Race and Descent as Social Categories in India

Author(s): André Béteille

Source: Daedalus , Spring, 1967, Vol. 96, No. 2, Color and Race (Spring, 1967), pp. 444-463
Published by: The MIT Press on behalf of American Academy of Arts & Sciences

Stable URL: https://www.jstor.org/stable/20027046

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

American Academy of Arts & Sciences and The MIT Press are collaborating with JSTOR to
digitize, preserve and extend access to Daedalus

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
ANDR? BETEILLE

Race and Descent as Social Categories in India

I
Although the range of ethnie diversity in India has few parallel
elsewhere, palpable physical differences have, by themselves, co
tributed very little to the country's recent social tensions. The most
significant membership groups in the society are, however, clearly
ones with which the individual identifies himself by reason of birt
and through sentiments of common blood and common ancestry
Such sentiments are likely to have special appeal in a society tha
was until recently relatively closed, and where membership in the
most significant groups could be acquired only by birth. Many out
side observers have echoed the view Max Weber set forward in the
opening sentence of his book on Indian religion: India is "a lan
... of the most inviolable organization by birth."1
Since the time of Manu, Hindu law-givers have attributed par
ticular qualities to particular lines of descent. Strict rules of la
and custom circumscribed intermixture of lines. To preserve the so
cial and cultural identity of each caste, marriages not only between
different castes but often those between subdivisions of the same
subcaste were forbidden. This rule nourished the idea that mem
bers of the same caste or subcaste were of common blood. Indeed,
Irawati Karve has tried to show that the minimal unit of endogam
within the caste system consists of a group of persons related b
real ties of kinship and affinity.2
The sentiment of common descent is not confined to the sub
caste viewed as the minimal unit of endogamy. It is often shared b
a group of related castes among which no intermarriage is pra
ticed. Thus, although the Tamil Brahmins are divided into nume
ous sections, there is a keen belief among them that all Brahmins
belong to a single stock. This is based on the common suppositio
that the subdivisions of today are the outcome of fissions within an
444

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
 Race and Descent as Social Categories in India

original, undifferentiated group?a supposition that may have some

historical validity, at least within a single cultural region. Depend
ing upon context and situation, the sentiment of common descent
can clearly be associated with either a broader or a narrower
group.
The organization of society on the basis of birth-status groups
and the general acceptance of a rank order ascribed by birth gave a
distinctive character not only to the absorption of alien ethnic ele
ments into Hindu society, but also to the place they acquired in it.
In the main, Hindu society seems to have grown by adding new
blocks to itself, while allowing these blocks to retain a measure of
autonomy and identity. Traditionally the unit of absorption has
been a community rather than an individual.8
This process often enabled an alien community to retain its re
ligious or sectarian identity while imbibing a variety of Indian cul
tural forms. No alien community could, in fact, function as a part of
Indian society without acquiring some of the structural properties
of castes. Many people have noticed a caste-like organization
among both Moslems and Christians in India. Also, reform move
ments that have sought to break through the closed system of castes
have themselves generally ended as castes. The Lingayats of South
India provide the best-known, though by no means the only, ex
ample of this.
The discreteness of the blocks out of which traditional Indian
society was built was maintained by a combination of ecological,
genetic, linguistic, and other cultural factors. The localization of a
community in a particular habitat came about because of the ab
sence of easy means of transport and communication. The concen
tration of large tribal blocks in certain parts of the country provides
a good example of this tendency. Genetic identity was maintained
at least in part by the rule of endogamy. The preservation of cul
tural identity was made possible by the pluralistic values of Hindu
society.
In the past and to a great extent in the present, the Indian's iden
tity has been fixed in a fairly definite manner by a variety of
bounded units?each characterized by not one, but a number of
different attributes. N. K. Bose writes :

The map of Calcutta . . . shows a highly differentiated texture. Ethnic

groups tend to cluster together in their own quarters. They are dis
tinguished from one another not only by language and culture but also by
broad differences in the way they make their living. Naturally there is a
445

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
ANDR? B?TEILLE

considerable amount of overlap, but this does not obscure the fact that
each ethnic group tends to pursue a particular range of occupations.

It can be said, therefore, that the diverse ethnic groups in the population
of the city have come to bear the same relation to one another as do the
castes in India as a whole.4

Because of this kind of clustering, social identities have been pre

served to a much greater extent in India than in the more rapidly
changing societies.
In the traditional system there was a close association between
caste, on the one hand, and economic and political power, on the
other. Each village or group of villages had one or a few dominant
castes whose members exercised control over both the land and the
political system. Concomitantly, there were other castes whose
members were almost wholly landless and powerless; most con
spicuous among these were the Untouchable castes. Sometimes the
gradations of political and economic power were almost as elab
orate as those of the caste hierarchy. Recent studies by anthropol
ogists show that the conflict among castes is often a conflict over
landownership.5 Such conflicts tend to be localized, however, be
cause the landowning caste varies from one area to another. In de
scribing the correspondence between caste and landownership in
Kerala, M. N. Srinivas writes:
At the top of the hierarchy were the Nambutri Brahmins who were non
cultivating owners (jenmi). The "high" N?yar castes were the non
cultivating lessees of Nambutri land on twelve-year leases (k?nam). The
agricultural labourers, both tied and free, came from the lower castes like
Cerum?n and Pulayan and from the P?nan tribes.6
Even today many castes bear occupational names. In the tra
ditional system, caste and occupation were closely associated,
particularly among artisan and service groups. Thus, potters,
blacksmiths, carpenters, barbers, and washermen were not only oc
cupational groups but also hereditary castes. The specific association
between caste and traditional occupation is breaking down, al
though a broad relationship continues to exist. Occupations are
graded by traditional values in an elaborate hierarchy in which
ideas of purity and pollution play a significant part. Upper castes
generally avoid manual occupations, particularly those that are con
sidered polluting or onerous. Scavengers and sweepers are still al
most wholly drawn from the ex-Untouchable castes.
The close association between caste and traditions of literacy
largely explains the extent to which a few upper castes came to
446

