Re: [tied] How many laryngeals?

On Thu, 23 Nov 2000 15:21:23 +0100, "Piotr Gasiorowski"
<gpiotr@...> wrote:

>>Is that referring to such forms as mehur, sehur?
>
> Yes. I'd analyse them differently. For example, the insistence that mehur must be derived from *meh1- 'measure' is oddly modern (we think of time in terms of minutes and hours). Why not *meih2-wr/n- from the root *meih2- 'pass, go by' (as in Latin meo: or Slavic mijati, mimo 'past, by')?

Eichner, claimimg that laryngeals do not colour a long *e:, derives
this from *me:h2wr, from the root *meh2- as in Latin ma:turus "ripe",
ma:ne "early" (IEW "gut, zu guter Zeit, rechtzeitig").

>>I find Rasmussen's analysis of Hitt. utne "land, country" <
>*h3ud(r/n)- ~ Grk. ou~das (*h3ud-) "ground, floor" ~ Arm. getin
>"ground" (*h3wed-) rather convincing. Maybe *h3u- was lost early in
>Hittite.
>
> But in *h2u-/*h2w- *h2 wasn't lost, which is a bit strange if the two sounds were similar.

But less strange if *h3 was a labialized phoneme (cf. the loss of *kw
before *u in Lat. ubi < *kwu-bhi).

> Rasmussen's etymology would still hold if you replaced *h3 with *h1 and assumed an o-grade initial syllable (*h1oud-) in Greek and possibly in Hittite. The Armenian form may be related (*h1wed-?) but certainly doesn't match the Greek one directly.

Rasmussen cites Benveniste "Origines de la formation des noms en
i.-e." for Greek substitution of -ar by -as. If the Arm. form points
to *Hwed-, the o-grade must be *Hwod- rather than *Howd-.

>>[On h1 = /h/: I have trouble seeing all forms with Brugmannian initial
>*e- as originally [he-] instead of at least sometimes [?e-] (with
>"automatic" glottal stop). However, they all seem to give *@- in the
>zero grade (e.g. Hitt. as-, ap-, ad- for the verbs "to be", "to take",
>"to eat"), so I can't resist splitting up *h1 into original /h/ and
>/?/ (or at least [?]).]
>
> Very likely, though hard to demonstrate.

Indeed. I vaguely remembered a *h1- > h- in Albanian, but that's just
a misprint in Beekes (p. 268 hidhur "bitter" < *h1idh-un-, should be
*h2idh-un-).

>>The origin of the Hitt. hi-conjugation is a whole other can of worms,
>but surely there are many hi-verbs with e-vocalism throughout, or with
>the most common hi-conjugation ablaut, i.e. -a- in the sg. and -e- in
>the plural. For the root *mel-, I gave my explanation here recently
>(*mwel- > *mel- or *mol-, zero grade *mul- as in Greek etc.).
>
> Of course, I couldn't cover too much detail in my posting. There is a whole series of interesting articles by Jasanoff, which more or less cover the topic of o-presents (with o/e or o/nil ablaut) and Hittite hi-conjugation. Your hypothesis about *mel- works for this particular root, but it's ad hoc character comes to light if one considers similarly behaving roots beginning with velars, like *k^onk- (Hittite ka:nki, Germanic *xanxiT), where you can't posit a labialised onset to explain the colour of the vowel.

Where can I find these articles by Jasanoff? I'm certainly not
claiming that every *-o- comes from *-e- preceeded by *-Cw- (that's
only a sporadic development, which works for *mel-, maybe for the
Luwian vb. endings -wan(i), -tan(i) [Hitt. -wen(i), -ten(i)] if they
were in origin *-mw-en(i), *-tw-en(i), and some other cases, such as,
quite consistently, *h3e). I was merely complaining about your ad-hoc
explanation of "my" root *m(w)el- as a "perfect" o-grade.


=======================
Miguel Carrasquer Vidal
mcv@...