Learning Center
Plans & pricing Sign in
Sign Out

Proto Indo-Aryans (T. Burrow, April1973 JRAS)


									                               THE PROTO-INDOARYANS
                                             By T. BURROW

IT IS NOW GENERALLY agreed by most authorities on the subject that the Aryan linguistic
vestiges in the Near East are to be connected specifically with Indo-Aryan, and not with
Iranian, and also that they do not represent a third, independent Aryan group, and are
not to be ascribed to the hypothetically reconstructed Proto-Aryan. This conclusion is
incorporated in the title of M. Mayrhofer's bibliography of the subject, Die Indo-Arier im
alten Vorderasien (Wiesbaden, 1966), and it can now be taken as the commonly accepted
view. It is based on the fact that where there is divergence between Iranian and Indo-Aryan,
and where such elements appear in the Near Eastern record, the latter always agrees with
Indo-Aryan. Such items are aika "one" and suriyas "sun", and the colour names parita-nnu
and pinkara-nnu which correspond to Sanskrit palita- "grey" and pingala- "reddish". The
evidence of vocabulary is supported by that of the four names of gods appearing in the
Hittite-Mitanni treaty, where the Vedic gods Mitra and Varuna, Indra, and the Nasatyas
can be clearly recognized. This combined evidence is sufficient to establish the conclusions
of Mayrhofer and others beyond reasonable doubt, and the arguments of A. Kammenhuber,
who later attempted to resuscitate the theory that the Aryans of the Near East were Proto-
Aryans, cannot be said to have been successful.1
     One of the most important contributions to this conclusion was P. Thieme's study of
the divine names in the Mitanni treaties published in I960.2 This article provided the final
demonstration that the gods of the Mitanni treaties are specifically Vedic gods, and that
they cannot be Proto-Aryan, a conclusion that was first argued by Sten Konow. 3 Some of
Thieme's conclusions will be examined below and criticized, since his article is the main
starting point of the present investigation, but the main principle which he formulates is
deserving of whole-hearted approval: "We cannot reconstruct Proto-Aryan religious
terms—and much less Proto-Aryan religious ideas—by simply and naively projecting
Rigvedic data into Proto-Aryan times. A reconstruction can be attempted only by a careful
confrontation of Vedic and Avestan terminology". The assumption which Thieme criticizes,
that the Vedic religion has faithfully preserved the Proto-Aryan religion, has indeed been
commonly held, and this is due to three reasons: the greater antiquity and better state of
preservation of the Veda as opposed to the Avesta, the undoubted fact that Zoroaster was
the founder of a new religion whereas no such revolutionary change occurred in ancient
India, and the fact that the Proto-Aryan period was conceived of as being not very much
earlier than the Vedic period. In spite of all this, however, as Thieme demonstrates, there
are matters in which the Veda shows innovation, and the Avesta is more original.
     A case in point is the relationship between the terms Varuria and Asura (> Ir. ahura-).
It has been commonly assumed that Varuria was an original Proto-Aryan deity out of whom
the Iranian Ahura Mazda eventually emerged. There is however no trace of Varuna in the
Iranian record, while on the other hand Asura/Ahura is common to both traditions, and
    A. Kammenhuber, Die Arier im vorderen Orient, Heidelberg, 1968.
    P. Thieme, "The 'Aryan' gods of the Mitanni treaties", JAOS, 60, I960, 301-317.
    Sten Konow, The Aryan gods of the Mitani people, Christiania, 1921.
124                                    THE PROTO-INDOARYANS

 consequently it appears to be reasonable to assume that the dvandva combination
 Midra-AhuralAhura-MiOra of the Avesta is more original than the Mitrd-Varurid of the
 Veda. There are some passages in the Rgveda, though not many, where a god Asura,
 distinguished from, but occasionally working with Mitra and Varuna, is mentioned, and
this god is to be traced back to Proto-Aryan times. The same conclusion as regards these
 divine names was later adopted by I. Gershevitch (JNES, XXIII, 1964, 12), although earlier,
in the introduction to his Avestan hymn to Mithra (Cambridge, 1959), he had been prepared
to assume an Iranian * Vouruna.
      Thus the Indo-Aryan nature of the "Aryan vestiges in the Near East" is established both
on the basis of language and religion, but the material is still remarkably small. The reason
for this is, as Mayrhofer observes (op. cit, 11), that it all originates from outside the Mitanni
kingdom itself, and if ever the archives of that state are uncovered, then a great increase
of such material is to be expected. In view of the large amount written on the subject not
much more can be extracted from what is already available, but the name of the capital of
the Mitanni state, Wassukanni, seems to be capable of explanation in terms of Old Indo-
Aryan, even though none of the proposals registered by Mayrhofer is satisfactory. Since
the interpretation should be from Indo-Aryan, we should see in the first member a noun
vasu "wealth", and not an adjective vasu- meaning "good" as in Iranian (Av. vohu-).4
In this case the second element is obviously khani- "mine", the whole meaning "mine of
wealth (i.e. precious metals)", and this explanation provides the motive which attracted
those Proto-Indoaryans to this region. It was an age when prospecting for metals, precious
or otherwise, was being actively pursued, and the Aryans were as much interested in this
activity as anybody else. It is understandable that when it came to their knowledge that a
new and large deposit of such materials had been found, it would stimulate them to attempt
to get control of that territory. It will be seen later that they need not have been situated
very far away, so that any such information would soon reach them.
      It is a noteworthy fact that the conclusion that the Aryans of the Near East are to be
connected with the Indo-Aryans can be founded on such a small amount of linguistic
material. In this connection a pertinent remark of Mayrhofer is worth quoting (op. cit., 23):
"Die Abweichungen der beiden arischen Dialekte voneinander in Einzelheiten der Wort-
wahl, Wortbildung und Semantik werden infolge des tiberwaltigenden Eindruckes ihrer
Ahnlichkeit oft nicht geniigend wahrgenommen". The differences between the two halves
of the Aryan language-group are particularly noticeable in the case of vocabulary, and the
result is that it needs only a small amount of such material to make it clear whether the
language involved belongs to Indo-Aryan or Iranian.
      Since the "Aryan" of the Near East is to be connected with Indo-Aryan it follows that
the division of Proto-Aryan into two branches, Indo-Aryan and Iranian, must have taken
place before those languages were established in their eventual homes, and not merely be
due to developments which took place within each of the two groups after the Indo-Aryans
had settled in India and the Iranians in Iran. This conclusion could only be shown to be
wrong if it could be shown that the Vedic Indians, having migrated all the way to the
'According to L. Renou (£tudes vidiques et panineennes, 9, Paris, 1961, 91) vasu- in the Rgveda is used
    as a neuter (sg. and pi.) in the sense of "wealth, riches" and in the masculine as the name of a class
    of gods, but not as an adjective meaning "good".
                                     THE PROTO-INDOARYANS                                        125

