On Mon, 24 Jan 2005 10:39:25 +0100, Piotr Gasiorowski
<
gpiotr@...> wrote:
>
>On 05-01-21 17:50, Miguel Carrasquer wrote:
>
>> Have you compared notes with Jens' theory about these
>> *VHi/*VH/*ViH/*Hi/*&-forms? I seem to recall a discussion a
>> long time ago, but I can't remember if anything was
>> resolved.
>
>Oh, yes. I was on summer vacation at the time, and unable to cope with
>the e-mail traffic on my occasional visits to Cybalist, so there wasn't
>much of a discussion on my side. My "smoothing theory" left Jens
>unconvinced (and I myself wouldn't any longer defend all its elements),
>although I think we both, as well as other people, agree at least about
>one thing: the last word about the "long diphthong roots" has not been
>said yet.
I was mainly concerned with the Slavic i/ê-verbs, and the
PIE "fientive/essive" in general, which lends itself to a
nice solution if we apply Jens' soundlaws to a formant
*-eh1i-.
LIV recognizes two verbal formations: the "fientive" (strong
mn-éh1-, weak *mn-h1-') and the "essive" (thematic
*lip-h1yé-). In view of the material presented in support
of those categories, I cannot accept the analysis as given
in LIV.
In the first place, there is only one category, not two.
The present/progressive forms of the essive are said to mean
"to be in a state" (Zustand des Subjekts), while the
aorist/punctual forms of the fientive are described as "to
enter a new state" (Eintritt des Subjekts in einen neuen
Zustand"). The two are related in an obvious way, and the
assignment of e.g. the OCS aorist bUdê to the fientive, with
the rest of the paradigm (bUz^doN, bUdêti) in the essive is
clearly artificial. The fact that both formations contain
*h1 also suggests that they are one, although the way in
which they are is not as obvious.
It *is* obvious that the Slavic i-present (-joN, -is^I,
-itI, -imU, -ite, -jeNtI) cannot be derived from the
proposed *-h1yé-, since that would have given -joN, -jes^I,
-jetI etc. It is also impossible to derive Lithuanian -i-
from that.
The reconstruction *-hyé- is apparently primarily based on
the Vedic passive -(:)yé:, -(:)yáse:, -(:)yáte:,
-(:)yá:mahe:, -(:)yádhwe:, -(:)yánte:. I haven't studied
the appropriate literature, but I'm guessing that additional
evidence for this reconstruction is believed to be the Greek
verbs in "stative" -eo: (< *-&1jo:) and the Germanic third
weak class (Goth. haban: haba, habais, habaiþ, with -ai- <
*-&1j-).
LIV doesn't specify if the essive had active or middle
endings or could have both. Based on the material
presented, Vedic and Tocharian (in the present) usually have
middle endings, while Balto-Slavic (of course), Germanic,
Latin, Celtic and Greek have mostly active endings.
Apart from the middle vs. active distinction between Vedic
on the one hand, Greek and Germanic on the other, the
formant may well be *-&1ye- in the latter two. However,
there are other explanations for the attested forms.
The Germanic present forms are:
Goth. ON. OE. OS. OHG.
haba hef(e) hæbbe hebbiu habe:m
habais hef(e)r hafas hab{e/a}s habe:s
habaiþ hef(e)r hafaþ hab{e/a}ð habe:t
habam hOfom habbaþ hebbiat/ habe:me:s
habaiþ hafeð ,, habbiat habe:t
haband hafa ,, ,, ,, habe:nt
My 1911 edition of Loewe's "Germanische Sprachwissenschaft"
derives all of these from *-e:(i)- with athematic endings.