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
 Race and Descent as Social Categories in India

dominate the variety of urban-middle-class occupations that de

veloped during the nineteenth century. Western education and,
through it, the new occupational structure were the preserve of a
few upper castes. In Madras, the civil service and the professions
were dominated almost completely by Brahmins until World War I;
the situation was broadly similar in Bombay. In Bengal, Brahmins,
Kayasthas, and Baidyas provided a very high proportion of recruits
to the new middle class. Higher education and salaried occupations
are more open today, but there is still a wide gap between the
castes with traditions of literacy and those without.
Western education gave an advantage not only to certain castes
but also to certain regions in the country. The British first estab
lished their rule in the three Presidencies of Bengal, Bombay, and
Madras, and their capitals became centers of Western education.
Bengalis who graduated from the University of Calcutta spread out
to other parts of the country and manned the professions and
bureaucracies there. Until recently sizable sections of the upper
middle class in large urban centers in the states of Assam, Bihar,
and Orissa were Bengalis. Likewise, Tamilians from Madras domi
nated the administration and the professions in Andhra and Mysore.
Since Independence, and particularly since the establishment of
linguistic states, there has been a certain amount of tension between
such outsiders and the "sons of the soil."
Certain religious communities also benefited more than others.
The Parsis, who are a small and highly urbanized community, be
came the most Westernized section of Indian society. The Chris
tians, in those areas where their economic position was not too
backward, took advantage of schools and colleges set up by mis
sionaries. In certain parts of Kerala they now occupy the highest
position in the new occupational structure. The Moslems, who are
a much larger group, at first lagged behind in education, and this
was to a large extent responsible for the backward nature of the
Moslem middle class in relation to the Hindus, particularly at the
turn of the century.
Today, no ethnic, religious, or linguistic group in India is eco
nomically homogeneous. In general, the larger the group, the more
heterogeneous it tends to be. The economic dominance of any such
group is almost always a local affair.

Religious and linguistic groupings as well as caste affiliations

447

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
ANDRE B?TEILLE

reinforce popular notions of descent in India. Each of the dozen or

so major linguistic groups has a "homeland" within the country,
and, in fact, the states that make up the Indian union consist es
sentially of people who speak the same language and share a com
mon history. The territorial concentration of language groups gives
a particular edge to regionalism in India.
Cutting across the linguistic divisions, and in many ways just as
important socially, are groupings based on religious affiliation. Al
though the Hindus constitute over 80 per cent of the population,
there are important Moslem and Christian minorities in several
regions. The concentration of people following a certain religious
faith in a particular social stratum or region helps to sharpen their
social identity. In East Bengal (now East Pakistan) prior to the
partition of India, the Hindu-Moslem conflict was greatly accentu
ated because the Hindus owned most of the land while the bulk of
the Moslems tilled it. Again, the fact that the Sikhs not only speak
a common language but are also concentrated in a particular ter
ritory has been made the basis of a demand for a separate state
(which has recently been conceded).
Although units based on regional, linguistic, sectarian, caste, or
lineage affiliation have played a very important part in traditional
Indian society and continue to do so today, they do not exhaust the
types of groups in modern Indian society. Universalistic groups?
such as those based on affiliation to classes, parties, or trade unions
?exist, but have, by their nature, little to do with race or descent
II
No significant social unit in India?whether based on language,
religion, or caste?is racially homogeneous; it is more common for
physical and social differences to intersect than to overlap. Studies
by P. C. Mahalanobis, C. R. Rao, D. N. Majumdar, S. S. Sarkar, and
others7 have shown how complex the racial pattern may be even in
a limited region or sector of society. Sarkar has demonstrated that
even tribal groups?generally assumed to be homogeneous?show
a great deal of internal variation when examined anthropometrically
and serologically.8
There is also a wide range of diversity in the distribution of phys
ical types. Whichever trait one considers?whether skin color, stat
ure, nasal index, or head form?one encounters a very broad spec
trum in which almost every variation is represented. In general,
physical types are not sharply differentiated but represented on a
448

