Re: How many velar series in P.I.E? (was Re: [tied] *kW- "?")

----- Original Message -----
From: "elmeras2000" <jer@...>
To: <cybalist@yahoogroups.com>
Sent: Sunday, September 18, 2005 6:38 AM
Subject: How many velar series in P.I.E? (was Re: [tied] *kW- "?")


> --- In cybalist@yahoogroups.com, "david_russell_watson" <liberty@...>
> wrote:
>
> > Ever since I read Lehmann's argument I've been puzzled why
> > more Indo-Europeanists (at least those on this list) don't
> > seem to accept it, and also why the Nostraticists (at least
> > those on Nostratic-L) are trying to take off from a P.I.E.
> > with three velar series when its earliest stages would have
> > had no more than two.
> >
> > What's wrong with Lehmann's view?
>
> First, that he does not give valid reasons. Second, that the facts are
> against his choice: Take Pokorny or LIV and you will find quite many
> roots beginning with *ke-, i.e. roots attested in satem languages and
> showing velar, not palatal before the reflex of *e.
>
> Jens

***
Patrick:

That Pokorny or LIV have PIE roots beginning with *ke- that do _not_ show
palatalized reflexes of dorsals in satem languages is quite immaterial.

The palatalization of dorsals before -*e happened in pre-PIE, before the
institution of the Ablaut vowel (*A = *e/*o/*Ø); said another way: pre-PIE
*e palatalizes, PIE Ablaut *e does not palatalize dorsals.

Even earlier, *k^ had been merely an allophone of *k before *e but by
pre-PIE, *k^ achieved phonemic status.

In centum languages, *k^ merged into *k.

Therefore, one may legitimately reconstruct PIE roots beginning with
*k^a(:)- and *k^o(:)- and *ke(:)-.

Of course, *k here is shorthand for all dorsals: *g, *gh, *k, *k(h),
_except_ velarized dorsals (*gW, *ghW, *kW, *k(h)W).

***