Indian Economic & Social History

Review
http://ier.sagepub.com/

State formation in South India, 850 -1280

James Heitzman
Indian Economic Social History Review 1987 24: 35
DOI: 10.1177/001946468702400102

The online version of this article can be found at:

http://ier.sagepub.com/content/24/1/35.citation

Published by:

http://www.sagepublications.com

Additional services and information for Indian Economic & Social History Review can be found at:

Email Alerts: http://ier.sagepub.com/cgi/alerts

Subscriptions: http://ier.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

State formation in South India, 850 -1280
James Heitzman
Villanova University, Pennsylvania

The present study utilises inscriptional records from the Chola period (AD
849-1279) in South India in order to address several major problems con-
nected with Chola kingship in particular and with early Indian polities in
general: (1) What were the mechanisms evolved by kings to exert their
authority over relatively large areas and disparate human groups? (2) To
what extent did the authority of kings vary over distance and in relation to
local conditions of geography and socio-economic organisation? (3) What
was the relationship between royal political systems and local or intermediate

powers?
I will describe the early state as a problem in intermediate authority, that
is, an attempt to unite large numbers of localised social and economic units
within varying ecological niches into overarching systems of political control
through the creation or remodelling of mediating social institutions or
communications. The goal of this description is a view of kingship as an
agency representing the dominating classes of an agrarian society but
challenging, through its impact on political and economic organisation, the
positions of dominant groups. The methodology employed for this study
utilises descriptive statistics and strict controls over variables of time and
space in order to relate changes in state institutions at the highest levels to
changes in the productive and extractive modes operating in local
environments.
The study addresses three models of state formation-bureaucratic,
segmentary, and feudal-that portray the state in pre-modem South Asia.
These models focus on the relationship of state institutions to local and
intermediate arenas of power, and their ultimate relationship to forms of

Acknowledgements: Many thanks to David Ludden and Carol Breckcnridge, wh ~ read earlier
drafts of this paper and contributed
helpful comments.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

36/

production and control of resources in an agrarian world.’ The models

disagree the nature 6f the relationship between royal administrative
on

organs, however defined, and intermediate or local power-holders. Earlier

scholars stressed central bureaucratic power,2 the students of feudalism
have stressed independent, intermediate political offices,3 and proponents
of the segmentary state have stressed’local authorities.’ Each model takes its
position within different approaches to time and space. The bureaucratic
approach is the most static, since it posits little structural change, or
motivation for change, from the time of the Mauryas (fourth-third centuries ,

Be) until the ’Muslim invasions’ of the eleventh century and beyond. The
1
Herman Kulke has discussed the three main models of pre-modern South Asian historio-
graphy in ’Fragmentation and Segmentation Versus Integration? Reflections on the Concepts
of Indian Feudalism and the Segmentary State in Indian History,’ Studies in History, Vol. 4,
No. 2 (1982), pp. 237-54. See also James Heitzman, ’Socio-Economic Formations in Premodern
South Asia: Case Studies and Methodology, Peasant Studies, Vol. 13, No. 1 (Fall 1985), pp.
47-60.
2
R.C. Majumdar ed., The History and Culture of the Indian People, Vol. 2, Bombay, 1951,
pp. 303-34; K.A. Nilakantha Sastri ed., A Comprehensive History of India, Vol. 2, Calcutta,
1957, pp. 50-66; Anant Sadashiv Altekar, The Rāshtrukūtas and Their Times, Poona, 1934, pp.
174-76; Krishna Murari, The Cālukyas of Kalyāni (from circa 973 A.D. to 1200 A.D., Delhi,
1977, pp. 191-98; Romila Thapar, Asoka and the Decline of the Mauryas, Delhi, 1961, pp.
98-123. Burton Stein has summarised South Indian versions of this historiography in ’The State
and the Agrarian Order in Medieval South India: A Historiographical Critique,’ in Burton
Stein ed., Essays on South India, New Delhi, 1975, pp. 65-69; Peasant State and Society in
Medieval South India, Delhi, 1980, pp. 254-64.
3
Ram Sharan Sharma, Political Ideas and Institutions in Ancient India, Delhi, 1959, pp.
199-230; Indian Feudalism: c. 300-1200, Calcutta, 1965; Sudras in Ancient India: A Social
History of the Lower Order Down to circa A.D. 600, Delhi. 1980, pp. 171,192-93,243,260-61;
Perspectives in Social and Economic History of Early India, New Delhi, 1983. pp. 128-56; ’How
Feudal was Indian Feudalism?’ Journal of Peasant Studies, Vol. 12, Nos. 2 and 3 (1985), pp.
33-36; D. D. Kosambi, The Culture and Civilisation of Ancient India in Historical Outline, New
Delhi, 1981, pp. 192-98; Ganesh Prasad Sinha, Post-Gupta Polity (A.D. 500-750): A Study of
the Growth of Feudal Elements and Feudal Administration, Calcutta, 1972, pp. 198-219, B.N.S.
Yadava, ’Secular Land Grants of the Post-Gupta Period and Some Aspects of the Growth of
Feudal Complex in Northern India.’ In D.C. Sircar ed., Land System and Feudalism in Ancient
India, Centre of Advanced Study in Ancient Indian History and Culture, Lectures and
Seminars No. I-B, Calcutta, 1966, pp. 72-94.
4
Burton Stein, The Segmentary State in South Indian History,’ in Richard G. Fox ed.,
Realm and Region in Traditional India, Durham, S.C., 1977; Peasant State and Society in
Medieval South India, Delhi, 1980, pp. 101-09, 134-40, 173-82, 270-72; ’Politics, Peasants and
the Deconstruction of Feudalism in Medieval India’, Journal of Peasant Studies, Vol. 12, Nos. 2
and 3, pp. 61-65; ’State Formation and Economy Reconsidered,’ Modern Asian Studies, Vol.
19, Part 3 (July 1985), pp. 393-400; Kenneth Hall, ’Peasant State and Society in Chola Times: A
View from the Tiruvidaimarudur Urban Complex,’ IESHR, Vol. 18, Nos. 3 and 4 (1982), pp.
393-96; R. Champakalakshmi, ’Peasant State and Society in Medieval South India: A Review
Article, IESHR, Vol. 18, Nos. 3 and 4, pp. 411-26; Vcnkata Ragotham, ’Religious Networks
and the Legitimations of Power in Fourteenth Century South India: A Study of Kumara
Kampana’s Politics of Intervention and Arbitration in Madhu Sen ed., Studies in Religion and
Change. New Delhi. 1983, pp. 150-66; David Ludden, Peasant History in South India. Princeton.
1985, pp. 26-40.214.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 /37

segmentary state of Burton Stein contains explicit references to spatial

variability, since it originates and flourishes in ’nuclear areas’ of peasant
farming communities within irrigated zones in distinction to non-irrigated.
drier zones; ritual control becomes more tenuous with increasing distance
from the nuclear zones, exhibiting central, intermediate and peripheral
forms of political authority. The dynamics for systemic change over time
are not as clear as spatial dynamics in the segmentary mudel. Feudalism, on
the other hand, offers a clear-cut chronological progression, from the crisis
of an ancient socio-economic and political formation in the carly Christian
era through subsequent fragmentation of state rights. The feudal hypothesis
remains weaker in its spatial aspects, since ’feudal’ relations may be described
almost anywhere in early South Asia, poorly articulated with local modes
and relations of production.6 6
In the following discussion, I will address the central problematic of
historical models of pre-modern state formation-the articulation of central,
intermediate and local authority-within parameters that explain spatial
variability and temporal change. The Chola dynasty is the choice for this
case study because the numerous records of the Chola period, amounting to
over 10,000 inscriptions, provide data that allow statistical analyses impossible
in most other areas of early South Asia for which data are scarcer and more
fragmentary. The rich Chola-period data may produce insights that may
then be extrapolated to other areas of South India or South-Asia as a whole,
where historical sources are less plentiful for early times.

