Re: [tied] *g'(h)- > d as aberrant outcome

09-05-2004 19:02, alexandru_mg3 wrote:

> Yes I know, but this means also before ( or latest during) the
> first Slavic Loans ...because during the Slavic loans we have s->s
> both in Romanian or Albanian.
>
> So we are here even before Alb . s->sh have been started, that
> means somewhere before 600 AC.
> This means that the Romanian loan (if it is a loan) has to be
> borrowed before this date. (and it could be just before but more
> probable long before...for this reason the substratual theory fit
> better in my opinion).

The _oldest_ Slavic loans show s --> sh, which allows us to date the
change _later_ than AD 600.

> The second issue that you have is to explain next today Alb. G' for
> Romanian 'ghe' (knowing that we are somewhere before 600AC).
>
> A proto-form that we know for that period is:
> Proto *gl - Rom. gl' (later Rom. ghe) - Alb. gj
>
> But you didn't indicate a proto-form *gl...
>
> On the other hand, a proto-form : *g'e would give toRom. ge (g^e)
> (see Latin Words)
>
> So trying to locate in time an Albanian G' for a Romanian 'ghe'
> both of them not derived from a proto-form *gl seems difficult to be
> done in Roman Times...

But Albanian <gj> has several sources. It can derive from *s in a
stressed syllable, from *j (as in the this case), from *g palatalised
before a fron vowel, _or_ from *gl. In the word we're discussing it
never passed through the stage *gl.

Piotr