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
 Race and Descent as Social Categories in India

continuum. Nowhere are the somatic differences between language

groups, religious communities, or castes so sharp as the differences
between Negroes and whites in the United States. Along with di
versity, there is a relative absence of polarity.
The high degree of overlap between adjacent segments and the
absence of polarity go a long way in explaining why there has been
little or no organized conflict between "racial" groups in Indian
society. In spite of the rigid rules of endogamy, no meaningful so
cial unit can be readily identified by its physical characteristics.
Clear physical types do not exist in India in the shape of concrete
groups as they do in the United States or South Africa. They are,
rather, constructs that enable anthropologists to order their data.
Knowledge of "racial" differences and their distribution in India
is, consequently, very limited. The anthropom?trie and serological
data available do not permit one to present a racial map of the
country with any confidence. Sir Herbert Risley, under whom the
1901 Census of India was conducted, made the first systematic at
tempt to classify the population by race.0 In his classification, Risley
tried to relate certain physical differences?in particular, those per
taining to the shape of the nose and skin color?to certain basic
features of social stratification. The main racial types in his scheme
of classification?Aryan, Dravidian, Aryo-Dravidian, Mongolo
Dravidian?were really linguistic or regional categories in disguise.
Risley's classification was criticized, in part, because of his failure to
distinguish clearly between physical and social categories.
Several classifications have been offered since, and of these B. S.
Guha's seems to be the most popular.10 Guha identified six principal
racial elements in the population of India: Negrito, Proto-Austra
loid, Mediterranean, Western Brachycephal, Palae-Mongoloid, and
Nordic, the first and last occurring only in very diluted forms, if at
all. Only the Proto-Australoid and Palae-Mongoloid groups have any
real correspondence with social divisions in the country, the first be
ing well represented in the aboriginal population of Central and
Southern India and the second in the people of Assam and the Him
alayan foothills. The Mediterranean and Western Brachycephal
types are too broad and general to have any close connection with
real social divisions.

Ill
Sharp physical differences may exist in India, but such differ
ences are rarely found between territorially or structurally adjacent
449

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
ANDR? B?TEILLE

sections of the population. One could no doubt make out a sharp

contrast between, say, the Kashmiri Brahmins and the Paraiyas of
Madras. These units do not, however, constitute by themselves a
meaningful field of social interaction. The social distance between
them is spanned by many groups whose existence serves to reduce
the contrasts between the two ends of the continuum.
When skin color is taken as the criterion, regional and caste dif
ferences appear to be of particular importance. The inhabitants of
the Northern states?particularly Punjab, Jammu and Kashmir, Ra
jasthan, and parts of Uttar Pradesh?are on the whole fairer than
those of the Southern states. Indeed, many North Indians have a
vague prejudice against South Indians because of their dark skin
color.11 Similarly, people from the topmost castes are generally
fairer than the Harijans. But there are numerous intermediate castes,
and among these one frequently encounters persons who are either
darker than some Harijans or fairer than many Brahmins. Further,
regional and caste differences often cut across each other. People
from the lower castes in North India tend to be, on the whole,
darker than those from some of the highest castes in the South.
Over the centuries, environmental factors have probably evened
out to some extent the physical differences within any one region.
As new ethnic elements were absorbed continuously and gradually,
they are likely to have undergone some change due to geographical
factors and probably also to a certain amount of intermixture. It
seems almost certain that some intermixture between castes through
unsanctioned unions has been a pervasive feature of traditional
Indian society. Only this would explain both the high degree of
overlap between structurally distant segments and the very con
siderable measure of heterogeneity within subcastes that in theory
have bred true for centuries.
A few subcastes are well known for the distinctive physical fea
tures of their members. The Chitpavan Brahmins of Maharashtra
and the Saraswat Brahmins of Mangalore have not only light skins
but occasionally light eyes as well. Beyond a certain point, how
ever, it is difficult to differentiate between facts and stereotypes.
Thus, the Iyengar Brahmins of Madras, though reputed to be fight
skinned, include a very large number of dark individuals.
Within the subcaste, a certain amount of selective breeding has
no doubt taken place for centuries. A light skin color is valued al
most universally. Wealthy landowning families often have a tradi
tion of seeking light-skinned brides from among poorer members of
450

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
 Race and Descent as Social Categories in India

their subcaste. It is very common to find a high concentration of

light-skinned people among established landowning families. This
feature is particularly conspicuous among aristocratic Moslem fami
lies in North India, which probably also contain a high component
of foreign blood.
Caste and aristocratic status as determined by landownership
often cut across each other. Non-Brahmin castes tend to be, on the
whole, dark skinned. Where established landowning families exist
among them, however, their members are often as light skinned as
the Brahmins and sometimes more so. Where members of aristo
cratic landowning f amilies also belong to the highest caste, they are
likely to be particularly fair.
Anthropologists have made some attempts to study systemati
cally the relationship between the physical diversity of the popula
tion and its social diversity. Sir Herbert Risley saw the caste system
as providing unique possibilities for the differentiation of physical
types. He considered this differentiation to be closely related to the
rank order of castes.12 In concrete terms, he argued that the social
rank of a caste varies inversely with the average nasal index of its
members; in other words, members of the upper castes are narrow
nosed while those of the lower castes are broad-nosed.
Risley's views have been challenged by Ghurye, Majumdar, and
others.18 A closer examination of even the limited anthropom?trie
material at hand shows that the relationship between the two sets
of factors is extremely complex and that Risley's generalization was
sweeping and hasty. Even within a region, the exceptions to the
correlation suggested by him are numerous. When inter-regional
comparisons are made, it seems to break down altogether.
Although Risley's correlation may be untenable from the an
thropom?trie point of view, he was no doubt correct in drawing at
tention to an important fact?that in India, as elsewhere, high so
cial values are attached to certain physical traits. Among those that
are valued most highly, fair skin color occupies a conspicuous po
sition.
In many Indian languages the words fair and beautiful are often
used synonymously. The folk literature places a high value on fair
skin color. The ideal bride, whose beauty and virtue are praised in
the songs sung at marriages, almost always has a light complexion.
A dark girl is often a liability to her family because of the difficulty
of arranging a marriage for her. Marriages among educated In
dians are sometimes arranged through advertisements in the news
451