The Study Areas in the Chola Heartland

The sources for Chola-period history are inscriptions found almost entirely
on the stone walls of temple structures in Tamil Nadu and surrounding
states. These records describe gifts to Brahman communities and to temples
in order to support ritual performances. Because almost all inscriptions
contain brief or lengthy preambles mentioning the reigning king and his
regnal year, approximately 90 per cent of the records engraved during the
period of Chola hegemony may be dated with great accuracy.’7
5
Stein describes nuclear areas as nadus, or agrarian zones clustered around common
irrigation facilities and containing distinct social or kinship zones (Peasant State and Society,
pp. 90-140). His interpretations rest on the work of Y. Subbarayalu, Political Geography of the
Chola Country (Madras, 1973, pp. 19-55: ’The Cola State,’ Studies in History, Vol. 4, No. 2
(July-Dec. 1982), p. 273.
6
See suggestions of the ’proto-feudal’ character of the Chola State in M.G.S. Narayanan’s
review of Noboru Karashima’s South Indian History and Society, in IESHR, Vol. 22, No. I
(Jan.-March 1985), pp. 95-101.
7 We must note the peculiar nature of the inscriptional record, which is predominantly
limited to records of donations for deities. On one hand, the inscriptions are a goldmine of
information on contemporary social. political, and economic organisation, which emerges
obliquely, and in a pristine form, from deeds that ostensibly record religious devotion. On the

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

38/

For purposes of statistical aggregation, the inscriptions fall into four

sub-periods, each lasting about a century, and conforming to discernible
changes in the political fortunes of the Chola dynasty and the formats of the
records themselves. Sub-period one (849-985) saw the rise of the Chola
dynasty within the central area of modern Tamil Nadu, with their capitals at
Tanjavur and Palaiyaru within the Kaveri river delta. Most inscriptions
during this time were short, concentrating on gifts of perpetual lamps to
temples in memory of deceased relatives. Sub-period two (985-1070) began
with the accession of the greatest of the Chola kings, Rajaraja I, who
extended Chola military power throughout Tamil Nadu and over much of
peninsular India. Under his immediate successors, Chola military expeditions
travelled as far as the Ganga river, Southeast Asia, and Sri Lanka. Inscriptions
during this time became more detailed and often included poetic preambles
(prasasti, meykkirti) praising in chronological order the accomplishments of
the kings. Sub-period three (1070-1178) began with the accession of
Kulottunga I, who inherited the thrones of the Chola empire and the
kingdom of the Eastern Chalukyas in modern Andhra Pradesh. Despite the .
union of two major royal lineages, Chola military power entered a period of
slow decline. Inscriptions from this time contain increasing information on
land transactions. During sub-period four (1178-1279) the Chola dynasty
collapsed. The Pandya dynasty based in the southern town of Madurai, the
Hoysala dynasty from modern Karnataka, and local chiefs from northern
Tamil Nadu divided up the Chola realm. Inscriptions record increasing land
transactions and greater numbers of local leaders arrogating to themselves
high titles and local powers.&dquo;8
The present approach, building on the lead of Karashima and others,9
concentrates on five discrete study areas, each containing large numbers of
individual villages, each representing varying ecological, political and

other hand, the necessity of concentrating historical inquiry on temple records constantly
focuses attention on a forum that was certainly important, but probably only a single environment
within a much larger universe of human activity. Even within the world of temple records,
accidents of history have left us a limited view of the large number of social transactions
occurring in the environs of religious institutions. In the present study, then, we analyse
samples of samples that represent central historical processes but do not exhaust the potential
for alternative social formations. -

8
The standard histories are K.A. Nilakantha Sastri. The Colas: A History of South India
from Prehistoric Times to the Fall of Vijayanagar, London, 1958, pp. 173-210; T.V. Sadasiva
Pandarathar, Pirkalac colar varalāru, Annamalainagar, 1958-61. The four-part chronology
used here appears first in B. Sitaraman, Noboru Karashima, and Y. Subbarayalu, ’A List of the
Tamil Inscriptions of the Chola Dynasty,’ Journal of Asian and African Studies (Tokyo), No.
11, 1976, p. 89; Noboru Karashima, Y. Subbarayalu and Toru Matsui, A Concordance of the
Names in the Cōla Inscriptions, Madurai, 1978, p. xiv.
9
Noboru Karashima ed., Socio-Cultural Change in Villages in Tiruchirappalli District,
Tamilnadu, India, Part I, Pre-Modern Period, Tokyo, 1983; South Indian History and Society:
Studies from Inscriptions A.D. 850-1800, Delhi, 1984.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 /39

historical features characteristic of the Chola polity (see Map 1.). The study
areas lie within the traditional ’circle of the Cholas’ (Cholamandalam),
where we may expect the fullest possible record of political processes
featuring the Chola kings. (1) Modem Kumbakonam taluk lies in the very
centre of the Chola polity, near the capital of Palaiyaru, always controlled by.
the Chola dynasty. Its economy was, and is, oriented to the production of
rice through artificial irrigation systems dependent on the Kaveri river and
its effluents. (2) Tiruchirappalli taluk lies on the southern bank of the Kaveri
river, where lands benefiting from riverine irrigation have supported a rice
economy and, in the Chola period, a number of important temples headed
by sacred Srirangam. This area was a traditional part of the Chola homeland,
but Pandya and Hoysala influence became more important as the Cholas
collapsed after 1220.’° (3) Tirutturaippundi taluk lies on the ocean to the
southeast, at the tail end of the irrigation channels carrying Kaveri river
water. This area is a good example of a political backwater, always integrated
within the Chola empire but having no impact on political affairs. (4)
Tirukkoyilur taluk has a mixed economic base, with a zone of rice cultivation
concentrated along the banks of the Pennai river and around scattered tanks
farther away from the river, but with large expanses of poorly-watered lands
supporting the cultivation of millets and animal husbandry. This area had
the characteristics of a march or border on the northern edge of the Chola
heartland. In the tenth century it was briefly overrun by Rashtrakuta armies
from the northwest, and in the early thirteenth century it fell under the
control of the rebellious Kopperuficinkan, whose kidnapping of Rajaraja
Chola III signalled the collapse of the Chola political order.&dquo; (5) Pudukkottai
refers to the area of modem Pudukkottai district lying north of the Vellar
river, the traditional southern boundary of Cholamandalam. This southern
march of the Cholas had no natural access to major riverine sources for
irrigation and remains the scene of dry cultivation and agriculture centred on
scattered man-made lakes. The Cholas conquered this area from the Pandya
rulers in the late ninth century, but the resurgent Pandyas took over again
after about 1220.
The five study areas exhibit an ecological and agricultural continuum
ranging from Kumbakonam and Tiruchirappalli taluks, with good availability
of irrigation waters, through the more intermediate or mixed zones
exemplified by Tirutturaippundi and Tirukkoyilur taluks, to the relatively
dry zone of Pudukkottai (see Table 1).’2 Politically and. historically, the five
10
K.R. Venkataraman, The Hoysalas in the Tamil Country, Annamalainagar, 1950, pp.
7-18; J. Duncan, M. Derrett, The Hoysalas: A Medieval Indian Royal Family, London, 1957,
pp. 105-28; Nilakantha Sastri, The Pāndyan Kingdom, Madras, 1972, pp. 124-37.
11
Altekar, op. cit., pp. 115-19; S.R. Balasubrahmanyam, Kopperuncinkan, Madras, 1965,
pp. 14-15, 124-37; Nilakantha Sastri, The Cōlas, pp. 128-34, 417-44.
12
For historical characterisation of wet and dry zones in Tamil Nadu, see David Ludden,
’Patronage and Irrigation in Tamil Nadu: A Long-term View,’ IESHR, Vol. 16, No. 3 (July-
Sept. 1979), pp. 349-65; Peasant History in South India, pp. 81-94; Christopher John Baker,

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

40/

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 41

areas display a somewhat similar continuum from those zones closer to the
centre of the Chola political system and controlled by the Chola kings for
longer periods, to the outlying zones where political fortunes and distance
from the centre contributed to shorter periods of control by the Chola kings.

TABLE I
Ommavat= of Fr11f! Shady An-

The methodology for this study involved the collection and reading of all
extant Chola-period inscriptions from the five study areas, the compilation
of references to key terms relating to state formation within the collected
inscriptions, and the breakdown of the distribution frequencies of those
terms within the four sub-periods of the Chola period.&dquo; Changes in the
distributions of the terms within the five study areas and over time suggest
relative differences in political integration which may then be correlated
with the ecological, political, and historical characteristics of the different
areas. The terms chosen for analysis refer to the king and the royal family,
officials of the king, taxation, and control over local property rights.

