Re: [tied] 1sg. -o: [was Re: IE Thematic Vowel Rule]

Miguel Carrasquer wrote:

> I have doubts about whether the doubly thematic subjunctive
> was pan-PIE or merely dialectal IE, let alone pre-PIE. If,
> as Jens has stated, and I tend to agree with him, the
> thematic present derives from the subjunctive, then the
> thematic thematic subjunctive must be late. And its
> composite nature was very much at the surface still (had it
> existed in Balto-Slavic, I have no doubt it would have had
> circumflex intonation).

Are you so sure _all_ thematic presents were original subjunctives? As
far as I recall, Jens has claimed that of the "plain" thematics like
*bHere/o- and, I think, of the reduplicated ones. What about
causative/iteratives in *-eje/o-, and the *-je/o- and *-sk^e/o- presents
-- all of them major classes of present stems in PIE? I see no reason to
deny them the status of ordinary presents in the protolanguage, whatever
their ultimate origin. It would be rash to assume that the hyperthematic
subjunctive _must_ be late and "merely dialectal" only because it's
secondary and transparently composite (the *-a-a- < *-e-e- still scans
as disyllabic in Avestan).

>>There would have been a contrast between this *-o: and later
>>contractions, or forms retaining a final consonant (even if the latter
>>could be dropped in sandhi). One could compare the 1sg. ending with the
>>animate thematic nom./acc.du. *-o:, if from analogical *-o-e. The Baltic
>>development is the same.
>
>
> For me, that's another compelling reason to believe that the
> dual ending does not in fact come from *-o-e.

The behaviour of vowels in open auslaut is often different from the
behaviour of the same vowels before a consonant. We have significantly
different environments in *-e-e-t and *-o-o or *-o-e, and we can't
expect the same behaviour of contracted vowels in both cases on an a
priori basis. The possibility I'm arguing for is that word-final *-o-V
was reduced to monosyllabic *-o: already in the protolanguage, which is
precisely why the duals in *-o: and the 1sg. *-o: lost their
transparency early and are so hard to analyse.

> Elsewhere, you said:
>
>>But I find it difficult to believe there was an *h3 in the
>>1sg. ending.
>
>
> Well, I'm reluctant too. For starters, I don't see any
> reason why the thematic endings should be based on anything
> different than the athematic ones. In particular, I don't
> like (becuase there is no motiovation for it) a connection
> between the thematic present and the stative/perfect/middle
> system.

My sentiments, precisely.

> However, if there *is* a connection, then *-h3 is perhaps
> not completely unexpected. The vocalism of the 1sg. stative
> points to *-ku in several branches of Southern Nostratic
> (Semitic -ku, Kartvelian xw-, Elamite -(h)u, -k), so perhaps
> PIE *-h2 (< *-ka) is the unexpected ending.

Unexpected or not, *h2, and definitely not *h3, is what we get in PIE.

Piotr