One could also postulate *-a:ns, *-o:ns,
*-ons > *-aNx > *uNx
> *u:, collapsing the last two stages with the development of the Acc.pl. of
u-stems (*su:nuns > *syny) and thus making for greater
parsimony.
As for the Acc.pl. of *jo- and *-ja: stems,
that is North Slavic *-je^ corresponding to South Slavic *-jeN, we could
have
*-jons, *-ja:ns > *-jaNx > *-jeNx
> *-je: (with dialectal denasalisation) ~ *-jeN
Some authors (e.g. Van Wijk and Kortlandt)
attempt to trace the South Slavic present participle in *-y and Northern forms
in *-a back to a common prototype. The former is easy to derive from PIE *-onts
if, as elsewhere, *ts > pre-Slavic *s (*-onts > *-ans > ... > *-y),
but the participle in *-a is rather puzzling.
Milewski (1948) proposes the following
scenario (for my taste, it involves far too much
prestidigitation):
*-o:nt(s) > *-a:nt > *-a:t (yielding
-a), dialectally *-a:(t)s (with analogic *s) > *-a:s (yielding
-y)
He also suggests (plausibly this time) that
Old Polish -eN (which is many times more frequent than -a) derives from neuter
*-ont (> *-aNt > *-uNt > *-oN), and assumes that the neuter form
was generalised for participles in *-jont-, which show a nasal (*-jeN)
throughout the Slavic branch. It may have been so, but an original *-jonts >
*-jaNs would have ended up as *-jeN anyway, at least in South
Slavic.
To conclude, I'd propose:
*-ont-s (masculine) > *-aNx > ...
> *-y
*-ont-i: (feminine) > *-aNti: >
*-oNtji (with *j of analogical origin)
*-ont (neuter) > *-aNt > ... >
*-oN
*-j-ont-s > ... > *-jeN (~ *je:?
perh. in Czech znaje, Russ. znaja, etc.)
*-j-ont-i: > ... >
*-jeNtji
*-j-ont > ... > *jeN
This leaves -a unexplained. Any
suggestions? The only possibility that I can think of is that the *-a of *znaja
< *zna:-ja: < *zna:-je: was generalised as the Nom.sg. ending of masculine
participles in some (not all) North Slavic dialects.
Piotr
P.S. I'd like to return to your
argument that forms like OCS robotU contain a reinforced jer in the penult
syllable. As argued already by Rozwadowski (1914), this explanation doesn't take
into account the dialectal distribution of the -o- forms. The development of
reinforced *U into o would not be surprising in West
Bulgarian/Macedonian dialects, but this -o- is also found in dialects that
generally preserve the back jer unchanged (West Bulgarian) as well as those in
which *U > e in strong positions, cf. Old Czech vec^eros
'tonight'.
I regard it as possible that *-x was lost
early in pre-clitic sandhi: *-ax#C- > *-a#C-. Perhaps the loss was
dissimilatory, e.g. *-ax#tax > *-a#tax > *-atux > *-otU. A regular
Nom.s.g in *-U could certainly be restored on the analogy of the isolation form,
but *-o survived in more-or-less idiomatic combinations.
It's more or less like the loss of "the"
after English "at" through phonetic simplification (at the > atte > at,
e.g. Chaucers atte beste > Modern English at best), subsequently reversed
everywhere except in fixed phrases (at home, at table, at stake, at least,
etc.). Shakespeare still used idioms like "at very heart/root". (On the other
hand, the Authorised Version has many instances of reintroduced "the" which
eventually didn't catch on: "at the last/length/first" etc.).
Piotr
----- Original Message -----
Sent: Saturday, November 25, 2000 3:01
AM
Subject: Re: [tied] Slavic
endings
On Fri, 24 Nov 2000 09:34:15 +0100, "Piotr
Gasiorowski"
<gpiotr@...>
wrote:
>You know what, Miguel? This collaborative result is more
reasonable than any explanation of Slavic endings I've seen in the literature.
It works in other respects, too, by accounting for *-a:s > *-a:x > *-u:x
> *-u: > *-y and several other inflectional problems as well. There's
nothing like a good dispute.
Indeed. Even in the more common case
that no agreement is reached,
it's still worthwile (not to mention fun), if
the dispute is good and
one can learn from one's "opponents".
To
continue the discussion: what I'm missing in the above is the nasal
element
that lurks somewhere in the development *-a:s > *-y. It's
obviously
there in the acc. plurals *-a:ns and *-ons (as well as in
the n-stem nom.sg.
-y < *-o:ns), but also, as you know, in the gen.sg.
and nom.pl. [if this
is not the acc. form] of the a:-stems, judging by
the ja:-stem forms in
-je~. At least in South Slavic (W. and E.
Slavic have *-e^).
We
have:
n-stems nom.sg. *-o:ns > *-o:~s >
*-a:~x
o-stems acc.pl. *-ons > *-o:~s >
*-a:~x
a:-stems acc.pl. *-a:ns > *-a:~s > *-a:~x
a:-stems
nom.pl. *-a:s
> *-a:x
a:-stems gen.sg.
*-a:s >
*-a:x
In South Slavic, the two remaining distinct endings seem to
have
merged as *-a:~x, by some kind of intrusive nasalization
(recalling
Sanskrit forms like nom.pl. -(y/v)a:m.s-as from roots in *-yas
and
*-was [i.e. comparatives and the pf.ptc.act.]).
For the rest, the
hard forms developed as you say: *-a:(~)x > *-u:x >
*-u: >
*-y.
The soft forms (whether in S. or W./E. Slavic) have *-ja:(~)x
>
*-je:(~)x, which I find somewhat puzzling (unlike *ja > *je, *ja:
does
not give *je:). Undoubtedly another effect of the *-x, but
here
clearly raising rather than backing, as I had described it earlier.