The King and the Royal F’antily

The kings personally appear very rarely in the Chola-period inscriptions, In
a few records pious donors instituted rituals producing merit for the

An Indian Rural Economy 1880-1955: The Tamilnadu Countryside, Delhi, 1984, pp. 22-34;
Hans-Georg Bohle, ’The Cauvery Delta: An Investigation into the History and Determinants
of Agrarian Development and Rural Underdevelopment,’ Indian Geographical Journal, Vol.
58, No. 1 (1983), pp. 29-46; Bewasserung und Gesellschaft in Cauvery-Delta (Südindien),
Wiesbaden, 1981, pp. 10-38.
13
’Chola-period inscriptions’ are all extant records from the five study areas containing
references to Chola kings and/or regnal years that situate those records within the time period
849-1279. Also included within the data base are Rashtrakuta inscriptions and pre-1250
inscriptions of Kopperucinkan from Tirukkoyilur taluk, and pre-1250 records of Pandyan or
Hoysala kings from Tiruchirappalli taluk and Pudukkottai. The following discussions of the
royal family and officials utilise only records inscribed under the authority of the Chola kings.
The later discussions of tax and property terms utilise the entire data base.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

42/

well-being or success of the king, indicating that some localities were officially
concerned over the ruler’s illnesses or military adventures.&dquo; In one case the
great Rajaraja I appears on a tour of inspection through the Tiruchirappalli
area, and in several cases the kings personally intervened in the arbitration
of local disputes.&dquo; But if we expand the concept of royal presence to include
members of the Chola lineage, and especially the wives of the kings and
princes, we obtain a larger 16population of records describing the activities of
the royal house as a whole. The policies of the Chola house fall into the two
main categories of donations and royal orders.
Donations include a number of gifts to temples by the king or other
members of his family in order to institute and support brahmanical sacrifices
and/or the worship of the god Siva. Most of these donations were large gifts
such as ornate shrines, images and ornaments of precious stones and metals
to adorn images, funding for the expensive and elaborate consecration
ceremonies initiating temple worship. Even in this sphere the kings them-
selves appeared relatively infrequently, since queens or princesses usually
performed the honours. The ritual leadership of the royal family appears quite
dominant in the large-scale donations, when the ostentatious public support
of the moral universe, in the name of personal piety, was the explicit message. &dquo;I
Royal orders occur in altered contexts featuring direct royal interference into
local temple affairs. A standard scenario for the issuance of a royal order began
with the presentation of an official request (vinnappam) by a person be,-Uing
high honorific titles, typically asking permission for the deferment of taxes on
agricultural land and their transfer for the funding of rituals at a specific temple.
The king, hearing the request, granted permission for tax deferment and
ordered that the official transfer of taxes to the temple be entered in public
records. Some inscriptions appear to be word-for-word transcriptions of
these interchanges, indicating that the king was personally involved.&dquo;
14
23: 252; IPS 169; KK 128; ARE 1917: 280. Following are the compendia of published
SII
inscriptions used in this study, with their abbreviations: Annual Reports on Epigraphy (ARE);
Epigraphia Indica (EI); Inscriptions of the Pudukkottai State (IPS) Kutantai Kalvettukkal (KK);
South Indian Inscriptions (SII); Tirutturaippunti kalvettukkal (TK). All ARE references
indicateunpublished records viewed in transcript or estampage form in the Epigraphy Office in
Mysore. My thanks to Dr K.V. Ramesh, Chief Epigraphist, and his staff for their assistance in
reading unpublished inscriptions.
15
SII23: 310; ARE 1908: 468; 1914: 109; 1918: 1; 1927: 229, 231, 1931-32: 70, 71.
16
George W. Spencer, ’When Queens Bore Gifts: Women as Temple Donors in the Chola
Period,’ in K.V. Raman ed., Srinidhih: Perspectives on Indian Archaeology. Art and Culture,
Madras, 1983, pp. 361-73.
17
Examples of the kings’ donations include SII 3: 276; 6: 28, 33; 23: 52, 307; KK 18, 43.
Important examples of donations by female members of the royal family include SII 3: 96; 4:
542; 8: 237; 13: 197; 23: 42; KK 144, 147, 154; ARE 1908: 459. Donations by members of the
royal family outnumber donations by the kings 67 to 11in the five study areas.
18
SII 5: 723; 7: 1012; IPS 141, 153; TK 39. Related phenomena are inscriptions describing
the exact location of the king or queen during the issuing of a royal order: SII5: 706; 8: 223: 23:
48. 272. 351:
ARE 1921: 533; 1934-35: 184; 1937-38: 3.21; 1961-62:429.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 /43

The royal order was called the tirumukam, or ‘sacred face’ of the king; when
the written form of a royal order reached local villages for implementation,
village leaders greeted and reverenced it as if it were the king himself.’9
Descriptions of court procedures show that, in keeping with the quasi-sacred
character of the royal order, careful transcription and checking of the
written forms was standard procedure. 20 It seems certain that even those
inscriptions beginning with the simple phrase ’according to the royal order’
(tirumukattup pati) reflect the final, locally inscribed versions of a decision-
making process revolving around the king personally. It is doubtful whether
the ceremonial issuance of the royal order indicated in all cases royal control
over local resources, since in many instances the official request by local
leaders may have marked local initiatives with the simple acquiescence of
the king.2’ Nevertheless, the centrality of the ruler in these transactions
demonstrates a marked interest in, and ultimate control over, the resource
allocations associated with temples at the local level. In distinction to royal
donations, usually involving one-time gifts within a cash nexus, royal orders
reflect penetration into local agrarian economy.
Table 2 portrays the number and relative frequency of inscriptions con-
training references to either donations of the Cholas or royal orders of the
kings. The spatial distributions reveal a clear concentration of the activities
of the kings and their family members within the two central areas of
Kumbakonam and Tiruchirappalli taluks, for these two areas yield 134 out
of 167 total records, or 80 per cent of all references. The other more distant
study areas yield very few references, especially in the category of donations
(9 out of 68 records, or 13 per cent), suggesting that the kings and other
members of their families rarely went outside the central areas around the
Kaveri river. Changes in the distributions over time reveal a consistent
decline in references to donations and a consistent increase in references to
royal orders, a trend visible in all study areas. In Kumbakonam taluk, for
example, royal orders increased from only three in the tenth century to
twenty-one in the thirteenth, while donations declined from a high of
twenty-eight in the eleventh century to only two in the thirteenth.
The original dominance of donations early in the career of the Cholas may
mirror an emphasis on displays of ritual primacy in the construction or
consecration of temples and the patronage of brahmanical rituals, a policy
that peaked during the period of greatest Chola power in the eleventh
century. But the eleventh century also witnessed a shift toward greater
penetration of the royal will into local arenas of power through the issuing of
19
Village leaders, along with the person requesting the royal order, met the order when it
arrived, placed it on their heads to show obeiscence, and then circumambulated donated lands
with the royal order, mounted on a female elephant ( II 72; 23: 264; ARE 1931-32: 74).
S3:
20
SII
8: 222, 223; ARE 1931-32: 74.
21
Burton Stein, ’All the Kings’ Mana: Perspectives on Kingship in Medieval South India,’ in
J. F. Richards ed., Kingship and Authority in South Asia, Madison, 1978, pp. 136, 144-46.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

44

TABLE 2
DirtribfItioru of Refen!ItCG 10 DoNIIiora anIl Royal
Orden of 1M ChoIiø ill FIVe Study Anm

* Ratios refer to the number of occurrences in each sub-period to the total of all inscriptions
in that sub-period.

royal orders, an activity still surrounded by an ideology of ritual primacy but

entailing an active royal role in the allocation of resources at the local level.
This policy of local involvement continued to grow even as the Chola state
contracted in the thirteenth century.

The Officials of the Chola Kings

The identification of ’officials’ working for the Chola kings is a continuing

problem that revolves around the interpretation of personal names in the
inscriptions. As they appear in the surviving records, personal names follow
a typical south Indian format: first comes a place intimately associated with
the person, perhaps his/her native place or a village where property is
owned, then comes the father’s/husband’s name, then the personal name.
Additional names and terms usually follow the personal name as aliases,
including a variety of honorific terms denoting high status (ar3yan) or
modelled after the epithets of the Chola overlords (e.g. Rdjardja cola
muventave!än), and a separate series of titles that seem to denote functions..
Historical scholarship has tended to concentrate on all these additional
terms as markers of participation in state structures. Older studies assumed

