--- In
cybalist@yahoogroups.com, mkapovic@... wrote:
> > --- In cybalist@yahoogroups.com, mkapovic@... wrote:
> >
> >> I think it's not at all questionable that IE was not a
> >> unified language. Every language has dialects (except maybe
those
> > that are
> >> spoken in only one village and similar cases) and so did IE. And
> > it's not
> >> at all strange that dialects have some different words, slightly
> > different
> >> morphology etc.
> >
> > But do we know that more than one dialect participated in the
> > expansion of IE into the daughter languages? Can we know that the
> > dialects of PIE did not just die out without affecting the
development
> > of a single norm that was the only one to be continued? If the IE
> > dialects just disappeared without leaving traces in the IE we
have
> > come to know they are of no interest to us. How does one decide
what
> > picture to imagine?
>
> That is ofcourse also possible. But I think that there are more
then
> plenty of differences which would point to a dialectal diversity
in PIE.
> Ofcourse, it is not impossible to think of some other scheme in
order to
> explain them, but it's not the simplest way in my opinion. But we
will
> probably never know without a time machine.
Our method is generally to work implicitly under the assumption that
there is only one PIE linguistic norm and that all IE daughter
languages are descended from it. And that method has been know to
produce majestic results. That does not exclude the existence of
cases where one is in doubt which form to posit for PIE. However, I
do not know of many cases where I feel I have to accept that more
than one possible reconstruction must be assigned for the
protolanguage. Mostly we just don't know, and the best guess is that
only one form is correct, if any. Even for a case like RV s'áye ->
AV s'éte and Luvian ziyari : Hitt. kittari I find it difficult to be
convinced that the parameter by which *k^éy-or and *k^éy-tor
differed was one of dialect. Would one PIE dialect be continued in
the Rigvedic dialect and Luvian, another in that of the Atharvaveda
and Hittite? Or was one form high-style archaic/archaizing and the
other one that of the common folk and their ill-behaving children,
or perhaps of women? I wouldn't know, but such variations would of
course live side by side and show interaction. A case form your
alley is the circumflex on Slavic contracted vowels which has
produced Proto-Slavic neoacute on the preceding syllable although
there are uncontracted forms in the daughter languages. I suppose
that can only be explained by assuming that contraction existed
before it became obligatory. I would of course like to know whether
to posit *wis- or *wi:s- 'poison' or both, and if both whether the
variation is one of territory or is due to some other parameter. If
it was a foreign word, it may not have had a norm at all. I'm afraid
we are mostly left with empty speculation, a frustrating point on
which we appear to agree.
Jens