ft ew ye

DS >, aed
7 ©
*> Ay
y
‘foe
°.

ry
Sie SOUTH INDIA UNDERR

Bae iheCholasdominau cheiil Vsdiae

| ee political scene‘from the ninth to the
thirteenth ¢centuries: ‘This book is an
authoritative and comprehensive study of
Chola state,
s society,pand economy.
a ae —¢

iiss - Southfiader theCholas explores the

fo state—society interactions in early medieval
‘south India. Tepresents an in-depth analysis -
5 of Tamil epigraphy. Using inscriptional
2 = ! ‘evidence from India and South-east Asia,
it analyses thesocio- economic milieu,
~merchant guilds,
g and.other sociological
See
Ba
ts

z aspects. gs OE ET

, ‘Subbarayalu ino. the revenue

sy stem, property rights, relations between
mt landowners, cultivators, and slaves and.
che structure ‘and character of the Chats
“state. He scrutinizes insdet:ail the ev olution
of organizations like Urar, Nattar, and
Periyanattar, social-classes like the left and
right hand divisions, and the merchant
militia: Bor the first. time an attempt is
~made here to quantify the revenue of a pre-
| Mughal Indian state.

=>. Based on a wealth of primary souitces

examined overa period: of thirty years,
this book will be indispensable to.schol ars,
“résearchers..and students ok south India and
eatly medieval history.
 Contents

List ofIllustrations
List of Tables and Appendices
Foreword by Noboru Karashima
Preface
Acknowledgements
List ofAbbreviations
Note on Transliteration

Introduction

SECTION I: EPIGRAPHY AND HISTORY

1. Tamil Epigraphy: Past and Present
2. The Social and Economic Milieu of the Pulangurichi
Inscriptions
3. The Merchant Guild Inscription at Barus in Sumatra
4. Sociological Aspects of the Personal Names and Titles
5. Interpreting Inscriptional Terminology

SECTION II: STATE AND SOCIETY

6. Land under Chola Rule: Measurement, Classification,
and Assessment
. An Overview of the Chola Revenue System
. Quantifying Land Revenue of the Chola State 100
. Sale Deeds and Property Rights 116
ook
= . Urar, Nattar, and Periyanattar 124
11. Alunganam and Milaparushai: Two Early Village Bodies 138
 vi Contents

12. Landholders, Cultivators, and Slaves 146

_ 28. Social Change and the Right and Left Hand Divisions 167
AB. Afjuvannam: A Maritime Trade Guild of Medieval Times 176
NG. Erivira-pattinam, Warriors, and the State 188
16. The Chola State 207
17. Characterizing the Chola State 248

Index 261
 CHAPTER 16

The Chola State*

A. Nilakanta Sastri’s The Colas still continues to be the most

uthoritative work on the Chéla dynastic history even after six
decades of its first publication. The learned scholar’s mastery of the
inscriptions which constitute the bulk of the primary sources for the
Chéla history was so thorough that it is difficult to disagree with him
on the interpretation of the inscriptional records generally. Any further
studies must certainly be based on this pioneer and classic study.
From the 1960s there has been again a marked revival of interest in
the Chdla studies. Besides Burton Stein who initiated a critical review
of K.A. Nilakanta Sastri’s work, Noboru Karashima, G.W. Spencer,
and a host of other scholars have contributed to further clarification
of different aspects of the Chola history, concentrating on matters
other than the dynastic history.” Certainly, there has occurred much
refinement in the methodology of studying the old records using new
conceptual tools. The present chapter reconsiders the Chola state in
the light of all the accumulated acknowledge.
The Political Process
The 400 odd years of the Chéla rule (ca. 850 CE to 1279 Cr) may be
divided into four analytical periods on the basis of certain prominent
political and social trends.’ Period 1 from ca. 850 CE to 985 CE was
the formative period. Period 2, 986-1070, comprising the reigns of
Rajaraja I, his son and grandsons, was the period of expansion and
consolidation. Period 3, 1071-1178, saw some break in the genealogy
due to the commencement of the rule of the Chalukya-Chdla line
and it was marked by some shrinkage in the territorial extent and the
gradual revival of locality chiefs. Period 4, 1179-1279, was noted

*This essay has been revised and adapted from “The Cola State’, Studies in History,
4 (2), 1982.
208 South India under the Cholas

for the full emergence of the chiefly families and the disintegration
and ultimate demise of the Chéla state. No information is available
to trace the antecedents of the Chéla dynasty which established its
rule in the Tanjavur delta about 850 Ce after wresting the area from
a Muttaraiyar chief. The Chélas may have been just one of the many
chiefly families that dotted the country in the ninth century and
earlier under the aegis of the Pallava rule. Of course, the dynasty
claimed descent from the Chola line mentioned in the earliest Tamil
literature some ten centuries earlier. By the early tenth century, the
Chélas became a powerful ruling dynasty with the exit of the Pallavas
who had been ruling the northern districts of Tamil Nadu till then.’
Very soon after this they acquired a mythical genealogy to connect
them with the Puranic traditions of northern India.’ In keeping
with this claim they encouraged vigorously Brahmana settlements
(brahmadeya) and canonized temples, mostly Saivite, of a pan-Indian
character.® During the four centuries of its existence the Chola state
was ruled over by twenty kings, excluding a few other members
who were just co-rulers. Except two or three minor breaks, the line
had a continuous, stable existence. Even the so-called Chalukya-
Chéla line that started in 1070 cE with Kuldttunga I, a great-
grandson of Rajaraja I through his daughter, produced only an
imperceptible break.
The Cholas started their career of expansion from what was called
the Chéla-nadu, their core territory, comprising the fertile Kaveri
delta. Within a few decades of its rise the Chola kingdom annexed the
Tondai-nadu, the core of the Pallava territory in northern Tamil Nadu.
But at its height of victory, about the middle of the tenth century,
it fell a prey to the onslaughts of the Rashtrakiita king, a powerful
neighbour on its north-west, and was temporarily crippled. After
restoration two decades later, the military expansion of the Chélas
faced no formidable rivals to the south of the Tungabhadra (Krishna)
river. Thereafter, throughout the eleventh century, wars became
big organized affairs. The arena of war covered a vast area from the
Krishna to Kanyakumari and even beyond the seas. The statement
by the reputed historian of the Chélas that ‘the whole country south
of the Tungabhadra was united and held as one state for a period of
two centuries and more’ may not be literally correct, but there was an
attempt by Rajaraja I to visualize such a thing.’
During Period 1 the Chéla king was not the sole ruler of the
country. There were a number of chiefs of small, ancient lineages
 The Chola State 209

maintaining some sort of subordinate relationship with the superior

Chola monarch. Those chiefs had their own territories clearly
demarcated and named after their family names like Vanak6-padi,
Mala-nadu, and such others.* Some of them ruled only small areas
covering one or two nddus (about 100 sq. km.). But some had
vast areas extending over two or more modern taluks. The chiefly
families not only had matrimonial relations among themselves but
also provided prospective queens to the Chola royal family many
a time. For all practical purposes the chiefs were masters in their
respective localities. They had their little courts, officers, and army.
They enjoyed fiscal powers and patronized local (canonized) temples
with large money and land gifts. They may be said to be subordinate
to the Chola king in the sense that they acknowledged the latter’s
overlordship by issuing their inscriptions in his name and executing
orders, if any, received from him. During wars they sent their armies
to assist the Chola overlord and some of them played personally a
leading role in the battlefield as commanders.
Throughout the eleventh century, however, inscriptions are
bewilderingly silent about the chiefs. Very few of the older chiefs
continue to be active in their territories. This is the century of imperial
expansion initiated by Rajaraja I (985-1014 ce) who was both a great
warrior and a great statesman. Very ingeniously he incorporated all
the chiefly territories in his new ‘valanadu’ scheme. The chiefs who
were deprived of their territories seem to have been recruited to the
bureaucracy elaborated by the same king. Also, a countrywide land
survey was for the first time undertaken in this reign.’ Both Rajaraja
I and his son Rajéndra I made Saivism as a sort of state religion and
encouraged a number of Saivite teachers to migrate from the north
and settle in the Tamil country. Rajéndra I (1012-1044 ce) advanced
his father’s imperial designs further by carrying arms across the seas
against the Sri Vijaya kingdom of Southeast Asia. Maritime trade
interests with China and Southeast Asia might have been a motivating
factor for this overseas expedition. Administratively speaking, his
reign is noted for the establishment of viceroyalties in the conquered
territories under the charge of his sons: Chdla~-Ganga in southern
Karnataka, Chéla-Pandya in the Pandya country in south Tamil
Nadu, and Chéla-Lankésvara in northern Sri Lanka.'® The period
covering the reigns of Rajéndra I and his three sons (1012~70 ce)
seems to have been a crucial time in the restructuring of society. The
emergence of the dual social divisions called Right and Left Hands
~
210 South India under the Cholas

should be attributed to the many wars of this phase which politicized

many erstwhile tribal communities which had remained till then
outside the plains society.!' We also notice conspicuous growth
of local and overseas trade activities carried out by the Ayyavole
merchant guild and its associates concomitant with the growth of the
Chola power and its territorial expansion.
Kuldttunga I (1070-1120), in spite of his great martial and
administrative capabilities, could not arrest the shrinkage of his
overlordly influence outside the central area during the later part
of his long reign. This happened due to an equally great warrior-
king, Vikramaditya VI, occupying the western Chalukya throne
on the north-west and due to the rise of the Pandyas on the south.
The Chola-Pandya viceroyalty is not heard of after 1070 ce. But
Kuldttunga tried to control the Pandya area by stationing military
outposts on important highways there. Hereafter, the Chola state
covered only the areas of Chéla-mandalam and Tondai-mandalam.
Even within this area, locality chiefs gradually started to gain upper
hand. Some of these chiefs, as their names and territories indicate,
claimed descent from some previous lines that had existed in Period
1. Others were obviously from the new lines. They were found almost
everywhere except in the core area comprising the Tanjavur District
and a portion of Tiruchirappalli District.'*
The common point relating to all these chiefs is that they were
generally in charge of the padikaval or watchmanship of their
respective localities, ranging from one village to several nadus.'*
For discharging that duty they were permitted to enjoy sizeable
revenue by way of the major tax called also padikdval levied on the
land produce and some minor levies on the local artisans. From
the nonchalant way the chiefs disposed of the revenue from these
taxes, it may be said that they were virtually independent rulers
of their localities, which they called their ‘nadu’, and royal orders
seldom penetrated into the borders of these chiefly domains." It
is also clear from some explicit evidence that most of these later-
day chiefs started as warriors playing a major role in the Chola
army of the eleventh century.'* Actually most of them and their
once martial clans became part of the settled communities only
during the eleventh century and after. The prominent warriors were
encouraged, it seems, by the Chola king, to become the ‘protectors
of their localities and were given considerable fiscal rights in the
areas under their jurisdiction.
 The Chola State 211

There were occasions when the bigger and powerful chiefs

aggrandized themselves at the cost of the smaller chiefs. Attempts
were made to avert such conflicts by political compacts (nilaimai-
tittu in Tamil) by the neighbouring chiefs by affirming fealty towards
each other and giving assurance for the continuation of the status quo
undisturbed.'® These political pacts clearly underline the fact that
these chiefs had their own army units under some captains called
agambadi-mudali. By the latter half of the twelfth century the Chéla
king had come to depend more and more on the armies of the chiefs.'”
The chiefs positioned on the borders were vacillating and equivocal in
their loyalties towards the Chola overlord. This became quite visible
when the Pandya king became powerful again so as to challenge the
Chola supremacy in the early decades of the thirteenth century. By
the middle of that century the Chéla rule was confined to its central
parts by the full emergence of powerful and independent chiefs like
the Kadava, the Bana and the Telugu Chéla. Not only this, there
cropped up the overbearing influence of the Hoysalas of Karnataka
who came first to assist the Chéla king Rajaraja II] when he was
taken war captive by the Kadava chief in 1230. Within a few decades
the Hoysalas established their own parallel rule in the Tiruchirappalli
area. The last Chola king Rajéndra III (1246-79) could retain only
the Kaveri delta under his nominal rule until his last year.
The King
In the Chola state, like in any other contemporary state, the king was
the central figure. A careful reading of the inscriptions points to the
fact that as the status of the king gradually rose from that of an ordinary
chief in the ninth century to the positionofan emperor in the eleventh
century, the usé of grandiose titles also became frequent. Whenever
his subjects referred to the king in the earlier decades they used the
designation peruman or peruman-adigal, adigal being an honorific
suffix. The basic form peruman (variant of perumagan) means ‘the big
son’ or ‘great man’. From the middle of the tenth century the social
distance between the king and his subjects began to increase as may be
observed from the many attributes given to the king.
~The first such term is the designation udaiyar meaning ‘our
possessor’ or ‘our lord’. After 1100, this form was elaborated into
ulagudaiya-perumal ot ulagudaiya-nayanar meaning ‘the lord of the
world’. In the beginning the title £6 meaning ‘king’ was prefixed to
the king’s coronation name (like ko-parakésari) by the royal scribes.
212 South India under the Cholas

