Miguel Carrasquer Vidal wrote:
> I think Jens would disagree there. In his theory, lengthened */e:/ would
> have given*/o:/, and the Osthoff shortening reduces that to */o/.
I can't speak for Jens, but acccording to my understanding of his theory
lengthened *e: gives *o: only when further lengthened into a trimoraic
vowel. This wouldn't have happened in my hypothetical *pe:nth2-z because
of the early appearance of a svarabhakti vowel breaking the obstruent
cluster:
*pe:nt&h2-z
The vowel abducts the extra mora, which is why we get *pé:nto:h2s >
*pénto:h2s (Osthoffian) rather than *pé::nto:h2s > *pó:nto:h2s >
*pónto:h2s. If the shortening is sufficiently old, the hysterodynamic
character of the word may be original rather than secondary.
By the way, probable further cognates in Germanic include *funsa-
'hastening, willing' < *funssa- < *pn.t(h2)-to-. The existence of such
obscured derivatives points to a long and complicated history of the
root in pre-Germanic, and in my opinion increases the likelihood that
*finþ-i/a- is a retention, not an innovation. Another instance where
Germanic preserves a verb lost elsewhere is *singW-i/a- < *sengWH-
'sing' (cf. Gk. ompHe: < *songWH-ah2).
Piotr