ocart’s perceptive discussion, it is possible to point
to the crucial place o f the city — this city o f victory — in the
total m oral order over which the Vijayanagara kings exercised sway.
The city, H ocart wrote, ‘never stands for anything specific; it
is never less than the whole world, and its parts are the parts o f the
world. . . .’63 And, persistently linked to the city and its establish­
ment is the goddess Durga. As Bhuvanesvari, ‘mistress of the
world’,64 the goddess was by tradition65 propitiated by Vidya-
ranya, or M adhvacharya, who is supposed to have been the pre­
ceptor o f the founders o f the city in a . d . 1336. This connection of
the great goddess and the city o f the Rayas was first presented by
William Taylor in his 1835 translation of several Tamil chronicles
60 K rishnadevaraya built a temple for V itthala which is considered one o f the
m ost beautiful in South India. However, this god was worshipped in Vijayanagara
before that time (F urther Sources o f Vijayanagara Flistory, v. 3, p. 47 and n.).
61 Sewell, op. cit., p. 264.
62 Ibid., p. 253 and Venkataram anayya, Studies in the T hird D ynasty, pp. 404-5.
This is another parallel with the asvam edha sacrifice over whose ten days of the
consecration of the horse, different gods are presented: Satapatha Brahm ana,
pt 5, pp. xxxi, 361-70.
63 H ocart, op. cit., p. 250.
64 M onier-W illiam s, A Sanskrit-E nglish D ictionary, p. 760.
65 The association of M adhvacharya with the Sangama founders o f Vijayanagara
in a . d . 1336 has been a contentious one; according to the official biography of the
sam pradaya, M adhva died in a . d . 1317 (B.N.K . Sharm a, S r i M a d h va ’s Teachings
in H is Own W ords, Bharatiya Vidya Bhavan, Bombay, 1970, pp. 4-9).
of later medieval times,66 and though the idea continues to be
accepted by m ost Vijayanagara scholars, none have exploited fully
the symbolic power of the relationships of goddess propitiation,
the Rayas and the city. While scholars have failed in this, the
successor states of the Vijayanagara in South India, notably the
Wodeyars of Mysore and the N ayaka kings of M adurai did not;
they m aintained this royal festival in their capitals in full richness.67
The im itation by successors of the Rayas in South India (and
possibly also the M aratha king Sivaji) of the ruler as an active
ritual ‘principal’ (to use H ocart’s term) was the result, in part at
least, of the perceived threat to Hindu institutions from Muslim
powers of peninsular India. Just as the expansion of the Bahmani
sultanate and its successors in the Deccan acquired a special
saliency in the ideological presentation of the Vijayanagara state
as symbols of danger to the dharm a which H indu kings were bound
to protect and nourish, Muslims of a later day continued to have a
special meaning, however dubious their actual threat may have been.
Actually, those who bore the brunt of Vijayanagara military
power were most often Hindu rulers, not Muslims. And, ironically
perhaps, the most strategically placed military units of the Vijaya­
nagara military formations were composed of Muslims, as is
generally conceded.68 This factor is often elicited to explain the
ultimate defeat of Vijayanagara arms in the sixteenth century;69
it nevertheless remains clear that Muslim contingents were res­
ponsible for at least part of the great early successes of the Vijaya­
nagara rulers against Hindu houses which they toppled. The
Vijayanagara state was however not in fact dedicated to different
principles of rule as might be supposed from the confrontation of
H indu and M uslim forces in the Deccan, whatever the im portance
of its dharmic ideology. To suppose otherwise is to transfer to
66 O riental H istorical M anuscripts, in the T am il L anguage: Translated with
A nnotations, by William Taylor, 2v., C.J. Taylor, M adras, 1835, v. 2, pp. 102-3.
67 N oted by Taylor, op. cit., p. 103; Abbe Dubois, op. cit., pp. 569-71; Rice,
M ysore: A G azetteer, v. 1, p. 378 and a description of its celebration at Seringa-
patam in September 1783, according to the report o f a captured English soldier.
68 Based upon evidence such as the a . d . 1430 inscription o f the time of Devaraya
referring to 10,000 turushka (Muslim) horsem en in his service (E .C ., v. 3, ‘Introduc­
tion', p. 23).
69 The battle which fatally weakened the Vijayanagara kingdom at Rakshasi-
Tangdi (Talikota) is supposed to have been decided by the desertion o f two Muslim
com m anders o f Ram a R aya’s arm y (N ilakanta Sastri, H istory o f South India, p. 238).
an earlier time the com m unal politics o f the twentieth century. It
was essentially a continuation of the segmentary state of the Cholas
in terms of its basic political character.
A nd, like the rule o f the Cholas, Vijayanagara power was often
quite rem ote after an initial intrusion of its forces into territories
ruled by H indu chiefs. M any parts o f the deep southern peninsula
continued to be ruled by members o f the same families whom the
Vijayanagara armies had conquered. This is particularly true
of the Pandyan territory through most of the fifteenth century.
In most other parts of Tamil country, the ancient territorial ter­
minology remained, and Telugu N ayakas and Brahmans placed
in positions o f supralocal agents for Vijayanagara authority.
Hence, to overstate the ideological element and to speak of a
newly constituted basis o f state power and legitimacy in the Vijaya­
nagara period would be to distort the historical evidence which we
possess o f the period. Pre-Vijayanagara forms proved by and large
to be both adequate and durable.
A nother feature of Vijayanagara rule which also invites com ­
parison with the impressive Chola kingship was the diversity of
the peoples under each. In a technical sense, it is the rule over
m any and different peoples which has justified the use of ‘empire’
or ‘im perial’ in connection with the Chola and Vijayanagara
states. F or the Tamil Cholas based in the rich Kaveri basin, the
earliest regions to be included in their expanding sovereignty were
the two secondary central zones of the Tamil p lain : Pandim andalam
and Tondaim andalam . But soon after, Chola sovereignty was
established over places of dominantly non-Tamil population with
ancient cultural traditions of their own.70 Chola influence was
extended northw ard to Vengi and north-westward to Garigavadi
well before it was fully established in the Tamil peripheral zone
of Kongu (m odern Coimbatore and Salem).
Similarly, the first Vijayanagara dynasty, shortly after the
establishment of their sovereignty over the northern portion of
what was to become the empire, moved to establish themselves in
Tamil country. The process was repeated in the second dynasty
70 According to one of the earliest extant works on K annada rhetoric, the
K avirajam arga of N ripatunga, c. a . d . 850, the area in which K annada was spoken

extended from the upper Kaveri to the G odavari (N ilakanta Sastri, A H istory o f
South India, pp. 375-6). Telugu inscriptions date from fifth century and K annada
inscriptions from the sixth (ibid., p. 387 and Sircar, Indian Epigraphy, pp. 48-9).
as well. W hether the Vijayanagara rulers are to be regarded as
essentially K annadigas or Telugus, or whether they are to be
regarded as both from the very beginning — a position taken
recently by P. Sree R am a Sarm a —71 the territorial scope of their
power included the entire southern peninsula.
Thus both kingships were firmly based in one part of the southern
peninsula from which they drew the major resources for sustaining
their military supremacy, namely their soldiery; but both also
achieved overlordship in other, well-populated and wealthy parts
of the macro region. Overlordship in these latter places appears
to have resulted from three different kinds of processes. One was
the result o f adventurous pillaging expeditions o f small groups of
Telugu w a rrio rs.o r by large invasions’, as Krishnaswami has
labelled them, o f Vijayanagara forces under royal commanders.
The second process was the transform ation of local chiefs into
n ayakas, thus constituting the cement of the new overlordship
and at the same time, the means of strengthening the control of
local chiefs. The result of these two processes in both the Chola
and Vijayanagara cases were similar. The third means of extending
the Vijayanagara overlordship over the southern peninsula is
different from anything seen in C hola tim es; this was the incorpora­
tion of the support and the followings of sectarian groupings in
all parts of the southern peninsula. With respect to this, of course,
the dharmic ideology of the Rayas was all-im portant.
Raids and invasions into territories remote from their prime
bases led to perm anent settlement in both cases. In the Chola
period this occurred when parts o f the m odern district of Kolar,
in K arnataka, were brought under the control o f Tamils o f Tondai-
m andalam as seen in the famous Mulbagal inscription of a .d . 1072
discussed above. A variation on this expansion process in the
Vijayanagara period is seen during the fifteenth century. Then,
Telugu warriors, without frontally challenging the Tamil chiefs
of many areas of older settlement, established themselves in more
remote parts of the Tamil plain abutting on the plain at the foothills
of the Eastern Ghats. The results of this latter form of expansion
are recorded in m odern census volumes showing a zone of Telugu
speakers running from north to south and splitting some of the
m odern Tamil N adu districts into a dom inantly Tamil-speaking
71 As noticed above, this question o f origins has m ost recently raised again by
Filliozat; the discussion by Sree R am a Sarm a: op. cit., p. 37.
eastern side and Telugu-speaking western side.72 In these sparsely
populated interstices, Telugu migrants o f the fifteenth and sixteenth
centuries not only found scope for political control, but they found
black soils with which they were familiar and for which they
possessed the means o f exploiting in ways superior to the older
Tamil residents o f these areas.73 It was the relatively high propor­
tion of Telugu warriors and settlers in these peripheral parts of the
Tamil plain that explains the placement of the two subordinate
‘capitals’ of Vijayanagara in Tondaim andalam : Padividu and
Tiruvadi. The striking pattern o f this Telugu expansion was
noted by early British adm inistrators as well as by epigraphists.74
But the unique record of this process is contained in the accounts
of the records collected by Colin MacKenzie in the late eighteenth
and early nineteenth centuries. Telugu-speaking Reddis displaced
Tamil-speaking local leaders as dom inant cultivating groups, and
this displacement brought with it changes in the language o f temple
inscriptions from Tamil to Telugu, as at the great Vaishnava
temple o f T irupati.75 This movement o f Telugu warriors into
Tamil country was itself part of a larger pattern of movements.
D uring the late fifteenth century, Reddis were seen to be moving
into the western parts of A ndhra from their home territories to the
east as recorded in a Kurnool inscription of Saluva N arasim ha’s
time.76 And it may be speculated that this m igration of small
groups o f Telugu warriors was not solely the result o f the attractions
o f new areas o f pillage and settlement, but pressure during the
preceding century from Orissan warriors, like themselves, who
were pressing southward into the Vengi region. An inscription
o f a .d . 1396-7, from Palnad taluk in western G untur district
records a conflict between Oriya speaking Vaishnavas and Telugus
in which the former, called badugulavaru (much as in Tamil country
Telugus were referred to vadugans or northern people), were
72 K. Srinavasaraghavan. 'A Geographical Study of the Vellore Basin’, Indian
Geographical Journal, M adras, v. 11, no.3 (October, 1936). p. 230. Based upon
observations in the 1901 Census, India Census Commissioner, Census o f India,
1901, R eport, p. 289.
7j M adras, The Central Agricultural Com mittee, ‘Some Suggestions for M adras
Ryots By Special C orrespondent of the “ M adras Mail” ’, Bulletin N o.3, M adras,
1906. p. 2; India Office Library, P/V 2148.
74 See the observation of the epigraphist Venkayya in A .R .E .. 1904, para. 14.
■w A .R .E .. 1904, para. 14.
76 A .R .E .. 1960-1, copperplate num ber 4 of 1960-1.
accorded full participation in the Vaishnava sectarian activity of
pallinadu .11
W hether by the m ovement of small groups of Telugu warriors
or by invasions o f Telugu armies, these fourteenth and fifteenth
century conquests did not result in an easy or firm control over
Tamil country by the Vijayanagara overlordship. Tamil country
had virtually to be reconquered in the late fifteenth century by
Saluva N arasim ha and his son, N arasa Nayaka, to seize back the
formerly conquered territories in Tondaim andalam as well as
Chola and Pandya countries from local chiefs.78 Invasions of
these territories subsequently by the Orissan king Kapilesvara
Gajapati and his son, K um ara Hamvira in a . d . 1463-4 did nothing
to strengthen the Vijayanagara hold on Tamil country.79 Later,
the great Krishnaraya had apparently80 to send another Telugu
army into Tamil country to refurbish, yet again, the Vijayanagara
overlordship. This resulted in what Krishnaswami claims to have
been a ‘m om entous change’, namely the replacement o f a system of
governors (m ahdm andalesvarai) by four military commanders,
n ayakas, presumably to act under the king’s orders with the assis­
tance of dependent warriors called palaiyagars.% 1 The residue
— and perhaps the only result — of these successive invasions
of Vijayanagara warriors into Tamil country was the creation
of a stratum of super chiefs who were either Telugus themselves
or Tamils allied to Telugus. It is these personages who constituted
the segmentary leadership in Tamil country during the Vijayanagara
period.
Looked upon as segmentary states, one of the most crucial
differences between the Chola and Vijayanagara states is the so-
called ‘n ayaka system’. The notion of a ‘n ayaka system’ can
scarcely be considered indigenous; it is a conception derived from
the first ‘outsiders’ whose attem pts to understand the Vijayanagara
polity we have, the sixteenth century Portuguese Nuniz and Paes.
As already observed, there is little either in the Vijayanagara
77 A .R .E ., 1910, para. 49 regarding 556 of 1909.
78 Krishnaswam i, op. cit., pp. 106-15.
7l) Ibid., p. 112.
80 ‘A pparently’ because the sole source of this ‘invasion’ is a docum ent from the
M ackenzie Collection, the ‘K arnataka Rajakkal Savistara C haritam ' cited, ibid..
pp. 193-4.
si Ibid., pp. 193-5.
 inscriptions or literary evidence to support the ordered political
relationships described by these chroniclers. Thus, the n ayaka
system as a system , may be as alien to the facts of Vijayanagara
political relations as the conception of feudalism which derives
in considerable part from notions of a n ayaka system. Those
military personages referred to in many Vijayanagara inscriptions
by the title ‘n a ya k a ’ are extremely im portant, however, for they
came to comprise the major connecting elements in the Vijaya­
nagara segmentary state. The emergence of the new stratum of
; supralocal warriors became a possibility with the raids and, later,
the invasions o f Telugus into various parts of the m acro region
from the middle of the fourteenth century; the stratum became
a reality when these powerful outsiders forged links to the diverse
locality populations they ruled while retaining certain ties to the
Telugu Rayas in Vijayanagara on the Tungabhadra.
N a ya k a s o f the Vijayanagara period are seen by m ost scholars
as warriors possessing an office conferred by the central Vijaya­
nagara government. The term , am aranayahkara, signifies an office
(-k a ra ) possessed by a military officer or chief (n a ya k a ) in command
(am ara) o f a body o f troops.82 The office o f n ayaka carried with it,
according to the conventional scholarly understanding of the
system, prebendal rights over land usually designated as am aram
tenure (or am aram akani or am aram ahale ), and it is supposed that,
perhaps, three-fourths of the villages of Tamil country were under
this form of tenure.83 The proportion of land under this tenurial
form for Vijayanagara as a whole is generally regarded to be about
the same.84 Two hundred nayakas were presumed to have existed
in the empire of the middle sixteenth century based upon the state­
m ent of Fernao Nuniz for the years 1535-7.85 However, Nuniz
names only eleven of the m ost im portant of these officials and
specifies the territory and revenue for which they were responsible
as well as the num ber and composition of the troops which they
were to m aintain.86
82 Krishnasw am i, op. cit., 179-80. D.C. Sircar (Indian Epigraphical Glossary
M otilal Banarsidass, Delhi, 1966) defines na ya ka as a royal officer or ruling chief
(p. 214).
83 Krishnaswam i, op. cit., p. 180.
V enkataram anayya, Studies in the T hird D ynasty, p. 180.
85 Sewell, op. cit., p. 389.
86 Ibid., pp. 384 ff.
 In an exhaustive search of inscriptional records of the Vijaya­
nagara period to about a .d . 1530, Krishnaswami was able to find
references to a larger num ber o f n ayakas in Tamil country. Bet­
ween a . d . 1371 and 1422, he finds six nayakas m entioned; between
a . d . 1440 and 1459, he finds six m ore; between a . d . 1465 and
1491, nine n ayakas are found; between a .d . 1491 and 1508, another
ten are found; and during K rishnaraya’s time, a . d . 1509 to 1530,
a total of twenty-seven are m entioned in inscriptions.87 These
Tam il records88 pertain to typical local, or segmentary leadership
activities including gifts to temples, repair and construction of
tanks, reclamation of waste, and the collection of dues from temples.
In none of these records, nor in any other Vijayanagara inscriptions,
are there references to payments by nayakas to the emperor or
his officials; except for the account of Nuniz (Paes does not refer
to these arrangements), there is no corroboration o f what have
been regarded as the ‘feudal obligations’ of nayakas.
The political history of the Vijayanagara state is essentially the
history of great Telugu n ayakas, their formidable military capabili­
ties, their patrim onial power, and their relations to religious leaders
in a new level of authority everywhere in the southern peninsula.
Each o f these aspects of Vijayanagara politics — except perhaps
the last — have been explored at least partially in the extant
historical literature w ithout attem pting to account for their exis­
tence or their interconnections. These factors are addressed below
without, it m ust be urged, claiming that the answers offered are
final.
However assiduous the effort by Vijayanagara historians to
elaborate the structure of politics of the time, the core of their
discussions has pivoted on the great n ayaka s o f the kingdom. All
evidence turns upon their exploits and their conflicts, whether
one considers the three genre o f literary sources — Hindu, Muslim,
and European — or the inscriptions of the several centuries o f the
87 Krishnasw am i, op. cit., pp. 181-6.
88 All the inscriptions noted by K rishnasw am i were in Tamil. It is in fact rare
for Vijayanagara records in Tamil country, until the time of-the N ayaka Kingdoms
o f Tanjavur and M adurai, to be in other languages. The N ayaka kingdoms did
occasionally inscribe stones in Telugu and Sanskrit as noted by Sircar, Indian
Epigraphy, pp. 47-8. A nother of the rare exceptions is that noted in the A .R .E .,
1905, referring to num ber 38 o f 1905, an undated inscription in K annada m ention­
ing one Lingappa, son of C hikka Koneri N ayaka, at Padaividu in Tondaim andalam
(m odern Tindivanam taluk, South Arcot).
 Vijayanagara state, or whether one relies, as Venkataramanayya
does, upon the local accounts collected by Colin Mackenzie.
There have been few efforts to go beyond, or behind, these great
political figures, to discover the structural framework within which
they operated. Ever and again, it is to the powerful personages
o f the empire on whom attention is riveted: the kings themselves,
o f course, their close w arrior kinsmen and other Telugu nayakas
upon whose military abilities all rested, and upon Brahm an agents
and commanders who, along with the peasant-caste nayakas,
played the m ost crucial political parts in the affairs of state. That
this attention from historians is warranted is simple to dem onstrate
in any o f the four Vijayanagara dynasties. Consider the period of
A chyutaraya’s succession, a time when the empire was at its greatest
strength following the reign of Krishnaraya.
Then, w arrior kinsmen were sources o f significant support to
Achyuta. When his brother, the great Krishnaraya died, A chyuta’s
position was secured against the powerful Aliya Ram araya, a
brother-in-law o f the late king, by two o f A chyuta’s own brothers-
in-law: Pedda and Chinna Salakaraju. The Salakaraju brothers
continued to serve Achyuta as among his most successful and
reliable generals as did another brother-in-law, Cevappa N ayaka.89
And, after the death o f Achyutaraya, in a .d . 1542, one o f the Sala-
karajus m urdered the late king’s nephew and successor, Venkata I.90
Throughout the third dynasty, the record o f m inor rebellions in
complicity with one or several great nayakas is a dismaying chapter
which is usually euphemistically discussed under the heading of
'police arrangem ents’.91 In Tamil country, intrigues am ong war­
riors linked to the royal house by agnatic o r affinal bonds was less
im portant than in the northern parts o f the empire, but they were not
absent. By the late sixteenth century the political arena o f these
Telugu political giants was the entire peninsula. The Brahm an
com m ander and minister Saluva N arasim ha Nayaka, or Sellappa,
who, with the Salakaraju brothers, assured the Vijayanagara throne
to Achyuta in a . d . 1529, was rewarded with control o f Tanjore, the
richest territory in the empire. Sellappa revolted against Achyuta
in A .D . 1531 in alliance with other n ayakas o f the south. Thereasons
for this revolt appear to have been the differences with Aliya
89 Venkataram anayya, S tudies in Third D ynasty, pp. 6, 13, 454.
90 N ilakanta Sastri, A H isto ry o f South India, p. 299.
91 V enkataram anayya, Studies in the T hird D ynasty, pp. 262-5.
R am araya; Sellappa had thwarted R am araya’s ambitions to the
throne at the death of K rishnaraya and was now being made to
pay for that by the still powerful R am araya.92
In explaining the dom inating character of the powerful Telugu
warriors to whom the title nayaka was affixed, two factors appear
most im portant. One was the sheer success of arms of these
warriors; the other is the significant role played by Tamil country
and the acceptance of Telugu rule by its chiefs during the Vijaya­
nagara period.
Reasons for the military success of Vijayanagara warriors against
their H indu and Muslim rivals are hardly considered in the existing
literature on the Vijayanagara state. This is peculiar since all have
differentiated the Vijayanagara state from others on the basis of
its m artial character and achievements. An unchanging dharmic
ideology is presumed to account for the successes o f the several
dynasties; yet, as is clear from the records of Vijayanagara, the
m ajor victims o f Vijayanagara military power were not Muslims
bu t Hindus, and a m ajor factor in this success were Muslim soliders
in Vijayanagara armies. Clearly, other kinds of explanations are
necessary.
One that would appear to deserve serious consideration is that
the success of Vijayanagara armies was a direct consequence of
their experience with and im itation o f Muslim armies, their tactics
and weapons. The founding brothers of the first dynasty had
served in Muslim armies, and it was against Muslim soldiers that
there was interm ittent conflict for two centuries. N or is the military
im portance of the Portuguese, with whom the Vijayanagara rulers
maintained friendly relations, to be ignored.
A second explanation, already briefly m entioned above as
differentiating the process of extending the overlordship of the
Chola and Vijayanagara kings, involved relationships between the
latter and their Telugu agents with the leaders of Vaishnava secta­
rian orders (sam pradayas ). As this relationship was dependent
upon the military capabilities o f Telugu warriors, it is to this
m atter that attention is first given.
Two military factors appear significant: one was the improve­
ment in cavalry warfare and the other, the use o f artillery. Horse
warfare is among the oldest elements o f Indo-Aryan culture and
92 Ibid., pp. 26-33 and Krishnaswam i, op. cit., pp. 200-5.
a well-recognized part o f armies in both the Sanskrit and Tamil
traditions.93 However, the mounted warrior appears to come into
his own in South India in the armies o f Vijayanagara. It has been
suggested th at the Vijayanagara kings were as famed as lo rd s of
the horse’ (asvapati ) as the imperial Gangas of Orissa were famed
as ‘lords of the elephant’ (g a ja p a ti ).94 This suggestion is supported
by the M ahanavam i festival already discussed.
At first glance, the im portance o f horses is puzzling since the
Vijayanagara rulers were completely dependent upon the im porta­
tion o f war-horses o f quality. Considerable notice has been taken
o f this trade in horses from Orm uz and other western Asian trade
centres by foreign com m entators since the time o f M arco Polo.95
According to the Portuguese visitors to Vijayanagara, and Nuniz
was there as a horse trader96 K rishnaraya purchased 13,000
A rabian (Ormuz) horses and country-bred horses each year. The
king kept the best of these for himself.97 Saluva Narasim ha, before
Krishnaraya, is reported to have paid a substantial sum for imported
horses whether dead or alive.98 The establishment of Muslim
powers in the Deccan and the long-standing hostility between them
and the Vijayanagara state must have eliminated or curtailed the
availability of horses bred in northern India or imported from
Central Asia.99 With country-bred horses of poor quality in South
India (in apparent contrast to those o f M aratha country so skil­
fully used by Sivaji and his successors somewhat later) the Vijaya­
nagara state sought and apparently attained a m onopoly of horses
fit for military use.
It must also have been true that military horses were a significant
political currency and source of political control by the Vijaya­
nagara rulers. The supply of strong horses would have extended
the range of effective political control of subordinates, local warriors
o f the time. There would thus be strong inducements on the part
93 M ahalingam , S outh Indian P olity, p. 250; this includes chariot and cavalry
warfare.
94 Sircar, Indian Epigraphy, p. 339 n.; also his ‘Suryavamsi G ajapatis of Orissa',
Indian H istorical Q uarterly, v. 33 (Septem ber 1957), p. 275.
95 The Travels o f M arco Polo, trans. by Ronald Latham , Penguin Books, Har-
m ondsw orth, 1958, p. 237.
96 Sewell, op. cit., p. 235; his ventures were not remunerative, it is noted.
97 According to Nuniz, ibid., pp. 381-2.
98 According to Nuniz, ibid., p. 307.
99 This trade is noted in M arco Polo, op. cit., p. 151.
o f local notables to assure themselves o f a supply of horses and,
especially, a source which remained accessible from one year to
another. Horse breeding and care were notoriously poor according
to foreign com m entators.100
O f course, warriors with strong m ounts constituted both a source
of strength and danger to the Rayas of Vijayanagara. The more
mobile and powerful subordinate chiefs were, the more effective
the army which could be brought to the field against the formidable
armies of Muslim and Hindu enemies. However, the same cavalry
capability could be and was used against the Vijayanagara kings
as the rebellions o f Telugu n ayakas instruct. This latter hazard
could be reduced in several ways. One was for the rulers to m ono­
polize access to superior military horses by paying a high price to
those im porting and trading in horses, even dead horses. A nother
m eans was to establish greater control over the coastal m arkets to
which horses came; this was apparently attem pted by Krishnaraya
and later under the forceful R am araya.101 Finally, mobile strength
could be checked in its long-term effects by strongly fortified garri­
sons of reliable soldiers of the sort to which reference is repeat edly
m ade in the poem, the A m u ktam alyada, attributed to Krishnaraya.
But there is another military factor which is almost totally
ignored by Vijayanagara historians; that is the use of artillery by
the Vijayanagara armies. The earliest experience of artillery by
Vijayanagara armies occurred in a late fourteenth century battle
between Bukka I and the Bahmani Sultan M uham m ad.102 But,
unlike the perfection of cavalry techniques which must have been
learned from Deccani Muslims, it is probable that the use of
artillery was more a consequence of later contact with the P ortu­
guese.103 Doubts have been raised about the proficiency in the
use of artillery by Hindus, as com pared to their Muslim opponents,
but this deficiency was off-set by the use o f Muslim and Portuguese
soldiers by the Rayas. We have the story, well-worn by its demons-
100 M ahalingam , S outh Indian P olity, p. 254. This was true notwithstanding the
attention in sastric literature to care o f horses (see V.R. R am achandra D ikshitar,
W ar in A ncient India, M acmillan & Co., Ltd. M adras 1944), pp. 174-9.
101 These expeditions and their results are still the subject of scholarly controversy
(see Krishnasw am i, op. cit., pp. 212-16, 233-5).
102 M ahalingam , S outh Indian P olity, pp. 262-3; the date of this battle is a . d .
1368.
103 This observation was m ade by R am achandra D ikshitar, op. cit., pp. 105-6.
tration of the religious tolerance o f the Vijayanagara rulers, of
Devaraya II keeping a K oran beside his throne so that his Muslim
soldiers could swear allegiance properly.104 Muslim soldiers served
in Vijayanagara armies from at least the early fifteenth century,
and from Paes and Nuniz a century later, there are descriptions of
the use of artillery as well as muskets and other weapons involving
gunpowder. The Portuguese accounts of the battle of Raichur in
a . d . 1520, record that the Muslim com m ander Salabat Khan used

artillery and his 500 Portuguese mercenaries also used guns. Against
these, K rishnaraya’s soldiers included musketeers, but there is no
reference to artillery. However, among the spoils of the R aya’s
victory were 400 heavy cannons and num erous smaller guns.105
There is little question therefore o f the development o f the use of
artillery and other firearms by Hindu soldiers during this time;
they augmented the firepower of Muslim and Portuguese auxiliaries
in Vijayanagara armies.
Changes in the form o f warfare in South India m ust have contri­
buted to the persistent success of the Vijayanagara rulers against
their H indu rivals. Increasingly effective cavalry and artillery
also explain the strategy of royal fortresses m anned by special
troops and commanded by dependable officers. Dozens of im por­
tant fortresses are mentioned in Vijayanagara inscriptions and in
the m ajor literary sources o f the sixteenth century.106 In inscrip­
tions, the Sanskrit term durga, ‘fort’, also acquired the additional
meaning o f the territory under the influence of a fortress, and the
title which is taken by Vijayanagara historians to mean ‘provincial
governor’ literally means ‘the officer (or chief) over a fort’: durga
dan n aik.l(>1 This designation emphasizes the im portance of forti-

i°4 N ilakanta Sastri, H isto ry o f South India, pp. 259-60.

105 Sewell, op.cit., pp. 327, 342-3. Saletore, op. cit., v. 1, p. 417, cites an anonym ­
ous work entitled, B akhair o f R am a R aya which purports to be a detailed description
of the Vijayanagara arm y at the time of the Battle of Rahshasa-Tangadi. It refers
to artillery pieces, casks o f gunpowder, and gunners. The report is unbelievable
owing to the num bers cited: 2343 ‘great guns’, 99 million casks of powder and 9
trillion bullocks. An earlier discussion o f the B akhair by S. Krishnaswami Aiyangar
m akes no reference to these num bers (A ncient India and South Indian H istory and
Culture, pp. 172-88; also Further Sources o f Vijayanagara H istory, v. 3, p. 207).
106 V enkataram anayya, Studies in the Third D ynasty, pp. 170-1 with special
reference to the R aya va caka o f the 16th C.
107 D annaik is a corrupt form of dandanayaka (Sircar, Indian Epigraphical Glossary,
p. 83).
fied places as well as the fact that the great officials of the Vijaya­
nagara state were not civil, but military officials. This last point
is somewhat surprising when it is noted that most of the durga
dannaiks were Brahm ans according to inscriptions. These referen­
ces thus corroborate the literary evidence of a policy of Krishnaraya
to place reliance on fortresses and to entrust them to Brahmans.
The didactic poem, the A m u ktam alyada, attributed to Krishnaraya,
gives an alm ost equivalent im portance to forts as to Brahm ans;
in fact these two subjects are treated together. Repeatedly, the
(royal?) poet instructs that fortresses are to be strong, well-manned,
and under the control o f B rahm ans.108
It is difficult to resist the tem ptation o f comparing the fortresses
of the Vijayanagara kings with the only other massive structures
o f the age under Brahm an custodianship, Hindu temples. This is
occasionally noted in the historical literature as in the case of the
hill temple of Sim hachalam (Vishakhapatnam district) in m odem
A ndhra-Pradesh.109 D uring the eighteenth century warfare among
English, French, and Muslim forces in what was then called ‘the
C arnatic’, temples were frequently used by all com batants. Orme
reported on the suitability o f temples for this purpose: ‘all, pago­
das on the coast of Coromandel are built on the same general
plan. . . a large area which is commonly square, is enclosed by a
wall 15 or 20 feet high. . . ,’110, and he referred to numerous temples
used as fortifications.111 It would be surprising if during the Vijaya­
nagara period, when the construction of great walled temples
reached full development, they were not used or were not seen as
potentially useful for military purposes by their builders, who
were for the m ost part locality magnates.
Superior military capability based upon cavalry and artillery
as well as strategically placed fortified places under the control
o f reliable troops and Brahm ans, especially Telugu Brahm ans:
these were the principal com ponents o f Vijayanagara authority
!08 A. Rangasvam i Sarasvati, ‘Política! Maxims o f the Em peror-Poet, K rishna­
devaraya’, Journal o f Indian H istory, v. 4 (January 1926), verse 207, p. 65; verse
255, pp. 72-3; verse 270, p. 74.
1()9 K. Sundaram , The Sim hachalam Tem ple Simhachalam [Temple] Devasthanam ,
Sim hachalam, 1969, pp. 135-6.
110 R obert Orme, A H isto ry o f the M ilita ry Transactions o f the British N ation
in Indostan fr o m the Year M D C C X L V , London, 1803, v. 1, p. 117.
111 Ibid., pp. 113, 127, 147, 171, and passim .
in the southern peninsula. Another element which is taken by
scholars to be part of the foundation o f Vijayanagara authority
was the n ayaka system, that is, the functions and relationships
among the several hundred Telugu warriors in control of substantial
territories distributed over the entire South Indian macro region
and presumably responsive to directives o f the Vijayanagara
rulers. Is this a correct understanding of the role of the nayakas'?
As already noticed, the extant historiography on the nayaka
system is at least confusing, even contradictory. The exact or even
approxim ate num ber o f such personages and their territorial
jurisdictions are unknow n; these questions have never been com pre­
hensively investigated. N aya k a s are described by the Portuguese
chroniclers of the sixteenth century, Paes and Nuniz, as agents of
the centralized control of the Rayas. The evidence of Vijayanagara
inscriptions and the later Mackenzie manuscripts present a different
picture; one o f territorial magnates pursuing political ends which,
at times at least, collided with the aims of the Rayas as these may
(indeed must) be inferred. Between these two opposed conceptions
there are other views of the n ayaka system, such as Krishnaswam i’s
recent presentation o f the system as ‘feudalism’. In the work of
N ilakanta Sastri, several views o f nayakas may be discerned. His
1946 publication, Further Sources o f V ijayanagara H isto ry, con­
trasts the Vijayanagara n ayakas before and after the battle of
Rakshasi-Tangadi in a .d . 1565, in the following words: ‘The
nayakas, who were absolutely dependent upon the royal will. . .
[until 1565] acquired a status of semi-independence.’112 Later,
in his H isto ry o f South India, Sastri implies a somewhat less strongly
centralized system before a .d . 1565. In his very brief discussion
o f Vijayanagara political, administrative, and military organization,
he states that in addition to a large standing army supported, in
part, from:

. . . crown lands, annual tributes from feudatories and provincial gover­

nors. . . military fiefs studded the whole length and the breadth of the
empire, each under a nayak or military leader authorized to collect
revenue and administer a specified area provided he maintained an
agreed number of elephants, horses, and troops ever ready to join the
imperial forces in war.113

112 Further Sources o f Vijayanagara H istory, v. 3, p. 299.

113 The 1955 edition pp. 295-6; 1958 edition, subsequent editions, pp. 296-7.
In a yet m ore recent work, published almost ten years after that
cited immediately above, the conception of nayakas as officials
under direct central supervision is drastically altered. Here,
Sastri writes:
V ijayanagar became the focus o f resurgent H indu culture and offered a
m ore successful resistence to Islam in this p a rt o f the country th an any­
w here else. So it was a long m ilitary vigil. As there was no room for
weak or incom petent rulers on the throne, there were revolutions resulting
in a change o f dynasty and renewal o f strength. The em pire is b e s t looked
upon as a m ilita ry confederacy o f m any chieftains cooperating u n d er the
leadership o f the biggest am ong th em .114

Here is a virtual denial o f anything approxim ating the centralized

political system o f which Sastri spoke in his 1946 work, except,
possibly, when an unusual w arrior — a Krishnadevaraya — occu­
pied the throne. Here, too, in this 1964 statement of Sastri, the
threat of Muslim dom ination is given continued saliency as a
reason for the politico-military changes of the Vijayanagara period.
However, Sastri’s most recent form ulation o f the essential char­
acter o f the ‘empire . . . as a military confederation o f many chief­
tains. . . .’ is an extreme position. It is a position which goes too
far in what may be seen as an effort to correct the early conception
o f a centralized p o lity ; it fails to distinguish the differential concerns
and capabilities of intrusive Telugu warriors from those of other
military chieftains, and it therefore fails to appreciate the way in
which the great Telugu warriors dom inate the political scene.
The Vijayanagara period is the age of Telugu military power and
glory. Most o f those possessing substantial military capability
were Telugus, and they comprise a new intermediary level of
authority in what continued to be a segmentary state in South
India.
As to the title, n ayaka, it occurs in K arnataka at least three
centuries before the establishment of the Vijayanagara state, and
it is found in A ndhra country at least two centuries before the
Vijayanagara state. Derrett refers to the term in inscriptions of the
late eleventh century and substitutes the English ‘captain’ implying
a military office.115 However, D errett mentions other evidence,
114 Sources o f Indian H istory, 1964, p. 79. Emphasis added to the original.
115 T he H oysajas, p. 25 with reference to Chalukya inscriptions of a . d . 1062 and
1068 and on p. 188 where he speaks of n a ya ka s as the lowest comm anders of foot
or cavalry.
notably an inscription of the middle twelfth century refers to a
nayaka of Holalkere (in m odem C hitaldrug district, K arnataka)
‘who recognized no overlord’, suggesting not a military office,
but a personage o f local power.116 Telugu literary works and
inscriptions from A ndhra of about the same time also mention
n ayakas in the same ambiguous fashion. K akatiya records of the
middle and late twelfth centuries refer to n a yaka o f specific localities
as dependants to great families o f local dom inance.117 There are
even uri-nayakulu and gram a ndyakulu, that is ‘village nayakas’,
mentioned in A ndhra.118 Later inscriptions o f the Kakatiya
rulers Rudram adevi ( a . d . 1259-95), the queen R udram ba (c.
a . d . 1273), Ambadeva, and Prataparudra ( a . d . 1295-1332) are
interpreted in such a m anner as to place nayakas into a formal
military organization. Thus, there is the ambiguous tradition of
the ‘seventy-five n a ya k a s’’ who died defending the Kakatiya queen
R udram ba’s claim to the throne against Ambadeva, and equally
vague references to a nayankara system.119
The term ‘n ayaka ’ is, in fact, a very ancient Sanskrit one denoting
a person o f prominence and leadership, particularly military
leadership.120 D uring the medieval period, in South India, n ayaka
refers to the bhakti relationship between god (n a ya k a ) and devotee
(n d yik a ). It is a term upon which too m uch has been permitted
to be borne by m odern historians concerned with the analysis
of the political system. When we may point to usage as diverse
as th at cited by D errett and others for warriors o f specific local
dom inance to that o f a G ajapati inscription o f the fifteenth century

'I s Ibid., p. 71.

