Re: [tied] The phonetic value of PIE *h3 and the 'drink' root.

From: Jens Elmegaard Rasmussen
Message: 14052
Date: 2002-07-17

On Wed, 17 Jul 2002, Miguel Carrasquer wrote:

> On Wed, 17 Jul 2002 14:40:51 +0200 (MET DST), Jens Elmegaard Rasmussen
> <jer@...> wrote:
> >The ablaut of Narten presents, reduplicated structures and subjunctives
> is
> >highly recessive.
> I'd like to know more about the ablaut of subjunctives.  My impression
> is that all or at least the vast majority of subjunctives have full
> grade throughout (following the bhávati-type rather than the
> tudáti-type).  Am I overlooking something?

We have too few and too inconsistent facts to formulate rules that work
blindly. I do take the form bhárati to be in origin the subjunctive of a
root aorist (the one to go with bibhárti, and incidentally retained in
the middle voice in Slavic aor. bIra 'took'). I do not quite know why the
subjunctive got this structure, for Narten presents have stavati,
taks.ati, and the s-aorist has stos.ati. The unanimous evidence of
reduplicated forms such as the present (sbj. dádhati), intensive
(carkirati) and perfect (Gk. eídomai, Skt. védati from *wé-[w]id-e-ti with
loss of the second /w/, thus also British Celtic) points to zero-grade
root. We expect *bhr-é- but find *bhér-e-. Now, since the subjunctive is
the "dependent clause form" par excellence, could it be the dependent form
of the verb, maked just like the dependent form of nouns, i.e. like the
adjectives with a thematic vowel? In that case, one could perhaps even
imagine the same double marking as in *twó- and *tewó- 'thy', i.e. with
*vrddhi*. Is the first -e- of *bhér-e- and inserted "extra mora", so that
the underlying form had a long /-e:-/ which ended up shortened and
accented as usual? Or is it just a matter of standardization of one of the
types, the one of the acrostatic verbs, which happened to coincide with
the basic root structure in the most simple way?

> >All the evidence we
> >have (which is little) indicate that the 1./2. pl. had final accent,
> >irrespective of the presence of reduplication; the 3pl perfect even has
> >final accent on a zero-grade morpheme which must be an innovation, even
> if
> >the form has been established before the breakup of PIE.
> At least for the reduplicated presents, I find it hard to conceive
> that final accentuation (not counting -i) in the dual/plural was *not*
> the original state (*bhi-bhr-méni (-més), *bhi-bhr-téni (-té(s)),
> **bhi-bhr-énti), as it is for unreduplicated presents.  The only thing
> that sets apart the 3pl. ending is that it starts with a vowel, but I
> must confess I have no idea why this would have triggered an accent
> shift from **bhV-bhr-énti to *bhé-bhr-n.ti (Skt. bíbhrati).

Both 1/2 pl *-me, *-te and 3pl *-ent are endings that are known to move
the accent: Ved. smas, stha, santi all have zero-grade of the root. I
therefore find it very easy to conceive that they acted in parallel
fashion also in the case of reduplication. And I find it equally easy to
imagine that the odd alternation distuptions caused by the reduplication
was mended in the parts of the paradigm that are less prone to harbour
archaisms: Gothic sind is an archaism, sijum and sijuth are not, or not
one of that degree. Some things give way to standardization more easily
than others.