TRANSFORMATIONS

comparative sfady of social transformations

CSST
WORKING PAPERS
The University of Michigan
Ann Arbor
 T H E ORIGINAL CASTE: POWER,
HISTORY, AND HIERARCHY
IN S O U T H ASIA

CSST Working CRSO Working

Paper #10 Paper # 3 6 7
November 1988
 THE ORIGINAL CASTE:
POWER, HISTORY, AND HIERARCHY IN SOUTH ASIA

Nicholas B. Dirks
The University of Michigan

Traditionally, power was what was seen, what was shown and
what was manifested and, paradoxically, found the principle
of its force in the movement by which it deployed that
force. Those on whom it was exercised could remain in the
shade; they received light only from that portion of power
that was conceded to them, or from the reflection of it that
for a moment they carried.
Michel Foucault, Discipline and Punish, (p. 187)

The Politics of Caste

In pre-colonial Hindu India, the king--both as an historical

figure and as a trope for the complex political dynamics
underlying the Indian social order--was a central ordering factor
in the social organization of caste. This statement directly
opposes the prevailing theories of comparative sociology, and in
particular the theoretical position of Louis Dumont (1980). As
is well known, Dumont holds that the political and economic
domains of social life in India are encompassed by the
nreligious.n The religious principle becomes articulated in
terms of the opposition of purity and impurity. For Dumont, the
Brahman represents the religious principle, inasmuch as the
Brahman represents the highest form of purity attainable by
Hindus. The king, while important and powerful, represents the
political domain, and is accordingly inferior to, and encompassed
by, the Brahman.
 There are in fact many textual confirmations of the view
that Brahmans, and the spiritual authority (brahma) that they
possess, are seen as higher, both relationally and ontologically,
than kings, and the temporal authority (ksatra) that is theirs.
However, these same texts provide evidence as well of what has
been called "the central conundrum of Indian social ideology"
(Trautmann 1981; also see Heesterman 1978). At times the king is
above the Brahman, as for example in the royal consecration
ceremony. At other times the Brahman appears to be superior to
the king, as for example in the Manu Dharma Sastras, and in
passages from the Mahabharata. This conundrum is often addressed
in terms of the postulation of two levels of truth, a higher
level in which the Brahman is clearly preeminent, the source of
everything else, and a lower level in which kings must protect
and sponsor Brahmans in order for them to exist, as gods, on
earth. Dumont's resolution of this conundrum extends the notion
of higher and lower truths from a classically Indic
epistemological contextuality to his well known ontological
separation of the religious from the political. The major
development of political thought in India, he contends, is the
secularization of kingship, that is the separation of the magico-
religious nature of kingship--preserved in the form of the royal
chaplain in particular and in the function of Brahmans in the
larger polity more generally--from the political aspects of
kingship, depicted, inter alia, in the Machiavellian Arthasastra
 While Dumont is not wrong to insist on radical differences
in the "ideologies" of India and the West, the irony is that the
way in which he postulates difference is based on a fundamentally
Western ideology, in which religion and politics must be
separated. Dumont 's ~ o i s
t ion in many ways caracatures the
Orientalist assumption that India is the spiritual east, devoid
of history, untouched by the politics of Oriental Despotisms.
Critics of Dumont have often accepted his basic epistemological
premises, but then reversed them. They take a materialist
perspective and view social relations in India in terms of power,
pure and simple (e.g., Berreman 1971). Recent work--often by
those committed to an ethnosociological approach to the
identification and description of cultural domains in India--has
suggested that this separation of religion from politics, like
many other dichotomies in Western social science, is
inappropriate at the level of ideological (or cultural) analysis
in Indian social thought (~arriottand Inden 1977; Appadurai
1981; McGilvray 1983; Dirks 1982, 1987). It is in this sense--as
also of course in the running critique of Dumont--that the
following analysis is ethnosociological.
Not only is there no fundamental ontological separation of a
"religious" from a "political" domain, but religious institutions
and activities are fundamental features of what wedescribe here
as the political system. Kings derive much of their power from
worship, and bestow their emblems and privileges in a cultural
atmosphere that is permeated by the language and attitudes of
worship. Further, temples are key institutions in the formation
of social communities (Appadurai and Breckenridge 1976), even
while they reflect structures of power worked out both in and
outside their own walls (Dirks 1987). In turn, temples represent
the preeminent position of the king by granting him the highest
honor in the temple, before even the learned (srotriya) brahman.
Religion does not encompass kingship any more than kingship
encompasses religion. There are not two distinct forms of power,
secular power had by kings and sacred power had by Brahmans.
Kings and Brahmans are both privileged but different forms of
divinity in a world in which all beings were, however distantly,
generated from the same ontological source. And power--whether
defined as a constellation of cultural conceits or as an analytic
concern--can not be restricted to a single domain of Indian
social life.
Dumont has suggested that caste is fundamentally religious,
and that religious principles actualize themselves in the domain
of purity and pollution. In my ethnohistorical study of a south
Indian kingdom in which Kallars were the royal caste and Brahmans
were heavily patronized according to scripturally mandated forms
of royal gifting activity, I have found that purity and pollution
are not the primary relational coordinates which endow hierarchy
with its meaning and substance. Royal honor (mariyatai, antastu)
combined with the notions of restriction, command, and order
(kattupatu, atikaram, orunku) are the key discursive components
which are embedded in and productive of the nature and order of
hierarchical relations.
 My analysis will, I hope, do more than simply contest Dumont
over the issue of which key terms underlie the structural logic
of hierarchy in South Asia. Indeed, I wish to reintroduce
concerns with power, hegemony, and history into studies of
culturally constructed structures of thought, whether
structuralist or ethnosociological. The forms and relations of
power in southern India efface our social scientific distinctions
of materialist etics from culturalist emics, for even the domain
of ritual action and language suggests the complex and
conjunctural foundations of hierarchical relations. At least
this is true among the Kallars of Pudukkottai, less affected
perhaps than most other groups by colonialism and the demise of
the old regime in the nineteenth century. For the concerns of
comparative sociology are not only the products of a nineteenth
century Orientalism, but also of the colonial intervention that
removed the politics from society and created a contradictory
form of civil society--with caste its fundamental institution--in
its place. It was not only convenient to render caste
independent of political variables; but necessary to do so in
order to rule an immensely complex society by a variety of
indirect means. Colonial sociology represented the eighteenth
century as decadent, and all legitimate Indian politics as past.
Under colonialism, caste was appropriated, and in many respects
reinvented, by the British. However, the British were able to
change caste because caste continued to be permeable to political
influence. Ethnohistorical reconstruction is thus important not
only for historians confronting new problems of data and
analysis, but for anthropologists who confront in their fieldwork
a social system that was decapitated by colonial rule.

