On Sun, 14 Oct 2001 21:04:42 +0200, "Piotr Gasiorowski"
>Because assimilation at a distance (like dissimilation) is typically sporadic rather than regular (BTW, what particular Germanic forms do you have in mind?).
Pokorny gives OHG giwahanen, "erwähnen", pf. giwuog, MHG wüegen "to
remember", OIce. va:ttr "testimony" (*wahtaz), va:tta "testify", Oice.
o:mun "voice", OE wo:m(a) "noise" (*wo:hm-).
>One could with equal reason ask why *loikW- gives *also* expected *laihW-, or why we have Norse ylgr 'she-wolf'.
*seikw- gives both Germanic *si:f- and *si:xw-, Gothic has <auhns>
"oven". Reflexes of *(l)yepr-, *leip- and *wlp- are also found
outside of Germanic. So the results are indeed rather inconsistent.
Still, the fact that the phenomenon occurs only in a particular
subgroup of roots with reconstructed *kw, sharing no common phonetic
environment, with Germanic showing consistently reflexes of *p (with
sporadic *kw > *hw / *gw) and the other languages (Osco-Umbrian and
Brythonic excepted, as their evidence cannot be relevant) showing
reflexes of *kw (with sporadic *p), makes me feel rather confident
that we are dealing with a separate (and rather rare) PIE phoneme.
Feel free to ignore, but it just so happens that postulating *pWV (<
**pu) in some of these words also allows interesting parallels to be
drawn with extra-IE material. I gave some comparanda for *pWetwores,
the ones for *(l)yepWr.(t)/*(l)yepWn(t)- "liver" are equally
interesting: Dolgopolsky (1998) : *l(ä|e)p.A "spleen" > Ham.Sem.:
East Cush. Afar <ale'fu:> [pl. a'lef-it] "spleen" /// West Chadic:
Sura <l.`lap>, Kofyar <láp>, Montol, Angas <lap> "spleen" || Uralic
*läppV (or *leppV) > Finno-Ugric: Cheremis lep& ~ lep / proto-Permian
*lOp > Votyak <lup>, SW Votyak <lup>, Ziryene <lop>, <lopt->, Upper
Sïsola Zyryene <lOp>, Yazvian <lop> / Lappish **DapDe (by assimilation
from *lap-De with a suffix -De) > Norw. Lapp daw'de ~ daD've (etc.) //
Teryugan Ostyak <La.p&tne> / Hungarian <lép> /// Samoyedic: Forest
Nenets <Laps'&> ~ <rab_s'&> id. || Altaic: Tungusic: Orok <lipc^e>
"spleen". Cf. for Uralic Sammallahti (1988): PFU *dä(pd|pp)ä "milt",
PFP: *däpdä: Lapp <daw'de>, Cheremis <leep(e)>, Votyak <lup>, Zyryene
<lop>; PUg. *däppä: Ostyak *<leepet>, Hungarian <lép>.
In view of PIE *yekWr(t) (**l^epWnt-), I would suggest a PN
reconstruction *lipunt- or *liput-n-.