117 W arfare am ong local m agnates in Palnadu (G untur district) o f the late twelfth
century, according to the P alnativirula Charita, was waged by n ayaka warriors
identified with specific places (e.g. M alyala, Kom aravelH); V enkataram anayya
and Som asekhara Sarm a, ‘K akatiyas o f W arangal’, in Yazdani, op. cit., v. 2, pp.
593, 596.
n s Inscriptions of the late twelfth century, N arasim ha Rao, op. cit., p. 121.
119 Yazdani; op. cit., v. 2, pp. 622, 630-1, 634. The following example casts funda­
m ental doubt upon the existence of som ething called the 'nayankara system ’ at
this time. In the tim e o f Prataparudra, the writers speak o f the king ‘remodelling the
n ayankara system, which appears to have come into vogue during the reigns o f his
predecessors, with a well-equipped army of 900,000 archers [!] besides cavalry and
elephants’, p. 644.
120 M onier-W illiams, A Sanskrit-E nglish D ictionary , p. 536, cites usage in the
M ahabharata.
in which the King, Kapilesvara, is called a n a ya k a ,121 it is necessary
to question the meaning generally ascribed by historians of Vijaya­
nagara: ‘one who holds land from the Vijayanagar kings on condi­
tion o f offering military service’.122 The m ore prudent reading of
the term n ayaka is the generalized designation of a powerful
warrior who was at times associated with the military enterprises
of kings, but who at all times was a territorial magnate in his own
right.
To the extent that it is possible to speak o f a n aya k a system,
this notion has to do with the existence of a new level of intermediary
authority in South India. The powerful com bination o f a technically
superior royal army, and strategically placed fortifications under
Brahm an com m anders constitute one part o f the Vijayanagara
political system. The other part consisted of Telugu nayakas, who,
with astonishing ease, established and m aintained their authority
over m ost o f the southern peninsula, especially Tamil country.
N a ya k a authority in Tamil country certainly hastened or perhaps
even completed the demise of those local institutions which together
provided each locality segment of the Chola state with basic
coherence: the local body of nattars acting corporately through
their territorial assembly, the nadu, or, latterly, co m b in e d w ith o th er
locality bodies in the greater nadu, the periyan adu; brahm adeyas
acting as the ritual and ideological cores o f each locality. That
these several institutions had already begun to lose their im portant
place in Tamil country as early as the twelfth century in some cases
seems clear. Their decline cannot be attributed to the Vijayanagara
state, but m ust be seen as the result of changes in Tamil society
and am ongst the Tamils themselves as discussed in chapter six.
Still, the scope for the politically integrative function o f ritual
authority o f the segmentary state remained. N either a Tamil state
n or a K arnataka state emerged to challenge the Vijayanagara state,
to attract the recognition of indigenous locality chiefs, or to rein­
force the claims o f the latter to their ancient locality control. That
recognition was instead extended to the Telugu nayakas whose
original presence in the m acro region outside A ndhra country
was as military agents of the Vijayanagara kings. These Telugu

121 E .I., v. 33 (1959-60), D.C. Sircar, ‘Two G rants of Raghudeva’, verse 10, p. 4.
122 Cited,in Sircar, Indian Epigraphical Glossary, p. 214, based on the usage of
M ahalingam , A dm inistrative and Social L ife in Vijayanagara.
warriors were not, however, to remain simply agents o f the Vijaya­
nagara kings; they could not because there was no political frame­
work through which an agency of this sort was capable of being
sustained. Telugu n ayakas quickly became locality figures in
their own right, encouraging the settlement o f other Telugus to
strengthen their control over local Tamil and K arnataka chiefs
as well as to buttress their relations with the distant but still inti­
m idating power of the Rayas. W ith respect to the Rayas, Telugu
nayakas continued to express their agency position (kartto).
Thus ensconced, they became a new intermediary level o f authority
within a changed, but nonetheless recognizable, segmentary state.
The Vijayanagara state was left intact and operated through
the sub-stratum o f indigenous chieftainships, as Venkataramanayya
observed. The earliest invasion of Tamil country under K um ara
K am pana resulted first in the defeat of, then the restoration to
authority of Pandyan chiefs in the far south.123 This pattern
persisted throughout the Tamil country. As long as the Tamil
chiefs did not seek to establish kingships which could compete
with that o f Vijayanagara as the legitimate source of ritual sovereign­
ty, it was an arrangem ent which satisfied the requirements o f the
Rayas. Where this was attem pted or threatened, as in the case of
the Bana chieftain Bhuvane Raviran Sam arakolahalan in the
1460s or, later in the century, the Pandyan Jatavarm an Kulasekara
Parakram a, this was treated as rebellion by the Telugu nayakas
of Tamil country with the support of the R ayas.124
In a similar way, the Rayas treated as rebellion and dealt harshly
with the claims of the great Telugu nayakas to the independent
status o f king. The latter assiduously m aintained their cultural
and linguistic distinctiveness in Tam il country and elsewhere they
also continued, on the whole, to claim to be royal agents. Thus,
they stood forth as effective representatives o f Vijayanagara suprem­
acy in the m acro region, and everywhere constituted a powerful
interm ediary level o f rule. As long as these nayakas issued inscrip­
tions as the agents and in the nam e o f the Rayas and as long as
they joined with the Rayas in the latters’ conflicts with Muslim
and H indu kings, these Telugu warriors contributed essential
strength to the Vijayanagara segmentary state. This they, and
123 Heras, A ravidu D yn a sty, pp. 107-8.
124 Krishnaswam i, op. cit., pp. 118-28, 157-60.
Tamil chiefs who also adopted the title o f nay aka, could do as
the territorial magnates they were. And, what is of great im­
portance, this interm ediary role was not a usurpation, an illegiti­
mate appropriation o f authority. Neither Telugu nor Tamil
chiefs were completely, nor even primarily, agents or officials of
the Vijayanagara state. Or if, as in the case of the four commanders
whom K rishnaraya sent ‘to pacify . . . the Nayakas in the Tamil
country’ as Krishnaswami puts it,125 such powerful men might
begin as agents or officials of the Rayas, but they very soon were
transform ed into centres of power independent of the Rayas and
were ultimately transform ed into fully competing 'N a ya k a king­
dom s’. Such transform ations to full independence from the
Rayas and, further, armed opposition were usurpations, however,
and these constituted a fundamental threat jo the Vijayanagara
segmentary state. :
It. was to cope with this basic instability o f relationships between
the Vijayanagara rulers and the new intermediary level o f Telugu
warriors that the significant place in the political system of Telugu
Brahmans must be seen.. Brahmans of the Vijayanagara period,
particularly Telugu Brahmans, had become political men as never
before. In contrast to the Brahman notables o f the Chola period
who were residents of brahm adeyas and custodians of canonical
learning, the great Brahm ans of the Vijayanagara period were men
of the court and adm inistration. It is true that the gh atikds of the
Chola period and earlier were places where Brahmans instructed
others, including other Brahmans in secular, even military, know­
ledge and that, coincident with these activities, made contributions
to the Chola segmentary state in ways already discussed. It is
further true that in personages like the brahm adhirdja are found
Brahmans with administrative careers and considerable political
influence. But, the contrast between the two periods with respect
to the role of Brahmans is nevertheless striking.
Referring to A ndhra during the Vijayanagara period, Venkatara-
mayya states:
T he m ajority o f the educated B rahm ans sought to enter the governm ent
service which offered them bright careers. They were especiaiiy trained to
becom e accountants and adm inistrators. The im perial secretarial was
alm ost entirely m anned by men o f this class. In the Telugu country,

125 Ibid., pp. 193-4.

 they separated from the o th er B rahm ans, and form ed a subcaste know n
as the N iy o g is . T here is reason to believe th at B rahm ans o f this class were
not very orthodox in the observance o f (heir religious rites. They became
m inisters, com m anders o f arm ies, an d governors o f provinces. Every
B rahm an m o th er w ished th at h er son should become a d u r g a d h ip a ti
o r governor o f a fo rt.¡26

The Brahman Vidyaranya and his kinsmen were ministers of the

founders of the Vijayanagara state. They provided not simply
political guidance,127 but a perhaps vital element of legitimacy to
the Sangam a brothers who were, according to the traditions o f the
dynasty, won back to Hinduism after having become Muslims.128 i
But, in keeping with the m artial character of the Vijayanagara
state, it was as military commanders that Brahmans were most
conspicuous. Until the end of Achyutaraya’s reign, Brahmans were
among the great military leaders. G opanaraya and Soman an-
danatha com m anded Vijayanagara forces in the conquest of
Tamil country during the first dynasty,129 and Brahman advisors
and commanders of K rishnaraya’s time were very prom inent:
Saluva Timma, Saluva Govindaraja, Rayasam (or Ayyaparasa),
Kondam arasu, Ram abhatlayya, Sellappa, Saluvayira, Narasim ha,
K aranika M angarasayya, and Bacharasayya.130 These and other
Brahmans were of the Telugu Brahman sub-caste called Niyogi.
As Venkataramanayya pointed out, this was a group which, despite
its prestigious secular activities — or perhaps because of these
activities — may have been forced to constitute an endogamous
group because they could not m aintain the high standards of
orthodoxy of Telugu Brahmans of an earlier time.131
M ost of the durga dannaiks were Brahmans at least until the
reign o f Achyutaraya and possibly after th at time. Their tasks
were to build and hold fortresses in various parts o f the empire,
and the resources for this central element of Vijayanagara role
were realized by income from villages appropriated for the purpose
126 Venkataram anayya, Studies in the T hird D ynasty, pp. 354-6.
127 V idyaranya was not simply the purohita of Bukka I, but a minister to whom
is attributed the Parasvara M adhaviva a text on legal adm inistration (ibid., p. 268).
128 N ilakanta Sastri, A H isto ry o f South India, p. 227.
129 Rangasvami Sarasvati, op. cit.. p. 85.
131) Ibid., Venkataram anayya, S tudies in the Third D ynasty, p. 356, and K rishna­
swami, op. cit., p. 53, and passim .
131 One thinks here o f the K arandai plates o f jRajendra Chola and the large num ber
o f Telugu Apastam ba Brahmans who were brought into the Kaveri, Basin.
by Vijayanagara military forces. Villages whose income was so
committed were distinguished by the term, bhandaravada , a term
taken to differentiate w hat are called ‘crown villages’, from villages
whose income was under the control o f local military persons,
or am aram villages. The preference for Brahmans as custodians
o f forts in the A m u ktam alyada was based upon two factors: Brah­
mans would serve faithfully, meaning, perhaps, that they had no
rule aspirations o f their own, and they would serve efficiently,
meaning they would be in competition with ‘kshatriya and sudra
officers’. Levity, or poetic licence may be attributed to the verse
in the A m u ktam alyada which refers to this striving o f Brahmans
for military accomplishment ‘lest they [Brahmans] should be laugh­
ed at by ksh atriya and sudra officers’.132 However, in a completely
serious way, competition between Brahmans serving the Rayas
and non-Brahm an warriors may be seen as having been essential.
Com parison with the Chola state is again instructive. D uring the
Chola period, Brahm ans became politically im portant as an
im portant ideological element of the segmentary state. Brahmans
then provided the means of replicating in every locality the dharmic
credentials o f rule for which the Cholas themselves created the
model in the Kaveri Basin. Legitimacy for the n a tta r and other
local chiefs depended partly on the exchange nexus between them­
selves and Brahm ans and those brahman-ical institutions found
in brahm adeyas: essentially the exchange of material resources
from the former for royal honours from the latter. The ability
of Brahmans to confer such honours derived from their recognized
ritual supremacy and from their non-competitiveness with establish­
ed secular rulers.
To sustain this exchange dui-ing Chola times, some of the most
valuable productive resources under the control of the nattar
— well-developed, irrigated land in the fertile plains — were
placed under the m anagement of Brahm an sabhas, permanently
and conspicuously alienated from local, peasant control. Such
grants to self-regulating Brahm an settlements did not diminish
the locality control of the n attar; on the contrary, these grants
132 V enkataram anayya, Studies in the T hird D ynasty, p. 154; citing verse 217.
The verse in full is given in Further Sources o f Vijayanagara H istory, v. 3, p. 154.
A king who confers nobility on a Brahm an prospers; for the Brahm an stands at
the post o f duty even at considerable risk, either to avoid the ridicule of the K satriya
and Sudra officers, or in em ulation o f the other Brahm an officers o f the king’s
service.
enhanced their power by validating the claim of the king-like
appropriateness o f n attar control.
The same attributes o f Brahmans which com m ended their
ideological role in Chola times facilitated their political roles in
Vijayanagara times, but with a difference. By the fifteenth century,
gift-giving (dana ) to Brahm ans for their exclusive enjoyment was
no longer a necessary nor perhaps even an appropriate form of
dharm ic activity. Gifts now went to temple deities for the protection
o f all, or gifts were given to sectarian leaders (acarya s , m ath adipatis )
for the guidance of all. Brahm ans as temple servants and as secta­
rian leaders were beneficiaries o f this dharmic largesse to be sure,
but the conception of the Brahman had surely changed. It was
not to individual Brahm ans or to groups of Brahmans for the
preservation of their learning as before (in the form of ekabhogam
or brahm adeya grants), but to Brahmans as integral elements of
the political system of the age as well as of the temple-sectarian
nexus o f b h a k ti Hinduism th at support was given. Arising from
the first were new political roles for Brahmans. These roles were
based, as before, upon the dignity and prestige which their birth
and learning continued to confer and upon their capacity to act
as reliable political agents. One such im portant, new political role
was collaboration with the Rayas (and in time with other great
political personages) in controlling the new stratum o f warrior
intermediaries between the kings and locality chiefs. A prime way
that this collaboration was realized was in the placing of strategically
located fortresses under Brahm an commanders. These Brahman
notables were considered the agents (dan nayakas) of the Rayas
in the territory under the influence o f the fortress. The Brahman-
commanded fortress therefore represented both the military supre­
macy o f the Rayas in a direct physical way and the ritual sovereignty
of the Rayas as an ordinary garrison could not. Arising from the
religious nexus were other new politically significant roles which
are considered below in the discussion of religion and sect during
the Vijayanagara period.
Bhandarvada, or ‘crown’ villages appear to have been the special
source o f income to defray the costs of building and maintaining
strategic fortifications under Brahm an com manders and at the
com m and o f the Rayas. In this sense, the bhandaravada village
replaced the brahm adeya of Chola times and symbolizes, in resource
terms, the changing political role of Brahm ans during the Vijaya-
nagara period. It was a political role for which Brahmans of the
time were uniquely endowed, and it was the most certain way of
controlling the new stratum of warriors of that time.
It is im portant to stress that this w arrior stratum was essentially
the creation o f the Vijayanagara rulers whose incursions into
territories outside the northern heartland of the kingdom were
responsible for establishing the w arrior intermediaries. Incursions
of Vijayanagara armies began with the first Vijayanagara kings
and continued through K rishnaraya’s reign when, according to
accounts in the Mackenzie manuscripts, the commanders of the
Vijayanagara army which invaded Tamil country, became the
nuclei o f the N a y a k a kingdoms o f the next century.133 From the
outset, in fact, these Telugu warriors themselves or their descen­
dants, became potential focuses o f power opposing that o f the
Rayas.
, Up to, and including the time o f K rishnaraya and Achyutaraya,
this use of Brahmans to check the opposition of Telugu soldiery
was successful. After the defeat at Talikota (or Rakshasi-Tangadi,
two villages closer to the actual battleground against Deccani
Muslims) in early 1565, this system broke down in the heartland of
the Vijayanagara state. It continued to operate in Tamil country
for another century, however, at least until the end of the ‘civil w ar’
am ong the great Telugu n ayakas in Tamil country in the first quarter
o f the seventeenth century.
While it is not possible to be certain that all or most of the Brah­
mans involved in imperial politics — as fortress commanders or
in other military and administrative roles —. were Telugus, there
is little question that the principal non-Brahm an military leaders
and many, if not most, o f their followers were men of Andhra
country. It was precisely this movement of Telugu warriors into
Tamil country and their constituting there a new intermediary
level of authority which gives the Vijayanagara segmentary state
one o f its distinctive characteristics. This intermediary ievel of
warriors, strangers in many parts of the empire, may not be seen
as ‘brokers’, to use a fashionable term in contem porary Indian
social science analysis. That is, they do not face in two directions
linking local and supralocal levels o f society and are thus not
valued and. rewarded by both levels.
133 Krishnaswam i, op. cit., p. 194.
 N a ya k a warriors were essentially conquerors who superimposed
their supralocality control over indigenous chiefs and expropriated
a portion of the wealth of the area for their support. They may have
established their locality control as part of an ‘invasion’ such as
that recounted in the Mackenzie m anuscripts for the period of
Krishnaraya, but they remained as supralocal rulers in their own
right, and with a continued identity as outsiders.
The ease with which this control was established over Tamil
localities and the stability o f this control virtually until the middle
of the eighteenth century in most places is as noteworthy as the
failure of these Telugu warriors to maintain stable relationships
with the great Rayas. It is another striking characteristic of the
Vijayanagara state that such serious resistenee for which there are
records do not speak of that of Tamil chiefs or the Tamil population
against the new supralocality élite o f Telugu warriors, but o f the
Telugu warriors against their overlords in the north. Places over
which the Telugu nayakas established their control remained what
they had been in the Chola state: segmentary parts o f a whole.
Most of the time until the end of the last Vijayanagara dynasty,
these localities recognized the Vijayanagara overlordship in
inscriptions. M any of these segments included numerous localities
of an earlier time, especially in the ancient settled territories of
Tanjavur and M adurai; other Vijayanagara segmentary territories,
such as Ginjee in Tamil country, were new nuclei attracting new
settlement from other places; and these three territories, plus
Ikkeri in K arnataka, became the core tracts of the new generation
of segmentary states in South India which followed the fall of the
Vijayanagara state in the seventeenth century - ■'

L a n d S ystem

Elaborate systems of land tenure emerged in most of South Indra

during the m ature period of the Vijayanagara state, that is by the
late fifteenth and sixteenth centuries. The principles of. these
systems and the bases upon which they operated were fundamentally
different from those of the Chola age.
Chola tenurial arrangements continued, even in the twelfth
and thirteenth centuries, to, be based, upon what might be called
‘ethnic territoriality’. That is, land was held or, more precisely,
income from it was realized by various individuals and groups
according to shared ethnicity within an autonom ous locality.
This locality, or nadu, was essentially that tract claimed, settled,
and cultivated by a peasant people possessing a common ethnic
identity, including a shared putative ancestry and, often, history
of m igration, a shared local loyalty, and a shared culture. These
were the n attar, the people o f the nadu. Somewhat separate form
n attar, but still a part of the core ethnic collectivity, were traders
in agricultural goods, the n agarattar, with a large measure o f control
over trade settlements within the locality.134 Similarly, rights
to landed income and substantial control over their settlements
were granted to others, notably Brahmans, by this ethnically defined
core population of a locality. Others who were permitted residential
privileges and a place in the local economy, but no rights to landed
income in most cases, were those who possessed skills and produced
goods which, while valued, were regarded as of an order lower than,
or at least different from agriculture. These non-agricultural,
artisan-traders appear to have been excluded from full membership
in nadu society and were accordingly disadvantaged until the
late eleventh century when references to idangai military forces
suggest an improvement in their fortunes. Others who were
excluded from ethnic m embership and from any rights, apart
from affiliational protection o f other locality groups, were untouch­
able labourers.
The Vijayanagara period affords the first view of a different
land system, one approxim ating that of the e a iy nineteenth cen­
tury. Individuals and, to a greater extent, members of small,
localized corporate groups were the prim ary holders of income
rights in the land during the fifteenth and sixteenth centuries.
In contrast to the well-defined corporate groups by whom land
income was regularly com m anded during the Chola period, those
of the later period were fragments of the earlier n attar groupings
and others in new forms of affiliation and operating within the
much narrower confines of largely autonom ous villages.
Three elements o f the land system o f the Vijayanagara period
serve to differentiate it from that o f the C hola period. First, the
basic unit o f agrarian organization was but a part o f w hat had
134 A ttention is directed to a recent doctoral thesis by Kenneth R. Hall, ‘The
N agaram as a M arketing Center in Early Medieval South India’, unpublished Ph.D
thesis, Dept, o f History, University o f Michigan, 1975.
previously constituted the nadu. This reduced unit m ay be called a
‘village’ in the sense in which that term was used during the British
period .to designate a m ajor agricultural settlement consisting of
one or m ore residential sites, all o f the fields of which were worked
under the direction, if not exclusively with the labour, o f those
residents. A m ajor share of income rights from the land, except
in the case o f some temple villages, perhaps, were similarly restricted
to persons o f the settlement. The older nadu, of which such village
settlements had been part in Chola times, continued to be a kinship
and marriage territory o f im portance for its agricultural peoples;
for them it was also continued as a cult territory. But, land m anage­
ment, that is the unit in which land, labour, and capital were
combined, had dim inished from the nadu to the village. Secondly,
those who m anaged agrarian activities were well-differentiated,
individual ‘big-men’, 135 not the anonymous n a tta r of Chola
times. Such powerful, local men were recruited from the dom inant
agricultural peoples of the immediate locality and often therefore
were of the ancient nattar. However, by the fifteenth century
everywhere in the macro region, there were few localities which
had not been penetrated by ‘outsiders’ as the dom inant agricul­
turists in some or many of its villages. ‘Outsiders’ in Tamil country
might include such diverse groups as Telugu Reddis in villages of
the Palar basin, in Tondaim andalam or the previously shunned
and feared Kallars in the villages of the Kaveri basin. In these
cases, the im portant men of villages (‘p edda ry o ts’’ and ‘m irasidars’
of the eighteenth and nineteenth centuries) might no longer be
Choliyar Vellalas or Tondaim andala Vellalas, but Reddis and
Kallars. But, whether they were of the ancient nattar or outsiders,
these im portant men of the village exercised local control of a new
and different sort from that of the Chola nattar. This is seen in the
third m ajor change of the Vijayanagara period, that is the rural
entrepreneurship o f village big men.
Together, these changes reflected a land system of far greater
complexity than that o f the Chola period, a system in which new
structural features are clear (the new unit o f land management,
135 The notion of ‘big m an’ is m eant to distinguish power derived from wealth
and achievement, not based upon office or birth, but ability; it is explored fully
by M arshall Sahlins in ‘P oor M an, Rich M an, Big-man, Chief: Political Types
in M elanesia and Polynesia’, Com parative Studies in S o ciety an d H istory, v. 5 (1963),
pp. 285-303.
the village; the new leadership of powerful village big men; and
the new activity of entrepreneurship which enriched the new leaders)
and in which these structural elements were combined in such a
way as to free land from the exclusive control o f ancient ethnic
groupings and to free villages from supralocal, territorial (nadu
and periyanadu) control.
These changes m arked the end of the ethnic territoriality of the
Chola period and resulted from a process whose dynamics can
first be located in the m ore advanced parts o f the Tamil country,
the central zones of the Choia segmentary state, during the twelfth
century. In that time and in those places occurred the most im­
pressive development o f temple-centred urbanization, trade, and
the emergence of m ore complex forms of localized society requiring
increasingly adept leadership. Com plicating this process during
and after the fourteenth century was the intrusion into all parts
of the macro region — but especially in the less advanced, in­
termediate and peripheral zones of the older segmentary state —
of Telugu warriors who, in time, came to comprise part of a new
intermediary leadership. With this new leadership came new
groups of cultivators, labourers, and mercantile groups. The new
political and economic m igrants into Tamil country of the four­
teenth century were attracted by its increasing wealth, and the
interests of these migrant peoples were best served by encouraging
the demise of ethnic territoriality in which they must forever be
outsiders. The success of these warriors and their followers in
establishing stable relationships with local Tamil and K arnatak
chiefs and peasant peoples during the fifteenth century, is the
most persuasive evidence that the earlier, forms o f ethnic organi­
zation had all but disappeared: Representatives of the ancient
n attar of Tamil country had undergone substantial changes in
their fortunes. Some few had risen to positions of chieftainship
and instituted largely endogamous lineages of chiefs (as seen in
the p a tta k k a ra r o f the K ongu Vellalas136) within older nattar
communities. Most other ancient, dom inant landed groups in
Tamil country and others like the okkulu of K arnataka ■were
content or compelled to reduce the scope of their control over
| !<> Noticed in a record o f the time o f M aiikarjuna. A .R .E . 1920, no.235 and in
one dated a . d . 1622 (A .R .E . 1920, no.239) and recently noted in the ethnographic
analysis by Brenda E.F. Beck, Peasant S o ciety in K onku. University o f British
Columbia Press, Vancouver. 1972, pp. 40 ff.
agrarian resources to quite atomized villages, it was left to the
congenia), supralocal authority of the new leadership of the Vijaya­
nagara segmentary state to safeguard and even to advance essential
forms of supralocal cooperation, including military protection,
the m aintenance o f sacred places, and the adjudication of con­
flicts which could not be resolved within village society. These
changes in agrarian organization are found reflected in the more
complex variety of land tenure found during the Vijayanagara
period as com pared to any previous time.
Tenurial forms during the Chola period were quite simple.
Ordinary land tenure was called, vellón vagai (literally, ‘cultivators'
share’) by which can only have been meant the shares of income
com m anded by the dom inant peasant group of a locality. Thus,
Subbarayalu states: ‘the N a tta r were none other than the repre­
sentatives of the Ve.llanvagai villages '. i37 This tenure accounted
for ‘the general class of villages’ according to A ppadorai138 and
‘the norm al type of tax-paying village’ according to K.A. Nilakanta
Sastri.139 In addition, there were special tenures pertaining almost
exclusively to eleemosynary grants of village income and, for
h rahm adeyas , a high degree o f autonom ous governance. Chola
copperplates and stone inscriptions which record grants to Brah­
mans of income from cultivated fields along with considerable
control over the government o f their settlements, it has been
argued here, represent special arrangements, the very uniqueness
of which required the detailed specificity found in such documents.
A part from these arrangements for groups of Brahmans, some
income from land, as well as from trade and other sources, was
granted to canonical temples as devadana. This was a relatively
m inor feature of the pre-fourteenth century, and support of non-
canonical temples or to non-Brahman ritual specialists, which was
continuous, is alm ost never referred to in inscriptions. Some in­
come from cultivated lands was deployed to support military
garrisons in the Chola country, but rarely elsewhere. In over­
whelming part during the Chola period cultivable land and the
income generated from it was under the control o f the nattar
to use and distribute as they saw fit. Nor is there evidence of
tenurial changes until the Vijayanagara period.
137 Subbarayalu, op. cit., pp. 34 if.
138 Appadorai, op. cit., v. 1, p. 154. *
139 The Colas, p. 571.
 It is usual for Vijayanagara historians to speak of three major
tenurial categories. These are: am ara, bhandaravada, and m anya.
In fact, these categories pertain to only one aspect of land tenure
during the Vijayanagara period, namely, a classification of the
disposition of the m ajor portion of the proceeds of agricultural
production on a village basis. Accordingly, the fewest villages
were bhandaravada or ‘crow n’ villages, some portion of whose
income went to support Vijayanagara fortresses in various parts
of the empire; a larger num ber of villages contributed to the
support o f Brahmans, temples, m athas as m anya villages; and the
greatest num ber o f villages — estim ated to comprise three-quar-
ters of all villages — yielded a portion of their income to those
with local dominance. The latter were designated as am aranayakas,
hence the tenurial category a m a ra m M 0
There are two difficulties with this classification. The first is
the contention that the Vijayanagara ruler was owner of the soil
and thus that the holders of m anya and am ara rights were ‘lessees’
o f conditional tenements. It is supposed that holders of these
rights were required, as a condition of holding the right, to perform
certain specified services (e.g., ritual or m ilitary); these holders
were also supposedly subject to unspecified restrictions regarding
heritability, transfer, or sale. Such a position is simply untenable
in the light o f inscriptional evidence of all sorts of transactions
involving land and in the light of much of the best scholarly in­
terpretation of this evidence. The second difficulty in this classi­
fication of land tenure into ‘crown villages’, ‘military service
villages’, and ‘beneficial villages’ is that there is no way of deter­
mining how many villages can be classified under any of these
tenure classes. The idea that seventy-five per cent of all villages
were military service villages is in no sense an estimate; it is, in
fact, an assertion, unfortified by contem porary evidence, that all
villages not designated ‘crown villages’ or alienated for religious
purposes — and there is no way to estimate those numbers —
were under presumed ‘feudal’ tenure.
N otw ithstanding these difficulties, however, it is im portant to
recognize that this classification system is based upon something
140 Krishnasw am i, op. cit., p. 180: ‘Nearly three fourths o f the land in Tam il
country was held by m ilitary chiefs on this tenure’. The term am aram is taken to
refer to military service tenure from the Sanskrit-derived Tamil word am ara (San­
skrit: sam ara, M onier-W illiam s, A Sanskrit-E nglish D ictionary, p. 1170).
real. That is the saliency of the village unit in references to land-
holding during the Vijayanagara period. One again notes the
contrast with Chola arrangements in which lands from a great
many villages were contributed to form a new brahm adeya settle­
ment. In the founding of the Anaimangalam brahm adeya re­
corded in the ‘Larger Leiden Plates’ of R ajendra C hola I, twenty-six
villages gave portions of their lands to the new settlement; the
K arandai Plates of the same ruler refer to fifty-seven contributing
villages.141 In the Tamil portion of the latter plates, these fifty-
seven villages are said to have been removed from the status of
veUan vagai and the rights (k a n iy-u d a iya r : literally ‘right to posses­
sion o f cultivable land’)142 o f the cultivators (k u d i ) cancelled.
Arrangements of the Vijayanagara period are very different; not
only are there no longer records of the establishment of settlements
like the great brahm adeyas of the Chola period, but there were
few transactions which refer to lands outside a single village. At
this time, the village has become the effective unit for land arrange­
ments. And, such arrangements include m ore than the ultimate
disposal o f income transfers beyond the village proper; they
pertain to that nexus of resource shares which a village had come
to be.
Vijayanagara inscriptions are concerned with new and public
claims upon shares of village income. To be stressed here is that
the ‘rights in land’ always refer to shares of income, not ‘dominion
in land’, and that in many cases such shares have always existed
but were not before given the public and formal status achieved
in the fifteenth and sixteenth centuries. Another point of note is
that the unit in which these income shares are mentioned is in­
variably that of a nam ed village.
Vijayanagara records delineate categories of income shares based
on the rights of the diverse groupings of which a village consisted."
The Tamil term am aram akani is understood as an ‘estate’ by
m ost Vijayanagara writers143 or, in N ilakanta Sastri’s usage, a
‘fie f.144 The particle, m akani in fact, denotes a fraction, one-
141 The ‘Larger Leiden Plates’ are discussed in E .I., v. 22 and on this m atter,
pp. 237-8. The "Karandai Plates’ are discussed in a typescript provided by the
Governm ent epigraphist, K .G . Krishnan.
142 S .I.T .I., ‘G lossary’, pp. xxiii, lxxxi.
143 V enkataram anayya, S tudies in the T hird D ynasty, p. 179; Krishnaswam i,
op. cit., p. 180; S .I .T .I ., ‘G lossary’, p. v (a m aram agani).
144 A H isto ry o f S outh India (1958 ed.), p. 297.
six teen th ,145 and is best u n d ersto o d as a conventionalized share
o f agricultural incom e or p ro d u ctio n paid by those in control o f
agrarian p ro d u ctio n to those with local political au thority (i.e.,
a mar a).
A m aram tenure o f the V ijayanagara period was m ilitary tenure
as m ost h istorians insist. Il was the control over land and its
p ro d u cts possessed by those w ith m ilitary ■capability. In this
lim ited sense, Sircar's .contention th at amaram tenure is am ong
the few exceptions to the total absence in early India o f anything
like ‘feudal ten u re’ m ay be accep ted .146 This alone can scarcely
justify feudal usage, and the rejection by Sircar o f the concept
o f feudalism has been supported in this discussion. But, the
analysis o f amaram tenure here cannot be m ade to fit the category
o f ia n d - lo r d ’ which Sircar insists is superior to the concept o f
‘feudalism ’ in early India. A m ajo r reason for this is th at the
concept o f ‘la n d lo rd ’ has n o t been defined by Sircar, not has the
encom passing legal an d political stru ctu re in which these ‘lan d ­
lo rd s’ w ere supposed to have functioned ever been clarified.147
N eith er ‘feudalism ’ n o r ‘la n d lo rd ism ’ seize the essential character
145 Tam il L exicon, v. 5, p. 3143; the one-sixteenth fraction is the sum o f the fraction
m u ' one-lw entieth, and kuni, one-eightieth; also. op. cit., v. 2, p. 859. Also, the
reference to ‘m a ; k a : n i' in S. Agesthiaiingam and S.V. Shanmugam, The Language
o f T am il Inscriptions 1250-1350 A .D ., A nnam alai University, Annam alainagar,
1970, p. 191.
146 D.C. Sircar. L and S y s te m and Feudalism, p. 57.
147 Sircar’s "land’ appeared to be created by the establishment of a 'free-holding’
as a result of ’relegating the king’s rights over the village to the donee (by a royal
charter)’, ibid., p. 20. The term "free-hold’ is old since such alienations according
to Sircar were sometimes completely free of taxes and obligations, were some limes
partly free of these, and were, at times, not exempted at all from obligations to the
king. Furtherm ore, the enum eration of twenty-one ‘free-hold’ rights on p. 21 con­
sist of perm utations of the conventional asta-hhogam (eight kinds o f benefits or
enjoyments of possession) which are not proof of a la n d lo rd ’ tenem ent or right.
Finally, p. 14 contains a confusing discussion of the role of the king which is crucial
in Sircar’s conception. In this discussion, the king is regarded as the ultim ate owner
of land and, he continues, ‘The royal charters . . . say that the free-holding was
created by the king at the request of a subordinate whose name was mentioned
in the (epigraphical) docum ent only when he was of sufficient importance. When
he (the subordinate) became more powerful, he would himself issue charters with
the king’s permission, and, when still m ore powerful, his charters would not even
m ention the name of the overlord. The next stage would of course be represented
by his [the subordinate’s] charters issued as an independent m onarch. Thus, a land­
lord became a king.’ !
o f amaranoyakas as highly local and' q u ite independent w arrior
chiefs. It is simply not the case th at amaranayakas were officials
o f the V ijayanagara governm ent. T here were to be sure great
m ilitary co m m anders — nayaka & w ho w ere also the m o st im ­
p o rtan t política! m overs of the fifteenth an d sixteenth centuries;
b u t they were few. T hey were also Telugu com m anders o f Vijay­
an ag ara arm ies w hose pow er derived from th eir m ilitary offices.
Secondarily, the pow er o f these Telugu grandees derived from
territorial bases in T am il co u n try ; these territories were the founda­
tion for the new generation o f segm entary states after the decline
o f the V ijayanagara state in the late sixteenth century. But the
nayaka title was also used by locality chiefs along w ith older,
local an d v ernacular form s o f address.
Amar am tenure consisted essentially o f th e pow er to distribute
o r redistribute the proceedings o f agrarian p ro d uction and to set
and collect those charges in kind or in cash to be paid to personages
w ith local au th o rity . These are pow ers enum erated by the m ost
able student o f m odern South Indian land tenure, S undararaja
Iyengar; to these pow ers, he no tes, the term amarakam is attached
during th e nineteenth a n d tw entieth cen tu ries.148 T he term amara
does indeed refer to w ar a n d som e amaranayakas w ere great
m ilitary com m anders. But, the term amaram, m ore broadly,
denotes the rights and pow ers of local m agnates.
A n o th e r im p o rtan t an d publicly acknow ledged category of
incom e shares is that paid to ‘village servants’ u pon w hom m uch
attentio n was lavished in early British records. A ccording to the
scholarly literatu re dealing w ith th e m edieval situation, those
w ho perform ed services for the agricultural p o p u lation o f a village
or a locality were called by the term ayagar in m any parts o f the
m acro region. This term m eans ‘those possessing incom e’ (ayam
in T am il).149 A ccording to the established h istoriography, áyagan ,
the body o f village ■servants, displaced village assemblies of the
C hola period [sabha a n d ur) as th e local m anagem ent institutions.
Thus, we have K rishnasw am i’s statem ent th a t: ‘. . . the in tro d u c­
tion o f the N áy an k ara Á yagar systems in the provincial and
local spheres by the V ijayanagar rulers b rought about the decline
and disappearance o f the local institutions in the Tam il co u n try ’.150
l4S Op. cit., p. 83.
149 S .I .T .I ., ‘G lossary’, p. x.
150 Krishnaswam i, op. cit.. p. 103.
 Village servants, or ayagars, consisted o f a num ber o f func­
tionaries of whom three are considered by Krishnaswam i151 and
others not to have existed formerly. These included the headman
(e.g. m an iyam reddi, or gauda ), accountant (e.g. karnam , senabho-
va), and watchman (e.g. ta la iya ri). To holders of each o f these
offices, a portion o f village income was allocated in the form of
rights in particular plots of village land. These plots were not
liable for regular tax payments, hence they were regarded as
m anya, or tax free152 though holders of such rights might be subject
to a fixed quit-rent paym ent.153 The same form of payment went
to other ‘village servants’ who possessed no governance or m anage­
m ent functions within a village, but did provide services essential
to the village community. Among these might be providers of
ritual services, such as washerm an and priest. For the m ost part
however, village servants were providers of goods and services
of a m ore ordinary sort. Among the latter would be the leather-
worker whose products included the leather bag used in lift ir­
rigation devices (m hote or kapila), potter, blacksmith, carpenter,
waterm an (n iran ikkar ) who controlled and m aintained irrigation
channels, bankers and money-lenders. In the nineteenth century,
these rights to income shares by ‘village servants’ were the subject
of intensive debate and some clarification in connection with
m irasi and inam rights. D uring the Vijayanagara period such
Persian words were not known or used; instead the Dravidian
or Sanskrit term s um bali, kodage, sro triy a are found. The meanings
appear to be the same. They refer to rights of income from agri­
cultural production which were exempt from the customary dues
on agricultural income in lieu of direct payments for services.
In those rare cases in which there were direct payments for services,
payments in kind were usually designated, dan yadaya ,154 those
in money, suvarnadaya or kasu k a d a m a i.155
W hat is distinctive about the ayagar system is not that the
complement o f skills and services existed in villages o f the Vijaya­
nagara period, but that special allotments of income shares from
151 Krishnaswam i, op. cit., p. 104; V enkataram anayya, Studies, pp. 160-8.
152 T am il L exicon, v. 5, p. 3044.
153 V enkataram anayya, Studies in the T hird D ynasty, p. 163.
154 S .I .T .I ., ‘G lossary’, p. xii.
155 V enkataram anayya, Studies in the Third D ynasty, p. 199 and S .I.T .I.,
‘G lossary’, p. xxvi.
 land and specified cash paym ents were for the first tim e generally
provided to those holding these ‘offices’. Village ‘adm inistrative’
offices existed in villages o f the C hola period as we know from some
o f the m ost elegant records o f th a t tim e.156 Sim ilarly, there were
carpenters, sm iths, persons responsible for the regulation and
m aintenance o f irrig atio n channels and village ta n k s providing
services vital to the C hola agricultural econom y ju st as there were
priests and w asherm en required for ritual purity. H owever,
livelihood shares o f village incom e for these services are first
m entioned in V ijayanagara records. It can only be assum ed th at
paym ents for such services in pre-V ijayanagara tim es were provided
inform ally by the nattar p atro n s o f the village an d locality to these
specialist clients. W hether the tran sfo rm atio n o f such paym ents
from inform al, essentially, patron-client ones, to public provision
o f specified incom e shares resulted from the dem and o f village
servants, w hose b argaining position m ay have im proved with the
decline in pow er o f the nadu, or the convenience o f village p atrons
cannot be judged. The im plications of either origin o f the change
are interesting, b u t the shift to this m ore form al, tenurial m ode of
paym ent underscores the tran sfo rm atio n from the anonym ous,
organic agrarian relationships o f the C hola period to the m ore
stratified and com plex relationships later.
Finally, there continued durin g the V ijayanagara period, as
earlier, to be paym ents from agricultural com m unities to those
w ho provided the highest o rd er o f ritual services. The form of
such paym ents is again some share o f village incom e in which the
recipients m ay have been liable to a small, fixed cess (later called
jodi) but w hich could also have been exem pt from all dues. These
continued to be called manya rights by an d large and were held by
B rahm ans as individuals or as m em bers o f c o rp o rate bodies specified
in term s like brahmadeya, devadaya, and mathapura .157
A distinctive category o f V ijayanagara land tenures related
to w hat m ight be called ‘ru ral developm ental entrepreneurship’,
in T am il co u n try an d A n d h ra, these tenures w ere called dasavanda
rig h ts; in K a rn a ta k a , they w ere called kattu-kodage rights. These
w ere special, private rights to a share o f the new productivity
i56The 'L arger Leiden Plates’ o f Rajendra 1 list such offices. See E .I., v. 22,
no.34, pp. 233. 237-8.
i-'7 V enkataram anayya, Studies in the Third D ynasty, pp. 180-4. The general
term um baii was also used for payments to Brahmans.
created by an investment in irrigation improvement in existing
agricultural villages.
A m aranayakas were among the most active rural entrepreneurs.
Perhaps these powerful men were seeking the maximum, protected
return to their descendants on resources they com m anded in many
parts of the macro region; a kind of insurance against the vagaries
of political fortune. In agrarian tracts where productivity could be
increased by a relatively m odest new construction of, or improve­
ment in, irrigation facilities, there was scope for agrarian entre-
prenuership or, as it might be termed, developmental activity.
Such activity seems to have occurred less in the regions of reliable
irrigation potential such as in the deltaic tracts of the Kistna-
G odavari or Kaveri, where irrigation construction and m aintenance
were well-established locality o r village functions with cesses and
personnel provided from quite early times.158 In tracts with little
or no potential for this scale of water management, such as the
dry upland tracts o f Rayalaseema or Pudukkottai, there is also
little evidence of this kind of activity; here, irrigation development
had to await the large-scale works o f the nineteenth and twentieth
centuries. It was rather in those places where hydrographic and
topographic features were such as to provide the basis for consider­
able productivity and settlement, but provided still greater scope
for development through investment in small-scale irrigation
works that this feature of the Vijayanagara period is m ost evident.
‘Developmental investment’ was undertaken by individuals of
means and local prominence in return for which they secured income
rights to a portion o f the enhanced productivity as personal,
heritable, and transferable property. This is the meaning o f the
tenurial forms called k a ttu -k o d a g e in K arnataka and clasabhan-
dam or dasavanda in Tamil country. U nder these arrangements,
it was usually stipulated that a share of increased product from the
construction of a tank or channel was to go to the cultivators of
the village in which the construction was carried out and a smaller
share was to be granted to the person who financed or otherwise
executed the construction.159 The share to the latter appears often
>58 Chola terms such as cen-nTr-vetti, refer to labour obligations for the purpose
of dredging irrigation channels, and m r-vllai , to a water cess, are among the most
frequently occurring terms found in the Kaveri basin according to Karashim a and
Sitaram an, op. cit., p. 91.
159 M ahalingam , E conom ic L ife in Vijayanagar. p. 52. K arnataka inscriptions
to have been about one-quarter. An inscription from Pelleru
village, A tm akur taluk, Nellore, dated a . d . 1622, involves a village
held on am uram tenure under the following specified arrangem ent:
. . . the term s for providing channels for the people this year are: the
produce raised on dry fields should be divided into four shares, of which
three should go to the ryot an d one to the estate every year. The grain
raised under the tanks should be divided into three shares, o f which two
should go to the ryot and one to the estate every year. In this m anner
w hoever is the ruler should cause to be d o n e .I6o

An earlier record from Mulbagal taluk, K olar district, K arnataka,

a part of the Gangavadi territory, provides elaborate description
of the dasavanda right. In a .d . 1496, a person constructed a tank
in a temple (d evadan a ) village under an arrangement with the
head priest of the N arasim ha temple. It was stipulated that the
builder of the tank would be entitled to three-tenths of the produce
from the rice land watered by the tank and, in addition to this, the
builder was granted income shares in dry cultivation of ragi.
From these shares, it was provided that the builder of the tank
would be responsible for repairing and m aintaining it under the
penalty of being liable to a special payment to the temple should
he fail to do so. It was m oreover provided that if additional irriga­
tion facilities were created in the village by this person, the same
arrangement would apply, and that any groves o f coconut or areca
or any permanent gardens were planted, tank water could be
utilized as on the rice lands. As a final provision, it was stated
that irrigation water for the plots charged with meeting the dasa­
vanda shares could only come after temple lands had been
watered.161
Some large temples which held income shares in villages as

referring to k a ttu kadage grants may be found, among other places, in E .C ., v. 10.
Inscriptions in the K olar D istrict, K olar taluk, no. 207, A .D . 1661 ; no. 219, a . d . 1663;
no. 220. a . d . 1628; no. 227, a . d . 1655 (?); Mulbagal taluk: no. 131. a . d . ¡407; no. 132,
a . d . 1494. An example of dasavanda in Andhra is the Telugu inscription from
K ondam arripaili, Kalahasti taluk, Chittoor, dated a . d . 1592 in Inscriptions o f
A ndhradesa, ed. M. Ram a Rao, v. 2, pt 1, Sri Venkateswara University Press, Tiru-
pati, 1969, no. 891, C hittoor District.
160 Butterworth and Venugopal Chetty, Nellore Inscriptions, v. 1, p. 264. Also
noted, I M P , v. 2, p. 1055. ‘E state’ appears to be the translation of am aram akani.
1(51 E .C ., v. 10, Mulbagal, n o . 172. The next record, n o . 173, dated a . d . 1503 per­
tains to the same person and temple providing for the share-cropping of the temple
lands which the form er undertook on approxim ately half-shares.
devadana m aintained irrigation works departm ents for the precise
purpose of utilizing money endowments for productive irrigation
works. During the Vijayanagara period ail temples received money
gifts and some received large am ounts. D onors of money were
often designated as recipients of a share of the food offerings made
to the god (prasadam ) as a part of the b h a k ti temple ritual of the
age. Entitlement to a share of prasadam , apart from being an
im portant temple honour (m ar iy a da i),162 was materially valuable.
P rasadam was sold, as it is to this day, by those who were entitled
to receive it, to pilgrims to be carried back to their homes by the
latter and shared with others in the same manner, and with the
same auspicious effect, as G anga water was and still is. The donor’s
share at the temple o f Venkatesvara, the premier temple o f the
later Vijayanagara period at Tirupati, was one-fourth of the food­
stuffs resulting from a particular investment in a new irrigation
facility in a temple village. A typical Tirupati inscription with
this provision may be cited. The Tamil record is dated 1 October,
A.D. 1536.163

H ail, Prosperity! This is the silasasanam executed by the S th an attar

(trustees) of Tirum alai tem ple in favour o f K oneri, son o f Sellan residing
at Palaverkadu village in the Saka year 1458 while Srim an M aharaja-
dhiraja R ajaparam esvara Sri VTrapratapa Sri Vira A chyutaraya M aha-
raya was ruling the kingdom , to wit.
Since you have paid the sum o f 3,200 n ar-p an am into the tem ple trea­
sury for the purpose o f p ro pitiating Sri V enkatesa with 2 firupponakam
[food offerings] daily as your ubh&iyam [donation] — we shall utilize
this sum o f 3,200 panam for the im provem ent o f the tanks and channels
in the tem ple villages and with the incom e obtained thereby, shall be
supplied from the tem ple-store for the preparation of 2 vellai-tiruppona-
kam , 2 m arakkal o f rice, m easured with the T irum alai tem ple measure,
1 u lakk u o f ghee, 1 u lakku o f green gram , salt, pepper, vegetables and
curds.
Y ou are hereby authorized to receive the q u arter share o f the offered
prasadam due to the donor. The rem aining prasadam shall be reserved
for distribution during early adaippu.

162 The work in progress of Carol Breckenridge on the m atter of temple honour
is very im portant here: the author is indebted to her and to Arjun A ppadurai on
the point. See their joint paper, 'The South Indian Temple: Redistribution, Honors
and A uthority', Contributions to Indian Sociology (New series: December, 1976).
163 T irum alai-T irupati D evasthanam Epigraphicai S eriesj hereafter: T .T .D .I., v. 4,
Inscriptions o f A ch yu ta ra ya 's Tim e (fro m 1530 A .D . to ¡542 A .D .j. Tirumalai-
Tirupati Devasthanam Press, M adurai, 1936, no.75, pp. 144-45.
 This arrangem ent shall continue to be in force thro u gh o u t the success­
ion o f your heirs, till the m oon and the sun shine.
In this m anner this deed is draw n up by the tem ple-accountant, Tiru-
ninga-udaiyan w ith the consent o f the Srivaishnavas. M ay these the
Srivaishnavas protect!