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
ANDR? B?TEILLE

papers; even a casual examination of the matrimonial columns of

such popular dailies as The Hindu, The Hindustan Times, or The
Hindustan Standard shows that virginity and a fight skin color are
among the most desirable qualities in a bride.
The caste system has given birth to a variety of stereotypes that
have a bearing on social conduct, although their influence on it is
less marked now than in the past. Some of these stereotypes dwell
on the physical features of the different castes, the upper castes
being always represented as fair and the lower castes as dark.
Further, a reversal of the assumed correlation is viewed as not only
unusual but sinister. A Kannada proverb cautions, 'Trust not a dark
Brahmin or a fair Holey a"; a North Indian proverb maintains, "A
dark Brahmin, a fair Chuhra, a woman with a beard?these three
are contrary to nature"; and another North Indian proverb runs,
"Do not cross a river with a black Brahmin or a fair Chamar."14
In spite of much evidence to the contrary, the belief that some
castes are fair and others dark is very widespread, particularly in
South India. During fieldwork in a Tanjore village, I lived among
a very orthodox section of Brahmins. The Tanjore Brahmins have
an overweening pride in their relatively light skin color. My hosts
often remarked that since I looked like a Brahmin, they had no very
strong objection to my living among them. Once I jokingly told one
of my hosts that when I came next, I would live among the Kallas?
a fairly low but influential Non-Brahmin caste?and pass myself off
as a Kalla. He scoffed at me, saying, "Nobody will believe you
you are not black like those fellows."
Refinement of features of the kind that is valued by the Tanjore
Brahmins is probably the outcome, in part, of a certain style of life.
In the rural areas the Brahmins generally lead a sedentary existence
and are not exposed to sun and rain to nearly the same extent as the
average non-Brahmin peasant. I found that one of the Brahmins in
the village where I lived could often identify a Kalla or a Harijan
by his looks. The same Brahmin pointed out, however, that a Kalla
friend of his, an educated lawyer from a neighboring town, looked
quite different from the Kallas in the village. He was refined enough
to pass for a Brahmin. Perhaps there is room for a cultural theory of
appearance that goes beyond the conventional anthropom?trie in
dices and probes into these subtle differences of carriage and ex
pression which have such an important bearing on social inter
change.
In a society where purity of descent is palpably associated with
452

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
 Race and Descent as Social Categories in India

diversity of physical type, features that do not correspond to the

ideal must be accommodated. Among South Indian Brahmins, a
dark girl has a low value in the marriage market. At the same time,
a dark Brahmin girl will always (or almost always) be preferred
among them to a Non-Brahmin girl, however fair. Wherever phys
ical differences cut across caste lines?as they frequently do?the lat
ter are assigned far greater weight in almost every institutional
sphere.
While there is clearly a preference for light skin color in almost
all sections of Indian society, it is difficult to say exactly how far
this preference influences social action. The most concrete expres
sion of it is to be found in the choice of marriage partners. The
choice of a fair bride (or groom) must, however, be made within
limits that are strictly defined by considerations of other kinds.
In certain parts of North India, Moslem women are often very
light skinned and have features that are positively valued. Never
theless, a Moslem bride, however fair, would not normally be ac
ceptable in a Hindu household. Likewise, a Kashmiri bride would
not be acceptable in a Tamil Brahmin household in spite of her very
light complexion. Thus, physical features of a particular kind are
only marginally important, other things (or, at least, certain other
things) being equal. Fair skin color has, for instance, much greater
weight in choosing a bride than a groom. In the case of the latter,
other qualities such as wealth, occupation, and education play an
important part. A dark-skinned son is not so much of a liability to a
middle-class family as a dark-skinned daughter, for he can more
easily acquire other socially desirable qualities.
The caste system in India has often been compared with the
system of Negro-white relations in the United States. Recently
there has been some argument as to whether the American system
can be viewed as a caste system. In a comparative study of "caste"
in India and the United States, Kingsley Davis differentiates
sharply between what he calls "racial" and "non-racial" caste sys
tems:

A non-racial caste system, such as the Hindu, is one in which the criterion
of caste status is primarily descent, symbolized in purely socio-economic
terms; while a racial system is one in which the criterion is primarily
physiognomic, usually chromatic, with socio-economic differences im
plied.15
Davis rejects the view that race has very much to do with the caste
system as it operates in India today.
453

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
ANDR? B?TEILLE

The hypothesis that the Hindu system began on a racial basis is unproven.
Even if true, however, it does not alter the fact that today this system is
for the most part purely a matter of descent rather than race, symbolized
in socio-economic terms.16

Davis has drawn attention to an important difference, but his

formulation of it is not wholly satisfactory. "Race" and "descent"
do not refer to mutually exclusive categories; in fact, their conno
tations overlap to a very considerable extent. One can argue that
descent is a social category, whereas race is a biological category.
This is, however, a distinction the anthropologist makes to facili
tate analysis of a certain kind. It is not clear how much such a dis
tinction helps in the analysis of native categories of thought in
which both descent and race are socially defined.