From the reign of Rajéndra I the title wdaiyar was thus prefixed. The
title chakravartti (emperor) came into use only from the reign of
Virarajéndra (1063-70 ce) and the more grandiloquent sribhuvana-
chakravartti (emperor of the three worlds) was adopted in the reign
of Kuléttunga I (1070-1120 ce). As for the names of the king
mentioned in inscriptions, they are only coronation names with
the suffix deva, for example, Rajaraja-déva. Déva (meaning god)
obviously suggests some divine attribute. An interesting feature of
the titles used for the kings is that many of them (except chakravartt
and tribhuvanachakravatti) had striking parallels in the titles used
for contemporary deities. Sometimes the king is even treated as the
comrade (talan) of some famous deity."
The numerous titles of the king, partly in Tamil and partly in
Sanskrit, were sometimes given as names to jewels presented to the
temples, to fields and places, to large territories and, above all, they
were prefixed to the titles of important persons and officials (as noted
in the previous section). That is, the king was always the fountain
of honour. . van
The Chola kings were all polygamous.'? More is heard of about the
queens of early kings than those of the later kings in inscriptions. The
early kings also maintained concubines quite openly. Some of the early
queens came from humble families and some were originally temple
girls. But from the beginning, taking wives from chiefly families was
very common. Thus, Uttama-chéla (970—5) had his queens from the
families of the Irungdlar, Viluppu, Miladu, and Paluvédu chiefs.”
These alliances must have figured high in the considerations for the
consolidation of Chéla rule in the early stages. Inscriptions from the
time of Rajaraja I are silent about the parentage of the queens.
The princes, like the queens, are prominent only in pre-Rajaraja
inscriptions figuring as donors. They were then referred to by the
honorific pillaiyar. Eleventh-century inscriptions refer to them
indirectly in the royal eulogies. Otherwise they remained anonymous
until they got a chance to rule. Succession to the throne was by
primogeniture. Deviation from this norm was rare and succession
disputes were minimal.
The coronation ceremony (called abhishéka) was celebrated by
brahmanic rituals.?' It seems the milaparudai of some Brahmana
villages took a leading role in the ceremonies and received land
gifts as abisheka-dakshinai.”* The coronation ceremony must have
been the most important occasion for the Brahmana scholars to
 The Chola State 213

compose and shower profusely titles (4iruda) in Sanskrit upon the

new king. |
_ The insignificant position assigned to the princes and queens in
inscriptions of Period 2 and later may partly be due to the glorified
image of the king systematically enhanced through panegyrics. In
the meykkirtti (eulogy) section of Rajaraja I’s inscriptions, the king's
achievements alone were praised. In the succeeding reigns, the entire
royal family was given encomiums but there too the king towers above
all. Only from the reign of Virarajéndra, the queen was given some
space in the royal eulogy. The king’s possession of the entire earth
and his heroic feats (personified as goddesses of earth and victory,
respectively) were stressed in the first part of the eulogies and brute
force was graciously acknowledged. Sanskrit eulogies (prasasti) found
in the copper-plate charters followed the somewhat ancient, ‘astra
tradition. Their aim was to praise the entire Chdla line, claiming
puranic antiquity for the dynasty tracing its origin from the sun.
Territory
Traditionally, the territory of the Chdla lineage had been called
Chéla-nadu (Ch6nadu being a variant). In the ninth century this
was confined to a narrow but fertile area, namely the Kaveri delta
from Tiruchirappalli on the west to the sea on the east. There existed
in the Tamil Country at this time other large politically defined
territories like the Tondai-nadu and Pandi-nadu, besides a number
of chiefly territories like Mala-nadu, Vanak6o-padi, and so on.” The
Chéla-nadu was the core territory of the Chola state when it was
established in the middle of the ninth century. And it continued to
be the most fertile and stable part of the state's territory to the end of
its existence whereas areas outside its limits were added to and lost in
the intervening years. For a major part of its existence, however, the
Chola state had sovereign rights over what was called Tondai-nadu
first and Tondai-mandalam or Jayangondachdla-mandalam later,
covering the northern part of Tamil Nadu. Only during the eleventh
century the southern part of Tamil Nadu, that is, the Pandya country
proper, and the southern districts of Karnataka were under the direct
control of the Chéla government. These territories were treated as
somewhat independent provinces. Actually, in the case of the Pandya
area it was administered under the viceroyalty of a Chola prince, with
the designation Chéla-Pandya, almost independent of the Chola
government for about five decades from 1020. The other territories
214 South India under the Cholas

beyond the above-mentioned area, such as the Andhra districts up

to the Krishna and northern Karnataka and Kerala, never became
integral parts of the Chola State. They were just tributary territories.
With the expansion of the Chola state some structural change
took place in its territory. In the early phase (from 850 to 1000) the
territory included in its fold many distinct chiefly territories besides
the core Chéla-nadu. Only in the reign of Rajaraja I, a clear attempt
was made to reorganize the territory so as to fully incorporate the
different chiefly domain within the territory of the growing Chola
state. Rajaraja I renamed all the conquered territories after his own
titles. The small chiefly territories were given new names which
had the suffix ‘valanadu’ for example, Mala-nadu alias Rajasraya-
valanadu, K6-nadu alias Kéralantaka-valanadu, and the like.”
Thus, the introduction of the valanadu nomenclature was a clever
idea of Rajaraja I to consolidate the conquered territories. This is
clearly attested to by the fact that the renaming coincided with the
disappearance of the chiefs from their original territories, which was
noted above. The valanddu set-up was introduced into the conquered
Pandya territory in the reign of his son, Rajéndra I. Only in the reign
of Kulottunga I this was extended to Tondai-mandalam and southern
Karnataka. Another innovation made by Rajaraja I for integrating
the territories was the concept of ‘mandalam’. The traditional
politico-geographical names of large territories were replaced by new
names taken again from Rajaraja’s many titles. Thus, Pandi-nadu,
the Pandya country, was called Rajaraja-mandalam,” Tondai-nadu
was called Jayahgonda-céla-mandalam, and so on. And Chéla-nadu
itself was called Chola~-mandalam.”° These were not just new names
as far as the Tamil country is concerned, rather these mandalams
incorporated the adjoining chiefly territories besides, of course,
their core territory (see Map 16.1).”” For example, Chola-mandalam
included the Chéla-nadu (the core territory) and the adjoining
Mala-nadu, Ko-nadu, Irungéla-padi, and such others. Thus, the
introduction of the mandalam nomenclature entailed significant
reorganization of the entire territory.
A seemingly stable feature of the Chola territory was the nadu,
which was the basic territorial unit in a major part of south India at
this time. Though the term nddu was also used in its general sense to
denote the larger political territories (like Chéla-nadu), it was more
specifically used to denote distinct agricultural micro-regions. Most
of the nddu names had the suffix nadu, some the suffix karram, and
 78°E 81°E
7S°E

12°N 12°N

oN
9°N

ON
ON

78°E SI°E
TS°E

Map 16.1: The Mandalams of the Chola State

Courtesy of the author
216 South India under the Cholas

a few had other suffixes like kulakkil, érikil and nilai. Unlike the
valanadu and mandalam names very few nddu names show royal
imprint in the Chola country. To judge from the meagre evidence
before the Chéla times it is clear that some of the nddus of the
Chéla times had a high antiquity. The nuclei of some other nadus
also might have formed long before the Chola rule.” An analysis
of the names, the topography and other particulars of the nddus
suggests that the na@du was basically a cluster or grouping of peasant
settlements formed about the nucleus of a common irrigation source
like a channel or tank and strengthened perhaps by kinship ties in
the initial stages.” A fair estimate gives a total of about 350 nadus for
the Chéla- and Tondai-mandalams. Excepting the extreme cases, the
nadus generally ranged in extent from 20 to 200 square kilometres.
It seems that in the drier areas the area of the nadu was larger than
that in the fertile areas. Nadus in fertile tracts, though smaller in
area, had denser settlements whereas those in the dry tracts had fewer
settlements. The valandadu set-up, as it was superimposed on the pre-
existing nddus, disturbed to some extent the territorial integrity of
the latter. Being a deliberate royal contrivance, the valanddus took for
their boundaries, wherever possible, some natural water channel. So
while a nadu lay on both the banks of such watercourse, a valanadu
lay on only one bank. Consequently, many a @du in the channel-rich
delta area, due to their lay on both the banks, came to be partitioned
among two different, adjoining valanddus.
The territorial element kottam over and above the nadu was peculiar
to Tondai-mandalam. This was a survival from the preceding Pallava
period. The number of nadus included by the different kéttams does
not show any uniform pattern while one big kdttam had 15 ndadus,
five had two or one each. About 150 nddus are found distributed
among 23 kottams. The introduction of the valanddu scheme into
Tondai-mandalam by Kulottunga I did not involve any structural
change as the kéttams seem to have been just renamed as valanddus.
The basic settlement comprising the habitation site and the
surrounding agricultural lands as well as forests, and hilly and pasture
lands was called a. The most common and undifferentiated type of
village was called vellanvagai, literally the ‘agricultural kind’, subject to
customary taxes and enjoying customary rights and privileges. In many
of these villages there were separate chéri or quarters for the paraiya
(the community of agricultural labour) and in some, separate quarters
for artisans (kanmafa) and for coconut gardeners cum toddy-tappers
 The Chola State 217

(ilava).*° In one or two villages there were separate quarters for the
washermen and water distributors (ta/aivdy). Unlike the vellanvagai
type, the eleemosynary and newly created commercial villages were
distinguished by the conferment of special names as well as special
privileges. Thus the brahmadeéya which was permitted full or partial tax
exemption was given a new name consisting of two components, the
first one being a title or name of the donor (usually a king or queen)
and the second being either mangalam or chaturvédimangalam, for
example, Kundavai-chaturvédimangalam.*' The name of dévadana
or temple village was characterized by the suffix-component nallir.
The settlement which was primarily mercantile was called nagaram
and distinguished by the name suffix puram. The villages of the
eleemosynary and other types, on their creation, became to a large
extent independent of the nadu in which they were a part earlier. Out
of about 1,300 villages, some 250 may be recognized as brahmadéya
villages, about fifty dévadana, about twenty-five nagaram, and the rest
as the general or vellanvagai villages.
It is now generally accepted that most of the 6rahmadeéyas in Tamil
Nadu, particularly those of the Chéla-mandalam had become well
established by the tenth century, mostly due to the interest shown
by the Pallava and Pandya rulers, if we go by the place names of the
brahmadeéyas.** The Chola dynasty might have added some new ones
in the beginning of its rule, but mostly it seems to have elaborated
the already existing ones. Of course, there are a few new gigantic
ventures, like the one created by the Karandai grant of Rajéndra |
in ca. 1020 in favour of 1080 Brahmana families invited from more
than a hundred old settlements spread over Tamil Nadu, a majority,
however, coming from within Chéla-mandalam.”
The creation of the brahmadéya while entailing reorganization
of the existing local irrigation network (like the criss-cross pattern
of feeder and drainage canals) also helped in making new canals to
bring fallow land under cultivation to support the large immigrant
population. For maintenance of the irrigation works, separate lands
were provided in each village under the name vettappéru (tenure for
the verti). There is clear evidence linking vettappéru lands to talaivay-
chanrar, that is, those in charge of maintenance of head sluices.”
References to such vettappéru lands are found more frequently in
pre-eleventh century inscriptions, both in Brahmana and non-
Brahmana villages. By the tenth century, the irrigation system in
the Kavéri delta seems to have become stabilized and thereafter only
 218 South India under the Cholas

the maintenance of the system was the major concern of the various
village communities.*° No new major irrigation works are referred to
in the eleventh century and after.
Some brahmadeya villages even attained a fully independent status,
denoted by the term tankaru or more popularly taniyir, ‘independent
village’. In the eleventh and twelfth centuries there were a number
of such prominent taniyirs. Generally, the brahmadeya villages had
attached to them many hamlets which were inhabited by the tenant
cultivators (ulukudi) of the Brahmana landholders. In the case of
some taniyurs in Chola-mandalam, their hamlets were grouped to
form new nddus. These new nadus which were named after kings had
the distinguishing suffix component périlamai-nadu (for example,
Kidarangonda-périlamai-nadu).*” Périlamai-nadu means the nddu of
the périlamai. The ulukudi or the tenant-cultivators were otherwise
called périlamai. These périlamai-nddus are peculiar to the Chdla-
mandalam, even though taniyaér is found in Tondai-mandalam in
good number.** From the royal titles in the first component of their
names, it may be suggested that this new territorial arrangement
around the taniyiir brahmadéaya was actually a royal creation, like
the valanadu; and the same titles would suggest that these périlamai-
nadus had their advent in the reign of Rajéndra I (1012-44).

Socio-economic Organization
The society of our period was more or less stratified. Though references
to this fact are sporadic in inscriptions the existence of castes with
certain hierarchy among them can be understood from the available
evidence. As may be seen from the following particulars any e!
the so-called castes started rather_as occupational groups. i
groups. Another important feature to be stressed is that though the
~ inferred hierarchy may apply as a broad outline to the entire Chola
period, the details changed over time. Caste-consciousness became
pronounced and explicit during the latter half, that is, the twelfth
and thirteenth centuries.*? Many of the tribal communities which
had been on the fringes of the plains society until the tenth century
began to be incorporated into the latter as new castes, such as the
palli, churutimdan, nattaman, and Srigdpala, in the eleventh century
and after.
Members of about sixteen castes figure in various inscriptions as
individual donors or in some other capacity. They are: brahmana,
vellala, manradi, idaiya, ‘merchant’ caste, palli, vétkd, tachchan,
 The Chola State 219

kanmalan, tattdan, kaikkola, ilamagan, kalla, kuchava, valaiya, and

véettuva.”° Ifwe take into account the temporal and spatial distribution
of these donors and also the quantum ofthe donations, the Brahmana
_and Vellala castes seem to be the two top influential castes over allthe
“area and throughout the Chola rule. fom Brahmanas enjoyed _
“a better position in the earlier half than in the later one,.Next to
‘these “two ‘castes the palfiis found in aprominent situation,-but-this
‘caste appeared only in the eleventh century and after and moreover
"it was confined to the South Arcot District and the adjoining areas.
“The herding caste (manradi and idaiya) and the merchant caste.may
Sicslead
be tethe next position. The potters and the other artisan
castes came next. Kaikkola, the future weaver caste, does not figure
in the above consideration because the caste members rarely appear
as individual donors, though we come across the kaikkdla soldiers
now and then.
Brahmanas occupied the first rank in the hierarchy due to their
ritual
superiority added to their landed. power..Wherever. the jatis
_were ranked specifically, the Brahmanas are found to be accorded
the first place.*! They were called perun-kudi, ‘the greater kudt’,
whereas all others were just ‘kudi’. The Vellalas, by virtue of their
land power, closely followed the Brahmanas. and numerically
‘they were more powerful tha
than_ the Brahmanas.. Uhe..merchant
caste was almost equal
in rank to the Vellala and it consisted of a
group
of sub-castes like chankarappadi (or sankarappadi), chaliya,
pattinavar, and the like, some of whom were themselves primary
producers besides being engaged in exchange trade. The strength
of the merchant groups, however, cannot be compared with that
“of the Vellala as they were confined only to mercantile settlements
“(nagaram) which were much smaller in number. Of course, they
“were in a good number in a few big political and urban centres
like Tanjavur, Gangaikondachola-puram, Kanchipuram, and such
others. The herding caste and artisan. castes.were.considered-as the
servicing groups, denoted by the term falanai or panimakkal.”
However, the herding groups occupied a better position than that of
the artisans. The paratya, the leather workers, and the hunters were
‘given the lowest place. The paraiya being the caste of agricultural
‘labour must have been quite widespread in the wetland. plains.
“However, with little property of their own and put under constraints
of segregation, these last groups of people never figure as donors
in temples.
 220 South India under the Cholas