While this is a typical record of the late fifteenth and early

sixteenth century temple complex of the hill shrine at Tirumalai
and the shrines at the base of the hill at Tirupali, this temple complex
was not typical. It was perhaps the greatest temple in South India.
Indeed, before it reached the peak of its fame in the sixteenth cen­
tury, 153 festival days were celebrated there, accounting for its
enorm ous attraction to pilgrims.164 Still, other large temples
m aintained a system o f productive investment o f endowment
funds in temple lands; and there were at the time few more secure
ways for temples to meet the responsibilities they incurred in accept­
ing funds for perpetual ritual services and there were equally few
ways in which those persons o f wealth, capable o f m aking such
endowments, could assure to themselves the prestige of temple
honours and a reliable return on some portion of their wealth
then could be realized through such things as the sale of p ra sa d a m .165
G rants of income from land to temples by these and other donors
had the further advantage o f placing some portion ,o f their land
under temple protection, yet another form of insurance.
Inscriptions from the shrines at Tirupati and Tirum alai afford
an unparalleled opportunity to take a measure of the complex
variety of persons who commanded land and m oney.166 A bout one
thousand stone inscriptions ranging from the ninth to the seven­
teenth centuries have been published by the Devasthanam of that
temple. Most of these date from the period a .d . 1450 to a . d . 1550,
a time when this sacred complex enjoyed substantial patronage
164 T .K .T . Viraraghavacharya, H istory o f Tirupati (T h e Tiruvengadam T em p le),
Tirum alai-Tirupati Devasthanam s, T irupati, 1954. v. 2, pp. 565-6.
105 For a more elaborate description o f these matters see: B. Stein, ‘The Economic
Function of a Medieval South Indian Tem ple’. Journal o f A sian Studies, v. 19
(February, 1960), pp. 163-76 and B. Stein, ‘The State, the Temple and Agricultural
Developm ent; A Study in Medieval South India’, The E conom ic W eekly A nnual,
Bombay, February 1961, pp. 179-87.
166 This discussion is based upon the au th o r’s unpublished doctoral thesis, en­
titled, ‘The Tirupati Tem ple: An Economic Study of a Medieval South Indian
Tem ple', University of Chicago, December, 1958. Published portions of this research
may be found in Stein, ‘The Econom ic Function of a Medieval South Indian Tem ple’.
from the rulers and officials of the Vijayanagara state. Between
a . d . 1456 and 1570, over one hundred villages and large sums of

money were granted to these shrines by some three hundred donors.

C oncentrating upon the m ajor p art of the sixteenth century, the
analysis o f these T irupati records provides im portant inform ation
on as well as raising questions about donorship and its increasingly
diverse sources. These records are, therefore, essential evidence
on social and economic complexity during the great days o f the
Vijayanagara state. Excluding the gifts to the temple from Vijaya­
nagara royal families, the following breakdown of donors and
endowments o f temple villages during the sixteenth century by
Vijayanagara regnal periods167 and tenurial type and num ber of
villages granted can be seen.

T able VIII-1
Village Endowments to the Tirupati Temple and Types
of Tenure by D onor Groups, 1509-68a

D onor G roup Type o f Tenure 1509-30 1530-42 1542-68 Total

A. State donors. Crown and Service

tenure.b 12 10 401 h 62-/2
B. Temple Eleemosynary
Functionaries Tenure 5 19‘/2 18i/, 43
C. Local Resi­ Some service,
dents and some peasant
M erchants. proprietor tenure. v2 6 3 9 '/ 2

Total 17»/2 351/, 62 115

a Based on 312 T irupati inscriptions. Miscellaneous land grants are not included,
b Differentiation of villages which m ight have been under bhandarvada tenure
from those under am aram tenure is not possible from the data. Theoretically,
grantors of both types would have required the assent of superordinate authorities
for such alienations, though there is no evidence that such assent was sought or
actually necessary.

F or this same period, using the same broad classification of donors,

money endowments entrusted to the stan atta r of the various shrines
at Tirupati and Tirum alai were as follows:
167 The period a . d . 1509-30, K rishnadevaraya; a .d . 1530-42, A ch^utadevaraya;
a .d. 1542-68, Sadasivaraya.
 Vijayanagara S ta te an d S o cie ty 431

T a b l e V III- 2

M onetary Endowments to the Tirupati Temple, 1509-68

1509-30 1530-42 1542-68

Value % of Value % of Value % of

Donors (panam ) Total (panam) Total (panam) T otal

A. STATE DO N O RS 33.0 65.0 20.5

Viceroy 30,675 6.5
Chief M inister 1,200 1.0
Com m ander-in-Chief 15,000 3.0
Generals 18.980 12.0 145,200 30.5 4,260 2.5
Royal Officers 19,990 13.0 11,010 2.0 1,580 1.0
Subordinate and
T ributary Officers 11,320 7.0 106,820 23.0 32,840 17.0

B. T EM PL E F U N C ­
TIO N A R IE S 26.0 24.0 23.5
Temple Priests 28,215 18.0 66,963 15.0 22,482 12.0
Musicians, poets
dancers 2,500 1.5 30,480 6.0 6,340 4.0
Scholars 2,520 1.5 4,185 1.0 5,747 3.0
Temple A ccountants 7,446 5.0 8,270 2.0 7,802 4.5

C. LOCAL RESID EN TS
AN D MERCHANTS 41.0 11.0 56.0
Citizens and M er­
chants o f T irupati 25,625 23.0 41,695 9.0 54,405 28.0
Private Devotees 27,560 18.0 10,293 2.0 54,150 28.0

Totals !145,356 100.0 470,591 ’ 100.0 189,606 100.0

100.0 100.0 100.0

The pattern o f donorship displayed in these two tables may appear

unexceptionable in all but one respect, that is the conspicuous
place of donors identified as temple priests and other devout
persons who resided around the precincts of the Tirupati/Tirum alai
temples. The fact that a variety o f im portant political personages
— some closely linked to the Rayas as officials and many other
men of seemingly independent prominence and power — con­
tributed substantially to the land and m onetary resources of the
shrines is, of course, unsurprising. These were, after all, precisely
 the persons with control over the distribution of village income
and changes in that distribution as well as those with the dharmic
responsibility to use their resources in this way. Somewhat more
interesting is the donative contribution of local residents and
m erchants o f the locality of Tirupati. The money endowments
of these people during the troubled reign of Sadasivaraya, a . d .
1542-68, constituted over one half of all of the money given in
that quarter century.
However, as noted above, it is the substantial level of money
and village endowments by religious leaders that must be considered
as most interesting. Almost forty per cent of all of the villages
granted to the temple and about twenty-five per cent of money
endowments came from those with religious titles and, as may be
judged from the inscriptions of the temple, im portant ritual func­
tions. This is a paradox which is scarcely dissolved by the observa­
tion that many, probably even most, villages were originally granted
to these religious personages by the category of donors labelled
‘state donors’, for it is not certain what is to be understood by this
complication in gifting. W hat were the advantages of or the needs
for the political donor (a chief or military official), or the priestly
or otherwise religiously m arked donor (temple priest or m athadi-
p a ti ), or the temple managers, possessing a village or some money,
to pass through the middle group of high ritual functionaries or
heads o f m ath a s on the way from the original donor to the temple?
This question will be reconsidered below, in the section on
‘temple and sect’; it will be sufficient here merely to assert that the
various kinds of religious leaders associated with this temple,
and perhaps with all large and im portant temples, are not to be
seen simply as ‘temple functionaries’, ‘employees’, as it were, of
the temple. Rather, the various categories of religious personages
— such as jTyars or m atha heads, achdryapurushas or learned and
often itinerant propagandists of Vaishnavism ; ek a k i Sri Vaishnavas
or celebate priests; and even temple accountants — may be seen
as im portant religious personalities in their own right. Such
persons would have followers in the countryside and in towns
who provided their leaders with the resources necessary to have
an im portant presence at such tem ples.168 These religious leaders
168 I again express my gratitude to two scholars whose work on this m atter has
persuaded m e: C arol Breckenridge and Arjun A ppadurai. Their work will be
considered more fully below.
may thus be viewed as personages whose religious roles conferred
com m and over substantial and redistributable resources; consider­
ing the evidence of the sixteenth century Tirupati, they were not
very different from the great political notables of the time. Such
sectarian and temple leaders may be viewed as conduits for resour­
ces marshalled from quite diverse social sources and ultimately
used in great temples to protect and augment sectarian interests
in ways that other entrepreneurial agents did; this deployment
of resources in temples also brought honours for the constituencies
served by these leaders. Thus, taking the categories o f donors
as shown in Table VIII-2 and adding to that diverse set o f resource
holders and redistributors those who anonymously provided
villages and money to their sectarian leaders for the support o f the
latter and the redeployment o f these resources as temple gifts —
considering all o f these appropriate holders and conveyors of
wealth — the fabric o f Vijayanagara society can be seen to have
become far m ore complex than what we are permitted to see during
the Chola period.
M ost, but not all, of the evidence which exists on the entrepre­
neurial and redistributive activities outlined above were related
to temples. This is the nature o f the evidence in almost all parts
of the m acro region. K arnataka offers something of an exception
to the general bias toward temple records. Num erous k a ttu
kodage rights are recorded on slabs o f stone which could be placed
anywhere whereas Tamil inscriptions, almost inevitably, are
inscribed on the basements and walls o f temple structures.169 In
whatever way these rights of the Vijayanagara period came to be
preserved, however, the purpose was to protect the entrepreneurial
beneficiary as well as to confer public recognition upon that activity
and the more generalized redistributive system o f which it was part.
These entrepreneurial and redistributional processes must be
fitted to extensive spatial networks in which centres, such as the
sacred complex at Tirupati, played a vital role. N otw ithstanding
these complex systems o f resource movements, agrarian m anage­
ment remained in the hands o f locally based men. Like the n altars
o f the Chola period, the chiefs and nayakds of the Vijayanagara
period were men of the locality, except, rarely, when they can be
identified as am ong the great Telugu n aya k a com m anders of
169 E .C ., v. 10, Kolar, ‘Introduction', p. xxiii.
Vijayanagara armies. The title, ‘n a ya k a ’ appears to have been
freely adopted and widely used by local magnates with no apparent
connection with Vijayanagara armies. This title signified an altered
form of local leadership which emphasized the greater scope for
individual achievement than can be observed in the evidence of the
Chola state and time. D uring the Chola period, and especially in
the more advanced, populous agrarian tracts, there is a pervasive
anonymity of leadership; it is the nattars acting as a corporate
body — the nadu — at times with the assent o f those local powerful
persons with the m uvendavelar title. In the Vijayanagara period,
it is as powerful individuals that we encounter local leadership.
The basis o f local authority remained the same: the dominant
local peasantry. M ilitary prowess and economic power were the
sources of local rural leadership in both periods, but the n ayakas
o f the Vijayanagara period m anaged more complex social and eco­
nomic universes. While this change in leadership can be noted in
parts o f the Tamil country as early as the twelfth century, it is a
com m on feature of the Vijayanagara period. G reater power and
prestige attached to this later form of leadership reflecting the
greater demands upon individual leaders as corporate organization
o f the earlier Chola age broke down. The scope for and rewards
to individual capability in the Vijayanagara period is sharply
contrastive with the C hola age.
Two essential points emerge from this recital of Vijayanagara
tenurial rights. First, land rights do not pertain to dom inion in
land, but to ‘property’ in share of income. Secondly, land continued
to be m anaged by the dom inant agricultural groups of a locality.
The Vijayangara land system, whether it is discussed in terms of
am aran ayakas or otherwise, is precisely about these two points;
so are the m ajor systems o f village organization described in the
late eighteenth and early nineteenth century British records.
It is appropriate to consider this latter, relatively full body o f docu­
m entation as a means of clarifying Vijayanagara arrangements.
The earliest direct experience o f the English East India Company
servants with land m anagem ent in South India came in what was
called, ‘the Jaghire’, the rural hinterland of F ort St. George, destined
to become the city of M adras. This territory o f 231 villages, 330
square miles, and a land revenue valued at Rs 250,000 (two and
one-half lakhs) was acquired as a jagir from the Nawab of Arcot
in September 1750 and was confirmed as an inam o r tax free
gift in an order (sanad) of the Mughal em peror in August 1765.170
The report of the East India Company official, Lionel Place, on
this tract in 1799 was the first to present that system of land rights
to become known as ‘mirasi system’. The following late nineteenth
century summary of Place’s description o f the system of ‘the
Jaghire’ is both accurate and interesting in its appreciation of that
system :
T he Jaghire was placed in 1794 in the hands o f M r. Lionel Place, and it
was from this gentlem an th at we are indebted for the first correct inform a­
tion as to landed tenures in South India. The villages o f the Jaghire were
discovered to be o f a class already [i.e., previously] described as dem oc­
ratic or m irasi. The villages, th at is to say, were co rpo rate bodies, with
an internal m unicipal constitution, and w ith the land the property
o f the corporation. T he land was sub-divided into shares which were
saleable, and still retained all the value o f real property. In each village
there were besides the corporate mem bers, cultivators holding as tenants
o f the co rporation and having o n their side prescriptive rights according
to ancient agreem ents. A gain there was a third class cultivating from
year to year w ithout any oth er privileges than th a t o f doing so. The
distinction between the shareholders and the tenants consisted in the
fact th at the latter could n o t sell his rights o f occupancy, n o r enjoy any o f
the various im m unities and advantages belonging to the form er as a
m em ber o f the corporation. M r. Place in m aking his settlem ents, dealt
with the w hole com m unities and n o t with any particu lar individual, an d left
it to the villagers to assess them selves individually. Each village chose
its representative o r representatives. There is every reason to suppose
th a t the jo in t village settlem ents o f M r. Place w ould have proved success-
fu l.n i
According to Place’s report o f 1799, the village lands o f ‘the
Jaghire’ were apportioned am ong 8,387 ‘meerassee’ holders in
no f f , e f i f t h R eport from the Select C om m ittee o f the H ouse o f C om m ons on the
A ffairs o f the E ast India Com pany, 28 July, 1812, ed. W alter Kelly Firminger,
Augustus M. Kelley, New Y ork, 1969; originally published Calcutta, 1918, v. 3,
p. xxi. Hereafter: The F ifth Report.
171 C.D. M aclean, Standing Inform ation Regarding the O fficial A dm inistration
o f the M adras Presidency in Each D epartm ent, in Illustration o f the Yearly A dm inistra­
tion R eports P repared Under the Orders o f Government, M adras, 1877; ‘Section II,
A dm inistration of the Land, Governm ent and Alienated L a n d ; (a) Historical Sketch’,
pp. 94-5. The reference in the quotation to villages ‘already described as dem o­
cratic. . . .’ is to the first im portant w ork subsequent to that o f Place on this form
of tenure, namely, F.W. Ellis, R eplies to Seventeen Q uestions P roposed by the Govern­
m en t o f F ort S t. George R elative to M irasi R ig h t; with Two Appendices (M adras,
1818) and. perhaps, the later work critical of that of Ellis. W .H. Bayley, M em oran­
dum on ‘M ira si T en u re’, London, 1856.
possession o f 15,995 shares o f income with another 1,827
shares either unclaimed or held appropriately by tenants
( ‘pyacarries’), or a total o f 17,822 shares.172 At almost
the same time that Place was reporting on the mirasi system of
‘the Jaghire’, other territories were being added to the C om pany’s
holdings in South India, including Cuddapah which was acquired
from the Nizam of Hyderabad as part of the Ceded Districts to
defray some of the costs of m aintaining a British force as provided
under the subsidiary alliance treaty between the Company and
Nizam. V isabadi was the name given to the system o f village
revenue encountered by Thom as M unro, the first Company
adm inistrator in Cuddapah. This term is a com bination of a
Dravidian word for the fraction ‘one-sixteenth’ (e.g. Tamil:
vis am ) and the Hindustani suffix ‘b a f , share or division;173 it
applies to coparcenery villages found not only in Cuddapah, but
in much of Telingana comprising Hyderabad State and modern
Cuddapah, A nantapur, Bellary, and Kurnool districts. According
to an early nineteenth century M adras Board of Revenue statement,
U nder this [visabadi] system, a fixed sum o f m oney was assessed on the
whole of the village for one or m ore years. A certain num ber o f the m ost
respectable ryots becam e answ erable for this am ount, each being res­
ponsible for his own separate portion thereof, and all for each other,
and the lands were divided by lot, as in the sam udayam villages o f the
Tam il country [a form o f m irasi tenure in which lands are held in com m on
but subject to periodic red istrib u tio n 174] the p o rtio n o f land to be occupied
by each being determ ined by the pro p o rtio n o f the rent for which he became
responsible . . . an d from this division o f the lands into shares the settle­
m ent took its nam e o f veesabuddy, namely a village settlem ent by shares
o f ready m oney.175

During the vigorous, and at times acrimonious, debates in

M adras between 1795 and 1817 over the kind of land revenue system
which was to prevail in the Presidency, village systems of the sort
described above were given serious study, and, for a time at least
172 The F ifth R ep o rt , v. 3, app. no. 16; ’E xtracts from R eport o f Mr. Place, Respect­
ing the Land Tenures in the Jaghire; dated 6th June 1799’, p. 165. Fractions in the
original have been deleted.
173 H .H . Wilson, A Glossary o f Judicial and R evenue Term s, Eastern Law House,
C alcutta, 1940. pp. 67, 549.
174 Ibid., p. 459.
175 C ited in S. Sundararaja Iyengar, L and Tenures in the M adras P residency
(S tu d e n ts ’ E d itio n ), M adras, 1933, pp. 81-2.
— betw een 1807 an d 1812 — seemed destined to becom e ‘the
M adras system ’. This, o f course, was in co n trast to, and to some
extent in conflict w ith, th e zam indari settlem ent o f Bengal. Village
settlem ents u nder a variety o f nam es were pressed by the revenue
officials;

the officers o f governm ent farm ed out the lands o f whole villages either
to the head inhabitants, w ho again sub-rented each field and settled with
each ryot, or the com m unity o f the village w ho settled am ong themselves
the land and rent w hich they were respectively to occupy and pay. The
apportioning o f rent and land was know n as am arakam . i76

In 1808, the M ad ras B oard o f Revenue p ronounced itself in favour

o f this form o f revenue system over that in force in Bengal or that
being p roposed by M u n ro as the ‘ry o tw ar’ or ‘kulw ar’ systems,
which were to triu m p h in the e n d .177
R ecent study o f the early phase o f the ryotw ar settlem ent in
M adras lends stro n g su p p o rt to the w isdom o f the M adras B oard
o f R evenue in insisting th a t the village settlem ent system was the
arrang em en t m ost closely a ttu n e d to the ag rarian order they
sought to co n tro l in the early nineteenth century. M ukherjee and
F rykenberg lay p a rtic u la r stress upon this point.
. . .im m ediately prior to the introduction o f C om pany rule, the prevailing
m ode o f land control was the village system. In this tradition, each indi­
vidual had been obliged to subm erge his own identity and to sacrifice
his own interests for the com m on weal o f the village, as determ ined by
the lord, the élite o f the village. Village affairs were controlled by persons
whose titles, by local custom , m ight be any o f the following : pedda raiyat,
reddi, kapu, dora, p a t el, kadim , mirasidar, gram atam u ; often the adjective
pedda, o r ‘g reat’ w as prefixed th ereto. These persons were the headm en,
the élite o f their villages, invariably, they were o f high and clean caste,
either B rahm ans or, m ore often, ‘yeom an-w arriors’. They stood between
the G overnm ent and the rest o f the village which was com posed mostly
o f labouring people o f low er caste.178

ne ibid., p. 83.
177 Maclean, S tanding Inform ation, p. 10!.
™ Nilmani M ukherjee and R obert Eric Frykenberg, ‘The Ryotwari System
and Social Organization in the M adras Presidency’, in L a n d C ontrol and Social
S tructure in Indian H istory, ed. Robert Eric Frykenberg, The University of Wis­
consin Press, M adison, Wisconsin. 1969, pp. 219-20. There is in this quotation
and elsewhere in this thoughtful essay a somewhat excessive appreciation o f the
harm onious village life of the time.
 Those whom Mukherjee and Frykenberg call the ‘yeoman-
w arrior’ or, elsewhere, ‘farm er-warrior castes’179 accurately reflects
the language o f early British adm inistrators faced with the task of
establishing Com pany rule. The experience of the earliest adm in­
istrators was essentially military, and the formal campaigns against
the well-organized field forces o f Hyder Ali and Tipu Sultan and
the M arathas were probably less influential on the views held by
these soldier-adm inistrators than the bitter, little battles with
‘poligars’ and other local magnates whose forces consisted of
peasant peoples fighting on and defending their own lands and
rights. The liquidation o f these rural warriors was strenuous
and effective, and their fate was often a brutal one.
In the southern p a rt o f Tam il country there were thirty-three Poligars,
for the realization o f w hose trib u te it was found necessary to ap point a
separate E uropean Collector. These Poligars fought desperately for w hat
they conceived to be their rights, and their reduction forms a notew orthy
incident in the m ilitary history o f the Presidency. O f those chiefs w ho
held their patrim onial estates for several generations, we find in the year
1803 thirteen only still in possession; the lands o f fourteen others were
under charge o f the E uropean C ollector, and six were forfeited, given
away or sold. In the districts ceded by the N izam in 1800, there were
eighty Poligars. These also resisted the G overnm ent, and had to be
reduced to subjection by force o f arm s. In 1807 they were found to have
been thus disposed o f; Pensioned 2, holding a Jaghire 1, residing on their
estates deprived o f auth o rity 23, m anaging their own estates 40, expelled
by force 6, in confinem ent 8, to tal 80.18«

British soldier-adm inistrators o f the late eighteenth and early

nineteenth centuries coped with the powerful com bination of
peasant-warrior, village leaders by persuasion and coercion. The
m ost successful, like M unro, developed a lively appreciation for the
capacity of the ‘yeom an-warrior’ to resist the imposition of Com ­
pany rule as well as their capacity to facilitate Company rule.
Once again, the findings o f Mukherjee and Frykenberg are im­
p ortant:
Evidence o f the continuous an d subtle influence o f village leaders
abounded [in the later nineteenth century]. In 1882, the H ead A ssistant
C ollector at C u d d ap ah reported th a t ryots still stood in respect and
scrupulously obeyed the ‘C hiefs’ or P edda Reddis. The P edda R eddi
was the only person in each locality w ho had effective control over ryots.

179 LOC. Cit.

180 M aclean, S tanding Inform ation , p. 96.
 The only peace and stability o f the area depended upon him . The R eddi
had only to express dissatisfaction at his own individual revenue settlem ent
and m ost other ryots (usually also o f the sam e caste) w ould follow his
exam ple, even if they stood to lose by doing so. W henever any difficulty
arose in com ing to an agreem ent, as in m aking an annual revenue settle­
m ent (jamabandi.) w ith village ryots, if the cause [of the difficulty] could
be discovered — and often it could n o t be — it w ould be found th at the
R eddi was responsible. As soon as he was satisfied, troubles would cease.
R yotw ari settlem ents, village settlements, or w hatever, his influence
persisted. A 'g o o d ’ (or cooperative) Reddi could be an asset to the G overn­
m ent ; a ‘b ad ' one could be its bane. It was reported th at persuasion by the
w hole staff o f the huzur kachahri, o r collector’s office, could not do in two
days w hat a R eddi could do in h alf an h o u r.181

It is of little wonder, as these authors point out, that revenue officials

o f the M adras government were, at the beginning o f the nineteenth
century, as suspicious o f peasant, village leaders, as they were to
become o f their Brahman revenue servants during the middle and
late parts of the century.1*2
In only one m atter were the early British adm inistrators wrong
with respect to the village leadership with which they had to deal.
T hat is, they believed that the ‘farm er-warrior castes’ exercising
village dominance were created by the chaos of the eighteenth
century or by the oppression o f Muslim rule. The British believed,
or in any case expressed the idea, that Muslim policies and the
oppression and chaos engendered by them justified the revenue
and judicial policies prom ulgated by the Company. In this, the
British adm inistrators and C ourt o f Directors in London were like
most of the historians o f Vijayanagara who consider that the
early Muslim incursions from the fourteenth century onwards
explain and justify Vijayanagara policies. The armed and powerful
village leadership o f the late eighteenth century is m ore truely the
vestige of the am aran ayaka system of the Vijayanagar period in
the particular form encountered by the British. Perhaps more
truely, this kind of local leadership dates from decline o f ethnic
territoriality and the nadu in the late Chola period.

Tem ple and S ect

The concluding section o f this chapter deals with a variety of

181 M ukherjee and Frykenberg, op. cit., pp. 222-3.
182 Ibid., p.’ 2.21.
issues only poorly apprehended by its title. The intention is not
to discuss ‘religion’ in the period from 1350 to 1700, though that
is a part of its subject m atter; it is rather to focus attention on the
medieval South Indian temple and the relationships between it
and various kinds of social and cultural features of that society.
The temple is viewed in relation to absolute and enduring religious
corporate groups such as the Terigalai Sri Vaishnava sam pradaya
as well as in relation to those relative and potential groupings to
which such term s as k a tja la id a r (temple donor) and idangai-
valangai (left and right caste divisions) are applied. M ore abstractly,
the temple is viewed in relation to m orality, that is, in relation to
those conceptions about how competing and conflicting claims
am ong persons and groups ought to be resolved. As in other
sections of this chapter, the purpose is less to summarize historical
interpretations on which there is wide agreement than to suggest
processes and relationships o f which we are presently less aware.
Underlying the conceptions which are presented here are a set
o f tensions which emerge from the preceeding discussion of the
political system and resources o f the Vijayanagara period and which
repeatedly called for m oral adjudication. The sources of these
tensions in the society of the Vijayanagara period should be made
explicit.
The nayankara system generated profound tensions for the
society o f later medieval South India. It was an intermediary level
of authority and power which was uncertainly poised between
two very different levels o f established authority: that o f a macro
regional kingship, at one level, and that of atomistic, micro regional
peasant societies, at another level. Because the South Indian
political system lacked the relatively firm principle of ruling clanship
of Kshatriya ideology and because it never developed a contractual
basis similar to European feudalism, the nayaka system was orien­
ted, Janus-like, both upward to the Rayas and downward to
thousands of peasant localities.
Vijayanagara nayakas were creatures o f the State, created by
the military superiority and purposes o f the Rayas, and therefore,
necessarily, involved in and vulnerable to the hazardous imperial
politics of this militaristic state. N otw ithstanding several kinds
of divergent interests, nayakas and Rayas shared one fundamental
interest: that o f preventing new regional kingships from coming
into existence. When this threatened, nayakas and Rayas together
 could and did cooperate in repressing such incipient states. But,
apart from this shared interest, there was little that clove the great
nayakas — or the many, m inor chiefs who assumed the title — to
the Rayas, and there was much that divided them from these kings.
To Teiugu nayakas bent upon m aintaining their often delicately
poised control over peoples and territories in which they were
outsiders, the prospect of an invasion by a Vijayanagara army —
as reportedly under Krishnadevaraya in the early sixteenth century
and m ore certainly under Sriranga III in the middle of the seven­
teenth century183 — and their possible displacement was an all too
real threat. The same threat was equally before Tamil chiefs as
known from a num ber of Mackenzie docum ents.184 Contrarily,
the Rayas could justly fear the com bination of n ayakas which
would thwart an even strong R aya’s ambition, but, more seriously,
could lead to the seizure of dynastic control. This occurred when
Saiuva Narasim ha, in the late fifteenth century, brought an end to
the first dynasty of Vijayanagara. Less focused combinations o f the
great nayakas could also constitute an alternative locus o f Vijaya­
nagara authority which brought about the civil wars o f the early
seventeenth century.
The policy of relying upon Brahmans as instrum ents of imperial
control, especially during the time o f Krishnadevaraya, as we have
it from the A m u kta m a lya d a and inscriptional evidence, appears
to have been successful under a king of his power and ability. But,
this proved a weak and unenduring solution to potential cleavages
in the political structure. The career and rebellion of Sellappa
make this clear.
Sellappa Saiuva Nay aka was a Brahman officer whose father
had served the deity Ekam baranatha at Kanchi. Sellappa rose
to political prominence under Krishnadevaraya, and he appeared
to have served that king’s interests in Ram nad. At T iruppattur
there are two inscriptions of a . d . 1510, extolling the virtues of this
Brahm an officer and recording endowments by him of two villages
— renamed in his honour as ‘Sellappapuram ’ — to a Siva tem ple.185
183 Krishnaswami, op. cit., pp. 193-4 and Ch. 15.
,!i4 E.g., 'A ccount of Kandava Rayan and Setu Rayan W ho Ruled from the Fort
of Tiruvidaiccuram in the Arcot D istrict’, M ackenzie M a n u scrip ts; Sum m aries
o f the H istorical M anuscripts in the M ackenzie C ollection, v. 1 (Tamil and
Malayalam), T.V. M ahalingam (ed.), University of M adras. M adras. 1972, p. 94.
185 Krishnaswam i, op. cit., p. 200.
Other honours were added in the remaining years of K rishnaraya’s
reign as he continued to hold high office along with other Brahman
notables586 in the far southern part of the peninsula; often Sel-
lappa’s inscriptions fail to even mention the King, though his loyal
relationship with Krishnaraya continued. M oreover, Sellappa
appears to have supported the troubled candidacy o f Achyutade-
varaya. However, early in the latter’s reign — one day after his
coronation according to the contem porary Sanskrit work, A chyuta-
rayabhudayam of R aianatha D indans187— Sellappa joined the
Travancore king U dayam artandavarm an, or Tiruvada (possibly
entering a m arriage alliance with that king), who had come to
control a substantial part o f the southern end o f the peninsula by
a .d . 1530. Achyutadevaraya reacted vigorously against their
com bination and used the occasion for a trium phal progress o f the
southern country, conferring gifts at Tirupati, Kalahasti, Tiru-
vannam alai and Srirangam .188 No very convincing explanations
o f Sellappa’s rebellion have come forward and none may be neces­
sary to explain the behaviour o f this able man apart from noticing
that the times offered num erous other examples of self-interested,
powerful men pursuing their am bitions. Against such m achina­
tions there was certain protection for neither Rayas nor nayakas,
thus the frequent political chaos o f the Vijayanagara empire that
m arks the strategy o f political self-preservation, of all against all.
U ncertain relationships between these great nayakas and the
Rayas was replicated downward from the nayakas. In Tamil
country many p f the m ost powerful nayakas were outsiders —
Telugus. ...Vho had won their locality control through the threat
of their superior military power based upon technical advantages
and in part on the following they held among their own peoples
whose migration had preceded their control or accompanied it.
All such migrants were outsiders like the nayakas. There were
also a variety o f local peasant allies offering collaboration with the
186 E.g., a prom inent poet in K rishnaraya’s court, AUasani Peddana was granted
n a ya n ka n t powers at Annur in Karirachi-Sima in m odern South Arcot ( A .R .E .,
1916, para. 66, p. 143),
i»7 S. Krishnasvvami Iyengar, Sources o f Vijayanagar H istory, n o .5 3, pp. 158-9,
cantos 4-6.
>** Ibid., canto 5 and V.N. Hari Rao, "A History of Trichtnopoly and Srirangam ’,
unpublished Ph. D Thesis, University of M adras, 1945; typescript consulted at
the University of Chicago Library.
n ayakas. Am ong such were those who perceived new opportun­
ities to alter their long-standing political relationships with
the ancient n attar to whom they had long been subservient. The
rise of the Pallis during the sixteenth century in modern South
Arcot and in Kongu is an example of this, so are the alliances bet­
ween the M adurai n ayakas and the M aravars. Equally, if not more
im portant, were the relations forged between intruding Telugu
warriors and a variety o f sectarian leaders in Tamil country, a
relationship which will be examined m ore fully below.
By these means, Tamil country became a region of great opportu­
nity for military adventure and imperial expansion during the
fifteenth century and earlier. But, the establishment of nayaka
authority there was made possible by yet another factor: the
prior reduction of various institutions o f local organization dating
from the Chola period. This was a precondition of n ayaka rule in
Tamil co u n try ; it was not, as Krishnaswami and others have argued,
a ccjnsequence of n ayaka rule. U nder the n aya k a s , no new political
forms o f institutions were established in place of the nadu and
periyan adu, the sabhas o f brahm adeyas, or the nagaram s of m er­
chants. The locality political superstructure having previously
been weakened, and in some places eliminated, there existed a
vacuum which provided the opportunity for the installation of
n ayaka rule. N or were political institutions linking nayakas
and their peasant subjects replaced by other political or adm inistra­
tive institutions. Hence, the absence of effective linkages between
the supralocal, intermediary authority of n ayakas and the peasant
peoples from whom they exacted tribute can be seen to have been
as great as the absence o f reliable linkages between the nayakas
and the Rayas.
At the base of these political tensions were changes in the ins­
titutional and m oral order whose origins date from transform ations
o f the twelfth century. W hat the Tamil historians N ilakanta
Sastri and Krishnaswami lament as the decay of the ancient self-
governing institutions under Vijayanagara (or more properly,
n ayaka) rule, and which they justify as a necessary cost of protect­
ing South Indian Hinduism and varnasram a-dharm a , is better
seen as internally generated change in the peasant society o f South
India in later C hola times. The transform ation o f some part of
the nadu leadership into chiefs over increasingly pluralistic local
societies, the emergence o f temple urban centres, the enhanced
im portance of mercantile and artisan groups with whom the
periyan adu chiefs forged new and strong relations, the increased
m onetization of the economy, all these were responsible for the
decay of the local institutions of Chola times and indeed the Chola
state itself.
The process continued into the fourteenth and fifteenth centuries
with the focus of local management of productive resources collaps­
ing more and m ore around villages controlled by segments or
lineages o f peasant clans o f a locality. D om inant, local agricultural
groups were thus less and less coherently linked together by ethnical­
ly defined political institutions such as the nadu and periyan adu ,
and increasingly by complex cultural and symbolic relationships
to wider networks of allegiance and identity. This is seen in a
genre of poetry celebrating Tamil peasant groups and in the crucial
role played by temples and sects.
Occasional references have been m ade in this study to a little-
recognized Tamil verse tradition extolling the qualities o f peasant
folk and the territories in which they lived. The poetic genre is a
satakam form, and the territory is the m an dalam ,189 The m andalam
term has a history among Tamils which is very ancient.. References
to it occur in the poetry of the Classical period,190 and these terri­
tories are again referred to in the Saiva doctrinal work Tiruman-
diram o f Tirum ular o f about the sixth century.191
In one stanza o f that work, the m andalam is given a strong moral
signature:
The five T am il M an d ala are T atvas (i.e., are regions w here the tru th has
been fully revealed); there roam wise men whose m inds have blossom ed. . .
and w ho have know n the ancient truths. . . . They have given utterance
to their know ledge as easily as they w ould throw ou t w ater thro u g h their
m ouths and this know ledge spread over the w hole of the five Tamil
m an d alas.192
The com m entary on the Tirum andiram takes the five mandalam s
to b e : Cholam andalam , Cheram andalam , Pandyam andalam , Ton-
189 An essay involving these elements by the author, entitled 'Circulation and
Historical Geography in Tamil C ountry', Journal o f A sian Studies, v. 37 (November,
1977), pp. 7-27.
190 Subrahm anian, Sahgam P olity, pp. 224-6.
191 C.V. N arayana Ayyar [Swami Sadananda], Origin and Early H istory of'Saivism
in S outh India, University o f M adras. M adras, 3936, p. 211.
I»2 Ibid., pp. 216-17.
daim andalam , and K ongum andalam .593 M andalam s are also used
during the Chola period to designate the several large territories
comprising the Tamil portions of the Chola state of the eleventh
and twelfth centuries.
However, the sa ta k a m s referred to her differ from older forms
of literature which only rarely refer to m andalam s and never to
their peasant inhabitants. Several of these works survive, each
set in a pattern o f one hundred verses. They exist for all parts of
Tamil country,194 as well as for southern K arnataka.195 To these
may be added a purdna of the territory in modern South Arcot
between Chola country and the central Tamil plain (or Tondai-
m andalam) which celebrates the dom inant Vanniyars, b u t. is
presented as an explanation o f the right-and left-hand castes.196
These works date from the seventeenth and eighteenth centuries
for the most part. In a num ber of ways, these satakam afford a
striking example of ethnic pride in and identification with their
territory by powerful peasant peoples.
In almost all cases, the sa ta k a m extoll the virtues or moral
superiority o f peasant peoples whose society was the core of the
contem porary social order upon whom all others depended. They
are called Vellalas in most cases.197 The C holam andala-Satakam
refers to the banner of the Choliyar which displays the ploughshare
(me//)198 and recalls with pride the support which the ancient
Choliyar gave to the late twelfth century poet Kam ban (‘K am ba’)
who produced the poem Erelupadu, a hymn to the plough and to the
Vellalas.199 Echoing K am ban’s appreciation of the agriculturist,
1« Ibid., p. 218.
194 The style of the sa ta ka m (Sanskrit — sataka, a poetic form of 100 verses:
T am il L exico n , p. 1254) is k a tta la ikka littu ra i according to Arokiaswami, Kongu
C ountry, p. 25, this being a verse o f four lines, each with sixteen syllables ( Tam il
L exico n , p. 647). The sa ta k a m s are: pandim andala-satakam ( P .S .) by Aiyain
Perum al Asiriyar o f M adurai (Sirkali, 1932); colam andala-satakam ( C .S .) by
A tm anathar D esikar o f Velur (1650-1728), ed. Som asundaradesikar (M ayunar,
1916); tondaim andala-satakam ( T .S .) by the Jaina Brahm an Jinendran or Kar-
m eghakavinar o f Vijayam angalam , ed. T.A. M uthusw am i K onar (Trichengodu,
1923,), cited in Arokiaswami, Kongu C ountry, p. 25.
195 K arm andala-satakam b y A raikilar of Avanasi ( K. S ) com m entary by P.A.
M uthuthanadavaraya Pillay (M adras, 1930).
196 Idahgai-valahgai puranam , Oriental M anuscripts Library, University of
M adras, no. D. 2793, palm leaf.
197 E.g., P .S ., v. 17 and C.S., v. 27; C .S ., v. 38; K .S ., v. 20-1, 23; T .S ., v. 84, 98.
19» C .S ., v. 38 and T .S ., v. 98.
i9y C .S ., v. 78; N ilakanta Sastri, The C olas , pp. 672.
the satakams make claims for the peasantry as people o f respectable
status whose wealth was lavished upon Brahm ans and canonical
gods and who were ‘pure’, sâtsüdras (Tamil = carcüttiran): ‘Chola-
m andalam is famous for its Chôliyâr who were the sâtsüdras
sprouted forth from the feet o f Brahm a and who gave the produce
their ploughs to all hum anity.’200
In p ro o f of their pretentions to ‘purity’, those honoured in the
satakams refer to m ost o f the sacred places in their territory and the
gods sheltered and worshipped there;201 they also claim to have
founded temples and mathas ,202 References are made to Manu
whose ordinances these peasants claim to uphold.203
A prom inent theme o f the satakams is the political importance
of the peasantry. The Pandimandala-Satakam states: ‘The vëlâkir
o f this mandala, give horses, elephants and chariots to the Pândya
kings and put the crown on their heads;’204 while in the Chola-
mandala-Satakam : ‘The Chôliyâr of this m andalam had the
rights of sthânikas [temple managers] and kdni [land holding] and
the privilege o f staying near the king. . .205 The sixty-four kudis
[lineages of Chôliyâr] had the right of crowning the Chôla king
Karakâla. ’206
The Karmandala-Satakam refers to the ‘three Tamil kings’207
whom those of the territory had served in their territory along with
rulers from northern K arnataka.208 However, the K aralar deny
th at any king ever subjugated them. Rather, kings like R ajaraja
the G reat were accepted as the king of the K aralar because of his
ardent devotion for the god Siva, and Rajendra I, whose title, the
verse points out, was gangaikondân (‘conqueror of the Ganga
country’) not gangankondân (‘conqueror o f the Ganga people’)
200 C .S ., v. 6. In T .S ., v. 97, the title cüttirar is claimed ( T am il L exicon, pp. 1342,
1561).
201 C .S ., v. 50, 78; K .S ., v. 26-6, 31, 40-4, 49-50, 60, 80; Y .S ., v. 84, 99.
202 K .S ., V. 60.
20Î C .S ., V. 29; T .S ., v. 56.
204 V, 17.
205 V. 57.
206 v . 61 ; see k u ti, am ong the meanings o f which is ‘family’ or ‘lineage’ (T a m il
L exicon, p. 968).
207 v . 27.
208 v. 28. K arnataka inscriptions occasionally use another word for m andalam ,
namely, p ani as in a Tam il record o f a . d . 1098 ( E .C ., v . 10, pt 1, no.426 Mulbaga!
Taluk, pp. 80-2).
by which name those of Karm andalam were also known.209
Karalars were friends of the great Bana chiefs and served them,
as well as greater kings.210
The satakam s have had little importance for historians because
they seldom contain evidence which would be of interest to those
concerned with conventional political history. A part from their
use in historical geography211 these poetic works have been dis­
missed as merely the expressions of a part of the peasantry of the
macro region — Vellalas — who shared a common northern
origin m yth.212
The generic terms k a rk a tta -vela la r or ‘Ganga-people’ refer to a
prim ary mythic origin in the Gangetic plain and a secondary one in
southern K arnataka — Gangavadi or Karm andalam . From here,
the verses state, these G anga people or kark a tta -vela la r, spread over
the entire m acro region.21 Pandim andalam was settled by these
folk after they h ad resided in Chola country according to satakam s
o f both Pandim andalam and C'holamandalam.214 Tondaim anda-
lam was also settled by folk of the Chola country according to the
sa ta k am of Tondaim andalam and other, somewhat later, tradi­
tions.215 Conquest is neither claimed in these verse collections
nor supported by them. There is no suggestion that the Tamil
plain was, at some time, overrun by more northern peoples. On
the contrary, even the Karalar, of what is now the southern portion
o f K arnataka, claim to be devoted to the protection and preserva­
tion of their Tamil language.216 Their boasted primacy seems
clearly to have arisen from close association with kings and great
chiefs and from their purity in ritual terms.
The satakam s provide an im portant, poetic insight into the
broadly shared ideology o f the leading stratum o f peasants in the
209 v . 28. Also see v. 32 where it is stated that karalar are the progeny of gahgadat-
tan or gahgeyan who was born near the river Ganga. hence they are known ‘¿ sgarigan.
21« K .S ., v. 88.
211 See Arokiaswami, Kongu C ountry , pp. 6, 13, passim .
212 N ilakanta Sastri, The Colas , p. 690, no.25; 'All vellalas are described as being
of the G angi-kula’.
213 See T .S ., v. 97; C .S ., v. 98; P .S.. v. 11; K .S ., v. 31-2.
214 P .S ., v. 5; C.S., v. 92.
215 T .S ., v. 97. Also, M ahalingam , M ackenzie M anuscripts ‘Account of Kandava
Rayan and Setu Rayan who Ruled from the Fort at Tiruvadaiccuram in the Arcot
D istrict’, p. 94. However, the P .S ., v. 9 and 10 claim that Vellalas were sent by
the Pandya king to Tondaim andalam .
216 K .S ., v. 60.
macro region. All claim ritual purity and respectability from their
support of varnasram adharm a according to sm riti and their support
of brahm anical religion; all claim respectable secular ra n k ; all state
with pride their precedence as the leading cultivators, upon which
pivot all their relationships: those with local chiefs (udaiyar), with
lesser cultivators (k u diyar ), with pastoralists (konar or idaiyar)
and with artisans and merchants of all kinds.217 Yet, these folk
were divided, during the seventeenth and eighteenth centuries,
into great territorial groupings, i.e. the mandalam s. The nadu
micro unit o f organization of the Chola period had, by that time,
or at least in the poetic forms of the saiak a m s, ceased to be as
significant as it once was; a larger territorial identification was now
claimed.218 T hat quite localized forms of peasant social structure
and agrarian organization had not, in fact, ceased altogether to be
significant in the time of the satakam s is clear from the earliest
British records of peasant peoples of the macro region.
Vellalas, who comprised the dom inant landed people from the
earliest times for which there are records, conform to a clear pattern
of territorial segmentation and have long done so.219 Vokkaligas
of K arnataka and K apus or Reddis of A ndhra also fall into the
same pattern. Vellalas of Tamil country have been divided into
four main groups according to ancient geographical terms as their
titles reflect: Choliya, Tondaim andala, Pandya, and Kongu.
Each m ajor sub-division has shared a common array of cultural
traits. These pertain to domestic and other ritual engagement
in ‘clean’ (sdtvik) occupations such as agriculture and trading;
each has had its own characteristic titles as *Pillai’ among the
Choliya and Pandya Vellalas, M udaliar among the Tondai­
m andala Vellalas, and K avanadan (or G ounda) among the Kongu
2)7 C .S ., v. 27; P .S ., v. 10. Also see T am il L exicon, p. 1204: konan.
218 References in the sa ta k a m s to locality units as the nadu or ko tta m are rare.
In the T .S ., v. 87, the '24 k o tta m s ’ are m entioned; the C .S ., v. 61 and 81 refers to
the ‘64 kudis as if these m eant the leading, localized lineages of the mandalam (cf.
Tam il L exico n , p. 968, ku ti); and the K .S ., most explicitly of all, in v. 7, mentions
the ‘32 nadus' of the western hill tracts of K arm andalam and the '64 nadus’ of the
eastern, plain or m aidan, tract o f this territory. The com m entator o f the kdr-
m andala-satakam notes o f verse 7 that the reference to 32 nadus is understood as
the 32,000 veldlar families and the 64 nadus as the leading families of the eastern
portion, thus, the reference is to karkattavelalar o f 96,000 and an explanation of
the ancient num erical designation for Gangavadi as a 96,000 territory.
219 Subrahm anian, Sarigam P olity, pp. 224-6.
Vellalas; and all contained numerous further divisions based
primarily upon territory.220 One nineteenth century estimate
of Vellala sub-divisions was 590, most being the result of slight
variations of a territorial sort.221 Sharing an equally prestigious
role in K arnataka, the Vokkaiigas are divided into a num ber of
territorial divisions. The G angadikara Vokkaiigas (Karalar) are
concentrated in the south-central portion o f K arnataka abutting
modern A ndhra having long shared the territory with Telugu
speaking R eddis; N onaba Vokkaiigas inhabit the tract on the
northern bank of the Tungabhadra, medieval N olam bavadi.222
Other Vokkaliga groups are similarly clustered in other parts of
modern K arnafaka.223
Divisions am ong the Kapus o f the Telugu-speaking portion of
the macro region show a strong territorial character. Panta
Reddis, one of the m ajor sub-divisions among this dominant
peasant group, were called ‘the fourteen com m unity' during and
after the fifteenth, century. O f the fourteen names, twelve may be
conclusively identified as territorial: Pakanati K apus of pakaadu
(between the Pennar and G undalakam m a rivers), Velanati Kapus
of valanadu (the northern portion of m odern G untur district),
M otati Kapus of m ottanddu (southern bank of the Krishna river),
M orasu Kapus, like M orasu Vokkaiigas, in K olar district of
eastern K arnataka, M unnuru K apus o f Muliki 300’ country.
Panta Kapus (of G udur taluk, Nellore district), Desati Kapus
(Ogeru river basin, G untur district), and the following Panta
Reddi groups are named for towns in the region: Ayodhya Kapus,
Oruganti (W arangal) Kapus, Kuricheti (Kurucedu, Vellore),
and G adikota (Cuddapah district).224
The formal territorial segmentation of Kapus may be identified
as early as the tenth century when a chief celebrated his investiture
with remissions to ‘K am pus’ o f Venadu.225 Telugu Kammas,
220 e . Thurston, C astes a nd Tribes o f Southern India, Governm ent Press, M adras,
1909, v. 7, pp. 374, 376-81.
221 M an u a l o f M a d ra s A dm inistration , Governm ent Press, M adras, 1893, v. 2,
app. no. X X XII, ‘M anners and C ustom s’, p. 228.
222 L.K. A nanthakrishna Iyer, The M ysore Tribes and C astes, Mysore University,
Mysore, 1935, v. 1, p. 125. Also see the useful maps prepared for that section.
223 Ibid., pp. 150-3.
224 M. Som asakhara Sarma, H istory o f the R eddi K ingdom s: Circa 1325 A .D .
to Circa 1448 A .D ., Andhra University, W altair, 1948, pp. 51-2.
225 E .I., v. 30, no.46.
also dom inant in medieval as well as m odem Andhra, are associated
in Pallava inscriptions with a portion o f the modern taluks of
Ongole and Bapatla in G untur district which was called ‘Kamma
R ashtra’.226
This elaborate mosaic of segmented territorial groupings is
further complicated over time. Ethnic groups whose historical
role with respect to control of the land has been param ount in
many localities o f the macro region since the eighteenth century
cannot simply be presumed to have been dom inant in these places
during earlier periods. Many changes have occurred. For example,
only in the fifteenth century is there evidence of tjie displacement
o f Vellalas as the dom inant land-controlling people of some parts
of northern Tondaim andalam by Telugu Reddis and Balijas or
by Tamil-speaking Vanniyars and their displacement from some
Kaveri delta lands by Kallars.227 N or is it supposed that there
was a simple relationship between locality or territory and those
persistent social interactions which together constitute territorial
social structures.
On the latter issue, it has been argued in the discussion of the
Pallava period that peasant society in the Coromandel plain
consisted of ranked groupings o f various kinds from Brahmans to
untouchables and that tracts outside of the ancient Kaveri tracts
and Tondaim andalam were only gradually incorporated into the
Corom andel form of peasant society. This conversion of many
localities of the macro region to peasant society was often
achieved under the dom ination o f peasant folk from elsewhere,
and every peasant locality contained a mix o f social groups, some
o f which might have been residents o f the territory from the earliest
times o f its existence as an identifiable territory while others were
relatively recent arrivals. Some villages in portions o f T ondai­
m andalam , particularly in the north, contain groups whose pro­
genitors settled the region almost a millenium ago, whereas many
villages in portions of Salem district contain groups whose arrival
226 A .R .E ., 1916-17, para. 3.
221 F.W . Ellis, R eplies to Seventeen Q uestions Proposed by the Government o f
Fort S t. George R elative to M irasi R ig h t; with Two A ppendices , Governm ent Gazette
Office, M adras, 1818, pp. xi-xiv; M. Arokiaswami, The E arly H isto ry o f the Vellar
B asin; with S pecial R eference to the Irukkuvels o f K odum balar, Am udha Nilayam,
M adras, 1954, p. 125; Diwan B ahadur T. Venkasam i Row, A M anual o f the D istrict
o f Tanjore in the M adras Presidency, M adras, 1883.
may be dated from the eighteenth century.228 F or most of the
macro region, therefore, the peasant society of a locality emerged
from the cumulative interaction of diverse, ranked social groups
many of whom were, in origin, strangers to the place. While all
such groups identified themselves, to some extent, with the locality,
often comprising endogamous sub-divisions within it, the locality
would inevitably become most clearly associated with those groups
in whom the m ajor authority over the land was vested - the
dom inant peasant people. Brahmans who were not exceptions to
this territorializing of social groups, now were lower ranked
peoples.229
Considering the complex configuration of localized ethnic
groups, num erous small-scale migrations and local conquests by
different peasant folk, and the many small territorial arenas where
all of this took place in the macro region, in what sense is it possible
to speak of a Vijayanagara society? If it refers to that segment of
Indian humanity under Vijayanagara rule, the notion of a Vijaya­
nagara society is a weak one. The Vijayanagara political system
provided some coherence, to be sure. Vijayanagara kings were
known in all parts of the macro region as great (and it must have
been hoped, distant) sovereigns to whom chiefs great and small
paid occasional homage; Vijayanagara fortresses under Brahman
commanders would have been known to many as they moved about
searching for land or trade or on pilgrimage. But the formal
political and administrative structure generated as much fission
as fusion. It is quite another kind of integration that provided a
significant measure of identification with some meaningful whole
for most people o f the m acro region in Vijayanagara times. This
was created by the Hindu temple and Hindu sect.
‘Tem ple’ and ‘sect’ are terms which serve poorly to convey the
meanings of k o y il and cam piradayam (Sanskrit: sarhpradayd)
because the English words necessarily carry meanings derived
228 The Baram ahal R ecords; Sec. Ill, 'in h ab itan ts’, Governm ent of M adras,
M adras, 1907, for evidence of this.
229 Distinctions were m ade between the dom inant peasant families and those
o f lesser im portance, velkudi-ulavar , or 'fallen vellala’s, according to Kanakasabhai,
op. cit., p. 113; Brahmans are discussed in M anual o f M adras A dm inistration , v. 2,
App. XXXII, 'M anners and Custom s', pp. 226-7; Thurston, op. cit.. v. 1, especially
pp. 334-93; also see N. Ram esan, Copper P late Inscriptions o f Andhra Pradesh
G overnm ent M u seu m , H yderabad, G overnm ent o f A ndhra Pradesh, H yderabad,
1962, v. 1, p. 192.
from the cultural context of their provenance. Such meanings lead
to distorted understandings of the relationships am ong Hindu
deities, devotees, and priestly and preceptoral intermediaries.
To some extent, of course, these distortions and their potential
for confusion are recognized. There nevertheless remains a ten­
dency to suppose certain likenesses with other religious traditions,
especially western ones, and these presumed similarities inflect
the terms ‘temple’ and ‘sect’.
In medieval South India, divinity attached to many kinds of
beings and sacredness attached to many kinds of places. The
term ‘temple’ may therefore be applied to a prodigious array of
places where some divinity was worshipped; all may be called
kdyil. Those places known as ‘temples’ to modern scholarship by
their designation as srTkoyil, by their distinctive architectural
features, by their staff of ritual functionaries, by their being ancient
resorts o f pilgrimage, and, indeed, by their inscriptions are but a
selection of the vast totality of places to which the term k d y il
is appropriately applied. There is no unambiguously coded place
or thing called a ‘tem ple’ as in Christianity there is a church.
N or, in the puranic Hinduism of medieval South India, is there
an institution like the church in Christianity or vihdra in Buddhism
in the sense of a bounded domain of action and meaning which
separates it from other domains, such as the church from the state,
the clergy from the laity. On the contrary, a temple was wherever a
group of devotees founded a deity and were co-sharers in its gene­
rosity, whether this was a great shrine like that o f Venkatesvara
at Tirupati or the tree shrine of a tutelary goddess. Generically,
a temple is a nexus of sharers consisting of a deity and its worship­
pers; its purpose is to protect and to transform the community
of worshippers by the boons o f the god; and its -means are the
transvaluing of substances — human and non-hum an — offered
to the god and returned to its devotees. The temple ‘institutiona­
lized’ — that is, considered as one among many institutions in
society as it was seen by British adm inistrators during the nineteenth
century and to an ex te:\ by modern scholars — is the temple
misconceived. Seen as a set o f sharing beneficiaries in the generos­
ity of a deity is to see the temple as it was understood by people
of the medieval age, and, conceived in this way, the temple was
perhaps above all else the model of the conception of shares and
sharing of that time. W hat shares of village income was for the
distribution and redistribution o f resources in the material order
o f the time, shares o f support to deities and prasadam from these
deities was in the moral order of the time. And just as it was
stated that there could be no complete village as a material order
without its eighteen a y agars holding service shares of land, so it
would have been that there could be no complete moral order
without the sharing and transactional nexus centring upon a god,
whether of a family, a ja ti, a clan or o f a hamlet, a village, a locality,
a kingdom.
A recent general essay on caste felicitously states certain of the
relations to which attention has been directed in the present dis­
cussion. It opens with the following sentence:
C aste systems are m oral systems th at differentiate and ran k the whole
population o f a society in co rp o rate units (castes) generally defined by
descent, m arriage, and occupation.230

