--- Piotr Gasiorowski <
gpiotr@...> wrote:
> Sean Whalen wrote:
>
> > The accent changes in verb par@.../p@...@t- "fight"
> > (Sanskrit prt- "battle" but there's no change in
> rt.
>
> As far as I can see, _all_ the Avestan examples of
> finite verb forms
> derived from this root are thematic middles with
> non-initial accent, so
> they simply fit the rule.
You've also explained examples that don't fit the
rule as the result of analogy. I don't know if you
agree with the following, but:
ml,d-doxW- "give softness > forgive, pardon"
ml.,d-doxW-
ml.,z-doxW-
ml.,z.-doxW-
ml.,z.-d.oxW-
m@...@da: - "forgive"; m@...@da "pardon"; m@...@dika-
"pity" (Avestan)
mr,d. - "forgive, pardon"; mr,d.i:ka- "favor"
(Sanskrit)
This change of d>z between l. (or r.) and a C plus
change d>z and t>s after g/gH/etc. makes it seem
likely to me that an optional change t>s after r.
could occur at this time in Avestan. There are very
few instances of s after r from inherited words so the
gap (many s>s. after i, u, k, g, etc., but almost no r
or l) could be filled by this change.
There is no r.-deletion before s. in the same
syllable (a:tar "fire (nom.)", tarna- "thirst") so
rule hrt> can be argued against if there's an example
of rt at the end of a syllable becoming r with no
coalescence. In compounds: ar-vacaNh@... "true speech
(acc.)" ar-uxða- "true-spoken" seem likely.
__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com