 Punjab from their earlier home, had then retraced their steps and undertaken yet another
 migration in the direction of the Near East. Konow was prepared to believe this, but there
 is no evidence for it, and it seems that a theory involving such complication can be safely
 ignored. Such an early separation of Indo-Aryan and Iranian has often been envisaged
 (e.g. M. Haug, whose opinions will be noted below), but it is only this Near Eastern evidence
 which renders it certain. A further conclusion following from this is that the date of the
 Pro to-Aryan period must be pushed back further than has often been thought, and probably
it cannot be brought down below 2000 B.C. at the latest. This will also leave ample time
 for innovations in the Vedic religion of the kind referred to above.
       It is not quite satisfactory to use the term Indo-Aryan for the Aryans of the Near East
 and their language if, as assumed above, those Aryans were never in India, and consequently
the term Proto-Indoaryan has been adopted in the title of this paper. In doing so it has
 been necessary to give the term a wider sense than it would have if it were used in a purely
 linguistic sense. Linguistically, Proto-Indoaryan refers to that stage of the language existing
 before the migrations into India and after the separation from Iranian. Its phonology can
 be partially reconstructed by a comparison of Indo-Aryan and Kafiri (which is not to be
regarded as a separate branch of Aryan, but as descended from this Proto-Indoaryan),
and it can be seen that to a large extent it had not altered from Proto-Aryan. Typical
changes that may be assumed to have taken place between this stage and Vedic are the
change of the short diphthongs ai and au to e and o, development of h from jh and zh,
emergence of the cerebrals, simplification of final consonant groups, and so forth.5 On
the other hand the term Proto-Indoaryan as used in this paper refers not only to the pre-
migration Indo-Aryans, but also to all those who did not make the migration to India, or
migrated elsewhere, and their descendants. Among these descendants are the Proto-Indo-
aryans of the Near East whose language had, as is well known (cf. satta "seven" < sapta),
evolved beyond the Proto-Indoaryan stage.6
      The identification of the Aryan traces in the Near East as Proto-Indoaryan has, in
Thieme's words, "considerable historical implications". The problem is also brought
considerably nearer to a solution by this identification, inasmuch as the range of possibilities
is now narrowed down. On the basis of what has been said it is possible to improve on the
statement made by R. A. Crossland on this subject: "The Indo-Iranian branch of the
prehistoric Indo-European-speaking people, or part of it, must have thus been living in an
area close to northern Mesopotamia, probably in northern Iran in the fifteenth century
B.C. at the latest" (CAH, I, 2, 838). The first of the two possibilities can be excluded, and the
"part" can be identified as a section of the Proto-Indoaryans. It is likely that before the
take-over of the kingdom of Mitanni, Proto-Aryans were settled in northern Iran, and
evidence in support of this will be provided later in this article.
      Mayrhofer briefly refers (op. cit., 40, n.) to the historical problem raised by the presence
of Proto-Indoaryans in the Near East, and mentions the three possibilities that have figured
in the discussion. The first of these—that the Aryans of the Near East, after they had

 Further examples in my The Sanskrit language, London, 1955, 33.
* In view of this admitted ambiguity of the term Proto-Indoaryan an alternative would be to speak of
    Western Indo-Aryan but unless much more material turns up in the Near East there is no urgency to
     settle the terminology.
126                                     THE PROTO-INDOARYANS

left or been driven out of Mitanni, proceeded eastward to colonize North-West India—
he rightly excludes as being out of the question. The theory, dismissed above, that the Aryans
of the Near East had actually come from India, does not appeal to him very much, though
he does not go so far as to exclude the possibility. The third alternative, which he considers
to be the most likely, is that the migrating Indo-Aryans had separated, probably in Iran,
into the two groups. This theory, which, as Mayrhofer says, is much the most likely,
is the one adopted by I. M. Diakonov.7 He considers (as many have done before him)
that Central Asia was the original habitat of the Indo-Iranians, and notes in passing that
Soviet archaeologists usually connect the Proto-Indo-Iranians with the Andronovo culture
of Kazakhstan and Southern Siberia. He proposes that the Indian tribes moved first from
this region and that some sections of them moved into Iran and from there to the Near East,
while the main mass migrated to India.
      This is the most likely account of the events that led to the appearance of the Indo-
Aryans in North-West India, and of their close relations in the Near East, at roughly the
same time. Nevertheless certain modifications of this statement are needed, principally
because it does not envisage the possibility of intermediate settlement. The colonization
of North-West India by the Indo-Aryans was an extensive operation, lasting over genera-
tions, which could only have been carried out on the strength of an extensive population
base immediately outside the sub-continent. That is to say that before these migrations
Proto-Indoaryans must have been in occupation of large tracts of eastern Iran and western
Afghanistan (such as Bactria, Areia (Haraiva), Arachosia, and Drangiana), which only
at a later period came into the possession of the Iranians. One would also not expect that
the migrations into India left these countries empty of Proto-Indoaryans, but rather that
this was a movement of the surplus population, so that when the Iranians took control
of this territory they would find the Proto-Indoaryans settled there, and that in due course
of time the latter would be absorbed into and merged with the later-coming Iranians.
      A slight trace of the earlier home of the Indo-Iranians has been preserved in some
river names which were transferred by them from the old to the new country. This is how
the relationship between the Indian Sarasvati and Iranian haraxvaiti- is to be explained.
There is not, as Bartholomae suggests, a common inheritance from Proto-Aryan sarasvati
here, since, as pointed out above, this region was not the home of the Proto-Aryans. On
the contrary Sarasvati is in the first place the Proto-Indoaryan name of the river in Iran,
which after the migration was transferred to the river in India. The Iranian name is a loan-
word from Proto-Indoaryan, with a substitution of h- for s-, occurring also in hindu-,
which will be discussed below. Another case is the river name sarayu-, which was transferred
first from Iran {haraiva-1haroyu-) to a river in North-West India, and then again from there
to a tributary of the Ganges in eastern India.
      The name of the Oxus is of interest from this point of view. The name occurs in
Sanskrit (yaksu-) in the Mahabharata and Kalidasa, in Greek from the period of Alexander,
and in Iranian only on Kushana coins. In a more general sense Khotanese bassa "streams"
is compared (H. W. Bailey, Khotanese texts, VI, Cambridge, 1967, 231). The stem vaksu-
can be derived from the root vah- in the sense of "to flow" with the suffix -su; cf. ddksu-

    1 . M. Diakonov, htoriya Midii, Moscow, 1956, 124-125.
                                       THE PROTO-INDOARYANS                                            127

