Re: PIE -*C-presents

From: tgpedersen
Message: 54013
Date: 2008-02-23

> >> It is also restricted to the third person.
> >
> >And according to Burrows and Jasanoff, so was the s-aorist, which
> >was the line of readoning I tried to carry on.
>
> My point was that it's a much more recent formation (as also
> witnessed by the fact that it's restricted to Indo-Iranian).

I don't get that line of reasoning. Why can't it be a sole survival
instead?

> >> Burrows may be right that
> >> it's in origin a nominal formation (neuter i-stems with
> >> o-grade), because that is indeed what the forms look like.
> >> The s-aorist forms (even in the third person) don't look
> >> much like masculine root nouns in the nominative at all.
> >
> >> Root nouns rarely have e(:)-grade,
> >
> >But if deverbal root nouns occurred partly free, with e-grade, and
> >partly bound in compounds with their verbal object, with o-grade,
> >the e/o-distribution caused by the position as tonic and posttonic
> >vowel, respectively, then, if such original root nouns were used in
> >preterite constructions, as I implied, they would invariably have
> >e-grade in that function, and, to avoid confusion, the o-grade
> >would have become
> >the mark of its non-verbal use, ie as a noun proper. Now we have an
> >explanation for that phenomenon.
> >
> >> and they always lose nominative *-s after a resonant.
> >
> >Awesome. Now we have an origin for root aorists too.
>
> Root aorists do not have e:-grade.

Do root nouns that lose nominative *-s after resonant still have
Szerémenyi lengthening?

> >> Furthermore, the *-s in
> >> the 3pl. is added to the verbal plural morpheme *-en >
> >> *-(e)r, not to any nominal plural. If the precursor of the
> >> s-aorist was ever a nominal form, it had already become a
> >> purely verbal form by the time of PIE.
> >
> >And that might be exactly why it was being mechanically applied to
> >the plural by analogy of the singular?
>
> I don't understand.

If the 3sg had -s (no other ending at first) the way to construct a
plural for it might be to add -s to an ending imported from the 3pl
perfect?


Torsten