A striking feature of the society was the existence of the caste-_

pra-village
basedcorporate bodies that functioned at village and supr
“Tevels.* Thus the urar, thesabha, and the nagarattar were the primary
corporatee bodies the ndttdar(or-ndttavar).at the —
nadu level. The arar was thecorporate body of Vellala landholders**.
ai)
“in general (vellanvagai ) villages, the sabha was the corporate body of ie
‘Brahmana landholders in Brahmana villages, and the nagaratiar was~
“that of the merchants cum landholders in the nagaram (mercantile)
‘villages. The nattar was a supra--village level body of landholders
representing mainly the general kind of villages and played a
fines role in the administration of the nddu localities vis-a-vis the
ing’s government.
The ar and nagaram bodies were constitutionally of a simplekind
whereas the sabha, particularly in
i the case of some bigbrahmadéyass
was an elaborate body associated with a number of administrative,
committees called variyam.~ Moreover, it had a éénstitutional
procedure influenced by the precepts of the dharmasastras.* In
certain Brahmana villages, two other corporate bodies functioned,
namely the milaparudai and the alunganam. They seem to have been
archaic bodies surviving in certain settlements even after the sabha
became more popular.*”
The economic status (that is, the landholding) was an important
criterion
for becoming a member of thése corporate bodies: This is
explicitly indicated for the sabha, where some additional educational
“qualification was also needed in some villages. In the case of non-__
Brahmana villages the landholding status of the members can be
understood on circumstantial evidence. In the Brahmana villages, the _
tenant-cultivators, known as ulukudi or périlamai, clearly constituteda
“fair proportion of the local society from the beginning as the Brahmana _
landholders did not cultivate their lands themselves. Subsequently in
—
i

number. Ethnically icoreeciliiiohan were closer to the landowning

non-Brahmanas though occupying a lower rank. Lower still than the
cultivators was the large number of paraiya and perhaps some Vellala
labourers too. These lowest groups were actually considered as slaves
(adimai) to the landholders and cultivators, both Brahmana and non-
Brahmana. Lacking ownership of any means of production, these
slave labourers were at the complete mercy of their possessors.
From the. beginning, individual ownership..was. encouraged in
Brahmana villages.“* On the other hand, ownership was communal
 The Chola State 221

in Vellala villages _un

until the tenth century... Thereafter, there is..a
lear change in i the _agrarian
a system, which is characterized. by the
breakdown of the communal landholding and. the emergence of
‘a
complex tenure system. The teasons for the growth of private
“landholding were many, such as increase of service tenures, import of
“much wealth into the Country due to military activities of the Chola _
_Kings, Targ émples, conferment of the " padikaval cor
~watchmanship rights on local chiefs, and other such reasons. The
net effect was the emergence offeudal tendencies with the rise of big
landholders, secular as well as religious. This situation is reflected in
the frequent use of the terms ani and parru. Though the kami right
may not be absolute ownership in property in the modern sense, it
came very close to that. The holders of the kami right (called kaniydlar)
could enjoy a privileged position in the society of the day.*’ Some of
them were big landlords owning each a big area, even whole villages.
An interesting development of this phase is that even in brahmadéyas
non-Brahmana ami-holders are prominently seen along with the
Brahmanas by the late tweifth century.’ This is in stark contrast to
the times of Rajaraja I in the early eleventh century when only some
service-tenure holdings of non-Brahmanas were permitted within the
Brahmana settlements.”
Temple villages provided much scope for the development of
a complex tenurial system. By the twelfth century, temples had
accumulated large extent of lands making the presiding deities of those
temples as ani holders. The kani of gods (tirunamattu-kani) could be
easily manipulated to become kani of big persons of the concerned
localities with the tacit approval of the king. And a part of the superior
right in these kami was also assigned by the king to the army people,
officials, and other retainers under various names with the suffix
component parru such as padai-parru, jivita-parru, vanniya-parru,
and the like. The parru-holders (called parralar), enjoyed generally
superior rights over and above the kami-holders and as representatives
of the government they were generally disliked by the latter.*’ Naturally,
conflicts and tension would result from this developing agrarian
complexity. Sometimes the kings interfered to resolve the conflicts by
controlling the aggressive kani- and parru-holders.”
Title-holders
The Chéla royal inscriptions give the impression that the elite of
the society were distinguished.by..appropriate titles..The titles may
222 South India under the Cholas

be broadly classified as titles based on caste or profession and those

relating to araiyanlrajan (chiefly) families. These two categories may
be further subdivided according to their prefixing royal titles or
territorial titles or both. The most prominent title was mavendavelan;
three persons out of every hundred of the Chola period held this
title. In Period 2 the percentage increased to nine. The title was
always found in a compound form, that is, suffixed to either the
king’s title (for example, Rajaraja-miavéndavélan) or to the name
of a nddu (Inga-nattu-mivéndavelan) or to both (for example,
Rajéndra-tiruvindalar-nattu-mivéndavélan). Nearly 94 per cent of
miivéndavéelans had the king’s title.” About 10 per cent are found
to be in the pattern of ‘king’s title + nadu + mivéndavélan’. The
connection between the king’s title and mivéndavélan is very striking.
Miavéendavélan may mean the ‘vélan of the three kings’. Vélan stands
for a member of the Vellala caste. The three kings were perhaps the
three traditional Tamil kings, Chéla, Pandya, and Chera. Most of
the mivéndavélans were also village-udaiyan. So mivéndavélans were
only leading Vellala landholders. From the available evidence it seems
that this title came into vogue only in the reign of Parantaka I (907-
55 cE). It may be suggested that Parantaka I, who was the first Chola
ruler who definitely attempted to conquer the entire Tamil country,
first introduced this title to win over the confidence of the traditional
landholding groups of the whole territory.
Of the remaining, the group ending in the component araiyan or
rajan deserves our attention. Araiyan and its hierarchical variants, adi-
araiyanladhi-rajan, pér-araiyan and marayan!maharajan (all meaning
‘the big araiyan), are found in a variety of compound titles, and
account for about thirteen per cent of the inscriptional population.
The general pattern of the araiyan titles may be represented as: king’s
title + caste/profession/chiefly family + araiyan. The titles given to
Brahmanas belong to this group; they were in the pattern ‘king’s title
+ brahma + araiyan’, for example, Rajéndracdla-brahma-marayan.
The Brahmana titles were actually the most numerous of this group.
Accomplished artistes or other professionals patronized by the palace
were conferred titles in the above pattern: for example, a dance master
would get the title Rajaraja-nritta-péraraiyan.
Then we have a number of chiefly (or pseudo-chiefly) titles like
pallavaraiyan, ganga-araiyan, and so on. Two of these chiefly titles,
namely vilupparaiyan and pallavaraiyan, and the brahma-marayan
titles are found to behave in identical manner. These three titles
 The Chola State 223

Table 16.1: Period-wise Distribution of Titles

No Title Period
oe ee
1. Udaiyan/kilan, and so on 49 293 («496 34
2. Miavéndavélan/vélan 14 205 33 6
3. Bhatta 6 35 3 ]
4. Brahma-rajan/Brahma-maharajan,
and so on 4 36 7 1
5. Marayan (excepting titles in no. 4) l 13 3 -
6. Pallavarayan and Vilupparaiyan 6 Thin, 23 15
7. Other araiyan titles 2 $2. 4B 4, fae
8. Nadu-mivéndavélan, and so on 3 23 6 3
9. Prefixing king’s titles > 7: ee b6
Total number of office holders 72 579 213 106
Source: Compiled by the author

accounted for about 40 per cent of the araiyan titles for all the periods
and in Period 2 for 70 per cent (See Table 16.1). They are mostly
appended to king’s titles and most of the holders of these titles held
good positions in the government (Table 16.2). So these titles seem
to have been conferred on officers of the administration by the king.
Taken as a whole the araiyan titles increased with time and showed
a much larger increase in Period 4. Many of these title-holders in
Periods 3 and 4 may be considered as chiefs in the sense that they
were, as holders of the padikaval rights, rulers of some territory,
though in many cases it was not bigger than a village or two. It may
be pertinent to add here that the prominent araiyan title-holders
were known by the general designation ‘rajakulavar (those of the
ruling clans) who enjoyed various service tenures during the twelfth
and thirteenth centuries and who very often created problems to the
landholders (kaniyalar), thereby attracting king’s admonitions.
Apart from the mavéndavélan and brahma-marayan titles based
on caste there were many other titles which were held by the people
of different castes who had some claim to a dignified position. The
term kon was associated with the herding communities, that is, the
idaiya and manradi. The pattern of these titles was ‘king's title + kan’
or ‘king’s title + nadu + kon’. As for the merchant communities the
suffixing terms were mayilatti and palan (Jaya-palan, Naya-palan)
besides chetti. The elite from the artisan castes were honoured by a
general method where the name of the particular caste suffixed to the
224 South India under the Cholas

attribute peru (‘great’ or ‘big’) was given as title in the pattern ‘king's
title + peru + caste’. For instance, in the case of a tachchan (carpenter)
it was Rajaraja-peru-tachchan.
Titles based on the territory n@du form a good proportion of the
official and non-official titles. The prominent combinations, some
of which have already been referred to above, are nadu-alvan, nadu-
velan, nadu-mivendavielan, nadu-kon, nadu-araiyan, nadu-bhatta,
nadu-vilupparaiyan, and ndadu-pallavaraiyan. Nadu-alvan which
became prominent from the middle of the eleventh century is found
associated with the padikaval system. There were also some officers
with this title suffixed to the king’s title Alvan means a ruler and is
equivalent to araiyan in the general sense. Nadu-bhatta being a temple
functionary stands apart from the present group. The other titles, with
or without the king’s title, were mostly titles of personages or officers.
It can be demonstrated from a few verifiable instances that the nadu
title was based upon the respective territorial unit nd@du where the
person hailed from.”
Officialdom
The king’s government had in its service a class of officials whose
putative functions have to be ascertained mostly from the names of
offices as generally there is no direct information on the actual work
they did. Sometimes even the names of offices have to be ascertained
indirectly. To start-with, there were only a few offices in the first
century of the Chola rule, namely adikari, srikariyam, and nadu-vagai.
In the reign of Uttama (970-86 ce), the puravu department, the d/ai
and the naduvirukkai appeared for the first time. In the next reign,
the conspicuous additions were the superior military offices, sénapati
and dandanayagam. Thereafter, there was some minor diversification
and from the reign of Kulottunga I a gradual decrease in the varieties
(see Table 16.2).
The term adikari (Sanskrit adhikadri) means one who wields
adikaram or authority. The verbal phrase adikaram-cheykira (‘one
who performs the adikdram’), which is used alternatively, makes this
meaning clear. In the context where this phrase occurs,” the officer is
referred to as ‘one who wields the authority on behalf of the king’. That
is, the adikari was, first of all, the servant of the king. The alternative
phrase nam-karumam-araykira®*® (one who looks after ‘our business’)
as used by the king to denote the adikari strengthens this idea. The
number of adikaris figuring in royal records may be taken as an
 The Chola State 225

Table 16.2: Correlation of Offices and Titles (All Periods Together)

No Name of No. of Udaiyar Muvenda- Brahma- Araiyan Nadu King’s
Office Offices velan rayan group group titles
1. Adikari 227 89 125 8 50 7 183
2. Sénapati 24 4 - 6 11 2 16
3. Dandanayagam 15 3 4 2 1] ] 12
4. Olai 85 24 27 10 29 7 48
5. Naduvirukkai 29 = - 27 ~ - l
6. Srikariyam 56 26 9 8 7 5 14
7. Naduvagai 18 l 2 4 6
8a Puravuvari-1 159 91 55 l 10 5 45
8b Puravuvari-2 126 90 17 l 7 2 19
8c Puravuvari-3 62 33 l - 3 l 4
Source: Compiled by the author

indication as to the strength of that office under each king. During

Period 1, there were a few adikaris and from the reign of Rajaraja I the
number increased and went up to about 100 in the reign of Rajéndra
II (1052-64 ce). The heyday of the adikaris was the eleventh century,
that is, Period 2. In the reign of Kuldttunga I (1070-1120 ce), the
number was considerably less and later the name itself disappeared.
The adikaris were divided into two sections, one called udankittam
and the other vidaiyil. The udankittam officers seem to have been the
courtiers, always accompanying the king” and the latter may be the
touring officers, executing the king’s commands (vidaz). The titles of
the adikdris are clearly suggestive of their importance. Nearly 80 per
cent of them had the king’s titles prefixed to their individual titles and
50 per cent of them were holders of mivéndavelan titles.
Tirumandira-6lai or simply 6lai was the royal scribe, personal
secretary, so to say, who committed to writing the oral orders
(mandiram) of the king on the spot. This office appeared at the end
of the first period (tenth century) and thereafter continued to exist
throughout the Chéla rule. The number of holders of this office
was fairly evenly distributed through the different reigns. Up to the
reign of Vikrama (1118-33 ce) there were some superior officers
called tirumandira-olai-nayagam, over and above the tirumandira-
dlai, who scrutinized the jottings of the latter before passing on the
same for execution. While the holders of mavéndavelan title were
dominant in the d/ai offices, those of the brahma title were so in the
superior 0/ai section.
226 South India under the Cholas

The office called naduvirukkai is always found mentioned along

with the adikari in royal records and the holders of this office were
all bhattas, that is, learned Brahmanas. This office was also peculiar to
Period 2. It may have been a judicial office.” The sénapati, the topmost
military office, existed only from the latter part of period 1 to the
beginning of Period 3. The brahma title-holders were proportionately
large in this office and mivéndavélan was conspicuously absent. The
second-ranking military office dandandyagam was found prominently
in Period 2 and less conspicuously in the subsequent periods.
The revenue department called puravuvari took shape in the latter
part of Period 1 and existed to the end of the ChOla rule. First, it was
simply called puravuvari, then puravuvari-tinaikkalam and still later
puravuvari-srikaranam. The latter two terms, obviously synonyms,
meant the ‘accounts department of puravuvari’.. There existed in
Period 1 some office called tinai to look after the revenue accounts.°!
It was just a small and simple office. Its successor, the puravuvari, on
the other hand, became elaborated and subdivided into some sections
from the reign of Uttama. The general hierarchy of those sections,”
which prevailed through the reign of Virarajéndra, as inferred
from the consistent order of their occurrence in royal inscriptions,
may be put as follows: puravuvari-tinaikkala-kankani, puravuvari-
tinatkkalam, or puravuvari-tinaikkala-ndyagam, varippottagam,
mugavetti, varippottaga-kanakku, variyilidu, and pattolai.
From the reign of Kuldttunga I, only a shorter version of this
hierarchy is mentioned: puravuvari-tinaikkala-ndyagam or its
synonym puravuvari-srikarana-nayagam and mugavetti. In many
cases of this phase, all the offices were lumped together in the general
term variyilar (meaning those of the vari department). The above
hierarchy is supported by the corresponding titles of the different
offices. The mivéndavelan titles as well as the associated king’s titles
are found to decrease gradually from the top section to the bottom
one. A striking feature is the almost non-existence of the brahma
titles among the officers of this department.
The office varikku-kiru-cheyvar meaning ‘those who make
settlement for taxes’ is found only in later inscriptions (Periods 3
and 4) along with the variyilar. Though the name would suggest its
treatment as a section of the puravuvari department, its contextual
position actually suggests otherwise. The office is mentioned in the
context where the term adikari is found in Period 2. This coupled
with the other fact noted above that the term adikari disappeared
 The Chola State 227

subsequent to Period 2 would suggest that the varikku-kiiru-cheyvar

was identical with or at least similar to the adikari.
At the nddu level there were some lower-level executive officers. Of
them, the prominent one was the nadu-vagai that was functioning
from the early tenth century to the end of the eleventh century.
Particularly, this office was very active during the reign of Rajaraja
I and Rajéndra I. The phrase nadu-vagai-cheykinra which was used
to denote the function of this office meant ‘one who made (tax)
settlement of the na@du’. Generally, a nadu-vagai officer was in charge
of a single ma@du. A few were in charge of more than one. The rarely
mentioned offices nddu-kankdni-ndyagam and nddu-kiiru may be
superior to but related to the na@du-vagai. After Kulottunga I, nddu-
level offices seem to have disappeared.
The srikariyam (literally the ‘sacred duty’) office was denoted
more by the verbal adjective srikariyam-cheykira (who performs the
Srikariyam) than by the noun form srikariyam. The holder of this
office was the manager of the temple on behalf of the government.
This office existed almost throughout the Chéla rule. Some of the
Srikariyam officers held this office only for particular occasions,

