On Wed, 04 Jun 2003 16:26:10 +0200 (MET DST), Jens Elmegaard Rasmussen
<
jer@...> wrote:
>On Sun, 1 Jun 2003, Miguel Carrasquer wrote:
>
>>
>> As to the -w- forms in the 1pl., besides Hitt. 1sg. -mi, 1pl. -wen(i)
>> we have Luwian 1sg. -wi, 1pl. -man(i), so it's unlikely that this has
>> anything to do with the dual (also, we'd expect -a- instead of -e-
>> vocalism and no final -n(i) in Hittite if the 1pl. form were derived
>> from 1du. *-wh2ás or *-wh2á).
>
>In Melchert's Cuneiform Luvian Lexicon I find seven forms of the 1pl, all
>ending in -u-un-ni. Not a single one has an -m-.
I was going by Silvia Luraghi (in Ramat & Ramat), who gives 1pl. pret. -man
(CLuwian), -min (HLuwian). I don't the status of the CLuwian form, but a
paper by Craig Melchert (
http://www.unc.edu/~melchert/MINA2.pdf) claims
that -min(a) was misidentified by Morpurgo Davies as a 1pl. ending. I have
not had the time to read the article in depth, but the case seems to be
convincing. The HLuwian form is then -unni/-wunni (Melchert actually lists
ten 1pl. forms (one of them, nu:dummeni "we desire" may be Hittite, the
others are aruwarunni "we lift", hizzawunni "we hand over", hu:iunni "we
run", iunni "we go", lu:nni, lunni "we take", li:wunni "we ?", mammalhunni
"we crush", marsunni "we ?", piwunni "we give").
>The bracket statement about the expected -a- in Hittite escapes me; in
>fact the 1.2.pl. sometimes do end in -wan(i), -ttan(i), but I don't see
>where the -h2-'s come in. Should they be reversed so as to fit Skt. 2du
>-thas, 2pl -tha? And if so, where *does* *-me come from? Or was there no
>such form?
As you know, I have reconstructed the dual endings as *-mW-h2-ás > *-wás,
*-t(W)-h2-ás > *-thás, *-(h2)-t-és > *-tés. However, in the light of my
new discoveries about Nostratic, I have to revise that a bit.
The Proto-Nostratic dual marker was *-ik-. In the status constructus this
gave -iN, with ergative -iku, absolutive -iki. The stative was:
1du. *-ik-mi-N ~ *-ik-mi-ki
2du. *-ik-ti-N ~ *-ik-ti-ki
3du *-iN ~ *-iki
This gives Uralic -min', -tin', -k(in') (from the status constructus form),
and Eskimo-Aleut -kuk, -tek, -k (from the absolutive form).
In PIE, I had assumed the dual stative did not survive. However, it now
seems to me that it may have merged with the plural (active), derived from
the status constructus of the personal pronoun (plural marker *-an ~ *-atu
~ *-ati):
plural active + dual stative > PIE
*-mu-án *-(k)-míng *-m(W)én(i) > *-mén(i)
*-tu-án *-(k)-tíng *-t(W)én(i) > *-tér/*-téni
*-án *-íng *-(y)én-t(i) > *-ér/*-énti
In the 1/2 plural, the active forms can also use the ergative endings:
1pl. *-mu-átu > *-més
2pl. *-tu-á[tu] > *-té
3pl. *-án > *-ér/*-ént
(Cf. Samoyed, where the pattern is *-m&t, *-D&t, *-ton)
The dual active would have suffixed the ergative forms of the personal
pronoun:
1du. *-mu-íku > *-mWóh3 > *-wo:
2du. *-tu-íku > *-tWóh3 > *-to:
This corresponds with the Balto-Slavic present dual endings -va:, -ta: (OCS
1du. -vê instead of *-va due to 1du. pronoun *wéh1).
The other active forms apparently appended the plural markers *-és
(present) or *-é (perfect/past), to make:
*-mWh3ós > *-wós *-mWh3ó > *-wó
*-tWh3ós > *-thós *-tWh3ó > *-thó
*-t-és > *-tés *-t-é > *-té
This explains:
Goth Skt OCS
1du -u -vas -va
2du -ts -thas -tha
3du -- -tas -te < *-té or *-tés
(Goth. from thematic *-o-wV(z), *-e-tVz). The Skt. 2du. form -tha was
transferred to the plural, to create a distinction between present -tha and
preterite -ta. The expected Skt. 3du. past form -ta was replaced for the
same reason: a distinction had to be made between 2pl. and 3du. For this
purpose (already in PIE), forms from the dual middle (there were plenty of
them) were borrowed into the dual active paradigms. We can see they are
middle forms because they end in *-m (which only in the middle can be a 2/3
person ending):
Grk (past) Toch Skt.
1du. -- -- --
2du. -ton -ton -- -tam
3du. -ton -te:n -teM -ta:m
The real middle forms are:
Skt:middle pr. middle past Grk:m.pr m.past
1du -vahe: -vahi -- --
2du -ithe:/-a:the: -itha:m/-a:tha:m -sthon -sthon
3du -ite:/-a:te: -ita:m/-a:ta:m -sthon -sthe:n
We can reconstruct:
1du. *-wVdhVi (Skt. -vahe:), *-wVdhi/*-wVdh& (Skt. -vahi)
2du. *-tom (Grk. -ton, Skt. -tam), *-tHom (Grk -sthon), *-HtHoi (Skt.
-Hthe:), *-HtHaHm (Skt. -Htha:m)
3du. *-tom (Grk. -ton, Toch -teM), *-tHom (Grk. -sthon), *-Htoi (Skt.
-Hte:), *-taHm (Grk. -te:n, Skt. -ta:m), *-HtaHm (Skt. -Hta:m)
1du. middle (Skt. only) seems analogical after 1pl. -mahe:, -mahi, but it
could be directly from 1du.x2pl. *-h2mWodh > *-wodh- > -vah-e:/-vahi.
The 2/3 du. forms I would derive from:
2du. *-tom, *-tHom 2du.x1sg. *-h2t-o-m
2du. *-HtHoi 2du.x3sg. *-h2t-o-i
2du. *-HtHaHm 2du.x1du. *-h2t-a-h2m
3du. *-tom (*-tHom) 3du.x1sg. *-h2-t-o-m
3du. *-Htoi 3du.x3sg. *-h2-t-o-i
3du. *-HtaHm 3du.x1du. *-h2-t-a-h2m
With the middle dual markers *-h2t- (< **-kt-) for 2nd person and *-h2- (<
*-k-) + -t- for 3rd person, with the usual middle endings *-m (originally 1
sg./pl. indirect object) and *-0 (+-i) (3rd. person indirect object). Note
the ending *-h2m for 1du. indirect object.
Other possible dual middle forms must have been:
1du. *-h2mWor(i), *-h2mWo(i), *-h2mWah2r(i), *-h2mWah2(i)
2du. *-h2tah2(i)
3du. *-h2tor(i), *-h2tah2r(i), *-h2tah2(i), *-h2todh(i)
I'll assume the Skt. pf. dual forms (1du. -vá, -áthur, átur) are an
analogical cross between 2/3du. *-tHV, *-te and 3pl.pf. *-rs > -úr.
=======================
Miguel Carrasquer Vidal
mcv@...