Here is an example -- one of many that could be cited -- showing that
Lord Renfrew of Kaimsthorn has little idea of what is possible in
linguistics and what is not. In _Archaeology and Language_ (1987) he
considers the pros and cons of two hypotheses of Indo-Iranian origins.
Hypothesis A: Neolithic Aryas (an early southeastward wave of advance
from the Iranian Plateau and unbroken Indo-Aryan continuity in the
Indus Valley from the early neolithic to the Harappan period and
beyond).
Hypothesis B: Mounted Nomads of the Steppe (incursion of pastoralist
nomads from the Eurasian steppes into the Middle East and India).
----------
From the linguistic point of view ... the two hypotheses are very
different. The first implies that the Indo-European languages of Iran
and India and Pakistan derive from precursors in eastern Anatolia and
further east, just as the first Indo-European languages of Europe
derived at the same time from precursors in central and western
Anatolia. The successors of the western Anatolian languages are the
Indo-European languages of Europe; the successor of the central
Anatolian languages was Hittite; and the successors of the easter
Anatolian languages were thus the Indo-Iranian languages. We might
expect, then, a number of resemblances between Hittite and the Indo-
Iranian languages. On the other hand, the languages of eastern Europe
would originally have born [sic!] little affinity with the Indo-
Iranian languages, although convergences could occur as a result of
steppeland influences on Iran and India at a later date. The original
separation would have taken place by 6500 BC.
On the other hand, Hypothesis B implies that the relationship between
the Indo-Iranian languages and those of eastern Europe would be very
much closer, with a common origin around 4000 BC. There are some
indications which might support such a view, for instance the old
classification of the Slavic languages within the eastern or _satem_
group -- but little emphasis is placed these days upon this
simplistic distinction. [Renfrew 1987: 206-207]
----------
Leaving aside purely extralinguistic difficulties (note the
incredible time-depths), hypothesis A would require us to believe
that the pervading structural similarities and common innovations of
the `European' languages and Indo-Iranian are due to convergence
(according to Renfrew, these groups arose independently and cannot
have met till many centuries after their separation from Proto-
Anatolian). Since Renfrew speculates that about 5000 BC Indo-European
languages were already spoken from northwestern France to Pakistan
and north India, far-reaching convergence must be proposed for such
an impossibly vast linguistic area (bypassing Anatolia). Renfrew
neglects this question completely. He does recognise the necessity of
accounting for the close affinities shared by the language groups of
eastern Europe and Indo-Iranian, but again attributes them to
secondary convergence -- taking place at a time when the ancestors of
the Baltic, Slavic, Iranian and Indo-Aryan languages were already
distributed from the forest-steppe zone of eastern Europe to the
Arabian Sea coast. He is not very specific about the source of the
supposed steppeland influences or about their nature and the
mechanism of their spread. Renfrew feels free to invoke convergence
whenever he needs to explain shared innovations that, if his
scenarios are to be believed, must not be attributed to common
descent. He is not troubled by the question of what conditions should
be met for convergence to be possible, or how to distinguish between
the effects of areal diffusion and shared ancestry -- a crucial
problem in historical linguistics.
Renfrew seems to find hypothesis A more satisfying and waves aside
the linguistic arguments that strongly support hypothesis B as old
and simplistic. He evidently does not realise that the (justified)
rejection of the idea of a neat eastern/western split within the Indo-
European family does not mean that the satem innovations can be
explained away as vague areal effects. He asserts that `[the
geographical position of Tocharian] has undermined confidence in the
satem/centum distinction, which like most very simple typological
distinctions is now felt to be too simplistic to carry much weight
[p.66]'. Such terminology is unacceptably loose. The presence or
absence of the satem shift -- a rather special phonological
innovation -- is not a _typological_ distinction; nor is it very
simple.
Piotr
--- In cybalist@..., celteuskara@... wrote:
Apart from all this Aryan indigenism nonsense [!] I was intrigued to
see Renfrew's name pop up here. I wonder could anyone tell me the
nature of his mistaken assumptions? I'm an archaeologist y'see, and
have often criticised Renfrew's Anatolian model, without being too
aware of the precise details.