On Tue, 30 Oct 2001 13:56:05 +0100, "Piotr Gasiorowski"
<
gpiotr@...> wrote:
>I can't offer a well-founded opinion on it. If you want me to speculate -- maybe some kind of "e-infix" in the present tense compensating for the immobility of stress in this type of aorist. At any rate, I wouldn't accept any solution that doesn't explain the athematic "Narten presents" at the same time (the *té:k^s-ti : *ték^s-enti type), and if it explained vowel alternations in Hittite conjugations as well I'd know it to be _the_ solution.
Ignoring some of the finer details, Hittite Umlaut in the active
conjugations doesn't pose much problems (sg: /e/, pl: /0/, just like
elsewhere). The difficult one is in the hi-conjugation, where we have
sg. /a/ vs. pl. /e/ (also in the preterit). If we equate the Hitt.
hi(ha)-preterite with the PIE perfect, the problem is that the perfect
has /o/ ~ /0/, not /o/ ~/e/. Then, there is also the temptation to
compare the Hitt. forms with <e> (if /e:/?) with the /e:/-perfects of
e.g. Latin and Germanic.
Given my, well, profound belief that PIE */o/ derives from earlier
**/a:/ (and */e/ from **/a/), the Hittite situation appears to be the
most logical: a lengthened grade singular /a:/ > /o/, vs. a full-grade
plural /a/ > /e/ (ultimately from a single paradigm with vowel /a:/,
stressed in the sg., unstressed in the pl., just like the active can
be derived from a paradigm with /a/, stressed in the sg., unstressed
in the pl.). So why the PIE perfect with zero grade in the plural?
I don't know. Maybe the reduplication had something to do with it,
maybe it's just analogical after the active (but why not go all the
way and normalize the act. sg. as well?).