On Wed, 18 Aug 2004 10:31:55 +0000, tgpedersen
<
tgpedersen@...> wrote:
>Some say there was another locative suffix -u (based on the loc.pl.
>-si, -su).
I don't. I say it's the same one (*-i), affected by Umlaut
after a *w, *xW (= *h3) or *sW (the original shape of the
plural suffix).
We have -i > -u in the loc.sg. of the u-stems (*-ew-i/*-ow-i
-> *-ewu/*-owu -> *-e:w/*-o:w; cf. in the i-stems *-ey-i ->
*-e:y). We have -i > -u in the dual (e.g. o-stem *-oy-h3-i
-> *-oyh3u -> Grk. -oiiu-n, -oiin). And we have -i > -u in
the plural *-sW-i > -su (Greek -si) [confirming that the
Armenian plural suffix -k` directly continues *-(e)sW, with
*sw > k` as is regular in Armenian].
>Perhaps the -u was used in the secondary inflection and later lost?
>The OCS I left its mark in Russian by palatalisation, the U left
>nothing; something similar is the case in Japanese. Would an -u
>explain the -U of 3rd sg. in OCS?
The inherited 3sg. ending is *-tI, as in Old Russian, modern
Ukrainian, etc.
The Old Novgorodian beech-bark inscriptions have both -tI
and -0 (-tU appears only late in the 14th. century, probably
under Russian/Muscovian influence). Zaliznjak says (p.
119):
"We can see some statistical correlation between the choice
of endings with or without -tI and the type of sentence: the
highest percentage of examples without -tI is found in
phrases that express a condition (introduced by a special
conditional conjunction or simply by the conjunction <a>);
it's also high in supplementary (pridatochnyj) dependent
clauses as well as in intentional (celevoj) and explanatory
(iz'jasnitel'nyj) ones; meanwhile in main or simple clauses
the share of examples with -tI and without -tI is roughly
the same."
This can be explained if -tI contines the old present
indicative *-(e)ti, while -0 (-e) continues the old
subjunctive *-et (there is another correlation between
endings without -tI and conjugation class: the zero endings
are more common in the thematic class).
The variation between -0 and -tU is comparable to the
variation between <ja> and <jazU> "I". Both *-t and *-g^ (>
-z) remained unaltered in Slavic until very late (in
contrast with other final consonants, which were dropped
early on, *-d even before Winter's Law). We therefore had
*jaz "I" and *beret "(that/if) he carries". When the
open-syllable rule finally imposed itself, such final
consonants were either dropped (ja, bere) or acquired a
prosthetic -U (jazU, beretU), or both. The old indicative
*bereti remains as beretI.
=======================
Miguel Carrasquer Vidal
mcv@...