Re: Divergence vs. convergence (was: Witzel and Sautsutras)

At 5:14:04 PM on Monday, October 22, 2012, Tavi wrote:

> --- In cybalist@yahoogroups.com, Jörg Rhiemeier <joerg_rhiemeier@...>
> wrote:

>> Sanskrit is one of the most conservative IE languages to
>> be sure, but its conservatism should not be overrated,
>> and the history of the standard model of PIE is a history
>> of emancipation from the Sanskrit model - more and more
>> features of Sanskrit were recognized as innovations of
>> the Indic branch. And now it turns out that we have to
>> posit a quite different Early PIE to account for the
>> divergent features of Anatolian.

>>> Not only "earlier" (in diachronical terms) but also
>>> "diverse" (in diatopical terms). Thus we've got (al
>>> least) two different "PIE"s. In my own model, the IE
>>> family is the result of the superimposition of several
>>> (proto-)languages due to contact and replacement
>>> processes over millenia. The classical genealogical tree
>>> model is simply inadequate.

>> Would you mind giving evidence for that? There are
>> well-established regular sound correspondences linking
>> the various Indo-European languages with each other, and
>> the best way of accounting for them is to posit a common
>> ancestor language that gradually diversified and broke
>> apart.

> You're contradicting yourself, as you previously said that
> "we have to posit a quite different Early PIE to account
> for the divergent features of Anatolian". So you're
> implictly recognizing a single proto-language doesn't work
> at all.

Nonsense. He's simply arguing for an early split between
Anatolian and the rest of IE.

>> If Indo-European was actually the outcome of "contact and
>> replacement processes", one would expect this to show in
>> numerous inconsistencies in the sound correspondences.
>> Where are they? Granted, there are a few words which look
>> "wrong" and are probably loanwords from other IE
>> languages, but if Indo-European was a convergence area,
>> such irregularities would have to be far more numerous,
>> especially in the derivational and inflectional
>> morphologies of the IE languages.

> As regarding morphology, there're some isoglosses such as
> -r passive suffix, the augment, etc. There're also many
> interesting phonetic isoglosses as regarding stops and the
> so-called "laryngeals", denasalization of initial nasals,
> an so on.

A characteristically Tavi-esque non-answer, throwing around
a lot of terminology without actually saying anything
concrete. I believe that Jörg has had some experience of
them in the past on the Zompist board, so it probably won't
come as any surprise.

>> Nobody denies that most Indo-European languages contain a
>> large number of lexemes that do not have reliable PIE
>> etymologies and are thus likely to be loanwords from
>> substratum languages, but that still does not make
>> Indo-European a convergence area.

> There's no simple way the more than 2,000 "roots"
> reconstructed by some IE-ists could belong to a single
> (proto-)language.

And your evidence for this unsupported opinion is?

Brian