Pre-PIE as a PIE substrate?

[X99Lynx at aol.com]

Way, way back on 14 Mar 2000, g_sandi at hotmail.com wrote:

<<What I find reasonable as a theory is Gimbutas's: the farming population of
much of Europe switched language because it was conquered by a horse-riding,
warlike elite who imposed its hero-worshipping ideology on it. A bit like
what the Hungarians did to the Slavic and other inhabitants of Hungary after
895 AD, or what the Turks did to the various inhabitants of Anatolia in
post-Classical times. The basic ethnic composition of both of these areas
remained pretty much the same, I believe, yet there was a language shift,
with plenty of substratal influence.>>

Hopefully the statue of limitations hasn't run out on this particular
exchange.

Putting aside for now the factual basis for horse-riding hero-worshippers
carrying IE into "much of Europe," there's a point here that raises a
linguistic question that might provide linguistic evidence of what actually
happened back then.

If "the farming population of much of Europe switched language" to IE (by
conquest or otherwise) from something other than IE, then of course there may
be that "substratal influence" to find.  Or, by analogy with the examples
given, "plenty of substratal influence."

But there is an interesting twist to this substrate issue proposed by none
other than Cavalli-Sforza:

"It should be noted that [Renfrew's] hypothesis is not incompatible with
Gimbutas' hypothesis.  It is perfectly possible that neolithic farmers
brought early Indo-European languages not only to Europe but also to southern
Russia, together with agriculture, and that their descendants, who developed
nomadic pastoralism in the Kurgan steppes carried their languages, which was
still of Indo-European origin but transformed from the original....  If this
fusion of hypothesis from Gimbutas and Renfrew is correct, the
proto-Indo-European language reconstructed by linguists on the basis of [sic]
modern Indo-European languages must be much closer to that which was spoken
in the Kurgan area some five or six millennia ago, than to the
pre-proto-Indo-European spoken by Anatolian farmers."  L. L. Cavalli-Sforza,
The Spread of Agriculture and Nomadic Pastoralism, from The Origins and
Spread of Agriculture and Pastoralism in Eurasia, edited by David R. Harris,
Smithsonian Institution Press (1996), 66.

Putting aside for the moment the question of how much Kurgan actually ever
came to most of Europe, the quote above suggests that if there is a substrate
to look for in early European IE languages -- as described above -- it might
consist of "pre-proto" IE languages.

My question is: what would such a substrate be like?  What would one look
for?  Thinking of other examples of where IE has folded over on itself, so to
speak -- where one IE language exhibits a substrate of an earlier IE language
-- where would one look for such a "pre-proto" substrate in PIE?  How would
one separate "pre-proto" features from "proto" features, since both would
ultimately be of the same origin?

Regards,
Steve Long