Re: [tied] IE genitive

On Sun, 20 Apr 2003, Miguel Carrasquer wrote:

> On Sun, 20 Apr 2003 02:06:29 +0200 (MET DST), Jens Elmegaard Rasmussen
> <jer@...> wrote:
>
> >It's a rule I was taught in Erlangen in the 70's. Of course it is not
> >better than the observations on which it is based, but that seems to be a
> >fair amount: Opposed to athematic 3pl *H1s-ent-i we have thematic 3pl
> >*bher-o-nt-i with -nt-, not -ent-;
>
> The other plural or dual forms (*-o-mes ~ *-o-men, *-e-te, etc.) do
> not reduce the vowel. In the 3rd pl., *bhero-ent may have been
> contracted to *bhero:nt, and then shortened again to *bheront (-V:CC >
> -VCC, as in the nominal nt-stems).

The rule-of-thumb says "anything capable of ablaut gradation appears in
the zero-grade if the thematic vowel precedes". One could also say that
"everything appears in its weakest allomorph if the thematic vowel occurs
earlier in the word". There are no further reduced forms of *-me/*-mos,
*-te. I know of no shortening rule like the one you are invoking. I do
know of a shortening, however, before -CC- + word-final nominative marker
(and probably also before -CC- + collective marker), but that has no
business here.

>
> >opposed to athematic 3sg opt.
> >*H1s-ieH1-t we have thematic 3sg opt. *bher-o-yH1-t with -yH1-, not
> >-yeH1-; opposed to statives based on athematic stems like *bhudh-eH1- we
> >have thematic-based *sene-H1- with -H1-, not -eH1-; and the gen. morpheme
> >*-os does appear reduced to /-s-/ in the middle of *te-s-yo (what -yo is
> >is a different matter);
>
> I prefer to reconstruct *tosyo (the first /o/ being the thematic
> vowel, lengthened to /o/ before the 'ending' -esyo, where the /e/ was
> indeed reduced by zero grade: toesyo > tosyo: I'm not sure whether the
> accentuation was *tó-esyo or *to-esyó). The athematic pronominal type
> would be represented by e.g. *kWesyo (from *kWís, not *kWós) or indeed
> *esyo itself (Nom.masc. *is).

The descriptive thing is that we do no have a long vowel here, so
_something_ has been reduced. I now of no "athematic pronominal type": the
"thematic vowel" means "stem-final vowel".

>
> >incidentally also in the pronominal gen.pl.
> >*-oy-s-oom (what -oom is is a different matter) and gen.du. *-oH3-s;
>
> But that's not the thematic vowel in the G.du., is it?

I have guessed it was, but it _is_ merely a guess. Klingenschmitt has made
the same guess, certainly independently (he would surely have made a
different guess if he had known I had already said the same thing). The
guess brings Gen.du. *-o-H3-s and Loc.du. *-o-H3-u in line with other
genitives and with the loc.pl. in *-s-u. That leaves the -o- which must,
then, belong to the stem and, conveniently, the most productive stem class
is precisely the one in -o-.

>
> >the
> >instr.sg *-VH1 is reduced in the pronominal form *te-H1, subst. *-o-H1. I
> >suspect the athematic 1sg middle secondary ending is *-H2a, while
> >the thematic primary ending is certainly *-a-H2-i without the desinential
> >vowel.
> >
> >The presumed counterexamples you mention are invalid, for the dative *-ey
> >and the nom.pl. *-es, and endings like *-bhyos, *-bhis, *-su do not
> >alternate at all.
>
> I didn't mention *-bhis or *-su. I do think the dative *-ei is in
> origin an Ablaut variant of locative *-i (stressed *-é(i) vs.
> unstressed *-e(i)), the same as instrumental *-éh1 (< **-ét) vs.
> ablative *-ot.

No you didn't, but they belong in the same basket, that of non-alternating
morphemes, that's why _I_ mentioned them. The relation between the dative
in unchanging *-ey and the loc. in zero or *-i with accent on the final
part of the _stem_ just cannot be reduced to anything deserving the name
ablaut. Why would ablaut distinguish cases?

>
> In none of the examples given do I see any special significance vis à
> vis zero grade for the position _after_ the thematic vowel. When a
> stressed ending is added after the thematic vowel, the vowel of the
> ending is not reduced (1pl. **-o-més, Dsg. **-o-éi, etc.), and the
> retraction of the stress back to the thematic vowel (1pl. *-ómes) is
> secondary. An unstressed /e/ in the ending *is* reduced, as it is
> everywhere else (ath. opt. 3pl. -yh1-ér, etc.).

Right, a lot of material is irrelevant, so the rule is based on the other
stuff. The 3pl opt. is *H1s-iH1-ent in GAv. x'ii&:n and OLat. sient.

>
> The thematic vowel _itself_ does of course behave most peculiarly with
> regards to the zero grade rule.

Yes, that's part of it. Saussure saw that already.

Jens