The Image of The Barbarian - Romila Thapar

The Image of the Barbarian in Early India
Author(s): Romila Thapar

Source: Comparative Studies in Society and History , Oct., 1971, Vol. 13, No. 4 (Oct.,
1971), pp. 408-436
Published by: Cambridge University Press
Stable URL: https://www.jstor.org/stable/178208
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
Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access
to Comparative Studies in Society and History
This content downloaded from

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
All use subject to https://about.jstor.org/terms
The Image of the Barbarian in
Early India
ROMILA THAPAR
Jawaharlal Nehru University, New Delhi
The concept of the barbarian in early India arises out of the

situation of the arrival of Indo-Aryan-speaking nomadic past
northern India who came into contact with the indigenous p
(possibly the remnants of the urban civilization of the Indus) and
them as barbarians. The earliest distinction made by the Aryan
was a linguistic distinction and, to a smaller extent, a physical dis
The Indo-Aryan speakers spoke Sanskrit whereas the indigenou
probably spoke Dravidian and Munda. However the distinction
one of binary opposition-in fact it admitted to many nuances and
of variation, hence the complication of trying to trace the histor
concept. The distinction was rarely clearly manifest and based
language, ethnic origins or culture. Political status, ritual st
economic power, all tended to blur the contours of the distinction
to this has been the confusion introduced by those who tend to id
language with race and who thereby see all speakers of Sanskrit as
of that nineteenth-century myth, the Aryan race.'
The Aryans,2 although unfamiliar with city civilization, did bri
them the central Asian horse and the light, spoke-wheeled chariot
gave them a military advantage over the local people still using ox
carts. Furthermore it is also believed that the Aryans either brou
them, or else were instrumental in the dispersion of, iron technolo
again was superior to the existing copper technology. It is likely t
cities of the Harappa culture had already declined or at least w
final stages of decline when the Aryans arrived. They were virtua
therefore with a series of chalcolithic cultures extending from th
valley to Rajasthan and across the Ganges valley. Their associa
iron technology would probably explain why they were so suc
1 E.g. Caldwell, A Comparative Grammar of the Dravidian or South Indian
Languages. Thus, all south Indian brdhmans who use Sanskrit were seen as origin
2 The use of the word 'Aryan' in this article refers to those peoples who spoke an
language. It has no ethnic connotation and is merely used as a more manageable form
phrase 'Aryan-speaking' with which it is synonymous.
408

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
IMAGE OF THE BARBARIAN IN EARLY INDIA 409
spreading the Indo-Aryan language system through a major part

northern India.3 The anomaly of a less civilized people referr
to the inheritors of a higher civilization as barbarians can thu
explained.
The word most frequently used in Sanskrit to describe the barbarian is
mleccha. Attempts have been made to derive the etymology of the word
from the root vac speech, hence one who is not familiar with the known
speech or is of alien speech.4 This also provides a clue to the early distinc-
tion being based on speech which fact is stressed in late works as well.5
The etymology however is false as mleccha represents a cultural event
rather than a linguistic fact. It has been suggested that mleccha may have
been derived from Me-luh-ha, the Sumerian name for an eastern land with
which the Sumerians had trading relations, possibly the people of the Indus
civilization.6 The Pali word for mleccha is milakkha, which relates even
more closely in phonetics to the Sumerian version.7 Buddhist sources
explain milakkha as referring to the non-Aryan people, the Andhra,
Tamil, etc.8 This is further substantiated by the Dharmasastra of
Jaimini in which he mentions certain mleccha words which are Sanskritized
versions of words occurring in the Dravidian languages. Thus the ety-
mology of mleccha would relate it to the indigenous inhabitants of northern
India at the time of the arrival of the Aryan-speaking peoples, a far more
plausible derivation than the earlier one. Another attempt derives mleccha
from the proto-Tibetan *mltse meaning 'tongue' and the Kukish mlei. This
would associate the early use of the word with the non-Aryan speaking
peoples living close to the Tibeto-Burman area.9
The verb mlech means 'to speak indistinctly'. It may have been an
onomatopoeic sound imitating the harshness of an alien tongue. Retroflex
3 For a discussion of the nature and impact of Aryan culture on existing cultures in northern
India, see Romila Thapar, Presidential Address, Ancient History Section, Proceedings of
the Indian History Congress, December 1969.
4 Categories of speech are demarcated in Vedic literature reflecting a considerable concern
for the correctness of speech. Satapatha Brdhamana, IV, 1, 3, 16; Kdthaka Samhita, I, 11, 5;
Taittiriya Samhitd, VI, 4, 7, 3; Maitrdyanf Samhita, III, 6, 8.
5 The Nydyamalavistdra. Manu, X, 43, distinguishes between mleccha-vdc and drya-vdc.
6 Recent exponents of this view are the Finnish scholars, Parpola et al., who have made this
identification basic to their reading of the Harappa script as proto-Dravidian, Decipherment
of the Proto-Dravidian Inscriptions of the Indus Civilisation, Copenhagen, 1969. An even more
recent reading is that of I. Mahadevan who reads two Harappan pictograms as *mil-ey which
becomes *mil-ec which in turn becomes mleccha in Sanskrit, all of which mean 'the resplendent
ones'-the assumption being that this was the name by which the Harappan people called
themselves. Journal of Tamil Studies, II, No. 1, 1970.
7 Vinaya Pitaka, III, 28.
8 Buddhaghosa's commentary explains it as 'Andha Damil, 'di'. The Jaimini Dharmas-
dstra gives a short list of mleccha words, I, 3, 10. These are all words used in the Dravidian
languages, but are given in this text in a slightly Sanskritized form-pika, nema, sata, tamaras,
meaning respectively, a bird, a half, a vessel, a red lotus. Panini mentions that the affix an
denoting descent occurs in the name of persons of the Andhaka, Vrspi or Kuru tribes, IV, 1,
115. The affix an in this context is characteristic of Dravidian languages.
9 R. Shafer, Ethnography of Ancient India, p. 23.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
410 ROMILA THAPAR
consonants are believed to have been assimiliated into Indo-Aryan fr

Dravidian. The earliest of the better-known grammarians, Panini, gi
a form of the word mli.ta as 'that which is spoken indistinctly or barba
rously' and treats it in its noun form as indistinct speech or a fore
language.10 Used as a noun, the word also has the rather significant asso
tion with copper and copper-coloured. This may have had some connecti
with the Aryan speakers introducing iron to Indian cultures erstwh
based on a copper technology."l From the early centuries A.D. onwards t
adjectival use of mleccha becomes quite frequent.12
The gradual emphasis on speech differentiation is apparent in the use
another range of words for barbarians which are clearly onomatop
and indicate an incomprehension of the language concerned; words s
as barbara, marmara and sarsara. The first may well be borrowed fr
the Greek barabaros, since it occurs in late works in Sanskrit and refers
people of the north who are said to be sinful, low and barbarous.13
word also occurs in Pali as babbhara and means 'people of an unknow
tongue'.14 Further variants in Sanskrit are bhara-bhara and balbala-karo
'to stammer or stutter'. Marmara and sarsara carry the same meaning an
are intended to convey the sounds of a halting and alien speech.
In the .Rg Veda, the earliest of the Vedic texts, there is no mention of
mleccha as such but there are references to the Dasa or the Dasyu,
local tribes who were conquered by the Aryan speakers and who were th
regarded as alien and barbaric. They are compared with demons, be
black-skinned (krSna-tvach) and snub-nosed, speaking a strange langu
(mrdra-vdc); they practise black-magic and do not perform the requ
sacrifices; they are treacherous and they live in fortified habitations.15 T
distinction of language and physical appearance is recorded. Societ
divided into two main groups, the Arya-varina and the Ddsa varzn
suggesting a rather simple division into 'us' and 'them' where politi
success justifies the superiority of the former over the latter.
That speech was the chief component in distinguishing the Aryan fro
the others is clearly indicated in a text from the later Vedic literature.
example of barbarian speech, that of the Asuras, is quoted in the Satapa
10 Astadhydyi, VII, 2, 18.
11 N. R. Bannerjee, The Iron Age in India.
12 Such as mleccha-desa (country), mleccha-bhdad (language), Mleccha-nivaha (hor
mleccha-bhojana (food-used by rice-eaters for non-rice-eaters, particularly those eating whe
mleccha-vdc (speech).
13 Mahdbhdrata, XII, 207, 65.
14 Majjhima Nikdya, I, 128.
15 Rg Veda, III, 12, 6; II, 12, 4; III, 34, 9; V, 29, 10; IV, 16, 9; I, 33, 4; IV, 16, 3; X, 22, 8
II, 20, 8; VI, 20, 10.
16 .Rg Veda, III, 34, 9; II, 24, 4; I, 104, 2. The word 'Varna' literally means 'colour' and
came to be used for varVa society or caste society. The word varna does not refer to the actual
caste of a person but to a more broadly differentiated group which some writers mistook for
caste. With the exception of the brdhmans and the ksatriyas the precise caste status of the other
two groups was never uniform.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
Brdhmana17 and is later quoted and discussed by a grammarian

fourth century B.C., Patanijali.18 It is evident from the example tha
barbarian speech in this case was a Prakrit dialect of eastern India.19
would also suggest that when the Aryans settled in the middle
valley the difference in speech was not only noticed but record
examined. The emphasis on language was important as the knowl
correct Sanskrit was crucial to the notion of being an Aryan, an
efficacy of the ritual hymns.
Having established a distinction in language, a demarcation w
made with regard to territory. Those areas where a mleccha bhd
guage) was spoken came to be regarded as the mleccha-desa or count
the mleccha, and this in theory at any rate, was clearly cordoned off
mleccha areas were impure lands not only because those who live
spoke an alien language but what was more important they did not
the correct rituals. These were lands where the sraddha ceremony (of
to ancestors on stipulated occasions) was not carried out, and w
people did not observe the laws of the varna. The pure land was
varta,21 traditionally the region inhabited by the Aryas, all e
mleccha-desa. Since the mleccha is ritually impure, Aryas visiting th
of the mleccha must perform prdyascitta or expiatory rites before th
be regarded as cleansed and fit for normal association again.22 The c
of ritual impurity relates to the functioning of caste and this pa
aspect of the image of the barbarian appears to be unique to early
culture. It was this dichotomy of purity-impurity which gave
significance to the role and status of the ritually pure-the Arya and
eminent amongst the Aryas, the brahman. If mleccha epitomi
barbarian, then Arya includes all that is noble and civilized. It is dou
that the term drya was ever used in an ethnic sense. In Sanskrit and
literature it is used primarily as a descriptive term or an honorific re
17 Satapatha Brdhmana, III, 2, 1, 23; which reads, te'surd dttavacaso he'lavo h