(dhdksu-). If so the word is Indo-Aryan, since the combination -ks- representing palatal
followed by s is a specifically Indian development. It has been thought that vaksu- might
be derived from vaks- "to increase", but though this might suit the Oxus, it does not fit
the meaning "stream" of the related Khotanese word, and so the derivation from vah- is
to be preferred. If so the name must have been given to the river by the earlier Indo-Aryan
population of Bactria, and taken over from them by the local Iranian population after
they had immigrated there. It does not, however, seem to have gained very wide currency
in Iranian.
     It is, however, in the sphere of the religious history of Iran that the presence there of
Proto-Aryans as forerunners of the Iranians can be most clearly demonstrated. In this
connexion it will be necessary to look again at some of Thieme's conclusions in the article
referred to above. Although he reconstructs a Proto-Aryan god Asura, Thieme does not
give any precise idea about the exact nature of this deity. An understanding of this god's
nature can, however, be obtained if we accept the conclusions reached by P. von Bradke
in the most detailed study of the term asura- that has been made.8 According to von Bradke
the nature of the god Asura corresponds to the meaning of the word asura-, which is
"lord, overlord, sovereign" (equivalent to the later Sanskrit isvara-), and the position of
Asura among the Proto-Aryan gods was that of sovereign ruler of those gods. Since the
word asura- is further found combined with Dyaus in the Veda, von Bradke goes further
and traces the Aryan sovereign god Asura back to the original Indo-European sky god,
who occupies a similar position in the related mythologies. This account seems to fit the
facts very well and it has the advantage of defining the concept of Asura, which otherwise
is not done. Having this meaning Iranian Ahura could very appropriately form part of the
combined name Ahura Mazda, which Zoroaster coined to express his conception of
the supreme deity when he founded the Mazdayasnian religion. On the other hand, in the
Veda Asura as a god had almost become obsolete, except in a few turns of phrase, and the
epithet asura- in the sense of "lord" could be applied to any great god whose power was
emphasized, even though it was most commonly used of Mitra and Varuna. In this use
also the term became obsolete by the end of the Rgvedic period.
     In these discussions much confusion has been caused by assuming a connection between
asura- "lord" and asu- "life" (particularly when the latter is rendered "spirit"). These two
should be kept clearly apart, as was pointed out long ago by J. Darmesteter,9 and has
recently been emphasized by B. Schlerath.10 On the one hand there is asu- "life", a straight-
forward derivative of as- "to be", and on the other hand Skt. asura-JAv. ahura- "sovereign,
lord", with which is connected further Av. ahu- "lord". Apart from its use as a divine
name, or in connection with names of gods, it is commonly used in Avestan as "lord" in
the ordinary human sense (Bartholomae's ahura-*), and also occasionally in the Veda.11
The Proto-Aryan word was at an early period borrowed into Finno-Ugrian; cf. Mordv.

   P. von Bradke, Dyaus Asura, Halle, 1885.
 • J. Darmesteter, Ormazd el Ahriman, Paris, 1877, 47; cf. also P. von Bradke, op. cit., 85.
   B. Schlerath, "Altindisch asu-, Awestisch ahu- und ahnlich klingende W6rter", in Pratidanam (Kuiper
      Festschrift), ed. J. C. Heesterman et al, The Hague, Paris, 1968, 142-53. For the other (in my opinion
      untenable) view, cf. H. Guntert, Der arische Weltkonig undHeiland, Halle, 1923,102, and J. Duchesne-
      Guillemin, TPS, 1946, 81.
   cf. W. Neisser, Zum Worterbuch des Rgveda, Leipzig, 1924-, I, 139-40.
 128                                      THE PROTO-INDOARYANS

azoro, azor "lord, master". The other connexions in Indo-European are with Hitt. hassu-
"king" and Lat. ems "master".
     Of the four divine names mentioned in the Mitanni treaty one, Mitra, is common to
the Vedic and Iranian traditions, one, Varuna, is unknown to the latter, and two, Indra
and Nasatya, do appear in the Iranian tradition, but only as demons. A list of principal
daevas in the Vendidad (10.9; 19.43) contains the names Indra, Na??hai#ya (= Nasatya),
and Saurva (Vedic l§arva; AB, VS.). Professor Thieme is prepared to allow these to have
been originally Iranian gods later demonized, but if that were so they would also have been
Proto-Aryan gods, as being common to the two traditions, and this would weaken the
argument against the gods of the Mitanni treaty being Proto-Aryan, since of the three only
Varuna would remain exclusively Vedic, and even he might have been lost only at a later
time in the Iranian tradition. There is, however, no reason to believe that Indra and
Nasatya ever belonged to the Iranian religious tradition. These three principal daevas of the
Vendidad should be classed, along with the gods of the Mitanni treaty (to which two out
of the three correspond), as Proto-Indoaryan deities. They are gods which were being
worshipped by the Proto-Indoaryans in eastern Iran when the Iranians took over the
country, and it is these gods, and others of which the names are not preserved, which were
condemned by Zoroaster. It has never been possible to account for the fact that the Iranian
gods, if they had been so uncompromisingly condemned by Zoroaster, should after a very
short time have been accepted back into the Mazdayasnian fold without any reservations;
or why, when this was done, the anti-daeva ideology should have remained just as strong
as ever, with these same gods in the forefront of the war against daevas. If, on the other
hand, we assume that the daevas who were condemned were the gods of the Proto-Indo-
aryans, then there is no question of the Iranian gods having somehow come back into
favour, since they were never condemned, and there is no contradiction between the
continuing intense hostility shown by the Mazdayasnian region towards daevas and the
worship of those Iranian gods.
     There has been a long discussion of the fact that the Proto-Aryan terms daiva- and
asura- ended up having different meanings in Indian and Iranian (Ind. deva- "god", asura-
"demon", as opposed to Iranian daeva- "demon", ahura- "(supreme) lord") and it has
usually been assumed, though wrongly, that the same reason should account for the changes
on both sides. Martin Haug12 assumed a schism between the Iranians and the Brahmanic
Aryans before their migration to India. If we substitute for this a "schism" between the
Iranians and those Proto-Indoaryans who were left behind in eastern Iran after the Indo-
Aryans had migrated to India, Haug's explanation fits very well as regards the degeneration
of the word daeva-. It does not, however, account for the change of meaning of the Sanskrit
word asura-, since this change took place in India at a time considerably later than the
migration, and there was no connexion between this and what happened in Iran to the
word daeva. Darmesteter13 opposed Haug's theory and asserted that there had been no
changes in the nature of the Indian or Iranian gods but merely an alteration in the usage
of words. As far as the history of the word asura- in Sanskrit is concerned, what he said is

     M. Haug, Essays on the sacred language, writings, and religion of the Parsees, Bombay, 1862, 248 ff.
     Daramesteter, op. cit., 261 ff.
                                      THE PROTO-INDOARYANS                                          129

perfectly true, but he was wrong about the daevas since those principal daevas mentioned
above are undoubtedly ancient gods who have been turned into demons.
     The theory that prevails nowadays, and has prevailed for some considerable time, is
quite different. According to this theory the Proto-Aryans had two sorts of gods, one
denoted by the term asura- and the other by the term daiva-, between whom there was
opposition. The result of this was that in Iran the daivas and in India the asuras were
eventually reduced to the state of demons. The theory is thus expounded by H. S. Nyberg:14
"Das arische Altertum kannte asura (= iranisch ahura) als eine Klasse von Gottwesen, die
einer andern Klasse mit dem Namen daiva (indisch deva, iranisch daeva-) nebengeordnet
war. In der religiosen Geschichte der Arier haben diese beiden Gottergruppen um die
Herrschaft gekampft. Die Entwicklung ist so verlaufen, dass in Indien die Devas siegten
und die Asuras zuerst zuriickdrangten, spater verdrangten, und auf die Stufe der Damonen
herabdruckten, wahrend dagegen in Iran die Ahuras den Sieg davontrugen und die Daevas
zu Damonen herabdruckten". This theory is accepted by Thieme (op. cit., 311), and by a
large number of other scholars,15 but it is false from beginning to end, and even a cursory
examination of the Vedic evidence shows it to be wrong. The theory is false because, as
Darmesteter pointed out long ago in his argument with Haug,16 no ancient Indian god
was ever turned into a demon. The Vedic gods who are considered to have belonged to the
asura class are Varuna, Mitra, and the Adityas, so the theory would imply that the Adityas
became demons, which is contrary to the facts. It is absurd to assume that a class of gods
called Asuras were relegated to the status of demons while every individual member of that
class remained a god. It is also wrong to speak of a class of gods known as Asuras since
the term is never so used. It is used of gods mainly in the singular as an epithet meaning
"lord", but in the plural, denoting a class, only of demons, and this not until the closing
phase of the Rgveda. On the other hand the term deva is used in the Veda and later of all
the gods, and there is no trace of its being restricted to any particular class.
     Just as the Vedic gods always remained gods, so the Asuras were always demons, and
so Darmesteter's view must be accepted, that it was only the meaning of the word which
was changed. This change became possible because the word asura- "lord" fell into disuse
in the ordinary language, though remaining familiar from the sacred texts. In particular the
phrase asurasya mdyd seems to have contributed to this change of meaning. It meant
originally the creative power of the supreme lord, but since mdyd in the sense of "magical
power" had been commonly spoken of in connexion with various individual demons, the
term was interpreted to mean "the magical power of the demon", and a new term asura-,
meaning the whole class of demons hostile to the gods, was invented.17
     Since the deva-god and asKra-god theory can be disproved on the Indian side, it
obviously cannot be used to explain developments in Iran. There is, of course, from this
side no evidence for it, since by definition it applies to the prehistoric period, and not to the
situation found in the Avestan texts. The Avestan literature provides us with the names of a
   H. S. Nyberg, Die Religionen des alten Iran, Leipzig, 1938, 96.
"e.g. E. Benveniste, The Persian religion, Paris, 1929, 40; J. Duchesne-Guillemin, La religion de I'lran
      ancien, Paris, 1962, 189; R. C. Zaehner, The dawn and twilight of Zoroastrianism, London, 1961, 36;
      J. Gonda, Change and continuity in Indian religions, The Hague, etc., 1965, 168.
   J. Darmesteter, op. cit., p. 269.
   cf. A, A. Macdonell, Vedic mythology, Strassburg, 1897, 156.
 130                                  THE PROTO-INDOARYANS

