Now that I'm fully awake, some points I forgot to address:
On Thu, 13 Jan 2005 06:17:13 +0000, willemvermeer
<
wrvermeer@...> wrote:
>And what about the locative singular?
I believe that Balto-Slavic mobility continues PIE mobility,
only "by different means" (to paraphrase von Clausewitz),
i.e. transferred also to the thematic (vowel) stems. The
athematic model for the mobility is a simplified pattern of
lateral mobility that merges the various patterns of PIE
mobility, such as:
PD HD
N *h2ák^mo:n *dhugh&2té:r
A *h2ák^monm. *dhugh&2térm.
V [*h2ák^mon] *dhugh&2tér
G *h2(a)k^ménos *dhugh&2trés
D *h2(a)k^m(e)néi *dhugh&2tréi
L *h2(a)k^méni *dhugh&2téri
I *h2(a)k^(m)néh1 *dhugh&2tréh1
Ab [*h2(a)k^ménod] [*dhugh&2térod]
N *h2ák^menes *dhugh&2téres
A *h2ák^menm.s *dhugh&2térm.s
G *h2(a)k^ménom *dhugh&2tróm
D *h2(a)k^mn.bh(i)ós *dhugh&2tr.bh(i)ós
L *h2(a)k^mn.sú *dhugh&2tr.sú
I *h2(a)k^mn.bhí(:)s *dhugh&2tr.bhí(:)s
NA *h2ák^menoh3 *dhugh&2téroh3
GL *h2(a)k^mn.(H)óu(s) *dhugh&2tr.Hóus
DI *h2(a)k^mn.bh(i)óh3 *dhugh&2tr.bh(i)óh3
We'd expect the sg. N., (V.), G., D., I., the pl. G., D.,
L., I. and the du. GL, DI to receive end stress; the sg. A.,
(V.), L, (Ab.), pl. N., A., and du. NA to have begin-stress.
In actual fact, the D. sg. is barytonic, not oxytonic. This
is not completely unexpected, as also in Sanskrit the dative
has retained its original oxytonicity much less strongly
than the instrumental (cf. ins.sg. prathiná:, pre:n.á:,
bhu:ná:, mahiná, dra:ghmá: with not only end-stress, but
loss of either /m/ or /n/ in the zero-grade of the suffix
-m(e)n-). Apparently, a PD form like *h2(a)k^mnéi had
become *h2(a)k^ménei, adopting the accentuation of the
locative *h2(a)k^méni.
The Balto-Slavic unified pattern will have been (for
non-neuters):
N *akmó:: *dug&té::
A *ákmenim *dúg&terim
V [*ákmen ?] [*dug&tér ?]
G *akmenés *dug&terés
D *ákmenei *dúg&terei
L *ákmeni *dúg&teri
I *akmené: *dug&tré:
Ab [*ákmeno ?] [*dúg&tero ?]
N *ákmenes *dúg&teres
A *ákmenins *dúg&terins
G *akmenó:m *dug&teró:m
D *akmen(i)mós *dug&ter(i)mós
L *akmen(i)s^ú *dug&ter(i)s^ú
I *akmen(i)mí:s^ *dug&ter(i)mí:s^
NA *ákmeno: *dúg&tero:
GL *akmenóu(s^) *dug&teróu(s^)
DI *akmen(i)mó: *dug&termó:
When mobility spread to the vowel stems with columnar stress
on the thematic vowel (by which I mean not only e/o, but in
this case also i, u), the pattern of the mobile C-stems was
mimicked, but with a few exceptions. For instance, the
o-stem N.pl. in *-àj retained its final stress (Lith. -ai~,
-ì). Another exception was apparently the loc.sg. of the i-
and u-stems. There, the expected forms **-éi-i and **-éu-i
(non-oxytonic) had been syncopated already in PIE times to
*-é:i, *-é:u (*-ó:u) (oxytonic). the i- and u-stems
therefore maintained oxytonesis in the loc. sg. (and *not*
in the dat. sg. which reflects non-oxytonic *-éi-ei,
*-éu-ei/*-óu-ei). There subsequently was a moderate
analogical spread of end-stress in the loc.sg. (e.g. to the
a:-stems, where dat. and loc. sg. had merged phonetically,
and end-stress in the loc.sg. was a welcome means of
distinguishing the two cases again).
>> On pp. 64-67 of Stang's accentology, there is a complicated
>> argument concerning the stress of the Lithuanian illatives
>> and allatives, used by some as evidence that the Lith.
>> mobile paradigms had originally had final stress everywhere.
>> Looking at the whole thing from my point of view, I don't
>> see any problem.
>
>I agree emphatically that there is no problem. In Lithuanian there
>still is a partial connection between the phonetic weight of an
>ending and its stressability in mobile paradigms. Disyllabic endings
>are always stressed. So if new disyllabic endings would appear (as
>happened in the case of the illative etc.) they would risk becoming
>end-stressed by analogy even if built on an originally stem-stressed
>form like the acc.sg.
So we all agree that these forms were originally
end-stressed. The question I was trying to answer was why
the stress got retracted, something which doesn't happen
with other disyllabic endings like -imì, -umì, -yjè, and,
even defying Hirt's law, -omìs, -osù (> -osè). [Except the
dat. pl. forms (-áms, -óms, etc.), which I don't fully
understand].
=======================
Miguel Carrasquer Vidal
mcv@...