Revision as of 13:46, 17 January 2016

The Aryans And Their Migration To India

So much weight has been given to the Aryans in the past that we shall couterbalance matters here by mentioning them only briefly. The general opinion is that the Aryans of the vedas lived in northwestern India. But at the time of the Mahabharata the centers of Aryan power were along the Sarasvati, the Yamuna, and the upper Ganges.

The general theory has been that the Indo-Iranians spread eastward, the Iranians stopping at the Indus while the Indo-Aryans went on and entered India by way of the Kabul River valley, overran the Panjab in vedic times, spread farther east in epic times, and during historical times pushed eastwards even into Assam. This theory made it a nice orderly west-to-east spread.

There was much to be said for such a theory. During historical times, invasions of India have been through the northwest, and from there extended eastward. The homeland of the Indo-Europeans has been placed somewhere in the great plains of Eurasia, and one would expect migrations to radiate from this center. Logical as this theory seems to be, it has certain defects.

First, although Sanskrit and Avestic are so closely related genetically that none doubts they were once one people, scholars have never succeeded in agreeing on their homeland more closely than that it was somewhere west or northwest of India.

Although the Iranians and Indo-Aryans have a number of common religious beliefs and gods, no common geographical knowledge has been adduced.

None of the vedic hymns treats at length of the passage of the Indus, which would be a necessary feat for the invasion of India from the west.

A glance at the ethnic-linguistic map will show that the Indo-Aryans at the period of the Mahabharata knew almost nothing of the peoples west of a narrow fringe on the further bank of the Indus River. And a survey of the geographical knowledge of the authors of the Rgveda shows that they too knew nothing beyond the tributaries of the Indus.

But the ethnic-linguistic map shows that the Indo-Aryans knew considerable about the peoples along the upper Indus valley in present Tibet.

The geographical locations outside of India that are most often connected with Indian religion and traditions are almost entirely in or north of the Himalayas, not west of the Indus.

Muir has noted references in the Rgveda indicating recollection of a colder country, expressions of "a hundred winters" (data(m) himdh), 1 as if winter were something to be dreaded.

These considerations had begun to arouse my scepticism regarding the accepted theory of the eastward migration, which I too had accepted for so many years without question. But it was while seeking information regarding the early inhabitants of Afghanistan that the solution became apparent.

The earliest account the Avesta has given us of geography is believed to be in the Vendidad, 2 Fargard I. Below I give a list of * 'countries" 3 with the pertinent information accompanying them in the Vendidad, and the interpretations of scholars: 4

1.Airyandm vaejah "Vaejah of the Aryas" (Vaejah, being the name of the country). Ten months winter, two of summer. River Daitya.

2.Gava in which live the Suyda (O. P. Suguda, Gr. Zoydiavtj, "Sog- dians").

3.*Maryu- 9 Mouru- (O. P. Margav-, Gr. MaQyiavr\, "Merv.").

4. Baxdl- (0. P. Bdxtri-, 5 Gr. /Sa^r^ta, M. Iran. Bdxl, Pers. Balx, "Baktria, Balkh.").

5. Nisaya-, between Mouru and Bax<5l. (Ntoaia, capital of Parthia; Christensen: the Nisaya of E. Iran mentioned by Ptolemy, "le Maimene moderne".)

6. Haroyu- (0. P. Haraiva-, "Herat"; Christensen: "Aqia or Mga'a 6

7. Vaekvrzta-, where Duzaka or Duhaka live. ("Kabul", Spiegel; "Sistan", Bournouf, Lassen, Haug; "Gandliara" Christensen; "Bargada" in Paropamisos, Darmesteter; Oichardes River crossing Scythia, S. Levi.)

8.Urvd-, with many pastures. ("Kabul," Haug, Lassen; between Hyrcania and Merv, Nyberg; between rivers Kurum and Gomal flowing in Indus, Geiger ; Urvada River flowing in Lake of Hamun in Sistan, Aurel Stein, Markwart, Herzfeld; about city of Ghazna on R. Ghazni, Christensen.)

1 Vol. 2, p. 323.

2 Or Videvdat, Av. vidaevo datem.

3 "Created," except for the first, by Ahura Mazdah.

4 The line following no. 5 is my own, placed there to facilitate the following discussion.

The latest interpretations I have seen of the following names are by Arthur Christensen, "Le premier chapitre du Vendidad et 1'histoire primitive des tribus iraniennes," Danske Videnskabernes Selskab, Hist.-filol. Meddelelser 29, no. 4 (1943).

6 Non-Persian form found in Old Persian; *BdxOri- would be the form expected for Old Persian (Benveniste, Bulletin de la Societe de Linguistique 47, fasc. 1 (1951), p. 22.

6 A fertile country watered by the Arios (Harirud) and Margos (Marvrud).

9. Xn9nta-, where live Vvhrkana- (O. P. Varkdna, Gr. 'YQxavia, "Hyr- cania"; "Gurgan", Spiegel; "Kandahar," Haug.)

10. Harax v aiti-, the beautiful. (O. P. Hara h uvatl- y Gr. Idga^coctfa, "Ara- chosia.")

11. Haetumant-, the brilliant, the glorious. (Gr. 'ETvpavdQos, Afg. Helmand, "Helmand" River.)

12. Raya-, with three forts (tribes). (O. P. Ragay-, Rayay-, Rajay-, Gr.

• Payiavrn, modern Rai in Media near Tehran.)

13. Caxra-, the strong, the faithful, (brigands, Nyberg; "Mazenderan" province, Christensen.)

14. Varzna-, with four ears, where Oraeta(o)na was born, who killed the dragon Dahdka. ("Guilan," Haug, Christensen; region of laxartes, Nyberg.)

