--- In
cybalist@yahoogroups.com, "Rob" <magwich78@...> wrote:
> So you surmise that accent was retracted where there was
underlyingly-
> long root vocalism?
Certainly. In the inventory of received IE paradimg types, long root
vocalism only occurs in acrostatic paradigms.
That does not quite exhaust the possibilities however, for some
lexemes do appear with a mobile accent, as *pó:d-s, *ped-ós, and
*yé:kW-r, *yekW-n-ós. In general opinion they are mostly taken to
have introduced mobility by analogy at a later date. I agree with
this scenario, although it could of course be wrong; but it could
hardly be simpler.
Jens