But many of them were permanent holders of this office having their
residence near the particular temples that were in their charge. Brisk
activity of this office is found only in the reigns of Uttama, Rajaraja
I, and Rajéndra I, that is, from ca. 970 cE to 1050 ce. Further, only
prominent temples had this officer. Among the titles of the srikariyam
officers, muvéndavélan was prominent and next only to that was the
Brahmana title bhatta.
Even a cursory checking of the contextual positions of the
above offices reveals the existence of a sort of hierarchy among the
different offices. This impression is corroborated by the evidence
relating to the vertical mobility of the different office-holders. About
twenty-five instances of vertical mobility are available for a corpus
of 1,400 officials. Most of them are found within the puravuvari
department. There are, thus, instances where the mugavetti in a
lower (eighth) position went up to become purvuvari-tinaikkala-
nayagam (third position). In a few instances, the top officer in the
puravuvari department became the adikari. There are cases where
the dandanayagam became the séndpati or adikari. The two terms,
perundaram (bigger grade) and chirudanam (smaller grade), which
were in vogue in the two earlier periods, particularly in Period 2,
228 South India under the Cholas

are supposed to denote the existence of a gradation of offices into

upper and lower categories.“ This is rather doubtful. Both these
attributes are found associated with military offices like sénapati
and dandandyagam and the military regiments. It may, therefore, be
suggested that the two terms were used only for a sort of grading the
military offices and not civil offices. The term panimagan, ‘servant’,
was used only for a brief while, in Period 2, to denote in general
terms all the lower ranking officers.
All the officers, big and small, were only the king’s servants doing his
business (7aja-kariyam). Their general designation koyirramar, meaning
‘the relatives of the palace’ or more exactly the ‘men of the palace’,
reveals the close personal relations of the officials with the king.” Later
variants of the term kéyirramar are komurravar, komarravar, and so on.
During the latter half, from about 1150, the (superior) officials were
generally called mudali (‘foremost person’) which is a general term
of honour rather than the name of an office. Further, these later-day
officers were given the honorific titles pillaiyar (‘son’) which was used in
Periods 1 and 2 for the princes only. It has been shown in the foregoing
account that there is a good correlation between offices and titles.
That is, a high office is correlated to a high-sounding title with a
prefixing king’s title. So the implication is that the title was conferred
on the officer by the king. The actual conferment of such a title is rarely
referred to in inscriptions directly.® This can, however, be demonstrated
by some circumstantial evidence. The king being the bestower of
honour, the titles used to change with the accession of a new king.
There are some instances of this kind among some prominent
officers. For example, the title of a sémapati, Kuravan Ulagalandan,
was changed from Rajaraja-maharajan held in the reign of Rajaraja
I to Rajéndra-chéla-jayamuri-nadalvan in the reign of his son and
successor Rajéndra I.°”
Military
The Chola rule was marked throughout by constant and endemic
warfare. The eulogies of the kings beginning from the reign of
Rajaraja I are mainly glorifications of their achievements on the
battlefields. As to the actual organization of the military there are
only stray pieces of information available in inscriptions. From the
names of the army regiments referred to now and then we may make
some inference about the mode of recruitment and the weaponry of
the particular unit.
 The Chola State 229

In the earlier half, the kaikkdla regiments were prominent. They

continued to exist in the latter half also but not so prominently. Each
being named after a title of the ruling king, these regiments were
called as ‘select’ units indicated by the verbal adjective terinda, for
example, Parantaka-terinda-kaikkolar, ‘the select kaikkdlas called
Parantaka’. The term kaikkéla was the name of a prominent weaving
caste in the post-Chdla centuries. But during the Chdla period there
is practically no evidence or clue to associate this term with the
weaving profession. The caste must have started as a military group
in the Chola period and transformed itself into a weaving caste in the
later centuries.
Another name which figures prominently in the Chola army
was the vélaikkdra. But this seems to have appeared only from the
reign of Rajaraja I and lasted up to the end of Period 3. The ‘right
hand’ vélaikkdra units of both the higher and lower grades are
found mentioned in the Tanjavur inscriptions of Rajaraja I. There
are two diametrically opposite views regarding the character of the
vélaikkara units based upon the same etymological interpretation.
The first view holds that they were mercenary troops recruited for
the occasion (vé/ai). The other view holds that these were the ‘most
permanent and dependable troops in the royal service ... ever ready
to defend the king and his cause sacrificing their lives when occasion
(vélai) arose’. For the mercenary character, the Polonnaruwa Tamil
inscription of the Sri Lanka king Vijayabahu dated 1110 CE is usually
cited.” This inscription refers to the vélaikkdra army as including a
number of subsections like valangat, idangai, chirudanam, and so on,
which were all obviously modelled on the Chola army and recruited
from the Tamil country. It has also been suggested that the big
mercantile bodies like the valafjiyar may have supplied the recruits
of this army to the Sri Lankan princes.’' This suggestion is debatable.
It is however possible that the Sri Lankan vélaikkara army may be
mercenary in character. But this does not support the view that the
Chéla army units were also mercenary in character. The presence
of so many ‘select’ vélaikkara regiments, preceding the Sri Lankan
evidence, would support the opposite argument. And we have
evidence from the twelfth and early thirteenth century inscriptions
for the chivalrous tradition of the vé/aikkara soldiers who pledged
their very lives for the safety of their masters.”
Among the other units, mention may be made of the parivarattar
(or parivaram) which consisted of guards attached to the palace. The
230 South India under the Cholas

regiment called viracola-anukkar (or vira-anukkar) seems to have

been connected with the protection of temples. A few regiments
were recruited from the neighbouring non-Tamil regions (Kannada
and Malayalam) and named after the respective regions, for example,
Kannadaka-kaduttalai (the Strong Heads of Karnadaka), Malaiyan-
orraichchévagar (the Malayala infantry). The infantry seems to have
formed a large chunk of the Chdla army. Among them were found
archers (villigal), sword-bearers (valilar) and spearmen (kondavar).
The term niyayattar (or niyayam) seems to have indicated the
military groups in a general way.”’ This is supported by the reference
to the niyayattar of both the perundanam and chirudanam grades as
possessing virabhdga lands, that is, lands assigned for ‘the enjoyment
of soldiers’.”4 Next to the infantry were the cavalry and elephantry.
Chariots were not significant. About the navy, very little is known
even though naval expeditions are mentioned very frequently in the
eleventh century. A hazy reconstruction of the naval fleet is possible
using some circumstantial evidence relating to merchant boats and
some evidence relating to Sri Lankan navy.” Each regiment had a
captain called nadyagam; this term was later replaced by padai-mudali.
Above the nayagams were the dandandyagam and the commander-in-
chief was the sénapati.
One reason for attributing mercenary character to the Chola
army is the non-availability of direct evidence as to the mode of its
recruitment and maintenance. It seems at least some regiments were
stationed near and attached to big temples. For their maintenance they
were assigned lands which were called virabhéga (warriors enjoyment)
and padai-parru (military holding). Of course there were cantonments
(padaividu) in the capital cities. Perhaps there were cantonments
elsewhere also with such names as kadagam or parigriham. Stationing
military outposts (milai-padai) in conquered territories is heard of only
in the reign of Kulottunga I (1070-1120 ce).
The evidence for a royal army is explicitly found only up to the
early twelfth century, particularly for Period 2. In its heyday, during
the eleventh century, the Chéla army was a huge body. A Chalukya
inscription of 1007 refers to the invading Chola army as consisting of
900,000 troops.”° Though the figure might be exaggerated, it conveys
the real magnitude of the Chola army at that time. During the period
prior to Rajaraja I the armies of the local chiefs assisted the king’s
army. This trend seems to have reappeared in the latter half along
 The Chola State 231

with the emergence of local, padikdval chiefs in most of the outlying

areas and thereafter the king’s army was to be at the mercy of these
chiefs. This fact is clearly borne out by the political compacts among
chiefs mentioned above.
The Bureaucracy
The king was the supreme head of the government, and the top
executive order always issued from the king. That order was called
tiru-dnai (the sacred command) and the royal letter carrying this order
was known as tiru-olai or sri-mugam (Sri-mukham). To act against
the tiru-dnai was a sacrilege to be attended by severe punishment.
Inscriptions which narrate in detail the passage of royal orders for
executive action do not supply us any clear idea about his court
except the fact that the king was generally surrounded by a number
of officers.” The Kalingattupparani, an early twelfth century literary
work refers to the gorgeous royal court of Kuléttunga I (ca. 1100 ce),
attended by his many queens, lady attendants (anukkimdr), musicians,
chiefs and subordinate kings (mandaligar), and officers.”* The work
further gives an interesting insight into the court routine. The king
in court was informed by the ##rwmandira-olai about the subordinate
kings waiting ready to pay the annual tribute. The king was enraged
to hear the absence of the Kalinga king among the gathering and so
immediately gave orders to take an expedition against the erring ruler.
The king’s order was obeyed forthwith. There is not even a hint that
either the king consulted any counsellors or anybody volunteered to
give some advice to the king. In inscriptions also, it is always the king
who speaks, though there were occasions when somebody represented
(vinnappam or vijhapti) some matter to the king. But generally, we do
not come across any occasion when the king was advised.
The royal inscriptions of the eleventh century mostly relating to
land grants mention a number of officers, which would suggest, as
stated by Nilakanta Sastri,” that a large and powerful bureaucracy
assisted the king in the task of administration. These inscriptions show
that the oral order of the king regarding a grant had to pass through
a hierarchy of officials before being put into action. First the king's
word was committed to writing by the d/ai, then it was checked by
a group of dlai-nayagam officers, and after the same was endorsed by
a group of adikari and naduvirukkai officers it was passed on to the
puravuvari department for entry in the revenue register called vari. A
232 South India under the Cholas

copy of this entry, called w/vari, was then sent to the concerned party
(such as the temple, local bodies, and other such parties). It is rather
difficult to generalize this procedure for every administrative action,
as these royal grants must have been related to special ceremonial
occasions.®° Otherwise the procedure looks a cumbersome one. There
was seemingly no division of labour among the different officers. The
endorsing adikaris, for instance, numbered even fifty at times.
There are other non-formal records relating to bureaucraticactivities,
which give us a clearer picture of the structure of the bureaucracy.
Most of the bureaucratic activities relate naturally to religious matters
due to the peculiar nature of the inscriptional source. An inscription
of 1013 at Tirumalapadi, Tiruchirappalli District may be cited as a
typical example.*! In it is recorded that the king (Rajaraja I) asked
his officers to prepare copies of the inscriptions found on the central
shrine of a temple under renovation. Upon this, two adikaris, one a
miwvendavelan, and another a bramadhirajar issued, to this effect, their
niyoga (orders) to the local sabhd and to a panimagan of the chirudanam
grade governing the village (d#ra/kinra). At the spot a séndpati, called
Mummudichdéla-brahma-marayan, acted as the nayagam (supervisor)
on behalf of the second mentioned adikari and a soldier acted as a
representative of the panimagan. Some thirteen years later (1026),
the then ruling king Rajéndra I asked his officers to get the above
inscriptions re-engraved on the walls of the renovated temple.*? This
time a dandanayagam, called Uttamacéla-brahma-marayan, sent the
communication (d/a1) to one of his subordinates, who was doing the
Srikariyam of the temple on his behalf. The srikdriyam officer in turn
was represented by a soldier while the actual re-engraving was made.
There are many such instances where orders for executive action were
passed on through hierarchically positioned officers. About a third of
these instances fell into a four-level (I to IV) category and the rest fell
into a three-level (I, II/III, 1V) category, as indicated in Table 16.3.

Table 16.3: Hierarchy in Executive Action

Level Official hierarchy represented by
I King
II adikari/senapati/dandanayagam/puravuvari-tinaikkalam/naduvirukkai
Ill srikariyam/nadu-vagai/seénapatiladikari
IV soldier/Brahmana/karanaml mugavetti
Source: Compiled by the author.
 The Chola State 233

The king of course occupied the top level. The adikari figured
prominently in the second level fairly often and the séndpati was the
next prominent official at this level while the dandandyagam rarely
figured here. In the third level, the srikariyam officer was frequently
seen and next to him the ndduvagai. Very rarely were the adikari,
senapati, and dandandyagam found at this level as subordinates to
senior officers of their own category. In the last level, soldiers were
found very often and so also some Brahmanas (in the case of temple
functions). The relation of the subordinate officer to his superior was
usually denoted by the term kanmi meaning agent; sometimes this
relation was expressed by the phrase ‘X acting on behalf of Y’.
In spite of the religious bias of the inscriptions, we can understand
something about the functioning of the bureaucracy in secular
contexts if we analyse the activities of the officials as known from
inscriptions. These activities may be arranged in the following
categories: collection of taxes and fines; local administration
concerned with the supervision of the affairs of the brahmadéya and
dévadana villages; management of temple affairs (s‘rikdriyam); and
police and judiciary.

The Collection of Revenue

An area where the bureaucracy was very much present was that of
revenue collection. It is now well understood that for the Chola
government the major land tax, generally known as irai or kadamai
was the most important source of revenue, collected in kind from
the primary landholders.*’ Besides the land tax, there were also some
cash levies, Zyam, pattam, and antarayam collected from merchants,
artisans, and such others. While the land tax was collected throughout
the Chola rule, the cash taxes became significant only during the
later half. The other taxes mentioned severally or grouped under the
broad term kudimai were mostly labour levies, in the form of corvee,
and demanded from the primary producers, namely cultivators and
artisans, and used locally by locality assemblies of landholders (like
nattar, urar, and sabha) or by temple bodies. It is in the fixing and
collection of the major land tax that the activities of the government
machinery are found conspicuously in the inscriptional records.
It is, however, not easy to give the exact process of revenue settlement
and collection. There is no doubt that there prevailed a regular annual
collection of taxes, particularly land tax as we have specific references
234 South India under the Cholas

to such a collection.** But it is not possible to say that the settlement

for these annual taxes was also made every year. There are, however,
references to periodical revisions of the total volume of revenue from
a revenue unit, be it a village or nddu. This revision took into account
new lands, if any, brought under cultivation, which was off and
on checked by undertaking actual measurement of the concerned
lands. During bad days, petitions were made to the higher officers
and sometimes even to the king by the taxpayers for tax remissions,
sometimes successfully, sometimes in vain.
It was seen above that there was a separate department, called
puravuvari-tinaikkalam, for the maintenance of revenue accounts.
But lacking sufficient empirical cases, the record-keeping functions
of the different sections of this department can be understood only
vaguely from the names of the offices:*° varipottaga-kanakku, ‘the
accounts of the tax register’; mugavetti, ‘the royal stamp’; variyilidu,
‘entry in the tax register’; pattdlai, ‘the writer of the palm leaf record’,
and other such offices. Actually one and the same word was used
to denote both the function and the functionary. That the officials
of this department were mainly engaged in the maintenance of
revenue accounts may be understood from the fact that they rarely
performed varied functions as the other officials like the adikari or
the sénapati did. The higher officials who usually made the revenue
settlements in the localities were the adikaris in the first two periods,
and the varikku-kiiru-cheyvar in the later two periods. Perhaps they
were assisted at the lower level by the na@du-vagai officers during the
earlier half. There are many instances in the later half where these
higher settlement officers acted very arbitrarily in fixing the taxes.”
Moreover, during that stage the taxes were mostly fixed and collected
by locality chiefs in outlying areas.
From many references to nattu-mudal or nattu-opati (the revenue
from the nadu), it may be understood that the na@du was the basic
revenue unit. After the introduction of the valanadu over and above
the nadu, the valanddu might have been treated as the basic unit.
Whether the basic territorial unit was the na@du or the valanadu, the
body which represented these territorial units to negotiate with the
government was the nattar, the corporate body of the landholders. In
the case of the Brahmana villages, the sabha directly negotiated with
the government. In many instances, particularly in the later half, the
nattar and the sabha together represented the taxpayers’ side. The
negotiated settlement between the government and the bodies of
 The Chola State 235