15. Hapta-hdndu- (O. P. Hindav-, the "Indus.") 1 .

16. People at the sources of the Rayhd. ("laxartes," Geiger, Nyberg; "Volga," Markwart; probably "Volga," Christensen.)

We shall leave no. 1 for discussion below. For the other places above the line, i.e., nos. 2 to 5, one may note general agreement among scholars on the identifications and very close phonetic agreement.

But for those below the line there is very general disagreement of scholars on the identifications, and phonetic similarity only in four, nos. 10 12, 15, and the last of these is not in Iran but in India.

I referred to the names above as those of ' 'countries," but Av. SoiOra- really means the territory occupied by a tribe. Now we may note a very peculiar thing in entries 6 to 16, on w^hich Iranists have so generally disagreed. Four of the names of "countries" agree with the Sanskrit names of rivers (Skr. s Av. h, qh) : 6. Haroyu-, O. P. Haraiva-, Iran. Harayu, Skr. Sarayu.

10. Harax v aitl-, O. P. Hara^uvatl- , Sarasvatl.

15. Hapta-hzndu-, O. P. Hindav-, Skr. Sapta-sindhu.

16. Rayhd, Skr. Rasa.

Scholars have already noted all these correspondences and they have remarked on the similarity of Iranian and Indo-Aryan place names. But one scholar seems to have observed one correspondence and another scholar another correspondence, and no one appears to have noticed that they are all names of rivers and that they are named going from east to west in India.

We may now note another peculiar thing about the Vendidad list: the number of names between nos. 6 and 10 corresponds exactly to the number of prominent rivers in India between the Sarayu and the Saras- vati; and between 10 and 15 to the number of big rivers between the

1 For the vowel of the Old Persian form, see Benveniste, op. cit. (p. 36, n. 5, above), p. 23.

Sarasvatl and the Indus. Thus, beginning with the Sarayu and going from east to west in India, we find the following correspondence:

6. Haroyu- y Skr. Sarayu

1. Vaekzwta-, Skr. Oangd

8. Urvd-, Skr. Yamuna

9. Xndnta-, Skr. Drsadvatl

10. Harax v aiti~, Skr. Sarasvatl

11. Haetumant-, Vedic $wraZn, Skr. SatadrA

12. Raya-, Vedic Parusni, Skr.Irdvatl

13. Caxra-, Vedic Asikni, Skr.Candrabhdgd

14. Varzna-, Skr. Vitastd

16. Hapta-hvndu-, Skr. Sapta-sindhu 16. Rayhd-, Skr. .Rasa

The Rasa River has not been definitively identified. It is probably a western or northern tributary of the Indus.

From the above list it seems obvious that the Iranians knew not only the Indus but all the great rivers of northern India from the Sarayu west to the Indus and the Rasa. 1

The theory of the eastward migration of the Indo-Aryans has blinded some previous investigators to the most ancient evidence left by both the Iranians and the Indo-Aryans. The Vendidad and the famous river hymn of the Rgveda both name the rivers from east to west. By the theory of eastward migration, this does not make sense. For according to that theory, the Iranians stopped when they came to the Indus and the vedic Aryans when they came to the Panjab. For every people, where they are is the center of the universe and from there they radiate out. If you were living in Calcutta and wished to tell your child about the great rivers of the world would you begin with the Amur or the Mississippi and finally get to the Ganges? Of course not. The only way in which the order of the rivers in the Vendidad and the river hymn of the Rgveda makes sense is as history, the rivers the Indo-Iranians passed in their westward migration.

The fact that the Iranians gave all but four of the rivers of northern India different names from those of their Indo-Aryan relatives is not surprising. For the much more closely related vedic Aryans and Sanskrit- speaking Aryans give entirely different names to two of the rivers of the Panjab (nos. 12,13), while the names of a third (no. 11) agree only on the consonants, not on the vowels.

The surprising thing is not that the Iranians and Indo-Aryans did not agree on the names of all the rivers, but that they agreed on some. And the question arises why they had the same names for the Sarayu,

1 Two other "countries," nos. 9 and 11, are specifically referred to as rivers (rud) in Iranian (Pahlevi, etc.) texts or commentaries. Sarasvati, Sindhu, and Rasa except for phonetic shifts that occurred in Iranian after the two people separated.

We can only conclude that the Indo-Iranians knew all these rivers before they separated. Where could they have lived where they would know and name these rivers, and yet not know the other rivers of northern India to which they each gave different names after their separation?

Since the list of rivers began with the Sarayu, this gives us a hint where to look. The headwaters of the Sarayu and of the Indus come within about 100 miles of each other a little south of Meru (Kailasa) about lake Manasa-sarovara. Both the mountain and the lake are famous in Indian tradition, and if the homeland of the Indo-Iranians was in this area, they would of course have had common names for the two rivers. But how did the Indo-Iranians come to have a common name for the Sarasvati? Only one other great river flowing south or west into India has its origin in the above area, the present Sutlej. During the time the Indo-Iranians lived about Meru, did they refer to the upper Sutlej as the Sarasvati? And, if so, why did they give the name to another river in India? Several explanations seem possible.

1. The upper Sutlej once flowed into the Sarasvati. Upon other grounds, Pusalker suggested that the Sarasvati was as large in the vedic age as the present Sutlej. 1 Such a great Sarasvati would explain why it commanded so much respect in ancient times, why a change in course of the upper part of the river would diminish the flow so that it dis- appeared in the desert, how some of the Kurus moved south from the Uttara Kuru country to lake Manasa-sarovara and then followed the Sarasvati into India and settled upon its banks.

