Re: Re[4]: [tied] Re: *a/*a: ablaut

On Mon, 18 Feb 2008 22:13:54 +0100, "fournet.arnaud"
<fournet.arnaud@...> wrote:

>No. The lengthening of /e/ in stressed position affects only
>Hittite. The other Anatolian languages (Luwian, Lydian,
>Lycian, Palaic) have different reflexes of /e/ and /e:/,
>e.g. Lydian bira "house" < *pé:r (cannot be from *pér).
>
>===========
>Can is the reflex of *pér in that language ?

*e gives <e> or <a>, depending on the accent (see Craig H.
Melchert, "Anatolian Historical Phonology".

You'll be interested to know that Melchert reconstructs
three different kinds of /e:/ Proto-Anatolian:

PIE PA Hitt./Pal. Luw./Lyc./Lyd.
*ei *e: e: ~ i:[*] i:
*e: *E: e: i:
*eh1 *æ: e: a:

[*] after velars

>Arnaud
>=============
>As the lengthening is clearly secondary, you may of course
>ignore it for macro-comparative purposes. Just don't pretend
>it doesn't exist.
>Miguel
>============
>I don't appreciate the way you worded this.
>I never told or hinted it does not exist.

Yes, you said it was an innovation in Greek (or
"Central-IE"). In fact, it exists in all of Indo-European
(Hitt. has^te:r "star", NApl. wido:r "waters", Latin fu:r
"thief", Germanic *fo:t- "foot", etc.).

>You are acknowledging it is secondary.
>The main problem is
>I don't see why your long /e:/ should
>not be secondary too.
>Arnaud
>=================

It's easy to see that a lengthening in the Nom.sg. of
animate nouns (caused by the nom. ending *-s) does not apply
to neuters, wich have a zero ending in the NAsg.

It has been suggested that the length in *pé:r could be due
to compensatory lengthening of an original *pérr, but that
argument applies only with some effort to *k^é:r < *k^érd
(against it we have Arm. sirt < *k^é:rd), and not at all to
*yé:kWr.t.


=======================
Miguel Carrasquer Vidal
miguelc@...