 large number of the pre-Zoroastrian deities of Iran, both those celebrated in the YaSts and
 those otherwise referred to, so much so that it seems reasonable to assume that we have a
 more or less complete inventory of the original Iranian pantheon. Among these divinities
 there is no sign of any such classification as that into Devas and Asuras, and since the list
 of gods retained by the Mazdayasnians is so large and complete there is hardly room for an
 additional lot who were condemned as daevas. It has been suggested that tribal differences
 might be involved,18 but when this aspect is considered, the names of the three principal
 daevas, corresponding to Vedic deities, show that the other tribe concerned was that of the
      There is another reason why the daevas condemned by Zoroaster cannot have been
 Iranian gods, and that is that the Proto-Iranian and pre-Zoroastrian Iranian word for
 "god" was not daiva- but baga-. This can be posited on the strength of the distribution of
 the word in the various Iranian languages: OPers. baga-, Av. baya-, Sogd. /?y-, etc. Reference
 was made above to the extent to which the Iranian and Indo-Aryan vocabularies differ
from each other, and the case of the word for "god" is one of the most important instances.
 The antiquity of this usage is demonstrated by the fact that it also appears in Slavonic
 (OS1. bogii "god", etc.), and this Slavonic word is not considered to be a loanword from
 Iranian.19 The use of bhaga- "assigner, apportioner" (in Sanskrit the name of a particular
deity) as an alternative word for "god", must, therefore, go back to an early period, and
it is characteristic that Indian and Iranian should diverge in this respect.
      In the Avesta the usual word for "god" is not baya-, but yazata-, and it is not difficult
to see what the relationship of these two terms is. As Duchesne-Guillemin has pointed
out,20 yazata- (originally an epithet "adorable" = Skt. yajata-) can be regarded as having
been invented by the Zoroastrians as a substitute for baga-. It cannot, of course, have been a
substitute for daeva-, since daevas were utterly condemned, and this fact provides further
confirmation that the Iranian word for "god" was only baga-. The reason for the substitution
was, no doubt, to indicate the changed status of the old Iranian gods, who were to occupy
a more subordinate position in the new religion of Ahura Mazda. In the event this was not
quite how things turned out, since the old bagas and new yazatas, not having been con-
demned by Zoroaster, as were the daevas of the Proto-Indoaryans, continued to be looked
upon by their worshippers as having much the same status as before, and to a large extent
the old name also was retained. On the other hand, where derivatives of yazata- occur
(e.g. NPers. Izad/yazddn, Saka, Khot. gyasta-, Tumsuq jezda-), these must, if the above
theory is correct, be traced back to this Zoroastrian coinage.
      If one were to rely merely on a reconstruction from later forms in the various Iranian
languages, and no other information was available, then one would without hesitation
reconstruct a Proto-Iranian daiva- "demon", since the representatives of this word are
even more widely distributed in the Iranian languages than are the words descended from
baga-. But such a conclusion would be mistaken, since it was Zoroaster who first declared
the daevas to be evil powers. This fact is vouched for not only by the tradition of the
Zoroastrians (cf. Yt. 13.89-90), but also by the nature of Zoroaster's own preaching against
   Nyberg, Die Religionen, 339, ascribed Indra, etc., t o the Medes.
" cf. M . Vasmer, Russisches etymohgisches Worterbuch, Heidelberg, 1950-8,1, 98.
   La religion, 166.
                                  THE PROTO-INDOARYANS                                     131

the daevas. It is generally agreed that in his references to daevas Zoroaster is not simply
saying that demons are bad—this would have been unnecessary, since everyone was
against demons—but that he is condemning certain deities who were being worshipped as
such, and who as a result of his propaganda were turned into demons. The fact that daiva-
"demon" is represented throughout Iranian can be taken as an indication of the wide
spread of Zoroastrian teaching among the early Iranians.
      Although the daevas were in origin the dethroned deities of the Proto-Indoaryans, and
this remained so for some time, eventually the term came to stand for a whole host of
demons, monsters, and evil spirits (some pre-Zoroastrian in origin), as well as for personi-
fications of vices, diseases, calamities, etc. The result is that not a great deal of information
is available about the dethroned gods, and only the three above mentioned can be identified
with certainty with Vedic gods. To these we may possibly add, with Nyberg,21 buti- which
would correspond to Skt. bhiiti- "prosperity" (also personified and equated with Laksmi).
      The demonizing of the Proto-Indoaryan gods was rendered possible by the fact that so
many of them, and particularly the important ones, had nothing corresponding to them in
the Iranian pantheon, and vice versa. Thus Indra, Varuna, Nasatya, and 3arva, which we
can say for certain were Proto-Indoaryan gods, had no correspondences in Iranian, and
there are other Vedic gods, who may also have been worshipped by the Proto-Indoaryans,
such as Visnu, Savitr, Pusan, etc., of which the same can be said. Thus in spite of certain
correspondences, such as Mitra/Mi#ra, etc., it could easily appear to the Iranians and to
Zoroaster that the daevas, taken as a whole, were a different set of gods from their own. In
the case of Mitra/Mi#ra, this god is so much more important in the Iranian scheme of
things than in the Vedic (and presumably Proto-Indoaryan), that the correspondence could
be ignored. Only in the case of the god Vayu, who occupied a place of similar importance
and similar character in both religions, does a genuine dilemma appear to have arisen,
and this has resulted in a good Vayu- and a bad Vayu-, a god and a demon, existing side
by side in the Mazdayasnian system.
      That Av. daeva- should be a loanword from Proto-Indoaryan is well accounted for by
the circumstances which existed in Zoroaster's time and in the time preceding it. To the
Iranians coming down from Central Asia the regions of Eastern Iran were an area of recent
occupation, and still largely populated by the Proto-Indoaryans who had preceded them.
To judge by their close relations in Vedic India the priestly organization of the latter is
likely to have been well developed and sophisticated, probably more so than that of the
incoming Iranians. It also seems clear, from the difficulties which Zoroaster had to over-
come, that the Iranian princes in this area respected the old religion and its representatives,
and to a large extent adopted those cults. Since he reacted so strongly against it, it is likely
that Zoroaster was brought up in this environment, though there is also the possibility that
when these regions were incorporated in Iran, the Proto-Indoaryan priesthood had also
infiltrated into old Iranian territory.
      In addition to the word daeva- there are some other words in Avestan which can be
regarded as loans from Proto-Indoaryan. One certain case is usig-, a priestly title correspond-
ing to Vedic usij-, which must originally have been a Proto-Indoaryan word for a certain