What was the course of this hypothetical great Sarasvati? After it issued from the Himalayas it was deflected southeastwards by the parallel Siwalik Hills until it reached the plains, whence it flowed south- westward through the present Saraswati-Ghaggar channels. Its course along the valley between the Himalayas and the Siwalik would have been approximately 1000 feet higher and considerably longer than that of the present Sutlej. But the Sutlej, having its source in the south slope of the Siwalik Hills, cut through that barrier and captured the "upper Sarasvatl."

That is a possible theory, in the opinion of Prof. John B. Leighly, of the department of Geography of the University of California, although he had only a rough commercial map for reference ; and of Arch C. Gerlach, chief of the Map Division, Library of Congress, after referring to two Survey of India series of maps at the scales of 1 : 63,360 and 1 : 126,720. Near Kalka, 30 48' N, 76 53' E, between the Siwalik Hills and the Himalayas, the Sirsa Nadi, a tributary of the Sutlej, comes very close to the Jhajra Nadi, a tributary of the Ghaggar, Mr. Gerlach wrote me. According to the above theory, these would be streams formed in the valley of the old Sarsavati after its upper course had been deflected into the Satadru (Sutlej).

Both Leighly and Gerlach thought a more likely place for the capture of the "upper Sarasvati" by the Satadru would be on the plains after the river emerges from the mountains, as 30 N, 74 E, where, as Gerlach wrote, "the plain is so flat and the streams so braided that the channels could, and apparently did, change many times. The original topography of the region is obscured by irrigation canals and ditches."

The dry bed of the Nyewal (or Naiwai) l lies midway between the Sutlej and Ghaggar where the latter approach within 60 miles of each other in very flat country. Two branches of the Nyewal reduce the greatest distance between stream beds at this place to 15 miles.

Gerlach concluded that "although this is not conclusive evidence that the Sutlej ever flowed into the bed of the Ghaggar, the possibility exists that it did." 2

So many questions of topography, geology, ichthyology, and geo- graphy are involved that I must leave the testing of the hypothesis to the scientists of India who have access to the field.

2. The Sutlej and Sarasvati were never joined. The Indo-Iranians called the upper Sutlej the Sarasvati when they lived in Tibet. A band of Aryas started to descend the Sutlej, which they called the Sarasvati, into India. But, perhaps because of steep canyons or floods, they took a short cut south over the mountains. On reaching the plains of India they came upon a large stream which they thought was the same one they had left in the mountains, and they called it the Sarasvati.

3.The Indo-Iranians knew the course of the upper Sutlej, which they called the Sarasvati, as far as the Himalayas when they lived in Tibet. When they invaded India, they descended the Sarayu and started their westward migration. They knew enough about astronomy to know that a certain stream they came upon in the plains of India was flowing south just below the most westerly point they knew on the course of the river they called Sarasvati when they lived in Tibet. They thought it was the same river and gave it the same name.

1 Naimal on the survey map available to the author.

2 I may not have been the first to suggest that the upper Sutlej once flowed into the Sarasvati. N. M. Billimora, "The Great Indian Desert . . .," Journ. Sind Hist. Soc. 8 (1947), p. 86, wrote as if it were a well-known theory be did not need to elucidate: "Formerly the Indus was deflected by the Rohri Hills directly into the Rann of Cutch, where it was joined by the river which was supposed to have formed a continuation of the Sutlej and Sarasvati through the now dried-up Hakra

(Wahind) canal" (The italics are mine.)

All these alternatives suppose that the Indo-Iranians called the upper Sutlej the Sarasvati. The first surmises that they called the Saras- vati in India by that name because it really was the same stream they knew in Tibet (the Sutlej -Sarasvati). The second and third alternatives suggest that the naming of the Indian stream was a case of mistaken identity.

We have not yet considered no. 1 in the list of "countries" of the Vendidad, the "Vaejah of the Aryas," where there were ten months of winter and two of summer. I am not a student of climates, but I would suppose that could refer to a homeland in the Hindu Kush, in the Kara- koram range, the Himalayas, or Meru. Only the Himalayas or Meru meet the requirements of being where the Indo-Iranians would know only the Sarayu, Sarasvati, and Sindhu, which one of the hymns of the Rgveda mentions as the "great rivers" a hymn that must have been composed when the Aryans still lived in Tibet. 1

The Daitya River of the "Vaejah of the Aryas" can only refer to the Skr. Lohitya (Brahmaputra), since it is the only other great river having its source near Lake Manasa-sarovara that is not otherwise accounted for. The Daityas of the Indo- Aryans were therefore originally those who lived on the Daitya river. According to the Indo- Aryans, Daityas lived on Sveta mountain which was in the east 2 .

From the homeland of the Aryas within the Meru (Kailasa)-Lake Manasa-sarovara region, what was the route of migration?

The Vendidad's first five countries would indicate that some Iranians migrated down the upper Indus to what was later known as Iran to Sogdiana, Merv, Bactria, and Nisaia. There does not seem to be much order here, and perhaps there never was. But it is possible that these countries did not occupy the same geographical position when the Vendidad was composed that they did in later Persian and Greek times. The Vendidad does not speak of the country of Sogdia but of Gava where Sogdians lived. We find Sogdians on the upper Indus in the Sogdian Ancient Letters and on down into Nestorian times.

We may guess that the Dardic branch of Aryas accompanied this migration and remained behind in the northwest of India. Possibly some Sanskrit-speaking Aryans also went down the upper Indus and settled in the upper Panjab, but I know of no direct evidence of it. Although the Indo-Aryans had a better knowledge of the upper Indus than of any other area outside India, they may have gained it by trade.

