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

--- In cybalist@yahoogroups.com, Piotr Gasiorowski <gpiotr@...> wrote:

>
> On 2007-01-23 17:05, mandicdavid wrote:
>
> > Not all yers became 'weak'. First they became lax (similar to

English

> > vowels in 'big' and 'pull'). Later some of them shifted to e/o

(cf.

> > the development of Latin short i and u) or schwa, which in some
> > languages subsequently yielded a.
> > The rest of them disappeared. Something similar happened in some
> > Croatian dialects, where the short i is frequently reduced or
> > dropped: vid (2sg imperative: look!) etc.
>
> It's the same in Polish. One extreme case is *vUz-Imi (Russ.

voz'mí) >

> wez'mi > wez'm > wez' meaning 'take!', where the palatalisation of

the

> final consonant of the "stranded" prefix is the only synchronic

trace of

> the verb root.

I think this didn't affect the Croatian form 'u`zmi' - unlike the
Polish, we don't 'like' complicated consonant clusters.
On the other hand, in some Croatian dialects there's an opposition
between the high and low tone in forms such as high-toned 'vid'
(instead of Standard vi``di, 'look!') and low-toned 'rec' (for ST.
r`eci 'say!').
Normally, there are rising and falling tones in Croatian, the rising
one arose as a result of a retraction of stress to the preceeding
syllable, while the high tone remained on the formerly stressed
syllable. Since the 'i' of r`eci (<recí) vanished, only the low tone
of the stressed syllable signalises the former presence of an ending.

>
> > This probably has something to do with metrical properties of

words

> > in PSl.
>
> There's a phenomenon known as Havlík's Law, operating to a varying
> extent in the separate languages: if there's a sequence of open
> syllables containing jers, every other jer is dropped beginning

with the

> last one (if final of followed by a syllable containing a full

vowel).

> In Old Polish Havlík's Law operated very regularly, producing
> odd-looking alternations like *xrIbItU/*xrIbIta > chrzbiet ~
> grzbiet/chrzepta 'back, ridge', for the most part levelled out in

Modern

> Polish (grzbiet/grzbietu).

I know of the Havlík's Law. For *xrIbItU, xrIbIta, we have hrbat,
hrpta 'ridge' - the 'r' is syllabic here - we always render RI/U as
syllabic R (krv, prst, trpjeti...).

In Croatian this law had a lot of exceptions: magla (mIgla), dani
(dIni), tama/tma (tUma)... In chakavian dialects they're even more
numerous: kade (kUdì, standard gdje), malin (stand. mlin, 'mill')...

>
> > The stressed yer weren't 'weak' - they were ordinary lax vowels.
> >
> > What I don't understand is how the word-final yers could

disappear

> > even if they were stressed.
>
> First, stress was retracted from them in accordance with the

trochaic

> prominence pattern that operated in strings of reduced syllables.
> There's something not unlike Havlík's Law in the treatment of

French

> schwa, which also has this tendency towards alternate loss and
> reinforcement.

Can you explain this 'trochaic prominence' issue. I'm not familiar
with it. Thanx.

David