It goes on:
Every hum an genus (and therefore every caste) is thought to have as the
shared and corporate property o f its m em bers a p articular substance
(e.g. sarTra, ‘body’, rakta, ‘b lo o d ’) em bodying its code for conduct (dhar-
ma). Every caste’s inborn code enjoins it to m aintain its substance and
m orality, its particular occupation, and its correct exchanges w ith other
castes.231
And continues :
The H indu gods to be w orshipped are related am ong themselves, like men,
by shared and exchanged n atu ral substance. They have no t merely
abstract qualities and representations like the Vedic gods, but also p a rti­
cular life-like images, biographies, bodily functions, and specialized
relations to men. They are attached by p articular codes o f w orship to
particular occasions, com m unities, and genera o f persons. . . 232 The
codes o f H indu w orship require the existence o f com plex local com m unities
o f castes. W orship can n o t proceed w ithout the priest to bring the living
substance o f the god into the image, m ade by an image m aker. . . . To
sponsor the w orship there m ust be local w orshippers o f m eans, typically
a ruler or a m an o f w ealth, w ho can by gifts entreat a priest to m ediate
w ith the god. There m ust be specialists o f appropriate castes. . . to feed,
attend, and entertain the god.233

230 Mckim M arriott and R onald inden, ‘Caste Systems’, Encyclopaedia Britannica ,
15th ed (1973). v. 3, p. 982.
231 Ibid., p. 983.
232 ibid., p. 985.
233 L O C . c i t .
 This conception of a moral order comprising communities of
Hindus, their gods, and priestly intermediaries sets the essential
framework for consideration o f the great temple in medieval South
India as a co-sharing body o f worshippers, in its insistence upon
priests, especially Brahmans, and a variegated corps of caste specia­
lists to serve the god, this conception obviously pertains not only
to great temples but also to the entire range o f shrines to which the
term k d y il is applied. Moreover, the structure of religious affiliation
of the period provided for the integration of worship and worship­
pers of even the most humble tutelary deity with the highest form of
bhakti, puranic worship of the age. This integration was made
possible by several m eans: pilgrimage, priestly affiliations, and,
m ost im portantly, sectarian organization. The great temples of
South India form the apex of this complex system of affiliational
linkages.
Temples serve the moral order o f medieval and modern South
Indian society in that they are the most im portant context in which
moral definition occurs. Temples define in the sense that they
determine or fix boundaries o f social groupings and social space
reckoned to m ark off the com m unity of co-sharers in the worship
of a particular god. A Tamil term for ‘lineage’ is p an gali , derived
from the verb pa n k u , ‘to share’, and the lineage in South Indian
kinship is no less delimited by shared devotion to a lineage tutelary
(kuladeva) than by genealogy. Similarly, a village is no less de­
limited by the worshippers o f its gram adeva than by its adm inistra­
tive boundaries. It is therefore interesting to notice that m ariyadai
is the term most frequently used to refer to the receipt of transvalued
substances from a god (prasadam ). This Sanskrit-derived word
(.m aryada) is glossed ‘honour’ in most usage. Thus, p a riva tta m ,
or the cloth which has part of the vestments of the god and distri­
buted as prasadam to worshippers, is k o yilm ariyadai, that is,
‘temple honour’ and it is worn on the head o f the recipient as a
a crown.234 M a rya d a has the meaning o f boundary, border, limit
(as in frontier or river bank); it also has a moral aspect as in the
phrase ‘limits of. . . propriety, rule, or custom .235 Hence, if in the
234 T am il L exicon, v. 4, p. 2518. The author gratefully records his indebtedness
to Carol Breckenridge for having brought this entire large and im portant issue to
his attention; discussions with Arjun A ppadurai and Kathi L. Rose were also valu­
able.
2';; M onier-W illiam s, A Sanskril-E ngtish D ictionary, p. 791.
transactional nexus of a god and its worshippers, which a temple
can be construed to be, the generosity of deity is expressed by con­
ferring ‘honours’ in the prasadam given to and accepted by devotees
of the god, then those who may offer to and receive from the god
must be seen as defined with respect to some criterion. Criteria
may be socially denotative; they may also be spatially denotative;
in all cases these criteria are viewed as morally valanced, that is
possessing a dimension of appropriateness in nature and in conduct.
Notwithstanding the im portance of temples in South Indian
culture during the Vijayanagara period and notwithstanding the
fact that m ost of what is known of the time comes from temple
inscriptions, little is known aggregatively about temples of the
m acro region. Almost nothing of their number, their spatial and
tem poral distribution, nor their cultic affiliations are known.
An opportunity to fill this gap in knowledge is provided for Tamil
country at least by the publication over the past decade of a series
o f volumes which form part of the 1961 Census conducted in the
M adras State. These data have been analysed by the present
writer and the results published elsewhere. Certain of the findings
and their im plications may be summarized here.236
2 » India, Census Commissioner. Census o f India, 1961, vol. IX , M adras State,
pt X l-D . ‘Tem ples of M adras State', 7 vols. The analysis of these temple data
appears under the tit 1c, 'Temples of Tamil Country, 1300-1750 A .D .', in a symposium
edited by the present writer dealing with various aspects of South Indian temples
in The Indian E conom ic and Social H istory Review , vol. 14, n o .l (1977); also pub­
lished under the title: S outh Indian T em ples: A n A nalytical R econsideration, ed.
Burton Stein, Vikas, New Delhi, 1978. Before considering the significance of these
data taken from the 1961 Census volumes o f Tem ples o f M adras S ta te (and Tem ples
o f Tam ilnadu), it is well to briefly catalogue the possible error factors which together
caution care in using the data and interpreting the findings derived from them.
Two m ajor problems are the reliability o f the m ethod used for collecting the temple
data and the definition o f 'im p o rtan t’ applied to the temples enum erated in the
census volumes. N o trained personnel were used in the census operations to verify
inform ation on the temples provided by questionaire. Hence, the age o f temples,
one o f the dimensions to which saliency is given in the present discussion cannot
be taken with the confidence that m ight have been wished. M oreover, some o f the
individual returns o f deity identification, as given in the printed volumes, are plainly
wrong (e.g., the M inakshi-Sundaresvara tem ple in M adurai is returned as dedicated
to a deity other than Siva, i.e., 'O ther Deity’. This too contributes uncertainty to
the findings discussed here. Even if such problems did not exist, another m ajor
source of error derives from the temples actually enum erated. These were required
to be "im portant’ shrines, not simply shrines of village deities or 'bajanai koyils’.
W hat constitutes 'im portant' here is problem atic of course, for all shrines are
 456
Tables V III-3 and V III-4 summarize data on 2035 temples which attained an im portant status between
a .d. 1300 and 1750. Figures V III-1 and VIII-2 provide a graphic and schematic display o f these tabular data.

Peasant State and Society

T a b l e VIII-3
Temples of Tamil N adu, a .d . 1300-1750, by Deity, M andalam and District3
N-2035 Figures in parentheses are percentages of totals.

Deity T otal by
Siva Vishnu M urugan Ganesa Amman Other District

T O N D A IM A N D A L A M
Chingleput Dt* 82(46) 62(34) 6(3) 8(4) 0(0) 23(13) 181(100)
N orth Arcot Dt** 77(38) 69(34) 9(4) 12(6) 21(10) 15(8) 203(100)

in Medieval South India

South Arcot Dt
Tindivanam T. 8(35) 11(50) 0(0) 1(5) 2(10) 0(0) 22(1000)
Total 167(41) 142(35) 15(4) 21(5) 23(6) 41(9) 406(100)
‘ Excluding M adras City
‘ ‘ Excluding Tiruvannam alai T.

‘im portant’ to some body of worshippers. In Chingleput district (vol. i) the 1961 Census temple volume estimates that about twenty
per cent of temples in the district were included in the designation ‘im portant tem ples’. The criteria used to determine the importance
of temples were that they be structural shrines of durable materials, that they be consecrated, and that they possess means for carrying out
worship under priests. The inclusion of some, but not all shrines, raises questions about coverage: another source of coverage error for
the Tamil country as a whole is the absence of inform ation on the temples of Tanjavur. Temples o f Tanjavur are absent because data on
the estim ated 1809 ‘im portant’ temples of that district had not been presented in a form com parable with temples of other Tamil districts.
This creates an im portant gap in total coverage and is the m ore im portant since T anjavur can have been expected to provide a substantial
num ber o f temples.

Deity Total by
Siva Vishnu M urugan Ganesa Amman Other District

N A D U V IL-N A D U
South Arcót Dt* 75(40)’ 57(30) 7(3) 16(8) 25(13) 11(6) 188(100)
Tiruchirapalli Dt
Peram balur T. 46(44) 38(37) KD 2(2) 11(11) 5(5) 103(100)
N orth Arcot Dt
Tiruvannam alai T. 3(33) 2(21) 0(0) 0(0) 1(11) 3(33) 9(100)
Salem Dt. A ttur T. 10(31) 7(22) 0(0) 4(13) 10(31) 1(3) 32(100)
Total 134(40) 101(32) 8(2) 22(6) 47(14) 20(6) 332(100)
*Excludittg C hidam baram T. and Tindivanam T.
KONGUM ANDALAM
Coim batore Dt 63(17) 88(23) 16(4) 37(10) 145(39) 25(7) 374(100) Vifayanagara
M adurai Dt, Paini T. 2(25) 4(51) 0(0) 0(0) 1(12) 1(12) 8(100)
Salem Dt, Tiruchengode T. 6(21) 9(32) 0(0) 2(7) 10(36) 1(4) 28(100)
N am akkal T. 13(19) 16(23) 1(1) 8(11) 25(36) 7(10) 70(100)
Tiruchirapalli D t, K arur T.
8(22) 9(24) 2(6) 0(0) 17(45) 1(3) 37(100)
Total 92(18) 126(24) 19(4) 47(9) 198(38) 35(7) 517(100)
State and Society

PA N D iM A N D A L A M
R am anathapuram Dt 65(37) 33(19) 10(6). 16(20) 34(9) 16(9) 174(100)
Tirunelveli Dt 89(30) 90(31) ¡5(5) 28(16) 48(10) 25(8) 295(100)
M adurai Dt * 32(19) 42(25) 9(5) 12(22) 38(7) 38(22) 171(100)
K anyakum ari Dt 19(14) 21(15) 0(0) 35(25) 36(25) 29(21) 140(100)
205(26) 188(24) 34(4) 91(12) 156(20) 108(14) 780(100)
‘ Excluding Palni T. Total
ABBREVIA TIO NS: Dt D1S1R1C1 T. - Taluk
457

‘‘India (Republic), Census Commissioner, Census oj India, 1961,\.\: V, M adras , pt. xt-D. Tem ples o f M adras S ta te: i., Chingleput and M adras
 458
T able V I I I - 4
Temples of Tamil Nadu, A.D . 1300-1750, by Deity, Period and M an dalam

Peasant State and Society

N =2035 Figures in parentheses are percentages of totals.

Deity Total by Total of

Siva Vishnu M urugan Ganesa Amman Other M andalam Period

a. .D . 1300-1450
a

N a = 179(9% of N)
29(47) 23(38) 2(3) 3(5) 2(3) 2(3) 61(100) (34)
Tondaim andalam
5(55) 4(45) 0(0) 0(0) 0(0) 0(0) 9(100) (5)
Naduvil-nadu
3(49) 1(17) 0(0) 1(17) 0(0) 1(17) 6(100) (3)
Kongum andalam
44(43) 23(22) 4(4) 3(3) 13(13) 16(15) 103(100) (58)
Pandim andalam

in M edieval South India

81(45) 51(29) 6(3) 7(4) 15(8) 19(11) 179(100) (100)
Total for Tam ilnadu

b. A .D . 1450-1550
N b =331 "(16% of N)
25(48) 16(30) 2(4) 5(9) 0(0) 5(9) 53(100) (16)
T ondaim andalam
24(43) 17(30) 3(5) 4(7) 6(11) 2(4) 56(100) (17)
Naduvil-nadu
24(24) 24(24) 6(6) 3(3) 35(36) 7(7) 99(100) (30)
K ongum andalam
37(30) 30(24) 6(5) 13(11) 19(15) 18(15) 123(100) (37)
Pandim andalam
110(33) 87(26) 17(5) 25(8) 60(18) 32(10) 331(100) (100)
Total for Tam ilnadu

City (1965): ii, Tiruchirapalli and South Arcot (1966): iii. Coim batore and Salem (1968): iv. N orth Arcot and Nilgm s (1968): v. Kanya
Kumari and Tirunelveli (1968); vi. Ram anathapuram and M adurai (1969): and vii. T hanjavur, pt (under the altered title of Temples of
Tam ilnadu’ (1971).

Deity Total by Total by

Siv Vishnu M urugan Ganesa Amman Other M andalam District

c. .D . 1550-1650
a

N c = 626 (31%; o f N)
T ondaim andalam 38(34) 38(34) 5(5) 6(5) . 8(7) 16(15) 111(100) (18)
Naduvil-nadu 46(41) 35(31), 3(3) 5(5) 17(15) 5(5) 111(100) (18)
Kongum andalam ■32(20) 42(27) 4(3) 12(8) 53(35) 10(7) 152(100) (24)
Pandim andalam ' 67(27) 61(24) 11(4) 30(12) 49(19) 34(14) 252(100) (40)
T otal for TamilnadtT 182(29) 23(4) 53(9) 127(20) 65(10) 626(100) (100)
d. A .D . 1650-1750
'S:
Njd -8 9 9 (44% o f N)
3
Tondaim andalam
5ÍS5
75(42) 65(36) 6(3) 7(4) 13(7) 15(8) 181(100) (20)
Naduvil-nadu
Kongum andalam
; 59(39)
34(13)
45(29)
59(23)
2(1)
9(3)
. 13(8)
31(12)
24(16) 13(8) 156(100) (17) I-s
100(43) 17(6) 260(100) (29)
Pandimandalam 57(19) 72(24) 13(4) 45(15) Co
75(25) 40(13)' 302(100) (33)
Total for Tam ilnadu 225(25) 241(27) 30(3) Ci
96(11) 222(25) 85(9) 899(100) (100)
TOTALS: a . d . 1300-1750 a
is
T ondaim andalam ' 167(41) 142(35) 15(4) 21(5)
ft,
. 23(6) 38(9) : 406(100) (20) Oo
Naduvil-nadu- 134(40) 101(32) 8(2) 22(6) 47(14) 20(6) 332(100) (16) rv
K ongum andalam ' 92(18) 126(24) 19(4) 47(9) 198(38) 35(7) 517(100) (26)
Pandim andalam : 205(26) 186(24) 34(4) 91(12) 156(20) 108(14) 780(100) (38)
Total for Tam ilnadu 598(29) 555(27) 76(4) 181(21) 424(21) 201(10) 2035(100) (100) 4^
SO
 Figure V III-1: Temples o f Tamil N adu, a . d . 1300-1750,
by M an dalam , Deity and Period
N = 2035 Source : Data of Table

DEITIES P E R IO D S
W/(HSiva E§3Ganesa a. 1 3 0 0 -1 4 5 0
b. 1 4 5 0 -1 5 5 0
E3 Vishnu n Amman
c. 1 5 5 0 -1 6 5 0
H |M u r u g a n 1 I Other d. 1 6 5 0 -1 7 5 0
Number of Temples

a b e d a b e d a b e d a b e d
Tondaimandalam Naduvil-Nadu Kongumandalam Pandimandalam
 Figure V III-2: Proportions o f Temples o f Tamil Nadu
by Deity, for F our Periods, a .d , 1300-1750
N=^2035 Source: Data o f Table
Percentage of Totai Tempies
 Variation in the num ber o f temples by m andalam is a conspicuous
feature in the tables and figures. The 2035 temples which attained
im portance between a . d . 1300 and 1750 are unevenly distributed
over the four m andalam s taking all o f the periods together : Tondai-
m andalam , 406 temples; Naduvil-nadu, 332 temples, Kongu-
m andalam , 517 tem ples; and Pandim andalam , 780 temples.
Respectively, the proportion o f each m andalam to the total for the
entire period is: twenty, sixteen, twenty-five, and thirty-nine
per cent. There are no reasons for assuming a uniform distribution
o f temples over these m andalam s , of course. Variations would
stem from differences in absolute populations, population densities,
and relative wealth. Tests o f such factors yields no significant
relationships which were m ore persuasive than the simple relation­
ships o f the num ber o f taluks and the num ber o f temples. This
latter relationship is shown by aggregating taluks according to the
ancient m andalam territories which yields the following :

T a b l e VIII-5
Temples, a .d . 1300-1750, by M odern Taluks Arranged by
M andalam s

Taluks Temples
M andalam N um ber P ercen t P e rc e n t Number

Tondaim andalam 19 27 2° 406

Naduvil-nadu 9 13 16 332
K ongum andalam 13 20 25 517
Pandim andalam 28 40 39 780
Total 69 100 100 2035

While there is considerable variation in the size and population

o f taluks of modern Tamil N adu State, the finding of a close
relationship o f taluks and temples in the period a . d . 1300 to 1750,
prom pts the suggestion that the m odern taluk may have an historical
validity n o t usually recognized, that m ost have been cultural areas
for the last five hundred years containing at least one temple of
im portance within its area.
The general increase in temple construction over all four m an­
dalam s in each period shown graphically in Fig.VIII-1, is somewhat
startling in only one particular. T hat is, the single instance o f a
decrease in Tondaim andalam in the period, a . d . 1450-1550. In
part, this is an artifact o f the data. There is no com parable inform a­
tion on temples for a substantial part of Tondaim andalam because
the northern taluks o f this ancient territory are included in the
m odem states o f K arnataka and A ndhra Pradesh which did not
conduct temple surveys. However, from contem porary inscrip-
tional evidence, it is obvious that northern Tondaim andalam , like
much of Kongum andalam , witnessed intensive temple-building
activity during the post-Chola period. And, it is precisely in the
period from a . d . 1450 to 1550 that the short-lived Saluva dynasty
o f Vijayanagara controlled the fortunes o f the empire from its
base in northern Tondaim andalam . The growth o f the Tirupati
temple in this period may be taken as symptomatic o f the general
support for existing and new temples o f this region by chiefs whose
fortunes rose with those o f the Saluvas and who drew resources
from a large region, including the southern parts o f Tondaim an­
dalam .237
A part from the anom aly o f Tondaim andalam in the fifteenth
century, the increase in structural temples in the four m andalam s
of the m acro region is striking as is variation in this growth. C on­
sidering variation from one period to the next, the doubling of
im portant tem ples. in Tondaim andalam and Pandim andalam
from a .d . 1550 to 1650 appears modest in com parison with the
sixfold increase in N aduvil-nadu between a .d . 1450 and 1550,
and the m ore dram atic sixteenfold increase in Kongum andalam
at the same time. These relatively short-termed variations may in
turn be com pared with changes in the rates among mandalam s
over all o f the periods. In Kongu the transform ation o f older
modest shrines into m ajor temples or the construction of new,
large temples increased by a factor o f forty between a . d . 1300 and
1750, and in Naduvil-nadu the increase was an impressive sixteen-
fold; whereas in Pandim andalam , with the largest num ber of
such temples, the increase between the first and the last periods
was about threefold, and in Tondaim andalam it was even less.
Finally, the 1961 census data reveal differences in the propor-
237 Stein, ‘Tirupati Tem ple’, pp. 53, 61-2, especially Table 3, p. 53 recording grants
o f land to the Tirupati temple between a . d . 830 and 1628 as 168 o f which all but
21 came between a . d . 1456 and 1564 and grants of money as 397 o f which all but
126 came during the same period.
tions and num ber of m ajor temples presided over by the six types
of deities utilized in the survey. Considering the entire period
o f four and a half centuries, 2035 temples analysed here show
Siva temples to comprise twenty-nine per cent o f the total, Vishnu
temples twenty-six per cent, M urugan temples four per cent,
Ganesh temples nine per cent, Amman temples twenty-one per cent,
and ‘Other Deities’ ten per cent. Each of the m andalam s contribute
to this array in quite different ways. Thus, in Tondaim andalam ,
Siva temples comprise forty-one per cent for all periods while in
Kongum andalam , Siva shrines comprised only eighteen per cent
o f the m ajor temples. In contrast, Amman temples in the former
territory comprised a mere six per cent whereas in Kongu, Amman
temples accounted for thirty-eight per cent of its m ajor temples
between a . d . 1300 and 1750..
Or, looked at slightly differently, the six types of deities can be
reduced to three classes o f gods to elicit certain differences. Siva
and Vishnu can be considered together as universal H indu gods
o f the highest order; M urugan and G anesha may be regarded as
secondary universal deities in medieval Hinduism ; while Amman
and ‘O ther Deities’ are essentially local, tutelary gods. Arranged
thus, Siva and Vishnu temples constitute the overwhelming majority
o f temples for Tondaim andalam over all o f the periods: seventy-six
per cent; and in Naduvil-nadu, these two deities accounted for
seventy-two per cent o f those presiding over m ajor temples. In
both m an dalam s , Siva temples dominate. By contrast, only forty-
two per cent o f the m ajor temples of K ongu were the resort o f Siva
or Vishnu, and here Vishnu temples predominated in about the
same proportion as Siva temples in Tondaim andalam and Naduvil-
nadu (i.e. ab out six per cent). In Pandim andalam Siva and Vishnu
temples in almost equal proportion, comprised half o f the m ajor
temples from a . d . 1300 to 1750.
M urugan and G anesha temples were unevenly distributed over
the four m andalam s considering the entire period. In Tondai­
m andalam these secondary, universal deities constituted nine
per cent o f all m ajor temples with G anesha temples three times as
numerous. The elephant-faced deity also dom inated over his
puranic brother as the object of m ajor temple worship in K ongu­
m andalam and in Naduvil-nadu where .these two deities com ­
prised thirteen and eight per cent o f temples respectively.
While all o f these findings o f the 1961 census survey o f temples
pertaining to temples o f the Vijayanagara period deserve attention,
concentration will be given to the interaction of Siva and Amman
temples.
. A detailed argum ent on the question o f the relationship between
Siva and Amman, has been given elsewhere by the author.238
A summary of that argum ent begins with the observation of what
appears to be the displacement or supercession o f m ajor temple
devoted to Siva by those devoted to goddesses. Such a notion
could be taken from Figure VIII-2 where there is shown to be a quite
startling reciprocal symmetry in the curves plotting the proportion
o f each of these types of temples over the four periods. This finding
jars conventional expectations based upon the dominance of Siva
worship among Tamilians from Chola times at least and the dyna­
mism of the Saiva Siddhanta movement during the medieval period.
An alternative to displacement as an explanation o f the finding
on Siva and Am m an deities is one of complementarity, and this
proposed explanation prom pts a useful consideration o f the re­
lationship of worship and sectarian affiliations of Tamilians during
the Vijayanagara period, the subject of the present section.
Complementarity in the relations between m ajor Siva temples
and m ajor shrines devoted to goddesses rests in two historical
features o f South Indian religion. One is the establishment of
separate d evi shrines within the precincts o f and thus as indepen­
dent components of Siva and Vishnu temples from the eleventh
century o n w ard ; the other is that Siva deities and goddesses alike
have served as territorial deities.
The progressive enhancem ent o f devi worship as a com ponent
o f the highest form o f puranic H indu praxis in South India and the
guardianship of goddesses over Tamilians combined to produce
the startling prominence o f Amman temples in parts o f Tamil
country, especially in Kongu and in Pandya countries. It is to be
stressed th at these often impressively enshrined goddesses are
correctly treated as independent deities by the 1961 census survey;
they are not regarded as the consorts of either Siva o r Vishnu. How­
ever, religious scholars have noted that though the m any goddess
shrines in Tamil country do shelter seemingly independent divinities,
these are often h.agiographically linked, at least weakly, to a m ajor
Vedic god, usually-Siva.239 This linkage may be considered to
238 See note 236.
2W W hilehead, op. cit., pp. 17ff, in which categorical distinctions between Vedic
strengthen the complem entarity hypothesis advanced here, especial­
ly when it is considered with other evidence which may be adduced
on religious affiliations in K ongu, where the highest proportion
of Amman shrines came into im portance during the Vijayanagara
period.
. How the increasingly num erous Amm an shrines were com-
plementarily linked to Siva shrines, cannot be ascertained from
evidence of the Vijayanagara period. The process must be inferred
with the assistance of late eighteenth century evidence from British
sources (e.g. the B aram ahal R ecords and Buchanan’s Journey. . . .)
and from more recent ethnographical research. From these post-
Vijayanagara vantage points, the contact of Amman shrines with
established canonical shrines seems to have been achieved through
affiliations o f priests and sectarian and pilgrimage networks.
Non-Brahm an priestly custodians (pandâram s) of m ost Amman
shrines appear to have become participants in the management of
great shrines o f Siva as well as Vishnu where Brahm an priests were
dom inant ritual functionaries. Entry into and participation within
these great shrines by non-Brahm an religious leaders was achieved
through endowments (k a tta la is) to these shrines in a process
recently elaborated by Carol A. Breckenridge. In addition, there
were connections o f such non-Brahm an religious leaders with
sectarian organizations through m athas located at the great shrines
of the canonical gods. Sectarian m athas were the seat o f sect
organizations, and there were both Brahm an and non-Brahman
orders. To the latter m ath as , under their m udaliyâr âcâryas re^-
ference . has. been made above (chapter VI). These institutions
were established during the later Chola period in several parts of
the. southern Tamil country and were regularly attached to major
Siva temples.240 Discussing these m athas, K.A. Nilakan-ta Sastri
noted that while they were first established only at major religous
centres,, they soon ,
deities and village deities are delineated, the form er possessing the following at­
tributes: they are ‘forces o f nature’, not ‘facts o f Hfe’; they are male, not female;
they are. worshipped according W h h a kti principles of ahim sa , not propitiated with
anim al sacrifices.; and they are officiated by Brahm an pujaris, not pandurann. How­
ever, in the lengthy story of the goddess Am m avaru or Ankam m a (pp. 126ff), this
goddess is consistently associated with Siva and symbols of Siva, especially (p.p.
132-8). Also, H. Zim m er, M y th s and S ym bols in Indian A rt and Civilization, H arper
Bros., T orchbook. New York, 1962, pp. 189-216.
240 A .R .E ., 1909, para. 53, pp. 101-2; Rajam anickam , op. cit, pp. 236-8.
 spread all over the land until alm ost every im p o rtan t tem ple cam e to
possess one o r m ore m at has functioning in close proxim ity to it. In course
o f tim e they grouped them selves aro u n d a lim ited num ber o f santanas
o r successors o f gurus, each having a central m at ha w hich was looked up
to for guidance by a num ber o f subordinate mathas in different places.241

He observes further that while the heads of these orders were often
ascetics who could own no property,
their mafhas often owned vast estates earm arked for their m aintenance
and the encouragem ent o f learning and the arts. . . . T he w ell-to-do
householder has ever been ready to m ake gifts (ddna) to the orders because
he was assured o f a good berth [s/c] in the oth er w orld as m uch for his
daría as the ascetic for his renunciation and austerity.242
W hat N ilakanta Sastri might further have pointed out in connec­
tion with the wealth and influence o f these heads o f religious orders
(m ath adipatis) is that they utilized their material resources to
contribute to the growth of temples, in the m acro region by their
often substantial endowments from the wealth which was conferred
upon them by their followers; they also constituted the clearest
linkage element am ong the various shrines o f Tam il country through
their influence with peasant followerships for whom access to
m ajor regional temples devoted to the worship of Vedic deities was
otherwise limited. The non-Brahm an santanas were the crucial
intermediaries between the priesthoods and worshippers of local
subcaste temples and the great temple centres of Tamil country.
It is therefore no coincidence that the proliferation o f major
structural shrines devoted tp peasant subcaste tutelaries and the
non-Brahm an m atha m ovement occurred at about the same time.
These two developments m ust be viewed as causally linked. Such
a view addresses the question of vertical connections among shrines
raised by the recent work o f Brenda Beck.243 It accords well with
the evidence from K ongu provided by the B aram ahal R ecords
and Buchanan, and it is a view which finds impressive inscriptional
and other support from the Vijayanagara period according to the
recent doctoral research of Arjun Appadurai whose writings have
shaped much o f the present discussion.
In a paper presented recently, Appadurai has offered a powerful
241 D evelopm ent o f Religion in South India, p,. 118.
2 « Ibid., p. U7.
243 Peasant S o ciety in K oñku, University o f British Colum bia Press, Vancouver
1973.
interpretation o f Vaishnava sectarianism during the Vijayanagara
period.244 His argum ent extends and bolsters suggestions outlined
above with respect to the relationship o f Siva and Amman temples.
This investigation of the vadagalai-tehgalai conflict and his original
form ulations of Vaishnava sectarian history in South India bear
centrally upon the question o f relationships among temples in the
m acro region, between temples and sectarian organizations, and
between both o f these and various levels o f political authority.
His findings address with particular cogency the m anner in which
various levels o f shrines during the Vijayanagara age achieved
and m aintained a coherent integration in the absence o f formal,
bureaucratic, ecclesiastical structures.
A ppadurai’s work on medieval Vaishnavism is part o f a larger
research work dealing with the Sri Parthasarathiswam i temple in
Triplicane during the nineteenth and twentieth centuries. In that
larger study, he argues that the distribution of temple honours is
central for an understanding o f Hindu temples in modern South,
Indian society. His research and that of Carol A ppadurai Breck-
enridge on the M inakshi-Sundaresvara and Ramesvaram temples245
have brought forward considerations o f the greatest importance
for understanding not only the m odern temple in its social context,
but the temple in its medieval setting as well.
The work o f these two scholars outlines a redistributive process
involving temple honours and m aterial resources. Appadurai
delineates four sets of transactions involving honours during the
Vijayanagara period: the sta n a tta r o f temples conferred honours
(e.g. p rasddam ) on w arrior chiefs and kings; kings and chiefs
244 W orkshop on Religion in South India at Bucknell University under the auspices
of the Conference on South Indian Religion, June, 1975 ('H o n o r and Conquest:
W arrior-K ings and Vaisnava Sectarianism in South India 1350-1700’). A slightly
altered version presented at a conference on ‘H onor and H onors’, University of
Chicago, June, 1975 (‘Why Think A bout H onors?’). This paper is published in
the symposium num ber of the Indian Social and Econom ic H istory R eview devoted
to South Indian temples and under the editorship of the present writer, mentioned
above, note 236.
245 In two papers e n title d : ‘Betel N ut and H o n o rs; Exchange Relationships
and Temple Entry in a South Indian Tem ple’ presented at the Association o f Asian
Studies meeting in San Francisco, March, 1975, and 'N otes to an H onor-Interested
G eneration: H onor in Sfluih India’ for a Conference on H onor and Honors,
University of Chicago, June 1975. Versions o f this work are published in Con­
tributions to Indian Sociology as well as the symposium num ber, edited by Burton
Stein, of the Indian Social and E conom ic H istory R eview ; note 236.
conferred honours, e.g. insignia and regalia upon sectarian leaders;
temple leaders conferred honours on sectarian leaders (e.g. p rasâ-
darri) ; and sectarian leaders conferred honours on chiefs and kings
(e.g. investiture). M aterial transactions among these various actors
involved endowments by chiefs and kings to temples and gifts
to sectarian leaders as well as endowments to temples by sectarian
leaders. These transactions in the media of honours and material
resources are equal or symmetrical with one exception.
T hat is, while between chiefs and kings, on the one hand,
and sectarian notables, on the other hand, honours are
exchanged symmetrically, material resources went only from
chiefs and kings to sectarian leaders. Appadurai asks whether, in
the relationship between chiefs and kings and sectarian leaders,
there is a transaction in some medium or of some sort to correspond
with the material resources conferred by the former upon the
latter. That is, what is the gain o f political personages in their
transactions with sectarian leaders? The answer: kings and
chiefs received political constituencies. Sectarian leaders were
‘crucial intermediaries for the introduction, extension, and legitima­
tion o f w arrior control over. . . [peoples] and regions. . . [which]
might otherwise have resisted conquest’. A ppadurai concludes:
P ut differently, it m ight be said th at the cerem onial exchanges o f h o n o u r
between w arrior-kings and sectarian leaders rendered public, stable and
culturally appropriate an exchange at the level o f politics and economics.
These w arrior-kings b artered the control o f agrarian resources gained by
m ilitary prowess, for access to the redistributive processes o f temples,
which were controlled by sectarian leaders. Conversely, in their struggles
with each other, and their own local and regional efforts to consolidate
their control over tem ples, sectarian leaders found the support o f these
w arrior-kings timely and profitable. Em pirically, and diachronically,
this relationship is neither simple n o r transparent. It is a complex sym bio­
sis in which m obile figures, o f b o th types, augm ented and sustained each
other.246
To the dynamism o f the Vijayanagara period engendered by
relations am ong temples, sects, and the w arrior élite, must be added
the modified functions of the right and left divisions of castes.
During the Vijayanagara period, the potential or relative character
of the right and left divisions of the Chola period continues ; the
246 Cited from A ppadurai’s doctoral thesis, entitled ‘Temple’ Politics in South
India; The Case of the Sri Parthasarathisw am i Tem ple’, ch. 2, pp. 14-15, Com ­
mittee on Social T hought, University of Chicago, 1976.
divisions are not corporate social units as are castes and sects247,
but divisions, like the segmentary state, underwent modifications.
Particularly the divisions were differentially linked to two com m an­
ding institutions of the age: temples and royal figures.
The question of functions o f the dual divisions during the Vijaya­
nagara period can scarcely be separated from the persistent com ­
position o f each division. During the Chola period, there is only
the most inferential evidence on their composition. However,
thanks to the evidence provided by foreigners in South India
during the eighteenth century and after, it is possible to specify in
considerable detail the composition of each division in various
parts o f the m acro region.
The most complete inventory o f the constituent social units of
the divisions becomes available in the late eighteenth and early
nineteenth centuries. At that, time, references to the dual divisions
are often to urban contexts utterly foreign to the Chola age. Thus,
while it would obviously be im proper to consider this eighteenth
and nineteenth century evidence for an analysis of the dual divisions
during the Chola age — that is to provide eighteenth century ans­
wers to tenth century questions — it is appropriate to consider
this later evidence for m ore proxim ate periods when urbanization
had in fact become as significant as it was during Vijayanagara
times. M oreover, the tem poral reach involved is not different
from that o f one o f the m aster craftsmen of rural historiography,
M arc Bloch, whose classic work on the agrarian system of France
to the eighteenth century provides examples o f the use o f the relati­
vely full evidence o f the eighteenth century to clarify much earlier
situations.248
European travellers and early adm inistrators are responsible
for m uch of the detailed understanding we possess about the func­
tions and com position o f the dual social divisions. Because these
divisions fall outside brahmanical culture — Brahmans being
24.7 W here 'sect’ is equivalent to sam pradaya , 'W hen a guru or set of gurus is
recognized by a series o f disciples, a particular cultural tradition (sam pradaya)
is established, and its authentic transmission is traced through a chain o f disciples,
who in turn become gurus to the next generation. In this sense, every Indian sect
or philosophical school forms a cultural tradition’, from his 'The R adha-K rishna
Bhajanas of M adras City’, in M ilton Singer (ed.), K rishna: M yth s, R ites, and A t­
titudes, University o f Chicago Press, Phoenix Books, Chicago, 1968, p. 108. Also:
M. Monier-W illiams, H induism , L ondon, 1906, pp. 135-6.
248 French R ural H isto ry; A n E ssay on its B a s k C haracteristics , University of
California Press, Berkeley, 1966.
excluded from the divisions249 — and notwithstanding the referen­
ces to the divisions in inscriptions of the Choia period and later,
the South Indian brahmanical tradition completely fails to note
the divisions. However, Europeans of the eighteenth and early
nineteenth centuries were as struck by the divisions as they were
nonplussed by them. The Abbe J.A. Dubois, in India from 1793
to 1823, commented upon the divisions in the following way.
T here is yet an o th er division m ore general th an any I have referred to
yet, nam ely, that into R ight-hand and Left-hand factions. T h is appears
to be bu t a m o d em invention, since it is n o t m entioned in any o f the ancient
books o f the country. . . I do n o t believe that any idea o f this baneful in­
stitution, as it exists at the present day, ever entered the heads o f those
wise lawgivers w ho consider th a t they had fo u n d in caste distinctions
the best guarantee for the observance o f the laws .which they prescribed
for the people.250