IV
The ideas of descent and race are inseparably combined in the
concept of jati.17 Although the word jati is understood by most
Western sociologists to mean "caste," it often has a much wider
connotation. In the Bengali language (and in some other North
Indian languages as well), it not only signifies caste, but also comes
closer to the meaning of "race" than perhaps any other ward in pop
ular usage. Basic to both race and caste is the idea of common des
cent from which, in fact, the Sanskrit word jati derives its root
meaning.18 This word is applied not only to "race" and caste, but to
practically every closed group (such as a linguistic or religious com
munity) that is believed to be based, however loosely, on common
descent.
Anthropologists have in recent years made significant advances
in the understanding of social structures through the analysis of
the meanings of indigenous category words. E. E. Evans-Pritchard's
analysis of the Nuer word for "spirit"19 and E. R. Leach's analysis
of the Trobriand word tabu20 offer cases in point. The term jati is a
basic category word in most of the Indian languages. An analysis
of its different meanings and their relationships is likely to offer
certain valuable insights into the principles of Indian social struc
ture.
The caste system evinces several levels of differentiation. To
those who participate in it, these levels have something basic in
common.21 The idea of "community" attaches to them all, although
not with the same degree of intensity. Consequently, the word jati
(and frequently the English term caste) is, according to the con
454

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
 Race and Descent as Social Categories in India

text, applied to a group of castes, a caste, a subcaste, or a subdi

vision of a subcaste. The term jati may even be extended to cover
the entire Hindu community when it is viewed as a unit in opposi
tion to other units of a like order. In Bengal, for instance, it is com
mon to use the word jati (or, rather, the colloquial form jat) to
differentiate not only between Brahmins and Shudras but also
between Hindus, Moslems, and Christians.
In one important sense, the difference between Hindus and
Moslems is of the same kind as that between Brahmins and Shudras,
or between various sections of Shudras. Hindus and Moslems (or
Christians) together form a system that is not very different from
the caste system. The idea of jati provides a clue to the under
standing of the similarities between the two. A person who has
lived in a mullicaste Hindu village in, say, Tamilnad will be struck
by the many features that this village shares in common with the
composite Hindu-Moslem villages still found in many districts of
Bengal. In a Tamil village Brahmins and Non-Brahmins live in
separate residential areas. Each unit is differentiated from the other
by certain peculiarities of dress, food, speech, and ritual, and each
generates a feeling of community.
Similarly, in a Bengal village, Hindus and Moslems live apart
and are differentiated from each other by distinctive social prac
tices and by a consciousness of community. Hindus frequently refer
to the entire Moslem community by the term jati even as they refer
to Untouchables and Tribals by the same term. In the villages in
which I lived in Bengal I frequently asked people about the jatis
present, and the list generally included?in addition to Brahmins,
Sadgopes, and Aghuris (the Hindu castes)?Moslems, Harijans,
and Adivasis.
In the urban areas, too, it is common for people to use the word
jati to differentiate Hindus from Moslems. In many instances, the
word "Hindu" has been entered against the information relating to
caste in a census or questionnaire. If there is so much confusion
even among educated Indians regarding the true referents of reli
gion, caste, or sect, I suggest it is because the social organization
of these different kinds of units has something fundamental in com
mon.

In areas where there is a concentration of Christians, as in

Kerala, the Christians often operate as caste-like units in relation to
Hindu castes. Members of smaller religious denominations such as
Jains, Kabirpanthis, or Brahmos are not only viewed as jatis but
455

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
ANDR? B?TEILLE

frequently enumerated as castes. (The English term caste has gen

erally a narrower referent than the word jati or its regional variant.)
The word jati thus refers not only to subcastes, castes, and caste
groups, but also to religious communities. In some North Indian
languages, it may refer, in addition, to a linguistic group that is a
unit at once broader and narrower than a religious community. The
Bengalis, for instance, commonly use the term jati to differentiate
between themselves and people from other regions such as Ori
yas, Assamese, and South Indians. There are stereotypes relating to
linguistic groups as well as to castes, the former being generally
derogatory.
Linguistic differences are sometimes expressed directly in a
caste idiom. Adrian Mayer has shown that differences between sub
castes in Central India are often based upon regional (that is, lin
guistic) affiliation. "Subcastes are mostly based on provincial dis
tinction; they will be the Malwi, the Gujarati or the Mewari sub
castes, coming from the parts of India bearing these names."22
Throughout Peninsular India the principal cleavages among Brah
mins derive from two factors: sectarian affiliation and linguistic
origin. An immigrant group that comes from a different linguistic
region is generally fitted into local society, but the fact that it is of
a separate stock is never entirely lost sight of.
The word jati may thus be applied to units based on race, lan
guage, and religion as well as to castes in the narrower sense of the
term. These different kinds of identity are easily confused as is il
lustrated by a common remark I used to hear in Bengal where I
grew up. It would be said of a person: "He is not a Bengali, he is a
Moslem (or Christian)." The speaker was, of course, a Bengali
Hindu, and the person spoken about a Bengali born in a different
religious community. The remark illustrates that in certain contexts
the Bengali Hindu views himself as the Bengali par excellence and
others as being different. It does not mean that he cannot in other
contexts view the Bengali Moslem as being of the same stock as he
himself.
Differences in religion, language, and caste give rise to tensions
and conflicts of various kinds. A noticeable feature of recent Indian
politics is the predominance of what is generally referred to as
"communalism." Communal politics has many forms, one or
another of which has been ascendant at different times. In its orig
inal use the term signified primarily the politics of religious com
munities and of Hindu-Moslem relations, in particular. Later the
456

