On Fri, 04 Jul 2003 06:04:20 +0000, Rob <magwich78@...> wrote:

>Proto-Uralic canonical word-
>order is reconstructed as SOV, which means that it had postpositions
>and modifiers preceding their heads. This is at odds, however, with
>the fact that PU also had possessive suffixes, since its word-order
>means that possessors would precede their possessions.

This doesn't apply to possessive suffixes. SOV languages have the order
Poss N, but NPoss (which can even be combined as Poss NPoss, as exemplified
in e.g. Turkish (Comrie p. 90): kadIn-In çavug-u "the woman's chicken",
literally <woman-GEN chicken-POSS> "woman-of chicken-her").

>If I may point out my reconstructed PU sentences again, we have:
>
>Weti ma elä mi, sä kala ma pexi.
>
>A literal translation would be "Water at carry me, (s)he fish at
>cook." Allowing for the possibility I mentioned above, and the use
>of *ma as an object marker, we have a somewhat smoother
>translation: "My carry(ing) water, (s)he cook(s) fish." From here, I
>think it's just a short step to an even better translation: "As I
>carry the water, she cooks the fish." Do you see the development
>here? This is a different argument than I presented before, and I
>think it's more plausible.

We can also take *ma as a genitive affix (as it is in Eskimo-Aleut, and in
the PIE personal pronouns). "She cooks the fish" is then literally "her
cooking of the fish" (kala-ma pexi-sä), and likewise "my carrying of the
water" (weti-mä kanta-mi) [*elä- is given by Sammallahti as PU for "to
live", the PU for "to carry" is *kanI-, *kanta-].


my house/my carrying *kota-mi *kanta-mi (=I carry it)
your house/your carrying *kota-di *kanta-di (=you carry it)
my houses/my carryings *kotad-mi *kantad-mi (=I carry them)
your houses/your carryings *kotad-di *kantad-di (=you carry them)
our house/our carrying *kota-mit *kanta-mit (=we carry it)
your house/your carrying *kota-dit *kanta-dit (=you carry it)
our houses/our carryings *kotad-mit *kantad-mit (=we carry them)
your houses/your carryings *kotad-dit *kantad-dit (=you carry them)

For the third person, things are a little bit more complex.

Given the fact that the possessum plural shows an ending *-j (plural
oblique in the absolute Auslaut), the original construction must have been
something like:

woman's house *näxi-GEN kota
woman's houses *näxi-GEN kotaj
women's house *näxi-GENpl kota
women's houses *näxi-GENpl kotaj

If the 3rd.person possessive had been added "redundantly" (as in Turkish)
we would have expected:

woman's house *näxi-n kota-sa
woman's houses *näxi-n kotaj-sa
women's house *näxi-j kota-san
women's houses *näxi-j kotaj-san

What we actually see, however, is that the possessor is in the nominative:

woman's house *näxi kota-sa
woman's houses *näxi kotaj-sa
women's house *näxit kota-san
women's houses *näxit kotaj-san

Literally: "the woman (women), her house(s)". Either the construction was
completely altered (gen. marker on the possessor replaced by poss.suffix on
the possessum), or the genitive was here originally marked by a suffix that
eventually gave -0 in the sg. *-t in the plural, i.e., merged with the
nominative.

>Regarding the development of possessive suffixes in the Uralic
>languages, I readily admit that your reconstruction is certainly
>possible. However, I still disagree that the Finnish plural
>pronominal suffixes had an ending in *-k. The *-k suffix forms the
>dual number; why would it also be employed as a pronominal plural
>(assuming that both are the same suffix)?

I don't think they are the same suffix. In the verbal/possessive/stative
endings, the dual is marked by *-in(') (Saami -&n, Ob-Ugric -&n, Nenets
-in, Selqup -i:(j)).

The reconstruction of the Finnic plural forms as 1pl. *-(m)mek, 2pl.
*-(t)tek is as far as I know uncontroversial in Uralistics. Saami shows
the -k in its plural forms, and even though Saami -k < *-k or *-t, Finnish
would not have dropped -t. Mari -na and -ta are also explained as *-nak,
*-tak. Mordvin shows the -k in the possessive forms (-mok, -nok, -nk), but
has dropped it in the stative/intransitive (-tano, -tado < *-t&-m&k,
*-t&-d&k). The Permic languages have dropped most final consonants, so
they tell us nothing. Hungarian has generalized -k as the plural marker
everywhere. Mansi has zero in the verbal endings (1pl. -w&, 2pl. -n&),
presumably from *-k, but *-n in the possessive. Khanty has the ending -Gw
in all 1pl. forms, the 2pl. has -(t&)n, except in the intransitive
conjugation, where it has -t&G. The 3pl. has -L, but -t in the
intransitive. Samoyed has -t in the 1/2pl., but -n in the 3pl.

