Witczak proposes (in 'The Pre-Germanic Substrata and Germanic
Maritime Vocabulary) an IE Germanic substrate with these properties:
PIE p t k > bh dh gh
PIE b d g > p t k
PIE nasals deleted before consonants
PIE w disappears before *y and front vowels
PIE o > a after w and m
PIE p b bh > m beween vowels
and example
PIE *weh1wérnah2 "squirrel"
Gaulo-Latin vi:verra
Welsh gwiwer
Breton gwiber
Gaelic feòrag
Lith. ve.verìs, voverìs
OPruss weware
OSl *veverIka
Polish wiewiórka
Czech veverka
>
Germanic substrate *ikwernan-
ON ikorni
Norw. ikorna
with folk etymology to Gmc *aik- "oak" > *aikwernan-
OHG aihhurno
Dutch eekhorn
OE a:cweorn-
I think Danish væver, ODa viver "nimble, agile" belongs to the first
form. Dansk Etymologisk Ordbog is unsure of where it belongs and
suggests also Engl. waver (?)
It seems to me PIE *weh1wer- must be somehow related to PIE *gWi-wo-
"alive" or *gWi-gWo- as I believe Piotr wrote it (to show it was
related to PIE *gWo-gW- "cow". If that is so, then *weh1wer- is
itself a substrate term, which means either that there are _two_ IE
substrate layers in Danish (one each for 'væver' and 'egern'), or
that the "alive" root occurred in a non-IE substrate in Germanic,
which means it wasn't native in PIE.
Torsten