" H. S. Nyberg, op. cit., 340.
132                                 THE PROTO-INDOARYANS

class of priest. Some other terms used in this context, namely kavi- and karapan-, are more
problematical. The term kavi- could, in certain contexts, have a reference to the Proto-
Indoaryan priesthood, but since it is also used of Iranian princes, it is simplest to take it
in this sense everywhere. The bad sense of the word kavi- is then due to the fact that the
earlier Iranian rulers had as a general rule supported the Proto-Indoaryan religion. It had,
however, also to have a good sense, since it was also the title of Zoroaster's patron Vistaspa-,
and also of the kings of the Kayanian dynasty, who were too illustrious to be lumped to-
gether with the kavis in general. Even though it may have been used only of Iranian princes,
it is a notable fact that this word occurs nowhere else in Iranian except in Avestan in this
one context, and it is not out of the question that this may have been a title taken over by
the incoming Iranians from their predecessors.
      The term karapan- was connected by Bartholomae with Skt. kalpa- "rite", and in that
case it could well have been taken over from Proto-Indoaryan. On the other hand it has
been proposed to see in this a derogatory term meaning "mumbler" derived from an
Iranian karp- "to mumble" (cf. Skt. krp-).22
      A small number of Proto-Indoaryan words in Avestan can be identified with some
probability among the "daevic" words which are a curious feature of that language. These
words were studied from this point of view by L. H. Gray in an article published in 1927.23
Gray noted that not all the antitheses involved in the two vocabularies are to be explained in
the same way, but nevertheless he made it clear that when the two vocabularies are examined
from an etymological point of view, a significant pattern emerges, namely that the "ahurian"
words as a general rule are the ones that have the widest representation throughout the
Iranian languages, while the "daevian" words have in many cases either no other cognates
in Iranian, or are represented only in a few marginal dialects. A good example of this
state of affairs is found in the daevian word karana- "ear". In Iranian the IE word for
"ear" is well preserved, and in Avestan (us-) it counts as an ahurian word. Beside it there is
Av. gaosa-, etc., which is a special development of Iranian, and which is used in Avestan
without any such differentiation. The third Avestan word karana- is a daevian word which
occurs nowhere else in Iranian and it corresponds to Skt. karna- "ear". This word in Avestan
can be accounted for as a loanword from Proto-Indoaryan, and its use as a daevian word is
due to the fact that it belonged to the language of the Joeva-worshippers. The same is
probably the case with daevian as(i-) "eye", as opposed to ahurian doiOra- and neutral
casman-, although the root appears in Avestan (aiwydxS-). Of the two words for "son" in
Indo-Aryan, sunu- and putra-, Iranian has only preserved the second, with the exception of the
Avestan daevian word hunu-. Here again we can see a loanword from Proto-Indoaryan. The
initial h- of Avestan presents a problem here, as it does in the river names haraxvaiti- and
hindu-. A possible explanation would be that the Proto-Indoaryan language of eastern Iran
(or part of it) in this period had undergone a change of s- to /?-, similar to that in Iranian.
There is a parallel common change in the development of / to r which is shared by the
Rgvedic dialect of Old Indo-Aryan and Iranian. Another example of this change in Gray's
material is found in gah- "to eat" (daevian) corresponding to Skt. ghas-. Another loanword
from Proto-Indoaryan is no doubt to be seen in garaSa "house" (daevian, as opposed to

" W. B. Henning, Zoroaster, Oxford, 1951, 45.
  L. H. Gray, "The 'Ahurian' and 'Daevian' vocabularies in the Avesta", JRAS, 1927, 427-41.
                                   THE PROTO-INDOARYANS                                     133

ahurian nmdna-), derived from Proto-Indoaryan *grdha- in a more original form than that
seen in Skt. grha-.
      Although many of the daevian words originate quite differently (e.g. zbarada- "foot"
is a derogatory term derived from zbar- = Skt. hvar- "to go crookedly"), there is no doubt
that there is a nucleus of these words which is to be explained, following Gray, as above,
and this constitutes another proof of the existence of a Proto-Indoaryan population in the
territories of eastern Iran controlled by the Iranians. In these territories a bilingual situation
must have prevailed for some time, with two languages current which were to a considerable
extent mutually intelligible. Although they form only a part of the daevian vocabulary it
is possible that the Proto-Indoaryan words of the type illustrated above formed a starting
point for the system of the two opposing vocabularies.
      The historical conclusions reached by Gray on the basis of this material are in essence
the same as those expressed in the present article, and so it is appropriate to quote the
relevant part of his article (op. cit., 439):
      "From Central Asia, through the open reaches east of the Caspian which have given
access to the Iranian plateau to conqueror after conqueror, there came at an undatable
period an invasion (or a series of invasions) of a people who spoke an Indo-Iranian dialect
which represented the Indo-European hard sibilant by s, and who termed their deities
*deivos. In their turn these invaders were gradually driven south by another invasion (or
series of invasions) by a kindred people (or peoples) who had changed the hard sibilant
s to h, who called their divinities ahuros, and who had, as kindred peoples often have,
somewhat different vocabularies. These /j-peoples (the later Iranians) gradually expelled
the ^-peoples until the latter finally made their way to India.
      "In their contacts the /j-peoples adopted some of the words of their j-enemies, but
since they already had corresponding terms of their own (e.g. us- as contrasted with karna-,
ahura- as opposed to deva-, etc.) they had no need to add them to their vocabulary except
as referring to their foes. Hence, these terms of the s-peoples, expelled or conquered,
formed no real part of the vocabulary of the A-peoples, as they did of the word-stock of the
i-peoples, and the former employed them only in a derogatory sense, applying them solely
to beings, human or superhuman, whom they hated and despised as hostile and malign".
      This statement needs correction only in three points. In thefirstplace, as already pointed
out, the Iranian word for god was not ahura- but baga-. Secondly it has to be insisted that
the prime responsibility for the Proto-Indoaryan daivas coming to be regarded as evil was
Zoroaster's, and it was not just some natural impersonal development. The third point
is the idea that the Indo-Aryans migrated to India because they were driven out of their
former habitat by the Iranians. That migration, which is associated with the destruction
of the Indus civilization, is far too early for such a theory to be plausible, and such a
powerful programme of invasion and conquest could only have been undertaken from a
secure base. This means that the Iranian occupation of eastern Iran is to be ascribed to a
period after those extensive migrations had been completed, and the "5-Aryans" whom the
Iranians came across were those who had remained in the territories from which those
migrations took place.
      The Avesta has no ethnic term to denote the Proto-Indoaryans, which is not surprising
since both they and the Iranians called themselves Aryans and spoke closely related languages.
134                                      THE PROTO-INDOARYANS

