Re: o-grade thoughts

> Jasanoff
> Hittite and the Indo-European verb:
> "
> As I have suggested elsewhere (Jasanoff 1994: 164), the o-timbre of
> the 3 sg. middle may have originated in a pre-PIE sound change that
> converted unaccented word-final *-e-r to *-o-r in the nascent
> primary ending.

...

> Once established before *-r in forms of
> the type *k^éy-o-r 'lies', an *-o- generated in this way would
> have been well positioned to replace **-e by analogy in the
> corresponding imperfect / injunctive **k^éy-e > *k^éy-o lay' and
> in oxytone forms of the type **dhugh-é-r > *dhugh-ó-r 'furnishes',
> impf./ inj. **dhugh-é > *dhugh-ó.
> "

Middle *-or has no e-grade. 3pl. *-erV- has no o-grade. But if they
have a common origin, that suffix must have had impersonal meaning
(which later split into senses of the type "it is said" and "they
say"). Impersonal forms exist only in one non-person, non-number
form. In PIE they might have been relegated to 3sg., if used in the
construction of a middle. Now it makes sense that the middle deviates
from the perfect only in 3sg -o/-e.

The Baltic Finnic languages have such an impersonal form. A few of
them have generated a true passive from the impersonal.

That works well with my idea that the mi-endings in their origin
deictics: (*-m "at me", *-s "at you", *-t "at him", *-n (-> *-r
in auslaut, cf the neuter -r/-n stems) "somewhere").


Torsten