Ethnohistory of an Indian Kinadom

The Kallars, like the Maravars, settled in mixed economy
zones (Ludden 1985) such as Pudukkottai on the borders of the
central political and economic regions of the south. In these
areas they quickly attained dominance in late medieval times by
exercising rights of protection (patikkaval) over local
communities and institutions. The Kallars were successful in
this role because their strongly kin and territory based social
structure and cultural valuation of heroism and honor were highly
conducive to the corporate control of the means of violence and
coercion. It was no accident that Kallars, like Maravars, were
often, when not granted rights of protection, the very groups
from which others sought protection.
The Tondaiman dynasty of Kallar kings wrested control over a
significant swath of the Pudukkottai region in central Tamil Nadu
in the last quarter of the seventeenth century. Whereas Kallars
had been branded as thieves in much early Tamil literature and as
criminals by the British under the Criminal Tribes Act, in
Pudukkottai--a little kingdom that became the only Princely State
in the Tamil region of southern India--they became the royal
caste. Kallars controlled much of the land, occupied the
greatest number of authoritative positions, particularly as
village and locality headmen and as miracidars, and ran the most
important temples as trustees. These temples were often their
lineage, village, or subcaste-territorial (natu) temples, in
which they received honors only after the king and Brahmans. In
short, Kallars were dominant not only in terms of their numbers,
but for economic, political, and ritual reasons.
Pudukkottai, which at its most extensive did not exceed 1200
square miles, was located in an exclusively rain-fed
agricultural zone right in the middle of the Tamil speaking
region of southern India, straddling the boundary between what
had been the two great medieval Tamil kingdoms. Ruled by Kallar
kings, it provides an excellent place to test many of the
proposals of Dumont, who, before he shaped the concerns of much
contemporary Indian anthropology in his general proposals in this
journal and in Homo Hierarchicus, portrayed Kallars in his major
ethnographic work in India as a ritually marginal group that
exemplified the Dravidian isolation of kinship from the influence
of caste hierarchy. But in Pudukkottai, less than one hundred
miles north of where Dumont conducted his fieldwork, Kallars were
kinas; they exercised every conceivable kind of dominance and
their social organization reflects this fact.
Pudukkottai rose, as did other little kingdoms throughout
southern India, within the context of a late medieval Hindu
political order. In both its emergence to and its maintenance of
power, it exemplified the social and military vitality of certain
productively marginal areas in the seventeenth and eighteenth
centuries, a period that has commonly been characterized as one
of decline and decadence. But the eighteenth century was not the
"black century:" the decentralization of political forms neither
a condemnation of the capabilities of'the Indian state nor a
natural prelude to British colonial rule. The British conquest
of the little kings in the south was anything but absentminded,
and there are indications not only that the economy was buoyant
in part because of the active court centres ruled by these little
kings, but that small and local level states were learning the
political, military, and administrative lessons that the French
and the English were learning at the same time. But win the
British did, and thus their version of the eighteenth century has
collaborated with a subsequent neglect of the seventeenth and
eighteenth centuries by Western and Indian historians to provide
the grist for comparative sociology's Indian mill.
Colonialism purposefully preserved many of the forms of the
old regime, nowhere more conspicuously than in the indirectly
ruled Princely States. But these forms were frozen, and only the
appearances of the old regime (without its vitally connected
political and social processes) were saved. Colonialism changed
things both more and less than has commonly been thought. While
introducing new forms of civil society and separating these forms
off from the colonial state, colonialism also arrested some of
the immediate disruptions of change by preserving many elements
of the old regime. But by freezing the wolf in sheep's clothing
it changed things fundamentally. Paradoxically, colonialism
seems to have created much of what is now accepted as Indian
"tradition," including an autonomous caste structure with the
Brahman clearly at the head, village based systems of exchange,
isolated ceremonial residues of the old regime state, and
fetishistic competition for ritual goods that no longer played a
vital role in the political system.
In my research on Pudukkottai, it took little study of local
land records to uncover the most surprising historical
characteristic of the political system: how little of the land
was taxed. ~ccordingto mid-nineteenth century records, less
than thirty percent of the cultivated land was either taxed (9%)
or given out from year to year on a share basis (18%) in which
one ninth of the produce was accorded to village servants and
four ninths each to the cultivator and the government. Seventy
percent of the cultivated land was inam, or tax-free. This mid
nineteenth century statistic was if anything far higher in the
eighteenth century, when there were at the very least another
five thousand military inams, i.e. forty percent more than the
total number of inams in the mid nineteenth century, before the
gradual dismantlement of the military system of the state.
Roughly thirty percent of the inams (numbers of inam units rather
than acreage) were for military retainers, their chiefs, and for
palace guards and servants; twenty five percent were for village
officers, artisans, and servants; and the remaining forty five
percent were for the support of temples, monasteries, rest and
feeding houses for Brahman priests and pilgrims, and land grants
to Brahman communities. In terms of acreage, roughly nineteen
percent of the alienated land was for military retainers et. al.,
seven percent for village officers, artisans, and servants, fifty
one percent for temples, monasteries, and charities, and twenty
two percent for Brahmans. Remember that these statistics reflect
a demilitarized political system, so that both the numbers and
percentages had earlier been far higher for military categories.
Remember also that this particular kingdom was ruled over by
kings of what have been said by most observers to be an unclean
caste, inappropriate for Hindu kingship, and therefore
inappropriate donors for Sanskritic temples and Brahmans.
This structure of privileged landholding reflects the
structure of political -power and socio-cultural participation
within state and village institutions. The chief landholders
were the great Kallar Jagirdars and Cervaikarars. The former
were collateral relations of the Raja. Jagir estates were
created for the two brothers of the Raja after a succession
dispute in 1730 severely threatened the stability of the state.
These collateral families kept these estates intact until their
settlement in the late nineteenth and early twentieth century.
The Jagirs were, in effect, mini kingdoms in their own right,
each containing a small court and a full set of inam grants,
including "military ones." Importantly, however, the jagirs were
not made up of contiguous villages and were therefore never
geographically isolable units.
Just below the Jagirdars came the Cervaikarars. All but one
of the Cervaikarars were of the same subcaste as the Raja, and
most had one or more affinal ties with the royal family. The
Cervaikarars were given large grants of land, titles, honors, and
emblems. Each of the Cervaikarars was awarded a specified number
of retainers, or amarakarars, to serve them at home, to go to
battle with them abroad, and to carry their honors and emblems to
 11