The fact that Gothic has -ai- where thematic forms have *-e-
and has -a- where thematic forms have *-o- in my opinion
indicates that the paradigm is thematic. But that doesn't
mean that OHG -e:- cannot come from unstressed -e:-, as well
as unstressed -ai-, as Loewe claims. I can't decide whether
the Germanic forms are to be derived from:
-ajo:, -ajesi, -ajeti, -ajomes, -ajete, -ajonti (> -(j)o:,
-ais, -aiþ, -(j)omz, -aiþ, -(j)and)
or from:
-e:jo:, -e:jesi, -e:jeti, -e:jomes, -e:jete, -e:jonti (>
-jo:, -e:s, -e:t, -jomz, -e:þ, -jand).
The latter forms would be in agreement with Italo-Celtic
(e.g. Latin -e:jo: > -eo:, -e:jes(i), -e:s, -e:jet(i) > -et,
-e:jomos > -e:mus, -e:jetes > -e:tis, -e:jont(i) > -eunt;
Celtic -i:- < -e:-), which is a priori more probable.
The Greek forms are discussed in Sihler, p. 514, but first
it is perhaps best to look at the Balto-Slavic forms.
Slavic -joN, -is^I, -itI, -imU, -ite, -jeNtI cannot come
from a thematic paradigm like Latin (and Germanic?)
-éh1i-o:, -éh1i-esi, -éh1i-eti, -éh1i-emos, -éh1i-etes,
-éh1i-onti (would have given Sl. -êjoN, -êjes^I, etc.). It
also cannot come from thematic -&1i-ó:, -&1i-ési, -&1i-éti,
-&1i-émos, -&1i-étes, -&1i-ónti, like Vedic (or Germanic?),
which would have given Sl. -joN, -jes^I, -jetI, etc.
The paradigm must have been athematic:
-éh1i-mi
-éh1i-si
-éh1i-ti
-eh1i-mós
-eh1i-tés
-eh1i-énti
Applying Jens' rules, we get:
-éih1mi
-éih1si
-éih1ti
-h1imós
-&1tés (> -h1ités, analogical)
-h1iénti
Slavic generalized the strong from -éih1-, Lithuanian the
weak form -h1i- (cf. the way in which Slavic has generalized
the strong suffix -né-, while Lithuanian has generalized the
infix -n-).
We have Slavic: -i:o: > -jo:, -i:-si > -is^I[1], -i:ti >
-itI[2], -i:-mos > -imU, -i:-tes > -ite, -i:enti > -eNtI.
[1] an athematic ending immediately explains Slavic -s^-.
[2] the subjunctive/injunctive in -t is absent (would have
become *-éh1i-t > *-ê(tU), which is too aberrant).
Lithuanian: -io: > -iù, -ii: > -ì, -iti > -i, -imes > -ime:,
-ites > -ite: (no 3pl.).
Greek (Aeolic) also has athematic forms in the essive
present (-e:mi, etc.). If we substitute the athematic forms
we used above in the table on p. 514 of Sihler, we get a
much better picture:
"stative" denominative
philéih1mi woikejo:
philéih1si woikejesi
philéih1ti woikejet
philhimén woikejomen
phil&1té woikejete
philhiénti woikejonti
Thematization (outside Aeolic, at least) of the stative
gives:
"stative" denominative
phileih1o: woikejo:
phileih1esi woikejesi[2]
phileih1et(i) woikejet[3]
phil&1jomen[1] woikejomen
phil&1(j)ete woikejete
phil&1jonti woikejonti
[1] *&1 of the 2sg. analogically extended to 1/3pl.?
[2] *-esi > *-ei + s > -eis
[3] (subj.?) *-et > -e + i > -ei
Which leads to the instant merger of the two paradigms .
As to the "fientive", the root aorist athematic paradigm:
*-éh1i-m
*-éh1i-s
*-éh1i-t
regularly gives:
*-éh1m
*-éh1s
*-éh1t
The plural should have *-h1i-mé, *-&1-té, *-h1j-é:r, besides
analogical *-eh1-mé, *-eh1-té, *-eh1-(j)-é:r(?).
=======================
Miguel Carrasquer Vidal
mcv@...