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 /45

that all honorific titles indicated a role within the administrative system of
the bureaucratic state, an approach that has continued with modifications
until today.22 Burton Stein’s segmentary state model has suggested that all
additional terms were pure honorifics perhaps bestowed by the kings but
empty of administrative functions outside those already performed by locality
leaders.23 In the present study I will avoid these two extreme viewpoints and
concentrate instead on the more limited group of terms that more obviously
point to supervisory functions .2&dquo; These functional terms are divisible into
two general categories.
The first group of terms appear at first sight to denote persons whose roles
may well include participation within an articulated administrative frame-
work. A category of men performing the ’settlement of the nadu’ (natu vakai
ceykinra) in inscriptions of the tenth and eleventh centuries is reminiscent of
the more recent officials who compiled ’settlements’ of land revenue in
British India. In a typical example, the ’chief superintendant’ (kankcini
ndyakan) called together temple officials at Tirunamanallur to determine
the amounts due annually from several villages and assemblies, and the
requisite allocations for temple deities.2-’ There are also ’administrators’
(adhikäri) appearing at court and also in more outlying areas, taking care of
a variety of supervisory tasks, along with ’leaders of the army’ (senapati).
For example, a record from Tiruvidaimarudur describes the intervention of
adhikdri Cirrinkan utaiyan, performer of ’sacred work’ (srikdryam) at the
temple. He hears various arguments concerning the revenue obligations of
the local community, orders leaders to produce relevant documents, and on
the basis of these records adjusts the scale of allocations for worship. 26 Then
there are a number of persons associated with the retinue (parivaram) of the
king and the royal family (e.g. pa1)i makan, or ’work son’), concentrated near
the palaces but also appearing as donors in other areas of the Chola Heartland. 27
22
Krishnaswami Aiyangar, op. cit., pp. 264-72, 376-77; Nilakantha Sastri, The Colas, pp.
462-74; Sadasiva Pandarathar, op. cit., pp. 477-86; R. Champakalakshmi, op. cit., pp. 415-17.
23
Burton Stein, Peasant State and Society, pp. 257-58, 270-73.
24
Earlier studies in this vein are Karashima, et al., A Concordance of the Names in the Cola
Inscriptions, Vol. 1, pp. xlv-lvi; Y. Subbarayalu, ’The State in Medieval South India 600-1350,’
Ph. D. dissertation, Madurai Kamaraj University, 1976, pp. 134-63, 192; ’The Cola State, pp.
281-85, 288-91.
25
ARE 1939-40: 228. See also SII 7:988; 8: 580(a); IPS 90; Transactions of the Archaeological
Society of South India, 1958-59, pp. 84-110.
26
3: 202, 203; SII 5:718. See also
SII SII7: 1000; 23: 225; IPS 234; TK
ARE 1914: 6. M.T.
95;
Rajakumar has stressed the importance of these types of interventions in local decision-making
for political unification under sacred kingship: ’Kōyil porulitdrattil-aracin panku,’ Cenkai
māvattavaralārruk karuttaranku, ed. R. Nagaswamy, Madras, 1978, pp. 124-29.
27
Most often these persons are part of the ’entourage’ (parivaram) of the king or other members
of the royal family (SII 5: 706, 723; 8: 234). Many of these persons appear as donors of memorial
lamps after the death of prince Rajaditya in battle in 949 (SII 7: 954-66). Other palace servants
worked as dancers, waiting women or accountants in one of the palaces (velam) in a capital of the
Cholas. Some of the associates of the rulers appear as ’intimates’ (anukkan) or ’friends’ (saciva) of
the kings; these latter persons often claim ranks higher than those of other palace servants,
23: 243, 286, 339: ARE 1927: 336).
including titles of lordship (araiyan, nayakan, udaiyan) (SII

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

46/

A close inspection of the circumstances in which all these ’officials’ appear

provides a composite picture of their functions. When the officials do not act
simply as donors in their own right, they act as arbitrators of local disputes,
perhaps resolving problems concerning amounts due for the god from
particular plots of land. Typically, an official comes into a village, investigates
(aroy) relevant documents and hears relevant testimony, and in the presence
of local assemblies delivers a judgement that has the force of law. With very
few exceptions, all persons bearing functional titles who perform such
investigationsr--nadu settlement performers, adhikdris, senäpatis-seem to
act either on their own initiative or with a generalised fiat from the Chola
king, and have no apparent relationship with royal administrative machinery
or with standardised procedures. There was a body of persons bearing these
titles congregated at the royal court, and it appears that as local disputes
arose and came to the court’s attention from this floating body of
someone

loyal, honourable men would receive the commission to handle the

problem.28 The various performers of tasks for the royal family fit the
pattern of a household staff rather than a ramified administrative organisa-
tion. Investigation of these ’official’ terminologies reveals, then, little
indication of a ramified bureaucratic system for ruling the Chola state, but
rather an ’extended court’ peopled by high-ranking associates of the king,
including the creatures of the king, scions of other noble families allied to the
Chola dynasty, and close relatives of the royal family. Despite the rather ad
hoc basis for the interventions of these royal representatives, they do function
as arms for royal penetration into local affairs and therefore perform crucial
roles for the extension of royal influence outside the framework of a
centralised bureaucracy.
Later inscriptions include references to a quite different group of royal
representatives described as members of the ’land revenue department’
(puravu vari tinaik kalam). While the agents of the extended court performed
wide ranging actions; members of the land revenue department focus on one
objective: the recording of land measurements and the amounts of taxes due
from lands for the royal government or its chosen beneficiaries. Often the
department appeared in records featuring royal orders, and several long
inscriptions portray the role of the department in effecting the king’s will and
its articulation with other political structures. When the king issued an
order, ic was transcribed and then witnessed by several high-ranking persons
in the royal presence, and often by a host of adhikäris who happened to be
present. Only then the royal writ came to members of the land revenue
department, who in distinction to the rather undifferentiated mass of
adhikäris appeared as holders of specific, hierarchically organised offices
associated with the processing of the royal border and its transmission to the
locality it affected. Tasks included the overseeing of the department, copying
28
These courtiers are the ’overlords who have joined together’ (utan Kūttattu atikārikal
).

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 47

of the record, and affixing of the royal seal, and in several cases linked with
local agents or with the activities of other accounting departments in the
palace.29 The rationality of procedures and recruitment in the land revenue
department must not be carried too far; some of the stages in document
processing seem quite generalised and even redundant, and a majority of
members in the department appear as ’possessors’ (u4aiydr) of landed
wealth in their own right, thus members of the class and status group that
produced the adhikáris.30 Nevertheless, the land revenue department exhibits
systematic, centralised features-a specific topic of activity, a chain of
command with differentiated, hierarchised roles, and subordination to the
orders of the king-that mark this as a bureaucratic state organ.
Table 3 portrays the relative frequencies of inscriptions from the five
study areas yielding references to either extended-court or bureaucratic
agents of the Chola kings. The spatial pattern displayed here is familiar from
the discussion of the royal family above; the central study areas of
Kumbakonam and Tiruchirappalli generally provide larger numbers of
references to officials than do the outlying taluks. The two former areas
alone yield 159 out of 239 total records, or 67 per cent of all references. The
general pattern over time again resembles patterns seen for changes in the
activities of the royal family; agents of the extended court form the largest
category in the early Chola period, but tend to decline subsequently, while
the bureaucratic agents of the land revenue department appear more fre-
quently after the beginning of the eleventh century. In Kumbakonam taluk,
for example, twenty-six records of the extended court in the tenth century
declined to three records in the thirteenth, while references to the land
revenue department simultaneously increased from none to twenty-one.
The data in Table 3 portray the progressive replacement of more arbitra-
tional, occasional administrative policies by those involving direct penetration
into local economy and precise determination of royal rights and their
allocations through royal orders. During the early stages of Chola rule, the
more decentralised policies were in order, as the polity rested on the ritual
supremacy of a generally distant king. After about 1000, with the triumphs
of Rajaraja I, royal policies continued to stress ritual leadership but began to
subtly change the rules of the game by introducing the land revenue depart-
ment and royal orders more frequently into local arenas. Ritual, arbitrational
forms were thus the hallmark of early political integration, more centralised
forms were the result of later and more formalised central control. The data
from Tiruchirappalli taluk provide the exception that proves the rule.
29
SII 6: 34; 8: 222, 223; 23: 288, 389, 292; 24: 46, 142; ARE 1931-32: 74, 115; 1978-79: 291.
30
Y. Subbarayalu, The State in Medieval South India,’ pp. 151-52. For discussions of
’possessors,’ see Noboru Karashima, South Indian, History and Society, pp. 15-23, 26-30, 57-58.
The concept of ’bureaucracy’ used here conforms to Max Weber’s idea of
rationality under legal
authority in The Theory of Social and Economic Organasation, Talcott Parsons ed., New York,
1964, pp. 328-45.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

48/

T.BBLE33
Distributions of Referercres to the Extended Court and
dte Land Revenue Deparonett in Five Study Areas

*
Ratios refer to the number oi occurrences in each sub-period to the total of all inscriptions
in that sub-period.