vadanta.y pdrdbabhabuh. The Kanva recension has a variant reading (Sacred Boo
East, XXVI; p. 31, n. 3) but the end result is similar.
18 Vydkarana Mahdbhasya, I, 1, 1, which reads, te'surd helayo helaya iti kurvantah p
habuh. In both cases the word for enemy, ari, uses '1' instead of the pure Indo-Ary
Asuras here referred to are a puzzle. They are described as demons, but also as a m
people whom the Aryans of the .Rg Veda had to contend with. Were they the peop
Harappa Culture or were they a branch of the Aryans who came from the southern
Iran? Archaeological remains in Chota Nagpur are associated by the local tribes wi
Asuras. Banerji Sastri, Journal of the Bihar Oriental Research Society, XII, pt. ii, 24
19 A characteristic of the Prakrit of eastern India as attested by the inscriptions of
that the 'r' sound changes into '1', J. Bloch, Les Inscriptions d'Asoka, p. 112.
20 Manu, II, 23; X, 45.
21 Arya-varta was traditionally the region inhabited by the dryas. Its precise geo
area is difficult to define as the concept was not static in history. Broadly speaking, h
the Ganges-Yamuna Doab and the plain of Kurukshetra to the north of Delhi would
correspond to drya-varta, in the strict sense. Some texts extend the definition to inclu
the entire Indo-Gangetic plain, e.g., Manu, II, 17-74.
22 Visnu, LXXXIV, 1-4.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
412 ROMILA THAPAR
to a respectable and honourable man.23 Ritual purity or the absence o

was used not to justify aggression against the barbarian, but to justif
laws of exclusion on the part of the arya.
The perspective from the south was rather different. The barbarian
defined as one whose language was incomprehensible. The arya was
often merely the northerner and the word was sometimes used sy
mously for Vadavar, also a person from the north.24 Later arya was u
in the sense of a noble, respected person. Curiously enough, one
synonyms given for arya in certain Tamil lexicons is mleccha, and it is
for those who cannot speak Tamil, separating them from the norther
tribes such as the Vadukar and the Malavar who live in the forests as
hunters and who rob travellers and also steal cattle from the neighbouring
settlements. Their language is alien and they use long and unlearned words.
Beyond them lies Dandakaranya (in the north-eastern Deccan) which is
part of arya-desa. This attitude compares favourably with modern tribes
of the Chota Nagpur region who refer to the neighbouring aryanized
Hindus (non-tribals) as diku, meaning foreigners, a word which was used
to great effect in recent years in the building up of a tribal political
movement, which sought to exclude the neighbours.
The relationship between the mleccha and the arya was conditioned by
all the different facets which went into the making of a caste society. There
was, first, a network of exogamous and endogamous kinship relations
(jati); second, a hierarchical ordering of occupations and a division of
labour which functioned on the basis of service relationships. The third
essential was the notion that every social group has a ritual status deter-
mined by the degree to which its occupation is clean or polluting. The ritual
status need not coincide with the actual socio-economic status. It can be
maintained that ritual status is expressed in the notion of varna with its
four categories of brahman (priest), ksatriya (warrior), vaisya (trader), and
asdra (cultivator). But for the purposes of the actual functioning of society,
jati (literally meaning 'birth') was the more significant unit. Fourth, each
group was associated with a geographical location. The mleccha had to
respond to each of these facets. Kinship relations were excluded and the
mleccha therefore formed their own mleccha jatis. No self-respecting drya
would marry into a mleccha family. Where the mlecchas in question were
technologically inferior, their occupation was low and this affected their
ritual status which was heavily weighted on the side of impurity and
therefore low. Consistency with regard to geographical location is evident
from the long periods of designating particular regions as mleccha-desa.
23 Manu, X, 45, 57; speaks of crya-vac and drya-rapa (noble speech and noble visage) where
drya is used in an adjectival form. The Pali ayya or ajja carries the same sense. The antonym of
andrya, ddsa or dasyu again carries the meaning of lacking in worthiness and respect and
cannot be taken in an ethnic sense alone.
24 S. K. Aiyangar, Some Contributions of South India to Indian Culture, pp. 1-42.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
Theoretically this seems to be a fairly clear situation. But in fact ther

were not only lapses from the theory but rarely did society functio
strict accordance with these rules although the facade of the rules
maintained. This has to be kept in mind when seeking information from
the sources. Whereas the Dharmasastras, being legal treatises and so
codes, maintain the theory and much of religious brahmanical literature
to conform to the theory, the non-brahmanical literature, particularly sec
literature, and epigraphic evidence provide pointers to the actual situati
By the latter half of the first millennium B.C. the picture had become
more complex. The amalgamation of existing local cultures, which
inevitable in the evolution of Aryan culture, created problems for
theorists of caste society. Not all social groups could be given a pre
varna status. The process of anuloma (hypergamy) and pratiloma (w
the mother is of a higher caste than the father) had to be conceded and
number of new and, inevitably, mixed castes (sanklrna jdti) were admit
to the theory of social order.25 They were given the rank of sudras. Of t
many came to be described as mleccha such as the Ambastha, Ugra
Nisadha among the anuloma26 and the Siuta, Magadha, Candala, Ayog
and Pulkasa among the pratiloma.27 Even within the sankTrna jdtis ther
a hierarchy of ranking as recorded in the Dharmasastras.28 Professional
they followed occupations which were regarded by the theorists as activit
associated with unclean tasks such as washermen, fishermen, pott
leather-workers, iron-smiths, basket-makers, hunters and scavengers.
That the members of the sankirna jatis did not necessarily in fact hav
a low social status is indicated by the sources. The Aitereya Brdhm
mentions an Ambastha king.29 The Taittereya Brihmana refers to t
material well-being of the Ugras, one of whom is mentioned as a ki
officer.30 Similarly the Suta and the Magadha were traditionally the bar
and the chroniclers, in fact the preservers of the early Indian histo
tradition. They were close to the king not only because of their profess
but we are told that the presence of the Sfita was essential to one of th
rites in a royal sacrifice.31 In contrast the case of the Candalas is exceptio
the emphasis being on impurity and not on a difference in culture. They w
regarded as so polluting that they had to live outside the village or town.
25 Manu, X, 10-12; 16-17.

26 Others included the Andhras, Abhira, Pulinda, Khasa, Magadha, Kirata, Malla. Gaut
Dharmasastra, IV, 4; Baudhdyana, I, 9, 3; Vasistha, XVIII, 9.
27 Gautama, IV, 15; Baudhayana, I, 8, 8; Vasistha, XVIII, 1-6. 28 Manu, X, 39.
29 Aitereya Brdhmana, VII, 21; The Ambastha tribe is frequently identified by modern
scholars with the Ambastanoi of Arrian and the Sambastoi of Diodorus. H. C. Raichaudhury,
Political History of Ancient India, p. 255.
30 Taittereya Brathmana, III, 8, 5.
31 Taittereya Samhita, I, 8, 9, 1-2; The sata was one of the ratnins at the rites of the vdjapeya
sacrifice.
32 PaiinI, II, 4, 10. R. S. Sharma, Sudras in Ancient India, p. 125, suggests that originally
they may have been an aboriginal tribe using their own dialect, the can.dla-bhasa.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
414 ROMILA THAPAR
One of the most interesting and yet at the same time ambiguous cases
the classification of a people as near-mleccha is that of the vrctyas. Ved
sources on the vratyas appear confused as to their exact status.33 L
legal literature uses the word vratya in the sense of 'degenerate'.34 Acco
ing to Vedic literature the vrdtyas were not brahmanical in culture
had a different language; but they did speak the language of the initiat
although with difficulty. Yet the vratyas were not dismissed as mle
and considerable efforts were made to try to circumvent this problem,
of them being the famous ritual of the vrdtyastoma, the rite by which t
vratya was purified and accepted into Aryan society.35 Clearly the vrat
were a powerful group whose power seems to have emanated from
religious sanction and who were therefore treated with a barely disguis
veneration by the authors of the Atharvaveda, but with some condescensi
by the authors of the Dharmasdstras.
The second half of the first millennium B.C. was also the period whic
saw the gradual but extensive urbanization of the Ganges valley. The riv
itself became the main channel of communication and trade with cities
rising on its banks. The agrarian settlements had also tended to lie closer
to the river. There were still large areas of uncleared forest, especially
nearer the hills where the Aryan agrarian economy had not reached. It was
now possible for the Aryan speakers to assume the role of the advanced
urban civilization based on technological and economic sophistication.
They could therefore regard with contempt the tribes living in the forests
who had remained at the food-gathering and hunting stage. Such techno-
logically inferior tribes as for example the Sabara, Pulinda, Mutiba and
Kirata constituted yet another category which came to be included in the
term mleccha.36 The distinction which is made in the epic Rdmcyazna
between the urban culture of the kingdom of Ayodhya based on a fairly
extensive agricultural economy can be contrasted with the hunting and
food-gathering culture of the enemies of Rama, the ridkas peoples.37 Very
often these tribes inhabited the fringes of Aryan culture and had to move
up into the hills with the gradual expansion of the agrarian economy. By
extension therefore the tribes on the frontiers also came to be called
mleccha, even in cases such as those of the Yavanas and the Kamboja
who were as civilized as the Aryans.38 Thus the use of the word mleccha
33 PafcavimSataka Brahmana, XVII, 1, 9; 53, 2. Apastambha Dharmasatra, XXII, 5, 4.
34 As for example the use Manzu makes of the term vratya-ksatriya or 'degenerate k$atriyas
when describing the Greeks, or vrdtya for those who have failed to fulfil their sacred duties,
X, 20; II, 39.
35 Atharvaveda, XV.
36 Also included were the Bedar, DaaSima, Matafiga, Pundra, Lambakarna, Ekapada,
Yaksa, Kinnara, KIkata, Nisada. Some of these are fanciful names-Long-ears, Single-footed;
some were celestial beings; but in the main both literature and epigraphs record the names of
many of these tribal peoples.
37 D. R. Chanana, Agriculture in the Ramayana.
38 Yaska in Nirukta, II, 2. Atharveda, V, 22, 14; Chandogya Upa.i$ad, VI, 14, 1, 2.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
had now been extended to include speakers of an alien language, s

