--- In
cybalist@yahoogroups.com, "fournet.arnaud" <fournet.arnaud@...>
wrote:
>
> Also w > m after ru:
>
> *dakYru 'tear'
> *dakYrwax > *dakY-u-wax 'tears'
> ==================
> A.F :
> An interesting possibility is that
> this morpheme is actually from *m? (glottalized m).
I do believe a glottalized m existed, but it merged with m long
before PIE.
I'd also say there was Nm > M (velar+bilabial > labiodental nasal)
and M>w between V or after a sonorant C. This is separate from the
changes of m>w in specific environments that occur in various IE
languages (not PIE); I'm fairly sure I have most of them covered
thoroughly, though I've barely discussed any but the Latin and Gmc.
changes here.
In Greek and similar languages various forms of x() or other C < x
in certain environments cause both a following mY>w and final n.>r. as:
*piixY-mYn., > *pi:-war > pi^ar 'fat'
*per-x-mYn., > *pers.-war > pei^rar 'end'
the plain x>s() also in:
*bhax+ 'shine, appear'
*bhanYax+ (present)
*bhanY-x-mYn., > *phanYsY-mYn., > phasma / phantasma
but
*dunYax+ 'bend back and forth, be strong' (present)
*dunY-x-mYn., > *dun-a-mnY, > dunami-s
with the difference in preceding a vs. u causing the dif. outcomes.
The determining V dif. can also be seen in:
*dexY+ 'bind'
*dexY-mYn., > *des.-mYn., > desmos
the i preserves xY in *piixY-mYn., long enough for it to participate
in the later change > w.
Indo-Iranian had some similar changes but often xY>y, xW>w, etc.
*+ixYn.o+s > *+iyn.o+ > -ina-
*+ixY-n.,+ > *+iy-n.,+ > -in-
*dexY-mYn., > *de-e-mYn., > da:man-
> As is :
> *sam? "sun" :
> PIE saw-(el/n) = Semitic sham(sh)
This must be from Nm as it denasalizes to either g or w in Uralic.