Dating the final IE unity (was: Re: GREEK PREHISTORY AND LANGUAGE)

[Larry Trask]

[ moderator re-formatted ]

Sean Crist writes:

>  With all of this background in mind, we can see the general ideological
>  framework which is leading Renfrew to seek the kind of solution which he
>  has done.  Renfrew wants to keep to an absolute minimum the number of
>  prehistoric migrations which we have to posit.  He'd probably prefer that
>  we not need to posit _any_ prehistoric migrations, but even Renfrew would
>  surely concede that the spread of a language over a large area implies a
>  population movement.

Actually, he doesn't, and he's quite explicit about it.

Renfrew allows language spread by wholesale population movement in just two
circumstances: the first settlement of previously uninhabited areas, and spread
by elite dominance.  Otherwise, he sees language spread as proceeding by
diffusion across established populations along with the spread of technology
and culture.  In particular, he sees the spread of IE as having proceeded in
this way.

>  Technology in agriculture, pottery, metallurgy, etc.
>  can plausibly spread by diffusion without a large-scale movement of
>  populations; but language (other than loan words) does not diffuse this
>  way.  Verb conjugations, etc. only spread when speakers move.

This is precisely the position which Renfrew attacks.  And, to be fair, it is
not difficult to point to cases in which prestige languages have spread,
displacing earlier languages, without massive population movements.

>  So what Renfrew is trying to do is kill two birds with one stone.  Suppose
>  we assume that the introduction of the agricultural cultural complex thru
>  Europe is the result of a population movement from Anatolia to the Balkans
>  during the 7th millenium BCE (which Renfrew accepts).  We're also forced
>  to say that the spread of the Indo-European languages implies a population
>  movement.  Renfrew's response is to try to collapse these two migrations
>  into one, thus keeping the number of migrations to a minimum.  If you're
>  assuming that it's a bad thing to posit prehistoric migrations, this is
>  the sort of solution you'd like to try for.

I think it's rather that Renfrew prefers to posit a minimum of population
movement and a maximum of diffusion.

>  Unfortunately, this solution cannot be made to work without ignoring a
>  huge amount of evidence.  An archaeologist here at Penn told me that he
>  was very "annoyed" at Renfrew for having put forward this view, and said
>  that if anyone with less than Renfrew's prestige had put it forward, it
>  simply would have been ignored as not worthy of consideration.  But
>  Renfrew's earlier contributions (i.e. to the notion of process
>  archaeology, to C-14 recalibrations, etc.) are so well respected that he
>  has to be answered.  He has been answered indeed.
>
>  Mallory (p. 164 ff.) discusses the whole issue at some length, and makes
>  the following arguments:
>
>  -Renfrew proposes that Anatolia is the homeland of the Indo-Europeans.  If
>  so, it is very odd the Hittite, Luwian, and Palaic existed in the midst of
>  languages (Hattic and Hurrian) which are not only not Indo-European but
>  which bear no typological resemblance to Indo-European. If there were any
>  language family which we might guess to be a sister of Indo-European, it
>  would be Finno-Ugric, which would argue for a Ukrainian homeland, not an
>  Anatolian one.

Well, as Mallory himself has pointed out at length, mainly in his article in
the first Blench and Spriggs volume, there are very many criteria which may be,
and have been, invoked in arguing for the location of a linguistic homeland --
some of them even mutually incompatible.  For an Anatolian homeland, the
presence of non-IE languages in Anatolia violates just one of these proposed
criteria, the exclusion principle.  But Anatolia is arguably supported by a
different principle, the linguistic relationship principle.

>  -There is clearly a substantial non-Indo-European substrate in Greek, both
>  in place names and in loan words.  This would be a bit surprising if
>  Indo-European speakers had been in the area since the beginning of the
>  Neolithic.

But Renfrew has expressly argued that many of these words are not substrate
words at all, but rather late borrowings into Greek long after Greece had
become Greek-speaking.

>  -Most importantly, placing the initial dispersion of the Indo-Europeans at
>  the beginning of the Neolithic around 6500 BCE in entirely incompatible
>  with the reconstructed Indo-European vocabulary.  Words such as yoke,
>  wheel, etc. are reconstructed for PIE, but this technology is not attested
>  until much, much later- namely, not much after 4000-3500 BCE, which is the
>  date which Mallory and others put forward as the final date of IE
>  linguistic unity.

This is exactly the point which troubles me the most.  But Renfrew has in fact
met it head on in his latest paper on the issue, presented at a symposium in
Cambridge last summer and due to be published in the proceedings early next
year.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk