View source for Ancient Indian Ethnography: Aryan Ethnic Classification

{| class="wikitable"
|-
|colspan="0"|<div style="font-size:100%">
This article is an extract from <br/>

ETHNOGRAPHY OF ANCIENT INDIA <br/>

BY <br/>

ROBERT SHAFER <br/>

With 2 maps <br/>

1954 <br/>

OTTO HARRAS SOWITZ . WIESBADEN <br/>
*****<br/>
''Indpaedia is an archive. It neither agrees nor disagrees ''<br/> 
''with the contents of this article.''<br/>
''Secondly, this has been scanned from a book. You can help by ''<br/> 
''sending the corrected version/ additional information to ''<br/>  
the Facebook page, [http://www.facebook.com/Indpaedia Indpaedia.com]. <br/> ''All information used will be duly acknowledged.'' 
 
</div>
|} 
[[Category:India|C]]
[[Category:Communities|C]]
=Aryan Attempts at Ethnic Classification=
We cannot take the genealogies quite as literally as Pargiter. Instead 
we must consider the earlier generations as eponymous ancestors of 
the race. Everyone is familiar with the Biblical attempt at classification 
by dividing peoples into the sons of Shem, Ham, and Japheth, the 
sons of Noah. I am not a Semitic scholar and I neither know nor care 
whether the Semitic classification was correct or not. We are interested 
here only in the accuracy of the Aryan attempt at classification, which 
only time will definitely decide. 

The Aryans derived all the important races of India from Manu, 
literally "man." The fact that they did not derive the Nisadas from 
Manu is an indication that they were no longer important in epic times. 
Or, since the authors of the Rgveda regarded the Dasyus with contempt, 
they may have expressed this by excluding them from descent from 
Manu. 

Saudyumnas. The Saudyumna stock was considered to be descend- 
ed from a biological freak, which may have been the Aryan means of 
expressing their feeling that this stock was not quite human. The identi- 
fication of the Saudyumna stock with the modern Mundic peoples seems 
reasonable, but it is not particularly important for our purposes because 
we can identify Mundic countries by other means. 

Constable's Atlas, plate 9, shows the Munda race extending right 
up to Bhagalpur and within 75 miles of Patna. 1 

1 For linguistic evidence that the Mundic stock has been in India a very 
long time, see Manfred Mayrhofer, "Neue Literatur zu den Substraten im Alt- 
indisehen," Archiv Orientdlnl, Vol. 18 (Symbolae Hrozny), no. 4 (1950), 367 371, 
with bibliographical references. 

Hamit Kosay's "Tiirkische Elemonte in den Munda -sprachen (Munda diller- 
indeki Tiirkce unsurlar)," Belleten (Turk Tarih Kurumu) 3 (1939), p. 107-126, 
presents comparisons between Turkish and Finno-Ugrian and Mundic, some of 
them sufficiently close phonetically to warrant further study. The works of W. F. 
von Hevesy, Munda- Magyar -Maori, An Indian Link between the Antipodes; New 
Tracks of Hungarian Origins (London, 1928), under pseudonym F. A. Uxbond, 
and Finnisch-Ugrisches aus Indien; es gibt keine austrische Sprachenfamilie das 
vorarische Indien teilweise finnisch-ugrisch (Vienna, Manz, 1932) are not available 
to me. A review of von Hevesy's works is given by P. W. Schmidt, ,,Die Stellung 
der Munda- Sprachen", BSOS 1 (1935), 729 738. Since work on the so-called 
Ural-Altaic and Austroasian families is now progressing somewhat under various 
scholars, soon we may more competently decide whether the parallels presented 
by Kosay and von Hevesy are purely accidental, whether they are loan words 

Manavas. The Aryans thought that the earliest race in India was 
the Manava, in two branches. That in the east, descended from Manu's 
son Iksvaku, occupied the kingdom of Ayodhya (later Kosala), and 
Videha, and still later Vaisali. The western branch, descended from 
Manu's son Saryati, occupied Anarta. 

We have very good reasons for believing that this Aryan classification 
expresses historical truth, for we know that there was a great city at 
Ayodhya before the Aryans came, and, since the excavation at Mohenjo- 
daro, we find that there was a very early urban civilization in the west 
of India, anteceding that of the Aryans. 

I am inclined to believe that the western Manavas at Anarta were 
Iranians. 1 The Iranians were the closest people on the west of India 
that we know anything about. And nearly half of the Mohenjo-daro 
skulls which could be classified belonged to the Mediterranean (Hamitic, 
Iranian) type, today a dominant element in the north of India and 
generally among the upper social classes. At Mohenjo-daro was found the 
sculptured head of a man with a short beard ("chin whiskers"). A "chin 
whisker" beard was a characteristic of many Persians in their sculptures, 
and the Pahlavas are distinguished in the Mahabharata from other 
people by their beards. 

Pandey 2 has protested against Pargiter's separation of the Manavas 
from the other peoples of India: "the Manavas are not described as a 
people distinct and different from the Ailas . . . Both are described as 
a people similar in race, language and religion." If the Manavas were 
Iranians, they would be similar in race, language, and religion to the 
Indo-Aryans, one of the Aila peoples. Moreover, Pandey 3 shows the 
Manavas to be Iranians when he states that the Sakas were descended 
from Narisyanta, son of Manu. 

If the Aryans were right in considering western and eastern Manavas 
as closely related, then the Kosalas, Videhas, and Vaisalas would also 
have been Iranians. But a past attempt to connect them with the Iranians 

and, if so, from what period, or whether they signify distant genetic relationship. 
If either of the last two possibilities should prove correct, they would be of con- 
siderable importance for the prehistory of India. 

1 Those inclined to daring geographical jumps may refer to P.-E. Dumont, 
"Indo- Aryan Names fromMitanni, Nuzi, and Syrian Documents," JAOS 67 (1947), 
251 3, and G. Dumezil, "Dieux cassites et dieux v6diques a propos d'un bronze 
du Louristan," Revue Hittite et Asianique 11, fasc. 52 (1950), pp. 18 37; both 
have bibliographical references. The Near East and India are also discussed in 
many of the works referred to in Constantin R^gamey's "Bibliographic analytique 
des travaux relatifs aux 616ments anaryens dans la civilisation et les langues do 
I'lnde," BEFEO 34 (1934), pp. 429-566. See also Wilhelm Brandenstein, ,,Die 
alten Inder in Vorderasien und die Chronologie des Rigveda", in Brandenstoin, 
Frtihgeschichte und Sprachwissenschaft (Vienna, 1948), pp. 134145. 

2 P. 101. 
8 P. 96. 
was not well received. Some traditional stories link the Eastern Manavas 
with Tibet. 1 We shall return to this question below. 

Ayava (Aila). The so-called Aila or Lunar race or, as I prefer to 
call it, the Ayava (from Ayu) appeared later, for there are two genera- 
tions between Manu and Ayu. Only the Kasis can be regarded as Ayava. 
Scholars have remarked ties between them and the Eastern Manavas, 
as Jala Jatukarnya, purohita of the kings of Kasi, Videha, and Kosala. 