landholders was called ottu (agreement). Though the settlement for

the taxes was made at the nddu level, the actual collection was made
at the village level.
It was mentioned above that the fixing and collection of the
kudimai group of taxes were done mostly by the locality bodies at
the village and locality level. At the same time these taxes, which
were mostly in the form of corvee labour, were demanded always
in the name of the king.** We do have a few cases where officials
are involved in fixing or demanding such levies. For example, in
Bahir, a brahmadéya settlement, an officer was present in 1027 when
the sabha of the place decided the quantum and mode of supply of
labour levies required for the maintenance of the village tank.*? There
is also another case in 1207 in Tanjavur District which says that an
officer unjustly demanded in excess of the prevailing commuted
amount towards the kattal-kasu, a part of the kudimai levies in a
brahmadeya village, ignoring a previous order of the king fixing it at a
lower quantum and that when this arrogant behaviour of the officer
was brought to his notice by another officer, the king ordered that his
original order should be adhered to.”
Though the collection of the land revenue by the Chéla government
is beyond doubt, the question relating to the transportation and storage
of the huge quantity of the collected grain could not be answered
satisfactorily. For, most of the land tax was paid in kind only. Only
in the case of taxes from merchants and artisans, money payments
were made and commutation of land tax into money was made
occasionally during the later half of the Chola rule. Some evidence is
there to say that some big towns and the capital cities had granaries
called nattu-pandaram or kottakaram for storing grains.”' Measuring
out the government paddy at the threshing floor and transportation
of the same to the government granaries was the responsibility of the
landholders who paid the tax, through their administrative bodies,
sabha or nattar.” As far as the cash taxes are concerned, they were
deposited into the government treasuries (karuvukalam, talam, and
the like) which were located in certain towns.” The cash collections,
however, formed only a minor share of the government revenue.
Local Administration
The presence of the king’s government in the localities may be studied
with reference to local officers of the government. There is some
evidence for mandalam-level administration during the imperial
236 South India under the Cholas

phase (Period 2). From the reign of Rajéndra | (1012-44 cE) some
viceroyalties were in existence for some five decades in the outlying
mandalams, \ike the Pandi-mandalam, the Ila-mandalam (Sri
Lanka), and the Ganga-mandalam. There is substantial evidence for
the viceroyalty of the Pandi-mandalam; the other viceroyalties were
not so conspicuous. The viceroys were generally the Chéla princes.
The office of mandala-mudali, the ‘Chief or Head of mandalam’ was
a short-lived one, it being confined to the reign of Kulottunga I only.
There are a few references to royal orders being communicated to the
mandala-mudali stationed in Ganga-mandalam (Mysore) and Pandi-
mandalam.™ Kulottunga I perhaps tried to replace the earlier viceroy
by the post of mandala-mudali manned by non-royal personnel. The
nadu-vagai officer at the nadu level has already been noticed. He was
certainly a government official, not a functionary of the corporate
body of the nadu or nattar. This office existed only during Periods 1
and 2 up to the reign of Kuldttunga I.
Whether the Chola government interfered in the functioning
of the local bodies is a moot point. We have a few instances in the
tenth century and then in the early thirteenth century where an
officer sat, at the bidding of the king, in the meetings of the sabha
of Brahmana villages. But this official presence did not go beyond
emphasizing some sastra-based solution to some problems referred
to the government for resolution by the local elite themselves.”
However, a village-level government official is heard of with reference
to the brahmadeya villages, until about the beginning of the twelfth
century. This official was variously called #r-dlvan, ir-paripalikkinra,
tiril-ninran or ur-mél-ninran, all meaning ‘one who administers the
village’.”° This may be due to the peculiar nature of the brahmadeéya
village, which required some special arrangement to administer it.
The concerned official seems to have functioned as a government tax
collector too.”
In the case of the general kind of villages, that is, the vellanvagai
villages, we do not come across any official similar to the ar-dlvan of
the brahmadeya villages. Of course there are a number of instances
where the village bodies refer to certain officers to whom they had to
pay taxes, like the #ril-tandininran (‘the one who was collecting taxes
in the village’), komurravar or mudaligal (the government officers
collectively). These government officers were only temporary visitors
in the villages, mainly for collecting taxes, not permanent village-
level government officers. The one regular village functionary who
 The Chola State 237

may be mistaken for a government officer was the ar-kanakku or

tir-karanam, the village accountant.
The ar-kanakku figured very often in the transactions of the
village bodies, like the arar, sabha, or nagaram, particularly where
an authoritative description of some village land was involved.
There is considerable explicit evidence to assert that the ar-kanakku
was a servant of the village bodies, not of the government. From
his designation madhyastha (the Tamil equivalent being kavidi), we
can suggest that the village accountant was supposed to be a neutral
person who was entrusted with preparing an unbiased account of
the village lands.** As one who knew intimately, and kept record
of, the accounts of the landholdings and the property rights inside
the village, the village accountant must be considered as the most
important executive link between the village and the government,
especially when it came to the question of tax collection. There are
a few inscriptions to lend support to this suggestion. An inscription
dated in 973 from Pullamangai, Tanjavur District records that the
madhyastha of the brahmadéya village had to forfeit his service tenure
land (kavidi-kani) since he failed to submit the proper accounts,
to a Sundaracéla-muttaraiyar, of the paddy dues of the village and
the money collected from the Vellalas and the Brahmanas of the
village.” Sundarac6la-muttaraiyar may easily be guessed as a superior
government officer, from the way his name (given in honorific
plural) is mentioned and from his title. Another inscription dated
1001 from Tiruvidaimarudir, Tanjavur District mentions that a
subordinate accountant (kil-kanakku) who was writing accounts in
a Brahmana village under a tala-pattolai (obviously one connected
with the puravuvari department) absconded without producing the
proper accounts and so his landed property (dmi) was ordered to
be confiscated (by the government). To this effect a superior officer
(most probably séndpati) issued the final execution order (kadaiyidu)
and sent some army personnel to the concerned village to carry out
the order.' Being subject to much coercion, the local sabha had to
sell away the lands of the defaulting accountant to pay the dues to
the government.
Prior to the time of the Cholas, temples were not controlled by
the government. They were managed mostly by local communities
through their communal assemblies, namely, the sabhd, the
miilaparushai, the arar, and other such assemblies. Even during
the Chéla period this communal management continued to a great
238 South India under the Cholas

extent. But the Chéla kings tried to control the temple administration
through the creation of the srikdriyam office. The management of
the temple was given the utmost care during the first and second
periods. As noted above, some prominent temples had permanent
srikariyam officers to supervise their affairs regularly. In the case of
most other temples periodical visits of officers for inspection, either
individually or as a team, served this purpose. The visiting officers
included adikari, senapati, dandandyagam, and very rarely a member
of the puravuvari department.
The creation of the dévadana or temple villages was also entrusted
to one of the above officers. These officers were ordered to fix the
boundaries of the newly created villages, to settle these villages with
tenants and apportion the income for various routine and special
expenses of the concerned temples. Again the officers were used to
tour the country to enquire into the affairs of such dévadana villages
and hear the grievances of the temple functionaries and the cultivating
tenants there. On many occasions, the local nadu-vagai officer was
also directed by the touring superior officers (adikari, senapati).'®'

Police and Judiciary

The king’s government had no special establishment other than the
army for policing the country. The local bodies seem to have had their
own policing arrangement in the first period. No such evidence is
forthcoming for Period 2. Being the peak period of imperial expansion
the military presence was greatly felt even in remote parts of the
country, from the commander-in-chief (sénapati) at the top to the
soldier at the bottom. We have instances where the military personnel
were employed to collect tax arrears! or even to punish a temple
priest who had somehow incurred the displeasure of the queen.'® In
the third and fourth periods, that is, during the twelfth—thirteenth
centuries, most of the outlying areas were put under the charge of
local chiefs, who provided the required protection, padikaval, in lieu of
some tax remuneration. Gradually the padikaval arrangement became
the only ‘government’ in most outlying localities as the effectiveness
of the Chéla government decreased.
If there was no separate establishment for policing, there was
neither one for dispensing justice. Nilakanta Sastri considered the
term niydayattar as a local level judicial body and took another term
dharmasana (‘the seat of justice’) which occurs frequently in early
inscriptions as the king’s court of justice.'°* Both of his views must be
 The Chola State 239

corrected. It was noted above that the term niydyattar was used as a
general designation for army people.'° The term dharmdsana occurs
only in the case of brahmadéya villages. Actually an inscription of
Uttaramallur, dated 994, clearly distinguishes it from ‘the king’s gate’
(rajadvara), that is, the king’s court.'®° This inscription classifies fines
(dandam) as those that might be levied at the king’s court, at the
dharmasana or at the vari (revenue office). The last one must be the
fines imposed in case of default in tax payment. The naduvirukkai, as
seen above, may be treated as a judicial office as all the incumbents
of this office are found to be bhattas, that is, learned Brahmanas
well-versed in dharmaSastra literature. But this office was a short-
lived one. In the few instances where an executive punishment
is mentioned, the punishing authority was any one of the high
officials, like the adikdri. Several cases of punishment executed by
some officers by the orders of the king are found in the first half of
the thirteenth century when cases of treason (rajadroham) against
the king committed by high officials (mostly military) were on the
increase. This is in a way supported by Songshi and other Chinese
annals of the twelfth-fourteenth centuries, which contains some
contemporary accounts of south India say that if any person is found
guilty of an offence in the Chola realm (Chu-lien Zhuwan) one of
the Court ministers punishes him, and so on.'” In all these cases,
the mode of punishment is found simple. When the acts of treason
were brought to the notice of the king, he ordered some of his higher
officers to confiscate the entire properties of the traitors and sell
them in public auction (peruvilai).'°* Usually the realized money was
asked to be deposited in some specified government treasury, called
variously karuvukalam, talam, tandal, and kundigai. Sometimes the
treasury had a jail attached to it, where the tax defaulter was put
in confinement.'”
Apart from the above judicial cases involving the government,
most of the available cases of judicial proceedings of the Chéla period
recorded in inscriptions are found to be communal or local in nature.
That is, disputes were adjudicated by the communal, corporate
bodies of the localities (for example, the néttar) or the caste assemblies
without reference to the king’s government. Case 1 from Narttamalai,
Pudukkottai District, dated 1056 relates to some decision of the
merchant guild Ayyavole-500 which enquired the death by suicide
of a woman of the chetti caste due to some bad treatment of her in-
laws.!!° Next, in South Arcot District there are several inscriptions
240 South India under the Cholas

where the pa/li-nattar or pannattar assembly is found to enquire and

give punishment when members of the pa/li caste are involved.'"'
Then we have the case of Chittiraméli-periyanattar, the bigger Vellala
assembly, which sat in judgement over a dispute involving two Vellala
brothers.''? In none of these judicial settlements, the government
interfered directly or indirectly. These judicial bodies behave more or
less as the later-day caste panchayats.
* OK OK OK

The salient features of the Chola state that may be recognized from
the information presented so far are as follows. It grew from being a
tiny state in the ninth century to a remarkably large one in the course
of the next two centuries or so. Its career falls into three phases: (i)
pre-imperial (850-985), (ii) imperial (986-1100 ce), and (iii) post-
imperial (1100-1250 ce). During the first phase of its existence it was
a small kingdom claiming sovereign rights over a restricted territory
and surrounded by several chiefly territories. At this time wars were
waged mainly to get tribute from vanquished rulers. No outside
territory was annexed to the core area. Local bodies and communities
with large vestiges of tribal features enjoyed their traditional power
without much governmental interference.
The imperial phase, inaugurated in the reign of Rajaraja I (985—
1014 cE), was marked with a powerful monarchy, growth of private
property in land and tendencies towards stratified relations in society.
The ruling strata were dominated by the Vellala and Brahmana
landholders. The king of this phase was very different from his
earlier counterparts. He was not only a great warrior but also a real
administrator assisted by an organized body of officials and army.
There were made deliberate royal efforts at centralization of the
political power. Territorial reorganization (like creation of valanddu
and mandalam), creation of an elaborate bureaucracy, extensive land
survey and detailed recording of land rights for revenue purposes,
government control of temples and brahmadeyas, and mobilization
of a big standing army for offensive and defensive purposes and also
for the maintenance of internal order were the salient features of this
centralization process.
The state control of the irrigation system, the very basis of the
agricultural economy of the times, can only be inferred, as there is no
direct evidence in the form of a separate government department or
special category of officers as in the case of the revenue matters. The
 The Chola State 241

irrigation network of the Kavéri delta, with its innumerable canals, had
reached its optimum level by the tenth century. The fact that several
of the sluices and canals, big and small, are named after the kings and
queens would support the proposition that most of the water courses
were royal creations or undertaken by royal initiatives. Thereafter,
their maintenance was the main lookout of the government. Though
local communities played a considerable role in the maintenance of
these irrigation sources, the corvee for these works, particularly for
those of big canals, was always demanded in the name of the king.
It is to be noted that the officials were actively mobile and not
tied to their localities. This highly mobile bureaucracy as well as the
army of the imperial phase certainly contributed to the emergence
of the complex agrarian system of the post-imperial phase of the
twelfth and thirteenth centuries. The consequent growth of new
landholding groups reduced proportionately the status and role of
the Vellala and the Brahmana landholders in the ruling stratum.
The other concomitant result of this situation was a fair degree of
feudalization of land relations. The centralized power structure of the
imperial phase, which was more dependent upon the ability of the
ruling king, could not continue under the changed circumstances.
There was obviously no imaginative revamping of the administration
to cope with the new situation. Fragmentation of political authority
with the emergence of various locality chiefs and multiplicity of
tax-collecting agencies imposed enormous tax burden on the actual
cultivators (ulukudi).''> Consequently, there developed an agrarian
crisis due to open confrontation between the big landholders and the
cultivators. The ultimate demise of the Chéla state may be attributed
more to this agrarian crisis than to any other factor.

