On Sun, 17 Feb 2008 13:34:55 -0000, "alexandru_mg3"
<
alexandru_mg3@...> wrote:
>--- In cybalist@yahoogroups.com, Miguel Carrasquer Vidal
><miguelc@...> wrote:
>>
>> On Sat, 16 Feb 2008 15:52:08 -0800 (PST), Rick McCallister
>> <gabaroo6958@...> wrote:
>>
>> >Thanx for explaining. No such law or traces in Gmc?
>>
>> Dybo did an article on vowel shortenings in Italo-Celtic and
>> Germanic, which look like a kind of mirror-image of Hirt's
>> law (where Balto-Slavic retracted the stress from e.g.
>> *suh1nús to *súh1nus, Germanic shortened the vowel in
>> *sunuz). The problem is that there are a number of
>> exceptions in Italo-Celtic with unexpected long vowel
>> (fu:mus is the one that comes to mind). In Germanic, the
>> rule only works if there is resonant between the laryngeal
>> and the stressed vowel.
>>
>> Dybo's conjecture is well worth a re-evaluation. Does
>> anyone have the original article at hand?
>
>
>The only free article on Internet that resumes this is
>
>http://users.ox.ac.uk/~mert1033/Dybo's%20law%20an%20ItalicCelticGermanic%20sound%20change.pdf
>
>
>Important is Schrijver(1991) reformulation for Germanic Italic and
>Celtic:
>
>"Any long vowel in a praetonic syllable is shortened before a
>Reasonant" ...but even so there are exceptions
>
>If you could find Dybo or Schrijver original articles please to tell
>me too.
Mult,umesc for the link, I hadn't seen it.
The matter is also discussed by Kortlandt here:
http://www.kortlandt.nl/publications/sa/sa_appc.pdf
http://www.kortlandt.nl/publications/sa/sa_appd.pdf
A few things to consider: the articles by Dybo,
Illich-Svitych, and Kortlandt's dissertation were written
before the discovery of Winter's law. This explains a number
of cases where Balto-Slavic has length but no retraction,
and Italo-Celtic(-Germanic) has no length.
The Francis-Normier law had, I think, also not been
discovered at the time. This law states that sequences of
syllabic resonant + *h2/*h3 were "broken" in Greek, Armenian
and Tocharian. That is, *r.h2 > ra:, *l.h2 > la:, *m.h2 >
ma:, *n.h2 > na:, *ih2 > ya:, *uh2 > wa: (the same for *h3:
*ih3 > yo:, *uh3 > wo:). It is my belief that the lack of
stress-retraction in cases like Slavic bylá (*bhuh2láh2),
z^ilá (*gWih3láh2) can be explained by this same phenomenon:
in Proto-Balto-Slavic, these forms were *bu&láh2, gi&láh2,
with vocalized laryngeal, and therefore no Hirt's law.
Subsequently, *i& and *u& were simplified to *i: and *u:
(with acute intonation), but the accent had remained mobile
(later still, Meillet's law caused the acute to become
circumflex in Slavic, yielding b^ylU, z^îlU, etc. This may
also be relevant for the Italo-Celtic and Germanic forms.
If Lat. ferus was shortened by Dybo's law, then the law
differs fundamentally from Hirt's law, which only retracts
the accent when a laryngeal is present in the pretonic
syllable. The Baltic forms z^ve:rís, zvêrs and Slavic zvê^rI
clearly point to a Balto-Slavic mobile *z'we:rís, acc.
*z'wé:rim, with a long vowel and no laryngeal. The word was
originally a C-stem *g^hwé:r "wild animal" (Grk. thé:r,
the:rós). It is hard to tell whether the /e:/ is
"fundamental" here, or the result of Szemerényi lengthening
in the Nsg. of a root *g^hwer-. Greek and Balto-Slavic have
generalized the /e:/, Latin has /e/. The paradigm may
originally have been *g^hwé:rs, *g^hwérm., *g^hurés, pl.
*g^hwéres (*e, lengthened in the Nsg.), or *g^hwé:rs,
*g^hwé:rm., *g^hurés, pl. *g^hwé:res (original *e:).
=======================
Miguel Carrasquer Vidal
miguelc@...