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
 Race and Descent as Social Categories in India

meaning was extended to cover "regionalism," "casteism," and "trib

alism."
Many Indians would deny the primacy of the communal ele
ment in Indian politics and argue instead that communalism?what
ever its form?can be readily reduced to more basic economic con
flicts. Communalism was no doubt heightened by the manner in
which the British transferred power to the Indians, but its struc
tural basis had deep roots in Indian society. The record of the con
flict between Hindus and Moslems from the creation of the Muslim
League in 1906 through the violence and bloodshed of 1946-47 to
the partition of the country in 1947 is too well known to bear de
tailed repetition. The one point that requires emphasis here is that
the Muslim League, basing its political program on the Two Na
tion Theory, maintained that Hindus and Moslems were different
by religion, culture, and race. They consequently urged that the
only rational course the British could adopt before leaving was to
divide the country between the Hindus and the Moslems. The In
dian National Congress under Gandhi, on the other hand, main
tained that Hindus and Moslems formed inseparable parts of a
single community and that partition would be suicidal to both.
That the Muslim League finally won its case and the country
was partitioned does not, of course, prove the objective validity of
its claims as against those of the Congress. Still, the events that took
place in the country just before and after its partition do show how
powerfully men can be moved by appeals to "community," "blood,"
a common way of life, and a common destiny.
As many had foreseen, the partition of India did not settle the
Hindu-Moslem question. India still has a Moslem population of
over 50 million people. The Muslim League not only survives as a
watchdog of their political interests, but is on occasion even used by
the ruling party for its electoral program. Moslems in India con
tinue to preserve their social and cultural identity as do many
other groups based on religious and other "communal" criteria. The
case of the Moslems is, however, somewhat unique. Many of them
have social and ideological finks with Pakistan, and also perhaps a
vague hope of finding a better life and a more secure home across
the frontier. While Moslems in India may not always live with the
feeling of belonging to an alien race, on occasion such feelings are
not only aroused but become the source of violence and bloodshed.
The Hindu-Moslem conflict provided the pattern of communal
politics in North India, but in the South this was provided largely
457

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
ANDR? B?TEILLE

by the Non-Brahmin movement. Kerala and certain parts of An

dhra are the only areas in South India that have organized Moslem
minorities. It is perhaps no accident that the Non-Brahmin move
ment was never quite so important in these areas as in Madras and
Mysore. That the conflicts between Hindus and Moslems and be
tween Brahmins and Non-Brahmins are examined within the same
framework does not, of course, mean that they are in all ways simi
lar or that they are mutually exclusive.
The Non-Brahmin movement was formally launched in 1916
with the issue of a Manifesto and the formation soon afterwards of
the Justice Party with its headquarters in Madras. The movement's
principal objective was to protect the interests of the Non-Brah
mins in education, employment, and political life against domina
tion by the Brahmins. It achieved considerable political success
within a short time and provided a meeting point for influential
Non-Brahmin leaders from all over South India. In fact, a dis
tinctive feature of the first phase of the Non-Brahmin movement in
the South was its success in cutting across linguistic divisions and
uniting in a common cause against the Brahmins people speaking
Tamil, Telugu, Kannada, and Malayalam.23
Soon after the formation of the Justice Party, its leaders began to
make representations to the British government for a share in the
affairs of the state proportional to their population in it. They ar
gued that the Madras Presidency was dominated by a small Brah
min oligarchy, that Brahmins and Non-Brahmins were different by
race and culture, and that unless the British government inter
vened no way could be found of reconciling their divergent inter
ests. In the words of the Memorandum presented by the Non
Brahmin delegation to the Joint Parliament Commission in Britain
in 1919:
Their customs and manners are essentially different, and even in the
matter of food, the two classes differ widely. Their interests are often not
identical. Over and above all this, there is the fundamental difference
which goes to the root of the whole problem?the two people belong to
two different races.2*

Before long, the political agitation of the Justice Party was

transferred to the social plane by the Self-Respect movement which
emerged in the mid-twenties under the leadership of the Non
Brahmin ex-Congressman E. V. Ramaswami Naicker. This move
ment urged Non-Brahmins to rid themselves of the social and
economic domination of the Brahmins, and it gathered strength
458