groups ranked as mixed castes, technologically backward tribes and
peoples along the frontiers.
The stabilizing of what were to be the Arya-lands and the mleccha-lan
took some time. In the .Rg Veda the geographical focus was the sapta-sind
(the Indus valley and the Punjab) with Sarasvati as the sacred river,
within a few centuries drya-varta is located in the Gariga-Yamfna D
with the Ganges becoming the sacred river. Together with the shift eas
wards of 'the pure land' the northern Punjab and the trans-Indus re
came to be regarded as mleccha-desa. Later Vedic literature speaks of
western Anava tribes as mlecchas and occupying northern Punjab, S
and eastern Rajasthan, as also the eastern Anava tribes occupying parts o
Bihar, Bengal and Orissa.39 The tribes of the north were mleccha ei
because they were located on the frontier such as the Gandhara an
Kamboja and therefore both their speech and culture had become c
taminated and differed from that of arya-varta, or else, as in the case of
Madras, they were once aryas but having forsaken the rituals were relega
to mleccha status. The latter was obviously an attempt to explain the co
tradiction of the earlier texts mentioning the tribe as aryas and the lat
texts, written when the arya-varta had shifted eastwards, referring to th
as mleccha.
That the northern region was once the land of 'the pure speech' is stated
with reference to the Udicya (northern region) where peoples such as the
Uttarakurus and the Kuru-Paiicalas are held up as the model in speech and
it is recommended that brahmans be sent there to learn the language.40
Buddhist literature describes Uttarakuru as a mythical paradise, a land
reminiscent of the utopian past when there were no institutions such as
private property and the family and when there was no need to work
because food was available from the trees and all man's desires were
satisfied.41 The later Puranic tradition echoes this description for we are
told that the land is covered with milk trees which eliminate the need for
cultivating food, that the women are beautiful like the apsards (celestial
nymphs) and that people are born as couples, presumably thereby intensi-
fying sexual pleasure.42 Possibly the brahmanical conception of Uttarakuru
as the land of the purest speech may have symbolized the brahmanical
utopia, a land of non-polluting peoples, observing all the required rituals
39 The western Anavas were the Yaudheyas, Ambastha, gibi, Sindhu, Sauvira, Kaikeya,
Madra, Vrsadarbha. The eastern Anavas were the people of Afiga, Vafga, Kalifga, Pundra
and Suhma. It has been suggested that the names ending in ahga are of Munda origin and
these tribes would therefore be pre-Aryan. P. C. Bagchi, ed., Pre-Aryan and Pre-Dravidian in
India.
40 Aitereya Brdhmana, VIII, 14, 23; Satapatha Brahmana, III, 2, 13, 15; Kausitaki
Brahamana, VII, 6.
41 Atanatlya Sutta, Digha Nikaya, III, p. 199 if.
42 Brahmanda PurandVa, II, 19, 24; III, 59, 46. Vayu, 91, 7; Matsya, 83, 34; 105, 20.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
4I6 ROMILA THAPAR
and speaking the purest language. Not surprisingly, of the tribes of arya-
varta by far the most significant are the Kuru-Paiicala.43 They emerge as
a confederation of a number of existing tribes earlier associated through
war and matrimonial alliances.
The Himalayan region was largely mleccha-desa since it was not only a
border region but was mainly inhabited by Tibeto-Mongoloid people and
the dissimilarity of language and culture would be indicative of difference.
The other mountainous region, that of the Vindhyas and their extensions,
is probably the most interesting from the point of view of geo-politics. The
Aravalli hills formed the natural watershed between the Indus and Ganges
valleys and this would be the natural frontier region between the two valleys.
For a long period up to the early centuries A.D. it was occupied by non-
Aryan tribal republics, which survived the general decline of republics in
the valley areas, and which were consequently the frontier for the Ganges
valley. The central Indian complex of the Vindhya and Satpura ranges
with the rivers Narmada, Tapti and Wainganga cutting through them and
the plateau areas of Chota Nagpur and Chatisgarh to the east has formed
throughout Indian history an ideal setting for the tribal peoples. It lent
itself easily to a pastoral and food-gathering economy with the possibilities
of agriculture in some parts of the river valleys and the proximity of rich
agricultural areas in the plains. With the expansion of Aryan culture and
the clearing of the forests in the Ganges valley the existing population of
the valley would have sought refuge in the central Indian highlands. Up
to about the middle of the first millennium A.D. the Vindhyan tribes lived
in comparative isolation totally unconcerned with the mleccha status con-
ferred upon them by the Aryans. The Chambal and Narmada valleys being
the main route from the urban centres of the Ganges valley to the western
ports (e.g. Bhrighukaccha, modern Broach) and the Deccan, the plundering
of trading caravans and travellers may well have provided the tribes with
extra comforts. Plundering was always a means of livelihood which they
could resort to, especially during periods of political disturbance. It is not
until the post A.D. 500 period that they begin to participate in the politics
of both northern and southern India.
The pre-Aryan settlement of eastern India is attested to by advanced
neolithic cultures and the chalcolithic copper hoards in Bihar and Bengal.44
Literary evidence dating to about the middle of the first millennium B.C.
indicates that the people of these areas spoke a non-Aryan language. The
43 The Kuru tribe had a well-known status and antiquity. They acquired fame through the
epic Mahabhdrata which concerns a family feud between the Kauravas and the Paitdavas,
both members of the Kuru lineage. The Paicalas were a confederation of five tribes. According
to bardic tradition the royal family of the Paficalas was an off-shoot of the Bharata family.
44 B. B. Lal, 'Further Copper Hoards from the Gangetic Basin .. .', Ancient India, No. 7,
1951, pp. 20 ff. S. P. Gupta, 'Indian Copper Hoards', Journal of the Bihar Research Society,
XLIX, 1963, pp, 147 ff,

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
boundary of Aryan control in the Ganges valley is perhaps referred to

striking story related about king Videga Mathava, the king of the Vid
who is said to have travelled with the god of fire, Agni, across the Ga
valley as far as the river Sadanira. Here he paused as the land to th
of the river had not been sanctified by Agni. Once this was done the
established the Videha people on the other bank and the lands to the e
of the Videhas were the mleccha-desa.45
Yet it was the mleccha-desa adjoining Videha, Magadha, which was to
play a leading role in Indian history during the subsequent millennium.
Magadha is described as the accursed land with a people of mixed caste
status. An expiatory rite is required from those who visit it and this
injunction is continuously repeated in the Dharmas'stras for many cen-
turies, right through the period when the state of Magadha was the centre
of empires and powerful kingdoms, viz., the Maurya and Gupta.46 The
other eastern peoples, those of Anga, Vanga and Kalinga were even more
polluting and required more elaborate expiatory rites.47
This was not the attitude however among the Jainas and Buddhists since
it was in these areas that the heterodox religions first gained ground, as
for example, Anga, which was an early centre of Jainism. The Jaina texts
clearly define the milakkhu as the Varvara, Sarvara and Pulinda tribes and
discourage monks and nuns from keeping their company.48 Buddhist
sources make no distinction between arya lands and mleccha lands when
describing the sixteen major states of northern India. Since the Buddha
himself preached in Magadhan Prakrit he would hardly have accepted the
term mleccha for the people of the region. A late Buddhist work mentions
the Magadha bhdsC as the speech of the Aryans indicating that Sanskrit
did finally come to be accepted in Magadha.49 The word milakkha is used
in Buddhist writing, and as we have seen, one very reliable definition of it
reads Andha Damil, adi, 'Andhras, Tamils, etc;' i.e. the people of the
peninsula.50 Milakkha is also used to describe those dryas who had lost
their status and the Kamboja are quoted as an example;51 also, foreigners
such as the Yavanas or Yonas whose status was high but who spoke an
alien language,52 and finally the tribes of the jungle, such as the Pulinda
and Kirata, where they are not only less civilized but again their language
45 Satapatha Brdhmana, I, 4, 1, 10.

46 Atharvaveda, XV, 2, 1-4; Baudhdyana Dharmasatra, I, 1, 32-3; Manu, X, 11.
47 Texts as late as the Mdrkandeya Purdna and the Yajhavalkya Smrti, III, 292, repeat the
need for the prayascitta.
48 Prajnapana Upahga, p. 397; Acarahga Sdtra, II, 3, 1; II, 11, 17.
49 Anguttara Nikdya, I, 213. The sixteen mahdjanapadas or major states are listed as
Gandhara, Kamboja, Kuru, Paiicala, Surasena, Matsya, Kosala, Kasi, Malla, Vrjji, Magadha,
Afiga, Vatsa, Cedi, Avanti, Asmaka.
50 Sammoha-vinodani, Vibhafiga commentary, 388; Manorathapurdni, Anguttara Commen-
tary, I, 409; Apdddna, II, 359; Sutta Nipdta, 977.
51 Jdtaka, VI, 208, 210. Cf. Manu, X, 44.
52 Summahgala Vildsini, I, 276; Sammoha-vinodani, 388.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
418 ROMILA THAPAR
is incomprehensible.53 It would seem from the Buddhist sources that lan-

guage was the most important criterion of differentiation. Ritual impurity
was not a major item in Buddhist thought, thus discrimination was not as
severe as in brahmanical writing.54 The Buddhists tended to underplay the
mleccha consciousness probably because of the Buddhist association with
the mleccha regions, these being the areas where it gained most ground.
Nevertheless even powerful rulers motivated by the Buddhist ethic such
as the Mauryan emperor Asoka (third century B.C.) could not disregard
the differentiation. His list of the tribal peoples in his empire recorded in
one of his inscriptions agrees closely with the lists of mleccha peoples
mentioned in other sources, although he does not actually call them
mleccha.55
Asoka makes a distinction between the tribal peoples and the forest
tribes, the latter having to be wooed by his officers in the context of a
paternalistic policy where he regards himself in the image of the father
and his subjects as his children. It would seem that the forest tribes did not
easily reconcile themselves to law and order.56 The same problem is
reflected in the Arthasastra, the treatise on political economy ascribed to
Kautalya the minister of ASoka's grandfather (fourth century B.C.).
Kautalya also distinguishes between the mleccha and the forest tribes
(aranyacdrah, atvikah). He recognizes the political advantages to be gained
from keeping the forest tribes happy since they had their own strongholds
and could be used effectively in campaigns. Furthermore it was necessary
to pay them off from time to time to prevent their resorting to plundering
and pillaging.57 Another source of the same period, the Indika of Megas-
thenes, the Seleucid ambassador to the Mauryan court, refers to the Indians
as surrounded by barbarian tribes, possibly a reference to drya-varta
surrounded by the mleccha-desa. Megasthenes adds that all these tribes
were indigenous but that they differed in mind and disposition from the
Indians.58
Although Megasthenes does not describe the Indians as barbarians, the
Indians undoubtedly regarded him as a mleccha. For the Indians, the
Greeks on every count were mlecchas. They were referred to by the term
53 Ibid; the ancestry of the Pulinda located in Ceylon alone, according to the Buddhist
sources, derives from the marriage of prince Vijaya with the demoness Kuveni.
54 The candala is known and mentioned in Buddhist sources but usually in the context of
his overcoming his low status although this is often done through the acquisition of some
spiritual power.
55 Major Rock Edict, XII. J. Bloch, Les Inscriptions d'Asoka, pp. 130 ff. Asoka lists the
Yona, Kamboja, Nabhaka, Bhoja, Pitinika, Andhra and Palida.
56 The Second Separate Edict. J. Bloch, Les Inscriptions d'Asoka, pp. 140 ff.
57 Arthasastra, II, 1; III, 16; VII, 8; VIII, 4; IX, 1; IX, 3; X, 2.
58 McCrindle, India as Described by Megasthenes and Arrian, pp. 20-1; McCrindle, India
as Described by Ktesias, pp. 23-4, 86. Earlier Greek writers such as Ktesias, the Greek
physician at the Persian court in the sixth century B.C., referred to the Indian king trading
cotton and weapons for fruit, dyes and gum with the Kynokephaloi or Kynomolgoi, a
barbarian tribe. The identity of this tribe has not been conclusively established as yet.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
Yavana, a back-formation from the Prakrit yona, which is said to