ritual occasions in the royal court and in temples. Lesser

chiefs, called Kurikarars, came from Kallar subcastes other than
royal one. Lands and privileges throughout the state were also
given to other Kallars, called in diminutive form Cervais, to
keep watch over villages and localities not dominated by loyal
Kallars (i.e., all groups other than the Vicenki Nattu Kallars
who were only finally brought under nominal control in a series
of wars in the late eighteenth century). The Cervais were mostly
members of the royal subcaste who had no affinal ties with the
royal family.
The royal family and court was itself protected by
Uriyakarars, all of whom were Akampatiyars, members of a non
Kallar caste which was aligned with the Kallars through ,'

membership in a special metacaste of three warrior groups along

with the Maravars called the mukkulattur. These royal protectors
in fact became a separate subcaste marked off terminologically
and affinally from other Akampatiyars in the region by virtue of
their connection with and service to the Raja. A number of
Uriyakarar chiefs had a prominent role in the kingdom. Like most
of the lesser chiefs, these chiefs were given extensive lands but
no formal group of military retainers under them.
In addition, within each village in the state, headmen were
given lands in recognition of their rights to local authority as
well as to render this representative of the state's power at
large. These headmen came from the locally dominant castes.
Kallars were dominant in the northern and eastern parts of the
state. Maravars had a significant presence in the south.
Arnpalams (the title for headman, literally meaning the central
common ground of the village, used by most of the castes in
~udukkottai)were also called miracidars after the mid-eighteenth
century. This new label, borrowed from Persian revenue
terminology, was used in an attempt to render local authority as
dependent as possible on recognition by the "bureaucratic" state.
Nonetheless, well into the twentieth century these local headmen
were often as powerful as small kings, with retinues and legends
sufficient to cause their power to be felt over significant areas
of the countryside.
Various village officials, artisans, and servants were also
given inam (more properly maniyam) lands by the state. In
addition to this land, each village servant was also rewarded
with shares of the village grain heap. Since the one-ninth share
of the harvest that was owed to village servants was taken from
the grain heap before its division into the Raja's and the
village's share, this classic jajmani payment was borne equally
by the village and the Raja. Thus, the sets of relations usually
characterized as jajmani, that is as an institution of the
village community alone, were sanctioned and underwritten not
only by the community but also by the king both through inams and
the share system.
Maniyam, the term used for many village grants, meant land
that was held free of tax, as well as privilege in a more general
sense. Maniyam derives from the Sanskrit manya, which means
honor and privi1eg.e. Many of the land grants to Brahmans were
called carvamaniyam, meaning completely tax free and honorable.
However, the term maniyam was not reserved for Brahmans, as
British categories which separated "religious" from
"nonreligious" grants implied. Indeed, in its most unmarked form
maniyam was sometimes used for inams in general. Maniyam was
also used in a more marked sense for land grants given to village
servants whose task was to maintain and operate irrigational
facilities, to village officers or headmen, to the pujaris or
priests of small village temples or shrines, and to inamdars
(holders of inams) who had such variable responsibilities as
blowing the conch for a village festival or tending a flower
garden which produced garlands of the village deities. These
maniyams reveal that royal grants sustained the entire structure
of local village ritual.
Even small locality temples were linked to the king through
the inam. These local temples organized the ritual systems of
villages, often constituting some of its fundamental cultural
coordinates as well: they demarcated boundaries, centers, the
relationships of social groups within the village, defining and
internally ranking lineages, subcastes, and castes. Service to
the temple was in many respects structurally equivalent to
service to the village community, even as most village service
inams specified services to both temples and the village, as
suggested for Sri Lanka as well in the work of Hocart, who saw
each village service group as a priesthood, and thus saw caste as
an institution that was simultaneously political and religious.
In addition to many inams granted to village and local
temples in the form of maniyams to local priests and village
servants, many inam grants were also made to Brahmans, temples,
and charities of various sorts. As is well know, the principal
sources for south Indian historiography are epigraphical records
of such grants, publically proclaimed because of the merit which
accrued to the donors from them and because of the centrality of
these gifts to the ideology of kingship. One of the fundamental
requirements of Indic kingship was that the king be a munificent
provider of fertile lands for Brahmans who would study and chant
the Vedas, perform sacrifices and provide ritual services for the
king so as to ensure and protect his prosperity and that of his
kingdom; for temples which were the centers of worship and for
festivals such as Dassara which renewed the sovereignty of the
.
king and regenerated the kingdom; and for cattirams (chatrams,
also called choultries, which were feeding, sometimes lodging,
houses for pilgrims) which provided sustenance and shelter for
itinerant Brahmans and pilgrims. The merit (punyam) of the king
who made the grant could be shared by all those who protected the
gift, a duty enjoined upon all subsequent kings. In spite of
Pudukkottai's marginal social and political position, 'it was well
endowed with temples and brahmanic institutions precisely because
of the prevailing force of royal ideology.
The underlying political base of any little kingdom in the
old regime was its military capacity. his capacity was in turn
based on structures of alliance and command, which were
articulated by gifts, privileges of varying kinds, and kinship.
No little kingdom could survive if it did not have an efficient
system of military mobilization. These systems were organized
around subordinate chieftains, connubial connections, and
privileged landholding rather than a centralized or
bureaucratically organized system of revenue collection and
military rule. Royal grants helped to sustain military
organization as well as local village ritual and an impressive
complex of larger temples and brahmanic settlements. The
political economy (by which I mean here the institution of
kingship, the distribution of authority, and the nature and
structure of resource allocation) was based on a logic of
redistribution that penetrated far and wide.
The gift of land without onerous burdens of taxation, the
occasional participation in wars in which honor and booty could
be won, and the organization of land and military rights in
relations of ritual clientage to chiefly and kingly patrons
resulted not only in a political system of great fluidity and
dynamism, but one in which individuals could vie for relative
distinction in a social system where honor was intimately tied up
with rank through interpenetrating forms of political and ritual
action. The valued constituents of sovereign authority were
differentially and partially shared through the redistributive
mechanisms of the gift. Service was offered as a way of entering
this redistributive system. Kinship (a relatively open and
inflected system) became the social base and expression of social
and political relations. Honor, in particular the emblems and
privileges that were given with each grant (itself a privilege),
but also the honors in temples that were procured through worship
and were ordered in relation to local and royal prerogatives, was
both the mediation and the mechanism by and through which
relations were established.
Thus I argue that the royal gift was basic to statecraft in
all the kingdoms of the old order in southern India. All gifts
were not the same; but they all shared one thing in common: they
were given by the king. The substance of the gift (the land
rights, titles, emblems, honors, and privileges of service,
usufruct, and command) was the partial sovereign substance of the
king. Participation in the king's sovereignty was not, however,
unranked, for the differential nature and contingent character of
all these entitlements provided the basis for the creation of a
political hierarchy. Ultimately, entitlements by their very
nature constituted hierarchy through a logic of variable
proximity to the king, to sovereignty itself. What Geertz has
written about Bali is true of the old order in south India: "The
whole of the negara (court life, the traditions that organized
it, the extractions that supported it, the privileges that
accompanied it) was essentially directed toward defining what
power was; and what power was what kings were."
My sense of the meaning of royal gifts was initially based
not only a reading of land records, which though they gave
histories of grants, revealed a thick infrastructure of gifting,
and suggested that land rights were necessarily conjoined with
other rights to privilege, service, and honor, were themselves
insufficiently explicit about the ideological content of the
system to permit a full or satisfactory interpretation. Rather,
I developed an understanding of kingly beneficence from textual
sources that depicted the centrality of gifts and their various
forms. Using eighteenth century texts (genealogies, chronicles,
ballads) as cultural discourses, I found persistent motifs,
events, narrative forms, tropes, and images, and I read the parts
they played in the poetics of power. I used this textualized
discourse not only, as at first I thought was all I could do, to
get a sense of how these Indians conceived their own past, but
also to demarcate the key elements of my subsequent inquiry, to
create an historiographical frame for understanding key
structures, events, and their relations: and I found that my
textualized readings were indeed realized in historical
processes. Thus I was able to identify and focus on the core
conceptions of sovereignty; the interpenetrating transactions in
gifts, service, and kinship; the structure and form of political
hegemony. I was able to understand what had previously been
obscured in the colonial writing on the little kings, or
poligars, that the adoption of Hindu forms of kingship by what
were said to be low caste (later often defined as criminal castes
under British rule) chiefs was not just an ideological ruse but
was rather reflected in the entire structure of the political
system; that rights to landholding were political rights, and
reflected the structure of the little kingdom at the same time
that they revealed the pervasive importance of royal honor; and
that the states were not absolute failures because of their lack
of emphasis on the bureaucratic demarcation of land rights and
the collection of revenue, but rather successful, vital, and to
the British highly threatening political systems because of the
interplay of rights and privileges to land, service, kinship,
honor, and local resources.