Unlike the other study areas, where manifestations of the extended court
were decreasing by 1070, in this area they increased dramatically. this
phenomenon may be traced to dislocations connected with the accession of
Kulottunga I and concomitant incursions of Hoysalas and Western Chalukyas
into the Tiruchirappalli area at that time.3’ The renewed instability of the
west resulted in an increase in displays of ritual sovereignty around Srirangam,
followed later in the twelfth century by an increased presence- of the land
revenue department. The segmentary state and the ritual polity Juis appear
as stages in political development which the Chola dynasty took steps to
abolish as far as they were able, once their military control was firmly
c;stahlisheci. The extent of the changes they were able to effect depended,
however, on distance from the centre of the polity and on the military
strength of opposing monarchs.
Taxes and Tax Collectors

Terms generally described as taxes (vari) usually appear in lists at the end of
inscriptions which record the transfer of rights to the produce from land.
31
Nilakantha Sastri, The Colas, pp. 305-10.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 49

The purpose of such donations was to endow tax-free land for religious
institutions, and the tax lists provide the names of various cesses defrayed
for the purposes of religious sacrifice and worship. 32 The following discussion
pursues the analysis of tax terminology-important for an understanding of
any state’s financesr-by defining major categories of taxes, tracing changes
in the distribution of major tax terms over time, and finally studying the
agencies most likely to collect these cesses in different study ai eas.
Two main divisions of land cesses existed during the Chola period in
relation to the processes of production and distribution of agrarian produce.&dquo;
( 1 ) At the local level, a variety of duties were incumbent on the controllers of
lands to pay for the annual expenses associated with maintenance of irrigation
facilities and local processes of self-government within the villages. (2)
Beyond the local level there were demands from superior agencies for
proportions of agrarian produce, in turn entailing several kinds of exactions.
Land taxes called kadamai were generally paid in kind, according to schedules
that were, at least in the central area of the Chola heartland, determinable
by the land revenue department. Additional payments in cash or kind were
necessary to defray the expenses arising from the collection of land taxes,
especially for the temporary maintenance of collection agents. The two main
divisions of agricultural cesses found expression in the categories of the
’upper share’ (melviram) due to the superior agents and comprising land
taxes in kind and collection expenses, and the ’lower share’ (kijväram)
retained by the controllers of the land and used to pay all cultivating
expenses. 3,1
32
These lists of terms only rarely provide hints concerning the meaning of individual terms,
and much of the work done on these terms has concentrated on glossaries or, more recently,
large scale studies of their distributions; T. V. Mahalingam, South Indian Polity, Madras, 1954,
pp. 421-28; T.N. Subrahmaniam, South Indian Temple Inscriptions, Vol. 3, Madras, 1953-57,
Part 2, Annexure; D.C. Sircar, Indian Epigraphical Glossary, Delhi, 1966; Noboru Karashima
and B. Sitaraman, ’Revenue Terms in Chola Inscriptions,’ Journal of Asian and African
Studies, No. 5 (1972), pp. 87-117. See also Nilakantha Sastri, The Cōlas, pp. 520-45;
P. Shanmugam, ’Revenue System under the Cholas (850-1279 A.D.),’ Ph.D. dissertation,
University of Madras, 1977; R. Tirumalai, Studies in the History of Ancient Townships of
Pudukkottai, Madras, 1981, pp. 199-268, 357-58.
33
This discussion passes over a variety of commercial cesses—significant transactions in the
dynamic economy of south India during the Chola period—levied on or collected by mercantile
groups. In the five study areas, such commercial cesses appear rather rarely, concentrated
mostly in Tirukkoyilur taluk. See James Heitzman, ’Gifts of Power: Temples, Politics and
Economy in Medieval South India,’ Ph.D. dissertation, University of Pennsylvania, 1985, pp.
388-92. This concentration may relate to an extended commercial network focused on Kanchipuram
to the north. See Kenneth Hall, Trade and Statecraft in the Age of the Colas, New Delhi, 1980, pp.
83-97, 123-30; Kenneth Hall and George W. Spencer, ’The Economy of Kanchipuram, A Sacred
Center in Early South India,’ Journal of Urban History, Vol. 6, No. 2 (Feb. 1980), pp. 128-33
34
These terms are rarely mentioned in the Chola-period inscriptions of the five study areas,
but the existence of the broad divisions between upper and lower shares is well-known (and
debated) in a number of studies and in modem parlance: A. Appadorai, Economic Conditions
in Southern India (1000-1500 A.D.), Madras, 1936, pp. 171-78; Nilakantha Sastri, The Colas,
pp. 522-27; Leonid Alayev, ’The System of Land-Rights in Southern India (900-1300 A.D.).’ in

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

50/

Aside from isolated and random references to fees for various forms of
village government,&dquo; the largest number of local cesses concerned expenses
for the annual or occasional upkeep of irrigation facilities, so crucial for an
economy dependent on water for rice cultivation. Terms relating to irrigation
dues carried with them connotations of local collection and expenditure, at
times in cash but mostly in personal labour; the frequency of these terms
peaks in the records of the early twelfth century but declines thereafter.36 A
separate term for ’dues of cultivators’ (kudimai) refers to similar types of
local personal labour, but occurs more often in contexts suggesting collection
by superior agencies outside the village; this term is rare in the early Chola
period but steadily increases in frequency during later times. 3*1 Similar
increases occur for the more generalised term viniyogam, which contextually
includes the cesses for local personal labour with an element of supra-village
control (see Table 4).38 The data on labour dues suggest a twelfth-century
decline in local responsibilities in relation to an increase in the involvement
of supra-village agencies.
The single greatest demand for agrarian produce was the land tax, katjamai
or more rarely irai.39 It is currently impossible to determine the percentage
of the yield demanded as kadamai, but there was an attempt during the
eleventh century to standardise the evaluation and measurement of rice land
and to impose standard land tax rates in the areas around the Kaveri river

Proceedings of the Fifth International Conference-Seminar of Tamil Studies, Vol. 2, Part 14,
Madras, International Association of Tamil Research, 1981, pp. 17-22; R. Tirumalai, op. cit.,
pp. 204-06; Kathleen Gough, Rural Society in Southeast India, Cambridge, 1981, pp. 110-11, 152.
35
These terms range from cesses for village government (ür ācci, ür itu vari, ür kalancu) or
nadu government (nätäcci) to fees for plough teams (erp pon), marriage (kannālak kānam) or
legal expenses (vivastai). The terms are scattered randomly through time and space, and there
are not more than six references to any individual term. We may assume that such local cesses
were ubiquitous, but rarely crop up in inscriptions because they were rarely defrayed in
donation arrangements.
"
Terms for irrigation dues include several kinds of vetti and vetinai (often glossed as ’forced
labour’), kulai and kurampu (terms for embankments), nir (water) or āl (manual) amanci,
muttaiyāl and cutti.
37
Most occurrences of the term kudimai refer to labour services for local agencies, especially
temples or other public buildings (e.g. SII 24: 58, 94; ARE 1927: 211, 355; 1931-32; 89). One
instance records kudimai levied by the nadu (IPS 327). A number of references describe
kudimai going for ’the sacred victorious gate’ (tiruk korra väcal), meaning either the temple or
the king’s palace but probably the latter ( II 64; ARE 1917: 276). How or when labour
S24:
services for the king were enforced remains unclear, although periodic military or construction
labour seems plausible.
38
There are viniyogams for the village, the Brahman assembly, and agricultural castes (TK
132, 190), and viniyogam including a variety of local labour cesses ( T212;K ARE 1927-28: 205).
See also iräja niyogam (SII
17: 540) and viniyogam including kudimai and kadamai (TK 163).
39 The term kadamai rarely appears in a generic sense as a ’duty’ to pay all kinds of taxes (e.g.
I
S 24:
I 53).ai takes the place of kadamai at times in Kumbakonam taluk (e.g. 8:611;
r
I II ARE
S
1917: 227) and often in Pudukkottai (e.g. IPS 135, 151, 158, 190):
40
Noboru Karashima, South Indian History and Society, pp. 94-105.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 51

TABLE 4
Distribution of Major Tax Terms in Five Study Areas

*
The ratios refer tax term to total inscriptions from each sub-period.

