On 2006-08-13 16:43, raonath wrote:
>> The tenseless (non-reporting) "injunctive" form *bHer-e-t is all
>> that we need as the pivot of the change. With the appropriate tense
>> markers it becomes a present (*bHer-e-t-i) or an imperfect (*e
>> bHer-e-t).
>
> That makes Hittite harder to explain: Why did the secondary endings
> become past, rather than remain tenseless?
Don't our PIE grammars exaggerate the importance of tense distinctions
(and primary vs. secondary endings) in the protolanguage? The least
marked form of the verb was in fact tenseless. What tense markers it
could receive depended on the aspect of a given verb stem. For example,
root verbs were inherently durative/iterative (*gWHen-) or
punctual/telic (*kWer-); the punctual ones couldn't be used in the PIE
"present" tense which had, more or less, the function of the English
present continuous, but the durative ones could be both "present
continuous" and "past continuous" (imperfect). In narration, the
particle *é was used to introduce each new event ("NOW he comes, and NOW
he kills the wolf"), and so it came to be interpreted as a "definite
past" marker. The *-i of the "present tense" marked action in progress
and was perhaps originally just an unaccented allomorph of *é, as argued
by Jens.
Hittite has done as much restructuring as the other branches. Having
developed a tense-based verb system and having abandoned aspect
distinctions, it could conjugate any verb in the present or the
preterite (unlike PIE), marking _both_ of them. It partly recycled the
PIE "present continuous" as the present and used (modified) "secondary"
endings for the preterite, but would you say that there is a markedness
contrast e.g. between <wek-mi, wek-ti, wek-zi> and <wek-un, wek-ta, wek-ta>?
> There are also a few other oddities: If primary endings were present
> markers from the beginning, why were they not applied to the "know"
> forms?
Meaning the PIE perfect? It was by nature non-modal and non-tense, so
e.g. it didn't form subjunctives or imperatives, didn't take on the
augment or the *-i ending (except in some of the daughter dialects,
where it acquired tense functions).
> And coming back to the original question, were there two subjectives,
> a "tenseless subjunctive" using secondary endings and a "present
> subjunctive" with primary endings? If so, what was the difference in
> their meanings, if any?
I could imagine various possible uses of the aorist subjunctive:
"tenseless irrealis" (*X-e-t 'he would/might X'), "narrative" (with the
augment: *é-X-e-t, perhaps 'he was about to X') and expressing "imminent
future" (*X-e-ti 'he's going to X'). The last use must hae been
extremely popular, judging from its consequences (the development of a
productive class of presents). But it's my personal opinion, not
necessarily relecting the "received wisdom" of the field.
Piotr