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

On 2007-05-16 17:20, tgpedersen wrote:

> Ahem. You have made your derivation more parsimonious by dodging your
> own argument from pompe < *kWonkWe < *kWenkWe < *penkWe, which is
> therefore also a three-step process.

But if *kWénkWe is accepted for common Italo-Celtic (whether a true
synapomorphy or an areal innovation), we get a neat scheme for both
branches, with *pompe and <qui:nqe> derived from the same PIt. form. The
overall picture _is_ parsimonious.

> But you're right about the (traditional) lack of *p in Proto-Celtic.
> So I'll propse a scenario like this:
>
> I Some IE people lives next to a people who can't say kW but
> substitutes p for it.
> II That people takes a piece of the p-people's land and get new slaves
> and girlfriends etc., who learn the new language, but say p.
> III Status: The High people are so disgusted with all these p's that
> they drop all p's. They also *kWe- > *kWo-. Now everyone in the
> conquered territory speaks q-Celtic.
> IV That new people takes the rest of the p-people's land and etc etc.
> V Status: The Low people in the newly conquered land say p for kW, but
> the High people are getting so used to it they say p too. Now everyone
> in the new conquered territory speaks p-Celtic.

OK, but this means abandoning the idea of *pémpe as a late IE (but
Pre-Italic and Pre-Celtic) thing. We're back to independent
branch-specific scenarios, whatever they are.

Incidentally, I have checked <cóic> in Thurneyssen's OIr. grammar and in
_IE Numerals_ (David Green in the latter just refers the reader to
Thurneyssen's exhaustive discussion of the problem). In a nutshell: the
word is problematic in that it reflects a long *o: that normally can't
go back to PCelt. *on before *k(W) (an additional complication is the
discrepancy between <cóic> and its compositional allomorph with a true
diphthong rather than a long vowel, but that part is not directly
relevant, so I'll skip it). While no solution seems to satisfy everyone,
Thurneyssen's optimal guess is that PCelt. *kWénkWe first developed
(regularly) into *kWe:gW- and only then did the long *e: undergo
rounding in the labiovelar environment, eventually yielding <cóic>
rather than *céic. So again we have a unique sequence of changes not
parallelled by the rounding in Osco-Umbrian.

Piotr