Amdvasava. This line at Kanyakubja was established at about 
the same time as the Ayava. One may guess from their geographical 
and genealogical position that both the Ayavas and Amavasavas were 
colonies sent out from Ayodhya. 

Two generations later, according to Pargiter, came the descendants 
of Yadu, Turvasu, Druhyu, Anu, and Puru. Pargiter assumed that the 
Anavas, Drauhyavas, Yadavas (and Haihayas), Taurvasavas, and 
Pauravas were all Aryan. But we have already noticed that Pargiter 's 
Aryans would leave little room for aryanization in northern India during 
the millenniums since the time of the Mahabharata. Direct evidence 
that these people were not Sanskrit-speaking Aryans is the fact that, 
according to the Rgveda, the Anus, Druhyus, Yadus, Turvasas, and Purus 
attacked the Aryans at the Parusni (the Ravi), so they must have been 

1 According to these tales Iksvaku had a son Vikuksi who succeeded him at 
Ayodhya. In six puranas, Ik$vaku had other sons, 50 of whom were kings of 
Uttarapatha (northern India) and 48 in Daksinapatha (the Deccan). But accor- 
ding to the other version in four puranas, Vikuksi had 15 sons who were kings 
north of Mem and 114 others south of Mem (Pargiter). 

Now, according to Tibetan tradition, the legendary kings of Tibet commenced 
with Gnya-k'ri Btsan-po, said to be a son of King Prasenajit of Kosala (Luciano 
Petech, "A Study on the Chronicles of Tibet," IHQ 13, sup. p. 19). 

These traditions may go back to a common source. Vincent A. Smith concluded 
from ancient sculptures at Barhut (Bharhut) and Safici and from certain insti- 
tutions that the Tibetan type was much more prominent in the plains of northern 
India 1800 to 2000 years ago than it is today. He also remarked that polyandry, 
such as Draupadi's marriage to the five Pandava brothers, is still a Tibetan and 
Himalayan custom (Oxford History of India [Oxford, 1923], p. 7). 

D. S. Trivoda seems to hold that the eastern Manavas were Aryans ("The 
Pre-Mauryan History of Bihar," JBRS 36 [1950], pts. 3-4, pp. 134-147; 37 
[1951], pts. 1 2, pp. 127152). But Triveda's reference to vaisya kings and their 
marriage with vaisyas may be interpreted as indicating a considerable yellow- 
skinned population of some social importance among the eastern Manavas (pp. 131, 
142). 

Chatterji (p. 35), evidently basing his conclusions on Dines Chandra Sirkar 
in the Vaisali Volume not available to me, inclines to believe that the eastern 
Manavas were Mongoloid or people of mixed blood with a strong Mongoloid element, 
who had adopted or were adopting Aryan speech. 

In the Vayupurana, MS e, the Ko&alas are regarded as foreign (Pargiter, 
Kali Age, pp. 2, 3). 
native tribes hostile to the Aryans. And in the Rgveda anu means a non- 
Aryan; 1 Druhyu is considered a derivative of druh- "be hostile to"; 
Turvada is a derivative of turv- "to overpower." No derivation has 
been found for Yadu and we may infer that it was a native name. Puru 
has been compared with purusa "man, people." Thus four of the five names 
indicate by their etymology or lack of etymology that they were applied 
to non- Aryan or hostile peoples. 

Kangacharya has noted that the Yadus and Turvasas are called 
Dasas in the vedas, 2 and that the Dasas lived in fortified villages, 
which was not true of the Aryans. 

And the Mahabharata states explicitly that the sons of Turvasu 
are the Yavanas and that the sons of Anu are the Mlecchas. 3 And 
no scholar has ever contended that the Yavanas or Mlecchas spoke 
Sanskrit. 

If these peoples were not Sanskrit-speaking Aryans, what then? 4 

We are hampered in this investigation by having almost no informa- 
tion about India's neighbors on the west or north. The only recent 
statement I have seen regarding the early population of Afghanistan 
is that of Antonio Pagliaro. He said that it was Iranian and that the 
earliest inhabitants were the Persian-speaking Tajiks. 5 The source of 
his information, however, seems to have been Mountstuart Elphinstone, 
An Account of the Kingdom ofCabaul (London, 1815). 6 Although Elphin- 
stone's conclusions seem to be based on some research and observation, 
one may doubt whether his conclusions would still stand after a modern 
investigation. 

We do have, however, one bit of information regarding the early 
inhabitants of Afghanistan in Herodotus' statement that in 516 B. C. 
Skylax, a Karian of Karuanda, set out under Darius' orders from the 

1 Since a number of Indian gods have been found in Near Eastern countries 
(see p. 16, n. 1), anyone interested in trying to connect Anu with the Babylonian 
sky god Anu of Uruk is welcome to the theory. Although I express another view 
of the Anavas below, our knowledge of ancient relationships is not so assured that 
one can be didactic. 

2 Even Rangacharya, interested primarily in the Aryans as "carriers of 
culture," observed that the battle on the Paru$m was between Sudas and the 
rulers of Matsya, Sivas, and Yakus, and the kings Turvas"a, Druhyu, Puru, Anu, 
Vikarnika, and perhaps Yadu and others, and remarked "it is obvious that some 
of these tribes were non-Aryan" (pp. 192193). 

8 Or. Ed. I. 80. 26; Boy I. 85. 

4 Chanda, although holding to the usual view that the Yadus, Purus, Druhyus, 
Anus, TurvaSas, Bharatas, and others were Aryans, noted that they were tradi- 
tionally akin to the dark section of the rsis, the Kaiivas and Visvamitras (p. 25). 
This is a contradiction, for then they would be Aryans who wore not Aryans, or 
they would be at least a mixed breed. 

6 Enciclopedia Italiana (Milan, 1929), under Afghanistan. 

6 Pp. 309 ff. 
city of Kaspaturos 1 in the country of Paktu (naxTvixij yfj)- 2 This 
can hardly have been anything but the country of the Paxtus, the 
northern Afghan dialect corresponding to Pastu in the south. It indicates 
that these people were already in Afghanistan at least before the time 
when Hopkins would date the composition of the Mahabharata. 3 

Yddavas. According to the 'Harivamsa, Yadu was the grandson 
of Iksvaku of the clan of Manu and of the daughter of the demon king 
Madhu, and Yadu had his capital in Anarta, which was part of the 
former kingdom of the demon Madhu. Madhu ruled at Madhu vana, but 
Satrughna slew his son Lavana, cut down Madhuvana forest and founded 
the city of Mathura. 4 

Thus, according to the Yadavas' own history, they were descended 
from the first race in India, the Manava, and from the demon king 
Madhu. How should this be interpreted? 

I have mentioned above that although I considered the Bhils to have 
once occupied all western India, I no longer believe they were dominant 
in that area at the time of the Mahabharata. For although proto-Austra- 
loid (pre-Dravida) skulls were found at Mohenjo-daro, they were in the 
minority in comparison with the Mediterranean (Iranian) type, and 
archaeologists have inferred that the former made up the slave or servant 
class. And, although a report on most of the burials at Harappa is still 
awaited, the dominant skull form in the earlier phase seems to have 
been the Australoid, but it does not appear to be dominant in the later 
phase. This would seem to indicate that the aborigines were being first 
subjected, then enslaved, and finally eliminated. 

