On Mon, 26 May 2003 03:01:07 +0200, Miguel Carrasquer <
mcv@...>
wrote:
>I don't think I'm re-analyzing all that much. The pronominal ablative
>ending was surely *-ot (as shown by Skt. mát, tvát), which would have
>given *e-ot, *kWe-ot as well as *to-ot, *kWo-ot (besides, to be sure,
>*esmo:t, *kWesmo:t, *tosmo:t and *kWosmo:t). That *to-ot and *kWo-ot
>yielded contracted *tõ:t and *kWõ:t is as expected, but the same does
>not apply necessarily to *eot and *kWeot. I see no reason why the
>Latin ablative eo: (despite its analogical long vowel) can't go back
>to PIE *eot.
PIE *eot(i) is attested as:
Skt. á:t
Lat. eo:
OCS jego
Lith jõ:
Hitt e:z
In Indo-Iranian, *eot became *aat, which was naturally contracted to
á:t.
Latin and Lithuanian have *eot > *eo:t, by analogy from o-stem *-o:t,
pronominal *to:t, *kWo:t. Perhaps South Slav jega also reflects *eo:.
Hittite has contraction *eo > e:, as opposed to *o: > a(:), as in the
regular ablative -az (< *-o:ti) [-e:z was subsequently transferred to
all pronominal paradigms at the expense of -az].
So we find more or less regular reflexes of *eot in 4 out of 10
branches of PIE, in fact in all the branches that have retained the
ablative. If I may use strong language: there is really really not a
shadow of doubt in my mind that this solves the problem of Slavic
jego.
=======================
Miguel Carrasquer Vidal
mcv@...