--- In
cybalist@yahoogroups.com, Piotr Gasiorowski <gpiotr@...>
wrote:
> Rob wrote:
>
> > What's your evidence? The forms you cited? Or is there more?
>
> There are a few more, like Skt. bHartám vs. the RV 1sg. aor.
> abHa:rs.am, subj. bHars.at
What is the t-element in _bhartám_? Again, the RV 1sg. aor.
_abha:rs.am_ seems to come from IE *e-bhe:rsm. The subjunctive form
_bhars.at_ may indeed reflect the earlier state of affairs, as it
seems to come from *bherset, meaning that the lengthened-grade was
not inherent in the root. Notice that, in the aorist subjunctive,
the /s/ is not in coda position and thus cannot cause any
compensatory lengthening. So I still conclude, as Sihler does, that
the lengthened-grade of the sigmatic aorist comes from lengthened
vowels of the 2sg and 3sg forms (e.g. *bherss > *bhe:rs, *bherst >
*bhe:rst) spread analogically to all persons and numbers. The
original paradigm would then have been (again with *bher-):
1sg *bhérsm
2sg *bhérss
3sg *bhérst
1pl *bhrsmés ~ *bhrsmén
2pl *bhrstés ~ *bhrstén
3pl *bhérsnt ~ *bhrsént
- Rob