--- In
cybalist@yahoogroups.com, Mate Kapoviæ <mkapovic@...> wrote:
> Wouldn't be so sure about Slav. dì´´verI with an acute. Derivatives
> (which rarely if ever go secondary) point to the circumflex - dì^verI.
Zaliznjak (p. 138) also gives a.p. c, while Dybo assumes both a.p. a
and a.p. c for the word (Slavjanskaja...: a.p. a on p. 18, but a.p. c
on p. 158; Morfonologizirovannyje...: both a.p. a and c on p. 114). If
we dismiss a.p. a as secondary, we would have to explain away the
Latvian word as secondary as well. Lithuanian is ambiguous: while DLKZ^
(mostly prescriptive by nature) gives both a.c. 1 and a.c. 3a, LKZ^
(trying to be purely descriptive) gives a.c. 3a only. That would mean
that DLKZ^'s a.c. 1 is a codified urban speech's innovation (the
puristic Vitkauskas' orthoepic dictionary gives 3a only, Vitkauskas
being an outstanding expert on the Lithuanian dialectal lexis), if it
were not a recently compiled (and thus not reflected in LKZ^)
dictionary of an archaic dialect of Zietela (an isolated Lithuanian
enclave in Belarus, now completely assimilated), which registers both
3a and 1 (with relevant examples). Another dialectal dictionary I got
to hand (that of a southern dialect of Dievenis^kiai) gives 3a only.
It looks like the word vacillated between mobile and fixed-stress
paradigms in Balto-Slavic.
Sergei