But there was another migration than that down the upper Indus. In the above list from the Vendidad, in no. 6, Hardyu is followed by viS hardzana which Bartholomae translated "wo die Hauser verlassen werden," but the passage has not been understood. It is clear now that in nos. 6 16 the author of the Vendidad is describing the second migration of Iranians from Tibet down the Sarayu into India and then west across the many rivers. And at the beginning of this migration, in no. 6, the Iranians have reached the Sarayu, "leaving their homes behind/'

The small bits of descriptive matter which accompany the names of the other "countries" from 6 to 16 could apply to many places and are, therefore, often not particularly significant. One can only show that some, at least, are not inconsistent with the interpretation given above.

8.The Urva- is described as having many pastures. I would surmise from Davies' rainfall map (p. 81) that the Yamuna valley would provide good pastures. There is also mention of evil invaders, who could have been the Kuru on the western bank.

9.On the Xnonta-, which I have identified with the Drsadvati, lived the Vohrkanas. Avesta whrka means "wolf", and Christensen suggested that the name of the tribe was totemic. Before the writer considered the geographical names of the Vendidad, he had placed the Yadava tribes Andhaka and Kaukura on Map 1 just below the Drsadvati and had suggested in the index that Kukura "dog" was totemic. I would here identify the Kaukuras and the Avesta Vohrkanas on geo- graphical and semantic grounds.

10.The Harax v aiti- is described as "the beautiful," which is the way the Indo-Aryans usually described the Sarasvati.

11.The Haetumant- is described as "the brilliant, the glorious," which would certainly apply as well, at least, to the Satadru as to the I would guess sluggish Helmand.

Here we may have partial phonetic correspondence of the names. Hans Krahe analyzes -mant- as the common Indo-Iranian suffix. 1 Sata-dru may be analyzed as "(flowing in) a hundred branches," but the vedic Sutudri is of unknown etymology. It is probable, however, that the vedic and later Sanskrit forms are compounds or contain a

suffix of some sort. The Iranian root would be Haetu- <*Setu-, correspond- ing fairly well phonetically with the first part of the vedic Sutu-, epic Sata-. A pre-Indo-Iranic name *Satu-> but with the sibilant initial between s and ^, and a as in at y may have been the native form.

12. The Raya- is said to have three forts or three tribes. Harappa was a fortified city beside what was once the bed of the Ravi before it changed its course, 2 and nearby is another site 3 which has apparent- ly not been excavated. If the meaning is three tribes, a glance at Map 1 will show that this could be an understatement by epic times.

1 Beitrdge zur Namenforschung (1951 1952), p. 161.

2 Mackay, Early Indus Civilization, p. 1.

3 Piggott, p. 137.

13. Parasara has Cakra, a people in the center, and the center in tortoise geography is sometimes very far west.

14. Reference is made to non- Aryan invaders. Here along the Vitasta is the home of the Kaikeyas and the Sindhu-Sauviras, the Anava tribes which we saw above were Mongoloid and probably Sino-Tibetan.

16. The Hapta-handu was described as having intense heat, but at the sources of the Ranha- was winter created by the daevas. The Rasa has not been properly identified. It was evidently a tributary flowing into the Indus from the right. The Vendidad description would indicate a northern tributary where people lived who followed the Indo- Aryan religion. There is also reference to Taofca invaders on the Ranha. Could this refer to the Tusaras? If so, the Rasa could be the present Shyok River of Kashmir. If not, it could be any one of a number of rivers flowing into the Indus from the north.

This river coming from a cold country has been excluded from consideration heretofore because it has not been satisfactorily identified, and from the meager data we have been able to note only some limiting conditions.

A number of cultural characteristics of the countries are mentioned in the Vendidad : pederasty on the Drsadvati, interrment of the dead on the Sarasvati, cremation on the Candrabhaga. I must leave these questions to the cultural anthropologists.

Textual evidence that the Indo-Aryans followed the same route as the Iranians to the northwest depends mainly on the river hymn of the Rgveda, where the rivers are named from east to west. 1

However, there is one important bit of supporting evidence. Indra slew the Aryas Arna and Citraratha on the far side of the Sarayu, 2 which is a long way from the vedic center of power in the Panjab.

From the above evidence we may reconstruct the migrations of the Aryas. Their homeland was within the Meru-Lake Manasa-sarovara region. They were probably pushed out of there by some more powerful people. Some migrated northwest down the upper Indus; and then the Iranians drove on into Iran, but left some behind, as the Sakas, Kambojas, Pahlavas, and perhaps some Sogdians; the Dardic branch remained in northwest India the Daradas, Kasmiras, and some of the Khasas (some having been left behind in the Himalayas of Nepal and Kumaon). Some of the Indo-Aryans may have followed this route and remained behind in the northwest, but there is very little to indicate it.

In another migration, the Iranians were the first to descend the Sarayu. They were followed by the Indo-Aryans who forced the Iranians

1 Until one comes to what are believed to be the tributaries of the Indus, where one finds the Rasa; Rgv. X. 75. 5, 6. Some of the tributaries, including the Rasa, are also named in V. 53. 9 together with the Sarayu. to scatter east and west, where they became the eastern and western Manavas, the "earliest" race in India. A second wave of Indo-Aryans forced the first Indo-Aryans to flee west to the less rich lands of the Panjab and it is from the latter that we have the vedic literature. The strongest of the Indo-Aryans, as the Paficalas, remained in the rich Ganges- Jumna valleys. The Indo-Aryans of the Ganges- Jumna and Panjab areas eventually drove the Iranians south, where they mixed with native elements and became the Yadavas and Haihayas, or west into Iran.

