residing in the south Balkans. They may have retreated from there under
pressure from Thracian tribes coming down from further north, from beyond
the Dnieper.^24 It was formerly assumed that Phrygians and Thracians were
closely related, and compound adjectives such as ‘Thraco-Phrygian’ used to
be freely used in various connections. In fact there is no special affinity
between the two. It is to be observed that Phrygian, like Greek, was a centum
language, whereas Thracian was satem, like Slavonic and Iranian.
Armenian too is a satem language, and not closely related to Phrygian, even
though both belong to the Graeco-Aryan group. In historical times the
Armenians were located far away to the east of the Phrygians, and Herodotus
(7. 73) was told that they were a Phrygian colony. Perhaps someone had
observed a similarity to the Phrygian in their language or culture. But if we set
aside this dubious western connection, their geographical situation is much
easier to understand on the hypothesis that they came there by way of the
Caucasus. They first appear in history in the seventh century ; there is no
sign of them earlier, despite our having Urartian inscriptions from the area
from the immediately preceding centuries. Their arrival may be connected
with the burning of the main Urartian fortresses in around 640.^25 This
was just at the time when the Cimmerians had come down from north of
the Caucasus and were causing havoc throughout Asia Minor. There seems
much to be said for the view that the Armenian influx was part of the same
movement.^26 If so, the Armenians had previously lived in the north-east
Pontic area, in the immediate neighbourhood of other satem-speakers such
as the Scythians (who drove out the Cimmerians according to Herodotus
- 15).
The Iranian and Indic languages are closely related to each other, and must
be traced back to a common Indo-Iranian or Aryan. The period of Indo-
Iranian unity may be put in the late third to early second millennium, and its
territory located north and east of the Caspian Sea. From an archaeological
point of view it seems a good fit with the Andronovo culture which developed
in northern Kazakhstan between 2300 and 2100 and later spread southwards
and eastwards.
Indic was already differentiated from Iranian by the sixteenth century,
when a horde of Aryan warriors established themselves as rulers of the land of
Mitanni in north Syria. Their personal names, their gods, and other evidence
of their speech show that they were Indic-speakers. We may suppose that
Indic had been the dialect of the southern Aryans, and that they had made a
major southward movement down the east side of the Caspian. Then, faced
(^24) Cf. Sergent (1995), 423–5. (^25) Paul Zimansky in Drews (2001), 23 f.
(^26) Cf. Feist (1913), 65; Schramm (1973), 164–217.
Introduction 9