On Tue, 27 Apr 2004 13:09:18 +0200, Sergejus Tarasovas
<
S.Tarasovas@...> wrote:
>> From: Miguel Carrasquer [mailto:mcv@...]
>
>>The -ija:- (> e:) in the
>> ja:-stems (e:-stems) is the result of Sievers (-CC-ja: >
>> -CC-ija: after heavy syllable).
>
>And what would play the role of -CC- in the protoform of, say, z^e~me:?
Nothing. The e:-stems probably also incorporate some nouns
with -e: from other origins, for which we can compare the
Latin e:-stems.
In the case of z^e~me:, besides Slavic zemja:, we have to
depart from an extension of the PIE word *dhg^hom- ~
*dhg^hem-. Perhaps the suffix here was originally athematic
feminine (unstressed) *-ih2, which becomes *-îh2 > *-ye:h2
after a _light_ syllable (cf. Lat. *mah2ter-îh2 >
ma:terie:(s)). Lithuanian z^e~me: would then continue
*(dh)g^hém-ye:h2 directly, while Slavic zemja comes from a
by-form with unstressed /e:/ reduced to /e/ (*g^hem-yeh2 >
*z^emja:).
In general, I would explain the rare PIE e:-stems as coming
from i-stems and ih2-stems after a light root.
proterodynamic
N *'-îs '-ye:s
G *-í-âs -éy-os
The dominant type (comparable to the amphidynamic type in
the C-stems) had a long root vowel, or -CC, so no
lengthening in the desinences:
N *'-is '-is
G *-í-as -éy-s
The hysterodynamic type went:
N *-ís -ís
G *-i-ás -yés
For the *ih2-stems, we have:
[PD => (y)e:-stems]
N *'-îh2 '-ye::
G *-íh2-âs -yá:os > -yá::s
[AD => de:vi:-type]
N *'-ih2 '-i:
G *-íh2-âs -(i)yá:os > -(i)yá::s
[HD => vrki:s-type]
N *-íh2-s -í:s
G *-ih2-ás -iyás
Both in the *-i and *-ih2 stems, heavy roots were apparently
much more common than light roots. The imbalance was
further aggravated by the fact that unstressed /e:/ was
shortened to /e/ in most PIE dialects, so that *'-ye:s
became *'-yes and further *'-is (merging with the other two
variants [perhaps maintaining a distinct gen. in *-yós]),
and *'-ye:h2 > *'-yeh2 > *-ya: (merging with the
*-y-ah2-stems).
If there was a similar phenomenon in the u-stems, then the
light-root proterodynamic forms in *-ûs > *-wos must have
merged with the o-stems (ending reinterpreted as *-w-os).
=======================
Miguel Carrasquer Vidal
mcv@...