real proto-lang

[petegray]

> it's better to understand those proto-languages as some tools of comparative
> linguistics than real languages that existed at some point of time.

This is the "abstractionist" way of understanding reconstruction.  It has
also been applied to reconstructed phonemes.   "Abstractionists" would claim
that PIE *bh, *kw  etc are simple symbols denoting the set of relationships
between the reflexes of *bh, kw  etc in the attested languages.

Fine and dandy, but we must ask if something is lost.  The abstractionist
position appears (from the little I know) to be losing ground rapidly,
because of the advances in our understanding of PIE based on a more realist
approach, which treats our reconstruction as a real language.

Firstly there is all the debate - on this list and elsewhere - about
homeland, where and who and what pots and what grave customs and so on.  If
the reconstructed language is merely a "tool of comparative linguistics"
then these questions are closed down.

Secondly there are questions of how such a language would actually work.
Typology comes in here, but not only typology.   The "new sound" of PIE
would not have arisen if *bh etc had been seen purely as an abstract tool
for comparisons.

So while at one level I relate to your idea, at another I think there are
advantages in treating our reconstructions as if they were actual spoken
human languages - even while admitting the limits of our knowledge.

Peter