stlatos wrote:
> Oh? What about *gYonYxWeye+ > *kYaniyi+ > OE cennan?
Wrong. OE cennan 'beget, conceive' contains a different root, *g^enh1-!
What you could quote instead is the OE "preterite-present" cunnan, pres.
cann 'know' (Mod.Eng. can), but the Gmc. "neo-full" grade is secondary,
back-formed from the IE nasal present *g^n.noh3- (hence cunn-).
> You also
> didn't object to:
>
> *kWen.-x-tro+ > Skt khanítra- 'spade'
>
> even though the root appears mostly as *kWn.ax+:
>
> G kná:o: 'I scrape'; MIr cnáim 'I chew/gnaw'
In both Greek and Celtic *n.h2 > na:.
> PIE allowed metathesis to correct difficult clusters, often the
> result of morpheme boundaries:
>
> *dhexY+ 'place, put'
>
> *kYerd+ 'heart'
>
> *kYerd+dhexY+ > *kYreddhexY+ 'place trust (heart) in > believe'
My personal theory about *k^red- is that it reflects an archaic locative
of *k^er-d-, *kr-ed(-i) meaning 'in/on the heart', which would imply
that the 'heart' word was morphologically complex.
Of course roots of the shape *CReC- occasionally show byforms with
*CeRC- (thus, e.g. *pleh1-, seen in more archaic derivatives -->
*pelh1-), but such forms are secondary and analogical, derived via the
ambiguous nil-grade *CR.C-, see Raimo Anttila's book on "Schwebeablaut".
Lith. génklas is precisely such a secondary full grade. The root of
'know' is certainly *g^noh3- and no **g^enh3- variant is reconstructible
for PIE. It's a very interesting root, by the way, and if you feel like
discussing it at some depth, I'm game.
Piotr