--- In
cybalist@yahoogroups.com, Piotr Gasiorowski
<piotr.gasiorowski@...> wrote:
> 26-01-04 10:04, Piotr Gasiorowski wrote:
>
> > I don't think short /a/ in Latin loans was nasalised. As for
> > rhotacism, words like Tosk fre/freri (Geg frê/frêni) < fre:num
differ
> > from <qen> in having no synchronic alternation between /n/ and /r/
> > within Tosk.
>
> Sorry, I didn' finish my thought. If early Tosk originally had
<qen> vs.
> expected *<qeri>, analogical levelling was possible because of the
> retention of the final n in the base form.
>
> Piotr
************
Retention of the final /n/ we have also in: Lat. fraxinus > (g.)
frashën, but Tosk <frashër> 'ash(-tree)', Lat. arena > (g.) rânë, but
Tosk <rëra> 'sand', Lat. ordo, -inis > (g.) urdhë, (t.)
urdhër 'order', Lat. orphanus > (g.) i vorfun, (t.) i varfër 'orphan'.
But, if we consider Illyrian (attested forms for <dog> are <can> and
<cand>) as intemediate stage of Albanian, I think that we will have
no difficulty to explain nor Albanian <qen>, niether Alb.
<qerr>/<qarr> 'cerris, oak'< *kWerkW-(attested Illyrian forms
<Kerkyra>/<Korkyra), the thesis that you object constantly.
Konushevci