Notes
1. K.A. Nilakanta Sastri, The Calas, 2nd edn, 1955 (1st edn, 2 vols, 1935-7),
University of Madras, Madras.
2. Burton Stein, Peasant State and Society in Medieval South India, Oxford
University Press, Delhi, 1980 (incorporates all his articles since 1965); Noboru
Karashima, South Indian History and Society: Studies from Inscriptions, AD 850-1800,
Oxford University Press, Delhi 1984 (incorporates all his articles since 1966).
B. Suresh (Pillai), ‘Historical and Cultural Geography and Ethnology of South
India with special reference to the Chola Inscriptions’ (unpublished PhD thesis),
Deccan College Post-graduate and Research Institute, Pune, 1965; George W.
Spencer, The Politics ofExpansion: The Chola Conquest of Sri Lanka and Sri Vijaya,
New Era Publications, Madras, 1983 (revised version of his PhD thesis, 1967); Y.
242 South India under the Cholas

Subbarayalu, Political Geography ofthe Chola Country, Tamil Nadu State Department
of Archaeology, Madras, 1973; Y. Subbarayalu, ‘State in Medieval South India, 600-—
1350, with Special Reference to the Pallava and Chéla Rules’ (unpublished PhD
thesis), Madurai Kamaraj University, Madurai, 1976; Y. Subbarayallu, “The Cola
State’, Studies in History, vol. iv, no. 2, 1982, pp. 265-306; Noboru Karashima, Y.
Subbarayalu, and Toru Matsui, A Concordance of the Names in the Cola Inscriptions,
Sarvodaya Ilakkiya Pannai, Madurai, 1978; Kenneth R. Hall, Trade and Statecraft in
the Age ofthe Cholas, Abhinav Publications, Delhi, 1980; P. Shanmugam, The Revenue
System ofthe Cholas, 850-1279, New Era publications, Madras, 1987 (revised version
of his 1977 PhD thesis); R. Tirumalai, Land Grants and Agrarian Reactions in Chola
and Pandya Times, University of Madras, Madras, 1987; R. Champakalakshmi,
‘The Study of Settlement Patterns in the Cola Period: Some Perspectives’, Man and
Environment, vol. 14, 1989, pp. 91-101; James Heitzman, Gifts of Power: Lordship
in an Early Indian State, Oxford University Press, New Delhi, 1997 (revised version
of his 1985 PhD thesis).
3. First used by Noboru Karashima and B. Sitaraman in their analysis of the
Chdla revenue terms, this periodization is followed in most of the subsequent studies
of Noboru Karashima (South Indian History and Society, pp. 71f) and his colleagues
and recently by James Heitzman, Gifts of Power: Lordship in an Early Indian State,
Oxford University Press, New Delhi, 1997.
4. Tamil Nadu, the name of the present linguistic state, where Tamil is the lingua
franca, is used here only as a convenient term of reference. It need not connote
exactly the ancient linguistic or political boundaries.
5. The Colas, p. 116. Also see George W. Spencer, ‘Heirs Apparent: Fiction
and Function in Chéla Mythical Genealogies’, Indian Economic and Social History
Review, vol. xxi, 1984, pp. 415-32.
6. R. Champakalakshmi, ‘Ideology and the State in Medieval South India’,
Mamidipudi Venkatarangaiah Memorial Lecture, Proceedings of the Andhra Pradesh
History Congress, 13th Session, Srisailam, 1989.
7. K.A. Nilakanta Sastri, A History of South India, 3rd edn, Oxford University
Press, Madras, 1966, p. 173.
8. Y. Subbarayalu, Political Geography ofthe Chola Country, State Department of
Archaeology, Tamil Nadu, Madras, 1973, pp. 72-83. The political organization of
the Paluvédu, Irukkuvél, and Malava chieftaincies is discussed in Y. Subbarayalu,
‘State in Medieval South India, 600-1350, with Special Reference to the Pallava and
Chola Rules’.
9. See Chapter 6 in this volume.
10. For the Chéla-Pandya viceroyalty see S//, vol. xiv, Introduction. A few
inscriptions of the Chola-Lankésvara viceroyalty have been discovered in Trincomelee
District in eastern Sri Lanka. S. Pathmanathan, The Kingdom ofJaffna, Colombo,
1978, pp. 38-9.
11. See Chapter 13 in this volume.
12. A brief historical account of the major chiefs in the later half of the Chéla rule
is given in Nilakanta Sastri, The Colas, pp. 400-7. For a discussion of the padikaval
system see Nobou Karashima, Ancient to Medieval: South Indian Society in India,
New Delhi, Oxford University Press, 2009, pp. 136-54.
 The Chola State 243

13. Sastri, The Colas, p. 534.

14. This is understood from the analysis of the royal (‘K6nérinmaikondan’)
inscriptions (See Chapter 5 in this volume). From the distribution of these
inscriptions it may be asserted that the king clearly avoided communicating his
orders into the localities under the chiefs. Y. Subbarayalu, ‘State in Medieval South
India’, p. 58, note 80.
15. For example, the case of some Churutiman (Surutiman) chiefs in the
Tiruchirappalli District may be cited. Noboru Karashima, South Indian History and
Society, pp. 21-31.
16. Annual Report on (South) Indian Epigraphy (ARE hereafter), 1934-5, p. 62;
South Indian Inscriptions (SII hereafter), vol. vii, nos 119 and 127; SII, vol. viii, no.
106; SII, vol. xvii, no. 245.
17. Epigraphia Indica (EI hereafter), vol. xxii, no. 14; S//, vol. vii, nos 149, 890,
and 1020; Nilakanta Sastri, The Colas, pp. 374-5, 407.
18. SI, vol. xvii, no. 599.
19. SII, vol. ii, no. 7; Nilakanta Sastri, The Colas, pp. 139, 186, 332.
20. SII, vol. xiii, no. 39; SII, vol. xix, no. 311.
21. T.V. Mahalingam, South Indian Polity, 2nd edn University of Madras,
Madras, 1967, pp. 38ff. Also Kalingattupparani (by Jayangondar), (ed.), Puliyur
Kesikan, Chennai, 1958, verse 264.
22. SII, vol. iv, no. 537; SII, vol. xiii, nos 44 and 46.
23. Y. Subbarayalu, Political Geography, pp. 72-80.
24. Ibid., pp. 56-7.
25. This was again changed a decade later as Rajaraja-pandinadu.
26. Ibid., pp. 14-17; Also see S//, vol. ii, Introduction.
27. Ibid., pp. 14, 83.
28. Some settlements, such as Milalai, Ollaiyair, Ambar, and so on, after which
some nddus were named are referred to in the Tamil literary works of the early
centuries CE.
29. Y. Subbarayalu, Political Geography, pp. 22, 30-2. Names of some ndadus (in
the Pandya country and Tondai-mandalam) with the suffix components kulakkil or
érikil meaning ‘under the tank’ confirm the suggestion that they formed about some
irrigation source. Kenneth R. Hall’s suggestion (Trade and Statecraft in the Age ofthe
Colas, p. 187) that the nadu was primarily a local marketing territory served by a
common nagaram market and that the economic functions of a nddu preceded all the
other structural features cannot be maintained as the growth of most of the nagarams
was concomitant with or consequent to agrarian expansion and large-scale temple
building activities of the Chola times. Further, only about 100 out of 500 and odd
nadus that had come into existence by the twelfth century had a nagaram settlement
within their boundaries. (Noboru Karashima, Y. Subbarayalu, and P. Shanmugam,
‘Nagaram during the Chéla and Pandyan Period: Commerce and Towns in the Tamil
Country ap 850-1350’, Indian Historical Review, 35 (1), 2008, pp. 1-33).
30. SII, vol. ii, nos 4 and 5. Also see Karashima, South Indian History and Society,
pp. 46-7.
31. Before the ninth century, only the ‘mangalam’ suffix was in use.
32. Burton Stein has reviewed extensively the historical literature pertaining to
244 South India under the Cholas

the brahmadeyas of the Pallava and Chéla times and has made a distribution map
of the sites (Peasant State and Society, pp. 141-72). See also R. Champakalakshmi,
‘Reappraisal of a Brahmanical Institution: The Brahmadéya and Its Ramifications
in Early Medieval South India’, in Kenneth R. Hall (ed.), Structure and Society
in Early South India: Essays in Honour of Noboru Karashima, Oxford University
Press, New Delhi, 2001, pp. 59-84; P. Shanmugam, ‘Brahmadeyas in the Pallava
Country’, in S. Rajagopal (ed.), Kaveri: Studies in Epigraphy, Archaeology, and History,
Panpattuveliyiittakam, Chennai, 2001, pp. 331-42; S. Rajavelu, “Migration of
Brahmins to Tamil Nadu under the Pallavas’, [bid., pp. 397-407.
33. K.G. Krishnan (ed.), Karandai Tamil Sangam Plate of Rajendrachola I
(Memoirs of the ASI, no. 79), New Delhi, 1984.
34. This interpretation of vettappéru is different from that of T.N. Subramanian
who takes it to denote ‘lands given to those who performed vedic sacrifices’.
(Transactions of the Archaeological Society of South India, 1958-9, pp. 91-2). See
Chapter 5 in this volume: section on vettappéru.
35. SII, vol. viii, no. 689; S//, xiii, no. 240.
36. This is also the impression of C.N. Subramanian’, Aspects of the History of
Agriculture in the Kaveri Delta, ca. 850 to 1600’, Unpublished MPhil Dissertation,
Jawaharlal Nehru University, New Delhi, 1984.
37. Political Geography ofthe Chola Country, pp. 45-6, 92-4.
38. For a detailed account of the taniyir village, see R. Champakalakshmi,
‘Reappraisal of a Brahmanical Institution’.
39. Y. Subbarayalu, “Social Change in Tamilnadu in the Twelfth and Thirteenth
centuries’, South Indian History Congress: Proceedings of II Annual Conference,
Trivandrum, 1981, 138-42. See also Chapter 13 in this volume.
40. Information derived from Noboru Karashima et al., A Concordance of the
Names in the Cola Inscriptions.
41. SII, vol. v, no. 1409; SII, vol. vii, no. 118.
42. SII, vol. iv, no. 223; SII, vol. xiii, nos 58and 215; S//, vol. xix, no. 19; ARE,
1914, no. 153.
43. The Colas, pp. 492 ff; Y. Subbarayalu, Political Geography, pp. 33-6.
44. There existed some difference between Kerala and Tamil Nadu in the
usage of this nomenclature. In Kerala, the terms arar (aralar) and nattar denoted
the Brahmana bodies also. See Kesavan Veluthat, Brahman Settlements in Kerala:
Historical Studies, Sandhya Publications, Calicut University, 1978, pp. 54-5; Rajan
Gurukkal, “The Socio-economic Milieu of the Kerala Temple’, Studies in History,
vol. ii, no. 1, 1980, p. 3.
45. Variyam is met with in a few nagaram settlements too.
46. K.V. Subrahmanya Aiyer, Historical Sketches of Ancient Dekhan, vol. ii,
Coimbatore, 1967, pp. 247, 260, 269, 275, where the author derives everything
connected with the sabha from the dharmasastra works.
47. See Chapter 11 in this volume.
48. Noboru Karashima, South Indian History and Society, pp. 3-16.
49. Ibid., pp. 21-31.
50. Ibid., pp. 26-7.
51. SII, vol. xxiii, nos 381 and 383.
 The Chola State 245

52. K.A. Nilakanta Sastri, The Colas, pp. 578-9.

53. ARE, 1943-4, no. 268. This feature became still more prominent in the latter
half of the thirteenth century when the Pandya rule replaced the Chdla one (S//, vol.
villi, nos 591 and 247).
54. Thus an inscription from Tirukkadaiyir, Tanjavur District (ca. 1152)
in the reign of Rajaraja II says, rather tersely, that ‘it is a warning of the king to
any rajakulavar (“those of ruling clans”) who would dare to purchase kami in the
dévadana \ands and also to any among the kudimakkal (“cultivators”) who would
dare to purchase such kami in excess of two véli’. SII, vol. xxii, no. 31. See Chapter
12 in this volume for a related discussion.
55. Noboru Karashima et al., A Concordance ofthe Names, vol. i, pp. xlvii-li.
56. Thus a Korramangalam-udaiyan Chiralan Kandarachchan who hailed from
Korramangalam in Tirunaraiyar-nadu had the title Irumudichdla-tirunaraiyar- -
nattu-vilupparaiyan (S//, vol. v, no. 1356).
57. SII, vol. xix, no. 218.
58. SII, vol. ii, no. 205, lines 37-49.
59. Nilakanta Sastri, The Colas, p. 472.
60. Two bhattas were deputed by Rajaraja I in 987 to punish the culprits who had
been involved fifteen years back in the murder of his elder brother and crown prince,
Aditya II (EY, vol. xxi, no. 27). These bhattas might have been naduvirukkai, though
they are not mentioned so. In the Leiden grant of Rajaraja I, five naduvirukkai
officers (all b/attas) are directed by the king to supervise the demarcation of a village
granted to a Buddhist institution (£/, vol. xxii, no. 34, lines 41-8).
61. SIZ, vol. viii, nos 618 and 620.
62. Detailed analysis of this is made in Y. Subbarayalu, ‘State in Medieval South
India’, pp. 143-8. See also P. Shanmugam, The Revenue System ofthe Chélas.
63. Y. Subbarayalu, “State in Medieval South India’, pp. 291-306, Appendix 14.
64. Nilakanta Sastri, The Colas, p. 463.
65. SII, vol. vi, no. 57; JPS, no. 135; viii, no. 252.
66. An official who built a temple on behalf of the queen dowager had a title
called Rajakésari-mivéndavélar, which is mentioned as pattam-kattina-pér, meaning
‘the title that was conferred on (him) (S//, vol. iii, no. 147. The prefix Rajakésari is
one of the alternative attributes of the Chola kings.
67. SII, vol. v, no. 723; SII, vol. ii, no. 95; EC, vol. xiv, Tn. 34. See also
Chapter 4 in this volume.
68. A later-day tradition of the kaikkola caste (for example, Thurston, Castes
and Tribes of South India, vol. iii, p. 31) which derives the alternative name sen-
kundar of this caste from a weapon called kundam (spear or lance) seems to preserve
the memories of its military origins. Another such military group turned weavers
is the niydyam ot niydyattar according to some early fifteenth century Vijayanagar
inscriptions relating to Right Hand and Left Hand groupings (ARE, 1914, no. 59;
ARE, 1917, no. 216).
69. K.A. Nilakanta Sastri, The Colas, p. 454.
70. EI, vol. xviii, pp. 330-8.
71. K. Indrapala, ‘South Indian Mercantile Communities in Ceylon, ca. 950-1250’,
The Ceylon Journal ofHistorical and Social Studies,New series 1, 1971, pp. 101-13.
246 South India under the Cholas

72. ARE, 1929-30, nos 267-9; ARE, 1934-5, p. 61.

73. The term niydyattar (niydyam is the basic form) is derived by Hultzsch from
nyasa and explained as a dedicated group of servants (SI//, vol. ii, p. 256 note). But it
seems that the emphasis was upon the military groups.
74. ARE, 1923, no. 224.
75. Y. Subbarayalu, ‘A Note on the Navy of the Chéla State’, in Hermann Kulke,
K. Kesavapany, and Vijay Sakhuja (eds), Nagapattinam to Svarnadwipa: Reflections
on the Naval Expeditions to Southeast Asia, Institute of Southeast Asian Studies,
Singapore, 2009, pp. 91-5.
76. EI, vol. xvi, p. 74.
77. According to Chau Ju Kua, four Court ministers were present at state
banquets. Nilakanta Sastri (ed.), Foreign Notices of South India: From Megasthenes
to Ma Huan, University of Madras, Madras, 1972, p. 143. For Songshi version, see .
Karashima, Ancient to Medieval, Chapter 13, p. 269.
78. Puliyur Kesikan (ed.), Kalingattupparani, Aruna Publications, Chennai,
1958, verses 314-42.
79. The Colas, p. 461.
80. These grants generally relate to the establishment of a new brahmadeéya or
dévadana.
81. SI, vol. v, no. 652.
82. Ibid., no. 651.
83. For details, see Chapters 7 and 8 in this volume.
84. For example, an inscription of Kuldttunga I dated in his 44th year (1114 CE)
from Tiruvadi, South Arcot district, records that the #rar of that village, failing other
means, sold some of their land to pay off the arrears in the kadamai dues of the 43rd
year (SII, vol. viii, no. 324). There are several such references to annual payments:
SIT, vol. vii, nos 96 and 97; SIJ, vol. viii, no. 303; S//, xvii, no. 590; SJ, vol. xxvi,
nos 663 and 693.
85. K.A. Nilakanta Sastri, The Calas, pp. 537-9.
86. Ibid., pp. 469-70.
87. South Indian Temple Inscriptions, nos 518-19. See also Nilakanta Sastri, The
Colas, p. 537.
88. See Chapter 7 in this volume.
89. Ibid., note 15.
90. SII, vol. viii, no. 263.
91. SI, vol. ii, nos 68-9; SI, vol., iv, no. 435.
92. See Chapter 12 in this volume. Also Chapter 5 in this volume: section on
kottakaram.
93. SII, vol. xxii, no. 54; SI, xxiii, nos 310 and 385.
94. EC, vol. xiv, Yl. 146; PS, no. 126.
95. The Colas, p. 499.
96. SII, iti, no. 156; SII, v, no. 652; SII, vol. vi, no. 292; SII, vol. xiii, nos 152
and 342.
97. SIT, vol. v, no. 702.
98. In the contexts we have only to agree with this suggestion of Nilakanta
Sastri, who rejected the meaning of ‘arbitrator’ which is usually assigned to the term
 The Chola State 247