from Ionia, suggesting that the Ionian Greeks were the earliest to
come into contact with India. Indian tradition however maintains that the
Yavanas originated from Turvasu the son of Yayati, associated with one
of the very early and important tribes of northern India.59 But this may
well be a late attempt to find the Greeks a respectable ancestry when their
role in the history of northern India became more than marginal. For the
Bactrian Greeks (or the Indo-Greeks as they are called in Indian history),
the gakas (Scythians) and the Kusanas aggravated the problem of having
to concede the existence of mleccha rulers. In spite of the dismal prophecies
of the ancient seers that the Kaliyuga (the period under discussion) would
initiate the rule of the low-caste, nevertheless the mleccha origin of these
rulers had to be faced.60 The problem was further complicated by the fact
that these rulers patronized and used Sanskrit as is evident from their
inscriptions and coins and they inter-married into the local ruling families.
The description of these areas as mleccha-desa was technically also prob-
lematical. The inscriptions of the Saka satraps (rulers and governors of
western India from c. 100 B.C. to A.D. 300) are not only composed in good
literary Sanskrit, but also assert with much vehemence that the kings are
doing their utmost to prevent the mixing of the castes and are protecting
the law of varna.61 Thus the two main criteria of barbarism could not
theoretically be said to prevail.
The mleccha both indigenous and foreign had acquired political power
and a new concept was necessary. It was probably largely to circumvent this
problem that the term vratya ksatriya (degenerate kvatriya) became current in
describing the origin and status of such peoples. It was maintained that in
origin they were of the ksatriya varna and that their degeneration was due
to the non-performance of sacred rites, or because of the wrath of the
brahmans when they ceased to perform the sacred rites.62 Among the
foreign rulers included as vratya ksatriyas were the Yavanas and the gakas
(Scythians).
The term Yavana was gradually extended to include not only the local
Greeks but any group of people coming from west Asia or the eastern
59 Matsya Purada, 34, 30; 50, 76.
60 Utpala's commentary on the Brhatsamhita, XIII, 3. describes the gakas as mleccha-
jatayo-rajanas and adds that the period of their destruction by Vikramaditya would be known
as Saka-kala.
61 A large number of early Sanskrit inscriptions come from the mleccha areas of north
and western India. Corpus Inscriptinum Indicarum, Vol. II. The Greeks had used Greek an
Prakrit or Sanskrit bilingually as on their coins: Obverse-Basileus Suthos Menandros,
Reverse-Maharajas Tradarasa Menamdrasa. Smith, Catalogue of Coins in the Indian Museum,
Calcutta, Vol. I, pp. 22 ff. Kusana coins show a slow but increasing adoption of Indian deities
particularly of the gaivite family. The gaka kings not only affirm their protection of the law of
varna but even record large donations of cows and villages and wealth to the brahmans.
Rudradaman's Junagadh Inscription, Epigraphia Indica, VIII, No. 6, pp. 44 ff.
62 Manu, X, 43-4; Mahabharata, Anusasana Parva, XXXV, p. 226; Asvamedha Parva,
XXIV, p. 74; Sabha Parva, XIV, p. 44; LII, p. 145.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
420 ROMILA THAPAR
Mediterranean. Much the same was to happen to the term Saka wit
reference to central Asia, but Yavana remained the more commonly used
one. Even in south India, traders from Rome and later the Arabs we
called Yavanas. Early Tamil literature has descriptions of the Yavan
settlements in the trading ports of the peninsula. The Yavanas here
referred to were also described as mleccha, since they spoke an alie
language which was so incomprehensible that it sounded as if their tongu
were cut off.63
Among the tribes of indigenous origin also referred to as vratya ksatriya
in some sources are listed the Dravida, Abhira, Sabara, Kirata, Malava
Sibi, Trigarta and Yaudheya. The majority of such tribes tended to be the
inhabitants of the Himalayan and Vindhyan region, traditionally call
the mleccha-desa. There is evidence from numismatic sources of the
increasing political importance of some of these tribes which would explain
their elevation to the status of vratya ksatriyas from being plain mlecchas.
The period from the first century B.C. to about the fourth century A.D. saw
the rise of a number of tribal republics in the Punjab and eastern Rajasthan,
in fact in and around the watershed between the Indus and Ganges valleys.
The Malava tribe, mentioned by the Greeks as the Malloi, established
themselves in the Jaipur area having migrated from the Ravi.64 The Sibi,
the Siboi of the Greeks, migrated to north-eastern Rajasthan.65 The
Trigarta referred to by Panini, were settled in the Ravi-Satlej Doab. The
Yaudheyas also referred to by Panini moved from Haryana northwards.66
The fact that these tribes were politically powerful after they had settled
in an area is clear from the use of the term janapada in the coin legends
indicating their assertion over the territory on which they had settled. The
Gupta conqueror Samudragupta, campaigning in the fourth century A.D.,
takes great pride in having destroyed the power of these tribal republics.67
The coin legends also clearly demonstrate that these tribal peoples were
now using Sanskrit.
In the middle of the first millennium A.D. when it was evident that
mleccha dynasties were dominating politics, the Puranic tradition (as it
was then recorded) had much to say on the problem of the mleccha.68
63 Kanakasabhai, The Tamils Eighteen Hundred Years Ago, pp. 37 ff. M. Subramaniam,
Pre-Pallava Tamil Index, p. 618.
64 McCrindle, Invasion of India by Alexander, p. 234; Mahabharata, Sabha Parva, XXX;
British Museum Catalogue of Indian Coins, p. cv. The legend reads, mdlava-ganasya-jaya.
65 McCrindle, The Invasion of India by Alexander, p. 232; Mahabharata, Sabha Parva, XXX;
Journal of the Numismatic Society of India, IX, p. 82; British Museum Catalogue, p. cxxiv;
the legend reads, sibi janapadasa.
66 Astadhydyi, V, 3, 116; Mahdbhdrata, Sabha Parva, XXX; the legend reads, trakataka
janapadasa. Astadhydyi, IV, 1, 178; British Museum Catalogue, pp. cxlix-cl. The legend reads,
yaudheya-bahadhanyake, and a fourth-century coin-mould reads, yaudheya-ganasya-jaya.
67 The Allahabad prasasti of Samudragupta. Corpus Inscriptionum Indicarum, III, pp. 6 ff.
68 The eighteen major Puranas were recorded from about the third century A.D. onwards.
They claim to be compendia of information orally transmitted over a period going back to
c. 3000 B.C. The texts deal with the mythologies of the creation of the universe, genealogies of

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
There is a general bewailing of the increase in mleccha influence which

associated with the prophecy that the Kaliyuga will see mleccha dom
nance.69 This will result in the establishment of the mleccha dharm
barbarous ordering of the universe when vice will be rampant, the author
of the sacred texts neglected, the sudras respected-in short, a comp
reversal of the world order as seen by the cryas.70 Passages such as thes
seem to express the sentiments of a small group fighting to preserve it
and prevent the change which is engulfing its world and its very existe
Not surprisingly the idea of the Saviour Deity is introduced in some of
Purdnas where it is stated that the god Visnu in his tenth incarnatio
Kalkin will ride through the world in an attempt to turn men back to t
path of virtue. Some of the mleccha peoples such as the Dravida, Sab
and Vrsala will be destroyed by Kalkin.71 But this was a temporary mea
sure as Puranic cosmology did not really envisage the coming of t
millennium since ultimately the entire universe was to be destroyed at
finale of the Kaliyuga.
It is curious that in spite of considerably increased communicat
between the Ganges valley and the peninsula and the spread of Sans
and of Aryan culture to the south, there is a persistence in regarding t
southern regions as mleccha-desa. The Andhras, for example, who h
ruled the northern Deccan for four centuries, are described as mle
kings and their lands unfit for the sraddha ceremony. At the same time
Andhra kings were claiming to be the protectors of the varna dharma,
the destroyers of the Sakas and Yavanas.72 That less concession was mad
to the southern kings as compared to the northern kings was partly du
distance and partly perhaps due to the belt of wild tribes inhabiting
Vindhyas who doubtless acted as a barrier.
However the attitude towards even these tribes was beginning to chan
and this is reflected partially in the genesis myths associated with t
69 Puranic cosmology envisages a cyclical movement of time and the world goes throu
period of four ages with the golden age at the start and an increase in evil through the dur
of the cycle. The last of the four is the Kaliyuga at the end of which evil will be prevalent
the mleccha all-powerful. Ultimately the entire universe will be totally destroyed after wh
a new universe will be created and the cycle will start again.
70 Vdyu Purana, 99; Bhdgvata, XII, 2, 12; 14, 38; II, 38; XII, 3, 25; 3, 35-6. Deprive
sacrificial activities the world will be reduced to mleccha-hood.
71 Matsya Purdna 47, 252; Vdyu, 98, 114; Brahmanda III, 14, 80; 22, 22; 73, 108; 35
IV, 29, 131.
72 Visvu Purana, IV, 24, 51; Brahmdnda, II, 16, 59; III, 14, 80; IV, 29, 131; Manu, X, 8
Yajiavalkya smrti, III, 292; Smrticandrika, I, 22-24. This is particularly contradictory in
case of the Puranas where a number of mleccha cults and rites had become incorporated
the recognized religion, particularly rites associated with the mother-goddess. For
reference to the gakas and Yavanas see, e.g., Nasik Cave Inscription, Epigraphia Ind
VIII, No. 8, pp. 60 ff.
kings and sages, social custom and religious practices generally pertaining to a particular
of which each Purana claims to be the sacred book. In fact much of the material reflects
contemporary attitudes at the time of the composition of the Purana. The genealogical sections
are in the form of a prophecy, an obvious attempt to claim antiquity.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
422 ROMILA THAPAR
origin. The most frequently referred to are the Niada. References to the
four varnas in Vedic literature includes mention of the Ni?ada who appe
to have been a non-Aryan tribe who succeeded in remaining outside Ary
control73 but had a low status in ritual ranking.74 They are genera
located in the region of the Narmada river or among the Vindhya a
Satpura mountains.75 They are described as being dark-skinned, f
featured with blood-shot eyes and of short stature.76 A series of myths
related regarding their origin.77 The variations apart, the main narrativ
states that they were born from the thigh of king Vena. The king V
was extremely wicked and flouted the sacred laws and the holy rites. Th
infuriated sages pierced him with the sharp ears of the kusa grass a
according to some versions, killed him. In order to avoid anarchy, si
the land was now without a king, they churned his left thigh and from
came a dark, ugly, short man, the ancestor of the Nisada, and in so
versions, the ancestor of the mleccha.78 Being unsatisfied with this resu
they then churned the right arm of Vena and from it emerged Prthu w
was crowned king and was so righteous that the earth was named af
him, Prthivi. Whatever the deeper meaning of these myths may be, it see
obvious that the original Ni?ada and Prthu represent two factions wh
may have fought for power. There also seems to be an association of gui
with the killing of Vena and the manner of the birth of Nisada suggests
that he may have been the rightful heir but was replaced by Prthu.
tribes with whom the Nisada are associated in these texts such as the
Bhila, Kol, etc; are often the tribes connected with the rise of new dynasties
in central India in the period after the eighth century A.D.
The Vindhyan region was the locale for the three tribes which came to
be mentioned almost as the synonyms for mleccha, the Kirata, Pulinda
and Sabara.79 The Kirata are described as a non-Aryan tribe living in the
hills and jungles of Magadha.80 The Mahabhdrata describes them as being
dressed in skins, eating fruit and roots and inflicting cruel wounds with
their weapons. Yet they were not as wild as the text would have us believe
73 In the RudrSdhyaya of the Yajurveda. Other degraded professions are the nomads,
carpenters, chariot-makers, potters, smiths, fowlers, dog-keepers and hunters. In this text
as also the Nirukta of Yaska they are mentioned as the fifth group after the four varnas.
III, 8; X, 3, 5-7.
74 Manu, X, 8, 18, 48. They were descended from the marriage between a brahman and a
sudra woman.
75 Garuda Purana, VI, 6; LV, 15; Padma, II, 27, 42-3; HarivamSa, XV, 27, 33.
76 Visnu Purana, I, 13.
77 Matsya Purana, 10, 4-10; Bhdgvata, IV, 13, 42, 47; Mahabhdrata, Santi Parva, 59.
78 Matsya Purdna, 10, 7.
79 The Amarakosa VII, 21; a lexicon of the post-Gupta period, in its definition of mleccha
mentions these three tribes and describes them as hunters and deer killers, living in mountain-
ous country, armed with bows and arrows and speaking an unintelligible language-the
conventional description of the mleccha by the time of the medieval period. Yet the location
of mleccha-desa in this text is not in central India but in northern India.
80 Jg Veda, III, 53, 14; Mahdbhdrata, Karia Parva, V, 9; Bhdgvata Purana, 11, 21, 8;
Manu, X. 44.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
MILES
100 0 300
Fig. 1