The opposition between the two sides is always spoken of in religious terms between
the Mazdayasnas, worshippers of Ahura Mazda and followers of Zoroaster on the one
 hand, and the Daevayasnas (which can include Iranians following the daeva religion) on
 the other. The conflict is one which is to be terminated by the defeat or conversion of the
 latter. Opposition to daevas and daevayasnas is ascribed in the Yasts even to ancient and
 legendary kings such as Yima and Haosyarjha. On the other hand references to wars against
daevayasnas undertaken by Vistaspa (Yt. 5.109, 113; 9.30-31), Jamaspa (Yt. 5.68-70), and
Vistaru of the Naotara family (Yt. 5.76-77) are likely to have a historical basis.
      In spite of the efforts at conversion on the part of the Mazdayasnians the worship of
the daevas continued to exist for a considerable period, much as later the Mazdayasnian
religion itself continued to exist in Islamic times. Christensen24 has drawn attention to a
number of passages in the Vendidad which make this clear, and which also illustrate the
disadvantaged position in which such communities of daevayasnas were placed. Thus the
faithful are encouraged to take possession of the lands, waters, and harvests of the daevayas-
nas (Vd. 19.26). In another passage (Vd. 7.36^0) it is laid down that one wishing to practise
the art of medicine should first try out his skill on the daevayasnas, and if they survive his
attentions, he should then be allowed to practise on Mazdayasnians.
      The survival of the cfa/va-religion over a long period is also attested by the daiva-
inscription of Xerxes. There has been much speculation as to identity of the daivaddna
which he had destroyed, but in view of the very precise meaning of the word daiva {daeva)
in the Iranian Zoroastrian tradition, namely that of gods so named by their worshippers,
it is best to adopt the opinion of Christensen26 that the old gods of the daevayasnas are
referred to. The identity of the daivaddna concerned, which would have been very interesting,
is unknown but it is of great importance in attesting the continued existence of the con-
demned religion in Iran so long after Zoroaster.
      The Avestan evidence so far discussed concerning the Proto-Indoaryans has dealt
with eastern Iran, but iii view of the appearance of Proto-Indoaryans in the Near East,
one would expect that they had also a base, or bases, in western Iran from which the
invasion of the Hurrian country was mounted, just as their migration to India depended
on their base in eastern Iran. The geographical horizon of the Avesta is almost exclusively
eastern Iranian, but it does have some references which indicate the presence of Proto-
Indoaryans in northern central Iran, bringing them within striking distance of the Near
East. These are the references which occur from time to time in the Avesta to the Mazanian
daevas. The adjective in question (jndzainya-) is derived from *mdzana-, the name of a coun-
try which happens not to occur as such in the Avesta, but whose location is indicated by
the fact that it has always been known to be connected with the country later known as
Mazandaran, i.e. the territory between the southern shore of the Caspian sea and the
Alburz mountain range. In the later tradition this figures prominently as a region hostile
to the Iranians and as a notorious home of Devs. The presence of daevas in Mazana indicates
the presence of daeva-worshippers, and since we have seen that the daeva-worshippers were
the Proto-Indoaryans, we can conclude that the Avestan references to Mazanian daevas
indicate their presence also in this region.
     A. Christensen, Essai sur la demonologie iranienne, Copenhagen, 1941, 37 ff.
     op. cit., 46.
                                  THE PROTO-INDOARYANS                                  135

     Although the names Mazana and Mazandaran are connected, they are not identical,
since the latter contains an additional element not present in the former. It has been
suggested that this extra element is to be connected with NPers. dar "door", but a more
likely explanation is got, in view of the mountainous nature of the country, if we see in it
Iranian dara- "ravine" (for which see H. W. Bailey, op. cit., 115). This distinction enables
us to give to Mazana a wider sense than that of Mazandaran, since before the Iranians
moved westward, we would expect the Proto-Indoaryans who were in control of central
northern Iran to have occupied also that territory to the south of the Alburz mountain range
which later became Media.
     The Mazanian daevas are sometimes mentioned by themselves in the Avesta, but more
often they are coupled with the dn//-adherents {drvant-) of Varana. For instance Haosya^ha
prays (Yt. 5.22) that he may slay two thirds of the Mazanian daevas and of the Varanian
</rw/-adherents (ya9a azam nijandni dva drisva mdzainyanqm daevanqm varanyanqmca
drvatqm). In this case (as opposed to the case of mdzainya-) the basic word varana- does
 occur in the Avesta as the name of a country (Yt. 5.33; Vd. 1.17), but there has been
considerable doubt about its identification. This problem seems now to have been satis-
factorily solved by W. B. Henning (BSOAS, XII, 1947-8, 52-3), who, basing himself on
the order in which this name appears in the list of countries in Vd. I, has identified it with
Skt. Varnu- and Gk. "Aopros (Buner). This conclusion is of great interest, since it enables
us to see what is the precise significance of the juxtaposition in the Avesta of the Mazanian
daevas and the Varanian <//-w/-adherents. We may take Varana to refer (as pars pro toto)
to the Indo-Aryans in general to the east of Iranian-occupied territory, and Mazana to the
still unconquered Proto-Indoaryans to the west. Between these two, who would originally
have formed a continuum, a broad wedge of territory has been taken over by the Iranians,
who now have two hostile frontiers separating them from their enemies, the worshippers
of the daevas. This is the significance of the above mentioned cliche concerning the Mazanian
daevas and the Varanian drvants, and since drvant- is by definition one who worships
daevas, the two terms refer in effect to the same thing.
      Whereas Mazana, under the alias of Mazandaran, continued to be famous in Iranian
legend and tradition, Varana was eventually forgotten, so much so that the adjective came
to be connected with Pahl. Varan, the demon of concupiscence. Even some modern inter-
preters have taken it in this sense, but the clear parallelism with mdzanya- in the passages
where the two are mentioned together shows that the name of the country must be referred
      The earlier history of the Zoroastrian religion was confined to eastern Iran as delimited
by the above-mentioned frontiers, but eventually the defences of the Proto-Indoaryans to
the west were overcome, and this was followed by massive Iranian immigration into central
and western Iran. In view of the extreme hostility of the Iranians towards the "Mazanian
daevas" this westward movement is not likely to have taken the form of a gradual and
piecemeal infiltration, but of a deliberately organized military campaign and crusade
against the heathen, and it is not unlikely that a reflection of these events is to be seen
in the later legendary account of the wars against Mazandaran in the Shahnamah, even
though these have been set back in time to the reign of Kai Kaiis. As a result the Medes
became masters of central northern Iran, and instead of Mazana we hear in the future only
 136                                    THE PROTO-INDOARYANS