That the Iranians had come to India ahead of the Indo-Aryans explains why they were the first in India, the direct descendants of Manu. Later the Indo-Aryans drove south and separated the Manavas into an eastern and western branch. That the Iranians were the first in India accounts for their not remembering Meru and many other things connected with Tibet. They had left there earlier and they were separated from contact with Tibet by Indo-Aryans.

The above remark applies to the Iranians who fled west, the western Manavas, perhaps the authors of the Avesta. The Iranians who fled east, the eastern Manavas, retained traditions about Tibet. 1

This outline of the invasions from Tibet explains why the Indo- Aryans regarded the Manavas, Yadavas, and Haihayas as foreigners, and yet wherever these Iranians or part-Iranians settled in India the languages spoken today are classified as Indie, as if descended from Sanskrit. The Iranians of India probably spoke dialects that were closer to Sanskrit than is Avestic recorded centuries later. Bringing these peoples to speak something close to Sanskrit or Prakrit was about like teaching a modern Roman to speak standard Italian.

We have seen above 2 that the Rgveda hymns frequently mention Indra overcoming foes, Dasa and/or Arya. Now, although at a later period the Sanskrit-speaking Aryans sometimes fought each other, it is much more likely that the invading vedic Aryans, surrounded by foes, fought most often against Iranian-speaking Aryas, and that it is the latter that Indra overthrew. 3

During the vedic period in the northwest some of the natives had no doubt been partially aryanized. Half the Kaikeyas, for example, were Aryan. But during the epic period, Pargiter has already noted that the Indo-Aryans were losing their control over the Panjab. Rather,

1 See p. 17 and n. 1.

2 P. 32.

3 B. K. Ghosh (Vedic India, 220221; observed that with the daiva- worshipping Aryans there came to India also asura-worshippers ; and from his mention of Zarathustra's condemnation of orgiastic festivities of the daiva- worshippers, one may conclude that the latter must have been present among the Iranians. Protecting the vedic Aryans not only from the Dasas but also from the asura- worshipping Aryas no doubt provided Indra with a full-time job.

we should say, there is no evidence that they ever had control over the Panjab. The Sanskrit-speaking Aryans who had been forced west to the Indus valley composed the weaker branch of the Indo-Aryans and the Rgveda shows them to be in continual combat with enemies* It mentions battles and defeat of enemies, but nothing to indicate any extensive and durable kingdom or any lasting peace. Then in epic times the greater part of the Indo-Aryans are concentrated between the Sarayu and the Sarasvati and they are inclined to look down upon the Indo-Aryans of the Panjab. That probably was nothing recent. The Ganges-Jumna Aryans had early defeated the vedic Aryans and forced the latter to flee west. They had probably always looked down their noses somewhat at the vedic Aryans, particularly when it came to war or statecraft.

So if we find a difference between the language of the vedas and that of the Mahabharata, if we find different names for some of the rivers of the Panjab in the hymns and in the epic, it is not due entirely to a difference of time but because they represent different branches of Indo-Aryans and different migrations.

The greater part of the aryanization between the Sarayu and the Sarasvati seems to have been carried out by the Vasavas, who were comparatively late. It is unlikely that the brahman and ksatriya varnas alone could have effected this, and one is led to the conclusion that they were accompanied by a large part of the Indo-Aryan population.

Mackay stated that the end of Mohenjo-daro may be fairly safely dated to the 17th century B.C. 1 In a more interpretative summary on the place of the Indus civilization in general history, Piggott implies invasions before 1500 B.C. 2 and the composition of the Rgveda some- where about 1440 1500 B.C. 3 in an apparent attempt to connect the two.

But the invasion of India toward the decline of Indus civilization may not have been by the Aryas alone. From the appearance of Map 1, the western Anavas surely had something to do with the eclipse of that culture. If they were the yellow race composing the vaisya varna, they were in India before the Aryas or they accompanied them.

Greeks in India

Since the theory of a central Asian homeland of the Indo-Europeans was evolved, "Tokharian" (a centum language) has been discovered north of Tibet. And here we find the homeland of the Indo-Iranians in Tibet. And now HansKrahe, in a study of Indo-European hydronomy, 4 makes the interesting observation:

1 Early Indus Civilization, p. 4. 2 P. 238. 8 P. 255.

4 SprachverwandtscJiaJt im alien Europa (Heidelberg, 1951), p. 24.

,,AuBerhalb dieses Raumes, der um es noch einmal zu sagen von der Westkiiste Europas bis tistlich der Weichsel, von Skandinavien bis zu den Alpen reicht, stehen die iibrigen indogermanischen Sprachen ; und damit stehen auBerhalb etwa auch das Griechische und die Griechen, auch sie natiirlich Indogermanen, aber nicht teilhaftig des Kreises, welcher die charakteristischen alteuropaischen Gewassernamen schuf. Nach allem, was wir wissen, sind sie fruher nach dem Siiden gekommen, als die Illyrier und die Veneter, fruher auch als die 'Italiker', wohl schon zu Beginn des 2. vorchristlichen Jahrtausends, ahnlich wie die ebenfalls indogermanischen Hethiter und ihre Verwandten in Klein- asien. Woher sie kamen, ist ungewiB. Vorhistorische Sitze haben sich weder fiir die Hellenen noch fur die Hethiter bisher auffinden lassen, auch nicht mit Hilfe der Namenforschung. Am wahrscheinlichsten diinkt uns, daB ihre Heimat Ostlich des hier umschriebenen und als alteuropaisch bezeichneten Bezirkes lag.*'

If students of Indo-European toponymy are looking to the east for the homeland of the Greeks and Hittites, perhaps we shall not be too bold in suggesting that the evidence of the Indie and Greek sources indicating Greeks in India before Alexander should be further investi- gated. The people of Nusa named a mountain above the city Mr]()6<; "thigh," because of Dionysos birth, according to Arrian. 1 But was that the reason for the name, or was it named after Mt. Meru which had been forgotten ?