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
424 ROMILA THAPAR
because they also brought as gifts to one of the heroes, sandalwood, aloes
wood, expensive skins, gold, perfume, rare animals and birds and t
thousand serving girls. They arrived riding on elephants.81 If the gif
amounted to even a portion of what is described then the Kiratas cannot
said to have had a primitive economy. Early texts speak of them as living
in the east but later texts give the Vindhyas as their place of residence.8
Their migration may have been due to the expansion of the agrarian sett
ments in the Ganges valley. The most interesting reference to them howev
is the famous literary work, the Kirdtarjuunya where significantly t
Kirata is identified with the god Siva and gives battle to Arjuna, one of t
heroes of the Mahabhdrata.83 South Indian sources as late as the seven-
teenth century continue to refer to them as living in the Vindhyas in a
semi-barbarous condition.84
The names Pulinda and Sabara in particular seem to have become generic
names for barbarian tribes. 85 Ptolemy uses the curious expression 'agrio-
phagoi', the eaters of wild things,86 in describing the Pulinda, and locates
them to the east of Malava. The Pulinda may have migrated from the
Mathura region to the Vindhyas for the same reasons as did the Kiratas.87
They too are described as being dwarf-sized, black in complexion like
burnt tree-trunks and living in forest caves.88 The Sabaras were also
located in the Vindhyan region.89 A ninth-century inscription mentions
81 Mahdbhdrata, Sabha Parva, LVII, 144.

82 Mdrkandeya Purdna, p. 284; Matsya Purana, 114, 307; a seventh-century author identifies
them with the Bhila and Lubhdhaka tribes of the Vindhyas and also connects them with the
Matafiga, the lawless hunters of the region, Dandin, Dasakumdracarita, III, 104; VIII, 203.
The name Matafiga is very curious and suggests a Munda-Dravidian combination. The
twelfth-century Pampa Rdmdyana of Abhinava Pampa, VII, 105-55, also refers to them.
83 Bhfravi's long poem, the Kirdtarjuniya, is based on an episode from the Mahdbhdrata
when Arjuna goes into the Himalayas and does penance. He finally meets the god Siva in the
form of a Kirata with whom he has a protracted fight, but eventually acquires the divine
weapons which he is seeking. It is interesting that the Kirata should be identified with Siva-
perhaps suggesting their worship of Siva, and also that it is through a Kirata that the great
hero Arjuna acquires the divine weapons.
84 Pampa Rdmdyana, Nijagunayogi's Vivekacintamani, pp. 423-4. Chikka Deva inscription
of the seventeenth century in Rice, Mysore and Coorgfrom its Inscriptions, p. 129.
85 Buddhist sources refer to the children of the demoness whom prince Vijaya married on
his arrival in Ceylon as the Pulinda and state that they lived in the interior of the island at a
place called Sabaragamuva (= gabaragrama, the village of the Sabaras ?), Mahdvansa, VII,
68; Vinaya Pitaka, I, 168. These have come to be associated with the primitive Veddah tribes
of Ceylon. In early brahmanical sources they are mentioned as a wild mountain tribe of the
Deccan, Aitereya Brahmana, VII, 18; Mahdbhdrata, Adi Parva, CLXXVII, 504. Later sources
connect them with the Bhilas, Kathdsaritasdgara, II, 12; Amarakosa, II, 20-1.
86 Ptolemy, VII, 1, 64; Ptolemy's phrase brings to mind the use of the Pisdca in Indian
literature which also carries the meaning of those who eat raw flesh. Its most obvious connec-
tion is with the famous Brhatkatha of Gunfdhya which was written in a Pisdca or goblin
language, and the location was the Vindhyas. Possibly the Pisdca language was that of these
mleccha tribes. Interestingly, it is often associated by some scholars with the north western
areas which may suggest a migration of some at least of these peoples from the northwest to
the Vindhyas. Keith, A History of Sanskrit Literature, pp. 266 ff.
87 Rdmdyana, IV, 3; XLIV, 12; Kathdsaritasagara, IV, 22.
88 Ndtyasdstra, XXI, 89; Brhatkathaslokasangraha, VIII, 31.
89 Rdmdyana, Adi Kanda, I, 59; Aranya Kanda, LXXVII, 6-32. Bana, Kadambari, p. 12.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
the mleccha along the Chambal river and a fifteenth-century inscrip

refers to the quelling of a revolt by the Sabaras inhabiting the Ch
valley.90 (This valley has remained throughout Indian history th
route from the Ganges valley to the north-western Deccan and a
centre of dacoity to this day. Perhaps the plundering of caravans was
lucrative for the area to develop any other substantial economy.) An
medieval adaptation of the Ramayana from the south speaks of
Sabara chief as a powerful ruler of mleccha-desa.91 It is not clear whe
this is poetic imagination or whether it reflects a real impression of
Sabaras as seen from a south Indian perspective. As late as the six
century the king Krsna Deva Raya of Vijayanagara writes in his m
on government that the Vindhyan tribes must be brought round to a
ing the administration by gaining their trust,92 a sentiment reminisce
the emperor Asoka. A Sabara tribe exists to the present day in we
Orissa. The Kol tribes preserve a traditional memory of the name
and the Sabari river in Chhatisgarh reflects an association with
tribes.
The authors of the Dharmasdstras continued to prescribe dire pu
ments for those who travelled in mleccha lands, yet this did not
people. Needless to say Indian traders (brahmans included) did t
extensively and profitably in mleccha lands, the performance o
expiatory prdyascitta on returning home providing a convenient solu
to the problem.93 However with the incursions of mleccha ruler
drya-varta itself, a new problem arose: the pure land was being turne
a mleccha land. This had happened in the case of the Yavanas wh
come a fair way into the Gafga-Yamuna Doab. It was to happen
with the coming of the central Asian Huns or Hunas as they were
in India. The solution to this problem in the words of the medieval co
mentator Medatithi was that if the varna laws were introduced into the
region (or continued to be maintained) then it would be fit for the perfor-
mance of sacrifices.94
The coming of the Huns was not a traumatic event in the history of
India. Its impact has perhaps been exaggerated owing to its continual
Comparison with the arrival of the Huns in Europe. Even the parallel which
90 Dholpur Inscription, Indian Antiquary, XIX, p. 35; Khadavada Inscription of the time
of Gyas Sahi of Mandu, Journal of the Bombay Branch of the Royal Asiatic Society, XXIII,
p. 12.
91 Rdmdyana, IV, 37-8. 92 Amuktamalyada, IV, 206.
93 Visnu Dharmasatra, 71, 59; 84, 2-4; Vasistha, 6, 41; Gautama, IX, 17; Atri, VII, 2. T
radddha ceremony was an essential rite for the drya since it concerned the offering of f
the spirits of the ancestors and thereby strengthened and re-affirmed kin-ties. It is clearl
in the above texts that the drya is prohibited from speaking with the mleccha, from lea
their language or from making journeys to a mleccha-desa since contact with the mlec
polluting. The journeys were regarded with particular disapproval since the Srdddha ce
could not be performed in such areas.
94 Medatithi, a tenth-century commentator, on Manu, II, 23.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
426 ROMILA THAPAR
is frequently drawn between the Huns dealing a death blow to the Roman
empire and the Hiunas doing the same to the Gupta empire (fourth-fifth
centuries A.D.) is not strictly comparable since the nature of the two
empires was different as also the cause of their decline. Northern India
was by now familiar with foreign invasions and government under mleccha
dynasties. The Hunas were known to inhabit the northern regions and are
sometimes mentioned together with the Cina (Chinese).95 The close of the
fifth century A.D. saw the Huna invasions of India under their chief
Toramana. The location of his inscription at Eran (Madhya Pradesh) and
the discovery of his seals at Kausambi (Uttar Pradesh) point to his having
controlled a substantial part of arya-varta.96 Hence the problem of living
in a region overrun by the mleccha referred to earlier. Toramana's son
Mihirakula lived up to the conventional image of the Hun. He is particu-
larly remembered for his cruelty which has become a part of northern
Indian folklore.97 His violence however was directed mainly against the
Buddhists and the Jainas, whose literature is replete with complaints about
him.98 He was however forced back from the Ganges valley and the Hfna
kingdom after him was reduced to a small area of northern India. The
Huna invasion itself did not produce any major changes in the life of
northern India, except at the topmost political level. Epigraphical evidence
suggests that the feudatories of the Gupta kings continued as the local
governors under Hfuna rule.99 Hiunas used Sanskrit as their official language
and patronized Hindu cults and sects.
The impact of the Huns was greater in other spheres. Hun activities in
central Asia affected north Indian trade which had close links with central
Asia. Furthermore in the wake of the Huns came a number of other tribes
and peoples from central Asia jostling for land and occupation in northern
India. This led to a migration of peoples in these parts which in turn upset
one of the stabilizing factors of the caste structure, the inter-relationship
between caste and locality. Some of these movements of peoples from the
north southwards can be traced in the place names and the caste names, as
in the case of the Gurjaras and Abhiras.o00
95 Mahdbhdrata, Adi Parva, 174, 38; Mahavastu, I, 135; Raghuvamsa, IV, 67-8.
96 Eran Stone Boar Inscription, Corpus Inscriptionum Indicarum, Vol. III, p. 158; G. R.
Sharma, Excavations at Kaudmsbi, pp. 15-16.
97 Rajatarahginf, I, 306-7; Kalhana calls him the 'god of destruction'.
98 E.g. Hsiian Tsang's descriptions: S. Beal, Buddhist Records of the Western World, I, pp.
171 ff.
99 Dhanyavisnu the brother of Matrvisou (visayapati of the Gupta king Budhagupta)
became the feudatory of Toramana. Cf. The Eran Inscription of Budhagupta, Corpus
Inscriptionum Indicarum, III, p. 89 with the Eran Stone Boar Inscription of Toramana, op. cit.,
p. 158. Budhagupta in his inscription is referred to merely as bhupati (king), whereas Toramana
takes the full imperial title of Mahdrajddhiraja and is described as 'the glorious', 'of great fame
and lustre' and 'ruling the earth'.
100 It is believed that the Gurjaras came from central India after the sixth century A.D. and
were of Tocharian extraction, D. R. Bhandarkar, Indian Antiquary, January 1911, p. 21-2;
A. C. Bannerjee, Lectures in Rauput History, p. 7; P. C. Bagchi, India and Central Asia, p. 17.
Place names in the Panjab-Gujerat, Gujeranwala, etc.,-suggest a settlement there as do the

