Re: [tied] Alb. =?UNKNOWN?Q?kat=EBr?=

17-11-03 19:39, alex wrote:

> but it seems easy enough to link them to Rom; at least dy, tre, gjashtë,
> shtate, tetë, nëntë, dhjetë ( if for 10 indeed "k^" got lost as Abdullah
> assume for some stops in certain conditions) don't appears to me to make
> some troubles in linking them, don't you find?
> The most interesting here seems to me by now the "gjash(të)" versus Rom.
> "Sase" ( cf. gjarpër-Sarpe).

For each of the numerals you list there are good reasons why it should
be regarded as inherited and why Latin origin can be ruled out. For
example, <shtatë> and <tetë> preserve the PIE location of stress,
changed in Latin. We'd get quite different forms if we took Latin as the
point of departure, e.g. okto: --> *oft (cf, Rom. opt).

Piotr