The people of Nusa were quite evidently playing up their Hellenism with Alexander in order to get off as easily as possible as their possession of ivy, which could not possibly have been brought overland from Greece; yet out of a considerable amount of myth and fiction which had no doubt grown up about Dionysos and Herakles, there is probably enough truth to require careful consideration whether some Greeks may not have remained in India, while others migrated to Greece in prehistoric times, so long before that they preserved no memory of the migration. Herakles was considered a native of the country by the Indians themselves, according to Arrian. 2

Modern scholars can be no more sceptical of the story of Greeks in India before Alexander than Strabo. And yet Strabo's argument 3 that the peoples through whose countries Dionysos passed would have preserved some record of the event is invalid because they did not pass from Greece to India, but remained in India. Rather we may point out that the Greeks themselves have no record of whence they came. Or perhaps we should say their record has not been correctly read, just as the Indo-Iranian records of their homeland have not been correctly read before.

1 Indica 1.6. 2 Indica 8. 4.

3 15. 1.6-9.

There is some ground for belief that the homeland of the "Tokhar- ians," the Indo-Iranians, and perhaps the Greeks was in eastern Asia, rather than in central Asia. This is not an essay to try to establish eastern Asia as the homeland of the Indo-Europeans, but only to present the evidence found on the two branches of the family with which we are concerned in investigating the ethnography of ancient India.

Situation before the Rebellion

We may here briefly review conditions in India before the Great Rebellion. Map 1 will show that the Anavas had all but submerged the Indo-Aryans in the Panjab, but they had themselves become partly aryanized.

We find traces of the Yadavas on the fringe of a vast area in the west : the Vatadhana and the Salva and the subdivisions of the latter along the Satadru; then the Kukura, Andhaka, Surasena, Abhisaha, Pataccara, and Usinara between the Drsadvati and the Ganges; below there Mathura, and the former Yadava kingdom of Cedi along the south bank of the Yamuna; and to the east Jathara, Kaukura, Bhoja, and Dasarna ; then coming west along the Purna were Nisadha, and Vidarbha ; and along the west coast, Dandaka, Surastra, Nairrta; and along the lower Indus, Anarta and Anupa. Within this peripheral area were the Haihayas, the younger branch of the Yadavas, along the Vindhyas. There is little doubt that the Yadavas and Haihayas had once been the chief power west and south of the Yamuna. *

But shortly before the Great Rebellion the Yadavas had conducted a long struggle with the chief power in the east, Jarasandha, king of the Magadhas, who had finally forced Krsna to flee from Mathura to the west. And two of the principal kingdoms, Matsya and Cedi, had become aryanized.

The Yadavas were probably already very weak before the Great Rebellion, partly from the losses sustained in the long war with Jara- sandha, partly from aryanization of some of the most important king- doms, partly from the scattered position of the remaining Yadavas and their lack of cohesion. To say that one was a Yadava still commanded respect in an Aryan, because the Yadavas had been the chief force in the west, and force always commanded respect in an Aryan. But the Mahabharata seems to know little about the Yadavas. It is not clear whether the Bhojas were Yadavas or Haihayas or both, whether Yuyu- dhana Satyaki and Cekitana were Satvatas or Vrsnis, or just what Yadava tribes fled to Anarta. 1 This confusion is perhaps in part because the

1 The Aitareya Brahmana seems to use Satvata for Yadava, or at least for the Yadavas south of the Satpura range (viii. 14). And at a later period Andhakas,

Mahabharata is not the book of the Yadavas, l but in part because of indifference about people who were no longer very important. 2

In the east the chief powers were the eastern Anava, to whom the Saudyumna stock offered some support, and the eastern Manava.

The Aryan strength was along the Yamuna and the upper Ganges. The Pandavas received very little aid from the Indo-Aryans of the northwest. The Indo-Aryans supporting the Pandavas were so far in the minority that they would have lost the war, according to their own account in the Mahabharata, if it had not been for their superior weapons which the rsis, of course, attributed to their gods 3 .

For the line-up of opposing forces, see Map 2. 4

The southern Dravidians are said to have been on the Pandava side. Perhaps the Dravidians had been annoyed by encroachments of the Yadavas further north and so favored the Pandavas. But of course there is enormous exaggeration in the Great Epic, 5 and probably

Vrnis, Dvaivavrdhas, and Mahabhojas were thought to be four branches of Satavatas (Law).

1 The confusion is perhaps worse in the Harivarhs'a, but that is later than the Mahabharata.

2 They continued to play an important role south of the Vindhyas.

3 For a discussion of the "divine" weapons of the Aryans, see E. W. Hopkins, "The Social and Military Position of the Ruling Caste in Ancient India ..." JAOS

13 (1887), 296ff., and "On Fire-Arms in Ancient India," ibid., p. cxciv ff., and especially V. R. Ramachandra Dikshitar, War in Ancient India (Univ. Madras, 1944), pp. 93 ff.

In a study of one of the weapons in the Pandava army, M. B. Emeneau in- ferred that the Sarriga, the bow of Kr^na, was probably a sinew-backed wooden bow with a strip of horn on its belly, giving much greater force. This bow seems to have originated in Siberia or Central Asia, and there is no evidence of it in the Rgvoda. (From an address before the Western Branch of the American Oriental Society, Nov. 24, 1951, with Prof. Emeneau's permission).