 of a truncated Mazandaran occupying the mountainous territory to the south of the
 Caspian sea.
      The name Mazandaran would indicate that it was to this mountainous region that
the worshippers of the Mazanian daevas retreated after their defeat by the Iranians. Since
it is not to be expected that the survivors of Mazana would in a short time disappear
altogether from the scene, the question arises whether they can be identified with any people
mentioned later in the classical sources. The people most likely to be so identified are the
Mardi (Gk. MdpSoi and "A/jLapSoi) because they are found in possession of that same
mountainous territory which is later known as Mazandaran. It was in this region that there
was a clash between the Mardians and Alexander (Quintus Curtius, VI, v, 11 if.). In addition
to this region the Mardians are also reported as living in two other regions, namely in the
vicinity of the Persians (these are mentioned by Herodotus, I, 125 and Aeschylus, Persae,
992) and also in the direction of Armenia.26 This distribution is interesting since it is con-
sistent with their having been displaced in various directions as a result of the Iranian advance
westwards. Classical authors have noted the distinction between these people and the
Persians (cf. Quintus Curtius, V, vi, 17: Mardorum gentem bellicosam et multum a ceteris
Persis cultu vitae abhorrentem), and although their distinctive life style was no doubt largely
conditioned by the mountainous nature of the territories they inhabited, there is probably
also, from what the classical authors say of them, an ethnic difference involved. As the
immediate neighbours of the northern Mardians have the name Anariacae (Strabo, XI, 8, 8),
the Mardi themselves are not likely to have been non-Aryans, and in view of the position
they occupy, and their distinctness from the Medes and Persians, there exists the possibility
of identifying them as the survivors of the Proto-Indoaryans of central northern Iran.
     In dealing with the history of the Proto-Indoaryans as predecessors of the Iranians in
Iran it is necessary to say something about the date of Zoroaster, partly because the com-
monly accepted date, c. 600 B.C., does not fit in with the theories proposed above, and
partly because if the date of the prophet can be roughly ascertained it will provide a chrono-
logical framework into which the events described can be placed. As regards the first point
it is obvious that the period of the later Median empire cannot be the time when the
Iranians first abandoned the cult of the Proto-Indoaryan daevas. This must have happened
much earlier, and not too long after the first occupation by the Iranians of eastern Iran.
     The usually accepted date of Zoroaster is based on a tradition dating from Sasanian
times which places him 258 years before Alexander. In spite of the arguments that have been
advanced against the creditworthiness of this tradition, it has continued to be accepted by
the majority opinion, largely because certain very distinguished and very influential scholars
have adopted it, notably A. Meillet in 1925 and W. B. Henning in 1951.27 Besides this date
an even later one has been canvassed, based on the identification of Zoroaster's patron
ViStaspa with the father of Darius.
     On the other hand many scholars have argued for a much earlier date, and the fact
that the later Iranian tradition is totally ignorant of the Median empire and almost totally
ignorant of the Achaemenians (and what it does know comes from the Alexander Romance)

" Pauly-Wissowa, Realencyclopadie der classischen Altertumswissenschaft, XIV, 1650.
   A. Meillet, Trois conferences sur les Gathas de I'Avesta, Paris. 1925; W. B, Henning, Zoroaster.
                                          THE PROTO-INDOARYANS                                             137

has justifiably been used as an argument against the reliability of this tradition.28 While
it is not known how this precise figure came to be adopted, the fact that it is so low can
be attributed to the ignorance of the later Zoroastrians about the history of the period be-
tween Zoroaster and Alexander. The authenticity of the later Zoroastrian tradition is
obviously not so well established as to outweigh any other arguments that might be adduced
in favour of an earlier date.
      One argument that has often been advanced in favour of an earlier date is the antiquity
of the language of the Gathas, which few would deny to be on the same level as that of
the Rgveda. The discrepancy between this and the proposed date of Zoroaster about 600 B.C.
is somewhat perfunctorily explained away by both Meillet and Henning with the statement
that languages evolve at a different rate. This might have to be accepted if it were known for
certain that Zoroaster lived c. 600 B.C., but it would still be a most remarkable phenomenon,
and one difficult to account for. Leaving aside this date the normal conclusion would be
that since the language of the Gathas is comparable in antiquity to that of the Rgveda, its
date would be expected to be in much the same period.
      Although the date c. 600 B.C. for Zoroaster is the commonly accepted view, there
have in recent years been some objections to it in various quarters, and those who do object
propose an earlier date.29 The question to be decided is how much earlier. The answer
to this question can be got by combining the historical information in the Fravardin Yast
with dates available from ancient Near Eastern sources. This Yast contains, in addition to
legendary material, a sort of gazetteer of the names of those individuals who had deserved
well of the Zoroastrian faith during the early period of its history, and these names can be
taken as largely historical. At the end of this section (Yt. 13, 143-4) there is a list of the
countries in which the Zoroastrian religion had spread: "We reverence the fravasis of the
followers of the religion (male and female) belonging to the Aryan, Turian, Sairimian,
Sainian, and Dahian lands". This enumeration is very interesting, since it shows that the
religion had spread from the Aryan country in the narrow sense, which we may take to
include the old Iranian territory, known as Aryana Vaejah, and the territory more recently
acquired in eastern Iran from the Proto-Indoaryans in a northerly direction, so that it was
also flourishing among such semi-nomadic and Scythian-type tribes as the Dahae and the
Sarmatians. On the other hand there is no mention of any expansion to western Iran, or
of the Medes and Persians. It cannot be argued that the term airya here might include
those western tribes, since this statement has to be considered in connexion with that
discussed above about the Mazanian daevas. In this text also there is a reference to with-
standing the Mazanian daevas and the Varanian drvants (Yt. 13, 137), which makes it clear
that as far as this text is concerned central and western Iran are still outside the Iranian and
Mazdayasnian fold. We can only conclude from this that when this statement was formulated,
the Iranians had not yet moved to the West, and that the Median and Persian sections of
the Iranian nation had not yet been constituted. This conclusion is backed up by the fact
that among all the considerable number of geographical references which are to be found in
the old Ya§ts the one solitary reference to central western Iran is the above mentioned adjec-
tive mdzainya-. All the other references are confined to eastern Iran or Central Asia, and
     e.g. by Edward Meyer, KZ, XLII, 1909, 1-2. Cf. also J. Duchesne-Guillemin, La religion, 136.
     For arguments for an earlier date see F. B. J. Kuiper, ///, v, 1961, 43, and the authorities quoted there.
JRAS, 1973, 2                                                                                                11
138                                     THE PROTO-INDOARYANS

taking them into account together with the detailed points just mentioned we may conclude
that the Iranians had not yet expanded beyond these two areas at the time when not only the
Fravardin Yast, but also the other principal Yasts were composed in their present form.30
Other points to be considered in this connexion are the fact that many of the geographical
names mentioned in the Yasts (e.g. Muza,31 Raozdya, etc.) are never heard of again, which
would be consistent with an early date, and the fact that the majority of the numerous
proper names listed in Yt. 13 are not found later in Iranian nomenclature. This would
be difficult to account for if the date of this text was towards the close of the Achaemenian
empire, as it would have to be on the assumption of c. 600 B.C. for Zoroaster's date. Also
unaccountable would be the almost total absence in this text of Iranian names known to
have been current at that time.
      The conclusion that the old Yasts and in particular the Fravardin Yast are to be dated
before the migration of the Iranians to central and western Iran provides a basis for calculat-
ing the date of Zoroaster, since we can begin from the earliest mention of the Medes and
Persians in the Assyrian annals, which occurs in the second half of the ninth century B.C.
(Parsua 844 B.C., Madai 836 B.C.). As these references are likely to be somewhat later than
the time of the actual migration itself we would probably not be going too far back if we set
a provisional date of c. 900 B.C. for the movement of the Iranians westwards. At some time
before this the principal old Yasts, including the Fravardin Yast, were composed, and it is
this latter text which provides the basis for further calculation, since it contains information
about the early history of the Zoroastrian religion. Of course, in general the difference in
language between the younger Avestan of the Yasts and GS0a-Avestan implies a consider-
able lapse of time, which is not likely to be less than 200 years. But in Yt. 13 we have
some more detailed material to go on, and the large number of names mentioned as
supporters of the faith can be held to imply no less a period. The text is particularly in-
formative about the family of Saena who was considered to be the most important of the
successors of Zoroaster. He himself is said to be the son of Ahum.stut and the names of
three individuals, representing three successive generations after Saena, are given: Ziyray,
Vitkavay, and Utayutay. We thus have five successive generations, which, allowing thirty
years for a generation, would cover 150 years of the history of the Zoroastrian religion
preceding the Fravardin Yast. Another consideration is the later tradition according to
which Saena lived between the years 100 and 200 of the Zoroastrian religion. That quite
a considerable period had elapsed between Zoroaster and Saena is indicated by the statement
of the Yast itself (Yt. 13, 97) that Saena was the first teacher of the Zoroastrian religion

