From: Arnaud Fournet
Message: 61583
Date: 2008-11-13
----- Original Message -----
From: G&P
>how do we establish the existence of H1 and H2
>if at the same time, we admit a purely vocalic a/a: and e/e: ablaut ?
.> All this being wrong, I'm afraid.
It's still an interesting question, to see how far we could establish any,
or two or three laryngeals without using vowel colouring. I agree with you
(of course) that vowel colouring is the main evidence for distinguishing
them. I'd go further, and ask if we could establish them without Greek.
But the question feels like the sort of thing one should know an answer to.
Peter
=============
Without the testimony of Greek,
I suppose there is no doubt we need at least one laryngeal.
I can think of Anatolian, Balto-slavic accentuation, LWs into non IE
languages, aspiration of some stops,
Maybe it's enough to prove we need at least two.
I think the "classical" theory has four pillars :
1. e and o are the only "real" vowels,
2. the root structure is at least CvC,
(This is most probably wrong, some words are e- with no H)
3. the color of the laryngeal is (most often) given by Greek,
4. non laryngeal long e: as in Nominative pHte:r is "syntactical"
(Cf Szeremenyi and I guess the better-versed of the list will provide
information)
Any time one starts tinkering with long vowels like a: without laryngeal,
I think it causes major havoc because the structure of the reconstructed
roots is not longer known.
If wag/wa:g is acceptable, then this root is just *w_g with a ablaut.
There is no reason to reconstruct it as *w_H2g with standard e ablaut.
But another issue is how reliable is Greek ?
I'm not sure Greek is as reliable as usually considered.
For example, it's fascinating that H(2)ente "in front of" cannot be compared
to Egyptien H._nt_
because H. (=H1 pharyngeal unvoiced) cannot be equated with H2.
You can choose from three options :
1. Not related words.
(The conservative viewpoint)
2. Something in Egyptian caused H2 to become H1
(But then what ??)
3. Something is wrong with H2 identification.
I once read Kurylowicz and my own reading is that
Proto-Greek once had alternations like
a a: from H2 eH2
a e: from H1 eH1
a o: from H3 eH3
That is too say a was H as in all the other IE languages.
But because of the long vowel,
the short a was changed to short e and o to match the long e: and o:
The result of that kind of reasoning is that
whenever proto-Greek did not have a long vowel on which to rely,
the short a was not "emended" and therefore /a/ stands for any H
We may also posit that there are three levels of stress :
- fully stressed Hxé = a e o
- fully unstressed H = a
- countertonic : Hxe = a
In that case, Egyptian H._nt can be equated with *H1ente
this word used as a preposition is in a countertonic position
the main stress being on the following noun.
There are plenty of other cases when gutturals in Chamito-Semitic "resist"
comparison with PIE laryngeals
because the color of H is not the right one.
This suggests that the "procedures" to ascribe the color of the laryngeals
are flawed somewhere.