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
 Race and Descent as Social Categories in India

during the 'thirties. In the 'forties Naicker created the Dravida

Kazhagam, a militant organization that openly preached a racialist
ideology and made the "Aryan" Brahmins their principal target. A
certain amount of violence was used against the Brahmins who
lived for some time under fears that seem to have been greatly
exaggerated. Indeed, I have heard Tamil Brahmins refer to them
selves (rather melodramatically) as the Jews of South India.
The Non-Brahmin movement had deep economic roots, as its
leaders were the first to point out. In general, the Brahmins enjoyed
a comfortable economic position as landowners, government of
ficials, and members of the professions. Still, economic differences
frequently cut across caste boundaries, and the Justice Party
counted in its ranks the biggest landowners in the Presidency.25 The
unique position of the Brahmins in South Indian society is, in fact,
based upon differences that are far more pervasive than the purely
economic ones. Nowhere else in the country are they so sharply dif
ferentiated in speech, dress, and appearance from the rest of the
population as in the South. These differences, along with the tra
ditions of their North Indian origin, lent themselves easily to the
myth that the Brahmins were "Aryans," living on the toil of the
"Dravidian" masses.
Regionalism is today viewed as perhaps the most serious threat
to the national unity of India. The demand for a separate Dravid
ian state has close historical links with the Non-Brahmin movement,
and was first put forward in an organized way by the Dravida
Kazhagam. Its principal proponent today is the Dravida Munnetra
Kazhagam (D.M.K.), the largest opposition party in Madras. Al
though the leaders of the D.M.K. claim to speak for all the four
South Indian states, the party's following outside Madras is very
limited.
The four Southern states?Andhra, Kerala, Madras, and Mysore
?share a number of distinctive features that mark them out from
the rest of India. The unity of the South is cultural as well as geo
graphical; the South Indian languages belong to the Dravidian
family, whereas those of the North are Indo-Aryan. The confusion
of linguistic and racial categories, which was such a common fea
ture of the earlier ethnology, has no doubt contributed to the popu
lar belief in the existence of a separate Dravidian race with Pen
insular India as its homeland. This belief has been worked into an
elaborate mythology by the Dravidian movement in its campaign
for a separate political status for the four South Indian states.
459

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
ANDR? B?TEILLE

Leaders of the D.M.K. lose no opportunity to demonstrate the

threat to the "national" identity of the South. This national identity
has been created largely out of a history and ethnology to which
fact and fancy have contributed in about equal measure. It is kept
alive by the repeated assertion of the historical, cultural, and racial
identity of the Dravidian people. As M. G. Ramachandran, the
noted film star and D.M.K. leader, wrote in Homeland, the official
organ of the party:
The Dravidians, constituting of [sic] the Tamils, Telugus, Malayalees
and Kanarese are a distinct race. They have a common linguistic origin.
. . . Their food habits are alike. They think alike, live alike, do alike and
act alike.26

Characteristically, the D.M.K. has also sought to justify its

separatist demands on grounds of economic and political domina
tion by the North. Economic development has been unequal in dif
ferent parts of India, and the North appears to have benefited more
than the South. This, however, is hardly the whole story. There are
states in North India that are backward in relation to Madras or
Mysore. Tamilians complain that commerce and industry in their
state is often controlled by outsiders, but the principal threat as
they view it today is of "Hindi imperialism." The imposition of
Hindi as the national language is likely to turn the tables on the
South Indians who have so far used their superior knowledge of
English to maintain a fairly comfortable position in government
and other services.
India's tribal population, over 30 million people, enjoys a cer
tain special legal and political status in the country. The tribal pop
ulation is by no means homogeneous, but sharply divided region
ally, culturally, and racially. In the past, the tribes were not only
fragmented, but lent themselves to slow, continuous absorption into
Hindu society. This tendency has today been reversed to some
extent, being replaced in part by a conscious effort to create and or
ganize a separate tribal identity in order to achieve certain eco
nomic and political objectives.27
The roots of pan-tribalism go back to the nineteenth century,
although the passage of the Government of India Act of 1935
helped to give it a much more organized form. By creating separate
Scheduled Areas for the tribal people, the British helped to sharpen
their social identity and to strengthen the feeling that they had a
different political destiny from the other Indians. Most tribal people
seem always to have had an ambivalent attitude toward the Hindus
460

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
 Race and Descent as Social Categories in India

under whose shadow they lived. On the one hand, there was a de
sire to emulate their superior material techniques and more so
phisticated cultural forms; on the other, an attitude of suspicion,
bitterness, and even hatred toward them on the ground that they
were aliens and exploiters. The Hos, one of the major tribes of
Chota Nagpur, use the word diku ( a corruption of dacoit, meaning
"brigand") to refer to the Hindus who are their neighbors.
In the Chotanagpur area of Bihar which has a large concentra
tion of tribal people, the Jharkhand Party was organized to put
forward the demand for a separate tribal homeland. Pan-tribalism
has developed its own symbols and mythology. The very words
used by the tribal people to characterize themselves (Adivasi,
"original inhabitants"; Adimjati, "original race") have a strong con
notation of folk appeal. Attempts have been made to recreate a
largely imaginary past in which the purity and vitality of tribal life
have not been sapped by aliens from the plains. It is probable that
tribal separatism has drawn part of its inspiration from the work of
Christian missionaries. A belief current among the tribal people of
Chotanagpur is that they are the descendants of one of the lost
tribes of Israel.

