--- In
cybalist@yahoogroups.com, "pielewe" <wrvermeer@...> wrote:
> ---
>
> How is it that those who hold that intervocalic *s yielded both
_sh_
> and _h_ in Albanian see the relationship between those reflexes?
>
>
> W.
There are mainly 2 Positions, Willem, regarding this issue:
1. "There wasn't any intervocalic s > h" (Hamp)
==============================================
In this case because s gave only sh in Alb., we need to suppose a
consonantic cluster with s and not a simple s, for all today's
intervocalic h.
The next step is to try to identify these clusters.
Knowing that we have the known PAlb. Rule: sk,sk' > ks > h, the
first re-constructed cluster that could be easy identified: is sk'
From here Piotr's derivation...
(I will ignore for instance an original k's that gave something
else).
Note: The issue here is that the Albanian words and the related
cognates easy support correct derivations without any k^ (see
Derksen's and Demiraj's derivations)
2. "There was an intervocalic s > h"
=====================================
2.a "intervocalic s>h => but only in special contexts"
======================================================
...where this 'special context' is defined by each scholar...more
or less similar.
Example:
's > h before an u' (Pedersen)
(my opinion: Seems in contrary that s > h not before u
see vesh 'ear' < PAlb *wausis or *wo:usi and here we have solid
cognates)
The issue here is to well identify this special context.
2.b "intervocalic s>h => in all contexts"
=========================================
In this case we need to accept that an original intervocal PIE *s
has already 2 ouputs in PAlb because finally we have h and S /sh/ on
today Albanian
Next to explain the 2 different outputs : will logically result that
we need 2 outputs for the intervocalic PIE *s also in PAlb times.
These ouputs are easy to be identified: S /sh/ and s (>h) already
in ProtoAlbanian.
Next we need to explain the source of this PAlb S /sh/ we can
proposed different origins: after u, after u,i etc...in all these
situations we arrive to have at least a partial Ruki Rule in PAlb.
Because Albanian belongs to the Satem Group this ancient S /sh/ will
not be a suprised
The issue here is to can demonstrate the contexts of this PAlb S
(difficult because any s passed to sh later)
==================================================================
II. For me this situation is very similar to a second one: PIE *sk
==================================================================
For the PIE *sk we have also 2 different outputs in Albanian : s and
shk (and I talked here only about inherited words)
Here the theories go in a similar direction:
===========================================================
II.1. "There wasn't any sk > sk (later shk) so all sk > h"
==========================================================
I saw that this theory is sustained on this forum by Piotr.
In this case the other output: sk(>shk) should be explains by
different contextual solution:
a. in initial position shk is explain based on a 'mobile-s'
That s- wasn't there when sk>h took place, and was added only later
b. in internal positions shk is explained by supposing a more
complex consonantic cluster like : mainly k(^)sk etc...that sometimes
is really the case without any doubt.
=================================================================
II. 2. "There was also sk > sk (>later shk) together with sk > h"
=================================================================
2.a "sk>sk => but only in special contexts"
======================================================
See Beekes derivations -> already discussed on this forum
2.b all the sk>h because others PIE *sk became already shk when
sk>ks>h took place
=====================================================================
To be honest this is mainly my suposition trying to link it
with the first issue and to globally solved both situations (see
'My hypothesis'.
Best Regards,
marius
P.S.
My hypothesis:
==============
If we accept that in PAlb we already have had a s and a S /sh/ all
the issues above will be solved using a single solution:
The already existing cluster shk or the simple intervocalic sh
remained unaffected by the new transformations:
1. sk > ks > h
2. intervocalic s > h (> zero)
(In this situation the additional transformation *si > shi was also
older that s>h, and next remained active for a very long time)
So when the Roman arrives: the inherited PIE *s remained only in some
consonatic clusters, like PIE *st,*skr - but in all the other
contexts we have S /sh/ and h in PAlb.
Finally all the remained inherited s passed to sh (sec V-VII)
Best Regards again,
Marius