The Cultural Poetics of Power

But I was still not altogether clear if, and if so how, a
logic of variable proximity to the king actually informed social
relations in the little kingdom. It was this, in a general
sense, that was the subject of a subsequent year's ethnographic
research. It was only in my fieldwork that I was able to find
that the forms of clan and caste structure that the British had
seen as organic growths from the Indian soil had in fact been
vitally transformed by the political histories of local level
chiefs. First, through inquiries conducted in the field I
determined that the political hierarchy was also, with certain
crucial exceptions, a social hierarchy. As I mentioned above,
the Jagirdars were collateral chiefs, the Cervaikarars affinally
connected warriors, the Kurikarars mostly Kallars but from
subcastes other than the royal subcaste, and so on down the line.
The Kallars themselves were, as Durnont also found for the
Pramalai Kallar, territorially segmented, but in Pudukkottai the
royal subcaste occupied a uniquely important position, dominating
all the other segments, or subcastes. The internal organization
of the royal subcaste, markedly different than all other
subcastes, itself reflected a systematic if sometimes paradoxical
inflection by political forces. And the settlement of Kallar
chiefs, both great and small, throughout areas of non-Kallar
settlement, as well as in the area inhabited by a large and often
unruly Kallar subcaste, effected-both the ideological and
instrumental dominance of royalty. These Kallars often had royal
retainers under them, privileged rights to local lands, and the
right to receive first honors on behalf of the Pudukkottai Raja
in all village and locality temples and festivals.
The hierarchical force of royalty was expressed in many
ways, not least through the comments of one of my principal
informants, the titular head of the royal clan or kuppam. "When
we assumed our royal status (antastu)," this man told me, "we
became, as it were, a royal family. Hence, we the five top
lineages of the clan, began to have affinal relations only with
royal families. So we became more elevated and dignified than
the other groups and other clans. While the influence and glory
of the Raja was high, the influence of those of us living in our
group also went up accordingly. Others who do not have marriage
ties with the five chief lineages also reside here but we
classify them at a lower level." All members of the royal
subcaste were loosely called rajapantu, meaning that they had a
connection with the raja. While this term was used to designate
all members of the subcaste in an unmarked sense, within the
subcaste itself there were multiple distinctions of rank, all of
them, as it turned out, having to do with proximity of the king.
In one particularly lucid discussion, my informant explained to
me the logic of hierarchy in Pudukkottai. "Why are we (meaning
royal Kallars) superior?" he asked rhetorically. "Because we
maintain control and order (kattupatu) in our community. We do
not allow widow remarriage and we abide by the moral codes of our
society strictly. Other Kallars may say that all Kallars are the
same. It is popularly assumed that all Kallars are thieves. But
we are not thieves. How can the ruling Kallars steal from
others? Our Kallars are pillars of the community, chiefs,
village leaders, politicians, and nobles. We have to maintain
law and order. How can we go off thieving. We decided that we
should lead a life of order and restriction. Others are not like
us. We live for honor and status. Our Kallars base our lives on
the temple and on marriage relations. Only if the temple and the
lineage are correct can we seek an alliance; Our honor is
displayed in the palace and in temples. When honor is measured,
in the same way the number of carats is measured in gold, will we
like less dignified groups taking seats on a par with us. No,
they are not fit to sit with us."
The preeminence of the royal subcaste is thus explained not
only through reference to the fact that the king hailed from this
subcaste, but by noting that this subcaste has the most rigidly
defined and maintained code for conduct of all subcastes. These
Kallars have the most order, and they enforced order through the
set of restrictions which are implied by the term kattupatu.
Kattupatu, which can be taken to mean code for conduct and
discipline, literally means something more like restriction, or
even constriction, deriving from the root kattu, meaning tied or
knotted. The code for conduct includes rigidly defined kinship
rules, some of which, like the Brahmanic prohibition of widow
remarriage, mark the royal Kallars off from all other Kallars and
suggest a kind of Sanskritization, others of which involve
working ones way through the myriad gradations of upper Kallar
society by trading political, social, and cultural capital back
and forth, often through affinal transactions. Kattupatu is a
term that is used frequently by all Kallar subcastes and indeed
all Pudukkottai castes, though only among Kallars, and
specifically within the royal subcaste of Kallars, does it have
the particular kind of inflection I just described. For all
these groups, though again most importantly for the royal
subcaste, kattupatu does not mean simply a code for conduct, but
a set of authoritative procedures which renders this code
enforceable within the community.
My informant's statements, and the general ideological
orientation they reflect, reveal the continuation of concerns
about the past reputation of Kallars as thieves, bandits,
outlaws. The ethnographic discourse here shares much in common
with the eighteenth century family histories that I mentioned
earlier. In these texts as in the statements of Kallar