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
Politically too the period from the sixth to the ninth century tend
be unstable in northern India, barring perhaps the reign of Hars
kingdoms of the northern Deccan were also beginning to take a p
interest in the areas adjoining the Vindhyas, which culminated
attempts of the Rastrakuta kings to capture and hold the city of Kan
In addition to this the system of making land grants to brahmans an
secular officials (to the latter in lieu of salary) was becoming more w
spread.101 In cases where the land was virgin the system resulted
expansion of the agrarian economy. The tribes of central India were
to adjust to both the population movements from the north as also t
encroaching agrarian economy often in the form of enforced settlem
of brahmans and agriculturalists. That this is also the period in whic
areas on the fringes of the Vindhyan uplands give rise to a num
principalities some of which play a major role in the politics of
India is not surprising. Some provided armies to neighbouring
others became the nuclei of new states which arose on the debris of
dynastic changes. The area continued to be a major artery of trade which
made it a prey to many ambitious dynasties and the scene of constant
battles. This uncertainty benefited the tribal peoples who exploited it t
secure power for themselves.102 However, many parts of central Indi
remained comparatively untouched by either the agrarian economy or
Aryan culture since pockets in this part of the sub-continent still harbour
Dravidian and Mundi-speaking tribes existing at a food-gathering stage
or at most, using primitive agriculture.
From the ninth century A.D. political power moved more recognizably
into the hands of the erstwhile feudatories, the recipients of land grants.
The new feudatories in turn became independent kings, granted land and
revenue in lieu of salaries to their officers, and to learned brahmans fo
the acquisition of religious merit. The legal sanction of the grant wa
generally recorded in an inscription in stone or on plates of copper, an
101 This situation is discussed by R. S. Sharma in his book, Indian Feudalism.

102 Ghatiyala Pillar Inscription, Epigraphia Indica, IX, p. 280.
presence of the Gujjar herdsmen in Kashmir. The Gurjara Pratiharas ruled in western Indi
and there is the more recent Gujerat as a name of western India. The existence of the Gujja
caste in Maharashtra points to a further movement towards the south; I. Karve, Hindu Society.
The Bad-Gujar clan survives among the Rajputs as also the brahman caste, Gujar-Gauda.
The Abhira are nomadic herdsmen who are believed to have migrated into India with th
Scythians. Some of them very soon rose to importance, such as the general Rudrabhfti,
Gunda Inscription of A.D. 181 in Epigraphia Indica, VIII, p. 188. They are located in the lowe
Indus and Kathiawar region, Bhdgvata Purana, 1, 10, 35; Periplus, 41; Ptolemy, VII, 1, 5
The Abhiras are described as mlecchas and sadras in status, Manu, X, 15; Mahdbhdsya, I, 2,
72. They gradually took over political power from the gakas and the Satavahanas and sprea
down the west coast of India where there is mention of the Konkanabhira, Brhatsamhita, 14,
12; 5, 42; 14, 18. Samudragupta in the Allahabad prasasti refers to the conquest of the
Abhlras, Corpus Inscriptionum Indicarum, III, 6 ff. A tenth-century Pratihara inscription
speaks of removing the menace of the Abhiras in western India, Ghatiyala Pillar Inscription
Epigraphia Indica, IX, p. 280.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
428 ROMILA THAPAR
the preamble to the grant contained the genealogy of the kings. The re-
markable fact of these genealogies is that most kings claim full ksatriya
status on the basis of a genealogical connection with the ancient royal
families, the Sfryavamsa (Solar lineage) and the Candravamsa (Lunar
lineage); or else there is the myth among some Rajput dynasties of the
ancestor having emerged from the sacrificial fire, the Agnikula lineage.
Such genealogical connections were claimed by the majority of the dyna-
sties of this time though not all.103 What is even more significant is that
most of these families are found on examination to be at least partially if
not wholly of non-Aryan origin.104 Thus instead of being described as
mleccha kings, they claim ksatriya status and have had genealogies
fabricated to prove the claim. Whereas the gakas and Yavanas were
denounced as vratya ksatryas and the Andhras were described as mleccha
kings, the kings of this period, some of whom came from mleccha stock
such as the Gonds and Gurjaras, are willingly accorded kpatriya status.
Why did the brahmans agree to this validation? It is possible that the
distinction between arya and mleccha had become blurred in actual practice
although the dharmasdstras continued to maintain it. The system of land
grants appears to have played a significant part. Brahman grantees were
often given land in virgin areas: thus they became the nuclei of Aryan
culture in non-Aryan regions.105 This process having started in the early
centuries A.D. not only resulted in more land coming under cultivation
but also Aryanized fresh regions. The return on the part of the brahman
may have been the fabrication of a genealogy for the new ruler.
The advantage of the fabricated genealogy was that mleccha antecedents
were soon overlooked or forgotten, particularly in those areas where the
mleccha had become powerful. In a ninth-century inscription of a Calukya
feudatory of the Pratihara king great pride is taken in 'freeing the earth
from the Hfuna peoples'.106 At almost the same time a Guhilla king of the
103 The Gafga and Candella dynasty claim Candravasmi descent, the Gurjara-Pratiharas
Saryavamsi descent and the Parmaras regard their ancestor as having emerged from the
Agnikula. The Gfhilas, the Calukyas of Venigi, the Cilukyas of Badami and the Calukyas of
Kalyani all claim solar descent, D. C. Sircar, 'The Guhila Claim of Solar Origin', The Journal
of Indian History, 1964, No. 42.
104 An example of this, which was a common condition, is discussed in D. C. Sircar, The
Guhilas of Kishkinda. Even the Khasa chiefs claim ksatriya status in the Bodh Gaya inscription,
Epigraphia Indica, XII, p. 30. The Pratihara claim to descend from Laksmapa the younger
brother of Rama who acted as a door-keeper (pratihdra) is very suspicious, Indian Antiquary,
January 1911, p. 23.
105 R. S. Sharma, 'Early Indian Feudalism', in Problems of Historical Writing in India
(S. Gopal and R. Thapar, ed.) p. 74. These ideas are further worked out in his Social Changes
in Early Medieval India.
The same policy was adopted by the Mughals who located colonists in these areas partly
to encourage them in the ways of Islam and of'civilization' and partly to keep a check on them,
particularly at the time of the Maratha-Mughal conflict when the Vindhyan tribes occupied
a strategic geographical position. It is not surprising that, during the period of British rule in
India, Christian missionaries were extremely active in these regions.
106 Una Pillar Inscription of Avanivarman II dated A.D. 899, Epigraphia Indica,
IX, p. 6 ff.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
Udaipur region proudly married the daughter of a Huina king.107 Yet th

founder of the Guhilla dynasty claimed to be a brahman. Marriage allianc
broke the kinship barrier and mleccha rulers became patrons of Sansk
learning and culture, so that they were as good as the dryas for all pract
purposes. Ultimately the Hunas came to be regarded as on a par with
Rajput clans and today the name survives merely as a caste name in t
Punjab.108 The degree of assimilation can be seen in the fact that t
accepted lexicon, the Amarakosa, in its definition of mleccha mere
lists the three tribes-the Kirata, Pulinda and Sabara.109 The names o
erstwhile mleccha tribes are defined according to occupations. Thus t
Abhiras are herdsmen, the Ambasthas physicians and scribes and t
Darada dealers in antidotes. The erstwhile mleccha-desa are described with
reference to their produce: thus Vafiga produces tin and Yavana-desa
horses fit for the asvamedha sacrifice.
The process of Sanskritization (the acquisition of Sanskritic culture and
higher ritual status) was usually spread over some centuries. The Bedars,
a mleccha tribe of the Deccan, are recorded in seventh-century A.D. sources
as molesting brahmans who had received land grants and settled in the
new areas.110 It is stated that these plundering raids had to be warded off
by the villagers themselves as the king could not enforce law and order in
those areas.1l This situation continued until about the thirteenth century.
Gradually the Bedar chiefs themselves were bought off with land grants
and other concessions.112 In periods of political confusion the chiefs began
to found independent principalities. Trouble between the Bahmani kings
and Vijayanagara was fully exploited and the Bedars not only plundered
the city of Vijayanagara in 1565 but strengthened their principalities.
Sanskritization continued apace and can be seen in the claim of the Bedar
kings to a high ritual status in the use of Sanskrit names such as mahdnayak-
dcarya, and also in the endowment made to the temple of Gopala-Krsna
by the Bedar chief in 1568 and ultimately in the fact that the famous
Saivite saint Kannappa was of Bedar origin.113
From about the ninth century onwards references to large numbers of
indigenous peoples as mleccha begin to decrease. Where they are men-
tioned and are other than the Vindhyan tribes, it is generally for a parti-
cular reason. The tenth-century Abhira king is called a mleccha because
he indulges in beef eating and plundering the pilgrims who visit the famous
107 Atpur Inscription of gaktikumar, Indian Antiquary, XXXIX, p. 191 ff.