The earliest enum eration o f the divisions was m ade by Pierre

Sonnerat, an agent of the French government in Asia from 1774
to 1781.251 Noting that ‘sudras’ were divided into right and left
hand divisions, Sonnerat listed fourteen groups of the right and
seven groups of the left. He also mentioned seven groups of
‘sudras’ who were totally separated fro m . either division. These
latter castes differ from those enumerated in other lists of the dual
division in consisting o f low status groups such as idaiyars (herds­
men), kusavars (potters), m uchchiyars and chitragaras (painters)
as well as fishermen, long-distance porters, hunters and other
forest folk.252 M ost others-w ho have provided lists o f the dual
divisions affiliate these particular ‘separated’ groups with the
right hand division and reserve the category of aloofness from or
249 A ppadurai notes that Brahm ans were 'faction leaders’ in the seventeenth
century right-left strife in M adras city. These Brahm ans were town m erchants
in com petition with other m erchant leaders and their factional strategies included
m aking use o f the divisional symbolism and recruitment presum ably to counter
the strategies o f their opponents. A ppadurai notes that Brahm an faction leaders
recruited support from any quarter, not from among established groupings. This
supports A ppadurai’s interpretation of context variability of the right-left divisions
in this colonial city where Brahm ans were involved in m ercantile activities and
thus contextualized in the divisions as m erchants. However, during the Chola
period, we hear o f no Brahm an m erchants, or artisans or others who might be
included within the divisions (A ppadurai, op. cit., pp.. 250-3).
250 H indu M anners, C ustom s and C erem onies , p. 24.
251 Voyage A u x Jndes Orientates et L a C hine , v. 1, Paris, 1782, pp. 95ff. C on­
sulted in the Janies Ford Bell C o llectio n , University o f M innesota.
252 Ibid., p. 61.
neutrality with respect to the divisions "to high status groups,
always Brahmans. The most exhaustive listing of groups of the
dual divisions was compiled under the direction o f Colin Mackenzie
during his career as soldier, engineer, and surveyor-general of the
M adras Presidency from 1786 to 1815. The work is entitled:
idahgai-valahgai jd tiy d r varalaru (H isto ry o f the L eft-H a n d and
R igh t-H an d C a stes).25i Here, the canonical ninety-eight castes
of the right-hand and the same num ber of left hand groups are
spoken of, though less than this conventional num ber are actually
listed. In addition, the flags and other symbols which each allegedly
used are inventoried. The form o f this Mackenzie docum ent’ like
others of that collection, affects an archaic style. Thus, the right
and left hand castes are enum erated in the M anu-like m anner of
caste generation through miscegenation. The listing of right and
left hand groups by Gustav Oppert, an early student of Indian
cultural history, was based upon a court record of the Chingleput
zillah m agistrate George Coleman, 25 July 1809. An Indian
judical official, A. Krishnaswamy Iyer, prepared the list for Oppert,
and the same list, with m inor exclusions, was published by Maclean
in the glossary volumes o f his M an ual o f M a d ra s A dm inistration.
The Maclean version of the divisions included sixty right hand
groups and seven left hand groups together with their symbols.254
A somewhat shorter list o f the constituent groups of the divisions
for the interior upland of Salem was prepared by the first British
adm inistrators of the tract called ‘the BaramahaF (modern Salem
district) in the 1790s. Forty right hand castes and eleven left hand
castes are enum erated here, but an alm ost equal number, forty in
all, are identified as ‘neutral’ (m adh yastam ) with respect to the
d u a l divisions. Among the ‘neutral’ castes are Brahmans who
appear to be excluded from all lists, but also many groups which
appear as right hand groups in other lists.255 The Sriniyasachari
list o f Tamil castes o f the right and left hand is based on a culling of
inscriptional and early- English records to elaborate the first list
provided by M. Srinivasa Iyengar in 1914. Srinivasachari’s dis-
253 'idangai-valangai jd tiy d r varalaru', Mackenzie M anuscript Collection, M adras
Oriental M anuscripts Library, no. R. 1572.
2-4 M anual o f the A dm inistration o f the M adras P residency in Illustration o f the
R ecords o f G overnm ent and the Yearly A dm inistrative R eports , Governm ent Press,
M adras, 1893, v. 3 , ‘G lossary’, pp. 1036-7. .
255 The B aram ahal R ecords; Section I l f ‘Inhabitants’, M adras, 1907, pp. v-vii.
cussion is the first by a m odem historian,256 For the southern
portion o f K arnataka, the lists o f Francis Buchanan and Lewis
Rice are im portant. Buchanan, reporting in 1800, restricted his
observations to the town and neighbourhood of Seringapatam in
m odern K arnataka state where he was commissioned by the Court
o f Directors o f the East India Com pany to report on the new terri­
tories acquired after the last M ysore war.257 In the late nineteenth
century Lewis Rice, the D irector o f Archaeological Researches
o f the state o f Mysore, enum erated almost the same eighteen
castes o f the right hand and nine o f the left hand am ong K annada
speakers on the basis o f census returns. Rice’s antiquarian interests
led him to search for antecedents of the divisions in K arnataka as
they existed in his day, but he had to acknowledge failure to find
docum entary evidence o f an earlier time.258 Evidence from among
Telugu groups are in fact enum erated in the lists o f Oppert, the
Mackenzie docum ent, and the Baramahal Records where they
are usually distinguished by the terms vadugu (‘northerner’) or
telinga. Only one discussion o f dual divisions am ong Telugus
exists, that of Subha Reddi in which nineteen right hand and seven
left hand groups are enum erated as a result o f this anthropologist’s
work with roving acrobats and other custodians o f caste lore.259
Table V III-7 presents only a portion of the confused evidence on
the composition of groups of the dual divisions as they were known
during the nineteenth century. From the hundreds o f castes
enum erated, only those with substantial numbers and those which
appear on at least two lists have been noted. M any of the groups
enum erated in the Mackenzie docum ent cannot be identified
with actual castes or tribes in the compendia o f Thurston and
A nanthakrishna Iyer.260 O ther groups are sub-divisions of castes
found in these compendia and are not listed separately.
256 C.S. Srinivasachari, "The Origin o f the Right and Left H and Caste Divisions’,
Journal o f the A ndhra H istorical Research S o ciety, v. 4 (1929). pp. 77-85.
257 A Journey from M adras through the Countries o f M ysore, Cartara. and M alabar,
3 vols., London, 1807. The decisions are referred to in v. 1, pp. 76-80.
258 B. Lewis Rice, M yso re; A G azetteer C om piled fo r G overnm ent, rev. ed., 2 vols.,
London, 1897. The discussion is in v.. ‘Mysore in G eneral', pp. 222-4.
259 N. Subha Reddi, ‘Com munity-Conflict among the Depressed Castes of
A ndhra’, M a n in India, v. 30, no.4 (Oct.-Dec., 1950), pp. 1-12. He makes the interest­
ing observation that while m ost Brahm ans remain aloof from the dual divisions
in A ndhra some Vaishnava Brahm ans associate with the division. This aberrant
feature is also m entioned in the M ackenzie history o f the dual divisions where Vai-
 Table V I I I -6

V a la n g a i-Id a n g a i D e s ig n a t io n s 3 '

Occupation Source. b.
Caste Nam e I II III IV V VI VII VIII

A G R IC U LTU R ISTS
Banajiga (K aravar) R R R R R R
M alaiyam an R R O
N attam an Vellala R R R o
Palli L L. L L L
Vokkaliga R R R
Reddi R O R O
Vaduga Vellala R R
Vellala R R O

M ERCHA NTS
Beri-chetti L L L L L L L
G ujarati R R
Jaina R R L O
K am m a R R O o
Kom ati R R/L R R R R/L R
Lambadi R R R R
L adar R R R
HERDERS
Golla R R R/L L O
K annidaiyar R R
K uruba R R R O

kanasa Brahmans are listed am ong the right hand division (ibid., p. 5).
260 Edgar Thurston, C astes and Tribes o f Southern India, 7 vols., M adras, 1909
and L.K. A nanthakrishna Iyer, The M yso re Tribes and Castes, 4 vols., Mysore,
1928-35.
a. Shown as usually spelled on at least two lists, these sixty-eight m ajor castes
are designated by the following symbols: ‘R ’ = ‘Right hand’ or valangal ; ‘L’--=‘left
hand’ or idangai; where castes are given as sometimes one or the other, ‘R /L ’; and
when they are given as neutral, 'O ’.
b. Sources: I: M anual o f the A dm inistration o f the M adras Presidency in Illus­
tration o f the R ecords o f G overnm ent and the Yearly A dm inistrative Reports, 3v.,
Governm ent Press, M adras. 1893, 3, ‘Glossary’, pp. 1036-7; IT. Pierre Sonnerat,
Voyage A u x Indes O rientates et L a C hine , Paris, 1782, pp. 95ff, consulted in the
James Ford Bell Collection, University of M innesota Library; III'. C.S. Sriniva-
sachari, ‘The Origin o f the Right and Left H and Caste Divisions’, J .A .H .R .S .,
4 (1929), pp. 77-85; IV : B. Lewis Rice, M yso re: A G azetteer Com piled fo r Govern­
m ent, 2 v., L ondon, 1897, 1, pp. 222-4; V: N. Subha Reddi, ‘Community-Conflict
 Occupation j n III IV V VI VII VIII
Caste N am e ___ ______________________________________ ______________
W EAVERS
Devanga L L L R
K aikkolar L R L L L L
Pattunulkarar R O
R angare R R R
Padm a Saiiyar R R R
Pattu Saiiyar R R R
Sedar R R
K annidiyan Seniyan R R
Vadugan Seniyan R R R

OILM EN
G aniga R R
Hegganiga L L
Ilai Vaniyar R O
N agara Vaniyar L L L
O ntierutu Vaniyar R O
Sekku Vaniyar R ^

PA IN TERS
C hitrakara R R ..
M uchchiyar R O R

ARTISANS
K am m alar L L L L L L L
Kanciyar R L
Vaddi R R
Veli T achchar R R

Am ong the Depressed Castes of A ndhra’, M an in India , 30, no. 4 (Oct.-Dec., 1950),
pp. 1-12; V I : Francis Buchanan, A Journey from M adras through the Countries o f
M ysore, Canara, a nd M ala b ar , 3v., London, 1807. 1, pp. 76-80; V II: idangai-valaiigai
jatiya r varalaru in the M ackenzie Mss. o f the M adras Oriental M anuscripts Library,
University of M adras, no. R. 1572; V III: F ort St. George, Governm ent of M adras,
The B aram ahal Records, sec. 3, ‘Inhabitants', Governm ent Press, M adras, 1907,
pp. v-vii.
 O c c u p a tio n " „
VI VII VIII
Caste Nam e

BASKETMAKERS
K uravar R O
M edara R R
V etakarar R L

PO TTERS
K um bara R R R
K usavar R O R R

M USICIANS
Dasigal R R
M elakarar R R

TO D D Y TA PPER S
Sannar R L R R
FISHERM EN
Besta R R R R
K araiyan R L
Pattanavar R O

LABOURERS
O ttiyar R O
U ppara R R R

W ASH ERM EN
Agasa R R R R R
V annar R R
BARBERS
A m battan R R /L R R
M angali R R
N ayinda R R R
ACROBATS
D om m ara R R

PRIESTS
M ariam m an Pujari R R
Vaduga Pandaram R O

H U N T ER S
Enati R R
Irular R O R

BEGGARS
K udukudupaigaral R R
 Occupation
1 II 111 IV V VI VII VIII
Caste Nam e
SCAVENGERS
Holeya R R
Pallar L L L
Paraiyar R R R R R

LEA THER W ORKERS

M adiga L L L L
Sakiligal L L L

The lists from which the above table was constructed raise
obvious problems. One is that only rarely do all or m ost of the
lists consider the same groups. Where this occurs, however,
there is moderately strong agreement (i.e. more than half agree
on the assignment of a division to particular groups), thus eight
of the m ajor caste groups o f the macro region may be designated
with confidence as o f the right or the left division (Banajigas,
Pallis, Beri Chettis, Kom atis, Kaikkolars, Kamm alars, Vannar,
and Paraiyars). M oreover, the compositions by language regions,
as those o f Rice for K arnataka and Subha Reddi for Andhra,
fail to register the variation which m ust have existed within such
regions.261 And, the sixty-eight caste groups comprise a small
fraction o f those in the region during the nineteenth century; the
1901 Census o f M adras Presidency enumerates over 400 caste
groups divided into twenty-one m ajor classes.262 In general,
sources o f the late nineteenth century neglect the division and thus
are o f no help on this m atter.263 One of the lists of the right and
261 Buchanan noted this variability when com paring compositions of the divisions
from Bangalore wilh those o f Seringapatam , op. cit., p. 79; F.W. Ellis, in an 1812
report on the divisions cited above, stated: . . throughout the country, where
these distinctions [of the right and left hand divisions] prevail, a general resemblance
in the customs regarding the Right and Left H and Castes [exists], yet this is liable to
so m any local variations, that there is scarcely a district or a town of note where
there is not some difference to be found in these customs . . .’, from M adras, Public
D epartm ent Consultations, v. 391, p. 1450, dated 6 M arch 1812.
262 Census o f India, 1901, v. 15. M adras, pt 2, ‘Index to Castes and Tribes’, pp.
156-7.
263 The M yso re G azetteer o f Rice is an exception, as are some o f the M anuals
and G azetteers of the M adras Presidency published in the late nineteenth and early
twentieth centuries. O f the latter, the following make some references to the dual
divisions: C.S. Crole, The Chingleput (L a te M adras) D istrict: A M anual, M adras,
1879, pp. 33-4; F.A. Nicholson, M anual o f the C oim batore D istrict in the M adras
left hand divisions used here is that o f Chingleput compiled for
an 1809 judgem ent by the magistrate George Coleman and used by
Oppert, as already noted. This district, which included the city
o f M adras, contained a high im migrant population of Telugu
and K annada speakers, and it may distort our expectations of
elaborateness for other places. Yet, the Coleman list is shorter
than a list subm itted about a year later for the same place to a
special committee of inquiry appointed by the Governm ent of
M adras. The report o f that committee in 1810 included petitions
from right hand persons enum erating 130 constituent castes in
their division, whereas, the Coleman list included only fifty-six
valangai castes.264
As might be anticipated, the affiliation of some castes according
to broad division is ambiguous. F or example, Kom ati merchants
are usually associated with the right hand division, but they are
referred to as sometimes o f the right and other times of the left
hand division by Sonnerat.265 Similar ambiguities are reported for
A m battan, barbers, by Sonnerat.266 K aikkolar weavers are retur­
ned a left hand group by some, but as a right hand group by Son­
nerat.267 The confusion in the K aikkolar case is mentioned in
early eighteenth century English record where it noted that the
Kaikkolars were ‘very fickle in their caste. . . sometimes of one
(division), sometimes o f another’.268 Fiat was resorted to in
1812 when it was ordered that in the territory under the adm inistra­
tion o f the East India Company, Kaikkolars were to be regarded
as belonging to the left hand division.269 This was an act of authori­
tative adjudication reminiscent o f the Vijayanagara period.

P residency , M adras, 1887, pp. 61-3; J.H . Nelson, The M adura C ountry: A M anual ,
M adras, 1868, pt 2, pp. 4-7; F.J. Richards, Salem G azetteer, M adras. 1918, pp.
125-6; and F.R. Hemingway, T richinopoly G azetteer, M adras, 1907, pp. 92-3.
These are brief references lacking, for the most part, in any appreciation for or
perception of the functions of the divisions. This is further substantiated by the
virtual absence o f any discussion of the divisions (one para. v. 7, p. 298) in Thurston,
op. cit., though m any castes are m entioned as affiliated with one or the other division.
264 M adras, Public D epartm ent Consultations, 6 M arch 1812, p. 1506. These
lists are not included in the report.
265 Op. cit., p. 55.
266 Ibid.
267 ibid.
2138 Referred to in M adras , Public D epartm ent Consultations, 6 March 1812,
p. 1416.
269 Ibid. .
 Sect affiliations played a part in the dual divisions o f the eighteenth
century and earlier. This was notably the case in A ndhra where
the divisions appear to have been late to develop.270 Here the
divisions, when they arose, did not use the designations ‘right and
left han d ’ but often used religious designations.271 Similarly,
Buchanan, in 1800, reported that in K arnataka artisans were
divided between two sects of Siva worshippers. Panchan Banajigas
were Lingayats and associated with the right hand along with
Lingayat agriculturists while non-Lingayat, but Saivite Panchalars
were leaders o f the left hand division.272
Problems posed by the eighteenth century lists are best understood
within the context of the eighteenth and early nineteenth century.
In only one essential respect can they be supposed to reflect condi­
tions of an earlier time. T hat is, the association o f the right hand
division was with agricultural activities and the left hand division
with mercantile and craft interests. This distinction between the
two divisions is suggested in the earliest records, but the distinction
is made m ore salient during the later period with the growth of
towns. U rban conditions throughout the M adras Presidency
during the eighteenth and nineteenth centuries produced stresses
and corresponding solidarity requirements which were anticipated
in the towns of the earlier Vijayanagara age. Teeming ‘Black
Tow n’ behind F o rt St. George in colonial M adras, where idangai-
valangai outbreaks regularly disrupted life, may have had few
earlier parallels in any earlier period in terms o f densities, diversity
o f residential groups, and economic and social competition. How­
ever, the stress o f progressive urbanization of South Indian society
from the twelfth century on was manifested in the relationships
between right and left caste groupings in many parts of the macro
region, and especially in the continually improving terms of ad­
vantage to the m obile craftsmen and traders who were the core of
idangai groups everywhere.
U rbanization as a factor in the evolution o f the South Indian
dual division of castes is given an im portance in the present discus-
270 K. Sundaram , S tudies in the E conom ic and Social C onditions o f M edieval
Andhra (A .D . 1000-1600), Triveni Publishers, M adras 1968, p. 89. He states that
medieval A ndhra was free o f this division, though the functional characteristics
o f the divisions appear clear from his discussion.
271 Subha Reddi, op. cit., p. 4.
272 Buchanan, op. cit., v. 1, pp. 77-8. -
 sion which is not found in two recent and im portant studies of the
right and left castes — that of Beck on K ongu and that of A ppadurai
on seventeenth century M adras. It has been argued here that the
bifurcated groupings o f middle and low castes into those o f the
right and those o f the left was based upon the essential' interests
of broad occupational groupings which gave rise to alliance systems
o f a potential or occasional kind; that conflict between the two
groupings was never the sole nor even a prim ary characteristic of
the divisions (though conflict was often expressed in terms o f right-
left caste confrontations); and that the m ajor activities o f these
potential groupings during the Chola period pertained to warfare
and colonization and to the assimilation of various peoples to the
expanding agrarian order. Temple urbanization of the later Chola
period was im portant in changing the power terms under which
mobile artisans and traders operated and therefore the bloc of
local groupings with whom such mobile groups were, on occasion,
allied.
A nother way o f conceiving the dual division in Chola times is to
differentiate two broad groupings o f local caste groups. First,
there was the cluster o f rural groups which focused upon the
redistributional nexus o f locality leaders — n attar or miivenda-
velar — who presided over nadu societies not simply as powerful
groups o f chiefs, but as ritually integrative yajam anas. These
were the valangai, local, and therefore variable, social formations
capable o f being mobilized for a variety o f purposes and cohering
not simply through shared and interdependent agrarian functions,
but also shared ritual affinities centred in the dom inant locality
leaders. C ontrasted with such locality clusters were those groups
whose livelihoods were not dependent upon fixed clienteles nor
centred upon the redistributional authority o f powerful local chief­
tains and clans, but rather oriented to an extensive consumer
network — any and all who could pay for their specialized products.
These included goods brought into localities through the itinerant
trade network o f the time, or goods produced by highly skilled
craftsmen (e.g., goldsmiths, stone and metal sculptors). This
second g ro u p ing — idangai — being tenuously connected to the
prevailing and norm ative system o f locality societies and, further,
tainted by association with ancient, heterodoxical religious affilia­
tions — Buddhist and Jaina — underwent a transform ation of
their standing beginning in the eleventh century in connection
with Chola-inspired temple construction, an activity widely imitated
by locality leaders throughout the macro region. The demand
from temples for the special craft skills and exotic ritual materials
controlled by idangai groups enabled them to claim increasingly
prestigious places in the society o f later Chola times.
Temples of later Chola Hinduism became the prime moral
centres of the age, and great temples caused the creation around
them of towns. Over time, such urban places incorporated and
centralized many functions previously dispersed over the country­
side: political authority, trade, and perhaps most im portant, the
earlier ritual primacy o f the local, yajam ana-n attar. Whereas
rural localities of n attar ritual primacy precluded idangai groups
from any but the m ost m arginal places, towns were places where
artisans, craftsmen, and m erchants along with those allied to them
in left caste clusters could be im portant and, indeed, thrive.
During the Vijayanagara period, this process o f urbanization
continued, responsive still to religious factors, but also to the
political and military centralization of the n ayakas and to the
significant economic functions which urban places acquired as a
result o f the deliberate autarchic economic policies which led,
am ong other things, to the destruction of the itinerant trade system
by the enterprising n ayakas. To the im portance of temple centres
as the focus o f ever-expanding pilgrimage networks and as the
seats o f the acharyas who com m anded large sectarian followings
was added the ever-present Vijayanagara kings. Their dharmic
rule was m ost succinctly expressed in their protection and support
o f these sacred centres; Vijayanagara kingship cannot be separated
from these purposes. Hence, it is not an accident o f preservation
that the historical record o f Vijayanagara kingship should exist
in temple inscriptions. It was here that the civil, as against the
military, aspect o f kingly rule was m ost clearly realized. Each
temple city was like the capital city itself; it was a world, and, as it
was the duty o f kings (rdjadharm a ) to protect and nourish the
world, the m scriptional record o f their custodianship is found
at all such centres. It is as stated by A.M. H o c a rt:
We have seen that king, priest, animal, tree, corpse, idol all represent
things it is desired to control. So do the temples and palaces in which
these cult-objects . . . are housed. So does the city in which stand the
temples and palaces; it is sacred in the same sense as they are, that is they
are equivalent to that on which the life o f the people depends. Only the
 city never stands for anything specific; it is never less than the whole
world, and its parts are the parts o f the world. . . ,21i
Certain o f the views o f the dual division o f Vijayanagara times
presented here are sympathetic with and find support from recent
work by Arjun A ppadurai and Brenda Beck. But there are im por­
tan t differences from these recent works too. Beck and Appadurai
have made im portant use o f the dual division to explain specific
aspects o f South Indian society: Beck, the bifurcation of Tamil
castes in the K ongu region and A ppadurai, conflicts in seventeenth
century M adras city. In the course of their work, each has de­
veloped a different set o f explanations o f the dual divisions in South
India, and each has emphasized different attributes of the divisions.
Thus, Beck sees the divisions as arising from two conditions:
(1) a long-standing differentiation of ‘immobile’ castes whose
claim to precedence was based upon landed wealth and control
of people through control of land and in contrast to which there
were ‘m obile’ castes whose wealth derived from crafts and trade;
(2) an equally well-established ambiguity in models of esteem
and precedence derived, on the one hand, from economic and
political dom inance with little concern about purity or ritual
standing, and, on the other hand, from precisely such a concern
as the basis o f precedence (in the fashion of K ongu Brahmans who
owned little land).
Beck then proposes that in K ongu and perhaps elsewhere in
South India where conditions are similar, two structures o f castes
within a region come into existence. In one o f these structures, fixed
landed groups develop strong clan organization and elaborate clan
and subcaste religious affiliations while in the other structure, taking
as its model the high status, but land-poor Brahmans o f Kongu,
m obile wealth is used to achieve ritual purity through strict em pha­
sis upon hierarchical social relations and exclusive devotion to
universal H indu deities. Here is a system o f stable equilibrium
expressed as an abstract and formal dualistic model in which a
single Kongu society is divided into essentially non-conflicting,
com plementary halves.
Appadurai proposes a different conception.274 He speaks of a
213 Kings a nd Councillors , p. 250.
274 A ppadurai, ‘Right and Left H and Castes', p. 225-6, no.22 contains a trenchent
critique o f Beck’s argum ent.
dual division as occurring within a single, unified, South Indian
cultural system ; however, for him the division is a conflict mecha­
nism, a way o f dealing with ‘conflicts, anomalies, and antago­
nisms’275 which derive from two different preconditions in South
Indian caste: (1) the high degree of territorial segmentation of
caste groups and especially the confined m arriage networks of
South India resulting from the conceptions of South Indians
o f the equal role o f both parents in im parting qualities to their
offspring; (2) the absence of varna categories called, or cognate
with, K shatriya and Vaishya. Ingeniously A ppadurai offers the
notion that the right and left divisions o f the social body is a somatic
m etaphor like that o f the horizontally divided body in the prototypic
image of the Purusha-sukta of the R ig Veda, a m etaphor which
simultaneously can express a great variety of social conflicts while
yet affirming the unity o f South Indian society. The formulation
o f this conception A ppadurai dates in the Chola period, and his
analysis o f the various traditions preserved about the divisions,
lead him to conclude that the dual division form ulation was invoked
at times o f weak political integration when regal adjudication was
inadequate to the task of settling disputes among castes except when
these profoundly threatened civil order.276
275 Ibid., p. 221.
276 Ibid., pp. 233-45 where these are summarized. According to Appadurai,
‘the explicit form ulation of the [‘root paradigm ’] scheme occurred under Chola
auspices' (p. 223). From that time it appeared ‘in varied contexts and served multiple
functions’. These included assim ilation o f peoples new to caste-agrarian society
o f the Chola period, m ilitary activities, status strivings of artisans to a position
of parity with Brahmans, and sectarian conflict between Jainas and Brahm ans in
medieval K arnataka. A ppadurai’s emphasis upon the heterogeneity of issues
which invoked the dual division is certainly sound and provides a rem inder of the
hazard of attributing single causes or functions to so enduring and widespread a
phenom enon as the dual social division of South Indian castes. But, even in the
elicited heterogeneity of A ppadurai’s examples, most pertain to divisions between
the broad occupational groupings upon which the conventional interpretation he
is seeking to correct is based (e.g. the Brahm an-Jaina conflict resolved in a . d . 1368,
p. 226). He may wish to argue — but does not explicitly — that these persistent
occupational alignments are incidental to the heterogeneous causes of conflict
involving the two divisions in pre- and early Vijayanagara times, but the overall
constancy o f the recruitm ent and com position o f the divisions throughout the m acro
region and over m any centuries bids that the more conventional interpretation
not be too hastily abandoned. This caution is reinforced by the fact that there is
no explicit annunciation of the vertically divided body image during or after Chola
times. This m etaphor, it m ust be recognized, is one contrived by A ppadurai as a
 Both of these rich and original studies achieve theoretical clarity
by appearing to convert perceived central tendencies into principles.
F or Beck, conflict assumes no serious place in the alignment of
K ongu Tamil castes as formal, exhaustive, and complementary
structures o f the right and left castes; for Appadurai, it is conflict
alone which created and sustained the divisions. In the present
analysis, the complem entarity of attributes of the divisions catalo­
gued by Beck and the persistent occupational or materialistic basis
for the m ajor groups in both divisions is taken as very significant.
However, the m anner in which the divisions bridge highly localistic
kinship territories and the explicit conflict m otif of many of the
valangai and idangai records, as argued by Appadurai, m ust be
considered very im portant also. Both complementarity and conflict
appear to characterize the divisions. A ppadurai also attributes
m ajor im portance to the fact of urbanization, an issue which Beck,
concerned with the Kongu peasantry, says little about. Appadurai
notes that the situation in seventeenth century colonial M adras
— the case which he examines in detail — does not constitute a
sharp break from urban situations of the Vijayanagara period
in being especially volatile social stress points.277 This assessment
by A ppadurai is the m ore impressive as it re-emphasizes the way
in which the Vijayanagara period provided the basic forms — in
rural land relations as well as relationships in towns — which
constitute the foundations o f m odern society in South India
during the later British period.
A ppadurai’s provocative discussion o f the dual division does
draw attention to a m atter of great im portance in relation to the
present analysis o f right and left castes during the Vijayanagara
period. That is, he has, in his discussion o f the dual divisions and
in his discussion o f the sectarian dualism among Srivaishnavas
also discussed above, directed attention to the significant role of
kings in the society o f the time. This he concludes from a close
study of Vijayanagara inscriptions and from his reading of a
variety of the traditions (‘indigenous explanations’) about the dual
divisions. The latter are succinctly summarized by him,278 and
he notes that all of these traditions relate to conflicts over emblems
means for m oderns to understand the divisions; it is not a form ulation of any South
Indians of the past.
277 Ibid., pp. 245-57.
™ Ibid., pp. 233-42.
and practices claimed and disputed between groups which led
to or may have threatened public order and thus prom pted authori­
tative adjudication, usually by a king.
Considerable in scrip tional evidence supports this point of
royal intervention as well as A ppadurai’s attributed cause for
conflict over largely symbolic objects: intensified urbanization
with the concom itant stress entailed in that.279
In Tamil country, royal intervention into both temple and
valangai-idangai affairs became continuous around the middle of
the fifteenth century. According to Krishnaswami and other
Vijayanagara scholars, this coincides with the m ajor period of
sustained intrusion of Vijayanagara authority into Tamil country.
Even before, however, during the late fourteenth century, such
intervention was noted in inscriptions with respect to the restoration
o f temple worship as has already been discussed. Not so well
recognized is a similar appreciation of the dharmic intervention
o f Vijayanagara kings or their agents in m atters pertaining to the
relations o f the right and left castes. The reconstructed Vishnu
shrine at D harapuram in K ongu country contains a record of the
late fourteenth century which pays the following tribute to the
dharmic kings o f Vijayanagara:
Affording protection to all and the enjoyment o f freedom and prosperity
to people o f all sects — Srivaishnavas o f the eighteen lands, Sri Mahes-
varas, to the right and left communities — may they so govern the country
foreover. To them this dharmam is dedicated by all Srivaishnavas.280

The late period o f the reign o f Devarya II ( a . d . 1422-46) provides

good docum entation of such dharmic intervention. Then, accord­
ing to a record from Takkolam (in m odern A rkonam taluk of
N orth Arcot) the Vijayanagara king issued an edict which was
circulated to the m ajor Siva and Vishnu temple centres of Tondai-
m andalam, including Tiruvorriyur, Kanchipuram , K alahasti, and
Tiruvalangadu, providing for the remission o f certain dues from
temple lands and their use by the temples.281 D evaraya’s successor,
V irapratapa, who reigned for about a year ( a .d . 1446-7),282 re­
established the rates o f dues to be paid by right and left castes to a
a™ Ibid., pp. 227-8.
280 Viraraghavacharya, ‘The Srivilliputtur Temple’, p 6, no.8, p. 1.
2SI A .R .E ., 1922, para. 45, p. 110 regarding 270/1921.
282 N ilakanta Saslri, H isto ry o f S outh India , p. 300.
level which had existed at an earlier time before these dues had
been increased and coercively collected, thus compelling the flight
o f many in the dual divisions.283 This occurred at villages around
the town of Tirukkoyilur and elsewhere in m odern South A rcot.284
A nother record of D evaraya’s reign comes from another temple
centre, Vriddhachalam , in South Arcot, which recounts a royal
order including the following passage:
members o f the Valahgai and Idangai sects met . . . and came to the
decision that, since the officers o f the king (rajanyas) and owners o f
JTvitas [tax-free prebends]285 oppressed. . . and the kaniyalan [holders
o f hereditary land rights] and the Brahmanas took the rajakaram (i.e.,
taxes), none o f Valangai and Idangai people should give them shelter and
that (none o f the people o f the two sects) bom in the country should
write accounts for them or agree to their proposals. If any proved a
traitor to the country (by acting against this settlement), he should be
stabbed. . . ,286

In discussing this Vriddhachalam inscription, the epigraphist

noted that similar records of Devaraya II are to be found in Tan-
javur and Tiruchirapalli.287
Such cooperation between the divisions may have been less
com m on than competition. M any Vijayanagara inscriptions refer
to the equalization o f the rights o f each division with the other,
from the Arunachalesvara temple at Tiruvannamalai, dated Saka
1340, contained an order of Devaraya I, possibly issued by his son,
to the effect that the idangai and valangai groups at Tiruvannam alai
should enjoy the same privileges.288 The same notion is conveyed
in a record of Achyutadevaraya which speaks of the ninety-
eight casts comprising both the right and left divisions.289
This adjudicated parity of privileges may reflect pressure
o f idangai demands arising out of their steadily augmented
urban power. Such a sense is conveyed in several records of
the later Vijayanagar period pertaining to special demands of
idangai groups. One from M adurai reports the action o f Virappa

283 A .R .E ., 1922, para. 46, pp. 110-11 regarding 476/1921.

284 A .R .E ., 1907, para. 55, p. 69, and A .R .E ., 1904-5, para. 30 regarding a record
at K ilur, T irukkoyilur taluk, South Arcot.
285 S .I .T .I ., ‘G lossary’, p. xx.
286 A .R .E ., 1918, para. 68, p. 163.
287 Ibid., regarding 92/1918, dated sa ka 1357.
288 A .R .E ., 1903 regarding 564/1902 and f.M .P ., v. 1, p. 94.
289 A .R .E ., 1906 regarding 103/1906, from Jam bai, South Arcot.
N ayaka in confirming the fivefold division of Kam m alars o f the left
division at the request of the latter.290 In another inscription from
Tiruvadi in C holam andalam , Chinnappa N ayaka exempted the
five classes of Kam m alars from certain taxes which had previously
been exacted under threat of force and had caused many o f the
idangai artisans to flee the town.291 Still another inscription from
T iruvam attur in modern Villupuram, South Arcot, refers to a
resolution by residents o f the place before K rishnappa Nayaka,
agent o f the Vijayanagara king Sriranga, establishing for the
K am m alars o f the vicinity the same rights and privileges as those
enjoyed by artisans o f the idangai in the towns of Padaividu, Senji,
Tiruvannam aiai, and K anchi.292
This evidence of idangai demands for political intervention to
achieve parity of privileges with their fellows in other places and
with right caste groups m ust occasionally have resulted from or
led to conflict as m ay be judged from an epigraph of the late four­
teenth century referring to conflict between right and left castes
in Tondaim andalam which endured for four years.293 Tondai-
m andalam also exhibited another form o f stress which appears
to have occasioned royal intervention during the latter Vijaya­
nagara period. This was competition between Tamil-speaking
Vellalas and Telugu-speaking Balijas for locality headships. The
perceptive Francis W. Ellis, in his 1818 discussion of m irasi rights,
reported that when Tondaim andalam came under the control
of the Bijapur sultanate, probably around a . d . 1650, following the
ejection by the Muslims of Sriranga III, the office o f rial tan cus­
tom arily held by Tamil Vellalas was reduced in im portance by the
creation o f a parallel local office called d e sa y i filled by Telugus.
N a tio n s of Tondaim andalam had previously regulated all disputes
pertaining to caste and were especially im portant in valcmgai
affairs in the seventy-nine nadus o f Tondaim andalam . D esdyis,
appointed from am ong Telugu Balijas, claimed joint jurisdiction
with the Tamil nattdns in an office which came to be called nddu-
d~sam. Idangai groups, Ellis reported, disputed the authority
of both.294
290 a .R .E ., 1918, regarding 55/1917.
291 A .R .E ., 1922, regarding 413/1921.
292 Ibid., 65/1922.
2 A .R .E ., 1906, regarding 422/1905 dated sakti 1305 from Gudim allur, N orth
Arcot.
294 Replies to Seventeen Q uestions P roposed by the G overnm ent o f Fort St. George
 While the frequency o f royal intervention into iocai, Tamil
affairs, including conflicts o f the right and left castes, certainly
increased during the fifteenth century, the few inscriptions pertain­
ing to these m atters continue to reflect the potential nature of the
divisions and the ability of these groupings to cooperate against
exploitation at the hands o f others: Brahmans, royal officials,
soldiers, and others.295 This the assembled right and left caste
groupings did just as they had done during the Chola period when
there were few signs o f royal adjudication.

R elative to M ira si R ight, G overnm ent Gazette Office, M adras, 1818, App. 1. ‘List
o f the Nadus, K ottam s and Villages of T ondam andalam ', p. ix, note.
21)5 See the undated, probably post-C hola record from Pennadam on the Vellar
River boundary o f C holam andalam , A .R .E ., 1928-9, para. 79, p. 90; also a late
Chola record, possibly Kulottunga 11! from A duturai, Tiruchirapalli, A .R .E .,
1913, regarding 34 of 1913.
 Bibliography

EPIG R A PH Y
General W orks
Indian Palaeography. Oxford: Clarendon Press, 1963. Pp. 297.
D a n i , A .H .
List o f Inscriptions Copied by the Office o f the
K r i s h n a m a c h a r l u , C.R.
Superintendent fo r Epigraphy. D elhi: Manager o f Publications, 1941.
Pp. 136.
S irc a r, D .C . Indian Epigraphical Glossary. D elhi: M otilal Banarsidass,
1966. Pp. 560.
------Indian Epigraphy. Delhi: Motilal Banarsidass, 1965. Pp. 475.

Serials
Epigraphia Indica. Delhi/Calcutta: Archaeological Survey o f India, 1892-.
Indian Antiquary. Bombay: 1872-1923.
Studies in Indian Epigraphy [Bharatiya Purabhilekha Patrika]; Journal o f
the Epigraphical Society o f India. Mysore: 1974-.

South Indian Epigraphy

G eneral
The Language o f Tamil Inscriptions,
A g e s t h i a l i n g o m and S .V . S h a n m u g a m ,
1250-1350 a . d . (Department o f Linguistics, Publication no. 23). Anna-
malainagar: Annamalai University, 1970. Pp. 288.
Annual Report on (South Indian) Epigraphy. Madras: Government Press,
1887-.
Elements o f South Indian Palaeography, from the Fourth to
B u r n e l l , A.C.
the Seventeenth Century a . d . , 2nd ed. London: Trubner and Co., 1878.
Pp. 147.
Journal
K a r a s h i m a , N o b o r u . ‘A Bibliography o f South Indian Epigraphy’.
o f Asian and African Studies (Institute for the Study o f Languages and
Cultures o f Asia and Africa, Tokyo), no. 6 (1973), 151-63.
Subject Index to the Annual Reports on South
K r i s h n a m a c h a r l u , C .R.
Indian Epigraphy from 1887 to 1936. Delhi: Manager o f Publications,
1940. Pp. 93.
M a h a l i n g a m , T.V. Early South Indian Palaeography. M adras: University
o f Madras, 1967. Pp. 341.
N a g a s w a m y , R. Seminar on Inscriptions, 1966. Madras: Books (India)
Private Ltd., 1968. Pp. 230.
S i v a r a m a m u r t i , C. Indian Epigraphy and South Indian Scripts (Bulletin
 o f the Madras Government Museum, New Series, ‘General Section’,
vol. III). Madras: Government Press, 1952.