(The Colas, p. 510). The madhyastha always figures as an accountant/writer obeying

the commands of the village sabhd; he never sits as an arbitrator.
99. SIT, vol. xix, no. 63.
100. SI/, vol. v, no. 723.
101. SUZ, iv, no. 391; SII, vol. v, nos 724 and 1378; SI/, vol. vii, no. 891; SI,
vol. xvii, 328; Tamaraippakkam Inscriptions, Tamil Nadu State Department of
Archaeology, Chennai, 2000.
102. S//, vol. v, no. 723; SI, vol. xxii, no. 27; ARE, 1908, no. 447.
103. SII, vol. iv, no. 391.
104. Nilakanta Sastri, The Colas, pp. 473-4.
105. See note 73 above.
106. FI, vol. xxii, no. 32.
107. Karashima, Ancient to Medieval, p. 264.
108. SII, vol. xxii, no. 54; S//, vol. xxiii, nos 310 and 385; Also see K.A. Nilakanta
Sastri, The Colas, pp. 126-7. See also Chapter 10 note 32.
109. SII, vol. xii, no. 224.
110. SI, vol. xvii, no. 389.
111. SI, vol. xvii, no. 200; SI, vol. vii, nos 68, 70, and 85.
112. Tamaraippakkam Inscriptions, Chennai, State Department of Archaeology,
Tarhil Nadu, 1999, pp. 54-6.
113. See Chapter 12 in this volume.
 CHAPTER 17

Characterizing the Chola State*

discussion relating to the characteristics of the Chéla State or any

ther medieval south Indian state, can only be tentative from the
very nature of the inscriptional data that it has to depend upon.
Matters relating to the temple get the major share in inscriptions as
they are primarily meant for recording charities to temples. In spite
of this bias, the basic details of several other things of contemporary
culture are discernible in the secular context. The evidence as available
in the present stage of our research and understanding gives a fair
idea about the remarkable changes in the economy and society of the
Tamil country through the four centuries of the Chéla rule} This long
duration saw the disappearance of the last vestiges of eaftier tribal
society leading to the emergence of stratified society, breakdown
of communal property and feudalization of landed relations, and
steady growth of kingly power accompanied with a deliberate policy
of centralization, elaboration of bureaucracy and standing army,
survey and careful recording of land rights for revenue collection,
territorial reorganization, and control of temples and Brahmana
villages. Any formulation relating to the political organization of
the Chola times must take into account these dynamic aspects.
The conception of the Chéla state that the previous generation of
scholars put forward may be said to be defective in that it did not
take adequate notice of the socio-economic changes over the four
centuries. Nilakanta Sastri, the doyen of Chéla history, viewed the
Chola State as an ‘imperial’ state associated with a Byzantine royalty
and a highly organized and thoroughly efficient bureaucracy. At the
same time the king’s government is said to have interfered in local
affairs minimally.'

*This has been revised and updated from the article “The Cola State’, Studies in
History, 4 (2), 1982.
 Characterizing the Chola State 249

The more one reads the contemporary records, the more one begins to admire
the nice balance struck between centralised control and local initiative, the
clear distinction, ever present, between the functions of the state and those
of the social group. The individual as such did not count. The problem of
‘the man versus the state’ never arose in a society that is best described as a
federation of groups.
The apparent contradiction in this conception was not noticed by
most south Indian historians until Burton Stein made a scathing
criticism of it. Stein presented, through a couple of articles an
alternative conception, that of the segmentary state, which he thought
would fit in with the prevailing agrarian order and the decentralized
nature of the polity providing scope for the active participation of
local institutions.” He later elaborated his idea fully in his book,
Peasant State and Society.’ Stein holds that two different, but related
formulations, namely, the pyramidal segmentary state and sacral
kingship, would provide the basis for a better understanding of the
medieval political system of south India and for a better explanation
of political evidence from that period.*
Taking the idea of the segmentary state from Aidan Southall’s
discussion of Alur, an East African tribal society, Stein presents his
formulation as follows: °
(i) In segmentary state, sovereignty is dual, consisting of actual political
sovereignty and ritual sovereignty. (ii) In it there may be numerous
‘centres’ of which one has primacy as a source of ritual sovereignty, but all
exercise actual political control over a part, or segment, of political system
encompassed by the state. (iii) The specialized ‘administrative staff’ is not an
exclusive feature of the primary centre but is found operating at and within
the segments of which the state consists. (iv) Subordinate levels or ‘zones’
of the segmentary state may be distinguished and the organization of these
is ‘pyramidal’.
In Stein’s view the basic segments of the south Indian medieval
segmentary political system were the nadus under the leadership of
chiefs who in the Chola period held titles such as udaiydr, arasar,
mummudi, or mivendavelar.® Segmentary systems, as understood by
anthropologists, are always associated with complementary opposition
within and among their segments.’ In view of ‘the profound
differences between African societies and those of India, including
the far weaker (but not absent) clan and lineage structures at the
lowest levels of society in India, and the overarching ideological and
250 South India under the Cholas

ritual integration achieved under Indian conceptions of kingship’,

Stein carefully adds that the dominant basis of opposition of the
nadu segments was not that of ethnically or culturally differentiated
peoples as in the case of African societies. 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 dominant castes from which
they had emerged, between locally dominant landed groups and
subordinate ones, between agricultural and non-agricultural groups,
between established castes of a locality and new-comers or outsiders,
and among sect and cult groups. Many of these oppositions took
concrete form in the Right and Left caste groupings.*
Analysing the territorial structure of the Chéla state into three
inferential zones as central, intermediate, and peripheral,’ Stein asserts
that effective territorial sovereignty of the Cholas was confined only
to the central zone, namely, the rich populous core of the Kavéri delta
and that beyond this region Chola sovereignty was an increasingly
ritual hegemony as the peripheral zones (Kongu and Gangavadi) of the
state were approached.'® The myth of Gangetic origin, the royal Siva
Cult which incorporated the local place and caste tutelary deities, the
network of Brahmanical institutions established throughout the realm
accompanied by impressive ceremonials, at times jointly participated
in by the Chola ruling house and locally dominant personages (nattar,
for instance), propagation of a standard symbolic system through
copper and stone inscriptions, and the like, are all viewed rather as
means used by the Chola rulers to affect ritual hegemony over the
numerous locality chieftains of the macro region."
In the segmentary state conception there is little scope either for a
bureaucracy or for a standing royal army. Just as locality institutions
provided most of the administrative functions required at the time,
so too, it must be supposed that the major forces involved in the
wars of the Chélas were supplied from the existing organizations of
the locality-based society of the time.'? The members of the so-called
bureaucracy are considered either as mere scribal functionaries or fully
independent locality chiefs.'* Stein further denies the existence of any
regular tax transfers from localities to the central government."
Stein's conception of the segmentary state is diametrically opposite
to the conception of a highly centralized state described by Nilakanta
Sastri. Even though in certain respects this conception looks a better
one to explain the nature of the Chéla polity with its vigorous
 Characterizing the Chola State 251

local institutions, it fails to take into account certain explicit data.

The analysis of the Chdla territory into central, intermediate, and
peripheral zones is a welcome feature to focus attention upon the
real nature of the Chola imperial claims. That the effective territorial
sovereignty of the Chélas did not extend much beyond the central
zone is supported by another specific study Spencer and Hall made
by mapping the distribution of inscriptions in the eleventh century.'°
But the zones have to be defined more rigorously as at least a part
of Stein’s intermediate zone, namely, Tondai-mandalam would fall
within the central zone if we consider the occurrence of numerous
inscriptions there throughout the four periods.
In his treatment of the nddus as the basic segments—crucial to
his segmentary state—Stein overlooks certain possible changes in
the structure of nadus over time. Though he distinguishes nadus
according to their hierarchical or tribal features he seems to assume
little change in the socio-economic structure basic to each naddu, but
his clarification about the oppositional elements within the nddu
segments would imply some change. Complementary opposition,
the characteristic feature of segmentary systems, was lacking between
and within the nddus as Stein himself implies this when he says that
the relations between opposing elements were often asymmetrical.'®
In the absence of balanced and opposed elements, would it be
appropriate to speak of pyramidal segmentation? Most of the basic
criteria used by Stein in classifying the nddus as central, intermediary,
and peripheral are speculative based upon very recent phenomena
which are projected many centuries backward. Ironically, Stein
whose critique attacked the telescoping of data from disparate times
and spaces in the earlier works on south India has committed the
same mistakes. His suggestion that the ethnic cohesiveness of nd@dus
was due to the spatially compressing character of the marriage system
which existed among the dominant, land-controlling peasantry as
well as among most other locality social groups may not be tenable.
It is found that most of the herders (idaiya) who took charge of the
animals of the Big Temple at Tanjavur in about 1014 are found to
have had a kinship network spread over distances ranging from 50
to 90 kilometres.'” As noted elsewhere, the ethnic cohesiveness of
the nadu micro-regions that might have existed during pre-Chola
times started to disintegrate even by the tenth century, due to the
growth of the Brahman settlements.'* This disintegration was further
accelerated by the imperial wars of the eleventh century which
252 South India under the Cholas

inducted many martial groups of the fringe areas into the fertile
plains. The administrative activities of the nastar became altogether
dormant in the central area by the end of the eleventh century due
to the overbearing presence of the Chola government there. On the
other hand, the nattar continued to be active in the dry tracts lacking
facilities for extensive irrigated agriculture where Stein would locate his
‘intermediate’ na@dus.'? It is again in these areas where the periyanattar
supra-local body of landholders became active.” This is rather
contrary to Stein’s assertion that the horizontal segmentation among
the sub-castes and clans along with the relatively sparse populations
supportable by mixed agricultural and pastoral utilization of these dry
lands made for fewer durable linkages among neighbouring @dus as
compared with ‘central’ nadus.*! This conflicting evidence about the
nadu segments makes the segmentary state thesis very weak.
Among the opposing elements are also mentioned the Right and
Left caste groupings. Stein has clarified the real nature of this dual
division in an excellent discussion of all the relevant material.” He
rightly treats them as relative or potential groupings of established
local groups rejecting the prevailing view that they were factions or
conflict groups. These groups became prominent only from the late
twelfth century even though they might have first appeared a century
earlier. They were moreover confined to certain localities outside the
central zone.”* Therefore, there is very little possibility of treating
the Right and Left caste groupings as basic segments or overarching
segments over a wide area. Thus the basic premises on which Stein’s
segmentary model rests are found to be flawed.
In his emphasis given to the locality chiefs Stein does not concede
any possibility of distinguishing ‘officials’ from chiefs. In his view, the
adikdris, the superior officials of the Chéla government, were only
lesser chiefs and did not have any administrative functions. Wherever
these and other higher officials are found in royal records they are
considered to be doing scribal functions producing ritual documents,
not bureaucratic orders.* This is in keeping with his other view
that the power of ‘officials’ (locality leaders according to Stein) was
not delegated by the king but was based upon their ties with local,
dominant peasantry from which they themselves originated.” It is
generally difficult to differentiate between the ‘ritual’ elements from
‘bureaucratic’ elements in the royal records due to the very nature
of the inscriptions, which are primarily documents of charities,
not of pure administration. Only some glimpses of the political
 Characterizing the Chola State 253

organization of the period can be obtained from such charity deeds.

In spite of this problem there are clear cases where we can certainly
distinguish officials from locality leaders. Stein does not try to sort
out the names of ‘titles of status’ from the names of ‘offices’. For
instance, mivéndavélan is the suffix component of a prominent title
of status whereas adikari is an office. It has been proved elsewhere
that there was a clear correlation between ‘offices’ and ‘titles’.2° In
the case of locality leaders no ‘offices’ are alluded to in inscriptions.
While locality leaders are confined to their localities, the official
personages are found to be highly mobile. For example, a royal grant
made in favour of a Vedic college established at Tirumukkudal in
Madurantakam taluk of Chengalpattu District was finally executed
by a nadu-kiru-aditari who hailed from a place near Nagapattinam
in Chéla-mandalam and by another subordinate official (designation
not given) who came from a place near Chennai in Jayangondachdla-
mandalam.” Most royal inscriptions would corroborate and add to
this evidence. Of course, Stein is well aware of such evidence but
he is not prepared to accept its implication. * There was definitely
some hierarchy, though shallow, in the structure of the bureaucracy.”
Clear administrative roles of many officials are also known,
though the roles were not as rationally defined as would fit in the
Weberian model.
There is substantial evidence to assert that the king’s government
could and did collect major taxes from most localities. Stein questions
the tax transfer from localities to the government on the ground that
most taxes, even the frequently appearing taxes, were paid in kind.*°
He underlines the fact, quite rightly, that none of the earlier scholars
has attempted to show how the kadamai in kind was converted into
money income for ‘the central government’.*! There is no denying of
the fact that the economy of Chéla times was still predominantly a
natural economy even though coinage, a state monopoly, had become
quite prevalent during the eleventh century and after. Though
large land transactions were made using gold bullion or money,
ordinary day-to-day transactions were carried on only in terms of
grain.” Inscriptions are certainly silent about the mode of transfer
and storage of the government revenue collected in grain. Some
mechanism must have existed for this, which is not explicit in the
inscriptional record. Each basic revenue unit (nddu for instance) may
have possessed a granary. In fact, a Tanjavur inscription of Rajaraja |
refers to the local granary (ullur-pandaram) and also the nadu-level
254 South India under the Cholas

