Re: [tied] Re: Latin is a q-Dialect having p- from kW , PIE is s

On 2007-05-16 11:02, tgpedersen wrote:

> This could instead be explained chronologically as sets of allomorphs:
>
> *penkWe ->
> *penkWe/*pempe (kW -> p allomorph) ->
> *kWenKwe/*pempe (p- -> kW by shibboleth reaction) ->
> *kWonkWe/*pempe (kWe- -> kWo) ->
> *kWenkWe/*kWonkWe/*pempe/*pompe by analogy (loss of shibboleth action?)

But P-Celtic *pempe can ONLY come from PCelt. *kWenkWe, as PCelt. had no
*p. It _can't_ continue pre-Celt. *pempe. If such a "dialectal PIE"
variant existed at all, it has left no traces in Celtic! You propose a
zigzag derivation (in effect, *PIE penkWe > *pempe > PCelt. *kWenkWe >
*pempe) to bypass the PCelt. anti-p filter, but that's less parsimonious
than the straightforward change of *penkWe > *PCelt. *kWenkWe > GBrit.
*pempe.

Piotr