108 Kanhadeprabandha of Padmanabha, a fifteenth-century work, mentions a Huia among
the list of Rajput jtgirdars, The Journal of Indian History, XXXVIII, p. 106.
109 Amarakosa, II, 10, 2; 5, 16; 8, 13; 4, 11; 4, 29; 2, 13.
l0o Rice, Mysore and Coorg from the Inscriptions, p. 5.
11 Epigraphia Carnatica, VII, p. 188; VI, p. 113-14.
112 The Gafiga king Kofigunivarman gave a grant in A.D. 887.
113 B. N. Saletore, Wild Tribes in Indian History, p. 81 ff.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
430 ROMILA THAPAR
temple at Somanatha.114 In eastern India there is the interesting inscr

tional reference to the kingdom of Kamarupa (Assam) being occup
a mleccha ruler, Salastambha, who starts a new dynasty.1ls We ar
told why he is a mleccha. Was he of tribal origin or did he have T
connections ?
Among the foreigners with whom there was a fair amount of contact,
especially through trade, were the Chinese, the Arabs and the Turks, all
of whom were of course considered mlecchas. Contact with the Chinese
goes back to the third century B.C. through trade in silk. Although silk
was greatly appreciated in India, the Chinese were firmly relegated to the
ranks of the barbarians and their land declared unfit for sraddha rites.116
They are often associated with the Kamboja and the Yavana (presumably
because of the central Asian connection) and with the Kirata and eastern
India-the two regions from which trade with China was conducted in
the early period.117 But the interest in China waned with the arrival of the
Turks on the northwestern borders of India and the Arabs in the west.
The Arabs are most frequently referred to as Yavanas and are regarded
as mleccha.118 The former relates to the fact that they came from west Asia
and were in a sense the inheritors of the earlier Yavana role in India. The
Turks are described correctly as Turuskas in some cases but more often
they too came under the general term mleccha or are called ?akas and
Yavanas.119 The latter was probably the result of their coming from the
same geographical direction as the earlier invaders. It would suggest that
to the Indian mind the Turks represented a historical continuity of the
Sakas and Yavanas. It does however point to a comparative lack of
interest in events across the frontiers of the sub-continent that the new
invaders should not have been clearly demarcated from the old. It is also
possible, however, that in using the old terms there was a sub-conscious
attempt on the part of the Indian rulers to compare themselves with earlier
kings who had tried to stem the tide of the Saka and Yavana invasions.
114 Ray, Dynastic History of Northern India, II, p. 941.
115 Bargaon Copper-plate of Ratnapala, Journal of the Asiatic Society of Bengal, p. 99;
Parbatiya plates of Vanmalaveramadeva, Epigraphia Indica, XXIX, pp 145 if. It has been
suggested that the name galastambha approximates a Sanskritized version of the name of the
Tibetan king, Sron-bstam-sgam-po.
116 Mahabhdrata, Adi Parva, 174, 38; Manu, X, 43-4; Matsya Purdna, 16, 16.
117 Brhatsamhita, V, 80; Mdrkan.deyaPurdna, 57, 39. Chinese interest in eastern India during
the seventh century A.D. is attested to in the reign of Harsa and by his contemporaries in Assam.
The pedestal inscription on the tomb of Tai Tsung mentions a diplomatic connection with
eastern India.
118 Gwalior Inscription of Nagabhatta I; Sagar Tal Inscription, Epigraphia Indica, XVIII,
p. 107 ff. An Arab attack on Kashmir in the eighth century is mentioned in the Rajatarahgini,
VIII, 2764.
119 Mahamadi Sahi Inscription, Epigraphia Indica, I, p. 93; Jaitrasimhadeva grant, Epi-
graphia Indica, XXXII, pp. 220 if.; Vilasa grant of Prolaya Nayaka, Epigraphia Indica,
XXXIII, pp. 239 ff.; Chitorgarh pralasti, Journal of the Bombay Branch of the Royal Asiatic
Society, XXIII, p. 49; Madras Museum Plates, Epigraphia Indica VIII, p. 9; Bhilsa Inscription
of Jayasimha, Epigraphia Indica XXXV, p. 187; Dantewara Inscription of A.D. 1703, Epi-
graphia Indica, IX, p. 164.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
Perhaps this degree of romanticism was essential to the me

ethos.
It was after all the same romanticism which led comparatively
kings to claim suzerainty over vast areas of the continent. The
recurring list of places which occurs in many of the inscriptions
period and becomes almost a convention and which reads
suzerainty over the mleccha, Afiga, Kalifga, Vafga, Odra, P
Karnata, Lata, Suhma, Gurjara, Krita and Cina....'120 It is not cl
this case who the mleccha were, whether they were the Arabs or indig
people, although it could well be that the word was used in an ad
sense to cover these places which were in the earlier tradition regard
mleccha-desa. A similar convention relates to the conquest of the
peoples and the capture of their hill forts such as Anarta, Malava, Ki
Turuska, Vatsa, Matsya, etc.121 The 'eulogy' style of inscriptions in w
these conventions are observed continued to be used even for the Turkish
Sultans after they had established their rule.
Mleccha as a term of exclusion also carried within it the possibility of
assimilation, in this case the process by which the norms of the sub-culture
find their way in varying degrees into the cultural main stream. Assimilation
can be achieved at various levels. The obvious forms are noticeable in
external habits such as names, dress, eating-habits and amusements. The
more subtle forms are those which can be seen in the framework of law
and of religious beliefs. The Sanskritizing of names was a common feature
among both indigenous and foreign mlecchas who slowly tried to move
away from their status of mleccha.l22 Very often in the case of ruling
families it took one or two generations to make the transition. In other
situations it took a longer time. The importation of foreign fashions is
evident from the terracotta and stone sculpture of various periods. The
tendency was to follow the dictates of the court circles. The couture of the
deities however was more rigidly bound by conventional forms. Assimila-
tion can also be seen in the appropriation of melodies and musical forms
associated with mleccha peoples into the mainstream of music.123 One of
the most direct forms of the expression of brahmanical ritual purity was on
the form and type of food which the brahman could eat. He was forbidden
to accept cooked food from any non-brahman.124 Eatables were ranked in
120 Bhaturya Inscription of Rajyapala, Epigraphia Indica, XXXIII, p. 150; Chitorgarh

prasasti of Rana Kumbhakarna, Journal of the Bombay Branch of the Royal Asiatic Society,
XXIII, p. 49; Balaghata Plate of Prithvisena II, Epigraphia Indica, IX, p. 270.
121 Sagar Tal Inscription of Mihira Bhoja, Epigraphia Indica, XVIII, p. 107.
122 gaka inscriptions reveal this very clearly as also the names of the Indo-Greeks, Epi-
graphia Indica, VII, p. 53, 55; Epigraphia Indica, VIII, 90; Archaeological Survey of Western
India, IV, pp. 92 ff.
123 Mention is made of the Gandhara and Kamboja melodies as also of gaka and Abhira
melodies, Panfcatantra, Apaniksetakanakam 55.
124 From this point of view at least Indian eating habits and rituals would form an ideal

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
432 ROMILA THAPAR
a carefully determined order of priority. Thus when the Punjab becam

mleccha area, its staple food was given a lower place in the hierarc
food-ranking. Whereas the Rg-Vedic Aryan had a staple diet of wheat
barley, by the twelfth century A.D. wheat was described in one lexico
'the food of the mlecchas' (Mleccha-bhojana) and rice became the
cereal.125 Onions and garlic were also regarded as the food of the mle
and therefore prohibited to the brdhman. One of the habits of the mle
which seriously defiled them was the fact that they drank alcohol an
the flesh of the cow, and this in later periods was strictly forbidden
Aryan twice-born.126
We have seen that an essential difference between the drya and
mleccha was that the latter did not conform to the law of varna. On one
occasion the god Indra is asked how the Yavanas, Sakas, Cinas, Kambojas,
Pulindas, etc., can be brought within the social pale, and he replies that if
they follow the dharma of the sastras (essentially the law of the var.ia),
they can be admitted.127 For the laws of the mleccha and the laws of the
dryas were distinct. As was the case with other jatis, the mleccha appear
to have had their own customary laws and functioned within the frame-
work of these. Within the law of the sastras a sharp differentiation was
maintained between the status and rights of the drya and the mleccha. A
significant and relevant example of this is that the mleccha is permitted to
sell or mortgage his own life and that of his offspring.128 But an drya can
never be subjected to slavery, except for very short periods when he is in
adverse circumstances.
An even more subtle form of assimilation was through the incorporation
of cults and cult-priests into the religious beliefs and rituals of the estab-
lished religions of the aryas. In the case of the Buddhists the problem was
easier since there was not the same stress on ritual ranking as among the
brdhmans. The gaka and Yavana rulers and particularly their queens who
were patrons of Buddhism were accepted as fully as other Indian ruling
families.129 For the indigenous mleccha the acceptance of Buddhism did
not necessitate the disavowal of earlier cults, since Buddhism has com-
monly assimilated local cults in its process of expansion. Buddhism itself
arose in mleccha areas and it is significant that the main strongholds of
125 Trikaipdasesa in Namalinganusasana of Amarakosa.

126 For the prohibition on onions and garlic, Manu, V, 19; for references to eating the flesh
of the cow, Jaimint, I, 3, 10 and Rdjatarahgini, VII, 1232.
127 Mahabhdrata, ganti Parva, LXV, 13-15.
128 Visnu Dharma.astra, 84, 4; ArthaaSstra, III, 13-15.
129 Mathura Lion Capital Inscription, Epigraphia Indica, IX, p. 141; Mandasor Inscription,
Epigraphia Indica, VIII, p. 95; Visnudatta Inscription, Epigraphia Indica, VIII, p. 88. Saka
kings often refer to themselves as dharmika on coin legends with the symbol of the Dharmacakra
on the coin.
subject for structuralist analysis, along the lines of the theories developed by L6vi-Strauss.
See Manu, IV, 205-25; 247-53; for laws regarding the acceptance of various kinds of food.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
Buddhism were in these areas. However, it tended to by-pass the tribe