I think we must interpret the Yadavas' account of their origin as 
that the Manavas (Iranians) 5 were subjugating the Bhils, and that the 
Yadavas were their descendants who had conquered the Bhils from 

1 But Kaspapuros in the country of the Gandarii in Hekataios frg. 295 (Felix 
Jacoby, Die Fragmente der griechischen Historiker, pt. 1 [Berlin, 1923]). Instead 
of y I have written u throughout in transcribing Greek geographical names. 

2 IV. 44. 

3 See above, p. 4, n. 2, 2d paragraph. Perhaps these people were the same 
as the Pakthas ( < *Pakhtu), who fought in the Battle of the Ten Kings of the 
Rgvoda. For the Sanskrit substitution of final -a for peoples for the final -u of 
the root of foreign names, see Huna in Index. 

4 Harivam6a 54. 

6 Sir Herbert Risloy attributed the broad, round-headed elements among the 
Gujaratis, Mara^his, and Coorgs to Sythian admixture (The People of India, 
2d ed., Calcutta and London, 1915). 

And J. H. Hutton thought the distribution of brachycephalics worked out by 
Guha in 1931 indicated a third millennium B. C. invasion from the direction of 
the Iranian plateau which penetrated down the west coast of India as far as Coorg 
and apparently threw out offshoots down the Jumna and Ganges valleys to the 
Ganges delta (Man 43 [1943], p. 44). The Yadavas covered most of the area out- 
lined by Hutton, but did not, apparently, extend as far east as the delta of the 
Ganges. 
Anarta to Mathura and married their women. For we must take the 
demon Madhu to symbolize some primitive people who occupied at an 
early period all the territory from Mathura (i.e., from the Tibeto-Burman 
occupied Ganges valley?) to Anarta. The only non- Aryan people to 
occupy any of this territory today are the Bhils. And Krsna "black" 
was the great Yadava prince of Anarta, formerly of Mathura. 

Yadu founded seven clans, according to the Yadava history, the 
Bhaima, l Kaukura or Satvata, Bhoja, Andjhaka, Yadava, Dasarha, and 
Vrsni. The Vrsnis and Andhakas were branded as vratyas in the Maha- 
bharata 2 and Krsna himself was so regarded, 3 and the Surastra kings 
were considered vratya and mostly sudra (i.e., black) in the puranas. 4 
Baudhayana states in his Dharmasutra 5 that the inhabitants of Anarta 
and Surastra are of mixed origin. Devala 6 said that he who visited 
the Saurastras should be initiated anew. The Periplus of the Erythrean 
Sea calls either the Saurastras, or the Abhiras, or both, men of great 
stature and black in color, 7 which would indicate a mixture of a tall 
race with the black Nisadas (Bhils), and further absorption of the latter. 

The Yadavas occupied former Madhava territory at Surastra, Anarta, 
and Mathura, and the outcast status was no doubt due to their inter- 
mixture with the Bhils. 

That some of these Yadavas of Manava-Madhava (Iranic-Bhil) 
descent had become aryanized, but not completely by the time of the 
Mahabharata is indicated by the statement of Karna that "the Salva 
cannot comprehend until the whole speech is uttered," the Salva being 
Yadavas. 

Haihayas. The Haihayas are considered a branch of the Yadavas, 
and we find Pahlavas, Sakas, and Kambojas (all Iranians) associated 
with the Haihayas in an attack on Ayodhya. 

The branches of the Haihayas were: Vitihotras, Bhojas, Saryatas, 
Avantis, Tundikeras, and Talajangas. The Vitihotras were thought of 
as foreign in the Vayupurana, MS e, 8 and the king of the Avantis as 
vratya and mostly sudra. 9 

From the position of the Haihayas within and north of the Vindhya 
range, I would infer that the base of the population was not Bhil, as 

1 Descendants of Bhima, who ruled over Surastra. 

2 VII. 143. 15. 

3 MRK, p. 95. 

4 Pargitor, Kali Age, pp. 54 55. 

5 I. 1.32-33. 

6 Quoted by Vijfianevara on Yajiiavalkya III. 292. 

7 Par. 41. Hjalmar Frisk gives the original Greek text in "Le P6riple do la 
mer Erythr6e," GOteborgs HOgskolas Ars skrift 33 (1927), 1 145. The text is in- 
definite in reference to the people. 

8 Pargiter, Kali Age, pp. 2, 3. 

9 Op. cit., 54-55. 
with the Yadavas, but either Kol or Gond, and hence the distinction 
between Haihayas and Yadavas. 

When the racial-linguistic map in this work had been filled in, it 
appeared that there were only a few scattered Nisada tribes left, and 
that the Yadavas, Haihayas, and Anavas had occupied nearly all the 
west. Hence we must assume that the Nisada-Madhava (Bhil) natives 
had become subjected and largely absorbed. 

Anavas. One of the most important non-Aryan groups in India 
was the Anava. The eastern Anava included Anga, Vanga, Kalinga, 
Pundra, and Suhma. The first three of these occupied territory farthest 
to the east, territory which we would expect most likely to be Tibeto- 
Burmic. x But Pundra looks more like a Mundic word and the country 
adjoined Mundic territory; it was probably occupied by Mundic peoples 
under a Tibeto-Burmic ruling caste. Kalinga itself was probably occupied 
by Tibeto-Burmans who drove the native Mundic speakers north from 
the seacoast before the period of the Mahabharata. Suhma and the 
variant Svaksa could be Tibeto-Burmic compounds, Su-hma, Swak-sa, 
but there is nothing particularly Tibeto-Burmic about their appearance. 2 

The western Anavas were the Yaudheyas, Ambasthas, Sibis, Sindhus, 
Sauviras, Kaikeyas, Madrakas, Vrsadarbhas, the kings of Navarastra, 
the lords of the city of Krmila, 3 and, according to some, the Bahlikas. 
The names of the western Anavas do not have the final -hga so characte- 
ristic of the names of the eastern Anavas. Hence one may doubt that 
the western Anavas spoke a Tibeto-Burmic language. 

But, although the East Himalayish dialect names usually end in 
-n, 4 none of the West Himalayish names of dialects end in -n, 5 so it 
may be due to similar reasons that the West Anava names did not end 
in -nga. Moreover, Karna said to &alya: "The regions are called by 

1 Baudhayana in his Dharmasutra (I. 1.32 33) said that the inhabitants 
of Ariga were of mixed origin, and that anyone who visited the Pundras, Varigas, 
or Kalirigas should offer a punastoma or sarvaprsthi. Devala (op. cit. in n. 6, 
p. 20 above) said anyone visiting the Arigas, Varigas, or Kalirigas should be ini- 
tiated anew. 

Risley attributed the broad, round-headed elements among Bengalis and 
Oriyas to a Mongoloid race (op. cit. in n. 5, p. 19 above). 

2 Almost any word can be Tibeto-Burmic if we assume it is a compound. 

3 Dines Chandra Sircar, "The Ancient City and District of Krmila," IHQ 17 
(1950), 138141, identified Krmila as probably the modern village of Valgudar, 
near Luckeosarai in the west of the Monghyr district. But Sircar does not state 
that this was the Western Anava city, and I doubt that the latter would be so 
far east. 