The plural markers are thus:

Finn Saami Mari Mordv Permic Hung Mansi Xanty Samoyed
poss
1/2 -0 -k -0 -k -0 -k -n -n -t
3 -0 -k -t -t,-k -0 -k -n -L -n
verbal
1/2 -0 -k -0 -0
3 -t -n,-k -t -0
intr
1/2 -0 -k -0 -G -t
3 -t -k -t -t -n
trans
1/2 -k -k -0 -n -t
3 -t(?) -k -n -L -n


The data can be resumed as:

stative/intransitive:
*-k/*-t (Finn,Saam,Mari,Mord,Mans,Xant)
[*-k/*-n (Saami) ]
[*-t/*-n (Samoyed) < tr.]

possessive/transitive:
*-k/*-k (Finn,Saam?,(Mord),(Hung))
[*-k/*-t (Finn,Saam?,Mari,Mord) < intr.]
*-k/*-n (Saam,Mans)
*-n/*-n (Mans)
*-n/*-L (Xant) [< *-n -n' ???]
*-t/*-n (Samoyed)

It's not easy to come up with a Proto-Uralic reconstruction. The
intransitive forms can be derived from *-k, *-t (Saami 3pl. -n is
transitive, Samoyed has generalized transitive *-t/*-n). The 3pl.
intransitive/stative must necessarily be identical with the nom.pl. of the
noun, so *-t is as expected. The *-k on the 1/2pl. forms can perhaps be
explained as an echo of the *-k of the stative personal endings (cf. PIE
1sg. *-h2 < *-k, 2sg. *-th2 < *-tk, 2pl. *-dhw- < *-t-ku).

The transitive must then come from 1/2pl. *-t, 3pl. *-n, as in Samoyed. In
Ob-Ugric, *-n was generalized in the possessive. In the 1/2pl. (and
sometimes in the 3pl. as well) *-k from the intransitive paradigms was
introduced in Finno(-Permic), Hungarian, and also in the Mansi verbal
forms. The 1/2pl. marker *-t is identical with the nominal plural marker
(nominative) *-t. 3pl. *-n is a pronominal plural marker (cf. the Finnic
pronouns with sg. t-, pl. n-).

>I am still skeptical that PU had a separate stative paradigm.

If Proto-Uralic did not distinguish verbs from nouns, it follows that
personal endings could be attached to nouns as well, as is still the case
in Samoyed and Mordvin (which have a stative). Hungarian and Ob-Ugric
don't have a stative, but do have a separate intransitive conjugation.

The ending 1sg. stative *-k is weakly attested (it was replaced in Northern
Samoyed by *-da-m(&), in Mordvin and Ob-Ugric (and sometimes in Hungarian)
by *-m), but the presence of Hungarian 1sg. stative/intransitive -k, Selqup
-k cannot be a coincidence in the light of Eskimo-Aleut 1sg. -ng (< *-k),
-k-a; PIE 1sg. stative *-h2; PAA 1sg. stative *-kV, etc.

Stative 2sg. *-n(g) is attested in Samoyed, Komi, Ob-Ugric, and may also
underlie Hung. -sz. I believe it reflects the Proto-Nostratic 2sg. stative
ending *-tkV ~ *-tk (by the nasal rule > *-nng > *-n&ng, *-ng, *-n).

The stative third person forms are of course identical to the sg., du. and
pl. of the noun (*-0, *-k [or *-gin'], *-t).

The 1/2 du. and pl. forms of the stative had the personal endings attached
to the du. / pl. forms of the noun/verb, which explains the Finnish pl.
forms in -mme, -tte (pl *-d-/-t- + personal endings). Mordvin has
interposed a connecting vowel: 1pl. stative *-t&-mo(k), 2pl. stative
*-t&-do(k).

I have exposed above the reasons why I think the plural marker *-k on
verbal/possessive endings may have originally arisen in the stative (this
is a Uralic innovation, AFAIK it has no parallel in other Nostratic lgs.)


=======================
Miguel Carrasquer Vidal
mcv@...