Serials
Damilica ( ‘Journal o f the Tamilnadu State Department o f Archaeology’).
Madras: 1970-.
Epigraphia Tamilica; A Journal o f Tamil Epigraphy ( ‘Jaffna [Sri Lanka]
Archaeological Society’). Ed. K. I n d r a p a l a . 1971-,
Transactions o f the Archaeological Society o f South India ( ‘Publication o f
the Archaeological Society o f South India’). Madras: 1955-.

G lossaries
M a h a l i n g a m , T.V. South Indian Polity. Madras: University o f Madras,
1967. ‘Glossary’, by R. G o p a l a n and Y. S u b b a r a y a l u . Pp. 26.
S u b r a m a n i a m , T.N. (ed.). South Indian Temple Inscriptions ( ‘Madras
Government Oriental Series’, no. 157). Madras: Government Oriental
Manuscripts Library (vol. Ill, pt II), 1957. ‘Epigraphical Glossary’,
pp. 96.

T ex ts an d T ranslations
South India
Catalogue o f Copper-Plate Grants in the Government Museum, Madras.
Ed. R . S r i n i v a s a R a g h a v a A y y a n g a r . Madras: Government Press,
1918. Pp. 89.
Selected South Indian Inscriptions, Tamil, Telugu Malayalam and Kannada.
Ed. V .R . R a m a c h a n d r a D i k s h i t a r . Madras: University o f Madras,
1952. Pp. 646.
The Historical Inscriptions collected till 1932 and Outlines
S e w e ll, R o b e rt.
o f Political History. Ed. S. K r i s h n a s w a m i A y y a n g a r . Madras:
University o f Madras, 1932. Pp. 451.
South Indian Inscriptions ( ‘Publication o f the Archaeological Survey o f
India’). Madras: Government Press, 1890-.
‘Topographical List o f Inscriptions Compiled under the Direction o f
T.V. Mahalingam.’ (Typescript volumes, University o f Madras, Depart­
ment o f Ancient History and Archaeology, Madras.)
A Topographical List o f the Inscriptions o f the Madras Presidency, collected
till 1915, with Notes and References. Ed. V. R a n g a c h a r y a . 3 vols.
Madras: Government Press, 1919. Pp. 2070.

T am il N adu
Copper-Plate Inscriptions belonging to the Sri Sankaracharya o f the Kama-
kotipitha. Ed. T.A. G o p i n a t h a R a o . Madras: Law Printing House,
1916. P. 137.
Chronological List o f inscriptions o f the Pudukkottai Slate arranged according
to Dynasties. Pudukkottai: State Press, 1929. Pp. 154.
Inscriptions in the Pudukkottai State, translated into English. Trans, by
K.R. S r i n i v a s a A i y a r . Pudukkottai: State Press. Pt 1 (1941), pp. 124;
pt 2(1946), pp. 121.
Inscriptions ( Texts) o f the Pudukkottai State arranged according to Dynasties.
Pudukkottai: State Press, 1929. Pp. 653.
South Indian Temple Inscriptions ( ‘Madras Government Oriental Series’,
no. 157). Ed. T .N . S u b r a m a n i a m , 3 vols. Madras: Government
Oriental Manuscripts Library, 1953-7. Pp. 1610.
Tamil-Brahmi Inscriptions. Ed. I r a v a t h a m M a h a d e v a n . Madras: Tamil-
nadu State Department o f Archaeology, 1970. Pp. 15.
Travencore Archaeological Series (‘Publication o f the Travencore Archaeolo­
gical Department’). 9 vols. Travancore: Government Press, 1910-41.

Andhra
Andhra-Pradesh Government Archaeological Series. Ed. P. S r e e n i v a s a c h a r ,
P.B. D e s a i, N . R a m e s a n , et al. Hyderabad: Government o f Andhra
Pradesh, 1961-.
Andhra-Pradesh Government Epigraphy Series. Ed. N . V e n k a t a r a m a n a y y a .
Hyderabad : Government o f Andhra Pradesh, 1968-.
A Collection o f the Inscriptions on Copper-Plates in the Nellore District.
Ed. A l a n B u t t e r w o r t h and V . V e n u g o p a u l . 3 pts. Madras: Govern­
ment Press, 1905. Pp. 1520.
Inscriptions o f Andhradesa (Sri Venkateswara Historical Series). Ed.
M . R a m a R a o . 2 vols. Tirupati: Sri Venkateswara University, 1967-8.
S ri T .T . [ Tirumalai-Tirupati] Devasthanam Epigraphical Series. Ed.
S. S u b r a h m a n y a S a s t r y and V . V i r a r a g h a v a c h a r y a . 6 vols. Madras:
Tirupati Sri Mahant’s Press, 1931-8.
Telengana Inscriptions. 2 vols. Hyderabad: Lakshmanaraya Parisodhaka
Mandali, 1935 and 1960. Pp. 429.
H yderabad Archaeological Series (Hyderabad State Archaeological Depart­
ment). 19 nos. Calcutta: Nizam ’s Government, 1915-56.

K a rn a taka
Annual Report o f the M ysore Archaeological Department (University o f
Mysore). Mysore: Government Branch Press, 1905-.
Epigraphia Carnatica (Mysore Archaeological Series). E d . B. L e w is R ic e ,
et al: 16 vols. Bangalore/M angalore/M ysore: Government Press,
1889-1955.
Karnatak Inscriptions. Ed. R.S. P a n c h a m u k h i , B.R. G o p a l , et al. Dharwar:
Kannada Research Institute, Karnatak University, 1941-.
Summaries o f [Kannada] Inscriptions (1943-44 to 1949-50). Ed. P.B. D e s a i.
Dharwar: Kannada Research Institute, Karnatak University, 1966.
 O fficial P ublications (N on -E pigraph ical)
M adras D istrict Manuals (variously titled, most published in Madras by the
Madras Government Press or the Lawrence Asylum Press).

North Arcot Cox, A r t h u r F. 1881.

St u a r t , H . A . 2 v o ls. 1894-5.
South Arcot G a r s t in , J. H . 1878.
B a l ig a , B .S . 1962.
Chingleput C r o l e , C . S. 1879.
Coimbatore N ic h o l s o n , F. A. 3 v o ls. 1855, 1887, 1898.
Godavari M o r r is , H enry 1874.
Madura N elso n , J. H . 1868.
Malabar Loga n , W . 3 vo ls. 1887, 1891.
Nellore B osw ell, J o h n A . C. 1873.
Salem D y k es, J a m e s W .B . 1853 ( L o n d o n ) .
L e F a n u , H . 2 v o ls. 1883.
R a m a s w a m i, A . 1967.
Tanjore V enkatasw am i R o w , T. 1883.
Trichinopoly M o o r e , L e w is 1878.
Tinnevelly St u a r t , A . J . 1879.

M ysore; A Gazetteer Compiled fo r Government. Ed. B. L e w is R ic e . 2 vols.

Westminister: Archibald Constable and C o., 1897. Pp. 1425.
M ysore Gazetteer. Ed. C. H a y a v a d a n a R a o . 5 vols. Bangalore: Govern­
ment Press, 1930.
The Travencore State Manual. Ed. V. N a g a m A i y a . 3 vols. Trivandrum:
Government Press, 1906.
The Travencore State Manual. E d . T.K. V e l u P i l l a i . 2 vols. Travancore:
The Government o f Travancore, 1940.
S c h o l b e r g , H e n r y . The D istrict Gazetteers o f British India. Zug, Switzer­
land: Inter Documentation Co., 1970. Pp. 131.

E thnological S eries
Castes and Tribes o f Southern India. E d . E d g a r T h u r s t o n a n d K . R a n g a -
c h a r i . 6 vols. Madras: Government Press, 1909.
The M ysore Tribes and Castes. Ed. L.K. A n a n t h a k r i s h n a I y e r . 4 vols.
Mysore: The M ysore University, 1935.

M iscellaneous R eports
An Alphabetical L ist o f the Feasts and the Holidays o f the Hindus and
 Bibliography 493

Muhammadans (Publication o f the Imperial Record Department).

Calcutta: Government Press, 1914.
Alphabetical List o f Villages in the Taluks and Districts o f the Madras
Presidency; Corrected up to September 1930. Madras: Government
Press, 1933.
Chiefs and Leading Families in the Madras Presidency. 2nd ed. Madras:
Government Press, 1923.
A Collection o f Papers on Mirasi Right, 1862. Ed. W. H u d d l e s t o n , notes
by VV.H. B a y l e y . Madras: H ig h Court Library, vol. no. 1599. Pp. 294.
Collectors’ Reports on the Deposed Poligars o f Udaiyarpalayam, Ariyalur,
and Toraiyur, 1816-62: Selections from the Records o f the Collectorate
o f Trichinopoly. Trichinopoly: Collectorate Press, 1867. Madras Re­
cord Office, Library, ASO (D ), vol. 358.
Replies to Seventeen Questions Proposed by the Government o f
E l l i s , F .W .
Fort St. George Relative to Mirasi Right, with Two Appendices. Madras:
Government, Gazette Office, 1818. Pp. 76.
The Fifth Report from the Select Committee o f the House o f Commons on the
Affairs o f the East India Company, 28th July, 1812. Ed. W a l t e r K e l l y
F i r m i n g e r . 3 vols. Reprint ed. N ew York: Augustus M. Kelley, 1969.
India. Archaeological Survey o f India. Annual Report, 1903-4. V. V e n k a y y a .
‘Irrigation in Southern India in Ancient Times.’ Calcutta: Government
Press, 1906. Pp. 202-11.
India. Government o f Bengal. Report o f the Land Revenue Commission,
Bengal. 5 vols. Alipore: Bengal Government Press, 1940-1.
India. Linguistic Survey. Linguistic Survey o f India. Ed. G .A . G r i e r s o n .
11 vols. Calcutta: Government Press, 1903-22.
India. Madras Presidency. The Baramahal Records (Records o f Fort
St. George). Madras: Government Press, 1907-33.
Mackenzie Manuscripts; Summaries o f the Historical Manuscripts o f the
M ackenzie Collection. Vol. 1 (Tamil and Malayalam). Ed. T.V. M a h a -
l i n g a m . Madras: University o f Madras, 1972.
M acqueen, P ercy [Curator, Madras Record Office], ‘Historical N otes,
South India.’ Typescript mss, Madras Record Office, Library no.
AHB 886, dated 5 Novem ber 1932. P . 113.
Mysore Stale in Maps (Government o f Mysore, Department o f Statistics).
Bangalore: Government Press, 1958.
On the Names o f Places in Tanjore. Ed. B .R . B r a n f i l l . Madras: Govern­
ment Press, n.d.; collected in 1877-8.

SECO N D A R Y W ORKS
A b d u l A z i z , M. ‘The Deccan in the 15th Century’. Proceedings o f the
Asiatic Society o f Bengal fo r 1924, New Series, vol. 21, 549-91.
‘Account o f the Province o f Râmnâd, Southern Peninsula o f India, compiled
from the “ Mackenzie Collection’” . Ed. H e n r y H a r k n e s s . Journal o f
the Royal Asiatic Society o f Great Britain and Ireland, vol. 3 (1836),
165-89.
A c h u t a R a u , D .S. ‘The Early Wodeyars o f Mysore and Their Cultural
Traditions.’ Quarterly Journal o f the Mythic Society ( ‘Culture and
Heritage Number’, 1956). Pp. 179-94.
A i y a k a n n u N a y a k a r , T. A Treatise on the Vanniyar Caste. Madras:
Albinion Press, 1891.
A l t e k a r , A.S. A History o f Village Communities in Western India (Publi­
cation o f the University o f Bombay Econom ic Series, no. 5). Madras :
Oxford University Press, 1927.
A p p a d o r a i, A. Economic Conditions in Southern India, ¡000-1500 a .d .
2 vols. Madras: University o f Madras, 1936.
A p p a d u r a i , A r j u n . ‘Right and Left Hand Castes in South India.’ The
Indian Economic and Social History Review, vol. 11, nos. 2 and 3 (1974),
216-60.
A r a s a r a t n a m , A . ‘Some Aspects Relating to the Establishment and Growth
o f European Settlements A long the Tamil C oast in the Seventeenth
Century’. Tamil Culture, vol. 7 (October 1958), 355-69.
A r o k i a s w a m i , M. The Early History o f the Vellar Basin, with Special
Reference to the Irukkuvels o f Kodumbalur. Madras : Amudha Nilayam
Ltd., 1954.
------ ‘The Inscriptional Term “ M uvendavelan” ’. Journal o f Indian History,
vol 34 (April 1956). Pp. 191-3.
------The Kongu Country. Madras: University o f Madras, 1956.
------ ‘Social Developments under the Imperial C olas.’ Transactions o f the
Archaeological Society o f South India (1956-7), 1-8.
B a d e n - P o w e l l , B a d e n H e n r y . The Indian Village Community. London:
Longmans, Green, 1896. P p . 456.
------The Land Systems o f British India. 3 vols. Oxford: Clarendon Press,
1892.
B a i l e y , F.G. ‘Closed Social Stratification in India.’ Archives Européennes
de Sociologie, vol. 4, no. 1 (1963), 107-24.
------ ‘For a Sociology o f India?’ Contributions to Indian Sociology, vol. 3
(1959), 80-101.
— —Political and Social Change in Orissa ; Orissa in 1959. Berkeley: Univer­
sity o f California Press, 1963.
------Tribe, Caste and Nation: A Study o f Political Activity and Political
Change in Highland Orissa. Oxford : Oxford University Press, 1960.
B a l a s u b r a h m a n y a m , S.R. Early Chola Art, Part One. Bombay: Asia
Publishing House, 1966.
Sir Asutosh Memo­
B a l k r i s h n a , P r i n c i p a l D o c t o r . ‘Interest and U sury.’
rial Volume. Patna: J.N. Samaddar, 1926-8. Pp. 277-301.
B a n a j i , D .R . Slavery in India. Bombay: University o f Bombay, 1933.
B a n e k j e a , J.N. Development o f Hindu Iconography. Calcutta: University
o f Calcutta, 1956.
B a r n a b a s , A.P. ‘Sanskritization.’ Economic W eekly (Bombay), vol. 13,
no. 15 (1961), 613-18.
B a r r e t t , D o u g l a s . Early Cola Architecture and Sculpture: 866-1014 a .d .
London: Faber and Faber Ltd., 1974.
B a s a v a R a j a , K .R . ‘Agraharas in Medieval Karnataka.’ Journal o f the
Karnatak University (H um anities), vol. 4 (June 1960). Pp. 106-14.
B a y l e y , W .H. Memorandum on 'M irasi Tenure’. London: August 1862.
Madras Record Office Library: S.L. XIII-A-I09. P. 17.
B e a l s , A l a n R. and B e r n a r d J. S i e g e l . Divisiveness and Social Conflict:
An Anthropological Approach. Bombay : Oxford University Press, 1967.
B e c k , B r e n d a E.F. Peasant Society in Konku. Vancouver: University o f
British Columbia, 1972.
B e id e lm a n , T.O. A Comparative Analysis o f the Jajmani System (Association
o f Asian Studies Monograph, N o . 8). Locust Valley, N ew Y o rk : 1959,
B e t e i l l e , A n d r e . Studies in Agrarian Social Structure. D elhi: Oxford
University Press, 1974.
B h a n d a r k a r , R.G. Vaisnavism, Saivism and Minor Religious Systems.
Strassbourg: Karl S . Traubner, 1913.
B h a t t a c h a r y a , S i v a p r a s a d . ‘Religious Practices o f the Hindus.’ The
Religion o f the Hindus. Ed. K e n n e t h N . M o r g a n . N ew York: Ronald
Press, 1953.
B h o o t a l i n g a m , M a t h u r a m . Movement in Stone (A Study o f Some Chola
Temples). N ew D elhi: Soumani Publications, 1969.
B l o c h , M a r k . French Rural H isto ry; An Essay on Its Basic Characteristics.
Berkeley: University o f California Press, 1966.
B o w e r , R e v . H . The Chintamani; First Book Called Namagal Ilambagam;
with a Commentary o f Nachinarkiniyar and with N otes and Analysis
in English. M adras: Christian Knowledge Society, 1868.
B o x e r , C.R. Salvador de Sä and the Struggle fo r Brazil and Angola, 1602-
1686. London: University o f London, 1952.
B r i g g s , L a w r e n c e P. The Ancient Khmer Empire (Transactions o f the
American Philosophical Society, vol. 41, pt 1). Philadelphia: American
Philosophical Society, 1951.
B r o w n , C h a r l e s P h i l i p (ed.). Three Treatises on M irasi Right by the
Late Francis W. Ellis, L t. Col. Blackburne, and Sir Thomas Munro,
with Rem arks by the Honourable, the Court o f Directors, 1822 and
1824. Madras: Christian Knowledge Society, 1852.
------ War o f the Rajas: H istory o f Anantapuram; Written in Telugu; in or
about the Years 1750-1810; Translated into English. Madras: Christian
Knowledge Society, 1853. P. 160.
B u c h a n a n , F r a n c i s . A Journey from M adras through the Countries o f
 Mysore, Canara, and Malabar. 3 vols. London: T. Cadell, W. Davis
and Black, Parry, and Kingsbury, 1807.
Remarks on the Ryotwari System as It Exists in the
B o u r d i l l o n , J .D .
Presidency o f Madras. Madras: 1853. India Office Tracts, N o. .567.
B ü h l e r , G . The Sacred Books o f the Aryas as Taught in the Schools o f
Apastamba, Gautama, Vasishta, and Baudhayana (Sacred Books o f the
East, vol. 14, pt 2). Ed. M a x M ü l l e r . Reprint edition. D elh i: Motilal
Banarsidass, 1965.
Hill Farms and Paddy Fields in Mainland South-East
B u r l i n g , R o b b in s .
Asia. Englewood Cliffs, N .J .: Prentice-Hall, 1965.
B u r r o w , T. and M .B . E m e n e a u . A Dravidian Etymological Dictionary.
Oxford: Clarendon Press, 1961.
C a ld w e ll, R. A Political and General History o f the District o f Tinnevelly
in the Presidency o f Madras. Madras: Government Press, 1881.
C a r p e n t e r , J. E s t l i n . Theism in Medieval India (Hibbert Lectures, Second
Series, 1919). London: W illiams and Norgate, 1921.
The Economic Condition o f India during the Sixteenth
C h a b l a n i , H.L.
Century. D elhi: Oxford B ook and Stationery C o., 1929.
C h a m p a k a l a k s h m i , R . ‘Jainism in South India.’ M. Lit. Thesis, University
o f Madras, 1958.
C h a n c h i a h , P. and R .M . B h u j a n g a . A History o f Telugu Literature (The
Heritage o f India Series). C alcutta: The A ssociation Press, n.d.
C h a n d r a s e k h a r a S a s t r i , S. ‘Econom ic Conditions under Hoysalas.’ The
Half-Yearly Mysore University Journal, vol. 2 (1928), 196-241.
C h a t t e r j i , B.R. Indian Cultural Influence in Cambodia. 2nd ed. Calcutta:
University o f Calcutta, 1964.
C h a y a n o v , A l e x a n d e r . The Theory o f Peasant Economy. Ed. D a n i e l
T h o r n e r ; translated from the Russian by R .E.F. S m ith . Hom ewood:
American Economic Association, 1966.
A B rief History o f the Kammas (Abridged and Translated
C h o u d a r y , K .B .
from Telugu). Guntur: The Author, 1954.
C h o u d h a r y , R. ‘Some Aspects o f Feudalism in Ancient India (D ow n to
the 14th Century a .d ., Based Mainly on Epigraphic Sources).’ Journal
o f Indian History, vol. 37, pt 3 (1959), 385-406; vol. 38, pt 1 (1960),
193-203.
C o d r i n g t o n , K. d e B. ‘Castes or N ation? A Mala Salvadi’s Badge o f
Office.’ Man, vol. 31 (October 1931), 205-7.
C o ed es, G e o rg e .The Indianized States o f Southeast Asia. Ed. W. V e l l a ;
trans. S.B. C o w i n g . H onolulu: East-West Center Press, 1968.
C o h n , B.S. and M c k im M a r r i o t t . ‘Networks, and Centers in the Integra­
tion o f Indian Civilization.’ Journal o f Social Research (Ranchi), vol. 1
(1958), 1-9.
cölamandala-satakam [¿y] Atmanathar Desikar \of\ Velur (1650-1728 a .d .) .
Ed. S o m a s u n d a r a r d e s i k a r . Mayunar: 1916.
C o rre ia -A fo n s o , J o h n . Jesuit Letters and Indian History. B om bay: Indian
Historical Research Institute, 1955.
The Functions o f Social Conflict. G len coe: Free Press, 1956.
C o s e r , L e w is .
Regions and Regionalism in South Asian Studies:
C r a n e , R o b e r t I. (ed.).
An Exploratory Study (Monograph N o. 5, Program in comparative
Studies on Southern Asia, Duke University). D urham : Duke University,
1967.
D a n v e r s , F r e d e r i c k C h a r l e s . 2 vols. The Portuguese in India. London:
W .H . Allen, 1894.
D a y , W i n i f r e d M. ‘Relative Permanence o f Former Boundaries in India.’
The Scottish Geographical Magazine, vol. 65, no. 3 (1949), 113-22.
D e r r e t t , J. D u n c a n M. The Hoysalas; A Medieval Indian Royal Family.
Madras: Oxford University Press, 1957.
------ ‘Two Inscriptions Concerning the Status o f Kammalas and the Applica­
tion o f Dharmasastra.’ Prof. K .A .’ Nilakanta Sastri 80th Birthday
Felicitation Volume. Madras: Felicitation Committee, 1971. Pp. 32-56.
D e s a i, D i n k a r . The Mahamandalesvaras under the Cdlukyas o f Kalyani.
Bombay: Indian Historical Research Institute, 1951.
D e s a i, P.B. Jainism in South India; and Some Jaina Epigraphs. Sholapur:
Jaina Samskrti Samrakshaka Sangha, 1957.
D i e h l , C a r l G. Instrument and Purpose; Studies o f Rites and Rituals in
South India. Lund: C .W .K . Gleerup, 1956.
D i k s h i t , G.S. ‘Economic Conditions in the Time o f Krishnadevaraya.’
Vijayanagara Sexcentenary Commemoration Volume. Dharwar: Kar­
natak Historical Research Society, 1936. Pp. 213-30.
Local Self-Government in Medieval Karnataka (Karnatak University
Historical Series, N o . 5). Dharwar: Karnatak University, 1964.
------ ‘Okkalu— A Corporate Body o f Medieval Karnataka.’ Journal o f the
Karanatak University (Humanities) , vol. 4 (June 1960), 58-63.
D i r k s , N i c h o l a s . ‘Political Authority and Structural Change in Early
South Indian History.’ The Indian Economic and Social History Review,
vol. 13 (1976).
D o r a i R a n g a s w a m y , M .A. The Religion and Philosophy o f the Tevdram.
2 vols. Madras: University o f Madras, 1958.
D u b o is , A b b e J.A. Hindu Manners, Customs and Ceremonies. Trans.
H .K . B e a u c h a m p . London: Oxford University Press, 1928.
D u m o n t , L o u is . ‘The Conception o f Kinship in Ancient India.’ Religion/Po­
litics and History in India; Collected Papers in Indian Sociology. The
Hague: M outon, 1970.
------Hierarchy and Marriage Alliance in South Indian Kinship (Occasional
Papers o f the Royal Anthropological Institute o f Great Britain and
Ireland, no. 12). London: Royal Anthropological Institute, 1957.
------Homo Hierarchicus; The Caste System and Its Implications. Trans.
M a r k S a i n s b u r y . Chicago: University o f Chicago Press, 1970.
------ ‘A N ote on Locality in Relation to Descent.’ Contributions to Indian
Sociology, vol. 7 (March 1964), Pp. 71-6.
------Une Sous-Caste de L ’Inde du S ud; Organisation Sociale et Religion
des Pramalai Kallar. Paris: M outon, 1957.
------ ‘The Village Community from Munro to M aine.’ Contributions to
Indian Sociology , vol. 9 (December 1966), 67-89.
D u rr, K.I. ‘The Historical Importance and Economic Aspect o f the South
Indian Temple.’ Journal o f the Andhra Historical Society , vol. 8, pts 1-4
(1933-4), 121-2.
------ ‘Studies in Vijayanagar Polity.’ Ibid., vol. 4, pts 1-4 (1930-1), 155-64.
E lm o r e , W .T . Dravidian Gods in Modern Hinduism (Reprinted from Univer­
sity Studies o f the University o f Nebraska, vol. 1 5 ,no. 1,1915). Madras:
The Christian Literature Society o f India, 1925.
E v a n s - P r î t c h a r d , E .E . The Nuer. N ew York: Oxford University Press,
reprint 1971.
F a b r i c i u s , J.P. A Dictionary ; Tamil and English. Tranquebar: Evangelical
Lutheran M ission Publishing H ouse, 1910.
F a l l e r s , L l o y d A. ‘Are African Cultivators to be called “Peasants” .’
Current Anthropology, vol. 2, no. 2 (1961), 108-10.
F a r q x j h a r , J.N. Modern Religious Movements in India. Delhi: Munshiram
Monoharlal, 1967.
------ An Outline o f the Religious Literature o f India. London : Oxford Univer­
sity Press, 1920.
F i l l i o z a t , J e a n . ‘Les D ivisions Sociales de l’Inde.’ Anthropologie des
Tamouls du Sud de l'Inde. (Publications Hors-Série de L’École Français
d’Extrême-Orient.) Pondicherry: Institut Français d’Iiidologie, n.d.
F illio z a t, V a su n d h a ra . L ’Epigraphie de Vijayanagar du début â 1377.
(Publications de L’Ecole Français d ’Extrême-Orient, vol. 91.) Paris:
1973.
F i r t h , R a y m o n d . ‘Factions in Indian and Ovèrseas Societies.’ British
Journal o f Sociology, vo!. 8, no. 4 (December 1957), 291ff.
------ a n d B a s i l Y a m e y . Capital Saving, and Credit in Peasant Societies.
Chicago: Aldine Press, 1965.
F ra z e r, R.W. ‘Dravidians.’ Encyclopedia o f Religion and Ethics. Ed.
Ja m e s H a s t i n g e s . N ew York: Charles Scribner’s Sons, 1925. Vol. 5,
21 - 8 .
F u c h s , S t e p h e n . The Gond and Bhumia o f Eastern Mandia. Bombay:
Asia Publishing H ouse, 1960.
G a j a r a j a M u d a l i a r , B.L. ‘Thuluva Vellalar Community o f Bangalore
Civil and Military Station.’ Quarterly Journal o f the M ythic Society,
vol. 2 (1910-11), 62-3.
G a n a p a t h y P i l l a i , W.E. Etaiyapuram Past and Present; being a Short
History o f the Estate from the Earliest Times. Madras: S.P.C.K. Press,
1890.
G h a n a m u r t h y , T.E. A Critical Study o f the Civakacintàmani. Coimbatore:
Kalai Kathir, 1966.
G o n d a , J. Ancient Indian Kinship from the Religion Point o f View. L e id e n :
E.J. B rill, 1966.
G o p a l , B.R. ‘The Capital City o f Pânumgal.’ Quarterly Journal o f the
M ythic Society, vol. 48 (1957-8), 55-60.
-------- ‘The Later Western Chàlukyas (From the Earliest Times to 1076 a . d . ) . ’
Unpublished Ph.D . thesis, University o f Mysore, 1961.
G o p a l , M .H. ‘Financial Administration o f Mysore under Tipu Sultan.’
Professor K.V. RangaswamiAiyangar Commemoration Volume. Madras:
G .S . Press, 1940.
G o p a l a n , R. History o f the Pallavas o f Kanchi (The Madras University
Historical Series, no. 3). Madras: University o f Madras, 1928.
G o p i n a t h a R a o , T.A. Elements o f Hindu Iconography. 2 vols. Madras:
The Law Printing H ouse, 1914.
G o v i n d a s a m y , M.S. The Role o f Feudatories in Pallava History. Anna-
malainagar: Annamalai University, 1965.
G r o s , F r a n ç o i s and R . N a g a s w a m y . Uttaramerur : Legendes, Histoire,
Monuments. Pondicherry: institut Français d’Indologie, 1970.
G u p t a , K .M . The Land System in South India between c. 800 a . d . and
1200 a . d . (in the Light o f Epigraphic and Literary Evidence) (Punjab
Oriental Series, no. 20). Lahore: M otilal Banarsidass, 1933.
H a b ib , I r f a n . The Agrarian System o f Mughal India (1556-1707) (Published
for the Department o f History, Aligarh University). London: Asia
Publishing H ouse, 1963.
H a l l , D .G .E . A- History o f South-East Asia. N ew York: St. Martin’s
Press, 1955.
H a l l , K e n n e t h R . ‘The Nagaram as a Marketing Center in Early Medieval
South India.’ Unpublished Ph.D. thesis, Department o f History,
University o f Michigan, 1975.
H a r i R a o , V.N. ‘A History o f Trichinopoly and Srirangam.’ Unpublished
Ph.D. thesis, University o f Madras, 1948. Appendix (A ): ‘Koil OJlugu
in English’. Pp. 177.
------ Kôil Olugu, The Chronicle o f the Srirangam Temple with Historical
Notes. Madras: R ochouse and Sons, 1961.
H a r l e , J a m e s C. Temple Gateways in South India; the Architecture and
Iconography o f the Cidambaram Gopuras. Oxford: Bruno Cassirer,
1963.
H a r t , G e o r g e L. h i. ‘Ancient Tamil Literature: Its Scholarly Past and
Future.’ Essays on South India (Asian Studies at Hawaii, no. 15).
Ed. B u r t o n S t e i n . H onolulu: The University Press o f Hawaii, 1975.
—— The Poems o f Ancient Tamils: Their Milieu and Their Sanskrit Counter­
parts. Berkeley: University o f California Press, 1975.
H a v e l l , E.B. The Ancient and Medieval Architecture o f India: A Study o f
 Indo-Aryan Civilization. London: John Murray, 1915.
H ayavadana R a o , C . ‘Early Indian Finance.’ The Indian Antiquary,
vol. 40 (September 1911), 265-72.
------History o f Mysore, 1399-1799. 2 vols. Bangalore: Government Press,
1943.
------ ‘The Place-Names o f M ysore.’ Quarterly Journal o f the Mythic Society,
vol. 5 (1915-16), 264-81.
------ ‘Rural Habitation in South India.’ Ibid., vol. 14 (1923), 318-23.
H e e s t e r m a n , J.C. The Ancient Indian Royal Consecration. ’S-Gravenhage:
Mouton, 1957.
H e r a s , R e v . H e n r y . The Aravidu Dynasty o f Vijayanagara. Madras:
B.G. Paul and C o., 1927.
------Studies in Pallava History. Madras: B.G. Paul and Co., 1933.
------ and V.K. B h a n d a r k a r . ‘Vijayanagara Empire: A Synthesis o f South
Indian Culture.’ Vijayanagara Sexcentenary Commemoration Volume.
Dharwar: Karnatak Historical Research Society, 1936. Pp. 29-38.
H i l l , S.C. Yusuf Khan: The Rebel Commandant. London: Longmans,
Green and C o., 1914.
H i r t h , F r e d r i c h and W .W . R o c k h i l l . Chau Ju-Kua: His Work on the
Chinese and Arab Trade in the Twelfth and Thirteenth Centuries, entitled
Chu-Fan-Chi. Translated from the Chinese. St. Petersburg: Imperial
Academy o f Science, 1911.
History o f the Sourashtras o f Southern India (Publication o f the Sourashtra
Literary Society). Madras: Graves, C ookson and C o., 1891. Pp. 20.
History o f Tipu Sultan; Being a Continuation o f the Neshani Hyduri by Mir
Hussain A li Khan Kirmani. Trans. C o l W. M i l e s . (Originally pub­
lished 1864 by the Oriental Translation Fund o f Gt. Britain and Ireland.)
Calcutta: Susil Gupta, 1958.
H o c a r t , A .M . Kings and Councillors; An Essay in the Comparative Anatomy
o f Human Society. Ed. R . N e e d h a m . C hicago: University o f Chicago
Press, 1970.
H o p k i n s , E .W . ‘T h e D iv in ity o f K in g s .’ Journal o f the American Oriental
Society, v o l. 51 (1931). Pp. 309-16.
H u l t z s c h , E. ‘A Copper-Plate, Grant o f Ganadeva o f Kondavidu, dated
Saka-Samvat 1377.’ The Indian Antiquary, vol. 20 (1891), 390-3.
---------‘Karnata Grants, N o. 1 : A Grant o f Venkata II o f 1636 a .d . ’ Ibid.,
vol. 13 (1884), 125-33.
------ ‘Kopparam Plates o f Pulakesin II.’ Epigraphia Indica, vol. 18, 257-60.
Postscript note b y R o b e r t S e w e l l , 260 -1 .
------ ‘Motupalli Pillar Inscription o f Ganapatideva, a . d . 1244-45.’ Ibid.,
vol. 12, 188-97.
------ ‘N ote on a Tamil Inscription in Siam’. The Journal o f the Royal Asiatic
Society o f Great Britain and Ireland, 1913, 337-39.
H u t t o n , J.F. Caste in India; Its Nature, Function, and Origins. 4th ed.
 Bombay: Oxford University Press, 1963.
idahgai-valahgai ja tiya r varala.ru [History o f the Left Hand and Right Hand
Castes]. Mackenzie Collection, Oriental Manuscripts Library, Madras.
D ocum ent no. R 1572.
idangai-valangai puranam [Ancient Story o f the Left Hand and Right Hand
Castes]. Mackenzie Collection, Oriental Manuscripts Library, Madras.
D ocum ent no. D 2793.
I l l a k k U v a n a r , S. Tholkappiyam in English; with Critical Studies. Madura:
Kural Neri Publishing House, 1963.
J a g a d i s a A y y a r , P.V. South Indian Festivities. Madras: Higginbothams,
Ltd., 1921.
------ South Indian Shrines. M adras: The Madras Times Publishing Co.,
1920.
J o u v e a u - D u b r e u i l , G. Dravidian Architecture. Ed. S. K r i s h n a s w a m i
A i y a n g a r . Madras: S.P.C.K. Press, 1917.
K a i l a s a p a t h y , K . Tamil Heroic Poetry. O xford: Clarendon Press, 1968.
K a n a k a s a b h a i P i l l a i , V. ‘Tamil Historical Texts, N o. 2, The Kaiingattu-
Parani.’ The Indian Antiquary, vol. 19 (1890), 329-40.
------ ‘Tamil Historical Texts, N o. 4, The Vikrama-Cholan-Ula.’ Ibid.,
vol. 22 (1893), 141-50.
------ The Tamils Eighteen Hundred Years Ago. Madras: Higginbothams,
Ltd., 1904.
K a n e , P.V. H istory o f Dharmasastra (Ancient and M edieval Religious and
Civil Law). 4 vols. (Government Oriental Series, Class B, no. 6.) Poona:
Bhandarkar Oriental Research Institute, 1930-53.
K a r a n d i k a r , D . K . E xtracts from the Yajnavalkya Smriti. B om bay: Taibhav
Press, 1913.
K a r a s h i m a , N o b o r u . ‘Allur and Isanamangalam: Two South Indian
Villages in C ola Times.’ Proceedings o f the First International Con­
ference Seminar o f Tamil Studies (International A ssociation o f Tamil
Studies). Kuala Lumpur: International Association o f Tamil Re­
search, 1968. Pp. 426-37.
------ ‘Four Tamil Inscriptions o f the Cola Period.’ The Tokyo Gakuho,
vol. 48, no. 2 (1965), 1-27 (in Japanese).
------ ‘History o f South A sia.’ Oriental Studies in Japan: Retrospect and
Prospect, 1963-1972. T okyo: The Centre for East Asian Cultural
Studies, 1974, pt II, 1-17.
------ ‘Minami indo shakai no re kishi teki hatten [Historical Development
o f South Indian Society].’ Indo shi okeru tochi seido to kenryoku Rozo.
Ed. T o r u M a t s u i a n d T o s h i o Y a m a z u k i . T okyo: Tokyo University
— Press, 1969. Pp. 73-105.
------ ‘The Power Structure o f the Cola R ule.’ Paper presented to the Ilnd
International Seminar-Conference o f Tamil Studies, Madras, 5 January
1968.
------ and B. S i t a r a m a n . ‘Revenue Terms in Chola Inscriptions.’ Journal
o f Asian and African Studies (Tokyo), no. 5 (1972), 87-117.
------ and Y . Subbarayalu. ‘A Statistical Study o f Personal Names in Tamil
Inscriptions (interim Reports [March 1975] and II [March 1976].’
Computational Analyses o f Asian and African Languages. N o. 1 and 3.
Ed. M a n t a r o J. H a s h i m o to . T okyo: N ational Interversity Research
Institute o f Asian and African Languages, 1975 and 1976. Pp. 3-17
and 9-21.
kärmandala-satakam [by] Araikilar [of] Avanasi. Ed. P.A. M u t h u t h a n a d a -
v a r a y a P i l l a y . Madras: 1930.
K a r m a r k a r , A.P. Cultural History o f Karnataka (Ancient and Medieval).
Dharwar: Karnataka VidyavaidhalkaSangha, 1947.
K a r v e , I r a w a t i . ‘India as a Cultural R egion.’ Indian Anthropology:
Essays in M emory o f D.N. Majumdar. Eds. T .N . M a d a n and G o p a l a -
S a r n a . Bombay: Asia Publishing House, 1962. Pp. 325-38.
------ Kinship Organization in India (Deccan College M onograph Series,
no. 11). Poona: D eccan College, 1953.
Kautilya Arthasastra. Trans. R. S h a m a s a s t r y . (Government Oriental Li­
brary Series, Bibliotheca Sanskrita, no. 37, pt II.) Bangalore: Govern­
ment Press, 1915.
K e a r n s , J.F. ‘The Right and Left Hand Castes.’ The Indian Antiquary,
vol. 5 (December 1876), 353-4.
konqumandala-satakam [by] Karmagakkavinar . Ed. Ti. A. M u t t u s a m i k -
k o n a r . Tiruchengodu: 1923.
K o s a m b i, D .D . ‘The Basis o f Ancient Indian History.’ Journal o f the
American Oriental Society, vol. 75 (1955), pt 1, 34-45, pt 2, 226-36.
------ The Culture and Civilization o f Ancient India. London: Routledge and
Kegan Paul, 1965.
------An Introduction to the Study o f Indian History. Bombay: Popular
Book D epot, 1956.
------ ‘The Working Class in the Amarakosa.’ Journal o f Oriental Research,
vol. 24 (1955), 57-69.
K o s m in s k y , E.A. Studies in the Agrarian History o f England in the Thirteenth
Century. Ed. R .H . H i l t o n . Oxford: Blackwell, 1956.
The Koyil Olugu (History o f the Srirangam Temple); Condensed English
Version. E d . T.S. P a r t h a s a r a t h y . Tirupati: Tirumalai-Tirupati
Devasthanams, 1954.
K r ä d e r , L. Formation o f the State (Foundations o f Modern Anthropology).
Englewood Cliffs, N .J .: Prentice-Hall, 1968.
K r a m r i s c h , S t e l l a . The Hindu Temple. 2 vols. Calcutta: University of
Calcutta, 1946.
K r i s h n a I y e r , K.V. ‘The Temple States o f Kerala.’ Proceedings o f the
Tenth All-India Oriental Conference, Tirupati. M adras: 1941. Pp. 507-15.
K r i s h n a m u r t h i , S.R. A Study o f the Cultural Developments in the Chola
 Period. Annamalainagar: Annamalai University, 1966.
K ris h n aR a o , B.V. ‘The Haihayas o f Palnad.’ Journal o f the Andhra
Historical Research Society, vol. 12 (1938-9), 29-44.
------ ‘K ota Chiefs o f Amaravati.’ Ibid., vol. 2 (1927), 34-49 and 270.
K r i s h n a S a s t r i , H. ‘The First Vijayanagara D ynasty; Its Viceroys and
M inisters.’ Archaeological Survey o f India, Annual Report, 1907-8.
Calcutta: Government Press, 1911. Pp. 235-54.
------ ‘Fiscal Administration Under Early C olas.’ Commemorative Essays
Presented to Sir Ramakrishna Gopal Bhandarkar. Poona: Bhandarkar
Oriental Research Institute, 1917.
K r i s h n a s w a m i A i y a n g a r , S. Ancient India and South Indian History and
Culture. 2 vols. (Poona Oriental Series, no. 74.) Poona: Oriental
Book Agency, 1941.
------ ‘The Character and Significance o f Vijayanagara in Indian History.’
Vijayanagara Sexcentenary Volume. Dharwar: Karnatak Historical
Research Society, 1936. Pp. 1-28.
—— Early History o f Vaishnavism in South India. M adras: Oxford University
Press, 1920.
------Evolution o f Hindu Administrative Institutions in South India (Sir William
Meyer Lectures, 1929-30). Madras: University o f Madras, 1932.
—.... A History o f the Holy Shrine o f Sri Venkatesa in Tirupati. 2 vols. (Sri
Venkatesa Oriental Institute Series.) Madras: Tirumalai-Tirupati D e-
vasthanam Committee, 1941.
------ (ed.). Paramasamhita [of the Pamcaratra] (Gaekwad’s Oriental Series,
no. 86). Baroda: Oriental Institute, 1940.
—— Sri Ramanujacharya : A Sketch o f His Life and Times. Madras:
G.A . Natesan and C o., n.d. (c\ 1906).
------‘Social Legislation under Hindu Governments.’ Quarterly Journal o f
the M ythic Society, vol. 6 (1915-16), 47-56.
K r i s h n a s w a m i P i l l a i , A. The Tamil Country under Vijayanagara (Anna­
malai Historical Series, no. 20). Annamalainagar: Annamalai U ni­
versity, 1964.
L a i r d , J.L . ‘On Coomrie Cultivation.’ The Indian Forester [Calcutta],
vol. 1 (1875-6), 11-16.
L a k s h m i n a r a y a n , R a o , N . ‘Some N ew Facts about Chdla History.’
Journal o f Oriental Research, vol. 19 (1950), 148-51.
L e a c h , E.R. Political Systems o f Highland Burma. Cambridge: Harvard
University Press, 1954.
L i n g a t , R o b e r t . The Classical Law o f India (Sponsored by the Center
for South and Southeast Asia Studies). Berkeley: University o f Cali­
fornia Press, 1973.
L o n g h u r s t , A .H . Hampi Ruins Described and Illustrated. Madras: Govern­
ment Press, 1917.
L o r e n z e n , D a v i d N . The Kapalikas and Kalamukkas: Two Lost Saivite
 Sects. New Delhi: Thomson Press, Ltd., 1972.
L ynch, O w en M . ‘Rural Cities in India: Continuities and Discontinuities.’
India and Ceylon: Unity and Diversity. Ed. P. M a s o n . London: Oxford
University Press, 1967. Pp. 142-58.
M c c o r m a c k , W.C. Kannada: A Cultural Introduction to the Spoken Styles
o f the Lanugage. Madison: University o f Wisconsin Press, 1966.
M a c d o n e l l , A.A. and A.B. K e i t h . Vedic Index o f Names and Subjects.
2 vols. London: J. Murray, 1912.
M a c k e n z i e , J.S.F. ‘Caste Insignia.’ The Indian Antiquary, vol. 4 (November
1876), 344-6.
M a c l e a n , C.D. Standing Information Regarding the Official Administration
o f the Madras Presidency in Each Department, in Illustration o f the
Yearly Administrative Reports Prepared Under the Orders o f Government.
(Madras: Government Press, 1877.)
The Mahabharata o f Krishna-Dwaipayana Vyasa. Trans. P.C. Roy. 12 vols.
Calcutta: Oriental Publishing Co., 195-.
M a h a l i n g a m , T .V . Administration and Social Life Under Vijayanagar.
Madras: University o f Madras, 1940.
------‘The Banas in South Indian History.’ Journal o f Indian History, vol. 29
(August 1951), 153-81 and 277-305.
------‘The Citrameli Periyanättär.’ Dr. R.P. Sethupillai Jubilee Commemora­
tion Volume. Madras: Palaniappa Bros., 1961. Pp. 195-9.
------Economic Life in the Vijayanagar Empire. Madras: University of
Madras, 1951.
------Känclpuram in Early South Indian History. Madras: Asia Publishing
House, 1969.
------‘The Nagesvarasvami Temple.’ Journal o f Indian History, vol. 45,
pt 1 (April 1967), 1-95.
------South Indian Polity. 2nd ed. Madras: University of Madras, 1967.
M a j o r , R.H. India in the Fifteenth Century. London: Hakluyt Society, 1857.
M a j u m d a r , R.C. Ancient Indian Colonization in Southeast Asia. Baroda:
M .S . University o f Baroda, 1955.
------Corporative Life in Ancient India. Calcutta: S.N. Sen, 1918.
Manusmrti; The Laws o f Manu, with the Bhäsya o f Medhätithi. 4 vols.
Trans. G a n g a n a t h a J h a . Calcutta: University o f Calcutta, 1922.
M a r r i o t t , M c k im . Caste Ranking and Community Structure in Five Regions
o f India and Pakistan (Deccan College Monograph Series, no. 23).
Poona: Deccan College, 1960.
------Village India: Studies in the Little Community. Chicago: University
of Chicago Press, 1955.
------and R o n a l d B. I n d e n . ‘Caste Systems.’ Encyclopaedia Britannica
(15th ed.), vol. 3. Pp. 982-91.
M a t s u i , T o o r u . ‘A Statistical Study o f Personal Names in Tamil; Inscrip­
tions (Interim Report III).’ Computational Analyses o f Asian and
 African Languages, no. 3. Ed. M a n t a . r o J . H a s h i m o to . Tokyo:
National Interuniversity Research Institute of Asian and African
Languages, March 1976. Pp. 21-39.
M a y e r , A d r i a n . ‘Quasi-Groups in the Study of Complex Societies.’ The
Social Anthropology o f Complex Societies. Ed. M. B a n t o n (Association
of Social Anthropologists Monograph, no. 4). New York: F. A . Praeger,
1966. Pp. 97-121.
M e n c h e r , J o a n P. "Kerala and Madras: A Comparative Study of Ecology
and Social Structure.' Ethnology, vol. 5, no. 2 (April 1966), 135-71.
M i n a k s h i , C. Administration and Social Life Under the Pallavas. Madras:
University of Madras, 1938.
------TTif Historical Sculptures o f the Vaikunlhaperumal Temple, KancT
(Memoirs o f the Archaeological Survey of India, no. 63). Delhi:
Manager of Publications, 1941.
M o n i e r - W i l l i a m s , M. Hinduism. London: Society for Promoting Christian
Knowledge, 1906.
------A Sanskrit-English Dictionary, Etymologically and Philologically
Arranged with Special Reference to Cognate Indo-European Languages.
Oxford: Clarendon Press, 1960.
M o r a e s , G e o r g e . The Kadamba Kula; A History o f Ancient and Medieval
Karnataka. Bombay: B.X. Furtado and Sons, 1931.
M o r e l a n d , W i l l i a m H a r r i s o n . The Agrarian System o f Moslem India.
Cambridge: W. Heffer, 1929.
------From Akbar to Aurangzeb; A Study in Indian Economic History. London:
Macmillan, 1920.
M o r g a n , F.E. ‘Irrigation.’ Southern India. Ed. S o m e r s e t P a y n e . London:
Foreign and Colonial Compiling Publication Co., 1914-15.
M o r r i s , M.D. and B u r t o n S t e i n . ‘The Economic History o f India; A
Bibliographical Essay.’ Journal o f Economic History, vol. 21 (June
1961), 179-201.
M u k h e r j e e , N i l m a n i . TheRyotwariSystem in Madras: 1792-1827. Calcutta:
Firma K.L. Mukhopadhyay, 1962,
------and R o b e r t E r i c F r y k e n b e r g . ‘The Ryotwari System and Social
Organization in the Madras Presidency.’ Land Control and Social
Structure in Indian History. E d . R o b e r t E r i c F r y k e n b e r g . Madison:
The University o f Wisconsin Press, 1969. Pp. 217-26.
M u k h e r j e e , R.K. Local Government in Ancient India. Oxford: Clarendon
Press, 1919.
M u l l a l y , F.S. Notes on the Criminal Tribes o f the Madras Presidency.
Madras: 1892.
N a l l a s v a m i P i l l a i , J.N. Studies in Saiva-Siddhanta. Madras: Meykandan
Press, 1911.
N a r a h a r a y y a , S.N. ‘Keladi Dynasty.’ Quarterly Journal o f the Mythic
Society, vol. 21 (1930-1), 72-87.
N a ra s im h a R a o , R . Corporate Life in Medieval Andhradesa. Secundarabad:
University Grants Commission, 1967.
N a r a y a n a A y y a r , C.V. [Swami Sadananda]. Origin and Early History o f
Saivism in South India. Madras: University o f Madras, 1936.
N e a l e , W a l t e r C. ‘Lane is to Rule.’ Land Control and Social Structure
in Indian History. E d . R o b e r t E r i c F r y k e n b e r g . Madison: University
o f Wisconsin Press, 1969. Pp. 3-17.
N i c h o l a s , R a l p h W. ‘Factions: A Comparative Analysis.’ Political
Systems and the Distribution o f Power. Ed. M. B a n t o n . (Association
o f Social Anthropologists, Monograph, no. 2.) London: Tavistock
Publications, 1969.
------‘Structure o f Politics in the Villages o f Southern Asia.’ Structure and
Change in Indian Society. E d s . M i l t o n S i n g e r a n d B e r n a r d S. C o h n .
C h ic a g o : A ld in e , 1968.
N i c h o l s o n , S id n e y . ‘Social Organization o f the Malas—An Outcaste
People.’ Journal o f the Royal Anthropological Institute o f Great Britain
and Ireland, vol. 56 (1926), 91-105.
N i l a k a n t a S a s t r i , K.A. ‘The Beginnings o f Intercourse Between India
and China.’ Indian Historical Quarterly, vol. 14 (June 1938), 380-7.
------‘Cola Legends; Being a Detailed Notice of Some Manuscripts.’ Journal
o f Oriental Research [Madras], vol. 4, 318-40.
------The Colas. 2 vols. (Madras University Historical Series, no. 9). Madras:
University o f Madras, 1935-7.
------The Colas. 2nd ed. Madras: University o f Madras, 1955. (This one
volume edition differs from the two-volume edition of 1935-7 in sum­
maries of Chola inscriptions.)
------Cultural Expansion o f India. Madras: G.S. Press, 1959.
------The Culture and History o f the Tamils. Calcutta: Firma K.L. Mukho-
padhyay, 1965.
------Development o f Religion in South India. Madras: Orient Longmans,
1963.
------‘The Economy o f a South Indian Temple in the Cola Period.’ Malaviya
Commemoration Volume. E d . A.B. D h r u v a . Benaras: Benaras Hindu
University, 1932. Pp. 305-19.
------(ed.). Foreign Notices o f South India; from Megasthenes to M a Huan.
Madras: University of Madras, 1939.
------A History o f South India from Prehistoric Times to the Fall o f Vijayanagar.
Madras: Oxford University Press, 1955.
------The History o f Sri Vijaya. Madras: University o f Madras, 1949.
------‘A Note on Velaikkarar.’ Journal o f the Royal Asiatic Society (Ceylon),
vol. 4 (New Series, 1954), 67-71.
------The Pandyan Kingdom: From the Earliest Times to the Sixteenth Century.
London: Luzac and Co., 1929.
------Sources o f Indian History (Under the Auspices o f the Heras Institute of
 Indian History and Culture). Madras: Asia Publishing Co., 1964.
------Studies in Cola History and Administration (Madras University His­
torical Series, no. 7). Madras: University o f Madras, 1932.
------‘Vaishnavism and the Early Alvars.’ Transactions o f the Archaeological
Society o f South India, Silver Jubilee Volume, 1962. Pp. 120-5.
------- a n d N. V e n k a t a r a m a n a y y a . Further Sources o f Vijayanagara History.
(Madras University Historical Series, no. 18.) Madras: University of
Madras, 1946.
------and T.N. S u b r a m a n i a m . ‘Hemavati Pillar Inscription o f Kulottunga-
chola (III), Year 2.’ Epigraphia Indica, vol. 31, pp. 269-76.
O a t e n , E d w a r d F. European Travellers in India. London: K. Paul, Trench,
Trabner, 1909.
O p p e r t , G u s t a v . On the Original Inhabitants o f Bharatavarsa or India.
Delhi: Oriental Publishers, 1972 (Orignially published Madras Journal
o f Literature and Science, 1887-8 and 1888-9). Pp. 711.
The Ordinances o f Manu. Ed. and Trans. A.C. B u r n e l l and E.W.H o p k in s .
London: 1884.
O r m e , R o b e r t . A History o f the M ilitary Transactions o f the British Nation
in Indostanfrom the Year 1659. 2 vols. London: John Nourse, 1778-80.
pandimandala-satakam [by] Aiyan Perumal Asirayar [of\ Madurai.
Sirkali: 1932.
P a n i k k a r , K.M. Geographical Factors in Indian History. Bombay: Bhara­
tiya Vidya Bhavan, 1959.
P a r a n j o t i , V i o l e t . Saiva Siddhanta in the Meykandar Sastra. London:
Luzac and Co., 1938.
P i l l a y , K.K. A Social History o f the Tamils. Madras: University of Madras,
1969.
------ South India and Ceylon (Sir William Meyer Lectures, 1958-9). Madras:
University of Madras, 1963.
------The Suclndram Temple. Madras: Kalakshetra Publications (Adyar),
1953.
P o c o c k , D a v i d F . ‘Notes on jajmani Relationships.’ Contributions to
Indian Sociology, vol. 6 (1962), 78-96.
------‘Sociologies: Urban and Rural.’ Ibid., vol. 4 (April 1960), 63-82.
P o l a n y i , K. et al. Trade and M arkets in Early Empires. Glencoe: Free
Press, 1957.
P u r i , B.N. Cities o f Ancient India. Meerut: Meenakshi Prakashan, 1966.
P u t t a i y a , B. ‘The Kempe Gowda Chiefs.’ Quarterly Journal o f the M ythic
Society, vol. 12 (1922-3), 723-41.
R a d h a K r i s h n a , K.N. Thirumalisolaimalai (Sri Alagar Kovil) Sthala
Purana. Madura: Sri Kallalagar Devasthanam, 1942.
R a g h a v a n , V. Cola Campit o f Virupaksa (Tanjore Saraswati Mahal Series,
no. 62). Tanjavur: T.M.S.S.M. Library, 1952. Pp. xi, 46.
------‘Notes on “Society and Life.’ ” Typescript Manuscript. Madras, 1968.
------'The “Vaisyavamsasudhâkara” of Kolâcala Mallinatha.’ New Indian
Antiquary (Bombay), vol. 2 (1939-40), 442-8.
------‘Variety and Integration in the Pattern of Indian Culture.’ Journal o f
Asian Studies, vol. 15 (August 1956), 497-506.
R a j a g o p a l a c h a r i a r , T. The Vaishnavite Reformers o f India; Critical
Sketches o f Their Lives and Writings. 2nd ed. Madras: G.A. Natesan
and Co., 1909.
R a j a m a n i c k a m , M. The Development o f Saivism in South India ( a .d . 300-
1300). Dharmapuram: Dharmapuram Adhinam, 1964.
R a j u , P.T. Telugu Literature (Andhra Literature) . (The P.E.N. Books,
IndianLiteralure.no. 15). Bombay: International Book Service, 1944.
R a m a c h a n d r a C h e t t i a r , C.M. ‘Place Names in Tamil Nad.’ Journal of
the Madras Geographical Association, vol. 13 (1938), 32-57.
------‘Rule o f Vijayanagara Over Kongu Country.’ Vijayanagara Sexcente­
nary Commemoration Volume. Dharwar : Karnatak Historical Research
Society, 1936. Pp. 275-88.
------‘Social Legislation in Ancient South India.’ Quarterly Journal o f the
Mythic Society, vol. 21 (1930-1), 341-53 and vol. 22 (1931-2), 65-71.
------Studies in Tamil Literature and History. London: Luzac and Co.,
1930.
R a m a c h a n d r a D i k s h i t a r , V .R . Tirukkural o f Tiruvalluvar (in Roman
Transliteration) with English Translation. Madras: The Adyar Library,
1949.
------War in Ancient India. Madras: Macmillan and Co., 1944.
R a m c h a n d r a i y a , O. and K. S u n d a r a m . ‘The Teliki Community of Me­
dieval Andhra.’ Journal o f Indian History, vol. 39, pt 1 (April 1961),
79-88.
R a m a c h a n d r a M u d a l i a r A v a r g a l , D.B. ‘Mudaliar.’ Quarterly Journal
o f the Mythic Society, vol. 10 (1919-20), 289-99.
R a m a c h a n d r a M u r t y , V.S. ‘The Authorship and Importance of the
Âmuktamâlyadà as a Source o f History in the Reign of Krishnadevaraya.
Vijayanagara Sexcentenary Commemoration Volume. Dharwar: Kar­
natak Historical Research Society, 1936. Pp. 207-12.
R a m a c h a n d r a R a o , P.B. ‘Poligars of Mysore and Their Civilization:
A Study." Quarterly Journal o f the Mythic Society, vol. 29 (1938-9),
150-62, 310-~25 and vol. 30 (1939-40), 175-91, 423-8.
R a m a k r i s h n a A y y e r , V.G. [‘Historicus’]. ‘The Economy of a South
Indian Temple [Chidambaram].’ Mss. submitted for the Sankara Par-
vathi Prize, University o f Madras, 1943-4. University of Madras
Library.
R a m a k r i s h n a R a o , B. ‘The Dasara Celebrations in Mysore.’ Quarterly
Journal o f the Mythic Society, vol. 11 (1921), 301-11.
R a m a m u r t h y , K. ‘Some Aspects of the Regional Geography of Tamilnad.’
Indian Geographical Journal, vol. 23, no. 2 (April-June 1948), 1-137.
R am a R ao , M. Krishnadeva Raya. New Delhi: National Book Trust,
National Biography Series, 1971.
------‘The Recherla Family.’ Journal o p the Andhra Historical Research
Society,‘vol. 5 (1931), 139-50.
R a m a n , K.V. The Early History o f the Madras Region. Madras: Amudha
Niiayam Private Ltd., 1959.
R a m a s w a m i A y y a n g a r , M.S. and B. S e s h a g i r i R a o . Studies in South
Indian Jainism. Madras: Hoe and Co., 1922.
R a m a s w a m i N a i d u , B. ‘Remarks on the Revenue System and Landed
Tenures o f the Provinces Under the Presidency o f Fort St. George
(communicated by J o h n H o d g s o n ) . ’ Journal o f the Royal Asiatic
Society o f Great Britain and Ireland, 1934, 292-306.
R a m e s h , K.V. A History o f South Kanara fFrom the Earliest Times to the
Fall o f Vijayanagara). Dharwar: Karnatak University, 1970.
R a n d h a w a , M.S., et al. Farmers o f India, vol. 2. Madras, Andhra Pradesh,
Mysore and Kerala. New Delhi: Indian Council o f Agricultural Re­
search, 1961.
R a n g a c h a r i , K . The Sri Vaishnava Brahmans (Bulletin of the Madras
Government Museum, New Series, General Section, vol. 2). Madras:
Government Press, 1931.
R a n g a c h a r i , V. ‘The History of Sri Vaishnavism; From the Death of
Sri Vedanta Desika to the Present Day.’ Quarterly Journal o f the
M ythic Society, vol. 7 (1916-17), 106-18, 197-209.
R a n g a n a t h a R a o , S. ‘The Belagutti Kaifiyats.’ Quarterly Journal o f the
M ythic Society, vol. 35 (1944), 61-71.
R a n g a s w a m i A i y a n g a r , K.V. Aspects o f Ancient Indian Economic Thought
(Manindra Chandra Lectures, 1927, Banares Hindu University).
Banares: Hindu University, 1934.
R a n g a s v a m i S a r a s v a t i , A. ‘Political Maxims of the Emperor Poet, Krishna­
deva Raya.’ Journal o f Indian History, vol. 6, pt 3 (1925), 61-88.
R a g h a v i a h , T. ‘The Economic Organization o f the Nellore District Under
Hindu Rule.’ Journal o f the Madras Economic Association, 1915, 31-52.
R a y a c h a u d h u r i , H.C. Political History o f Ancient India. 4th ed. Calcutta:
University of Calcutta, 1938.
R a y c h a u d h u r i , T a p a n . Jan Company in Coromandel, 1605-1690 (Verhan-
delingen van Het Koninklijk Instituut voor Taal-, Land-, en Volken-
kunde, Deel 38). ’S-Gravenhage: Martinus Nijhoff, 1962.
R e d f i e l d , R o b e r t . Peasant Society and Culture: An Anthropological
Approach to Civilization. Chicago: University o f Chicago Press, 1956.
R e n a u d o t , E u s e b iu s . Ancient Accounts o f India and China by Two Moham­
medan Travellers Who Went to These Parts in the 9th Century. T ans.
from the Arabic. London: 1733.
R ic e , B. L e w is . ‘The Ganga Kings.’ Madras Journal o f Literature and Science
(1878), 138-48.
 ------‘Gangavadi.’ Commemorative Essays Presented to Sir Ramakhshna
Gopal Bhandarkar. Poona: Bhandarkar. Pp. 237-48.
------‘A Jain Vaisnava Compact.’ The Indian Antiquary, vol. 14 (1884),
233-5.
------‘Mudyanur Plates of Saka 261 of the Bana King Malladeva-Nandi-
varman.’ The Indian Antiquary, vol. 15 (June 1886).
------Mysore and Coorg from the Inscriptions. London: Archibald Constable
and Co., Ltd., 1919.
R i c h a r d s , F.J. 'Money-Lending.’ Quarterly Journal o f the Mythic Society,
vol. 4 (1914), 122-3.
------‘Race Drift in South India.’ The Indian Antiquary, vol. 59 (1930),
211-18.
R i t t i , S h r i n i v a s . ‘The Belavola-Desa.’ Jagadguru Tontadurya College
Miscellany, vol. 4 [Reprint; no date or publication information]. Pp. 11.
------‘Some Women Administrators of Medieval Karnataka.’ Quarterly
Journal o f the M ythic Society, vol. 48 (1957-8), 43-54.
R u d o l p h , L l o y d I. and S u s a n n e . The Modernity o f Tradition; Political
Development in India. Chicago: University o f Chicago Press, 1967.
S a h l i n s , M a r s h a l l . ‘Poor Man, Rich Man, Big-man, Chief: Political
Types in Melanesia and Polynesia.’ Comparative Studies in Society
and History, vol. 5 (1963), 285-303.
------‘The Segmentary Lineage: An Organization o f Predatory Expansion.’
Comparative Political Systems. Eds. R o n a l d C o h e n and J o h n M i d ­
d l e t o n . New York: The National History Press, 1967. Pp. 89-121.
------Stone-Age Economics. Chicago: Aldine, 1972.
S a l e t o r e , B.A. Ancient Karnataka, vol. I, History o f Tuluva. Poona:
Oriental Book Agency, 1936.
------Social and Political Life in the Vijayanagara Empire ( a . d . 1346-1646).
2 vols. Madras: B.G. Paul, 1934.
------The Wild Tribes in Indian History. Lahore: Motiial Banarsidass, 1935.
S a l e t o r e , R.N. ‘Features o f Bedara Administration in South India.’ Journal
o f the Andhra Historical Research Society, vol. 11 (1937-8), 149-53.
S a r a d a R a j u , A . Economic Conditions in the Madras Presidency, 1800-1850
(Madras University Economic Series, no. 5). Madras: University of
Madras, 1941.
S a s t r y , K.R.R. ‘The Madura Gilds— A Type.’ Journal o f the Indian Econo­
mic Society, vol. 1 (1922), 128-37, 183-96.
------South Indian Gilds. Madras: Indian Publishing House, Ltd., 1925.
Pp. 48.
The Satapatha Brahmana According to the Text o f the Madhyandina School.
Trans. J. E g g e i.in e (Sacred Books o f the East. Ed. M . M u l l e r .
vol. 44, p t 5). Delhi: M o tila i Banarsidass, 1963.
S a t h y a n a t h a A i y e r , R. [also: S a t h i a n a t h a i e r ] . History o f the Nayaks
o f Madura (The Madras University Historical Series, no. 2, ed.
 B ib liography 5]1