Vindhyas probably because the nature of their cults, stressing violence
shedding of blood at sacrifices, precluded easy acceptance into Buddh
The brahmanical religion did not remain rigid either. The Bha
tradition in Vaisnavism and Saivism which emerged in the early cent
A.D. stressing the personal devotion, bhakti, of the worshipper
individual deity, made the religion more flexible and more easily exp
able. It was this tradition of brahmanism that could and did attract
mlecchas. The Greek Heliodorus records his devotion to Visnu and sp
of himself as a member of the Bhagavata sect.130 The Hiunas ap
have been quite acceptable to both the major sects of Hinduism. Tora
was a Vaipnavite and was a patron of those who worshipped the
(boar) incarnation of Visnu. As a royal patron he was the direct succe
to one of the Gupta emperors who had earlier donated a cave t
worship at a place not too far from the site of Toramana's inscriptio
Mihirakula was such an ardent Saivite that he was led to an extreme
intolerance of the Buddhists and Jainas, again a tradition which is recorded
of earlier rulers of Kashmir.132 Perhaps the Sun and Fire cults of the Hiunas
acted as a bridge towards their acceptance of and by Hinduism. With the
strengthening of the Bhagavata tradition there was a proliferation of new
sects, some of which in their social attitudes were recognizably anti-
brahmanical, such as the Saiva Siddhantas and others which maintained
a flexible attitude to caste such as the Lifgayatas. As in the case of the
Buddhists and Jainas, such sects did not discriminate between airya and
mleccha peoples, and for the latter this became an avenue of entry into
Aryan society, since ultimately many of these sects became independent
castes within the varna system.
In the case of the indigenous mleccha many of the cults were slowly
absorbed into the main cultural tradition. Of these perhaps the most
obvious were the fertility cults, especially those devoted to the worship of
the mother goddess, and the phallus (lingam) and snake cults.133 These
cults were not totally foreign to brahmanism, but in the period after the
fifth century A.D. they began to play a more dominant role in the evolution
of Hinduism.134 The mother goddess, Devi, in various manifestations
130 Journal of the Royal Asiatic Society, 1909, pp. 1053 ff.
131 Eran Stone Boar Inscription. The varaha cave is at Udayagiri.
132 As for example the reference to Jalauka in the Rajatarangini, I, 108-52.
133 The snake cult or worship of the Naga is attested to in literature as well as in the
archaeological remains of a multitude of naga shrines. It is frequently seen as the symbol of
the chthonic goddess, of the ancestors and of lunar and fertility cults, and is commonly found
even to this day in the Himalayan and Vindhyan regions. In the historical period it gained
considerable respectability particularly in the peninsula.
134 There is mention in the Rg Veda of the pre-Aryan cults such as the worship of the phallus,
giSnadevah, and the existence of sorceresses, yatumati, practising magic. The Harappan
evidence clearly indicates the worship of the mother goddess which was new to the Aryan
religion.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
434 ROMILA THAPAR
appears to have been the most popular deity among the mleccha. Vindh
vasini, one of the names for the consort of Siva, was worshipped b
Sabaras, Barbaras and Pulindas.135 The name itself means 'she
inhabits the Vindhyas', and clearly she was in origin a mountain godde
She is said to be commonly worshipped by brigands, and the rites invo
the eating of meat and the drinking of wine.136 In another form s
described as the goddess of the outcastes who bring her oblation
sacrificed animals.137 Elsewhere she is identified with Narayani and Du
both well-known manifestations of Siva's wife and both repeated
associated with the mleccha tribes in early literature.138 The name Sav
meaning a Sabara woman, occurs as the name of a goddess in a med
work.139 The Savarotsava or Festival of the Sabaras was a bacchanalian
gathering of the tribe, as well it might have been with a fertility cult as its
focus. The Kirata worshipped the goddess Candika, yet another manifesta-
tion of Siva's wife Durga, a more fearsome form of the goddess being
responsible for the destruction of the buffalo-demon Mahisasura. The
Devi Mahdtmya, one of the more important sources on the mother-goddess
cult, suggests an eastern if not Tibetan origin for the birth of the goddess
Candi.140 By the medieval period the cults of Durga and Candi had been
absorbed into classical Hinduism. In fact, a substantial part of Hinduism
itself had undergone transformation with the popularity of the Sakti-Sakta
cults and Tantricism.
Nor were the cult priests left behind. Depending on the status of the
cult they would enter the hierarchy of brahmanism. As the cult became
refined and found a niche in classical Hinduism the cult priest would also
become Sanskritized and be given ritual status in the brahman varna. This
would account for the existence of contradictory categories such as the
Ambastha brahman and the Abhira brahman. It would also explain the
gradual evolution in status of the Maga brahmans who are said to have
come from Sakadvipa in the west.141 They are at first looked down upon
and not admitted to all the sraddha ceremonies. This may have been
because they were soothsayers and astrologers rather than genuine
brdhmans or else because of their association with the sun cult, which,
being a more powerful religious force in western Asia, may have been
regarded as somewhat foreign.142 But gradually their position improved
135 Harivamsa, II, 22, 59. 136 Ibid., II, 22, 53-4.
137 Ibid., II, 3, 12. She is sometimes described as krsnachavisdma krsna (as black as can be),
adorned with peacock feathers and with dishevelled hair. Bana, writing in the seventh century
A.D. when speaking of the mleccha tribe of the Vindyas, describes a Durga temple, Kddambari,
p. 331. Of the Pulindas said to be living in the Vindhyan region, an eleventh-century text states
that their king adores the cruel Devi, offers her human victims and pillages the caravans,
Kathdsaritasdgara, IV, 22.
138 Harivamsa, II, 58; DaSakumaracarita, I, 14; VI, 149; VIII, 206.
139 Vakpati, Gaucavdho, V, 305. 140 Mdrkan.deya Purana, LXXXII, 10-18.
141 Bhavisya Purana, II, 26; I, 39. Samba Purana, 27, 28.
142 Mahdbhdrata, Anu Parva, XC, 11. Manu, III, 162.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
when they were patronized by the royal courts, especially at Thane

Kamarupa, and they were regarded as the proper people to in
consecrate images of Surya the sun-god.143 Their association with
cult remained constant. However they still married into non-b
castes such as the Bhojas and the Yadavas. By the medieval period h
they were treated with considerable respect. The curious legend
are told about the origin of certain brahman families such as those
Chitpavans who virtually walked in from the sea,144 would also
that these were families of cult priests who were gradually assimil
the Hindu social structure.
There was however one facet in the concept of the barbarian which was
absent-the notion of the pagan. This did finally arrive in India but never
became an intrinsic part of the Indian notion since the form of the indi-
genous Indian religions had no use for this concept. It was applied by the
Muslims who came to India to the non-Muslim inhabitants of India. They
were regarded as pagans and by extension less civilized. From about the
fifteenth century onwards, when Turkish and Afghan rule had been estab-
lished in virtually all parts of the sub-continent, the Muslims at all levels
of society came to be described more extensively as mleccha. They were
mleccha partly because they were foreign in origin, but what was more
important they spoke an alien language (either Arabic, Turki or Persian)
and they could not conform to the laws of varna since Islamic laws demand
an egalitarian society. Certainly they did not observe the rules of ritual
purity. Gradually however the social organization of the Muslims began
to approximate that of the Hindus in that various castes evolved and
became similar to Hindu castes in many matters. A focusbof separation
was then provided by the distinctively theological quality of Islam which
took on a forceful shape alien to Indian notions of religion. It is also
possible that since a sizeable proportion of conversions to Islam in India
were from the lower castes (conversion to a non-caste religion being one
of the traditional methods of trying to by-pass caste), this also encouraged
the use of the description mleccha.
The most significant clue to assimilation lies not so much in the loss of
ethnic identity as in participation in the sense of the past. There is the
mutual appropriation of the past on the part of two groups where the group
with the weaker historical tradition accepts the stronger tradition. This
was certainly the case with the foreign peoples who settled in India and
with the indigenous tribes. Sanskritization implied the acceptance of the
historical tradition to the same degree as the organization of the tribe
according to the laws of varna and jati. Hence the importance of genealo-
143 Brhatsamhita, LX, 9.

144 Maureen Patterson, 'Chitpavan Brahman Family Histories', in Structure and Change in
Indian Society, (ed.) Milton Singer and B. Cohn.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC
436 ROMILA THAPAR
gies in the process of both historical and social validation. Yet this sense
of the past was in itself the result of assimilation at various points in time
and was given direction by the elements which went into the making of the
social fabric. Islamic historiography however brought with it its own highly
developed philosophy of the past which had little in common with tradi-
tional Indian historiography except that they were both powerful traditions
within the culture.
It is perhaps the very contradiction in the Indian concept of the bar-
barian which makes it distinctively different from that of Europe. The
perception of differences-linguistic, cultural and physical-set the
barbarian apart. The separateness was seen not so much in terms of
what the barbarians did as in the fact that they did not observe the norms
of ritual purity and were to that extent polluted. The lack of description
of the mleccha, comparatively speaking, was based on the assumption that
no self-respecting man would associate with them as long as they were
designated as mleccha. In a sense, this was the ultimate in segregation.
Theoretically this position was maintained throughout. Yet in practice
not only were concessions made, as for example, in the notion of the
vrdtya-ksatriya, but large numbers of mleccha peoples were incorporated
into the social, political and religious system and were in fact the pro-
genitors of many of the essentials of Indian culture. It would be a moot
point as to whether this could be called a culture which excludes the
barbarian.

173.54.27.229 on Wed, 05 May 2021 19:09:36 UTC

The Image of The Barbarian - Romila Thapar

Uploaded by

Copyright:

Available Formats

The Image of The Barbarian - Romila Thapar

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Image of The Barbarian - Romila Thapar

Uploaded by

Copyright:

Available Formats

The Image of the Barbarian in Early India

Author(s): Romila Thapar

Stable URL: https://www.jstor.org/stable/178208

This content downloaded from

Jawaharlal Nehru University, New Delhi

The concept of the barbarian in early India arises out of the

This content downloaded from

spreading the Indo-Aryan language system through a major part

This content downloaded from

consonants are believed to have been assimiliated into Indo-Aryan fr

This content downloaded from

Brdhmana17 and is later quoted and discussed by a grammarian

17 Satapatha Brdhmana, III, 2, 1, 23; which reads, te'surd dttavacaso he'lavo h

This content downloaded from

to a respectable and honourable man.23 Ritual purity or the absence o

This content downloaded from

Theoretically this seems to be a fairly clear situation. But in fact ther

25 Manu, X, 10-12; 16-17.

This content downloaded from

This content downloaded from

had now been extended to include speakers of an alien language, s

This content downloaded from

This content downloaded from

boundary of Aryan control in the Ganges valley is perhaps referred to

45 Satapatha Brdhmana, I, 4, 1, 10.

This content downloaded from

is incomprehensible.53 It would seem from the Buddhist sources that lan-

This content downloaded from

Yavana, a back-formation from the Prakrit yona, which is said to

This content downloaded from

This content downloaded from

There is a general bewailing of the increase in mleccha influence which

This content downloaded from

This content downloaded from

This content downloaded from

81 Mahdbhdrata, Sabha Parva, LVII, 144.

This content downloaded from

the mleccha along the Chambal river and a fifteenth-century inscrip

This content downloaded from

This content downloaded from

101 This situation is discussed by R. S. Sharma in his book, Indian Feudalism.

This content downloaded from

This content downloaded from

Udaipur region proudly married the daughter of a Huina king.107 Yet th

107 Atpur Inscription of gaktikumar, Indian Antiquary, XXXIX, p. 191 ff.

This content downloaded from

temple at Somanatha.114 In eastern India there is the interesting inscr

This content downloaded from

Perhaps this degree of romanticism was essential to the me

120 Bhaturya Inscription of Rajyapala, Epigraphia Indica, XXXIII, p. 150; Chitorgarh

This content downloaded from

a carefully determined order of priority. Thus when the Punjab becam

125 Trikaipdasesa in Namalinganusasana of Amarakosa.

This content downloaded from

Buddhism were in these areas. However, it tended to by-pass the tribe

This content downloaded from

This content downloaded from

when they were patronized by the royal courts, especially at Thane

143 Brhatsamhita, LX, 9.