informants the acknowledged past becomes totally transformed by
the attainment of kingship. Again, there is an implicit
opposition between the representation of the activities of
thieves and the activities of kings. In this case, the royal
duty of protecting and subduing disorder is combined with an
ideology of order and restriction which organizes and becomes the
subtext of the social relations of the subcaste.
The very word Kallar means thief in Tamil, and no one, at
least none of my informants, disputes the fact that at certain
places and times particular groups of Kallars engaged in
predatory activity. In fact they constantly bring it up with a
certain kind of relish, suggesting only partial embarrassment
about their past. And as I and others have written elsewhere,
this in itself is not necessarily a problem in any case, since
predation was often the principal means used to accumulate wealth
by kings who were not so concerned about a regularized tax base
as the British became when they began in the early nineteenth
century to gain most of their profit from land revenue. In
Pudukkottai, Kallars attained their position of royal authority
in the first place by providing protection to local communities
and institutions, and this is amply documented in inscriptions
recording protection contracts from the fourteenth century on.
Indeed, Kallars were given rights of protection because of their
capacity to control, and to a very large extent monopolize, the
means of violence, and there is much in Indian tradition to
suggest that the opposition of bandit and king is a complementary
opposition. But it is, culturally speaking, an opposition: the
violence of the bandit is illegitimate, and it represents and
causes the disorder that the legitimate violence of the king must
control. Kings are not only legitimate, they define the realm of
the legitimate. And the way in which the royal subcaste
organizes its social relations makes it impossible that they
could be thieves, or affected in any way by this general
reputation. The royal subcaste is headed by a king, and it
provides almost all the royal nobles of the kingdom. The
fundamental duty of these members of the elite is to subdue
disorder, destroy lawlessness, and enforce law and order, both
within the kingdom at large, and within the subcaste itself. And
of course, as kings, by virtue of defining what is legitimate,
they define disorder too.
But this is true in more than just the obvious, or for that
matter the Foucaultian, sense, as I discovered when doing
fieldwork with other caste groups. For most castes, there is a
steady decrease of order as one goes down the caste hierarchy, in
the sense defined by my Kallar informants and assimilated with
only minimal dissent by many non-Kallar informants. Maravars,
for example, who in all fundamental respects were like Kallars
except for the crucial fact that they were not kings, had found
it impossible to organize their social relations in the larger
territorial units, the natus, that Kallars had, and the Maravars
lamented this loudly and frequently. Indeed, the Maravars
themselves attributed this disorder not just to the general
decline and fall of the world in general, but more particularly
to their loss of political control. For other groups, there was
not only a noticeable decline in order, and a laxness in defining
and maintaining the kattupatu, there was also a decline in the
autonomy to define what order was. Untouchable groups, for
example, took the locality, and sometimes the lineage, names of
their dominant caste patrons. Whereas other groups had
traditional rights called kaniyatci to land, honor, etc.,
untouchables told me that for them kaniyatci only meant the right
to serve their patrons. The fundamental structures of their
social relations were inscribed by the hegemony of the dominant
classes. Notions of honor, order, royalty, and command have been
operationalized in the practices that produce and reproduce
hierarchy. These practices (embedded as they are in cultural
forms and historical processes) are themselves based on
structures of power as well as on the hegemonic nature of
cultural constructions of power.
If it be argued that my interpretation, though perhaps true
for marginal regions like Pudukkottai, can hardly apply generally
to south India, let alone to the subcontinent as a whole, I reply
that is precisely the marginality of Pudukkottai that makes it
possible to detect the forces that were at work elsewhere.
Because Pudukkottai was not brought under patrimonial control
(neither that of the Islamic rulers in the south nor later that
of the ~ritish)caste was never set completely loose from
kingship. Many current theories of caste, particularly those
emphasizing Brahmanic obsessions concerning purity and impurity,
but also those aspects of ethnosociological theory that stress
the proper and improper mixing of substances, are in large parts
artifacts of colonialism, referring to a situation in which the
position of the king and the historical dynamic of royal power
has been displaced, and sometimes destroyed. However much
Dumont's theory is predicated on an a.priori separation of what
he describes as the domains of religion and politics, which the
former encompassing the latter, he was also almost certainly
influenced by an ethnographic reality in which kingship played
only a very small, residual role. As for early ethnosociological
proposals about caste, Inden has himself recently noted that his
early work is largely derived from texts which were generated
only after the demise of kingship as a powerful cultural
institution (Inden 1983). The texts, he now says, reflected new
traditions which attempted to deal with the problem of regulating
caste interaction in an environment in which there was no longer
a king.