S. K r i s h n a s w a m i A i y a n g a r ) . Madras: Oxford University Press, 1924.

------Studies in the Ancient History ofTondamandalam (University o f Madras,
The Sankara-Parvathi Lectures, 1943-4). Madras: Rochouse and Sons,
1944. Pp. 71.
------Tamilaham in the 17th Century (Sir William Meyer Lectures, ¡954-5,
University o f Madras). Madras: University o f Madras, 1956.
S a u n d e r s , A.J. ‘The Saurastra Community of Madura.’ journal o f the
Madras University, vol. 1 (1928), 69-92.
S esha I y e n g a r , T.R. Dravidian India, vol. 1. Madras: C. C o o m a ra sw a m y
Naidu and Sons, 1933.
S e t h u P i l l a i , R.P. ‘Some Historical Place-Names in South India.' Pro­
fessor K.V. Rangaswami Aiyangar Commemoration Volume. Madras:
G.S. Press, 1940. Pp. 257-63.
S e w e l l , R o b e r t . A Forgotten Empire (Vijayanagar) ; A Contribution to
the History o f India. London: Swan Sonnenschein and Co., Ltd., 1900.
S h a n , A.M. and M.N. S r i n i v a s . ‘The Myth o f Self-Sufficiency of the
Indian Village.’ The Economic Weekly (Bombay), 10 September 1960,
1376.
S h a h a n i , J i v a t r a m S a v i t r i . ‘A Comparative Study o f Traditional Political
Organization o f Kerala and Punjab.’ Unpublished Ph.D. thesis, Univer­
sity o f London, School o f Oriental and African Studies, 1965.
S h a m a s h a s t r i , R. ‘Economical Philosophy o f the Ancient Indians.’ Annals
o f the Bhandarkar Oriental Research Institute, no. 12 (1930-1), 25-39.
------‘Sivananda’s Life o f Appaya Dikshit.’ Quarterly Journal o f the Mythic
Society, vol. 11 (1920-1), 113-21.
S h a r m a , R.S. Indian Feudalism: c. 300-1200. Calcutta: University of
Calcutta, 1965.
S h e r w a n i , H.K. The Bahmani Kingdom (Short Studies in Indian History,
Government o f India). Bombay: National Information and Publica­
tions Ltd., 1947.
S i n g a r a v e l u , S . Social Life o f the Tamils. Kuala Lumpur: Department
o f Indian Studies, University o f Malaya, 1966.
S i n g e r , M i l t o n (ed.). Krishna: M yths, Rites, and Attitudes. Chicago:
Phoenix Books, 1968.
S i r c a r , D.C. The Guhilas o f Kiskindhd. Calcutta: Sanskrit College, 1965.
------Landlordism and Tenancy in Ancient and Medieval India as Revealed by
Epigraphical Records (Dr Radha Kumud Mookerji Endowment Lec­
tures, University o f Lucknow, 1964). Lucknow: University o f Lucknow,
1969.
------(ed.). Land System and Feudalism in Ancient India (Centre of Advanced
Study in Ancient Indian History and Culture, University of Calcutta).
Calcutta: University o f Calcutta, 1966. Pp. 139.
------Studies in the Geography o f Ancient and Medieval India. D elhi: Motilal
Banarsidass, 1960.
------‘Suryavamsi Gajapatis of Orissa.’ Indian Historical Quarterly, vol. 33
(September 1957), 275-87.
S i v a n a n d a , S w a m i. Hindu Feasts and Festivals; their Philosophy. Rishi-
kesh, U .P .: The Sivanand Press, 1947.
S i v a r a m a m u r t i , C. Royal Conquests and Cultural Migrations in South
India and the Deccan. Calcutta: Indian Museum, 1955.
------South Indian Bronzes. New Delhi: Lalit Kalä Akademi, 1963.
S o m a s e k h a r a S a r m a , M . History o f the Reddi Kingdoms (circa 1325 a .d .
to circa 1448 a . d . ) (Andhra University Series, no. 38). Waltair, A.P.:
Andhra University, 1948.
S o m a s u n d a r a m P i l l a i , J.M. Five Tamil Idylls o f Paththup-pättu; Being
Studies and Translations. Madras: Golden and Co., 1947.
------The Great Temple at Tanjore. Madras: Golden and Co., 1935.
------A History o f Tamil Literature. Annamalainagar: Annamalai University,
1967.
S o n n e r a t , P i e r r e . Voyage aux Indes Orientales et La Chine. Paris: 1782
[James Ford Bell Library, University o f Minnesota],
S o u n d a r a R a j a n , K.V. ‘Kaustubha Prasada— N e w Light on Jayakha
Tantra.’ Journal o f the Oriental Institute, vol. 17, no. 1 (1967), 73-5.
Sources o f Vijayanagar History ; Selected and Edited for the University.
Ed. S. K r i s h n a s w a m i A i y a n g a r . (Madras University Historical Series.)
Madras: University o f Madras, 1919.
S o u t h a l l , A.W. Alur Society. Cambridge: W. Heffer, 1956.
S p a t e , O.H.K. India and Pakistan; A General and Regional Geography.
London: Methuen and Co., 1954.
S p e n c e r , G e o r g e W. ‘The Politics of Plunder; The Cholas in Eleventh
Century Ceylon.’ Journal o f Asian Studies, vol. 35 (May 1976), 405-19.
------‘Religious Networks and Royal Influence in Eleventh Century South
India.’ Journal o f the Economic and Social History o f the Orient, vol. 12,
pt 1 (1969), 42-57.
------‘Royal Leadership and Imperial Conquest in Medieval South India:
The Naval Expedition of Rajendra Chola I, c. 1025 a . d . ’ Unpublished
Ph.D. thesis, Department of History, University o f California, Berkeley,
1967.
------‘The Sacred Geography of the Tamil Shaivite Hymns.’ NU M EN:
International Review o f the History o f Religion, vol. 17 (December 1970),
232-44.
------‘Temple Money-Lending and Livestock Redistribution in Early Tanjore.’
The Indian Economic and Social History Review, vol. 5 (1968), 277-93.
S r e e n i v a s a c h a r , P. ‘Polipadu Grant o f Krishna-Deva-Raya.’ Journal o f
the Andhra Historical Research Society, vol. 10, pts 1-4(1936-7), 121-42.
- —‘Taxation in Vijayanagara.’ Indian Historical Quarterly, vol. 12 (1936),
443-55.
Sree R am a Sa rm a, P. ‘Saiuva Dynasty of Vijayanagar.’ Unpublished Ph.D.
thesis, Osmania University, Hyderabad, 1972.
S r i k a n t a i y a , S . ‘Life in the Hoysala Period.’ Quarterly Journal o f the
Mythic Society, vol. 8 (1918), 97-117.
S r i k a n t h a S a s t r i , S . Early Gangas o f Talakad. Bangalore: The Author,
1952.
The Sri M ad Devi Bhagavatam: Trans. S w a m i V i j n a n a n a n d a (H.P. C h a t -
t e r j i ) .(Sacred Books o f the Hindus, vol. 2 6 .) Allahabad: Panani Office,
1 9 2 1 -3 .
S r in iv a s , M.N. India’s Villages. London: Asia Publishing House, 1 9 6 0 .
------‘A Note on Sanskritization and Westernization.’ Journal o f Asian
Studies, vol. 15 (April 1956), 481-96.
S r i n i v a s a I y e n g a r , K.R. ‘Dravidian Language and Literature; Saiva
Siddhanta Literature.’ History and Culture o f the Indian People. Ed.
R.C. M a j u m d a r . Vol. 5. Bombay: Bharatiya Vidya Bhavan, 1966.
Pp. 361-88.
S r i n i v a s a A i y a n g a r , M. Tamil Studies; or Essays in the History o f the
Tamil People, Language, Religion, and Literature. Madras: Guardian
Press, 1914.
S r i n i v a s a c h a r i , C.S. ‘The Origin o f the Right and Left Hand Caste Divi­
sions.’ Journal o f the Andhra Historical Research Society, vol. 4 (1929),
77-85.
S r i n i v a s a I y e n g a r , P.T. History o f the Tamils; From the Earliest Times
to 600 a . d . (Published under the orders o f the Syndicate o f the Madras
University). Madras: C. Coomaraswamy and Sons, 1929.
S r i n i v a s a n , K.R. Some Aspects o f Religion as Revealed by Early Monuments
and Literature o f the South (Sankara Parvati Endowment Lectures
1959-60). Madras: University of Madras, 1960. Pp. 78.
------‘Tirukamakottam.’ Proceedings o f the Thirteenth All-India Oriental
Conference. Lucknow: Nagpur University, 1946. Pt 3, pp. 50-6.
S r i n i v a s a r a g h a v a n , K. ‘A Geographical Study of the Vellore Basin.’
Indian Geographical Journal, vol. 11, no. 3 ( 1 9 3 6 ) , 2 1 3 - 3 5 .
S t a a l , J.F. Nambudiri Veda Recitation. ’S-Gravenhage: Mouton, 1961.
S t e i n , B u r t o n . ‘All the Kings’ Mana: Perspectives on Kingship in Medieval
South India.’ Kingship and Authority in South Asia. Ed. J o h n F .
R i c h a r d s (Center of South Asian Studies, University of Wisconsin).
Madison: University of Wisconsin Press, 1977.
------(ed.). South Indian Temples; An Analytical Study. New Delhi: Vikas,
1978. Published as a supplemental volume of The Indian Economic
and Social History Review (Delhi), vol. 14, no. 1 (1977).
------‘Brahman and Peasant in Early South Indian History’, The Adyar
Library Bulletin (Dr V. Raghavan Felicitation Volume). Vols. 3 1 -2
( 1 9 6 7 - 8 ) . P p . 2 2 9 -6 9 .
------‘Circulation and Historical Geography in Tamil Country.’ Journal o f
Asian Studies, vol. 37 Traditional 1977 (November 1977), 7-27.
------‘Economic Functions of a Medieval South Indian Temple.’ Journal
o f Asian Studies, vol. 19, no. 2 (1960), 163-76.
------‘Medieval Coromandel Trade.’ Merchants and Scholars. Ed. J o h n
P a r k e r . Minneapolis: University of Minnesota Press, 1965. Pp. 49-62.
------‘The Segmentary State in South Indian History.’ Realm and Region
in India. E d . R o b e r t F o x . Durham: Duke University Press,
------‘The State and the Agrarian Order in Medieval South India: A Histo­
riographical Critique.’ Essays on South India. Ed. B u r t o n S t e i n .
(Asian Studies at Hawaii, no. 15.) Honolulu: The University Press of
Hawaii, 1975. Pp. 64-92.
------‘Social Mobility and Medieval South Indian Sects.’ Social Mobility
in the Caste System o f India; An Interdisciplinary Symposium. Ed.
Ja m e s S i l v e r b e r g . (Comparative Studies in Society and History; Sup­
plement III.) The Hague: Mouton, 1967. Pp. 78-95.
------‘Temples and Agricultural Development in Medieval South India.’
Economic Weekly Annual (Bombay), vol. 13 (February 1961).
Pp. 179-89.
S t o k e s , H.J. ‘The Custom of “Kareiyld” or Periodic Redistribution of
Land in Tanjore.’ The Indian Antiquary, vol. 3 (1874), 65-9.
S u b b a R a o , R . ‘History of Padmanaiks.’ Journal o f the Andhra Historical
Research Society, vol. 5 (1930-1), 50 if.
S u b b a r a o , B. The Personality o f India (M.S. University Archaeology
Series, no. 3). Baroda: MS., University of Baroda, 1956.
S u b b a r a o , N. ‘A S tu d y o f the Historical Geography of Walajapet.’ Educa­
tional Review (Madras), June 1932, pp. 2-14.
S u b b a r a y a l u , Y . Political Geography o f the Chola Country. M a d ra s :
T a m iln a d u S ta te D e p a rtm e n t o f A rc h ae o lo g y , 1973.
S u b h a R e d d i, N . ‘Community Conflict among the Depressed Castes of
Andhra.’ Man in India, vol. 30, no. 4 (1950), 1-12.
S u b r a h m a n i a n , N. ‘The Brahmans in Tamil Country (in Ancient and
Medieval Times).’ Lectures presented at the University of Madras,
18-19 March 1968; typescript of the lectures provided by the author.
------Pre-Pallavan Tamil Index; Index o f Historical Material in Pre-Pallavan
Tamil Literature (Madras University Historical Series, no. 23). Madras:
University of Madras, 1966.
------Sahgam Polity; The Administration and Social Life o f the Sahgam
Tamils. Madras: Asia Publishing House, 1966.
S u b r a h m a n y a A i y e r , K.V. Historical Sketches o f Ancient Dekhan. 2 vols.
Vol. 1: Madras: The Modern Printing Works, 1917; vol.’ 2: ed.
K.S. Vaidyanathan; Coimbatore: Coimbatore Printing Works, 1967.
------‘Largest Provincial Organizations in Ancient India.’ Quarterly Journal
 o f the M ythic Society, vol. 45 (New Series), no. 1, 29-47, no. 2, 70-98,
no. 4, 270-86; vol. 46, no. 1, 8-22.
------‘Seven Vatteluttu Inscriptions from the Kongu Country.’ Epigraphia
Indica, vol. 30, no. 19, pp. 95-512.
S u b r a m a n i a m , T.N. ‘The Huzur office Plates.’ Travencore Archaeological
Series, vol. 1, pp. 1-14.
------‘Pallankôvil Jaina Copper-Plate Grant of Early Pallava Period’. Trans­
actions o f the Archaeological Society o f South India, 1958-59, pp. 41-83.
------The Pallavas o f Kanchi in South-East Asia. Madras : The Swadesamitran
Ltd., 1967.
S u b r a m a n i a n , K.R. 'New Light on the Pallava Period.’ Maharaja’s College
Magazine (Vizianagaram), vol. 6, no. S, pp. 1-9.
S u n d a r a r a j a I y e n g a r , S. Land Tenures in the Madras Presidency. 2nd ed.
Madras: Commercial Press, 1921.
S u n d a r a m , K. The Simhachalam Temple. SimhachaSam, A. P . : Simha-
chalam Devasthanam, 1969.
------- Studies in Economic and Social Conditions o f Medieval Andhra ( a .d .
1000-1600). Madras: Triveni Publishers, 1968.
S u r e s h , B. ‘Historical and Cultural Geography and Ethnology of South
India (with Special Reference to Cola (Chola) Inscriptions).’ Unpub­
lished Ph.D. thesis, Deccan College Post-Graduate and Research
Institute, University of Poona, Poona, 1965.
------‘Raajaraajeesvaram at Tancaavuur.’ Proceedings o f the First Inter­
national Conference Seminar o f la m il Studies, Kuala Lumpur, 1964.
Kuala Lumpur: International Association of Tamil Research, 1968.
S w a m i n a t h a n , K.D. ‘The Horse Traders o f Malai Mandalam.’ Journal
o f Indian History, vol. 32, pt 2 (1954), 139-43.
------The Nâyakas o f Ikkëri. Madras: P. Varadachary and Co., 1957.
------‘Tiruvalisvaram : A Study o f Its History and Inscriptions.’ Unpublished
Ph.D. thesis, University of London, School of Oriental and African
Studies, London, 1964.
T a y l o r , R e v . W i l l i a m . Catalogue Raisonnée o f Oriental Manuscripts in
the Library o f the (L ate) College, Fort Saint George. 3 vols. Madras:
Fort S t. George Gazette Press, 1857-62.
------‘Marava-Jathi-Vernanam; From Unpublished Mackenzie Manuscripts
in Possession of the Asiatic Department o f the Madras Literary Society.’
The Madras Journal o f Literature and Science, vol. 4 (1836), 350-61.
------ (ed. and trans.). Oriental Historical Manuscripts. 2 vols. Madras:
C.S. Taylor, 1835.
------‘Translation of the Kongu-Desa-Rajaka!.' The Madras Journal o f
Literature and Science, vol. 14, pt I (1847), pp. 1-66.
T h a p a r , R o m i l a . Asoka and the Decline o f the Mauryas. Oxford : Oxford
University Press, 1961.
 B.M. ‘Traditional Limits and Sub-Divisions o f the Tamil
T h ir u n a r a n a n ,
Region.’ K.M . Rangaswamy Ayyangar Commemoration Volume. Madras:
G.S. Press, 1940. Pp. 159-69.
T h o m a s , P. Epics, M yths and Legends o f India. 8th ed. Bombay: Tarapore-
vala and Co., n.d.
T h o rn e r, D a n ie l. The A g ra ria n P rosp ect in In d ia . D e lh i : U n iv e r s ity
P re s s , 1959.
tondaimandala-satakam [by] Paddikkasuppulavar (1686-1723). Madras:
1913.
The Travels o f Marco Polo. Translated by R o n a l d L a t h a m . Harmond-
sworth, Middlesex: Penguin Books, 1958.
V a i d y a n a t h a n , K.S. ‘Ancient Geographical Divisions of Tamil Nadu.’
Handbook (Kaiyédu), Ilnd International Conference-Seminar o f Tamil
Studies. Madras: International Association for Tamil Studies, 1968.
Pp. 231-7.
------‘The Growth of the Independence o f the Kàdavarâyas.’ Quarterly
Journal o f the Mythic Society, vol. 37 (1947), 125-35.
------‘Mala-Nadu.’ Quarterly Journal o f the Mythic Society (Culture and
Heritage Number), (1956), pp. 225-60.
------‘Munda-Rashtra.’ Ibid., vol. 43 (1952-3), 110-22.
------‘Some Ancient Territorial Divisions.’ Ibid., vol. 40, no. 3 (1950), 81-96.
------‘Pascima-Paka-Rashtra or Mer-Pakkai-Nadu.’ Ibid., vol. 43 (1952-3),
54-65.
------‘The Members of the Ancient South Indian Army (sênai): Their
Assembly and Its Functions.’ Ibid., vol. 32 (1941-2), 127-45, 293-303,
399-407.
V a n L e u r , J.C. Indonesian Trade and Society. New York: Institute of
Pacific Relations, 1955.
V a r a d a c h a r i , K.C. Aspects o f Bhakti. Mysore: University o f Mysore, 1956.
V e d a c h a l a m , S w a m i. vêlâlar nâgarigam [The Civilization of Ancient
Vellalas], Pallavaram: T.M. Press, 1927.
V e n k a t a r a m a A f y a r , C.P. Kamban and His Art. Madras: C. Coomara-
swamy and Sons, 1913.
V e n k a t a r a m a A y y a r , K .R . ‘Medieval Trade, Craft and Merchant Guilds
in South India.' Journal o f Indian History, vol. 25 (1947), 269-80.
V e n k a t a r a m a n , K.R. Hoysalas in Tamil Country (Annamalai University
Historical Series, No. 7). Annamalainagar: Annamalai University,
1950.
------‘More about the Ainnürruvar.’ Proceedings o f Tenth All-India Oriental
Conference, Tirupati, 1940. Madras: 1941. Pp. 349-53.
V e n k a t a r a m a n n a y y a , N. The Early Muslim Expansion in South India.
Madras: University of Madras, 1942.
------The Eastern Câlukyas o f VëngT. Madras: Vedam Venkataraya Sastry,
1950.
------An Essay on the Origin o f the South Indian Temple. Madras: Methodist
Publishing House, 1930.
------Kampiliand Vijayanagar. Madras: Christian Literature, Society, 1929.
------Studies in the History o f the Third Dynasty o f Vijayanagara. Madras:
University of Madras, 1935.
------(ed.). Velugotivdrivamsdvali (Bulletin of the Department of Indian
History and Archaeology, no. 6). Ed. K.A. N i l a k a n t a S a s t r i . Madras:
University of Madras, 1939.
------and M. S o m a s e k h a r a S aR m a. ‘Kakatiyas o f Warangal.’ The Early
History o f the Deccan. 2 vols. Ed. G. Y a z d a n i . Oxford: Oxford
University Press, 1960. Pp. 575-715.
V e n k a t a s u b b a A i y e r , V. ‘Establishment of the Vijayanagara Rule in the
Tamil Country.’ Vijayanagara Sexcentenary Commemoration Volume.
Dharwar: Karnatak Historical Research Society, 1936. Pp. 169-79.
------Kalahasti and Its Inscriptions.’ Quarterly Journal o f the Mythic Society,
vol. 16 (1925-6), 20-4, 213-18.
V e n k a y y a , V . ‘Territorial Divisions of Rajaraja Chola.’ South Indian
Inscriptions [q.v.], vol. 2, pt 5. Pp. 21-9.
V i r a r a g h a v a c h a r y a , T.K.T. History o f Tirupati (The Tiruvengadam
Temple). 2 vols. Tirumali-Tirupati Devasthanams, 1953.
------‘The Srivilliputtur Temple of Sudikkodutta Nachchiyar.’ Tirumalai-
Tirupati Devasthanams M onthly Bulletin, vol. 6, nos. 3-11.
V i t h i a n a n t h a n , S. ‘The Pattuppattu; A Historical, Social, and Linguistic
Study.’ Unpublished Ph.D. thesis, University of London, School of
Oriental and African Studies, 1950.
W h e a t l e y , P a u l . ‘Desultory Remarks on the Ancient History of the Malay
Peninsula.’ Malayan and Indonesian Studies: Essays Presented to Sir
Richard Winstedt. Ed. J o h n B a s t i n . Oxford: Clarendon Press, 1964.
------The Golden Khersonese; Studies in the Historical Geography o f the Malay
Peninsula before a . d . 1500- Kuala Lumpur: University o f Malaya
Press, 1961.
W h i t e h e a d , H e n r y . The Village Gods o f South India. Oxford: Oxford
University Press, 1916.
W i l k s , L t . C o l . M a r k . Historical Sketches o f the South o f India. 2 n d e d .,
2 v o ls . M a d r a s : H ig g in b o th a m a n d C o ., 1869.
W ils o n , H e r b e r t M. Irrigation in India (U.S. House o f Representatives,
The Executive Documents o f the First Session o f the Fifty-Second Con­
gress, 1891-92, vol 18). Washington: Government Printing Office, 1892.
W i l s o n , H o r a c e H . Essays and Lectures in the Religion o f the Hindus. Ed.
R . R o s t . London: Trubner and Co., 1862.
------The Mackenzie Collection; A Descriptive Catalogue o f the Oriental
Manuscripts and Other Articles Illustrative o f the Literature, History,
Statistics and Antiquities o f the South o f India. 2nd ed. Madras: Hig­
ginbotham and Co., 1882.
 W ittfo g e l, K a r l A. Oriental Despotism; A Comparative Study o f Total
Power. New Haven: Yale U n iv e rsity Press, 1957.
W o l f , E r i c R. Peasants (Foundations of Modern Anthropology Series).
Englewood, N.J.: Prentice-Hall, Inc., 1966.
WOLTERS, O .W . Early Indonesian Commerce. A Study in the Origins o f
Srivijaya. Ithaca: Cornell University Press, 1967.
Y a z d a n i , G. The Early History o f the Deccan. 2 vols. Oxford: Oxford
University Press, 1960.
Z im m e r, H. M yths and Symbols in Indian Art and Civilization. New York:
Harper and Brothers, 1962.
Z v e l e b i l , K. The Smile o f Murugan; On Tamil Literature in South India.
Leiden: E.J. Brill, 1973.
------Tamil in 550 a . d . / An Interpretation o f Early Inscriptional Tamil. Prague:
Oriental Institute, 1964.
 Index

