Re: Balto-Slavic C-stems / long vowel endings

--- In cybalist@yahoogroups.com, "mcarrasquer" <miguelc@...> wrote:

>
> --- In cybalist@yahoogroups.com, "tgpedersen" <tgpedersen@> wrote:
> >
> > --- In cybalist@yahoogroups.com, "mandicdavid" davidmandic@ wrote:
> > >
> > > Something similar happened in some
> > > Croatian dialects, where the short i is frequently reduced or
> > > dropped: vid (2sg imperative: look!) etc.
> > > This probably has something to do with metrical properties of
> > > words
> > > in PSl.
> >
> > cf Russian derz^í!, vs búd'!, PSl *-í vs *´-I,
>
> They are both PSl. -i.

Why the jer in Russian then? How do you know OCS bo,di isn't analogical?

> > ie PIE *-éi vs *´-i.
>
> PIE *-oih1-s

Leskien: Handbuch der altbulgarischen Sprache
"
Im Imperativ endet der Stamm (das Element vor den Personalendungen) im
Singular auf -i-, im Plural auf -ê- (beides = indog. oi); dies
Verhältnis bleibt ungestört in Kl. I und II: 2. sg. nesi, 2. pl.
nesête; 2. sg. dvigni, 2. pl. dvignête; in Kl. III müssen die Plural-
und Dualformen wegen der vorangehenden palatalen Konsonanten (nach §
26) statt -ê- ein -i- erhalten: 2. sg. pis^i, 2. plur. pis^ite (über
eine Nebenform pis^ate (s. § 131 a). Bei den Verben der Kl. IV geht
-i- durch alle Personen, z. B. 2. sg. xvali, 2. plur. xvalite. Über
die besondern Imperativformen der Kl. V s. § 140fg.
"

I realize that the consensus since Leskien is that the Slavic
imperative is the PIE optative, which must be the reason for the
curious -s of the PIE *-oih1-s you cite. However, in Sanskrit it is
only the 2sg impv that has its own form, the 2pl impv is an
injunctive, and similarly it might only have been the 2pl of Slavic
which used the optative (at least in classes I and II), while thematic
2sg impv might be from *-e-i).

> > I think there is something methodologically wrong with a language
> > with 'normal' and 'supershort' vowels. Aren't linguists
> > backprojecting the present state of affairs onto PSlav. (or even
> > to ChSl.)? How about renaming them 'long' and 'short' as they are
> > named in any other language with two vowel lengths, which would
> > make PSlav i/u into i:/u: and I/U into i/u?
>
> That was indeed the state of affairs _before_ the rise of the yers.
> In late Common Slavic, however, former /i:/ and /au/ > /u:/ had been
> shortened to /i/ and /u/ in certain contexts, and we had long /i:/,
> short /i/ and short lax /I/ (likewise /u:/, /u/, /U/).

Estonian has super-long V::, long V:, and short V. My little book on
Low German claims the same for that language. Slavic is somewhere in
the middle, geographically.

> > That means stressed 'jers' (I question-mark them
> > now, they are i/u) were not reduced to ghostly I/U, like the
> > unstressed ones were,
>
> All /i/ and /u/ were laxed, independently of the place of the
> stress.
>
> > and with that formulation we need no
> > 'reinforcing' of stressed 'jers'
>
> There is no "reinforcing of stressed yers".

That's what I keep telling Piotr ;-)

> Yers are strong or weak according to Havlík's law, which works
> independently of the place of the stress. Before Havlík's law, all
> stressed yers had already lost the stress to the preceding vowel. Of
> course, that vowel could be a yer, so there were still some stressed
> yers left (there would have to be, in words with nothing but yers).

In other words, it moved to the preceding syllable, and if that
contained a jer and was not the first syllable, it moved again,
iteratively?


Torsten