Re: [tied] Re: PIE i- and u-stems again

On Mon, 19 Feb 2007 00:02:35 +0100, Piotr Gasiorowski
<gpiotr@...> wrote:

>On 2007-02-17 10:39, Miguel Carrasquer Vidal wrote:
>
>> I don't think so: the "lengthening" of the thematic vowel
>> (i.e. e > o) took place before zero grade, as shown for
>> instance by *-o-syo < *-%-esyo. If the full grade of the
>> suffix was *-eh1(i)- (and I don't believe in suffixes of the
>> shape -C-), a thematic vowel before it would have gone to
>> /o/.
>
>I beg to differ here. The way I see it, a suffix of the shape *-eC-
>would have been realised as *-C- postvocalically well before the rise of
>the "normal" zero grade. Examples of pronominal instr.sg. like Goth.
>hWe: suggest *kWe-h1 to me, and as gen.sg. *kWesjo can hardly be
>anything but *kWe-sjo, I'd say that the stem-final vowel was originally
>realised as *-e- in such cases, the shift to *-o- in substantives being
>analogical.

There is no shift to *-o- in substantives, as the *-o- is
well-attested in the pronominal forms themselves. *kWeh1 and
*kWesyo have /e/ because these forms simply do not contain a
thematic vowel: they come from *kWis, n. *kWid, f. *kWih2.
The interrogative pronoun is somewhat special because it
retains both variants of the pronominal paradigm (sometimes
mixed up in the individual languages): "athematic" *kWis and
"thematic" *kWos, n. *kWod, f. *kWah2. Most pronouns have
either one form or the other (e.g. *k^is (but thematic kas
in Hittite) is usually athematic, *so ~ *to- thematic,
etc.).

>> I'm also not aware of a formal distinction anywhere
>> between deverbative essive/fientives (e.g. sêdêti) and
>> deadjectival essive/fientives (e.g. slabêti). Is there?
>
>Probably not. But what does it prove?

It suggests to me that the deadjectival verbs are built upon
the athematic root of the adjective, which would compromise
the link with the "cato-group".

>> As I think I said earlier here, I would reconstruct the
>> original paradigm of these verbs as athematic (as still
>> preserved in Lithuanian and Aeolic Greek):
>>
>> present: aorist:
>> *-éih1-mi *-éh1-m
>> *-éih1-si *-éh1-s
>> *-éih1-ti *-éh1-t
>> *-h1i-més *-h1i-mé
>> *-h1i-té *-h1i-té
>> *-h1i-énti *-h1i-é:r
>>
>> Most languages have thematized the type, Slavic generalizing
>> *-eih1-e/o- in the present stem and *-eh1- in the infinitive
>> and aorist stem. The type is still basically athematic in
>> Baltic, with generalized *-h1i- in the present, *-eh1- in
>> the infinitive. Greek generalized *-eh1- to the present
>> (Aeolic <phile:mi>, thematized in Attic <phileo:>). In
>> Sankrit it may have been 3pl. *-h1i-ánti, reinterpreted as
>> thematic ye/yo-stem *-h1-yá-nti, which led to *-(h1)yo:,
>> *-(h1)yesi, etc.
>
>I can't say I understand the structure of this paradigm, and in
>particular the shifty *i, and I can't believe anything so odd could be
>original.

The different variants are the result of soundlaws described
by Jens in [I would have to go upstairs to look up the
reference]. From memory: stressed *éh1i becomes *éih1
before single medial C, *éh1 before CC or final C, and
remains as *éh1i before V. Unstressed we have *h1i before
voiced (including V), *&1 before unvoiced (or at least /t/)
[the 2pl. forms are analogical]. The soundlaws are pertinent
to the froms belonging to the problem of the "long
diphthongs", but when I plugged them into the
essive/fientive verbal endings, the Balto-Slavic paradigms
just came rolling out (Slavic has generalized singular
*-éih1- in the present, Baltic plural *-h1i-, like Slavic
sg. *-né- vs. Baltic pl. *-n- in the n-(infix)-verbs; both
have *-eh1- in the aorist/infinitive system). Excuse me for
getting away with the impression that this has to be the
correct explanation. Especially since the forms proposed in
LIV (fientive *-eh1-/*-h1-, essive *-h1yé-) are utterly
inadequate to explain the Balto-Slavic forms.

>I'd sooner accept an athematic by-form of the present stem
>without the *-je-, i.e. *-e-h1-mi ~ *-e-h1-jo: (similarly for
>deadjectival factitives in *-a-h2(-je)-?).

Such athematic forms as Aeolic phile:mi or OHG habe:m
continue generalized *-eh1- from the aorist (or imperfect).
Apparently, outside of the Balto-Slavic area, the original
paradigm was indeed perceived as too odd...

=======================
Miguel Carrasquer Vidal
miguelc@...