During later centuries the growing land revenue department was concerned
primarily with the determination of kadamai rates on agricultural land of
various sorts. Occurrences of the term kadamai increase constantly in the
inscriptions of the Chola period. (see Table 4.) Parallelling this increase was
a rise in the occurrence of a general term for the ’large tax’ (peruvari) which
1
apparently applies to the cesses of the upper share or kadamai.4
Expenses for the collection of superior taxation fell into the early concept
of eccoru, a term referring to various supplies of cooked rice provided for
officials.°2 This term was prevalent during the early Chola period but steadily
declined over time (see Table 4). Simultaneously another term referring to
’intermediate income’ (antariyam) steadily increased in frequency,
contextually connected to several other words suggesting expenses of the
threshing floor and ratios of produce. The intermediate income was a tax in
cash coupled consistently with the land tax in kind, and determined by the
land revenue department as part of the share for the Chola kings. 43 The
decline of eccoru, a term connoting occasional provisions tendered in the
41
Whereas the term peruvari does not occur in the data base in contexts that explain its
meaning, references to the paired term for ’small tax’ (cilvari) often include descriptions of
other, local cesses comprised within it. The term ’lower tax’ (kïl irai) reinforces a parallel
between large taxes as part of the upper share and small taxes as part of the lower share (IPS 90;
KK 72, 73, 74; ARE 1931-32: 93).
42
P. Shanmugam, op. cit., pp. 39-40.
43
Heitzman, op. cit., pp. 395-97. For explicit references to royal collection of antaräyam,
6: 33; 23: 49; ARE 1911: 211.
see SII

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

52

village, relates to the r.3e ot antarayam, a cess linked to a more standardised

and centrally managed tax structure.
Table 4 portrays the frequencies of important tax terminology during the
four temporal divisions of the Chola period. The central changes in taxation
during that time are visible by treating the six niajor terms as three separate
pairs. Eccoru and antarQyam, referring to the expenses of superior tax
collection, show an obvious inverse relationship, with a steady decrease in
the frequency of the more nebulous and perhaps occasional eccoru cess and
a steady increase in the frequency of antarayam, a cess determ1ned by the
revenue department of the Chola kings. A similar relationship exists between
the aggregated terminology connected with local irrigation cesses and the
term kudimai which was more closely associated with collection by superior
stateagencies. As the term kudimai came into more general use, increasing
from one occurrence in the tenth century to thirty-nine in the thirteenth
century, the local terms declined after the late twelfth century. Meanwhile,
relative frequencies of the land tax (kadamai) and other general tax terms
(peru vari and viniyogam), officially the prerogatives of the Chola kings,
continually increased over time, occurring by the thirteenth century in 15

per cent of all inscriptions.

A consistent pattern emerges from the study of major tax terms during the
Chola period, pointing toward a greater concentration of tax-collecting
power in the hands of superior agencies and a decline in the importance of
cesses collected and officially controlled by village administrations. These

findings suggest an increasing penetration of the village environment by

outside agencies officially subordinate to the Chola kings. Comparison of
tax developments with the changes in the actions of the Cholas themselves
and the duties of their official representatives supports the view of ever
greater royal authority and control of local agricultural environments during
the time after around AD 1000.
Kumbakonam, Tiruchirappalli, and Tirutturaippundi taluks, all partic-
ipating to various degrees in the irrigation ecoiiomy connected with Kaveri
river water, provide no indications that agencies other than those of the
Chola kings (especially the land revenue department) were in charge of the
collection of taxes. Central royal control was tied to the presence of the kings
themselves and the concentration of their officials within the geographical
area of the Kaveri river basin. This centrality of state control is less visible in
the more outlying study areas.
The Tirukkoyilur area was the centre of operations for subordinate families
even at the dawn of written records, when Malaiyaman chiefs ruled there at
the beginning of the Christian era and defied the ancient Chola, Pandya and
Chera kings. During the late tenth century, as Rajaraja I consolidated the
empire, a family of Miladu chiefs came to power in Tirukkoyilur and
retained local prerogatives in return for service to their Chola overlords.
After 1070, the leaders of the Tirukkoyilur area called themselves

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 /53

fdalaiyaman, hearkening back to the glories of their more ancient

predecessors, and supported the Chola cause to the bitter end around
1250.&dquo; 1’hese latter Malaiyaman chiefs appear in numerous instances as
collectors of special ’protection’ taxes (på{iikåval), cesses almost unknown
from the other four study areas during the Chola period. 41 In addition,
several cases suggest that the Malaiyamdn family at times collected land
taxes (kadamai) and other cesses associated elsewhere with the superior
rights supposedly due to the kings.&dquo; It appears that the Chola kings made
arrangements with locally powerful families in the Tirukkoyilur area,
allowing a certain amount of local autonomy in finances in return for
military support and the handling of local administration.
In dry Pudukkottai to the south, some of the administrative and tax-
collecting activities elsewhere associated with agents of the king remained
the responsibility of the nattar, or assembly of nadu leaders. A large per-
centage of references to the nfufr within the five study areas comes from
Pudukkottai alone. despite the fact that this area has yielded relatively few
inscriptions.47 When the nå!!år met together in Pudukkottai, they sometimes
performed the arbitrational activities of adhikiris or allocated taxes in the
manner of the land revenue department.48 But within Pudukkottai they did
not always act in this way. Larger percentages of inscriptions refer to their
activities in the tenth and the thirteenth centuries, with a distinct fall in
references during the eleventh and twelfth centuries. 49 It is surely no
coincidence that the periods of their decline correspond to the times of
greatest Chola power, while the periods of their greatest visibility were times
when royal authority was most tentative. The Chola kings and their officials
extinguished the primacy of local assemblies when they consolidated power
in the south, but as the Chola kings and their administrative organs retreated
the local assemblies bounced back to establish their own local dominance.
It is possible now to discern three types of supra-local political authority
competing with varying success for tax revenues extracted from the village
economies. At times of imperial weakness, dry and tank-irrigated zones,
with their discrete and fragile economic bases, supported nattar assemblies
44
S. Srinivasan, ’A Study of the History of Tirukkoyilur down to A.D. 1600,’ Ph.D. disserta-
tion, Karnatak University, Dharwad, 1980, pp. ff. 117-19, 147
45
Y. Subbarayalu, ’The State in Medieval South India, pp. 53-54; P. Shanmugam, op. cit., pp.
60-61: Karashima and Sitaraman, op. cit., p. 91. Within the inscriptions of the five study areas,
references to pādikāval number 54, with 47 occurrences from post-1070 Tirukkoyilur taluk.
46
ARE 1906: 158; 1934-35: 135, 186, 190.
47
Total numbers of Chola-period inscriptions from the five study areas are as follows:
Kumbakonam taluk—743; Tiruchirappalli taluk—541; Tirukkoyilur taluk—631; Tirutturaippundi
taluk—231; Pudukkottai—245. References to the nāttār in the data base total 43; references to
the nāttār from Pudukkottai alone number 21, or 49 per cent.
48
17:
SII 462, 540; IPS 125, 285.
49
Thirteen total references to nāttār occur in records pre-dating 985, seventeen references
occur in records post-dating 1178. The time of greatest royal power (985-1070) yields only five
references to nāttār.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 54/

which balanced local political tensions and performed arbitrational functions.

The limited zones of riverine irrigation around the Pennai river supported
the lineages which acted as little dynasties, complete with ancient legends
and the poetic praises surrounding royal famihes.10 The big irrigation systems
within the Kaveri river basin supported the imperial Cholas, the biggest
kings. The Cholas were so big that their military power and ritual pre-
eminence overshadowed the outlying areas of central Tamil Nadu. They
destroyed for some time the effectiveness of discrete ndndr assemblies and
reduced to subordination the smaller ’kings’ of the smaller river systems.
The boundaries of these three types of political institutions were never fixed;
the outlying parts of the Chola country experienced different mixes of
centralised, chiefly or nattar administration, depending on the energy of the
centre and the resistance of the localities. In this way the Chola agents and
the nå!!år acted side-by-side in Pudukkottai, or Chola agents, nattar and
Malaiyaman chiefs acted side-by-side in Tirukkoyilur taluk. The general
tendency seen in the variables of royal activity, official presence, and
important taxation favours a period of increasing central dominance from at
least 1000 to 1150, as morse. ’segmentary’ or ’feudal’ political organisations
succumbed to royal dominance. After 1150, the political forms typical of
local ecology and economy reemerged as significant arbiters of revenue
allocation and political power.

Changing Property Relations in the Chola Heartland

Necessary causes of the Chola collapse were invasions by external enemies

(the Hoysalas and Pandyas) and rebellions of chiefs in the northern parts of
the Tamil country. But were these sufficient causes for the decline of the
Chola state? The evidence indicates that the Cholas were engaging in ever
more effective involvement in local economic spheres and the manipulation
of tax revenues as time went on. Royal orders and central taxation generally
increased in all study areas until the thirteenth century, although con-
centrated as usual in the central areas of the empire. If the dominance of the
Chola polity rested on the ability of the kings to direct or mediate the
allocation of local resources, then the very moment of its collapse was
paradoxically the time of its greatest central authority. The power base that
had allowed the Cholas to dominate their neighbours for so long failed them
against those same enemies. Some weaknesses in that power base may have
underlain this failure.
Evidence from the inscriptions concerning property relations holds a key
to an understanding of changing local power during the Chola period and the
infrastructural developments that actually weakened royal control as it
50
R. Nagaswamy, Tirukkōyilūrppāttu, Madras, discusses the extended praises of the land
and rulers of Tirukkoyilur in SII
7:863.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 /55

expanded. The evidence concerns frequency distributions of a crucial term

for control over property-kani.
The term kani, related to the verb ’to see’ (kan), has immediate connota-
tions of overseeing some right or thing, and in later times referred to
’possession, right of possession, hereditary right.’&dquo; Within the five study
areas, the term occurs within three main contexts: (1) Occupational kani
describes a situation in which property or possession entailed performance
of a specified duty within a village or a temple. The relatively few extant
references to this type of right point to a general custom of granting village
lands as the property of occupational specialists-accountants, security
personnel, musicians, doctors-conditional upon their perfornnance of their
duties. -12 (2) Kani in temples refers to the enjoyment of properties or
prerogatives in connection with membership in a temple staff. The per-
formance of rituals or administrative tasks in a temple depended on support
of personnel through grants of temple land which remained the personal
property of the holders as long as they performed their stipulated temple
functions. 13 (3) Kdni in land appears in three forms: (a) Donors possessed
their own kani land, inherited it, or purchased it from third parties, or
alienated it to religious institutions with all rights to cultivation..54 (b) The
land tax in kind is called kani kadan, or the ’dues from kdni,’ indicating that
royal cesses were due from land that was officially possessed by persons or
corporate bodies.&dquo; (c) Temple lands were ’property of the holy name’
(tirun4mattuk kini), obtained at times through alienation of the rights of the
donors described above.sb
Several aspects of the term kdni indicate that it refers to the private
property of individuals. Donors had the right to inherit, bequeath, alienate
or subdivide their kdni land. Possessors of kani in temples, typically

