On Wed, 04 Jun 2003 23:27:40 +0200, Miguel Carrasquer <
mcv@...> wrote:
>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 just realize that 1du.x3du. *-h2mWah2 > *-wa: and 2du.x3du. *-h2táh2 >
*-t(H)á: may also underlie Balto-Slavic *-wa: and *-ta:, so that there is
no need for the unextended dual forms *-mWóh3 and *-t(W)óh3. If at least
Slavic 2du. -ta is from middle *-h2tah2, then that would free Lithuanian
-va: and -ta: to be from *-wós and *-tHós, in the same unspecified way that
Lithuanian -me: and -te: are likely to be from *-més and *-tés (and 2sg. -ì
< *-e:i < *-esi ??).
In that case, we only have two active variants:
Goth Skt OCS Lith
*-mWh3ós > *-wós -u ? -vas -va:
*-tWh3ós > *-thós -ts -thas -ta:
*-t-és > *-tés -tas -te ?
*-mWh3ó > *-wó -u ? -va
*-tWh3ó > *-thó (-tha)
*-t-é > *-té -te ?
And all other forms are from the middle dual, or analogical.
=======================
Miguel Carrasquer Vidal
mcv@...