4 See p. 14 above. 

6 I have not investigated why none of these names end in -?i, whether they 
are all native names that by accident end in some other phoneme than -?i, whether 
they are native names that have been aryanized, or whether they are names 
bestowed on them by others. Whichever it might be, the same may have been 
true of the western Anava names. 
the name of the Arattas. The people residing there are called Bahllkas." 1 
We can understand that. Dakota is an American Indian name, but the 
Dakotans of today are practically all Americans without a drop of 
Indian blood. 

Hence I am tentatively giving the author of the passage in the 
Mahabharata the benefit of the doubt when he considered the western 
and eastern Anavas closely related and both Mleccha. When he remarked 
that the sons of Anu were the Mlecchas but the sons of Turvasu were 
Yavanas, he was evidently trying to make a racial-linguistic distinction. 
He did not consider the Yavanas to be Mlecchas. And when he called 
people of the Panjab and of Bengal both Mlecchas, I assume that the 
Aryans' idea of race and speech was at least fair and that he had reason 
for it, perhaps Mongoloid features, perhaps a difference of tones in 
their languages as compared with the tones of Sanskrit or Dardic. 

Pandey says that Anu occupied the northern Ganga-Yamuna doab, 
but that the Anavas were forced to the northwest in the Panjab by 
Emperor Mandhatr of Ayodhya. We have seen above that the Tibeto- 
Burmic peoples occupied at least the eastern part of the Ganges valley. 
If Anu occupied the upper Ganges in very early times, it would narrow 
the geographical gap between the eastern and the western Anavas. 

We have also noted above that the vaisyas were yellow, so enough 
Mongoloid types must have lived in the north of India at the time of the 
Aryan invasions to form a yellow varna. And, according to Ranga- 
charya, 2 Sudas annexed the territory of the Anus after the battle on the 
Parusm, so the Anavas would already be living in the Panjab at the 
time of the Rgveda. 

As far south as Mohenjo-daro one burial was found where the skull 
clearly belongs to the Mongolian type, comparable to the modern Naga 
type. This is what we should expect when Mohenjo-daro fell from power 
the gradual infiltration of western Tibeto-Burmans (western Anavas) 
down the Indus valley. 

Pargiter has noted that the Aryans were losing their grip on their 
vedic home on the upper tributaries of the Indus toward the close of the 
epic period and even before that, and the most likely non- Aryan invaders, 
we may guess, would be Tibeto-Burmic tribes now living in the hills 
north of there. 3 

1 Mbh. VIII. 45. 

2 P. 194. 

3 For the Mongolian element in northern India today, see B. S. Guha, Racial 
Elements in the Population, Oxford Pamphlets on Indian Affairs, no. 22 (Oxford 
Univ. Press, 1944), map on inside covers. 

I do not insist that the western Anavas were Tibeto-Burmans. They may have 
been Turks, Mongols, or some other people. During the last century it was customary 
for scholars to call any unclassified central Asians Turanians, now it is the custom 
of ethnologists to call them Mongoloid. I have not too much confidence in either, 
It is true that we have no outside evidence that Tibeto-Burmans 
inhabited the western part of southern Tibet at the time of the Maha- 
bharata. 

From a manuscript of the 2an-zun language from Tun-huang, 
ca. 8th century, F. W. Thomas inferred that the language was West 
Himalayish. 1 If his interpretation of scattered words proves correct, 
I would classify it more narrowly as being closest genetically to the 
Almora branch of West Himalayish. Thomas believed the speakers of 
this language occupied the Kailasa region before 644 A.D. 2 From 
Kailasa a branch of this people could have followed the present upper 
Sutlej to the Himalayas in early times. 

One bit of evidence comes closer to epic times. In the Sogdian " Ancient 
Letters", referring to the region in Tibet north of the Panjab, occurs 
the word 'nkp; which Pelliot interpreted Ang-pa 3 This certainly looks 
like a Tibeto-Burmic word, the postposition -pa occuring very frequently 
with Tibetan stems. 

Baudhayana 4 said that the Sindhus and Sauviras are of mixed 
origin and that he who visited the Sauviras should offer a punastoma 
or sarvaprsthi; and Devala said anyone who had visited the Sindhus 
or Sauviras should be initiated anew. The fact that only the Sindhu- 
Sauviras were singled out from all the western Anavas indicates that all 
but those on the western border had been aryanized by the time the 
above works were written. 

The word Mleccha was used to refer to both the eastern and western 
Anavas. Although Mleccha came in time to be used for almost any 
non-Aryan and even for Aryans of impure speech, I have the impression 
that in the Mahabharata it was most often used for people living in 
close proximity to regions now occupied by peoples speaking Tibeto- 
Burmic languages, and that it originally referred to these peoples. In 
time it came to mean something like "foreigner," but that was after most 
of the Anavas had become assimilated. 

I suspect that skr. Mleccha, referring to the indistinct speech of 
some non-Aryans, is taken from proto-Bodish (proto-Tibetan) *mlte 
"tongue," Old Bodish lte, Kukish generally *mlei, the combination 
of initial consonants (*mltd-) being simplified hi various ways in different 
Tibeto-Burmic languages. Aspiration cannot occur after / in Old Bodish, 
and the proto-Bodish form may have been *mlts f e for all we know, so 

but we can hardly expect the ancient Aryans classification to be more precise 
than that of modern scientists. And such evidence as we have seems to indicate 
that the western Anavas were more likely Tibeto-Burmans than Turks or Mongols. 

1 JRAS, 1926, p. 506. 

2 Nam, An Ancient Language of the Sino-Tibetan Borderland (London, 1948), 
p. 14 and map at end of volume. 

3 See Clna in Index. 

4 Dharmasutra I. 1. 32 33: ca. 4th century B.C., according to Bhandarkar. 
the cch of Skr. Mleccha may come nearer the primitive affricate than 
anything preserved in the Tibeto-Burmic languages. Since *mlcche 
would be an impossible combination in Sanskrit, mleccha would be as 
close as a Sanskrit speaker could come to it. 

Tibet is the one country outside of India that the Aryans knew 
something about. During the Mahabharata period the hill tribes of the 
Himalayas were continually having to be subdued. And as late as A. D. 
646 7, Chinese and Tibetan forces invaded northern India through 
Nepal. 1 But early contacts between the two lands have been almost 
entirely neglected by scholars, 2 so I shall add a few observations. 

At the beginning of Tibetan history the capital was south of Lhasa 
at Yar-kluns (Yar-lun), almost at the border of Bhutan and on the 
edge of a valley at the head of which is a pass leading toward Assam 
and the Brahmaputra valley. My suggestion of the Tamralipta being 
the ( 'red-faced' ' Tibetans is therefore not improbable for the epic 
period. 

The word for "horse" provides an interesting example of early 
contacts between Tibeto-Burmans and India. Vedic adva has been re- 
placed in most of the Indie languages by Skr. ghota or its phonetic 
equivalent, which appears in the Srautasutra of Apastamba. 

