> The point of course is : how can you tell whether the
> specific characteristics of Slavic, those which make
> it a distinct group of IE languages, already existed
> at the time of Mycenaean Greek. Germanic for instance
> (if one considers the Grimm shift essential for its
> identification) likely did not yet exist at that time.
This is an interesting question. It has, of course, been considered before,
and it is the reason many linguists now speak of PIE as a cluster of related
dialects, rather than a single language. Your question reduces to whether
or not it is possible to think of PIE as a single undifferentiated language
at any stage (allowing for early removal of some dialects, if you wish -
like Hittite). Or must we think of PIE already containing the
differentiation that we see reflected in the full-blown language groups of
IE?
Vocabulary certainly does not seem to have been universally shared. Is it
easier to explain this by vocabulary loss, or by the concept of close
dialects with slightly different vocabularies (as we find in modern related
dialects)?
Morphology is not universally shared. Can we really derive all IE
morphology from a single origin, or is it easier to think in terms of
related dialects, with slightly different morphologies, as we see in modern
languages with dialects?
Which is more like real language, a dialect cluster, or a single uniform
undifferentiated speech?
There is a different (and much more trivial) question about labels. Is
pre-Germanic Germanic? Is pre-Slavic Slavic? Who cares? Call it what you
like. The real question is whether pre-Germanic and pre-Slavic were ever
absolutely identical. I'm arguing no.
Peter