Qn Dumont: The Politics of Hierarchv

Politics, as we define it here, has both to do with the
processes by which authority is constituted at each level of
representation and with the linkages of the constituent groups in
society to the king (usually through the authoritative figures
who represent their social groups). Politics has a territorial
dimension, but is not exhausted by territorial forms. In the
royal subcaste of Kallars in ~udukkottai,the intervention of the
king changed and reconstructed (as well as decomposed) the
internal order of the system, affecting both social and
territorial forms. Even in other subcastes, less directly
influenced by kingship, social organization was only
understandable within a framework which is fundamentally
political, realized over time (i.e., in history).
It is my argument here that structures of power play a
central role in the social organization of caste and kinship,
that politics is fundamental to the process of hierarchialization
and the formation of units of identity. Durnont has great
difficulty with the notion that kinship can be politicized. When
he does see hierarchical tendencies develop in the domain of
kinship he blames them on the ideology of caste, which has to do
not with politics but with purity and pollution. Dumont's
elevation of alliance as the fundamental principle of south
Indian kinship is in large part because alliance mitigates the
asymmetrical effects of marriage relations through the
generalized exchange of marriage.partners within the endogamous
group. Hierarchy creeps to the borders of the endogamous group,
but only enters in the sense that it can bring about the creation
of new endogamous subdivisions. Nonetheless, even though Dumont
suggests (Dumont 1957a; 1957b) the powerful role of political
dominance in creating alliances and particular marriage patterns,
he explains any such endogamous subdivisions by saying that they
arise through bastardy or the differential status of wives in a
polygynous marriage. Endogamous groups develop within previous
endogamous groups only because of the lower status attached to
marriages with women from outside the proper alliance group.
Politics not only occupies a subordinate position in
Dumont's general theory, but is eclipsed on the one side by the
preeminence of kinship, invaded by social bastardy and caste
hierarchy, and on the by caste, which elevates the -brahman, and
attendant principles of purity and pollution, above the king.
Caste, and the hierarchical principle it entails, is fundamental
because it is religious, and in Indian social thought, according
to Dumont, the religious encompasses the social, the economic,
and the political.
Dumont therefore sees caste authority and political
authority as fundamentally different. He writes that "the notion
of caste and of a superior caste exhausts all available
transcendence. Properly speaking, a people's headman can only be
someone of another caste. If the headman is one of their own,
then to some degree they are all headmen" (Dumont 1957b, 206).
This is true in Pudukkottai in that headmen are at one level
simply primus inter pares in their social group. However, by
virtue of their connection to the king, they do "transcend," at
some level, their own community. Most importantly, the king
himself is at one level simply a Kallar, and not the highest one
at that. But by virtue of his kingship, not caste transcendence,
he is also the transcendent overlord of the entire kingdom.
Hierarchy in Pudukkottai concerns transcendence in the
context of kingship, where the king is both a member of a
segmentary lineage system and the overlord of the entire kingdom.
What would seem contradictory to Dumont is the paradox upon which
the entire caste system rests. Kinship is inflected, at its
core, by politics; and politics is nothing more than the curious
paradox of a king who encompasses all even as he is one of his
own metonyrns. In the social and political world of the little
kingdom, this meant that the king was an overlord, but one who
was nonetheless always embroiled in the strategic concerns of
kinship, status hierarchy, protection and warfare, and in the
maximization of his own honor and sovereign authority within the
little kingdom and in a wider world of other kingdoms and greater
overlords.
Part of Durnont's resistance to acknowledging the political
inflection of caste and kinship may result from the political
marginality of the Pramalai Kallars, a marginality rendering them
far more similar to the unruly Vicenki Nattu Kallars who lived in
the northwestern part of Pudukkottai State than to the royal
subcaste. With both the Pramalai Kallars and the VN Kallars, the
lack of well developed affinal boundaries corresponding to
discrete territorial units, as well as of a distinct sense of the
hierarchy of groups, can perhaps be explained by their incomplete
incorporation within (and therefore inflection by) the political
system of a little kingdom. Everywhere in Tamil Nadu the Kallars
had highly developed notions of territory, but subcaste
organization achieved its particular level of territorial
segmentation and hierarchical articulation in Pudukkottai alone.
And only within the royal subcaste of Pudukkottai did Kallars
develop the pronounced and complex forms of territorial bounding
and hierarchical marking that they did, and which I describe in
great detail elsewhere (1987). Kingship does make a difference.
Some of Dumontls theoretical problems stem from the fact
that he does not pursue an interest in the ethnohistorical
reconstruction of the Pramalai Kallars. He is aware of the
modern decline of headmanship, and that it no longer expresses
itself as fully as it might once have done in the social logic of
Pramalai organization. Characteristically acute, he senses a
correlation with recent political change: "If authority rests on
external sanction, it is to be expected that it cannot maintain
-itself without formal government recognition" (~umont1957b,
203). Unfortunately, he does not consider the possibility that
colonialism, and the attendant break down of the old regime, have
much to do with the development of the separation of religion and
politics which he has identified and reified into a timeless
Indian social theory. A combination of theoretical program,
ethnographic "accident," and historical disinterest have
conspired to render Durnont's understanding of the Kallars,
however brilliant, limited in fundamental ways.
Here and elsewhere I have argued that the social relations
that made up Indian society, far from being "essentialist"
structures predicated on the transcendence of a set of religious
principles, were permeated by "political" inflections, meanings,
and imperatives. Caste, as it is still portrayed in much current
anthropological literature, is a colonial construction,
reminiscent only in some ways of the social forms that preceded
colonial intervention. The structural relations that made up the
"caste system" in Pudukkottai thus reflect--albeit with the
distortions of ethnohistorical time--the ideological proposals of
my informants. These ideological statements consistently
referred to the historical means by and through which meaning was
constructed and maintained. Caste, if ever it had an original
form, was inscribed from the "beginning" by the relations and
conceits of power. And in medieval and early modern south India,
it is clear that Geertz was indeed right: power was what kings
were.