William Ridgeway concluded from the description of the coloring 
of horses in the Rgveda that they had chestnut heads and backs, shading 
off into dun on the lower parts of the body in the manner of the wild 
horse of the steppes known as Przewalski's horse, and that they were 
of Mongolian or Upper Asiatic stock. 3 And Agrawala said horses of 
partridge color came from the Uttara Kuru country. 

But in a detailed examination of the question of the taming of the 
horse, Bengt Lundholm 4 agreed with the assertion of B. T. Rumiantseff 
that there was no connection between Przewalski's horse and the Mon- 

1 For map, see Helmut Hoffmann, "Tibets Eintritt in die Universalgeschichte," 
Saeculum, Vol. 1, pt. 2, pp. 268 269. 

2 Notable exceptions are Banikanta Kakati, Assamese, Its Formation and 
Development (Gauhati, 1941), pp. 47 53, 55 56; and Suniti Kumar Chatterji, 
"Kirata-jana-krti, The Indo -Mongoloids: Their Contribution to the History and 
Culture of India," JEASB-Let. 16 (1950), no. 2. And Jules Bloch observed: "In 
India gender disappears completely from the eastern (Aryan) languages only, 
and in fact there and there only the question is undoubtedly of the action of a 
Tibeto-Burman substratum" (Pre- Aryan and Pre-Dravidian in India, Calcutta, 
1929, pp. 44-45). 

3 The Origin and Influence of the Thoroughbred Horse (Cambridge, 1905), 
pp. 150ff. 

4 "Abstammung und Domestikation des Hauspferdes," Zoologiska bidrag frdn 
Uppsala 27 (1947), 1 287; see especially section on "Die Frage der zentral- 
asiatischen Zahmung," pp. 160ff. His conclusions are based largely upon the 
work he refers to as B. T. Rumjancev, 1936, "Origin of the Domestic Horse," 
Bull. Ac. Sci. UR SS, 2-3, which I have not been able to find. 
golian domestic horse, but that the Tarpan was the ancestor of the 
Asiatic domestic horse. 

Ridgeway remarked that the modern ponies of Bhutan, Nepal, 
and Spiti are in the main Mongolian, and that the Mongolian breed 
from the Himalayas kept steadily streaming into Hindustan in the 
thirteenth century and that we do not know how long previously. 

Since Ridgeway presented no such detailed analysis as Lundholm 
and Lundholm presented no detailed analysis of the horses of India, 
no definite conclusions can be drawn here. But the two points of view 
do not seem in the main to be irreconcilable with each other or with 
the linguistic evidence. 

From the little we know about horses in early India, Ridgeway could 
be right about the northwestern horses of the Rgveda being the Prze- 
walski horse of the steppes, which was called asva, a term of Indo- 
European origin. And later the Mongolian horse was brought down 
from the north and called by its northern name. 

For in Tsangla, the language closest genetically to Tibetan and 
spoken in Bhutan just south of Tibet, the word for "horse" is kurtd. 
In Tibetan itself, only the final part of the word, rta, has been preserved, 
probably because of an accent on the original final syllable. But the 
full form is preserved in Mundic: Savara kurtd, Gadaba krutd (with 
metathesis of u and r). The reconstruction *ghurta seems preferable for 
the forms outside of Dravidian, rather than Jules Bloch's *ghutr-, 1 
not only because of the Tibeto-Burmic and Mundic forms but also 
because Wackernagel more often derives Skr. t from *liquid + t rather 
than the reverse. 

Thus the linguistic evidence indicates that the Sanskrit (not vedic) 
horse was introduced into India by way of Tibet. 

Kaikeyas. The Kaikeyas are usually reckoned as Western Anavas. 
But Jaina writers tell us that half of Kekaya was Aryan, 2 while late 
vedic literature refers to the Kaikeyas as a very ancient people. 3 These 
statements account for seeming inconsistencies in the Mahabharata, 
where the Kaikeyas fought on both sides in the Great Rebellion. The 
Kaikeyas referred to in vedic literature were Western Anava. But Sudas 
had annexed the Anava territory and Aryans had lived in the Panjab 
in vedic times and apparently some remained in Kekaya. The princes 
of Kekaya were Aryans who went over to the Pandavas, but the Kaikeya 
population was predominantly Anava and fought on the side of the 
Kauravas. 

Taurvasavas. Pargiter inferred that since Turvasu received the 
territory in the southeast he must have ruled the area around Rewa, 
having subdued the Karusas. But Pandey wrote that although Turvasu 

1 Op. cit. (p. 24, n. 2 above), pp. 46 ff. 

2 Indian Antiquary (1891), p. 375. 3 Rang. 185. 
was given the southeast, his descendants moved northwest. This agrees 
with the statement in the Mahabharata that the sons of Turvasu were 
the Yavanas, whom scholars have generally placed in the northwest. 
Pargiter must have missed this reference in the epic, for no scholar 
has suggested that the Yavanas were Karusas or had any connection with 
the latter. 

Since the Turvasas took part in the vedic battle on the Parusni they 
lived in the Pan jab or near there. The Harivamsa leads to the inference 
that the Yavanas lived in the Himalayas: King Mucukunda slept in a 
cave in the Himalayas until awakened by Kalayavana, king of the 
Yavanas and son of Gargya, priest of the king of Trigarta, and of a 
milkmaid, really an apsara. 1 And in another tale, $alva, king of the 
Soubhas, was going to fly back to his capital anyway so he was selected 
to take the message to Kalayavana. Many tribes such as the Sakas, 
Tusaras, Daradas, Tanganas, Paradas, Khasas, Pandavas, etc., who 
dwelt in the hills, went over to Kalayavana's army. 2 Hence I believe 
that previous investigators have correctly placed the Yavanas in the 
upper Panjab. 

But while we can place the Taurvasava-Yavanas fairly accurately 
on the map, we have a problem in trying to define them ethnically. It 
has generally been held that the Yavanas were Greeks, because the 
peoples of the Near East referred to the Greeks by this or very similar 
forms. But if the Turvasus were Yavanas and the Yavanas were Greeks, 
the Turvasas of the Rgveda could not have been Alexander's Greeks 
fighting the Indo- Aryans on the Parusni in the battle of the ten kings. 

There would seem to be a broken link somewhere : either the vedic 
Turvasas were not the same as the epic Turvasus, or the Yavanas were 
not the sons of Turvasu, or the Yavanas were not Greeks. A good many 
scholars have accepted the latter solution, with some justification, I 
believe. For it is no more likely that all the population that went by the 
name of Yavana was Greek than that all the population of an "Aryan" 
kingdom was in fact Aryan. Previous scholars have noted that the 
Yavanas were most often associated with Sakas and Kambojas, both 
speakers of Iranic languages, and it is probable that the Yavanas ruled 
over some subject Iranians. 3 

But one may doubt that this is the solution of our problem. Sudhakar 
Chattopadhyaya has remarked 4 that Panini formulated the rule that 
the feminine of Yavana is Yavandni, that this was commented on by 
Katyayana, and the comment paraphrased by Patafijali. The author 

1 Chap. 35. 

2 Chaps. 109-114. 

3 This rule over subject Iranians would need to be very ancient to account 
for the close association of Yadu and TurvaSa in vedic times. 