Accuta Vikkanta, 78 Andhra plain, 4, 33, 34, 36, 44;

Achyutadevaraya, 486 southern parts of, 57
Achyutaraya, 375, 411 Annamalai, 36, 42
A g a m a : vaikhanasagama, 238; Aparaka, 388
paiicaratra, 238 Apasthamba sutra, 352
Agamic prescriptions, 85 Appadorai A., 248. 249, 258, 262,
Agesthialingam, S., 422 379-80, 419
Agni, 342 Appadurai, Arjun, 54n., 179, 204,
Agraharamt 145 383n„ 428, 432n., 467, 468-9,
Agrarian relations, 16; historically, 47ln., 480, 483n., 482-5
13, 16; between Brahman villages Appar (Tirunavukkarasu), 78, 295
and nattar, 164 Arasanimangalam, 157
Agrarian system, 14, 24, 63-89, 173, Arcot (South and North), 4, 27
254-65; historically, 16; and gar­
Arokiaswami, A., 45, 110, 111, 112,
den cultivation, 26, 27, 28; dry,
113, 115, 199, 304, 311, 314
wet and swidden land in, 26, 28;
Artisan-traders, 104, 171, 175, 252,
and agricultural strategies, 27, 28;
362; k il kalanaigal, 181; kollar,
and technology, 29; and peasants,
188, 250; nattar, 181, 250; taccan,
34; and Brahmin-peasant rela­
181, 250; koliyar, 181, 250
tionship, 63, 69; instability of,
Arulnandi, 235
216
Ashoka, 31, 45
Ahobilam, 239
Asvamëdha, 277, 386, 387, 388
Aiyer, K.V. Subrahmanya, 123, 217,
Âvani. 191, 194, 195, 204
22!, 227, 229, 231, 251, 322, 374
Âyagar, system, 423-4
Ajnapati, 383
Alangildi, 197 Ayyangar, S. Krishnaswami, 367,373
Ayyavali, 248
Alavandar, 233
Alayev, L.B., 226 Baden-Powell, B. H., 13
Alvar, 240 Bahmani state, 381, 392
Amarakosa, 197 Bahmani Sultan Muhammad, 402
Amaram tenure, 375 377, 422-3, 427 Bailey, F. G„ 90, 118, 242n.
Amaramakani, 42 1 Balagey, see Caste
Amaranayaka, 377, 422, 423, 426 Balasubrahmanyam, S. R., 335
Ambuliyar river, 98 Banavâsi, 34
Amman, 237 Banaiôsi-12,000, 108, 128
Am uktam aiyada, 402 Banerjea, J. N., 337
Anaimangalam. 122, 360 Bangalore, 34
Anantanarayana, 360 Baramahal, the, 42
Anbil, 228, 235 Beck, Brenda E. F., 176n., 315, 418,
Andhra-mamlala. 195 458. 467, 480, 482, 484
Bezwadu, 20, 248 distribution of, 305-6, 307;
B hakti, 53, 81. 86; and folk deities, association of Chojas with, 344;
84; ideological components of, 87; not exclusively Brahman settle­
and Sankara, 342 ments, 344; institutions of higher
Bhandarkar, R. G., 30 learning in, 345; functions of,
Bhubaneshwar, 44 349; and inscriptiona! preambles.
Bhuvane Raviran Samarakolahalan, 356
409 Brahnian(s). 4, 9, 11, 28, 46, 48, 50.
Bloch. Marc. 470 51. 63, 183, 186, 198, 229, 404,
Bombay-Karnataka, 43, 57, 59. 61. .441; and non-Brahmans, 5,6, U,
62 49, 212, 231, 239, 240: their in­
Boxer, C. R., 376n. fluence on Dravidian societies, 5 1;
Brahma, 342 and high non-Brahman alliance,
Brahmadeyas. 3, 51, 53, 106, 120. 51; as secular authority, 52, 141,
123, 141-72, 181, 197, 215, 229. 340; and locality power, 53: and
231, 232, 250, 252, 305, 367. 410; high non-Brahman castes, 54;
conditions for existence of, 142; and rural Brahmanical institu­
identified with the plain, 142: and tions and assemblies, 64. 73, 82;
divergent aims o f peasants and their relations with peasants,
Brahmans. 142; and peasant 71, 82; and cultivating groups,
villages, 143; their situation and 72; Brahman-peasant alliance,
characteristic names, 145; super­ 83, 87. 253; and ‘clean’ non-
vision and administration of, Brahmans, 85; and marriage
145-7; sanljarappadi, 148; and system, 102; and gotra 103;
non-Brahman villages, 150; distri­ their secular authority, 141: in
bution of Brahman villages in, relation to nadu and nattar, 143.
150. Map IV-1; and conditions of 150, 163: MaJayali Brahmans.
agriculture. 151: ‘central place’ 161; Vadama Smarta Brahmans,
in, 152. 154, 158; and ritual 210; their monopoly over higher
functions, 152; hagiographieal sacral functions, 211; Vaikanasas,
importance of, 153; educational 211; and false allegations against
functions of, 154; mercantile and other Brahmans. 224; their
artisan groups in, 155; inscrip­ participation in non-Brahman
tions found in, 155; their indebt­ rituals. 230; teachers and
edness to temples, 157: and students, 230; and Sanskrit,
transfer o f rights, 159; change 230-1; and non-Brahman culture.
o f names of, 160; external 231; Aradhya Brahmans. 235;
intervention in, 160; residential and excommunication, 235;
groups in, 162; voluntary basis of, North Indian Brahmans, 322;
166; offices o f mahasabha in, 169; relation of, with Chola rulers,
and cultivation, 169; and 343; as commandants and terri­
relationship between sabhas and torial ‘governors’, 371; Telugu
urs, 170; role of, 170; life-style in, Brahmans, 410; Niyogi, 411:
171; and mahasabhas, 217; ideological role of, 413
Brahman-peasant alliance, 83, 87,
253 Chettis,Nattukkottai, 301
Brahman-sai-Sudra alliance, 50 Chidambaram, 198, 235, 246
Brahmarayan, 114, 343 Chiefs: titles of, 278; power of,
Breckenridge, Carol A., 428n.,432n., 271; o f the Chola state, 283; at
454n., 466, 468 Kodambalur, 303; as hunters and
Briggs, L. P., 336 pastoralists, 304
Brihadesvara temple, 49, 230, 256, Chinese Buddhist pilgrims, 241
341 Chingleput, 26, 27
British: territorial power, 50; land ChittiraméJivinijagar, 220, 222
policies, 53; Brahmans as depend­ Chola(s), 36, 63, 70, 93, 94, 143,
able servants of, 53; institutions, 147, 173, 239, 241, 366; as rulers
204; officials between the 17th C. and kings, 40, 45, 57, 367;
and early 19th C.. 206 overlordship, 34, 40, 57, 98,
Buchanan, Francis, 473 240-1, 252-3; pillaging raids of,
Buddhadatta, 78 40, 41; boundaries o f their macro
Buddhism/Buddhist, 73, 77, 78, 79, region, 57; inscriptions of, 102,
80, 81, 83, 86; shrines, 120; 114, 181; regional names of, (15;
institutions, 208 families o f chiefs in C. times,
Bukka Raya, 202 115, 116; ritual sovereignty, 116;
Bureaucracy, 190 and 550 nadus o f Cholamanda-
Bureaucratic government, 13 lam, 151; army of, 188; and
velaikkarar, 189; regiments of,
Cakravartin, 70. 88 190; peasant military units, 190,
Cambodian kingdom. 37 214; idangai forces of, 191; land-
Campiratayam, 230 tax of. 192; status o f artisans
Cardamom, 36, 42 and merchants, 199, 200, 208;
Castes: principles, 9, 10, II; and urbanization, 204; agricultural
jajmarii system, 10, 177; ‘the expansion, 205; dominant peasant
eighteen castes’, 11; tripartite
groups in C. period, 208; territo­
horizontal division of, 54; o f
rial segmentation o f society in C.
macro region, 55; and nddu, 174,
period, 208; ambiguous position
South Indian associations of,
of non-Brahmans in C. period,
1X0; lower, 204; socio-political 208; expanding agrarian system
associations of, 205; and ‘Sudra’, of, 208; putative centralized state
212; see also Right-hand and structure of, 218; increased
Left-hand divisions o f castes demand from holders o f kilváram
Cattaiavan, 212 in C. period, 226; Saivite and
Chalukyas, 59, 294; o f Badami, 295 Vaishnavite doctrine of, 230;
Chalukya Vikramaditya U, 295 temples of, 232-3, 245; archi­
Champa, 336, tecture of, 246; 2nd major
Chamundi, the goddess, 320 zone o f segmentary state, 297;
Chan Ju-kua, 25! and Pandya country, 299; an­
Chan Ta-kuan, 336 cient tradition o f kingship, 300;
 and intermediate zone, 302; parakesari in, 363; dharmic
distribution o f brahmadeyas and character o f kingship in, 364;
temple inscriptions in C. period, locality segment in, 408; ideolo­
305, 307; peripheral zone o f C. gical role o f Brahmans in, 413
segmentary state, 318; origin Chola, Sundara, 257
myth of, 322; Arya Brahmans, Cholaganga tank, 362
322; three ancient kingships of, Cholamandalam, 91, 92, 94, 95, 97.
323; and cult o f Siva, 323; and 98, 109, 214, 216, 261, 446
Vijayanagara society, 367; king­ Cilappadikaram, 71, 75, 204, 215
ship, 367; land tenure in, 415, 418 Citrameli, 220, 228, 228n.
Chola state, 3, 22, 23, 216, 380; Citrameli-periyanattar, 221, 227,
decline and fall of, 216; historio­ 239, 251, 252
graphy of, 254; self-governing Classical times (Sangam era), 75,
institutions in, 255; army and navy 114, 209; Tamil poems in, 88;
of, 257; administrative structure great families o f chiefs in, 114;
of, 257-64, 340, 358; and agrarian and poets, 209; two disparate
order, 258; revenue terms in, 261; social orders 209; urban centres,
and emergence o f Tondaiman- in, during, 251
dalam, 294; irrigation in, 318; Coedes, G., 336
effective territorial sovereignty Coimbatore, 4, 26, 27, 28, 35, 42,
of, 321; hegemony of, 321; core 43, 54, 58, 126, 216
territory of, 321, 322; alteration Complementary opposition, 274
o f kingship in, 334; king as Conti, Nicolo, 384, 388
divine incarnation in, 334; Conventional numbers, 251; Ganga-
sacral Hindu kingship in, 335; vadi ‘96,000’, 129; ‘108’ families,
funerary temples of, 335, 338, 149; ‘eighteen’, 218-19
339, 364; ritual hegemony of, Copperplates, 131, 132
339, 341, 344; dharmic incor­ Coromandel plain, 27, 31, 32, 33,
poration in, 339; and brahma- 36, 37, 39, 40, 41, 42, 44, 69, 73,
deya, 344; as solar line of 76, 79, 81, 82, 83, 102, 107, 127,
Kshatriyas, 344; and sastric and 141, 143, 172, 310; entrepots, 38
ritual scholars, 345; and salais,
345-6; and m eykklrtti, 350; and Dana-S asana, 46
inscriptions, 355; ritual hegemony Dánda and art ha, 267, 280, 362
in, 356, 357, 358; ideological Dani, A. H., 33
framework and canonical style of Darasuram, 246
workmanship in, 357; eulogistic Dasavanda, 427
reference to royal house in, 358; Day, W. M ., 32
administrative specialists in, 360; Deccan, 30, 31, 32, 35; cultural
scribal function in, 360; ad­ sphere in, 60
ministrative specialists in, 360; Derrett, Duncan M., 278, 317, 318,
pyramidal segmentation of, 361; 319, 406, 407
massing and incorporative process Desai, D., 219
of, 362; titles o f rdjakesari and Descent system, 55
Dëvadana, 53, 111, 120, 158, 16i . Gana (puga), 22
231, 243 Ganapntideva, 59
Devalige ‘70’, 122 Gangas, 108, 310, 316, 317, 318;
Devarâja, 256, 335, model of kingship for Kongu
Deva Raya II, 384, 485 chiefs, 317
Devaram, 311 Gangadikara, 319
Devi Bhagavatam Purana, 390 Gaiigaikondacholapuram, 230, 256,
Devi worship, 237, 238, 239, 245 334, 341, 362
Dharmapuram matha, 236 Gangavadi, 92, 108, 309, 310, 317,
Dharmapuri, 320 318, 319, 393, 447; minor kingly
Dharmasâstra, 23, 301, 245, 267. lineage, 316; central domain of
278, 279, 355, 363 Hoysala state, 318; karmandalam,
Dikshit, Appaya, 201 320
Dikshit, G. S., 122, 129, 219 Gangavadi ‘96, 000’, 108, 126, 128,
Dirks, Nicholas, 296, 383 129, 130
Dorosamudra, 200 Gangetic plain, 31, 32, 37, 48
Dravidian, 5, 7, 35, 36, 51, 64, 100. Ghatikas, 230, 232, 351; at Kanchi.
101 347
Dubois, Abbé J. A., 471 God-king, see Divaraja
Dumont, L., 242n., 267, 390n. Godavari, R., 26. 35
Durga, 238 Gudavari-Kistna basins, 31, 36, 44
Durgadatmaik, 403, 404. 4 11 Gonda, J., 276. 277. 279. 280. 281.
Durkheim, E., 269. 384n„ 390n.
Göpuram, 245, 246
Ellis, Francis W„ 435, 487 Grand An ¡cut, 24
East India Company, 366 Gros, F., 156
‘Eastern Ghat’, 36 Guilds (srem). 22.49. 249: see SreiiT
Edagey, see Caste Guhau, 237
Ennayiram, ¡53, J60, 230 345
Êkabhôgam, 149 Habib, Irfan, 15
Evans-Pritchard, E. E., 270 Hall, D. G. E„ 335
Hall, R. Kenneth, 416
Fallers, L., 8 Hampi, ruins. 390
Festivals, 384-6 Harle. J. C„ 246, 248
Fiiliozat, Vasundara, 373 Harsha, 48
Firth, R., 12 Heesterman, J. C., 277n., 281
Fleet, J. F., 128 Hemmanahalli, 122
‘Folk’ culture, 5; ‘folk’ tradition, 5, Heras, H„ 373
7; folk and high tradition, 6; Hocart, A. M„ 276, 277, 280, 281,
assimilation of folk elements, 384n„ 390, 392, 481
239 Hopkins, E. W., 278, 280, 281, 334
Frykenberg, R. E., 437-8 Hoysalas, 77, 202, 226. 381; their
incursion into the lower Kaveri.
Gajapati inscription, 407 319
Hoysala Ballala II, 319 Iyer, Ananthakrishna, 60
Hsü'an Tsang, 73, 78, 297
Hultzsch, E., 190, 199, 346 Jaina(s), 34, 46, 77, 78, 79, 80, 83,
Huzur Office plates o f Trivandrum, 86, 119, 202, 203, 231, 298, 316;
348 shrines, 120; institutions, 208
Jambukesvara, 246
idangai, 222 Janaka, king of Videha, 348n.
Idangai-edangai, see Right-hand Janapada, 362
and left-hand division o f caste Jatavarman Kulasekara Parakrama,
Idangai-magamai, 202 409
Idangai-vari, 202 Jatis, 249
Idangai-valangai Jatiyar Varalaru. Jyeshtha, 238
184
Ilangovels, 115 Kadamai, 226, 258, 262, 263
Inden, Ronald, 453n. Kadamba Mayursarmam, 347
Indian kingship, 41; sacral character Kadaram, 40
of, 276 Kadavarayans, 115
Indo-Aryan civilization, 37 KSdu, 91
Tndo-Aryan culture, 101, 171 KaikkMar, 186, 199, 478
Indo-Gangetic plain, 31 Kailaspathy, 351, 362
Indonesia, 38 Kakatiya kingdom, 59, 407
Inscriptions, 46, 353, 354; nadu Kalabhra, 76, 77. 78, 81. 84. 88. 295
signatories of, 121; stone, 131; Kalamukhas, 81
copperplate, 131, 132; Sanskrit vs Kalahasti, 127
non-Sanskrit plates, 59, 132; Kalanai, 249
distribution of brahmadeya and Kalinga, 36; see also Orissa
temple, 305; circumambulation Kalittokai, 74
o f land in, 349; and m ey kk u tti, Kallar, 28, 50, 109, 110. 118. 139.
350-1; eulogistic preambles in, 301, 304, 417
355; prototype for the m evkkirtti, Kamma Rashtra, 450
in, 356; standardization of, 356; Kammas, 85
ceremonies recorded in, 358; Kammalars, 138, 199, 201, 227, 248,
circumambulation by female 487
elephant recorded:in, 360; attest­ Kammanacheri, 199,
ing signature in, 361; puranic and Kanchi, 38, 68, 171, 188, 230,
sastric, 362; results o f sacred 241, 247, 351
transformations recorded in, 363; Kandalur, 345, 349
invocation of royal protection in, Kane, P. V., 245
363; Chola epigraphs in, 365 Kanimurrütju, 181, 231, 352
Inscriptional prasasti, 296 Kannada, 34, 60
Inscriptional preambles, 24, 250 Kanya Kumari, 107, 322, 355
Irrigation, 17, 24, 27 Kanyapidariyar, 322
Irrukkuvels, 116, 303, 304 Kapalikas, 81
Itinerant guild. 39 Kapava Nayaka, 381
Kapilesvara (king), 408 Karandai Plates o f Rajendra 1, 343
Kapus, 131, 448, 449 352, 354, 356
Karalar, 320 Kotigu-Chola, 115, 304.
Karashima, N., 151, 260, 262 Kosambi, D. D., 6, 100
Karmandalam, 34 Kosminsky, E.A.. 14
Karmandala-Satakanu 101, J30, 319 Koyil Olugu, 233
321. 446 Kra Isthmus, 38
Karnataka, 4, 28, 32, 33, 34, 35, 43, Kräder, L., 268
58, 61, 88. 92, 108, 110, 111, 126, Krishna, 85
145, 172, 200, 202, 208, 310; Krishnadeva Raya. 76
southern parts of, 57 ; dialect Krishnan. K. G.. 232, 249, 352n..
division in, 60; and ‘eighteen 354n.
Phana Group’, 60,61; as transi­ Krishnaraya, 371, 375, 401
tional zone, 61; inscriptions, 110; Krishnaswami, A., 374. 378, 379.
numerical designations for ter­ 3 9 4 ,410,443,485
ritories, 128; artisans of, 200; and Kroeber, A. L., 6, 8
Nolambavadi, 316; and Bana- Ksaira, 24, 267, 296, 339. 360
vasi, 316 Kshatriyas, 47. 48. 50. 70. 71,
Karve, Irawati, 33 213, 279
Kaveri, R., 24, 26, 32, 36, 42, 49, KudigaU 198
57, 60, 69, 93, 97. 99, 107, Kudimcn, 263, 264
114, 162, 258, 294, 367; oriented Kudiväram, 168
districts of. 61 Kudumbm, 148, 224, 224n
Kaveripatnam, 38 Kulottunga 1, 35. 69. 126. 127. 128.
Kaveripattinam, 68 243, 257, 283
Kerala, 32, 33. 48. 58 Kulottunga III. 115. 246, 247. 283.
KU-kalanai, 198, 315
KTlvaram, 168 Kumbakonam, 243n.
Kingship, sacred, 23, 24 Kula. 22
Kistna, drainage basin o f the, 60 Kulasekhara, 299
Kistna-Godavari, 32, 35, 69; see Kulluka, 279
also Godavari-Kistna Kumara Kampana, 409
Kodambalur, 303, 304
Kolar, 34 Lakshmi, 238
K oliyar, 250 ‘Larger Leiden Plates’, 49. 121. 122.
Kollar-, 250 354, 356 359, 360
Kolleru (Collari) lake, 27 Lingat, R., 267, 278, 279, 280
Kollidam, 93 Literature, classical, 65. 100
Konadu, 302, 3Q4 Longhurst. A. H., 385, 389
Kongu country, 110, 115, 118, Ludden, David, 34n.
137, 138, 216, 309-11, 314-15, Lynch, Owen M., 242n.
485
Kongumandalam, 2 5 1 Machlipatnam. 38
Kongumandal a-Satakam, 316 Macleane. C. D.. 201
MacKenzie, Colin, 395, 472 Merchants. 260, 252, 282
Mackenzie Collection. 379, 441 Meykandar, 235
Macro region; see under South M e y k u tti, 47
India Migration, 368
‘Madhyasta’, 121 Minakshi-Sundaresvara temple. 468
M ad h vast ank aranattan, 121 Minchirai, 349
Madurai, 68, 78," 107,171 Moreland, W.H., 15
Mahadeva temple, 152. 156 M u d a li, 225
Mahajana, 146 M udaliydr, 236
Mahalingam, T.V., 98, 259, 260, Mughals, 14, 15, 16
276. 283, 374, 379 Mukherjee, N., 437-8
Mahanavaml festival, 384, 385, 386, M um m udi, 323
387, 398, 401 Munro, Thomas., 259, 336. 438
Mahanavami Tirtha, 399 Murudaiyaru, 97
Maharashtra, deccan lavas of, 43 Murugan, 84
Mahavamsa, 299 Muslims, 87; conquest in 14th C..
Maha.sa.bha, 82, 146, 149, 158, 161. 59; control over Indian Ocean.
162, 169, 247 39; and Coromandel, 39; Muslim
Mahendravarman I, 78, 80, 295 states o f the Deccan, 372
Malabar Coast, 107 Muttaraiyar, 115
Maldives, 40 Mu-vendar, 45
M alaiyamakkal, 222 Mu-venda-vel&r, 45, 111, 112, 113.
Mandalam, 261, 494 114, 117, 193, 194. 195, 323, 359
Mannargudi, 217, 222, 223, 225. Mysore, 34, 42, 43 ,
227, 228
Mananilanallur, 146 Nachchiyar, 238
Mandapam, 245 Nadars, 205
Manigrdmattar, 38 Nadu, 3, 90-140, 173, 206, 207, 216,
Manimekalai, 215, 276 222, 223, 250, 342; and massing
Mantri, 111, 112, 273 o f primary local segments, 22;
Marai, 235 defined, 90-1; and kurram and
Marathas, 31 kottam , 91; in classical poems, 91;
Maravars, 28, 50, 74, 109, 110, 114. as unit o f local administration,
115, 139, 301, 302, 304 98; as source o f names, 104;
Markanam, 38 and ‘founding’ villages, 105; and
Marriage, 9, 55, 103, 104 classical period, 107; in southern
Marriott, Mckim, 33, 453n. Karnataka, 108, 122; administra­
Marudam, 74 tive functions of, 111; nayagam
Mathas, 142, 195, 200. 229. 230. of, 117; and variation, 117; and
236,' 252 agrarian operations, 118; and
Mayurvarman, 34 vellan-vagai (peasant share
Medhatitfai, 363 villages), 123; and nattar, 124;
Melvaram, 167, 192, 226 ‘central n a d u s 134-5; inter­
Mercantile groups. 18 mediate nadus’, 135-9; peripheral
 nadus, 138, 139; Brahmans in, Nellore, 35, 222, 232
143; as unit o f peasant organi­ Nicholas, R., 178
zation, 167; vertical integration N ilaippadai, 127
of, 207; and vishaya, 218; and Nilgiris, 36, 42
medieval South Indian state, 282; Niyayattars, 186
and pyramidal articulation of Nolambavadi-32,000, 108
segments, 282 Non-Brahman, respectable, 103, 198,
Naduvil-nadu, 91, 92, 95, 97, 98, 231
307 Non-peasant (people), 73, 76; and
Nagapattinam, 81 maravar and eyinar, 75; warriors,
Nagara, 123, 180, 245, 250, 251, 79'
282 ‘Northern Circars’, 43
Nagarattar, 144, 181, 219, 227, Nulambavadi, 126,127
228; linked subordinately to Nuniz, Fernao, 375, 376, 385, 386,
brahmadeya, 251 396, 401
Nagaswamy, R., 156
Naldyira Prabandham, 81 O kkalu, 122, 129, 418
Nallur, 153 Oppert, Gustav, 201, 203, 472
Nammalvar’s Tirumoli, 190 Orissa, 33, 44
NQnadesi (merchants), 180, 227
Nandivarman I, 197 Paes, Domingo, 375, 376, 385, 387,
Nandivarman II Pallavamalla, 80, 396
282, 294, 295, 334, 362 Palar, the, 36, 42, 60
Narasimhachar, R., 373 Palaiyakkarar, 50
N attam akkal, 187, 188, 222 Palasige-12,000, 108
Nattar, 105, 110, 111, 117, 120, 123, Pallamkoyil, 119
124, 131, 167, 181, 206, 207, 214, Pallava(s), 37, 55, 63-5, 68, 70, 81,
218, 223, 239, 282, 352, 358, 361, 82, 84, 88, 97, 118, 150, 202, 208,
412, 416-17, 418; transfer arrange­ 241, 294, 316, 323, 339, 364; and
ments, 119; o f Tondaimandalam, bhakti sects, 64; Prakrit records
125; differentiation of, from other of, 65; military power, 66; society,
important groups, 130; a corporate 67; as southern variant o f Aryan
entity, 131, 229; and royal grants, civilization, 69; and territorial
230 segmentation of, 210; and agres­
Nava-ratri festival, 390 tic labourers, 211; expansion o f
Navaka, 369, 379, 396, 406, 407, 408, wet rice cultivation, 211; and
409, 410, 415, 440-1, 442, 443; forest dwellers, 211; and pumi-
kings of Madurai, 392; warriors, pputirar, 212; and village arti­
415 sans and service groups, 212;^
N dyaka system, 375, 396, 397-400, overlordship in Tondaimandalam,
434, 440 294; and Jainism, 295; inscription,
Nayanars, 240 296; conversion o f many macro
Hayankara system, 370, 376 regions,450
Nayar ‘Kshatriyahood’, 48 Palli, 188
Palliccanda,, 181, 231 212; support o f Brahman ritual
Pallippadais, 338 and learning, 239
Paini, 36* 42 Peasant societies, 2, 7, 8, 10, 11,
PcMchanamuvaru, 200 15, 90; concept of, 12, 14; and
Panchatantra, agamic, 5 culture, 55; defined, 67-8; expan­
Panchayattars, 138 sion of, 83
Panchan Banajigas, 479 Peasantry : and Brahmans, 7, 69,
Pandimandalam, 92, 137, 251 72-3, 82-3, 86-7; and the external
Pandimandala-satakam, 446 world, 18, 20; armed, 214; and
Pandya(s), 36, 77, 84, 93, 297-9; support o f non-Vedic ritual
regal tradition of, 297; warriors, and learning, 240;. in Tondai-
traders and rulers of, 298-9; mandalam and Naduvil-nadu,
supreme power in southern penin­ 305; Palli, 306; and the Gangas
sula, 300; o f peripheral tracts, and Gangadikara, 318; and
301 Karalar, 320
Paraiyar, 209 Penner, R., 36, 60, 93, 97, 99, 316
Parantaka I, 146, 283 Penugonda, 248
Pasupatas, 81 Periyanadu, see Supra-local assem­
Pasupata mat ha, 343 blies
Pasupata Saivism, 342 Periyanâttar, 221, 222, 225, 227, 252
Patikam (poetic epilogue), 350 282; the sub-culture of, 229-41;
Patirruppattu, 350 mediators between Brahmans and
P attak karar, 315, 418 peasantry, 225
Pattakkaran, 137, 315 Periyapuranam, 341
Pattinam, 242, 251 Pillai, Krishnaswami, 369
Pattinayar, 250 Pillay, K. K., 247
P attupattu, 55 Piranmalai, 227, 228, 234, 251
Peasant, passim; localities, 8, 10, 19, Place, Lionel, 435
22; household, 11, 16, 20; and Pocock, David F., 242n.
English peasantry, 11; relations ‘Poligar’, 50
with land, 11; agrarian relations, Polonnaruva; 188, 199
16; economics, 17; family, 18, 19; Ponnaiyar, 36, 42, 305, 306
village, 19; movement, 19, 310; Ponnani (R.), 42
community, 20; conservativeness, Prasasti, 47
20; external and internal pres­ Pre-Pallavan era, 64
sures on, 21; ecotypes, 25, 26; Puduchcheri, 38
expansion o f peoples, 35; hierar- Pümiputtirar, 187
chial ordering of, 71; and non­ Purtugonda, 200
peasant folk, 69, 75, 79, 81, Puram, 242, 251, 362
82, 83, 84; and religious sects, Puranânüru, 91, 298
79-81, 84-7; relations with Puri, 44
Brahmans, 82; village assembly
(ur), 82; and Brahman-peasant Raghavan, V., 6n.
alliance, 82-3, 86-7; absconding, Rajadhiraja II, 283
Rajamanickam, 236, 237 and trade activities, 179; ‘relative’
Rajaraja I, 35, 98, 99, 116, 125, 151, and ‘potential’ groupings of, 180-
155, 188, 207, 214, 249, 256, 263, 1; South Indian caste associations
310, 321, 340; armies of, 174 of, 180; division of, 182; relative
Rajaraja II, 126 work with, 213; lack of
Rajaraja III, 218, 223, 237 emphasis upon stratified relations
Rajaraja Chola, 39, 46, 57, 33-4, 380 among, 184; and shared insignia
Rajarajesvara temple and shrine, and symbols, 185; and occu­
155, 338, 364 pational interests, 185-6; and
Rajaraja-vijayam, 338 Brahmans, 186,206; and military
Râjaràjësvara-nâtakam 338 activities, 188-9, 204, 206; and
Rajendra I, 126,’ 188, 207, 256, 283, mercantile and craft occupations,
299 196,206; and possible explanation
Rajendra Chola I, 39, 40, 46, 125, o f these designations, 201-2;
126, 256, 283, 334 groupings, 205; and artisan-
Ramanuja, 233, 246 traders, 206, 215; gaps in
Ramayana, 334, 387, 389 evidence about, 207; origins of,
Ramesvara temple, 468 207-15; and ‘stranger’ versus
Ramnad, Sethupati of, 302 ‘indigenous’ classification, 209;
Rangacharya, V, 306 and left-hand caste, 249
Rao, M. Rama, 389n. Rose, Kathi L., 454n.
Rashtrakuta, 294
Sabha, 47, 48, 49, 121, 141, 144,
Rathakära inscriptions, 196-8, 215
166, 412
Razzak, Abdur, 384, 386
Reddi, Subha, 473 Sacrifice (yajna), 363
Sacral kingship, 280-1
Reddis, 85, 213, 448, 449; Telugu
Reddis, 417 Sahlins, Marshall, 417
Redfield, R., 6 Saiva, 212
Saiva Siddhanta, 7, 235, 236, 239
Revenue, 13; land, 14
240, 341
Rice, Lewis, 128, 473
Saivite, 78, 79
Right-hand and left-hand divisions
o f castes, 54, 61, 72, 85, 104, 173- Saivaite Panchalars, 479
215, 282,485; dual divisions in, Salai, 346-7, 349, 351
61, 185, 368; ‘Nine P h a m s' and Salem, 4, 27, 28, 35, 42, 43, 54, 58,
Eighteen Phanas of, 61; balagey 216
and edagey, 174, in classical Saletore, B. A., 367, 373
poetry, 175; self-organized, 176; Saliyar, 250
and assimilation o f new peoples, Saluva Narasimha, 395, 401
176, 182-3, 206; and idangai, 176 Sambandar, 80
(in Tamil country), 248; coloni­ Sampradaya, 185, 230
zation of, 177, 204; as a corporate Samudragupta, 41, 48
institution, 177; conflictful rela­ Sankaracharya, 337
tions between, 177; and Chola Sahkarappadiyar, 249, 250
inscriptions regarding, 179, 181; Sangam, 33, 34
Sangam era, see Classical times hierarchial power structure in,
Sankarappddi, 156, 162 272; administrative staff of, 273;
Sankarappâdiyâr, 198 conception of, 274; ritual
Sanskrit: language and ideas, 51 ; incorporation of, 276; and fragile
institutions of, and non-Brahman character o f medieval Indian
learning, 52; inscriptions, 65; states, 281; and Tamil plain, 305;
knowledge, 240; ghatikas and and Tondaimandalam, 305; as an
salais, 345, 346 inferred structure, 321
Sanskritization, process of, 66 Sendamangalam, 235
Santana, 236 Sethupati, 302
Santâmamudlaiyârs, 239 setft, 103
Sapta-matrikas, 238 ‘Seven Mothers’, 238
Sarma, P. Sree Rama, 373n., 374n., Sewell, Robert, 366
394 Shanmugam, S. V., 422n.
Sastri, Krishna, 190, 249 Share (pangu) rights, 137
Sastri, K. A. Nilakanta, 66n., 77, Shiyali Taluk, 38
98, 106, 111, 112, 113, 147, 165, Siddhesvaramatha, 44
166, 170, 189, 190, 191, 192, 193, Simhavishnu, 63
196, 229, 245, 253, 255, 256, 260, Singer, Milton, 470n.
263, 276, 367, 369, 372, 373, 374, Sirkar, D. C„ 353, 374, 376, 377
380, 406, 419, 421, 443, 466, 467; Sitaraman, B., 260, 262
centralist-bureaucratic bias of, Sivaji, 392
_ 283 Sivalinga, 337
Satapatha Brâhmana, 387 Sivaramamurti, C., 338
Satakam, 444, 445-8 Smriti, 259
Satavahanas, 31, 37 Sonnerat, Pierre, 471
Sattanur, 158, 162 South Arcot, 176, ?16, 236
Sattra, 280 South India, 32-4; physical elements
Segmentary state, 3, 8, 21, 22, 23, of, 35-7, 41-4; political, cultural,
44, 63, 274, 280, 282, 362, 366, and social elements of, 44-62,
367; pyramidal segmentation of, 64; macro region of, 48, 49,
21; ‘ritual sovereignty’ of, 35, 366; fragmented character of,
46, 273; dharmic kingship in, 173; South Indian Caste
45; and chieftains, 45; massing Association in, 180; social status
o f localities in, 90; characteristics o f non-Brahmans in, 207; seg­
of, 265; ritual hegemony in, 266; mentary society in, 223; and
dual territorial sovereignty in, temples, 246; progressive urbani­
266, 268; nadus- in, 270; opposing zation in, 479
elements of, 271; mHvendavejar South Indian macro region, 48, 49,
in, 271; transformation of, 271; 54, 55, 173, 366
central and peripheral relation­ Southall, Aidan, 23, 265, 321, 340
ship in, 272; distinction bet­ Spate, O. H. K., 33
ween pyramidal structure and Spencer, W. George, 4in ., 164, 338
Sreriî, see Guilds Tamil inscriptions, 4, 34, 65
Sri Lanka, 38, 39, 40, 188 Tamil-Brahmi inscriptions, 78
Srikakulam, 44 Tamilaham, 107 ,
Srinivasachari, C. S., 186, 472 Taniyur, 152, 170
Srirangam, 81, 233, 239, 487 ■ Tanjavur, 27, 37,. 40, 46, 50, 54.
Srivaishnavas, 85 236, 241, 249, 258, 341
Srivijaya, 40 Tanküra, 152, 170
Srivilliputtur, 171 Tanks, 24-5
Stein, Burton, 429 Tattanur, 359
Stone inscriptions, 132 Taylor, William, 391
Subbarayalu, Y., 91, 92, 93, 94, 95, Technology, 24
97, 98, 99, 123, 419 Telikis, 200, 248
Subramaniam, T.N ., 124, 253, 345, Telingana, 58-9; Kakatiya kingdom
346, 348, 355, 356 of, 59
Subramanian, N., 209, 240n. Temples, 46, 229, 252, 455; and
Suchindram, 161, 247 locality authorities, 229; Hindu,
Sudra, 85, 212, 233 231; aniconic propensities of,
Sundaram, K., 248 232; early Chola, 232; Brah-
Supra-Iocal assemblies, 216-41; manical, 233; and Sudras and
their functions, 217-23; and non-Brahman leaders, 233; and
itinerant merchant groups, 223, supralocal assemblies, 234; for
228; resources of, 225; as chief female deity, 237-8; urban
supporters of temples, 226, 234; characteristics of, 243; Vedic, 244;
and cultural development, 229-41; Brihadesvara, 245, 337; architec­
their support to Brahman institu­ tural changes and enlargement of,
tions, 230 in Chola period, 244-8; and urban
Suresh, B„ 102, 103, 114, 115 settlements, 246; and urban popu­
lation, 246; and urbanization, 248;
Tagadur, 320 and worship o f Vedic Shiva, 323;
Tambraparni, 36 and aniconic form o f worship,
Tamil, 88; plain, 33, 34, 323; funerary, 335; strategic con­
305; poetry, 33, 88; civilization, siderations for site of, 334; Bri­
35; classical works, 55; five-fold hadesvara sivalinga in, 337; and
division, o f ancient speakers, 56; pallippadai or samadhi shrine,
literature, chronology of, 56; 337; and the sacred Ganga, 341;
country, 57; speakers, poetic situ­ and moral order in South Indian
ations of, 100; Sanskrit sutra of society, 454; analytical informa­
raurava agama into, 231; hymns, tion about, 20-35; temples of
232; as a vehicle o f teaching, Tamilnadu, 456-63 (tables); of
235 Aditya I, 364; o f devi worship,
Tamil country, 57; and zone of 465
Telugu speakers, 394; and Vijaya- Tenali, 27
nagara leadership, 396, 400 Territorial segmentation, 71, 74, 99,
 102, 104; and marriage and kin­ Town-life, 241-2; rise of, during
ship networks, 101; tripartite divi­ 12th Cent, 250; and trade, 252
sions of, 103 Tribhuvani, 158, 230, 345
Tevaram, 81 Tribhuvanam, 246
Thomas, P., 337. Tripurantaka, 338
Thorner, Daniel. 14 Tungabhadra, R., 316, 367, 381
Three kingships mu-vendar, 45 Turner, Victor, 203n.
Tinai, 55, 100 UJavars, 71, 74
Tiruchchattimurram, 236 Umapathi, 235
Tiruchchengattangudi. 236 Ur, 82, 121, 166, 282
Tiruchirapalli, 233 Urar, 144
Tirukkalukkunram, 247 Urban milieu, 241-53; also see
Tirukamakottam, 238 Town-life
Tirukkoyilur. 187. 220, 234. 486 Urban-rural dichotomy, 6
Tirumangai. 81 Urrattur, 182, 184, 202, 204
Tirumalai, 246, 247 Uttaramerur, 146, 147, 149, 150,
Tirumukkudal, 345 153, 156, 157, 160, 161, 162, 163,
Tirunavukkarasar, 236 165, 168,
Tirupati, 246, 395, 428 Uyyakondan-Udaiyar, 197
Tiruppalatturai, 236
Tiruvanaikkaval, 246 Vadavanpatti, 258
Tiruvannamalai, 486 Vadugan, 50, 189, 395
Tiruvidaimarudur, 236 Vaigai, 36
Tiruvadutturai, 158. 230, 263, 345 Vaikanasa, 121, 157, 211, 238
Tiruvalangadu, 161, 162 Valanadu, 306
Tiruvalangadu Plates o f Rajendra Valarpuram, 235
C h o la l, 122, 132, 354. 360, 362 Vaishnava Sampradaya, Sri, 211
Tiruvallam, 188 Vaishnavites, 78, 79
Tiruvanaikkaval, 237 Vaishnavism, 233
Tirutturaipundi, 237 Valangai-idangai, see Right-hand
Tiruvenkadu, 188 and left-hand divisions o f castes
Tiruvorriyur, 230, 247, 342 V alahgai velaikkara, 124
Titles (administrative), 111-14, 117 Vanniyar, 445
Tolkappiyam , 107 Varahanadi, 97
Tondanur, 122 Varanjuram, 183, 187, 204, 222
Tondaimandalam, 82, 88, 91, 92, Variyam. 146, 166, 168
109, 125* 'l 27, 137, 161, 171, 188, Varna, 70, 103
191, 193, 204, 214, 216, 268, 295. Velaikkarar, 189-90, 209, 214
307, 309, 319, 394, 447, 487; Vellalas, 50, 84, 85, 103, 131,187,
Jayarigondacholamandalam, 125; 188, 199, 213, 224, 231, 235, 236.
as a ‘48,000 country’, 127, 129; 239, 304, 448-9, 487; of Tondai­
and peasant settlements, 210; mandalam. 237,417; of Choliyar,
chiefs of, 297 417
Vellans, 250 481; diversity o f people in, 393;
Veljan-vagai, 123 Teiugu speakers and warriors in
Veliaru, 93, 97, 304 Tamil country, 394-6, 406-15;
Vembarur, 174 military components and organi­
Venadu, 45, 58, 351 zation in, 396-405; temples and
Vengadam Hiils, 107 fortresses in, 404-5; land system
Vengi, 59, 200, 393 and tenurial forms in, 415-39;
Vengur, 187 ‘village’ in, 417, 419, 420; temples
Vennir Vellala, 350 and irrigation in, 427-8; temples
Venkayya, V., 190, 214 and sects in, 432; 439-41; trans­
Venkataramanayya, N., 367, 373, formation o f nadu leadership into
375, 376, 378, 379, 409, 410, 411 chiefs in, 443; and mandalam,
Vettiperu, 231 444-5; and satakam, 444-8; and
Vidyaranya, 411 Kapus, 449; temples and society
Vijayamangalam, 235 in, 455-6; devi worship in, 465;
Vijayanagara State and society: Shiva and Amma shrines in, com­
nature o f political system, 366, pared, 465-6; mathas in, 466-7;
367, 380-1, 396-7, 409, 414; and Vaishnavism in, 468; Right and
new elements in Tamil society, left-hand division o f castes in,
366; comparisons and contrasts 469-87; urbanization in, 481; and
with Chola society, 367-9; nature royal intervention, 485, 488
o f kingship, 367, 383-4, 386-91; Vira Saivite, 60, 235, 240
and religious institutions, 367-8; Virarajendra Chola, 322, 355
migration and military conquests, Virapratapa, 485
368, 380-2, 395; martial charac­ Visavadi, 436
ter of, 369, 370, 383, 400-401; Vishakapatnam, 43
and ‘feudalism’, 369, 374-9; and Vishayattar, 220, 227
nayankara system, 369-71, 376, Vokkaligas, 131,3 1 9 ,4 4 8 ,4 4 9
394, 397-400, 405-10, 415, 433, Vriddhachalam, 486
440-1, 442-3; viewed as military Vydparin, 198, 250
state, 370, 372, 400; political
and military roles o f Brahmans Warriors, 12
in, 370-1, 410-14, 441-2; political Warriors o f South India; peasant
systems of, 371-405; and Deccan origin of, 71; Teiugu, 87
Muslim power, 372; founding dy­ ‘Western ghats, 36
nasty and. origin of, 373-4; ex­ Wodeyars o f Mysore, 392
pansion o f V. overlordship, 380-
2, 394; expansion of, 382; festi­ Yadavas, 59
vals in, 384-90; Muslim contin­ Yajnavalkyasmriti, 101, 353
gents in V. army, 392, 400, 402-3; Yajur Veda, 5, 352
and dharmic ideology, 392, 400, Yudhisthira, 387