4 For the Kaurava supporters the writer has also used the list in Cr. Ed. VIII. 4, Roy VIII. 5, which is quite orderless and hence has not been included in the appendix. A few names have also been added from Pargiter's article on "The Nations of India at the Battle between the Pandavas and Kauravas," JRAS (1908), 309 336. Participants in the war, whose geographical location were too uncertain to place on Map 2 were: Kaurava -AsVataka, Arecaka (?), Kanana (?), Govasa, Cambupa ( ? ), Cicchila ( ? ), Dasamiya, Paribhadraka, Prasthala, Mavellaka, Lalittha, Varmila (?), Vikarna, Venika, Sreni, Sarhsthana; Pandava Kum- bharsa(?), Tittira, Nakula, Narayana, Prabhadraka (see Paribh- in Kuru army), Madaka, Salveya, Somaka; on both sides Da6eraka, Pi&aca, Raksasa, Naga.

5 See p. 1, n. 2. An example closer to the subject under discussion is afforded by a comparison of the digvijayas (list 2 below) with the dyuta list (no. 3 below). The digvijayas show the Pandavas conquering all India. But the tribute list indi- cates that the digvijayas were probably actually only punitive expeditions against foreign peoples in the Panjab, the Himalayas, and the east. Some scattered refe- rences to better-known kingdoms follow the list published hero; these perhaps represent contributions to help carry out the expedition; and were put down by later poets as tribute. The tribute list suggests that the Pandavas began their career as military adventurers and semi -mercenaries.

some of this was in the earlier Kuru epic. It is difficult to see how some of the widely scattered Kaurava forces could have rendered effective aid. Since this is primarily a work on ethnography, the array of countries in the Great Rebellion has been of primary interest to show the position of the Aryans, not to make a critical analysis of the allied powers.

Aryans Knew Little Beyond India

The only country outside of India that the Indo-Aryans seem to have known was Tibet as far east as Lake Manasa-sarovara. To the west they knew only a narrow fringe along the farther bank of the Indus and along the Kabul River. Toward the east they seem to have heard only of the river Lohita (Brahmaputra). The stories about Prag- jyotisa are conflicting. The center of the early state of that name was probably not in Assam, as generally believed, but somewhere in the upper Panjab, the Himalayas, or Tibet. It appears to have been a powerul empire extending from the Cinas and Yavanas on the west to or beyond the Kiratas on the east. Its center probably changed to Assam in the post-epic period. To the south the only country outside of India proper which was known to epic Aryans was Ceylon .

To eliminate some enticing theories about the Aryans' extensive geographical knowledge outside of India, we may state that Bahlika was not Balkh, Cma not China, Roman not Roman.

The Aryans had more than enough to occupy themselves in trying to conquer northern India. They were not greatly disturbed by events outside of that vast subcontinent until some foreign tribe actually reached its borders.

Aryans Knew Little Beyond India

The only country outside of India that the Indo-Aryans seem to have known was Tibet as far east as Lake Manasa-sarovara. To the west they knew only a narrow fringe along the farther bank of the Indus and along the Kabul River. Toward the east they seem to have heard only of the river Lohita (Brahmaputra). The stories about Prag- jyotisa are conflicting. The center of the early state of that name was probably not in Assam, as generally believed, but somewhere in the upper Panjab, the Himalayas, or Tibet. It appears to have been a powerul empire extending from the Cinas and Yavanas on the west to or beyond the Kiratas on the east. Its center probably changed to Assam in the post-epic period. To the south the only country outside of India proper which was known to epic Aryans was Ceylon .

To eliminate some enticing theories about the Aryans' extensive geographical knowledge outside of India, we may state that Bahlika was not Balkh, Cma not China, Roman not Roman.

The Aryans had more than enough to occupy themselves in trying to conquer northern India. They were not greatly disturbed by events outside of that vast subcontinent until some foreign tribe actually reached its borders.

Means of Linguistic Identification

Languages cannot be spoken over great areas without some differenti- ation and I have given the designations of languages of the present time to some of the ancient tribes, particularly in the Dravidian territory, assuming that, although the differentiation was not as great as today, it would provide some guide to the reader. I have differentiated between northern Tamil and southern Tamil, although the knowledge of Tamil dialects is insufficient to provide an actual linguistic criterion. Since the comparative grammar of the Dravidian languages has not pro- gressed far, the assignment of ancient tribes to a particular Dravidian language is necessarily a temporary makeshift. 1

1 Some countries basically Kanarese in language may have been classified otherwise here. The study of R. V. Jahagirdar, "Kanarese Influences on Old Marathi with Special Reference to Jnanesvari," ABORI 11, pt. 4 (1930), 374397, is not specific enough on dates and places to show whether the Kanarese influence was due to an early Kanarese substratum or to late contact with Kanarese.

Sometimes the identification of the language or race of a certain country or tribe hangs from a more tenuous thread of evidence than I should like it to, but then it is because this was the only clew available. Thus I have found no information regarding the race or language of the Kunti. But there is a reference to the language of the Avanti, as if it were something different from Sanskrit. Avanti and Mahismati both end in -ti and both are the names of Haihaya kingdoms. Kunti also ends in -ti and it lies between the Haihaya kingdom of Tundikera and the Haihaya-Yadava kingdom of Bhoja. So I have marked the kingdom of Kunti as Haihaya.

