The palatal sham :)

From: Richard Wordingham
Message: 30947
Date: 2004-02-11

--- In, "P&G" <petegray@...> wrote:
> >If all Latin words with non-ablauting /a/ are borrowed, then all
> >Latin words with <ca-> are borrowed. But the occurrence of /a/
> >supposed plain velars in Latin was the main argument for the
> >existence of those plain velars in PIE! So:
> >Plain velars don't exist!
> (b) The main argument for the existence of plain velars in PIE is
> correspondence:
> Centum /k/ ~ Satem /k/
> It has nothing to do with the vowel that follows.

That argument on its own, of course, relies on Satem *not* being a
tight-knit group. If it were, one could suspect dialect admixture,
as in the case of British s- ~ st- > Welsh h- ~ s-, but with several
words showing s- > s-, such as _saith_ 'seven'. (There was some
discussion of this group of anomalies at et seq.)

(Albanian digression - or reversion - resisted by request :)

> (a) A number of ca- words in Latin show an impeccable PIE pedigree,
eg capio
> ~ Germanic have.

And if one looks in Pokorny, one will find Sanskrit, Old Persian,
Albanian and Greek cognates, together implying *k, not *k^.

(Romanian digression resisted by request :)

> canis ~ kuo:n in Greek.

Hardly a regular correspondence! I didn't think it was even
explained, but rather just assumed to be a loanword if related to the
other dog words. The other possibility is that it derives from
*ken 'young, fresh' (Pokorny #907), and so originally meant 'puppy'.
I don't know where the /a/ of Latin _canis_ and Greek _kainos_ 'new,
recent' comes from, though. Or should we be asking about the /e/?