Bthnohistory and Ethnosocioloay

"We need history, but not the way a spoiled loafer in the
garden of knowledge needs it."
Nietzsche, Of the Use and Abuse of Historv.
 When I first began to use the term ethnohistory to describe
the particular blend of history and anthropology that I sought to
practice in my study of India, I thought that "ethno" should do
the same thing to history that it seemed to be doing to
sociology. Of course the place was Chicago, the time the mid-
seventies, and the word was culture. But even then, and even
despite the fact that in my work I sought to construct my sense
of what it meant to do history in light of "indigenous"
historical texts, ethnohistory struggled against itself. Not
only did ethnohistorians seem constantly to pose the questions
about epistemological mediation that began only rather later to
deconstruct the original assumptions of ethnosociology (questions
such as, how does an outsider attain access to or re-present a
culturally specific form of knowledge?), but culture as a domain
was much more difficult to lock up (or off) as a separate area of
inquiry. The injunction, "Always historicize! ," seemed aiways
already there. But, then as now, it was not always clear what
the injunction meant.
Originally, ethnohistory meant the reconstruction of the
history of an area and people who had no written history. As
such, it was used to denote in particular the field of studies
concerning the past of American Indians, and secondarily of other
so-called primitive or pre-literate societies. But, as many have
since demonstrated, ethnohistory cannot be restricted to the
unwritten or oral sources for history in most parts of the world
where texts and written sources exist, even if they do not seem
to penetrate some sectors of society. In India, as in many other
places, there are no pure oral traditions: texts have provided
the basis for tradition as often as the other way round. Indeed,
both texts and traditions relate not only to each other, but also
to historical processes of production and social forms of
contextualization, interpretation, and certification.
Ethnohistory in India is clearly not about the history of
primitive or preliterate people.
As suggested above, ethnohistory is also not simply a gloss
for a cultural analysis of historical sensibilities in India,
whether embodied in texts or traditions. However, part of the
task of ethnohistory is to contest the dominant voice of history,
which in India has always been a Western voice. This voice has
always disparaged India, insisting that the relative absence of
chronological political narrative and the unsettling presence of
myth and fancy are indicative of an underdeveloped sense of
history. Ethnohistory can therefore assist in the project of
recuperating a multiplicity of historical voices, revealing for
India an active, vital, and integral historiographical industry.
I have also argued that ethnohistory can help determine a
culturally specific set of relevancies, moments, and narrative
forms to expand and alter the sense of how to think about India's
past. But this past is never contained solely within the texts
or traditions that would be used for this task. If
ethnohistory is used to situate history, it is always seen as
itself situated in history (see also Dening 1980, 38).
Thus the difficulties in anthropologizing history are not
simply removed by the inverse call to historicize anthropology,
for we never seem to reach explicit agreement about what history
actually means. But in this case the problem and the solution
are integrally interrelated. For if an investigation into the
culture of history has both the strengths and the weaknesses of
ethnosociology, an exploration of what is involved in the history
of culture can assist in making a creative critique of culture
theory, whether in ethnosociology or elsewhere. Not only has
ethnosociology been insufficiently clear about the
epistemological privilege it assumes in its claim to re-present
indigenous forms of knowledge, it has completely excluded a wide
range of historical questions, as also any consideration of the
relations of knowledge and power beyond a narrow cultural domain.
This is not the place, however, to summarize the arguments
of Gramsci, Williams, Bakhtin, Bourdieu, Foucault, and other
theorists who have helped specify and problematize the
historicity of culture. It must suffice here to note that when
culture is situated in history rather than opposed to it the
concept of culture inevitably opens up. I began by using culture
as a method and conceit to privilege the discursive claims of my
Kallar informants in relation to Dumont's proposals about the
nature of hierarchy in India. That is to say, I began by
participating in the ethnosociological claim that if you
investigate native terms and meanings you will find that
hierarchy is about x and not about y. However, the cultural
statements of my informants subverted the autonomy of a presumed
cultural domain, and in particular the opposition between the
political/historical and the cultural/religious. At the same
time, the injunction to historicize, vague though it sometimes
seemed, enticed me to enter the web of power, knowledge, and
history that constituted both the world of reference as well as
the necessary conditions for contemporary cultural discourse.
Culture thus was a conceit that deconstructed itself through its
own historical reference, for culture distilled and displayed
(and often displaced) the historical legacy of its own hegemonic
ascendence.
The necessarily ambivalent position of history within any
ethnohistorical project provides critical access to much of the
current theoretical debate about culture. But ethnohistory
should not simply disparage ethnosociology. For in calling
attention to the hegemony of Western social science,
ethnosociology both set some of the conditions for this kind of
critique in Indian studies, and may yet have the last laugh. The
theoretical concerns about culture articulated above are as
Western in their dominant figures and intellectual histories as
the more positivist social science from which we tried to free
ourselves in previous decades. However, if "history" teaches us
anything at all, it should at least help us dispose of the idea
that culture can exist outside of history, however much this
history--and I suspect any history--is always mediated through a
multiplicity of cultural forms.
 REFERENCES
Appadurai, Arjun. 1981. Rorship and conflict under colonial
rule: A south Indian case. Cambridge: Cambridge University
Press.
Appadurai, Arjun, and Carol Breckenridge. 1976. The south
Indian temple: Authority, honor and redistribution. In
Contributions to Indian Socioloav n.s. 10 (2):187-211.
Dening, G. 1980. Islands and Beaches: discourse on a silent
1774-1880. Melbourne: Melbourne University
Press.
Dirks, Nicholas B. 1979. The structure and meaning of political
relations in a south Indian little kingdom. Contributions to
Indian Socioloay 13(2):169-204.
.
1982. The pasts of a palaiyakarar: The ethnohistory
of south Indian little king. Journal of Asian Studies 41
(4):655-683.

kinadom.
.
1987. The hollow crown: ethnohistorv of an Indian
Cambridge: Cambridge University Press.
Dumont, Louis. 1957. Hierarchy and marriaae alliance in South
Indian kinship. Occasional papers of the Royal ~nthropological
Institute. 12. London.
. 1962. The conception of kingship in ancient India.
Contributions to Indian Socioloav 6. Reprinted in Dumont 1980.
. 1980. Homo hierarchicus: The caste system and
implications. Chicago: University of Chicago Press.
its

.
1986.
University Press.
A south Indian subcaste. Delhi: Oxford

Heesterman, J. C. 1978. The conundrum of the king's authority.

In Kinaship and authority in South Asia, ed. J. F. Richards's.
Madison: University of Wisconsin Publications Series.
Inden, R. 1983. Orientalism and India. Modern ~ s i a nstudies.
Lingat, R. 1973. The classical law of India. Berkeley:
University of California Press.
Ludden, David E. 1985. Peasant historv in south India.
Princeton: Princeton University Press.
Marriott, McKim, and Ronald Inden. 1977. Toward an
ethnosociology of caste systems. In The new wind: Chanainq
identities in South Asia, ed. Ken David. Mouton: The Hague and
Paris.
McGilvray, D. 1982. Mukkuvar vannimai: Tamil caste and
matriclan ideology in Batticaloa, Sri Lanka. Caste ideology and
interaction, ed. D. B. McGilvray. Cambridge: Cambridge
University Press.
Trautmann, Thomas R. 1981. Dravidian kinship. Cambridge:
Cambridge University Press.
 L-
PROGRAM ON THE
C O M P A R A T I V E STUDY O F SOCIAL TRANSFORMATIONS
UNIVERSITY O F M I C H I G A N

kmuuNG PAPW SERIES

The Program on the Comparative Study of Social Transformations is an inter-

d i s c i p l i n a r y research program a t the University of Michigan. Its faculty
associates a r e drawn primarily from the departments of Anthropology, History,
and S ~ ~ i o l o g ybut t a l s o include members of several other programs i n the
humanities and s o c i a l sciences. Its mission is t o stimulate new inter-
d i s c i p l i n a r y thinking and research about a l l kinds of s o c i a l transformations
i n a wide range of pr$sent and past s o c i e t i e s . CSST Working Papers report
current research by faculty and araduate student associates of the program;
many w i l l be published elsewhere a f t e r revision. Working Papers a r e available
f o r a fee of $1.00 for papers under 40 pages and f o r $2.00 f o r longer papers.
The program w i l l photocopy out-of-print Workinu Papers a t cost ($.05 per
page.) To request copies of Working Papers, write t o Comparative Study of
Social Transformations, 4010 LSA Building, University of Michigan, Ann Arbor,
M I 48109-1382 o r c a l l (313) 936-1595.

