On Thu, 29 Nov 2001 02:35:52, "Glen Gordon" <glengordon01@...>
wrote:

>First off, it's hard to discuss anything seriously when you
>only offer the following:
>
> 1) ProtoEskimo *-nka < *-t-m-ka
> (Can I buy a vowel, Vanna?)

Let's review the Yupik system again then shall we?

Intr. Poss.sg. Poss.du. Poss.pl.
1. -Na -ka -k-ka -n-ka
2. -ten -n -ke-n -te-n
3. -q -a -k -i

1. -ku-k -pu-k -k-pu-k -ppu-k
2. -te-k -te-k -k-te-k -tte-k
3. -k -a-k -ke-k -kke-k

1. -ku-t -pu-t -k-pu-t -ppu-t
2. -ci -ci -k-ci -cci
3. -t -a-t -ke-t -i-t

In the light of the Samoyed paradigm, we can reconstruct an original
regular system with object number markers *-0- (sg.), *-k- (du.),
*-t-/*-j (pl.), followed by subject markers *-m, *-t, *-sa, followed
by subject number markers *-0 (sg.), *-k (du.), *-t (pl.). In full:

Intr/Poss.sg. Poss.du. Poss.pl.
1. -0-m-0 -k-m-0 -t-m-0
2. -0-t-0 -k-t-0 -t-t-0
3. -0-sa-0 -k-sa-0 -i-sa-0

1. -0-me-k -k-me-k -t-me-k
2. -0-te-k -k-te-k -t-te-k
3. -0-sa-k -k-sa-k -i-sa-k

1. -0-me-t -k-me-t -t-me-t
2. -0-te-t -k-te-t -t-te-t
3. -0-sa-t -k-sa-t -i-sa-t

In the first person, *-ka has been added to or has replaced *-m-,
resulting in intr. -Na (*-m +-ka), poss.sg. -ka, poss.du. -k-ka,
poss.pl. -nka, where the -n- (as in Samoyed) results from *-t-m.

In the second person, the intr. has (nom.) -ten, (acc.) -ken, based on
the independent form *tken (Aleut txin) (Fortescue and Seefloth
reconstruct *tk-en < *tk-et, but in the light of the EA auslaut-laws
(*-tt > -n, *-t > -t, *-n > -n), I would prefer *tken-t). The
transitive forms end in -n, which points to *-tt, not *-t, so
Seefloth's analysis (*-n in free variation with *-t) must be rejected.
If the form were *-t, one might also have expected 2sg>pl. *-t-t to
have regularly developed into -ci (as in the 2pl.). The *-t has been
replaced with a more complex cluster, resulting in *-tt, *-k(e)tt,
*-t(e)tt (with cluster-buster schwas) -> -n, -ken, -ten. I would
argue this is the old stative *-tk (> *-tt).

In the third person, the intransitive (*-0) has added "singulative" -q
from the nouns and substantives (roots ending in -q drop it before the
dual and plural endings, allowing -q to be reanalyzed as a singular
ending). The objective ending *-sa becomes *-a, and then disappears
in the dual and plural, but underlies trans.sg. -a.

The 1du./pl. shows intr. -kuk, -kut vs. transitive -puk, -put. The
transitive forms are regularly from *-m&-k, *-m&-t (regressive
denasalization and progressive labialization). The intransitive forms
are apparently influenced by the independent forms in *t(k)-, as in
the 2sg. (Fortescue) or by analogy from the singular (Seefloth).

> 3) Aleut t`in < *tkin based on AfroAsiatic ???
> (Note: /ti-/ is _known_ to mean "here" as in
> /ti-N/ "I am here". There is no underlying *****tin-)

The Aleut independent pronouns are tiN, txin, tin, where tx- is the
second person marker (e.g. verbal -xtxen, pl. -xtic^e). They can be
derived from *ti-kk (or *tin-k), *tki-tt (or *tkin-t) and *ki-tt (or
*tin-0). What's the reference for "here"?

>Now for the less annoying remainder of your post:
>
> 1. In IndoTyrrhenian, the *-i plural is used residually
> for pronouns and numerals (IE *wei- "we", *trei- "three",
> dwix "two", etc). The idea of a constructus and absolutus
> plural might be worthy of thought but then, how does
> it connect with the alternation in the third person
> that you mention in plural object possessives?

In Uralic and EA, oblique plural *-ti developed into *-j in final
position, but remained as *-t- when followed by the person/number
markers in the objective conjugation. In PIE final *-ti also became
*-j, as in *trei- "three" < PN *tiláti. The opposition between nom.
*-atu and oblique *-ati is very clear in the pPIE personal pronouns,
where we have:

1. *mu *mu-átu, *mu-áti
*mu-má *mu-atu-má
2. *tu *tu-átu, *tu-áti
*tu-má *tu-atu-má
3. (*su)
*su-má *su-atu-má

Developing into:

1. (*mu) *mesW, *mWei
*mwé > *mé *nsmé [-> *nos]
2. *tu *sWesW > *usW, *sWei
*twé *sWsWmé > *usmé [-> *wos]
3. *swé *smé

(For the soundlaw *sW(e)sW > *usW, compare the numeral "6", Armenian
<vec`> < *usWeks, Greek dial. (w)eks, OPr. <uschts> "6th").

We have thus: Armenian mek`, Lith. mes (*mesW) vs. Goth. weis, Hitt.
we:s (*wei-es).
Gothic ju:s, Lith ju:s (*(y)usW) vs. Celtic *xwei-, Hitt sumes
(*swei-es), etc.

> 2. Back up your unsubstantiated assertion that
> Nenets /-v/ < *-m with real examples. It's funny but I
> just don't recall there being an accusative in */-v/
> (Nb: Uralic *-m).

Mikola:

"Das PxSg1 *-mV ist in allen sam. Sprachen erhalten, im Tundranenz.
wurde es allerdings teilweise vom PxDu1 (*-min') verdraengt, auch das
waldnenz. PxSg1 stammt aus dem Dualparadigma, wo es urspruenglich das
Px fuer den Gen. war. Aehnliche Prozesse gingen auch im Enz. vor sich
(-b', -j).
PS -mV > nenz. T -w, (-m'i), W -m, (-j) | enz. Ch -bo, B -o, (Te)
(-b', -j), (Pr) (-j?), -mo | ngan. -m& | selk. N -m, (Pr, Hel) -m/-p,
-mI, M -m, -p, S -ù, -u | kam. -m, koib. -m, mot. -m, -ma, tg. -m."

etc.


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