On 2006-08-16 14:22, Abdullah Konushevci wrote:
> Because Lubotsky as Hittite cognate takes e-ip-zi 'to grab, catch'
> and Latin co-e:pi:'to start, unertake', then his reconstruction
> looks much plausible, for vocalized *H1 has as result /e/.
_Vocalised_ *h1- (i.e. *&1) yields Hitt. a-, as in asanzi 'they are'.
The point here is that _consonantal_ *h1 is lost in Hittite, so <epzi,
appanzi> appear to reflect *h1ep-ti, *&1p-ónti (for *h1p-énti). Hitt <e>
would not exclude a preceding *h2 if we were dealing with a Narten-type
root like *h2e:p-/*h2ap- (long *e: wasn't coloured by *h2), but in that
case the initial *h2 would have been preserved.
Piotr