4 "The Tribal Immigrations in Achaemenid India/* IHQ 25 (1949), 272. 
postulates that Panini lived in the pre-Alexandrian period, possibly in 
the 5th century B. C. How, then could Panini formulate a rule about 
Yavana if the Yavanas only entered India with Alexander after Panini's 
death? Chattopadhyaya concluded that the grammarian must either 
have heard of the lonians from the Persians or that there was a Greek 
colony in India by Panini's time. 

I have found no evidence to indicate that the early Indians gained 
any knowledge about the peoples west of Kabul from the Persians or 
anyone else. And Chattopadhyaya himself decided in favour of his 
second alternative when he pointed out that Arrian said that the Nysians 
were not of the Indian race but were descended from men who came 
into India with Dionysos. And the next year, S. B. Chaudhuri also 
thought 1 that the Yavanas were presumably known to India before 
Alexander's time, judging by Panini's reference to the writing of the 
Yavanas, Arrian's account of Nysa between the Kophen (Kabul) River 
and the Indus pointing to a Greek colony on the frontiers of India 
before the invasion of Alexander, 2 and the numismatic evidence that 
a Greek colony existed in some outlying province of India about 550 B.C. 

The weakness of this Indian theory would seem to be that western 
scholars have generally accepted a later date for Panini than the fifth 
century B.C. I have impression that British scholars have generally 
accepted ca. 400 B.C., and German scholars ca. 350 B.C. These dates 
are sometimes given with great assurance and no authority. But where 
an authority is cited, it always seems to be Bohtlingk. The Kathasarit- 
sagara (beginning of 12th cent.) mentions Panini as a pupil of Varsa 
and Varsa as living under the rule of king Nanda, father of Chandragupta ; 
and from this BGhtlingk deduced "DaB Panini im 4. Jahrhundert gelebt 
hat, ist, wie wir spater sehen werden, nicht ganz unwahrscheinlich". 3 
Bohtlingk suggested this date with more reserves than even the above 
citation suggests. 4 But if we take the dates ca. 400 B.C. or 350 B.C., 
which have been accepted by western scholars upon so insecure a basis, 
still it is rather surprising to find Panini making rules for the formation 
of the feminine of Yavana if the Yavanas only came to India with 
Alexander in 326 B.C., for Bohtlingk himself thought Yavanas must 
have been known in India for some time for Panini to make rules about 
their name. To allow for that we should have to date the great gram- 
marian's work at least as late as 300 B.C., and we should have to assume 
that he, Katyayana, and Patanjali were hastily recording a dying 

1 "Huns, Yavanas and Kambojas," IHQ 26 (1950), p. 120, n. 23. 

2 Strabo, 15. 1. 6 9; Arrian, index Dionysus. 

3 Panini's Orammatik (Leipzig, 1887), p. viii. 

4 Albrecht Weber criticized Bohtlingk's dating but his own date of A. D. 140 
rests partly on tales and the assumption that Yavanas came to India only with 
Alexander (The History of Indian Literature [London, 1882], p. 217). 

28 Aryan Attempts at Ethnic Classification 

language, since Asoka erected his pillar and rock edicts about 250 B.C. 
in the local prakrits. If we cannot definitely date Panini's life at present, 
we have at least seen that if we do not accept the early dating of Indian 
scholars, we still have difficulties with ' 'accepted'' western dating and 
problems if we date the grammarian after Alexander. 

Since we know almost nothing of the ancient history of Central 
Asia, we cannot superciliously dismiss evidence presented by both 
Sanskrit (vedic, epic, Panini?) and Greek sources that Greeks were in 
India before the time of Alexander. We shall consider the matter further 
below. 1 

Since Skr. Yavana corresponds best with the hypothetical Homeric 
form *'lafoves, the later Sanskrit Yauna with the earliest actual Greek 
form 'laoves, and the Prakrit Yona with classical Greek "loves, one might 
deduce that Sanskrit must have borrowed Yavana, at least by the time 
of Homer long before Alexander brought Greeks to India. 

Perhaps that is true. But I incline to believe that the evidence from 
the names is entirely inconclusive. Since the name had a rather wide 
extension in southern Asia, the question is one of some importance not 
only in India but in the Near East as well, and it seems worth while 
to digress long enough to review it briefly. 

The old Sanskrit pronunciation of Yavana was Yawana or Yauana, 2 
which corresponds very well with Hebrew Ydwdn, Assyrian Idmanu 3 
(pronounced lauanu) 4 as the name of the Greeks. The assumption 
has been that the Asiatics applied the name of one branch of the Greeks 
the lonians to all the Greeks. 

The oldest Greek forms were *Iaove$ for the people (Iliad), 'laovin 
for the territory. 5 Since the digamma had disappeared in Ionic by 
the eighth century, the reconstructions * y lafoves; and **Idfovia have been 
generally accepted. As these would be phonetically idwones and iawonia, 
the correspondence with Hebrew Yawan, Skr. Yavana (i.e., Yawana) is 
nearly perfect except for the o in Greek, which has not been explained. 

However, we shall see below 6 that the Greeks transcribed Skr. va 
as o, and I have noted elsewhere that they transcribed Lycian wa as o 
and Lycian awa as w 7 The indications are, then, that Greek o and o> 
were not simple vowels but a diphthong and triphthong respectively: 
wa or 09, and awa or aoe. Thus 'Idoves was actually pronounced idwanes 

1 Pp. 45 f. 

2 Whitney, Skr. Gram., 2d ed., sec. 57. 

3 la-a-ma-nu. 

4 m after vowel > u in Babylonian; Brockelmann, Semitische Sprachwissen- 
schafi, p. 67. 

' 5 Solon by Ar. Pol. Ath. 5. 

6 P. 115. 

7 "Greek Transcription of Lukian," Jahrbuch f. kleinasiatische Forschung 2 
(1951), 5. 
or iaoenes'lowei; was actually pronounced iawanes or iaovnes, almost 
the only difference being the position of the accent. That the last vowel 
of the stem was a schwa best explains the Assyrian lamnaia 1 (pro- 
nounced iaunaia), Old Persian Yauna, Elam. (i)-ia-u-na, compared with 
Skr. Yavana, Hebrew Ydwdn. The Hebrews and Indians heard the schwa 
as d; the Persians and Elamites did not "hear" it. 

The Sanskrit form Yauna would then probably be a loan word from 
Persian, Prakrit Yona by regular phonetic shift. 

Thus the evidence from the name Yavana in old Sanskrit is neither 
favourable nor unfavourable to Greeks being in India before Alexander. 
The Greeks changed the spelling of the name for lonians, but not the 
pronunciation from the time of Homer to Alexander, and hence the 
Indians would have had the same transcription of the word Yavana. 

Drauhyavas. The Druhyus founded Gandhara in the northwest 
where the native population was Dardic, and one might expect the 
Drauhyavas also to be Dardic. But the Mahabharata states that they 
were Bhojas, by which one may infer that the author was trying to 
imply that by origin they were Iranic. Since Dardic is linguistically 
between Indie and Iranic, it would require a better linguist than could 
be found among the early Aryans to make such a fine distinction. To 
them Dards were probably considered so close to Iranians that the 
nearest thing to a synonym they could think of was Bhoja (Yadava, 
Haihaya). 