granary (nattu-pandaram) of the Big Temple.” It may be remembered

here that Rajaraja assigned to this temple the government revenue
(mostly in paddy) from a number of villages spread over the entire
Kavéri delta. Though this and some related questions require more
intensive study, official involvement in taxation for government
cannot be disputed. This is the one field for which we have quite
impressive evidence. There was in fact a separate department
of revenue.
Stein categorically denies the existence of any standing army. He
thinks that the earlier existing non-royal military units controlled
and led by the dominant peasantry were recruited to the military
adventures of the Chélas.*4 Perhaps Stein bases his idea for the
peasant militia on the supposition that the Choha military ‘regiments
were named after the valanddu (territory) from which they came.”
This supposition is certainly wrong and is even contradictory to
what he himself has understood, correctly of course, of those names
as named after the ‘many pseudonyms taken by Rajaraja’.°° There
is certainly no regiment that is named after a valanddu territory.”
Stein relies also on another piece of evidence, that is a 1073 CE
inscription of Kulottunga I, copies of which are found at two places
in Kolar District, which would suggest some close association of the
valangai regiment with the periyanattar body. But the interpretation
of this inscription is problematic.** Even if for the nonce we accept
the existence of peasant militia on this evidence, it is found almost
at the end of the imperial wars. Actually, the peasantry grew
powerful subsequent to the time of the imperial wars. Further,
there is no evidence for the active functioning of supra-local peasant
organizations like the periyanattar in the central Chola area, that is,
the Kavéri delta. Rather, the periyandattar is found active only in the
peripheral areas during the twelfth and thirteenth centuries, which
was indicated above. It is during these latter centuries that the supra-
local mercantile organization, the Ayyavole-500 is also found to
possess its own contingents of guards.*” On the other hand, during
most part of the eleventh century, the time of imperial expansion,
the royal army was definitely present in the countryside. As shown
elsewhere, military personnel, from the commander-in-chief at the
top to the soldier at the bottom, were delegated powers to perform
many civil functions.*° The armies of the locality chiefs practically
disappeared during this century. So Stein either fails to take note of
certain obvious facts or ignores them in his enthusiastic application
 Characterizing the Chola State 255

of the segmentary concept to the Chéla state. A proper historical

perspective is also lost sight of in his treatment of the entire Chola
rule as one time unit which obscures the agrarian changes which he
himself has highlighted in many places.*!
Another consideration of the Chdla state is that of Kathleen
Gough.** This Marxist scholar considers the Chéla state as an
‘Archaic State’, which itself is the second of the three stages within
the state level of socio-cultural evolution, namely, the Early State, the
Archaic State, and the modern industrial Nations State. The three-
stage scheme, according to Gough, is actually a modified version of
that developed by Darcy Ribeiro.** According to Gough the Early
and Archaic states were characterized by a social formation which
Marx called the Asiatic Mode of Production (AMP). Gough also
includes feudalism, particularly of Western Europe and Japan, as a
special type under the Archaic State stage. The Chola society is said
to have corresponded to the Archaic State model coupled with the
AMP in many respects. Briefly they are:
(i) Little or no private property, all land being held nominally under the
authority of the monarch with the ‘owners’ holding merely the right of
possession at king’s pleasure; (ii) land tax and rent were identical, they
being the ‘upper share’ of the produce; (iii) the ruling strata and army
and religious institutions enjoyed the ‘upper share’ of specially assigned
prebendal estates; (iv) the majority of villages were held in joint possession
by communes of peasants or Vellalas; (v) prevalence of general slavery
associated with wet rice cultivation; (vi) large-scale irrigation works made
and controlled by the government; (vii) generally self-sufficient villages
which relied on the government only for irrigation, military protection
and large religious spectacles; and (viii) though the king was a despot in
theory, his power was in fact limited by the virtual self-sufficiency of the
village ‘republics’.
In order to avoid the usual criticism that the AMP model implied a
stagnant society, Gough suggests that a modified and refined model
of AMP would allow for greater social change, social stratification,
and commodity production which are actually evidenced in the
Tanjavur area, the central part of the Chola state. Gough's paper is
certainly of wider theoretical interest as it argues that there was an
AMP in the Hindu kingdoms of most of southern India prior to
foreign conquest. But the question will arise whether one should
retain the AMP with so many qualifications which make it very
different from and contradictory to Marx’s original proposition.
256 South India under the Cholas

Gough's overall understanding of the social and agrarian organization

of the Chola times as one characterized by sharp differences between
land-controlling ruling class and the large serf-like, sharecropping
landless cultivators has to be appreciated. But her empirical model of
Tanjavur of the Chéla times is bristled with much misinformed data
taken from an outdated study of the area, namely, that of Gupta.”
Her assertions that there was no private property and that there
was no distinction between land tax and rent cannot be maintained
now.“ There are many other sweeping statements like ‘(i) The
“slaves” were allotted by the state to the villages ... (ii) The merchants
and artisans were heavily taxed and under state supervision, etc.’,
for which there is no supporting evidence. It seems she has drawn
many such conclusions or assumptions taking the whole corpus of
pre-British evidence as one unit without temporal analysis, even
though she tries to contrast the socio-economic formation of the
post-Chéla times with that of the Chdla times. Lastly, though she
is not in agreement with Burton Stein’s criticism of the concept of
centralized bureaucratic Chéla state expounded by Nilakanta Sastri
and others, her own characterization of that state as a theocratic state
looks closer to Stein’s ritual polity. And she very cleverly glosses over
her contradictory stand in the following statement: “The problem of
combining a strong central government with largely self-sufficient
rural regions disappears if (as I think we must) we regard the temples,
monasteries, and other holders of prebendal estates as branches of the
government rather than as separate from and opposed to it’.*”
Let us consider whether the Chola State may be taken as an ‘Early
State’ as defined by Henri J.M. Claessen and Peter Skalnik. These
scholars have provided a big array of structural characteristics of the
Early State after analysing twenty-one case studies. The Early State is
defined as a centralized socio-political organization for the regulation
of social relations in a complex stratified society divided into at
least two basic strata, or emergent classes, namely, the rulers and
the ruled, whose relations are characterized by political dominance
of the former and tributary obligations of the latter, legitimized by
a common ideology of which reciprocity is the basic principle.**
The salient characteristics of the Early State may be summarized as
follows:* It has a predominantly agricultural economy, supplemented
by trade and a market system. There are full-time specialists. The
access to material resources is unequal. The upper (ruling) stratum
generally has tribute as its main source of income. Tax, however, is
 Characterizing the Chola State 257

paid by all social categories. The position of the sovereign is based

upon a mythical charter and genealogy which connects him with the
supernatural forces. The ideology of the early state is based upon the
concept of reciprocity; all categories of subjects provide the sovereign
with goods and services (tribute and tax), while the sovereign for
his part is responsible for his subjects’ protection, law, and order,
and the bestowal of benevolences. The priesthood supports the state
ideology. For the government a system of delegation of tasks and
powers is evolved. There are numerous functionaries fulfilling tasks
in the governmental apparatus. The organization of the early state
reveals a tendency to function syncretically, implying that most of its
functionaries fulfil more than one task. Most state activities appear to
be multi-purpose. The government is oriented towards centralization
and the establishment of centralized power. The government devotes
much of its attention to its own legitimation. The sacral character of
the ruler’s position is the most important constituent of his leadership.
He is considered as the guarantor of the state’s prosperity. There is a
tendency on the part of the emergent class of the ruling stratum to
live more or less apart from the other citizens of the state. To provide
for the survival of the early state, pre-state patterns will have to be
either suppressed or converted into institutions of the early state. The
early state is thus illustrative of a continuous process in its regulation
of the relations between the emergent social classes of the rulers and
the ruled.
The authors of the ‘early state’ divide it into three types, namely,
inchoate, typical, and transitional on the basis of such criteria as the
degree of development of trade and markets, the mode of succession to
important functions, the occurrence of private ownership of land, the
method of remuneration of functionaries, the degree of development
of the judicial and taxation systems. The first two types, inchoate and
typical, would go with the AMP as they lacked private ownership of
land and the imposition of taxes on the common people was based
for the greater part on their allegiance to the sacral ruler, and justified
by vestiges of reciprocity.
There are obviously some basic similarities between the Archaic
State of Darcy Ribeiro/Kathleen Gough and the Early State just
mentioned. The latter is described more elaborately with a number
of defining structural characteristics. Many of these characteristics
may be recognized in the Chola State: the pivotal role of the king,
a complex, stratified society divided into at least two basic strata,
258 South India under the Cholas

namely, the ruler and the ruled, presence of full-time government

functionaries doing more than one task at a time, government revenue
accruing from both tax and tribute, and government tending towards
centralization and devoting much of its activities to achieve its
own legitimization.
As the characteristics of the Early State are too numerous and too
general, some of those characteristics can be recognized in any pre-
modern state. Even then, the Chéla state cannot fit in exactly into
one of the three types of the ‘early state’ as some specific characteristics
of all the three types are found mixed in the Chola state. While
private ownership of land and a regular taxation system which are
two of the specific criteria for the ‘transitional state’ are found in the
Chola state, the other associated criteria like salaried functionaries
who were mostly appointed, codification of laws and punishments,
formal judges, and so on, were absent here. In the Chola government
(particularly during the middle phase), even though appointment
of functionaries to high offices may be guessed, the principle of
heredity also would have played a greater role.® It may be noted
that the ruling stratum of the imperial phase was drawn mostly from
the landholding Vellala and Brahman castes and to some extent
from the rich trading castes. Further, supra-local level trade and
markets, which are considered to be the criteria of the typical
and transitional states, were found in the Chéla country during the
last phase (Periods 3 and 4). But at that time the Chola state could
not command them due to its otherwise weak character. From such
divergences in the observed phenomena we have to say that the three
types of early state may not be found in a strict historical order. And
so it is difficult to explain the historical process of the Chéla state by
applying the classificatory criteria of the ‘early state’.
Notes
1. K.A. Nilkanta Sastri, The Colas, 2nd edn, 1955 (1st edn, 2 vols, 1935-7),
University of Madras, Madras, p. 462.
2. Burton Stein, “The State and the Agrarian Order in Medieval South India: A
Historiographical Critique’, in Burton Stein (ed.), Essays on South India, University
of Hawaii, 1975, pp. 64-91; Burton Stein, “The Segmentary State in South Indian
History’, in Richard C. Fox (ed.), Realm and Region in Traditional India, C. Fox,
Vikas Publishing House Pvt. Ltd., New Delhi, 1977, pp. 1-51.
3. Buston Stern, Peasant State and Society in Medieval South India, Oxford
University Press, New Delhi, 1980.
4. Ibid., pp. 280-1.
 Characterizing the Chola State 259

5. Ibid., p. 274.
6. Ibid., pp. 114, 271.
7. John Middleton and David Tait (eds), Tribes Without Rulers: Studies in African
Segmentary Systems, Routledge and Kegan Paul, London, 1970, p. 77.
8. Stern, Peasant State, p. 271.
9. Ibid., pp. 285fF.
10. Ibid., p. 321.
11. Ibid., pp. 339, 357.
12. Ibid., pp. 190-1.
13. Ibid., p. 284.
14. Ibid., pp. 258-64.
15. George W. Spencer and Kenneth R. Hall, “Toward an Analysis of Dynastic
Hinterlands: The Imperial Chélas of the 11th Century South India’, Asian Profile,
2 (1), February 1974, pp. 51-62. Another similar study is Thomas R. Trautmann,
etal., “The Study of South Indian Inscriptions’, in Robert E. Frykenberg and Pauline
Kolenda (eds), Studies ofSouth India: An Anthology ofRecent Research and Scholarship,
New Era Publications, Madras, 1985, pp. 1-29.
16. Stern, Peasant State, p. 271.
17. South Indian Inscriptions (SII hereafter), vol. ii, nos 94-5. There is another
interesting and rare piece of evidence relating to the wives of a big official called
Ammaiyappan alias Pallavarayan who hailed from Palaiyanur in North Arcot District.
His four wives belonged to four different places which seem to lie in different nadus,
his mother belonged to another village and his sister was married to one residing in
yet another village (Epigraphia Indica (EI hereafter], vol. xxi, no. 31).
18. See Chapter 16: sections on Territory and Socio-economic Organization.
19. Stern, Peasant State, pp. 135-6.
20. See Chapter 10 and Noboru Karashima and Y. Subbarayalu, “The Emergence
of the Periyanadu Assembly in South India during the Chéla and Pandyan Periods’,
International Journal ofAsian Studies, | (1), 2004, pp. 87-108.
21. Stern, Peasant State, p. 136.
22. Ibid., pp. 173-215.
23. See Chapter 13 in this volume.
24. Ibid., pp. 357-61.
25. Ibid., p. 117.
26. See Chapters 4 and 16 in this volume.
27. El, xxi, no. 38.
28. Peasant State, p. 359.
29. See Chapter 16, section on Officialdom.
30. Stern, Peasant State, pp. 258-64.
31. Ibid., p. 259.
32. Nilakanta Sastri, The Colas, pp. 559-62.
33. SII, vol. ii, nos 68 and 69.
34. Peasant State, p. 189.
35. Ibid., p. 333.
36. Ibid., p. 189.
37. Stein seems to have been misled in this regard by B. Suresh Pillai, “The
260 South India under the Cholas

Rajarajeesvaram at Tancavur’, Proceedings ofthe First International Conference Seminar

of Tamil Studies, Kuala Lumpur, 1964, 1ATR, 1968, pp. 437-50.
38. Stern, Peasant State, p. 191. Stein has discussed this inscription elaborately to
emphasize his point that the valangai (vélaikkarar) army was the military organization
of the Coromandel peasantry. He also takes the authors of this record as the dominant
peasantry from Tondai-mandalam (that is, Jayangondachdla-mandalam) who
colonized the Kolar area in the wake of the Chéla conquest in the eleventh century.
(Peasant State, pp. 124-7, 190-5). Though the colonization thesis may be plausible,
the other related discussion cannot be sustained as it is based on a faulty text and
translation of the inscription. Moreover, this inscription has to be interpreted in
the light of a similar, but earlier, inscription found at Tamaraippakkam, where the
valangai army was not present. See Chapter 10, section on Periyanattar.
39. See Chapter 15 in this volume.
40. See Chapter 16, Sections on Military and Police, and Judiciary.
41. Some of these points are also discussed from different angles by the learned
reviewers of Stein’s book: R. Champakalakshmi, ‘Peasant State and Society: A Review
Article’, Indian Economic and Societal History Review, vol. xviii, 1981, pp. 411-36;
Vijaya Ramaswamy, “Peasant State and Society in Medieval South India’, Studies
in History, vol. iv, no. 2,1982, pp. 307-19; D.N. Jha, ‘Validity of the Brahmana
Peasant Alliance and the Segmentary State in Medieval South India’, /ndian Historical
Review, vol. viii, nos 1-2, 1981-2, pp. 74-94; Noboru Karashima, South Indian
History and Society, 1984, pp. xxiv—xxviii; M.G.S. Narayanan, ‘Presidential Address:
Socio-Economic History—New Trends, A Case Study of the Chéla Times’, South
Indian History Congress, Mysore, 1985.
42. Kathleen Gough, ‘Modes of Production in Southern India’, Economic and
Political Weekly, 34, Annual Number, 1980, pp. 337-54; See also Kathleen Gough,
Rural Society in Southeast India, Cambridge University Press, Cambridge, 1981, pp.
105-13.
43. Darcy Ribeiro, The Civilizational Process, Trans. by Betty Meggers, Harper
Torch Books, 1968.
44. Ibid., pp. 343-7.
45. K.M. Gupta, The Land System ofSouth India between About 800 and 1200 ap,
Motilal Banarsidass, Lahore, 1933.
46. See Chapter 7 in this volume.
47. Kathleen Gough, Rural Society, p. 424, note 4.
48. Henri J.M. Claessen and Peter Skalnik (eds), The Early State, Mouton, The
Hague, 1976, p. 640.
49. Ibid., pp. 637-9.
50. See Chapter 4 in this volume.