--- In
cybalist@yahoogroups.com, "Grzegorz Jagodzinski"
<grzegorj2000@...> wrote:
> Now my question. Are similar spontaneous dispalatalizations common
among the
> IE family? Among the languages of the world? And a related question
- if
> such spontaneous dispalatalizations are documented, even if rare,
why should
> we reject the hypothesis of the existence of palatal (or at least
prevelar)
> series as long ago as in the common IE period?
There are dispalatalisations and dispalatalisations. In a sound like
English <sh> or Proto-Slavic *s^ palatality is only a secondary
feature colouring an essentially postalveolar articulation. The body
of the tongue is "domed", i.e. bunched up towards the hard palate, but
the main constriction is apical or laminal. The acoustic difference
between such a fricative and its "flat" counterpart (= Polish <sz>) is
hardly drastic, and the lowering of the body of the tongue may occur
without causing the phoneme in question to lose its identity (in other
words, doesn't give rise to any comprehension problems).
To "dispalatalise" a true palatal, however, would mean to take its
primary place of articulation away (like "delabialising" /b/, for
example). One wonders what would remain after such an operation -- the
glottal component? A change from a palatal stop to a velar one is
therefore an example of retraction, not of dispalatalisation, just as
the reverse change is fronting, not "develarisation". The natural
evolutionary dynamics in the case of true palatal stops makes them
prone to further fronting rather than retraction, since, because of
their very long contact area and their tendency to develop audible
palatal fricative release they are more readily confusible by
listeners with prepalatal or postalveolar affricates than with velars.
> Btw., it is often thought that the process of k^, g^, g^h > k, g, gh
looks
> inbelievable as only fronting development of palatal is attested.
But if I
> am not mistaken, a similar process "prevelar > velar" is known from
some
> Polish dialects, including the urban dialect of Warsaw. And, if it is
> possible in Polish, why not in PIE?
>
> Of course, the difference between Polish velars and prevelars
(incorrectly
> termed "palatals" in some sources) is less than between IPA [k] and
[c]. But
> why should we believe that PIE k^, g^, g^h were palatals? Couldn't
they be
> just prevelars instead? If yes, their retreating and mixing with the
> original velars in Centum would be easier to believe in. And the Polish
> dialectal processes (does anybody know examples from other
languages?) would
> be the needed attestation.
>
Of course they _could_ be prevelars, and since prevelars are modified
velars (articulated near the front of the velum and accompanied by
palatalised release), the modification may be cancelled as easily as
the palatal accompaniment of coronal obstruents. You are absolutely
right that a change from such a segment to a plain velar is much
easier to believe in than the spontaneous retraction of a true palatal
(IPA [c]).
Piotr