Pauravas. We have not yet discussed who the Pauravas were. One 
can only say that they were not Aryans. The Purus w r ere fighting the 
Aryans at Parusni and they were described in the Rgveda as mrdhravdcah 
"of hostile speech," otherwise applied only to Dasyus. 2 The Vayupurana 
MS e, held the Pauravas to be foreigners. 3 In the royal Paurava line 
were names such as Dhundhu, Kuru, and Jahnu for which no certain 
derivation has been found and others which seem out of line with the 
usual Sanskrit names of persons. 

The Bharatas were also against Sudas, and he humiliated them along 
with the Turvasas and Druhyus. The Harivamsa regarded the Bharatas 
as Talajanghas (i.e., Haihayas) along with Vitihotras, Avantis, Bhojas, 
and Tundikeras. 

There were many generations between Puru and Ajamidha, whose 
four sons founded royal lines in Hastinapura, Northern Paiicala, Southern 
Paiicala, and Kanyakubja. And it was some generations later before 
Kuru was king at Hastinapura, and then 16 more down to Dhrtarastra 
and Pandu. This gave a good deal of time for aryanization. Both the 
Kauravas and Panda vas claimed descent from Puru, and the Panda vas 
were almost certainly Aryans, probably the latest and fairest to enter 

la-am-na-a-a. 2 MRK, p. 26. 3 Pargiter, Kali Age, pp. 2 3. 

India, one may infer from their name. 1 But an upstart monarchy 
must attach its genealogy where it commands respect, and the Kurus 
had been the greatest power in India before the Great Rebellion, so it 
was logical for the Pandavas to claim descent from them. 2 

The Kurus are generally accepted as Aryans, and scholars went to 
the Kuru-Paiicala country to study the purest speech. 3 Let us assume 
that the upper varnas of the Kurus were Aryan. But how about the 
greater part of the population? I have heard that the best English in 
the world is spoken at Trinity College, Dublin, but one cannot say that 
the Irish as a whole speak the purest English. 

A strong reason for doubting that the bulk of the population of the 
Kaurava nation was Aryan is that the Kauravas rallied the support of 
nearly all the non- Aryans, the Pandavas the support of nearly all the 
Aryans. 

After the battle of the Mahabharata the Pancalas seem to have 
annexed the honored name of the Kurus. 4 The name Kuru and Kaurava 
seems to have soon disappeared from the political map of India. What 
became of the Kuru people ? If one looks at a modern linguistic map 
one may note three small patches labeled Kur, which one may compare 
with ancient Kuru, and the larger patch labeled Kur-ku\ also Korwa 
which we may compare with ancient Kaurava.^ 

These are Mundic languages. Edward Tuite Dalton wrote 6 that, 
although the various divisions of the Kaurs or Kauravas were much 

1 After the Aryans had won the war, they revised the Kuru epic to make 
themselves the heroes. So certainly they were not glorifying foreigners above all 
others when they praised the Pandavas as their leaders. It is true the epic does 
make a foreigner, or semi-foreigner, Krsna, into a god; but that, according to 
Hopkins, occurred in a still later revision (see p. 4, n. 2, par. 2). And one may 
suggest that Krsna's deification took place after enough of the native population 
had been accepted by the Aryans and had attained a position in society strong 
enough to force acceptance of one of its heroes as a god. 

Some scholars have been misled by the Pandavas practice of polyandry to 
the inference that they were not Aryans. Anthropologists are continually warning 
that we must distinguish between race, language, and culture and then are 
constantly using one to back up their assertions about the other. Their advice 
is sound, however, if they seldom practice it. Professor Kirfel has pointed out 
to me that polyandry is found in Tibet and in the Nilgiris, and probably no scholar 
will contend that the inhabitants of the Nilgiris are Tibetans. 

2 As Rangacharya (p. 198) remarked of the Bharatas "traditions came to 
be invented by the chroniclers of this and later epochs to prove that the Bharatas 
were the progenitors of all the important dynasties of India". 

3 Vedic Index, II, p. 279. 4 See Geography, entry no. 1. 

6 See Constable's Hand Atlas of India, ed. John George Bartholomew (West- 
minster, 1893), plate 10. Korwa is a dialect of Kherwari spoken in Chota Nagpur 
and South Midnapur. Kurku is spoken in the Satpura and Mahadeo Hills. One 
may also note that a vagrant tribe of Madras is called both Kurru and Korava. 

6 Descriptive Ethnology of Bengal (Calcutta, 1872), p. 136. 
scattered and had little contact with each other, they all tenaciously 
clung to the tradition that they are descended from the Kauravas of the 
Mahabharata who had taken refuge in the hill country of Central India, 
and that this was firmly believed by all castes of their Hindu neighbors, 
and that the latter regarded them as their brethren. 1 Dalton thought 
that only the upper castes of the Kauravas were Aryan. From the line-up 
of the non- Aryans on the side of the Kurus I had concluded that the 
upper Kuru varnas were more likely aryanized natives. At any rate 
the ksatriya varna of the Kurus had been almost exterminated, leaving 
only the base of the population. What was the race or language of this 
base? A(lfred) Ludwig first interpreted kuru as "red," 2 later as 
"brown." 3 And one may note that two of Sudyumna's sons ruled over 
Utkala and Gaya, the first Mundic down to the present time and the 
second Mundic in early times, while the third son ruled over the east, 
including the Northern Kurus. 4 This would lead to the inference that 
the Northern Kurus were also Mundic. 

The Aryan genealogies are the most confused and imperfect of any. 
Since the Aryans were an uncultured people without writing and since 
the system of writing they found in India was not the type easy to 
learn, some of the people of India may have had a literature in which 
they recorded their genealogies, while the Aryans had to transmit their 
traditions orally and this was more faulty. When the Aryans had developed 
as far as the period of the Great Epic we can understand the anxiety 
of their rulers to attach their family tree to one of a family of higher 
culture and greater antiquity. By that time they were taking over the 
literature of their once more civilized neighbors. The Panda vas attached 
their folklore to that of the Kauravas, and at the same time attached 
their genealogy to that of their former enemies. 5 Modern archaeological 
work has just begun in India, and we should not assume that blanks 
in our knowledge were blanks in fact. Some of the gaps in our knowledge 
of the development of writing in India between the Indus civilization 

1 Dalton hit close to the truth when he suggested that the war of the Pandavas 
and Kauravas was not a family quarrel but a struggle for supremacy between 
an Aryan and a "Turanian" nation, bvit he was ahead of his time and his suggestion 
went unheeded. 

2 "uber die mythische Grundlago des Mahabharata," Vestnik Kralovske 
Ceske Spoletnosti Nduk (Sitzungsber. kgl. bohmischen Qesellschaft der Wiss.), 1895, 
no. 9, p. 4. 