1 "Prqram i n Comparative Study of Social Transformations," by William H,

Se-11, J ~ . ITerrence J. McDonald, Sherry B. Ortner, and Jeffery M.
~ a i g ,e May 1987, 15 pages. Also CRSO Working Paper #344.

"Labor History, Uneven Development and the Autonomy of P o l i t i c s : The

Dockworkers of Nineteenth-Century Marseille," by William H. Sewell, Jr.,
J u l y 1987, 45 paues. Also CRSO Working Paper #346.
'

"Cbffee, Coppert and Class Conflict i n Central America and Chile: A

Critique of Z e i t l i n ' s Civil Wars in Chile and Zeitlin and R a t c l i f f ' s
Landlords and C a p i t a l i s t s , " by J e f f e r y M. Paige, September 1987, 10
pages. Also CRSO Working Paper #347,

"In Search of the Bourgeois Revolution: The P a r t i c u l a r i t i e s of German

Historyt" by Geoffrey Eley, September 1987, 61 pages. Also CRSO Working
Paper 8350,

"The Burdens of Urban History: The Theory of the S t a t e i n Recent

American Social History," by Terrence McDonald, May 1988, 50 pacjes.
Also CRSO Working Paper 8355.

" History, Sociology, and Theories ,of Organization ," by Mayer N. Zald t
May 19881 42 pages. Also CRSO Workinu Paper #357.

"Have Social Historians Lost the Civil War? Some Preliminary

Demographic Speculations," by Maris A. V i n o v ~ k i sM
~ay 1988, 55 pacjes.
Also CRSO Working Paper 8358.

"Revolution and the Agrarian Bourgeoisie i n ~ i c a r a g u a ~by

" Jeffery M.
Paise, 42 pages. Also CRSO Working Paper 8363,

"Nationalism and Class a s Factors in the Revolution of 1917," by Ronald

G. Sunyt October 1988# 42 pages. Also CRSO Working Paper 8365.
10 "The Original Caste: Peer, History, and Hierarchy i n South Asia," by
Nicholas B. Dirks, October 1988, 30 pages. Also CRSO Workinq Paper
8367.

11 "The Invention of Caste: Civil Society i n Colonial India," by Nicholas

B. Dirks, October 1988, 24 pages. Also CRSO Working Paper #368.

12 " Sociolouy a s a Discipline : Quasi-Science and Quasi-Humanities ," by

Mayer Zald, October 19881 43 pages. Also CRSO Working Paper #369.

13 "Constraints on Professional mter i n Soviet-~ype Society : Insights

from the Solidarity Period i n Poland," by Michael D. Kennedy and Konrad
Sadkowski, November 1988, 37 pages. Also CRSO Working Paper #371.

14 "Evolutionary Changes i n Chinese Culture," by Martin K. Whyte, November

1988, 20 pages. Also CRSO Working Paper #372.

15 "World Market, Class Conflict, and Rural Coercion i n mst-Colonial

Buenos A i r e s , " by Karl Monsma, November 1988, 22 pages. Also CRSO
Working' Paper 8373.

16 "Ritual and Resistance: Subversion a s a Social Fact," by Nicholas B.

Dirks, December 1988, 39 pages. Also CRSO Working Paper #375.

17 "Social Transformations of Gender i n Andean South America: A Working

Annotated Bibliography," by Janise Hurtig, December 1988, 24 pages.
Also CRSO Working Paper #376.
 CENTER FOR RESEARCH ON SOCIAL ORGANIZATION
WORKING PAPER SERIES

CRSO Working Papers report current research and reflection by affiliates of

the Center. Working papers which are still in print are available for a fee of
$2.00 for any paper under 100 pages and $4.00 for papers over 100 pages. The
Center will photocopy out-of-print working papers at cost (five cents per page).
To request copies of working papers, or for further information about the Center,
write us at 4501 LS&A Building, Ann Arbor, hlichigan, 48109, or call (313)764-7487.

359 "Innovative Participation in Neighborhood Service Organizations," by Barry

Checkoway, August 1988, 32 pages. Also PCNA Working Paper #12.

360 "Facing the Challenge of Diversity and Multiculturalism," by Judith H. Katz,

Ed.D., August 1988, 55 pages. Also PCMA Working Paper #13.

361 "Changes in Mate Choice in Chengdu," by Martin King Whyte, September 1988,
32 pages.

362 "Changing Organizati.ona1 Racism: A Workshop for University Staff Leaders,"

by Mark Chesler and Cheryl Hyde, October 1988, 68 pages. Also PCMA Working
Paper #14.

363 "Revolution and the Agrarian Bourgeoisie in Nicaragua," by Jeffery M. Paige,

42 pages. Also CSST Working Paper #8.

364 "Psychosocial and Social-Environmental Problems and Needs for Support

Services of Off-Treatment Survivors of Childhood Cancer," by Mark Chesler,
October 1988, 87 pages.

365 "Nationalism and Class as Factors in the Revolution of 1917," by Ronald G.

Suny, October'1988, 42 pages. Also CSST Working Paper #9.

366 "Organizations, Industry, and the State," by David Jacobs, Michael Useem,
and Mayer N. Zald, October 1988, 34 pages.

367 "The Original Caste: Power, History, and Hierarchy in South Asia," by
Nicholas B. Dirks, November 1988, 30 pages. Also CSST Working Paper #lo.

368 "The Invention of Caste: Civil Society in Colonial India," by Nicholas 8.

Dirks, November 1988, 24 pages. Also CSST Working Paper #11.

369 "Sociology as a Discipline: Quasi-Science and Quasi-Humanities," by Mayer

Zald, October 1988, 43 pages. Also CSST Working Paper #12.

370 "Mediating Ethnic Minority Conflict in the Classroom: The Case of Blacks
and Jews,'' by David Schoem and Marshall Stevenson, November 1988, 44 pages.
Also P(%MA Working Paper #15.

371 "Constraints on Professional Power in Soviet-Type Society: Insights from

the Solidarity Period in Poland," by Michael D. Kennedy and Konrad
Sadkowski, November 1988, 37 pages. Also CSST Working Paper $13.

372 "Evolutionary Changes in Chinese Culture," by Martin K. Whyte, November

1988, 20 pages. Also CSST Working Paper #14.

373 "World Market, Class Conflict, and Rural Coercion in Post-Colonial Buenos
Aires," by Karl Monsrna, November 1988, 22 pages. Also CSST Working Paper
#15.