V
The groupings discussed here have very little correspon
with race in the technical sense of the term. But sociological
sis is concerned not so much with the scientific accuracy of id
with their social and political consequences. Physical anthr
gists are now gradually coming to discard the term race itsel
to use in its place such "neutral" terms as population or br
unit, but the sociologist cannot afford the luxury of confini
analysis to value-neutral categories. He must penetrate the co
the fundamental values and categories of a society to reveal t
meanings?however vague and contradictory?and to sho
they govern and direct social action.
Fundamental categories such as race or jati are often inher
ambiguous. This ambiguity enables people to use the idea to i
different?even conflicting?loyalties in different situations. T
it may be invoked to unite Bengali Hindus and Moslems a
the Assamese, and, in a different historical context, to divid
dus from Moslems in Bengal.
Although the word jati has perhaps a wider referent th
English term race, in popular usage race has a very broad ref
461

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
ANDR? B?TEILLE

indeed. To speak of a "Dravida-jati" would appear to be no more

unreasonable than to speak of a Semitic race or a Slavonic race.
The idea of an Aryan race has been invoked by political movements
in both India and Europe, but, ironically, almost diametrically op
posite values have been placed on it.

References

1. Max Weber, The Religion of India, Hinduism and Buddhism, trans. H. H.

Gerth and Don Martindale (Glencoe, 1958), p. 3.
2. Irawati Karve, Kinship Organization in India (Poona, 1953).

3. N. K. Bose, "The Hindu Method of Tribal Absorption," Science and Cul

ture, Vol. 7, No. 7 (October, 1941), pp. 188-94.

4. N. K. Bose, "Calcutta: A Premature Metropolis," Scientific American, Vol.

213, No. 3 (September, 1965), p. 102.
5. Bernard S. Cohn, "The Changing Status of a Depressed Caste," Village
India, ed. M. Marriott (Chicago, 1955); Dagfin Silversen, When Caste
Barriers Fall (New York, 1963).

6. M. N. Srinivas, "Social Structure," The Gazeteer of India, Vol. 1 (New

Delhi, 1965), pp. 511-12.
7. P. C. Mahanlanobis, "A Revision of Risley's Anthropom?trie Data,**
Samkhya, Vol. 1 (1933), pp. 76-105; C. R. Rao and D. N. Majumdar,
Race Elements in Bengal (Bombay, 1960); D. N. Majumdar, Races and
Cultures of India (Bombay, 1958); S. S. Sarkar, Aboriginal Races of India
(Calcutta, 1954).
8. Sarkar, Aboriginal Races of India.
9. Herbert Risley, The People of India (Calcutta, 1908).

10. B. S. Guha, "Racial Elements in the Population," Oxford Pamphlets on

Indian Affairs, No. 22 (Bombay, 1944).
11. See, for instance, Seminar, No. 23 (July, 1961), pp. 10-11.
12. Risley, The People of India.

13. G. S. Ghurye, Class, Caste, and Occupation (Bombay, 1961); Majumdar,

Race Elements in Bengal.
14. Quoted in Risley, The People of India, p. xxviii.

15. Kingsley Davis, "Intermarriage in Caste Society," American Anthropol

ogist, Vol. 43 (1941), pp. 386-87.
16. Ibid., p. 387n.
462

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms
 Race and Descent as Social Categories in India

17. Thus far I have tried to keep close to the technical or anthropological
meaning of the term race. But this meaning is itself ambiguous and has
undergone much change in the last few decades. Many physical anthro
pologists regard the idea of race as having outlived its utility in scientific
analysis. Be that as it may, this idea, however vaguely held, continues
to exercise a powerful influence over the human mind. I, therefore, turn
to a consideration of the categories that correspond most closely to "race"
in popular Indian thought.
18. The word jati is derived from fan, "to give birth to."

19. E. E. Evans-Pritchard, Nuer Religion (Oxford, 1956).

20. E. R. Leach, "Concerning Trobriand Clans and the Kinship Category

Tabu," The Developmental Cycle in Domestic Groups, ed. Jack Goody
(London, 1958), pp. 120-45.
21. I have discussed the different meanings of the word jati in relation to the
caste system in "A Note on the Referents of Caste," European Journal of
Sociology, Vol. 5 (1964).
22. Adrian C. Mayer, Caste and Kinship in Central India (London, 1960),
p. 13.
23. Andr? B?teille, "Caste and Politics in Tamilnad," forthcoming.
24. Quoted in G. V. Subba Rao, Life and Times of Sir K. V. Reddi Naidu
( Rajahmundry, 1957), p. 46.
25. B?teille, "Caste and Politics in Tamilnad."

26. Homeland, May 28, 1961.

27. Andr? B?teille, "The Future of the Backward Classes, the Competing De
mands of Status and Power," Perspectives. Supplement to the Indian Jour
nal of Public Administration, Vol. 11, No. 1 (1965), pp. 1-39.

463

This content downloaded from

110.137.100.39 on Wed, 14 Feb 2024 03:00:51 +00:00
All use subject to https://about.jstor.org/terms