More than 25% but less than 30% seems to be the exact figure cited for
Germanic. I've seen attempts to suggest at least part of the Germanic
substratum is from another IE language, tho' this does not seem
particularly popular. The overwhelming consensus seems to be that (1)
the Germanic substratum is quite real, (2) covers some interesting
semantic areas IE itself was poor in (as with sea-faring terms), and
(3) most certainly is wholly assimilated into Germanic, following all
the usual rules for Germanic words and can be discerned only by
comparative studies to non-Germanic IE languages.
My reading always alludes to the problems in Germanic, but no one
seems to come right out and explain it all. I gather Germanic is most
likely the second oldest known split in IE (after Anatolic), and that
it seems to have enjoyed a long period of independent internal
development (i.e., unaffected by other IE languages), but nonetheless
had a close relationship with a non-IE language. The possibility that
the majority of the ancestors of proto-Germanic speakers had replaced
a native non-IE language with what became Germanic is real; a parallel
can be found in Central Asia, where (so I've read), if you scratch the
surface of the Turkic languages, you'll find an Iranian substratum.
Greek is messier. We know there is at least one non-IE substratal
language (what you call this language can start a fight). For an IE
language, the word might not be 'substratum', but superstratum or
adstratum, and we are likely talking of several such langauges, one of
which would seem to be Anatolian; there are also poorly understood
apparent connections with Illyrian, Phrygian and even Thracian.
Macedonian offers its own mysteries (is it Greek or is it something
else?). Greek also borrowed words as adstratal items, a la the current
relationship of American English and Spanish.
The possibility that Greek has been in Greece since about 3000 BCE is
actually possible to believe. This would make it quite nearly
autochthonous, in that Greece and Thrace were essentially depopulated
in the centuries before this date. The Mycenaean warrior elites, then,
would be intrusive non-Greek (but likely IE-speaking) elements, one
which was quickly assimilated into Greek.
--- In cybalist@..., george knysh <gknysh@...> wrote:
> There was an earlier (and presumably still open)
> discussion in this forum about whether Germanic
> contains as much as 30% non-IE material, or whether
> this is an inflated percentage. I remember Mallory
> ("In search...", 1989) mentioning that perhaps one-
> half of the old Greek lexicon could not yet be
> confirmed as IE, and that the % was even higher in
> Hittite. Has anyone attempted to make similar
> estimations with respect to other IE languages, and if
> so, what were the results (even if not fully corroborated)?