Re: Limitations of the compartive method

On 2008-01-27 01:14, mkelkar2003 wrote:

> So PIE must have been a real spoken language then.

What else could it be? With the obvious reservation: _reconstructed_ PIE
is a set of hypotheses about a "real spoken language". It's obvious that
the reconstruction is neither complete nor perfect. You may compare that
to the museum reconstruction of a dinosaur. It gives you a general idea
of what the actual animal was like, but it doesn't breathe of run about,
and if it has spots or stripes, their shape and colour are part of the
artist's vision, not warranted by anything we can deduce from fossilised
bones.

> I have two diffrent responses for this:
>
> 1. Buying into your argument that PIE must have existed "Just as there
> was a language called Latin."
>
> sorry for the long link
>
> <http://www.arch.cam.ac.uk/~pah1003/loe/Eng/Papers/06_12_07_DataShow_PaulHeggarty_TreesOrWebsSplitsOrWaves.pdf>
>
> Slide 31: The heartland where we know Latin was spoken or written i.e.
> Italy or France historically are the "homeland."
>
> and slide 53: where "Indo-Iranian" occupies that position hence India
> is the homeland of this single language.

I suspect you've got the numbers wrong. I can't see any analogy between
(31) and (53). Anyway, diagrams like (31) are hardly useful in
pinpointing the homeland.

> 2. Not buying into your argument that PIE must have existed "Just as
> there was a language called Latin." Refer to slide 50 the green line
> on top of the Romance languages has a 100% posteror probablilty but
> the higher node immediately prceeding that has only 46% and 2 nodes
> before that only 44%.
>
> I am going with position 2!

[Sigh] It only means that the (Romance + Germanic) and (Romance +
Germanic + Celtic) clades are weakly supported by this particular
analysis (which everybody realises anyway). The support for Romance and
Germanic in isolation is 100 (which is again common knowledge). That's
why it's perfectly legitimate to speak of Proto-Germanic, Proto-Celtic
and Proto-Romance, but substantially less legitimate to speak of
"Proto-Italo-Celto-Germanic". The validity of Romance is obvious,
especially as the parent language is directly attested. The validity of
IE as a genetic unit may be less evident, but has nevertheless been
soundly established and everything we have learnt for the last 150 years
or so has only strengthened this validity.