> Jens wrote:
>I have NOT proposed a regular sound law "s > j" for Albanian. I have
>however, proposed an event of lenition of the expected product of
*s->,
>viz. gj-, to Alb. j-. A good example is jo-në 'our' (fem.) from
*sa: >+
>enclit. *nos. Since IE *y- also gives Alb. gj-, I would assume that
>ju 'you' (pl.) is also a reduced form. And the verb to be certainly
>also qualifies, so janë 'they are' could be a reduced form of
>(expected) *gjanë from *senti or *sonti.
There are at least two words in Albanian : a-jo 'she this one' and
ju 'you' that present the treatement : PIE *y- > Alb. j- and in the
same time (could) have clear cognates Skt. ayam and Lith acc. jus.
On the other hand we have 2 Albanians rules that don't fit with
a supposed PIE *y > j
1. yV- > gjV-
2. sV-/stressed > gjV -
I think that yV- > gjV- wasn't applied in all contexts in order to
make room for PIE *y > Alb j too:
So PIE *yV- > Alb. jV- when:
V=a:
Alb. a-jo Rom. a-ia < PAlb a ja: < PIE ya:
V=u
Alb. ju < PIE *yus
Note:
This could have more sense if the intermediary stage of a: was *wa
(as I suppose that it was the case)
PAlb a: > wa > wë > Alb. o
(showing a treatement of *wa similar with the Romanian treatement :
una > ua > uã > uo > o, but that finished in Albanian very earlier)
In this case the rules would be become ONLY:
PIE *y/w,u > j (but not u: etc...) that has sense.
(This give us also an indication regarding different timeframes).
In all other cases :
PIE *yV- > Alb. gjV-
If true, this could help us to decide better for other contexts:
when the original PIE was from sV- or it was yV- for an Alb gj-?
I also think that:
jo-në < PIE *(a) ja: nos
y-në < PIE *(a) (e)i nos
and so-në < PIE *k'ja:i nos (see sonte < *k'ja:j + nokWtai )
so jo-në is not from *sa:- but from an original *y -> *a ya: nos
Best Regards,
Marius Alexandru