On Sat, 27 Jan 2001 12:01:38 CET, "Piotr Gasiorowski"
<
gpiotr@...> wrote:
>I wonder, however, if there are any
>known systems with /ts/ and without /s/. I would
>think (though I may be wrong) that affricates
>imply at least one coronal fricative.
To my recollection, I have never encountered such a system outside
South Dravidian (where [s] is merely an allophone of /c/, e.g. in
Tamil).
There is the case of Proto-Semitic, where the sounds traditionally
reconstructed as */s/, */z/ and */s./ must in fact to be reinterpreted
as */c/ */dz/ and */c./ (thus also explaining the relative rarity of
PSem */s/ = */c/). Similarly, the interdental and lateral fricatives
*/t_/, */d_/ and */t_./ (= Arabic /z./), and */s'/ and */d./ seem to
go back to affricates */c^/, */dz^/, */c^./, */tL/ and */tL./.
However, in this reinterpretation, PSem trad. */s^/ (= Arabic /s/)
turns out to be plain */s/ (while we have PSem */s^/ for the sound
that gives Akkadian /s^/ and /h/ elsewhere [<s^uwa> ~ <huwa> "he"]).
Plus cases where trad. */s'/ seems to go back to fricative */L/ rather
than affricate */tL/.
However, from the point of view of historical sound change, a system
with coronal affricates but no fricative(s) would not seem to be that
strange. Usually, the affricates historically derive from dental or
velar stops, mostly through some kind of palatalization (*k > *k^ > *c
or *t > *t^> *c^, etc.). The affricates in turn develop into
fricatives (occasionally into stops) (*c > *s or *c^ > *T (> *s ~ t),
etc.). A language without fricatives could in theory develop
affricates *before* getting fricatives. The fact that I can't think
of many language with, say, /c/ but no /s/, may be rather due to the
fact that there are so few languages that lack /s/.
=======================
Miguel Carrasquer Vidal
mcv@...