These Haihaya countries lie on the Narmada except for Avanti, the capital of which lay back from the river, although the country itself probably extended down to it. When I found the city the name of which is usually written Tripura was also on the Narmada, I suspected that the name had been aryanized, that it had originally been Ti-pura "the city of the Ti" or "the city on the Ti." Later I found the spelling Tipuri in the earliest record of the word, on coins of the second or third century B.C., and I classed Tipura as Haihaya. Afterward I learned that "Tripura" had been the capital of a great Haihaya state. *

Among the Western Anavas the evidence for classification is some- times more insecure than one would wish. The Yaudheyas and the Kai- keyas are known to be Western Anavas. And in the same general region as these are the Gramaneyas and the Madhyamikeyas, on whose origin we have no information. I put them on the map as Anavas because of the ending -eya and their geographic nearness to known Anavas. Later, I found that on coins of about the second century B.C. is found the legend Majhyamikaya-ibijanapadasa, in which the Madhyamikayas are associated with the Sibi, known Anavas. 2

It was probably not -eya, but -ya, which was characteristic of many Western Anava names ; for Kaikeya is also spelled Kekaya, and Madhya- mikeya was also spelled Madhyamakdya. Hence I have classed Tarksya as Western Anava: and I suspect that the Matsya, despite the Sanskrit appearance of the name, were also.

Except for Haihaya, tribal names ending in -ya seem to occur almost exclusively in the region of the Indus and its tributaries, where the known Anavas are found thickest on the map. And Haihaya is obviously not a root with a suffix -ya, but a reduplicated word Haihai to which Sanskrit added the usual stem ending -a for peoples, just as it added -a to Tibeto-Burmic An, Van, Ka-lin or Klin, noted above.

But could not a similar thing be said of the other ethnic terms ending in -ya ? Yes, there is nothing linguistically to prevent our deriving Yaudheya

1 Alexander Cunningham, Report of Arch. Survey of India 9 (1879), 54 57.

2 These coins were found in and about Chitor in southern Rajputana, which Allan recognized was south of the location of these tribes in the early literature. from a Tibeto-Burmic *Yau-dei, the -mikeya of Madhyamikeya from

• Mi-kei, Grdmaneya from *Gra-ma Nei, or Matsya from *Ma-tsi or *Mi-

ts r i. There is, however, still a difference between these terms and *Haihai. For the latter is reduplicated, and the other terms are not.

Two other ethnic names in -ya, Sukatya (?) and Saubalya, will be found outside the Punjab region on the ethnic map. In the Geography (see Appendix, 1. 9 11) they follow Matsya, and I surmise that they refer either to parts of Matsya or to Matsya colonies in either case to Western Anavas.

After completing the ethnic-linguistic map, I found a considerable concentration of ethnic names ending in -tha or -tha in the Himalayas in the northeastern region of the Five Rivers. Only one people whose name had such an ending, the Ambastha, had been assigned by Pargiter to one of the ethnic divisions devised by the Indo- Aryans the Western Anava. But we have found above that the Western Anava were probably Tibeto-Burmans of the West Himalayish linguistic stock, and one of the names, Kulattha, was applied to a people who were geographically close to the Kulindas, the ancient West Himalayans, or who were a tribe of that nation. And the first part of the names Kulattha &ndKulinda are the same ; compare modern Kulu. Hence I have classed all these ethnic names in -tha or -tha as Western Anava.

In classifying Mundic countries, the ending -kola has been used by previous writers. But some peoples of eastern India have been classed as Mundic because their names look more likely to be Mundic than Eastern Anava, as Kundlvisa, applied to a people in territory contiguous to Utkala.

Thus, in all these examples the linguistic means of classification does not stand alone but is supported by geographical considerations. And the test of any hypothesis is whether it proves fruitful in further research, and the linguistic means has led to further confirmation in the case of Ti-pura and Madhyamikeya, as remarked above.

Yet, for the older scholar or the ethnologist, it is hardly necessary to add that any classification made here is only a beginning. If we take a modern illustration, England was named from the Angles; but the Angles were only one of the ethnic components that went to make up ancient England: Kelts, Romans, Angles, Saxons, Jutes, and probably many a small band or occasional immigrant that we know nothing of. Similarly France was named from the Germanic Franks, but the French are generally thought of as Gauls (Kelts) or Latins. If we had no other knowledge of the ethnic composition of ancient England or France, the linguistic means would give us a valuable clue but an incomplete picture.

And so none of the countries of ancient India were without racial mixture or without different races and sometimes different languages within the same kingdom. Some indication of this has been given in the discussion of the varnas of epic times, or in occasional references to the mixed nature of the races of a kingdom or tribe. Hence this study is not definitive, but only a beginning.

In the index I have given some racial designations based on the work of modern physical anthropologists. In southern India particularly, the work of von Eickstedt seems to offer some ground for the division of states and peoples when the Dravidian languages do not. These modern racial terms are cited rather sparingly, however, partly because the two latest authorities known to this writer, von Eickstedt and Guha, are far from agreement on all points; partly because Guha gives at least two racial designations for each area on his map, without sufficient explanation to make identification possible ; partly because of the small size of von Eickstedt's and Guha's racial maps and lack of sufficient identifying geographical features to determine the precise areas.

I have followed von Eickstedt in southern India, because he appears to have been more familiar with that area and because of the simplicity of his terms, and where I have given modern racial terms for northern India I have followed Guha because of his finer distinctions. Somatic measurements for small areas, such as those covered by some of the early nations or tribes, are often wanting, and I have generally preferred to leave to ethnologists the correlation with modern racial terms in northern India.

Some approximate correlations between von Eickstedt's and Guha's terms are

von Eickstedt Guha

(graceful)Indid Mediterranean

Gondid Nisadic

Melanid Palae-Mediterranean

Malid Proto-Australoid