   The interpretation of bawray- in Yt. 5.39 as "Babylon", accepted by Bartholomae, should be abandoned
     in favour of Nyberg's opinion that the reference is to a "beaver-land", the beaver being an animal
     sacred to Aradvl Sura Anahita. There is no conceivable reason why this legend, one of the most ancient
     of the Iranian legends, should be connected with Babylon, and it is unlikely that the Iranians had ever
     heard of Babylon at the time when it arose. A difficulty would occur if the Raya mentioned in the
     Avesta were the Median Raga mentioned by Darius, but Gershevitch has cogently argued (JNES,
     XXIII, 1964, 36-7) that ZarathuStrian Raya of the Avesta is to be located in the East, and that Median
     Raga was named after it. The process of transferring names from east to west is common, and illustrated
     by such examples as Hara Baraz applied to the Alburz and Caliasta to lake Urmiya.
   Possibly to be connected with Skt. Mujavant- (Eilers). This would be a frontier region and it is interesting
     that DaStayni, who belonged to this country, appears to have an Indo-Aryan name. The word agni- is
     foreign to Iranian, and the first member, dasta-, can be interpreted as a past participle passive from the
     Vedic root das- "to worship".
                                          THE PROTO-INDOARYANS                              139

to have a hundred pupils, from which we learn that there had been previous teachers in the
period since Zoroaster, and the religion must have had a considerable development for
such a number to be possible. If we take the hundred years between Zoroaster and Saena
and add them to the period represented by the four generations from Saena to Utayutay,
we arrive at a period of 200 years at the very minimum between Zoroaster and the com-
position of this Yast. Adding this to the date 900 B.C. suggested above as the time of the
movement west of the Iranians, we obtain 1100 B.C. as the lowest possible date for the
founding of the Zoroastrian religion. Since, however, the Iranian movement to the west
might have been somewhat earlier than 900 B.C., and since there was probably some
interval between the composition of the Yast and that event, and also in the other direction
between its composition and the time of Utayutay, the date could be a good deal earlier.
The conclusion then is that Zoroaster's date must be at least half a millennium earlier than
that generally accepted, and quite likely it could be as much as 600 years earlier.
     Before Zoroaster a considerable period must have elapsed since the occupation of
eastern Iran by the Iranians, during which the Proto-Indoaryan religion was still respected
by the new rulers. According to later tradition the famous pre-Zoroastrian Kavi dynasty
was connected with the region of Seistan, and this would appear to be confirmed by the
geographical details in Yt. 19.65 ff.32 If we allow 150 years for this dynasty as Christensen
does, 33 we can get some idea of the length of the period involved, since according to this
evidence they must have been ruling in what was previously Proto-Indoaryan territory.
We may conclude the Iranian conquest of eastern Iran was an event that took place not
later than the fourteenth century B.C, and it thus coincided with the period when the
Proto-Indoaryans had their furthest extension westwards. At this time large numbers of
Indo-Aryans had migrated into India, and as a result of these two migrations the position
of the Proto-Indoaryans in their original base in eastern Iran must have been considerably
weakened, thus providing the Iranians with an opportunity to move in and take over.
     The Proto-Indoaryans did not maintain their position in the Near East for very long,
and after the period of the Mitanni kingdom they disappear from the record in this region.
An indication of their presence in the neighbouring region of Persian Azerbaidjan is perhaps
indicated by the name of Lake Urmiya (now Rezaiyeh). Although this name does not occur
before the Islamic period it is thought to be much more ancient because of the mention in an
Assyrian source of a place called Urmieate in its vicinity.34 Since the Proto-Indoaryans
must have been in north-western Iran in order to reach the Mitanni country, it will not be
unreasonable to suggest an Indo-Aryan etymology for this name, if one is available. Such
an etymology is available if we compare Skt. urmi- "wave" and urmya- "undulating, wavy",
which would provide a suitable descriptive name for the lake. This is a case where the
phonology of Indo-Aryan and Iranian have diverged quite widely (cf. Av. varami- "wave")
and it is interesting that the name of this lake, if the above etymology is correct, should go
so clearly with Indo-Aryan.
     The situation in this region in the first half of the first millennium B.C. is one in which
most of its northern part is under the control of Urartu. To the south-east of the lake lies
   cf. A. Christensen, Les Kayanides, Copenhagen, 1932, 22-23.
" op. cit., 34.
34                          1
     cf. Encyclopaedia of Islam , IV, 1032.
140                               THE PROTO-INDOARYANS

the kingdom of Manna, the names of whose kings are not of Aryan appearance (or other-
wise identifiable). In addition a large number of smaller cities and principalities are men-
tioned in Assyrian sources. Among this onomastic material some Aryan names turn up,
which could be interpreted as Iranian, but since the Proto-Indoaryans had penetrated so
far west they might also be explained as emanating from this source. Such was the opinion
of P. Kretschmer concerning the name of Bagdatti of Uisdis who was defeated and killed
by Sargon in 716 B.C.35 If this spelling of the name is reliable it goes better with Indo-Aryan
Bhagadatta- than with Iranian Bagaddta- on account of the double -U-. Another Aryan
name from this region, mentioned as early as 827 B.C., is Artasari and this name also may be
Proto-Indoaryan rather than Iranian. It is not unlikely that among the miscellaneous popu-
lation of this region some sections of the Proto-Indoaryans remained at this period, and it
is possible that if further documentation becomes available more names of this kind may
come to light.
     The presence of Indoaryans, or in the terminology used in this article Proto-Indoaryans,
in the Near East was puzzling on account of the later geographical distribution of the
Indo-Aryan and Iranian languages, the former being to the east in Northern India, and
the latter to the west in Iran. This distribution resulted from the events sketched above, but
at an earlier period the relative distribution of these two moieties of the Aryan language
group was between north and south. At this time the north-western part of the Indian
subcontinent was only a part of the territory occupied by the southern half of the Aryans,
and from here it stretched into eastern Iran and across the northern part of central and
western Iran to within striking distance of the Near East. To the north, in Central Asia,
lay the Iranians, who, beginning about 1400 B.C., moved south and by degrees took over
the territory previously occupied by the Proto-Indoaryans.
     KZ, LV, 1927, 100.

To top