Re: [tied] Re: Vrddhi in sigmatic aorist

On Tue, 30 Oct 2001 14:06:42 -0000, "Sergejus Tarasovas"
<S.Tarasovas@...> wrote:

>--- In cybalist@..., Miguel Carrasquer Vidal <mcv@...> wrote:
>> My conclusion would be that the lengthening originates in 2 & 3 sg.
>> active only: *bher-s-s and *bher-s-t becoming *bhe:rs and *bhe:rst.
>
>I understand *bher-s-s > *bher-r-s > *bhe:rs. But how would you apply
>that model to *bher-s-t? > *bher-r-t > *bhe:r-t > (analogical
>restitution of -s-) *bhe:r-s-t?

Contrary to Szemerényi, I don't think the /s/ has got to go to effect
the lengthening (cf. Rasmussen's counterexamples *gWo:us and *wo:kws).

>>In OCS, we have the strange situation that the
>> paradigm shows vrddhied forms everywhere *but* in the 2 & 3 sg.
>> (ne^sU, nese, nese, ne^somU, ne^ste, ne^se,), as these forms have
>been
>> substituted by thematic root-aorists (nese < *nes-e-s, *nes-e-t).
>
>Some time ago Piotr argued *r and other sonorants block the RUKI-rule
>in Slavic (*usro > *(j)ustro, not **(j)uxro). But what about Slavic
>aorists like *re^xU (< *re:ksm.)? Why *m doesn't block?

It's the vocalic allophone here, which may be significant.