3 In a review of Das Mahabharata als Epos und als Rechtsbuch ... by Joseph 
Dahlmann, in ibid., 1896, no. 5, p. 104. 

4 Pusalker, p. 273. 

5 If the early Kuru epic was in a non- Aryan language, Vyasa was particularly 
capable of translating it ; for he was the son of a brahman-rsi and a Dasa princess 
and he had married the Kuru queens of his step-brother, and was no doubt polyglot 
(Suniti Kumar Chatterji, "Krishna Dvaipayana Vyasa and Krishna Vasudeva," 
JRASB-Let. 16 [1950], 80). 
and Asoka will no doubt be filled and we shall find that some of the 
people of India had written records. l 

Although I am very sceptical of the Pauravas and their supposed 
descendants, the Bharatas and Kauravas, being Aryans, I think their 
upper classes were probably aryanized, and on the ethnic-linguistic map 
I have given these people the benefit of the doubt and regarded them 
as Aryan. 

Arattas. Scholars have generally assumed that Dasas and Dasyus 
referred to natives in general. This is erroneous. The Dasyus are referred 
to contemptuously. They are slaughtered or thrown down; they are 
made servile. They were identified above with the Bhils. But, while 
the Dasas are enemies who may be hated, dispersed, overcome, or slain, 
they are never held in contempt. Hence the vedic hymns frequently 
refer to Indra overcoming foes, Dasa and / or Arya. 2 The Dasas and 
Aryas were on a social level, above the Dasyu-Bhils. 

Sudas destroyed all strongholds and seven castles after the battle 
on the Parusm (Ravi). 3 And while some of the people of the Panjab 
were primitive people fighting with clubs, we know the Dasas had both 
forts and wealth. 4 We may now connect the ancient Harappa civili- 

1 Professor Kirfel suggested to me that remains of the Kurus may bo found 
in the present people of Kulu, the r of Kuru having been changed to /. This is 
a brilliant suggestion; for Kulu is just across the Sutlej from where I placed 
Kuru-Jarigala on Map 1. Professor Kirfel remarked "Der Wandel von r in I ist 
weit verbreitet. Nach einem Aufsatz von Karl Ammer ... findot sich das I schon 
im Rgveda hi vielen Worten, wo man eigentlich ein r erwarten wiirde. Wenn ich 
nicht irre, findet sich in den A6oka-Inschriften das I uberall da, wo man ein r 
erwarten wiirde. In sprachlicher Hinsicht wiirde also wohl keine Schwierigkeit 
vorliegen." 

Professor Kirfel referred to the works of Karl Eugen von Ujfalvy, Aus dem 
westlichen Himalaya. Erlebnisse und Forschungen (Leipzig, 1884), 34ff., and Karl 
Ammer, "Die L-Formen im Rgveda," WZKM (1948). Unfortunately neither of 
these are available to me, but the suggestion should certainly be further investi- 
gated. 

2 See Muir, Vol. 2, p. 361. 3 Rgv. VII. 18. 13. 
4 See Vedic Index, I, pp. 356, 358. 

The only reference I have seen as to whom the Indo- Aryans thought of as 
Dasas is by Manu, who is so late and so formal that 1 do not believe he can be 
generally accepted as more than suggestive regarding actual conditions during 
the vedic period. He said that the sons of a sudra (i.e., black) father and vaisya 
(i.e., yellow) mother was an Ayogava (x. 12), and that the son of a Nisada (black) 
father and an Ayogava mother was a Margava or Dasa. About all we can conclude 
from this is that, in Manu's time, a good many centuries after the term arose 
with the Aryan invasion of India, the people called Dasas were more black than 
yellow. 

But it is interesting to compare Manu's other term for this people, Margava, 
with the Iranian name *Maryu, Margava "Merv." And the Brhatsarhhita has a 
Margara, which may be an error for Margava, in the south-west (i.e., northwest). 

Phonetically corresponding to Skr. Ddsa, the Iranians had the word Dahae, 
zation on the Ravi with the vedic Dasas and the epic Arattas. For 
Kama stated that the regions were called by the names of the Arattas 
but that the people residing there were called Bahlikas. One might infer 
from this that the Arattas were extinct, and that the Bahlikas were 
occupying their territory, but that was not true. Rangacharya 1 has 
summed up the evidence on the Arattas, noting that Baudhayana in his 
Srautasutra 2 "clearly says that the lands of the Arattas and Gandharis 
in the Panjab and further west were not suitable for orthodox Aryans, 
and that those who went there should perform the catustoma. In his 
Karmasutras he is even plainer. He says they were of mixed blood . . . 
The Mahabharata says that none should spend more than two days in 
the land of the Arattas." 

The report of the biggest find of Harappa burials has not yet been 
published, so that it would be premature to attempt to define the ethnic 
composition of the Arattas here. On other grounds the writer had conclud- 
ed that the Bahlikas were a mixture of Manava and Anava. One may 
now tentatively interpret this as the partial absorption of the Arattas, 
centered around Harappa, by the western Anavas coming down from 
the north. 

But, some reader may object, you have made nearly all the nations 
of India non- Aryan. 

Well, I believe that is as it should be. One cannot expect some tens 
of thousands of Aryan invaders to have subjugated and aryanized a 
land of perhaps some tens of millions in a few centuries. But in India, 
where classes were divided, we must look at things somewhat differently. 
The Aryans were at the time of the Mahabharata still the invaders and 
conquerers. 

When the Mongols conquered India and China, the people still 
remained Hindu and Chinese, and the situation of the Aryans in India 
at the time of the Mahabharata was not very dissimilar to that of the 
later Mongols. When I state that a certain people was Yadava, hence 
Bhil-Iranian, I am referring to the greater part of the population; the 
the name of a people living in Alexander's time on the southwest shore of the 
Caspian Sea (see Davies, map 3). But Tacitus wrote ad flumen Sinden, quod Dahas 
Ariosque disterminat (Ann. XI. 10), and Paul Kretschmer thought Tacitus was 
referring to the Sindhu as the boundary between the Aryas and the Dahae (KZ 55 
[1927], pp. 102 103). If so, the question arises who was on which side of the Indus 
and when. 

In the Avesta, dahaka refers to impious persons and Azi Dahaka was an impious 
dragon (A. Christensen, Danske Vidensk. Sels. Medd. 27, no. 1, pp. 24f., 20ff.). 
They seem, therefore, not to have followed the Iranian religion. 

1 P. 197. 

2 XIII. 13. At I. 1. 33 he said that anyone who has visited the Arattas shall 
offer a punastoma or sarvaprsthi. 
ruling classes may have been aryanized natives, of mixed race, or pure 
Aryan. * And in designating a kingdom as Anava I am suggesting that 
the bulk of the population was Mleccha (Tibeto-Burman) although the 
ksatriyas and brahmans may have been of Aryan descent or aryanized 
natives. For I believe the early genealogies were trying to express what 
the people were, not the ruling classes. 

When the Aryans conquered a native people, it did not necessarily 
mean that they followed an Aryan policy thereafter. When the Normans 
conquered England, they continued to talk French but their policies 
were no longer French but English and often anti-French. And so if 
the Yadavas and Anavas had Aryan rulers, they usually did not line up 
with other Aryan rulers in the Great Rebellion but united against them. 

1 In his study of "Vedic, Sanskrit, and Prakrit" (JAGS 32 [1912], 414-428), 
Walter Petersen concluded that Sanskrit was a caste language from the beginning 
meaning, no doubt, from the earliest record of it.