Re: [tied] Greek substratum words

On 2006-10-18 00:28, Joao S. Lopes wrote:

> The words below, all from Greek, have similar vocalism and fall into the
> same declension:
>
> Phorkys (g. phorkynos)
> Telkhis (g. telkhinos)
> delphis (g. delphinos)
> rhis (g. rhinos)
>
> e-i and o-u seem to point to
> Phorkys (g. phorkynos) < *pHarkun-
> Telkhis (g. telkhinos) < *t(H)alkHin-
> delphis (g. delphinos) < *dalpHin-

Why not *gWelpHi:n-?

> rhis < *Hrin-

*sri:n-?

In the -í:s/-î:nos type (also aktí:s 'ray, light', glo:kHí:s/glo:kHí:n
'projecting point', o:di:s 'pang(s) of childbirth') we have long /i:/.
These words _look_ like compounds with an i-stem as the first member,
followed by something like *-h1(e)n- (assuming the analogical origin of
the Gk. nom.sg.). They may correspond to free-occurring u-stems, cf.
*gWelbHu- 'womb' and RV aktú- 'ray, light; night' vs. *gWelbHi-h1n-, etc.

Piotr