On 2008-04-07 02:14, Patrick Ryan wrote:
> A very great number of the PIE roots that we find are *C + one of the stem
> formants above:
>
> -*H, -*I/y, -*u/w
>
> Another large number combine two of these primary stem formants:
>
> *CVHi-, *CVHu-
>
> Without reaching outside of PIE for answers, is it possible to determine
> what the earliest meaning of these formants?
There's nothing *inside* IE to suggest that roots like *//dHeh1-//,
*//steh2-//, *//deh3-// contain an extension. On the contrary, all the
available evidence shows them to be indivisible, and the same is true of
roots like *//werh1-// 'speak', *//perh2-// 'trade', *//gWerh3-//
'swallow, devour'. A root like *werh1- has no relatives of the form
*wer- or *werC- (where *C is something else than *h1).
Cases like *//drem- ~ *dreh2- ~ *drew-// 'run' suggest an analysis in
terms of pre-PIE **d(e)r-éC-, but the meaning of the extensions can
hardly be recovered, as the roots in question are practically synonymous
and the unextended **der- is not attested. In my article on *gWih3w- I
point out that *-ew- recurs in verbs indicating movement (such as
*drew-, *plew-, *srew- and my hypothetical *gWew-).
"Long diphthongal" roots like *//dHeh1j-// 'suck', *//deh2j-//
'distribute', *//peh3j-// 'drink' are reconstructed by Jasanoff (2003)
as a special category of "h2e-conjugation i-presents", coming in two
flavours: mobile (3sg. *dHéh1-j-e, 3pl. *dH[h1]-j-énti) and static
(*spé:h2-j-e/*spéh2-j-n.ti 'be sated' [J.'s reconstrucion]). Jasanoff
hypothesises that this *-j- was restricted to laryngeal-final roots
while other roots made *-je/o- presents; he only observes but does not
explain this complementary distribution. If he is right, the *-i-
"extension" ha no meaning, just a _function_ (as a present-stem forming
element). It's all a new line of research, though, still disputed and
far from generally accepted.
Possible traces of a PIE u-extension are visible in a few present stems
like *//terh2w-// 'overcome, cross over' (Skt. tú:rvati < *tr.h2-w-é/ó-)
and *//wel(h)w-// 'turn' (Lat. volvo: etc.), but there are so few of
them that general conclusions about the meaning or function of this *w
are hard to draw (in verbs of movement like these two and *//dHenh2w-//
'move off' it could perhaps be identified with my *-ew- extension).
Piotr