Brahman, at times seem to have manipulated their land as their own private
51
Tamil Lexicon, p. 859; Noboru Karashima, South Indian History and Society, p.18. Other
recent discussions of the term kāni are Peter A. Granda, ’The "Gift after Purchase" in
Vijayanagara Inscriptions,’ Journal of the Epigraphical Society of India, Vol. 6 (1979), pp.
25-31; ’Property Rights and Land Control in Tamil Nadu: 1350-1600’, Ph. D. dissertation,
University of Michigan, 1984, pp. 89-110,142-44,221-22,409; Dharma Kumar, ’Private Property in
Asia? The Case of Medieval South India,’ Comparative Studies in Society and History, Vol. 27,
No. 2 (April 1985), pp. 340-66; David Ludden, Peasant History, pp. 38-39, 85-88.
52 In the five study areas, there are twenty-five total instances of this type of kāni holding. For
more extended discussions of all types of kāni, see Heitzman, ’Gifts of Power,’ pp. 123-47.
53
Possessors of these rights are typically called the ’Siva Brahmans possessing kāni’ (kāni
utaiya civapirāmmanar) in Chola-period inscriptions. There are fifty-six total instances of these
terms in the data base.
54
In a typical transfer of i,ān a Brahman assembly sold waste land )
k pāl in their village to a
(
certain Tanttottam udaiyān, who enjoyed the land as his i, ān with crops of his choice, paying
k
amai and kudimai taxes (
d
ka S23:II 303).
55 There are fifteen
total references to i ān kadan in the data base, fourteen from Kumbakonam
k
and Tiruchirappalli taluks alone.
SII 23: 187; TK 131.213,223; ARE 1939-40: 242, 389; 1978-79: B
56 293.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

56

property..57 Taxation depended on payment by the official possessor of käni

land. Temples possessed kani lands in their own right as corporate owners.
-

On the other hand, other contextual aspects of kani imbued the individual
rights it conferred within a variety of social duties. Occupational kini and
kciiii in temples were contingent upon the performance of public services,
while kuni in land necessitated the payment of taxes. Official title to land
also masked a variety of social limitations on individuals initiative. Privately
possessed holdings were shares (pangu) within village communities
dominated by assemblies of village notables who often made collective
decisions over land use, including at times the essential questions of irrigation
waters, agricultural labour, or cropping. 58 Inheritance customs put forward
a single person as official possessor of property, but there were limits to the

ability of that person to alienate or alter property without consideration of

family dependents .51 Embedded within ramified systems of social claims,
and perhaps existing in the absence of a land market, kipi nonetheless
referred to an indigenous structure of ownership that was invoked for
officially private purposes in a number of Chola-period records.
Table 5 displays the changes in the relative frequencies of references to

TABLE 5
Total Nuntbers and PeTœI’II/IgØ of Records Referring to Iui1:ù in Five Study Amm

*
Note that there are no reference to kä1)i from the inscriptions of sub-period 1 (849-985).
The ratio refers to the number of records from each sub-penod containing references to
kii1)i divided by the total number of records from that sub-period.
ARE 1906: 130, 146. 147. See also ’days’ in the temple as transferable rights of Brahmans
57
A
(1911:
RE 267; 1914: 46; 1931-32: 115).
58I udden, Peasant History, pp. 86-89. Kathleen Gough has interpreted the system of shares
as a type of village communal property, with the king as landowner, in ’Modes of Production in
Southern India,’ Economic and Political Weekly, Vol. 15, No. 1 (Jan.-Mar. 1980), pp. 343-46.
Lately Burton Stein has stressed ’communal’ ownership in ’Politics. Peasants.’ pp. 75. 82.
59J. Duncan M. Derrett, Essays in Classical und Modern Law, Vol. 2 Leiden, 1977. pp.
21-25. 86-91; Gunther-Dietz Sontheimer and J. Duncan M. Derrett, Der Begriff des Eigeutums
in Hindurecht,’ in J. Duncan M. Derrett, Gunther-Dietz Sontheimer and Graham Smith,
Beiträge zu Indischem Rechtsdienken. Wiesbaden, 1979, pp. 90-93, 103; Yvonne Bongert, ’La
notion de Propriété dans l’Inde.’ Travaux et Recherches de l’Institut de Droit comparé de
23. Etudes des Droit Contemporain. 1962, pp. 156-59), 162.
l Université de Paris,

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 /57

kä1)iin the inscriptions from the five study areas. Notable is the lack of
references before 985, and the relatively few occurrences until about 1070.
The relative and absolute number of references continue to increase
dramatically thereafter, especially in those areas more closely connected to
riverine irrigation. For example, in Tiruchirappalli taluk two references in
the eleventh century increased to sixty-one references in the thirteenth
century. The percentages of records containing references to kani eventually
outstrip those records mentioning the king, his officials and even centrally-
oriented taxation. Compare the greatest presence of the iand revenue,
department in Kumbakonam taluk (appearing in 17 per cent of the records)
with the much higher figure for kani (75 per cent) in the thirteenth century
(Tables 3 and 5).
The interpretation of this terminological shift depends on the connections
between word usage and local property control. Karashima has seen the
growth of kä1)i in Tiruchirappalli as part of the consolidation of larger,
private domains in the hands of ’possessors’ (uçlaiyär) during the later Chola
period, at: influxes of plundered wealth and religious donations disrupted
earlier communal properties within villages Transfers of kani land to
temples within the five study areas do reveal a predominance of ’possessors’
among secular donors .6 Although the phenomenon of property differentia-
tion may indeed have been important in the twelfth and thirteenth centunes
the more generalised increases in kä1.1i references may reflect a greater
specification of individual rights rather than origination of a completely new
system of property rights.
A key may lie in parallel although rarer increases in terms specifying
cultivators’ rights (ktidi kiiii), the official recognition of the right to cultivate
land even without title to that land as a pnvate possession. Records especially
frcm Pudukkottai describe donation arrangements with provisos that the
cultivators of donated land may not be excluded (kucli ninkd; from tenancy
under the new temple owners. At times it seems that those permanent
cultivators are the donors themselves. 12 These provisos may provide a clue
to procedures underlying many other land donations featuring kani transfers.
Official changes in ownership of property could entail retention by previous
owners or their previous tenants of considerable rights to agrarian produce.

Increasing specification of-ktilit rights was thus part of a larger movement

toward greater specification of all kinus of property rights, in order to insure the
preservation of local privileges within the framework of religious endowments.
60
Noboru Karashima.
South Indian History and Society . pp. 21-35.
61
Out of 65 recorded transfers of kāni land in the data basen, 43 instances (74 per cent)
involved persons whose names included the honoritic titles for ’possessor’ (udaiyār), ’elder’
kilavan or ’lord’ ).
.
(
) araivan
(
62
A total of 26 references to the term ku
i nāngā in the data base includes 16 references from
d
post-1070 Pudukkottai. SIISee donor-cultivators
23: in 257: Transactions of the Archaeological
Society of South India. 1958-59. pp. 84-110. Plate 13.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

58

A major cause of class differentiation and terminological change may lie

in the activities of the Chola state that we have traced in this paper. The
kings gradually and inexorably altered the rules of the political game by
moving into local arenas such as Pudukkottai, traditionally dominated by
local leaders or naftcr. The main weapon for this penetration was increasingly
a taxation system attempting to effectively control allocations of large

proportions of agrarian produce. The kings simultaneously continued their

support of ritual integration by building and endowing temples, a policy that
in earlier stages of political consolidation was crucial for political integration
of local allies. The nattar, appearing less frequently as arbiters of local
administration, were increasingly threatened in their own backyards.
Paradoxically, a way out was provided by the kings themselves, and lay in
religious donations. A local possessor could on one hand establish his own
position as a protector of the moral universe and supporter of the Chola
kings, and ensure continued control over the distribution of local agrarian
produce, by alienating lands to temples. On the other hand he could create
provisos entailing continued rights to cultivation for himself or his client
cultivators, or retain unofficial rights to appointment of ritual specialists or
allocation of sacralised food connected with the ritual enactments supported .
by donated lands. 13 Most donated lands were made tax-free by royal orders
after local requests, or were made tax-free through lump sums given by
donors to defray all future taxes; it may have been worth an initial expense to
eliminate later royal taxation in an environment where those taxes showed
signs of rising further. Alienation of titles to temples in these ways did not
entail a decrease in the burden of the upper share extracted from the
producers, but signified a redirection of control over its uses that ultimately
favoured local authorities.
’rhe donation of lands to temples, and the verification in inscriptions of
the particular rights enjoyed by all participating parties, were thus the signs
of an increasing flight from royal control and the creation of tax shelters in
religious institutions. The greater implementation of a centralised Chola
state thus led to ever greater alienation of officially individual rights to
temples, and the progressive starvation of the central state at its time of
greatest need. In this view the local leaders were major actors in the growth
of religious institutions, especially temples, and their need for differentiated
and specified property rights spurred on a widespread terminological change
that mirrored renewed political localism. But as the central state fell apart,
temple endowments expanded until the temples themselves became the
greatest institutions in South India, major landlords and political forces in
themselves.

63
Arjun Appadurai, ’Kings, Sects and Temples in South India,’ IESHR, Vol. 14, No. 1
(Jan.-Mar. 1977), pp. 47-73; Arjun Appadurai and Carol Breckenridge, The South Indian
Temple: Authority, Honour and Redistribution,’ Cuntributiom to Indian Sociology, Vol. 10,
No. 2 (July-Dec. 1976), pp. 187-211.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 59

Conclusion

The study of the Chola period has revealed two main policies of royal
political unification. The first and earliest policy displays the characteristics
of the ritual, segmentary state, through which the Chola kings attempted to
unite disparate and fairly autonomous local leaders under a single, mediating
agency through ritual means. Keys to this mediating royal role were (1) the
manifestation of royal protection over religious institutions (Brahman
settlements and temples) through establishment of tax-free revenue grants
and construction of shrines, and (2) periodic arbitration of local disputes,
typically involving religious institutions, either in person or more often
through representatives who were drawn from allied local elite groups. A
second and later policy involved a tightening of royal control over local
resources through the recruitment of elites into a more bureaucratic tax
collection agency, and the implementation of tax collection or reallocation
within the rich agricultural zones that supported religious institutions. This
second policy reveals a drive toward increased revenue extraction and
greater centralized control within the core area of the empire. The Chola
kings remained ritual leaders, but aspired to be managers in the ArthilSåstra
style.
The success of royal integrative policies depended on local variables of
geography. The most striking feature of the data presented here is the rapid
decrease in the penetration of all aspects of royal influence with increasing
distance from the centre of the polity. Even within the outer reaches of
Cholamandalam, the core area of an extended polity, the kings were more
likely to strike deals with local leadership than to implement a centralised
administrative apparatus. The nature of local leadership in turn varied
according to ecological characteristics that underlay varying productive
regimes. Smaller riverine tracts, with their relatively greater and assured
agricultural surpluses, supported dominant lineages that appropriated some
of the ritual or administrative characteristics of kings. When the Cholas
overran peripheral riverine tracts, they reinstalled or created dominant
lineages which supported their overlords in return for continuing local
.autonomy, in a process that resembles more closely ’feudal’ political sub-
ordination. In drier zones, with discrete and more insecure productive
regimes based on rain-fed fields or small artificial lakes, the kings en-
countered collective assemblies of many local power-holders. The local
assemblies became insignificant as the waxing royal system absorbed leader-
ship into roles as nadu ,settlement officials, adhikiris, and eventually tax
department members. But as royal power waned, the assemblies of local
leaders again came into view as forums for articulating and adjusting the
disputes of the dry zones.
The three configurations of political dominance-royal centralisation,
’feudal’ subordination, or ndttar assembly-thus rested ultimately on the
ecology and modes of production that underlay them in different areas.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

60

The determinative impact of ecological features intersected with the historical

processes whereby the Chola kings, through their military successes, con-
solidated enough power to allow a greater penetration of their own agents
within the peripheral zones during the eleventh and twelfth centuries.
Despite the importance of the varying interactions of these superior
political organisations or alliances for the historical development of the
Chola state, the underlying dynamics of state formation rested on the ability
of those superior agencies to coordinate the aspirations of elites emerging
directly from the village level. Political and economic leadership within the
predominantly agrarian economy rested on the possession of land and/or
rights to the produce from land. The contexts of the term~kani in the Chola
inscriptions suggest that power over land and its produce was not communal
(despite a variety of collective controls), but was instcad divisible into a
number of officially determinable legal rights or ownerships. Differential
access to these rights raised up locally-dominant kinship groups and
individuals representing their interests-individuals who entered nattar
assemblies, who called themselves ’possessors,’ and who ultimately interacted
with superior state agents. The policies of these local leaders included the
preservation or extension of their authority over agricultural resources.
When superior state organisations encouraged these local prerogatives,
local leaders offered support; thus the early policies of the Chola kings, who
entered rarely into local affairs but offered avenues for local legitimation
and opportunities for booty, enlisted support for imperil policies. But when
a superior state attempted to manipulate the village economy more directly,
the arbiters of village affairs were naturally willing to abolish that state; thus
policies of royal centralisation called forth the creation elf tax shelters and
ultimately open rebellion by the thirteenth century. In these ways the
behaviour of local elites within fertile agricultural tracts was crucial to the
political fortunes of premodern dynasts in South India.
The results of the present study exemplify the qualitative differences
between ’nuclear areas’-zones of rich alluvial soil and abundant water-
and zones where stable agriculture was less rewarding or more insecure. The
emergence of the Kaveri river delta as a dynamic agricultural tract was the
salient feature underlying the hegemony of the Chola kings, who in turn
initiated projects designed to stimulate further agrarian expansion. The
dynamism of the Kaveri delta fuelled an :mperialism that further stimulated
inveatnrent in land reclamat.on and irrigation expansion within other, lesser
riverine tracts and in peripheral areas. The impact of the Chola state was,
then. to provide formats for the expansion of leaders within the agrarian
society of Tamil Nadu. Several levels of leadership represented the upper
level of a hierarchically organized production process mat exploited
advantages of land and water to create fertile agricultural tracts. The Chola
period was thus a time when the riverine zones were being filled up, when
complex, unequal social management brought techmques of lanci and water

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011

 /61

exploitation toward the limits of the riverine zones. The political figures in
this drama were them several layers of nobles and landowners, resting on the
fruits of peasant cultivation, who interacted with the kings and constituted
state institutions. The Chola polity was an ’early state’&dquo; in the sense that its
agrarian base and the political power of its elites were in an early stage of
expansion.
What were the legacies of the Chola polity as an early state? Perhaps the
most striking images of secular change were the numerous and beautiful
temples that dotted the landscape of fertile agricultural zones in south
India-the legacy of an imperial policy that encouraged ritual. manifestations
of temporal authority. Behind the temples was an expanded agrarian and
commercial base that found its expression in the urbanised environment thut
grew up around the holy sites and in larger areas of green fields at harvest
time tilled by peasant cultivators. Four centuries of relative peace and
encouragement of local initiatives had spurred or. the medieval expansion,
but the Chola kings had also instituted a new level of governmental involve-
ment in the fertile nadus, including periodic overseeing of economic activities,
the mediation of centrally-sanctioned officials, and the collection of taxes.
The accomplishments of the Chola period--expansion, urbanisation, and
central involvement in local affairs-were not lost during the succeeding
Pandyan and Vijayanagara periods, but were exploited by later dynasts to
build larger and more impressive state structures.

64
See discussions of the features of the early state in H.J. M. Claessen and Peter Skalnik, The
EarlyState, The Hague, 1976. The authors conclude that political organisation was a relatively
closed system standing outside and exploiting large number of local communities, a feature
corresponding to Marx’s ideas of the Asiatic mode (pp. 546-54.604-06.642-43). The emphasis
of the present study on intermediate authorities suggests that a variety of elite groups emerging
from village-level production rotations could interact with the central state apparatus and wield
effective power in their own right in the early state.

Downloaded from ier.sagepub.com at Universitet I Oslo on April 29, 2011