Typology and the phonetics of laryngeals

[Alexander S. Nikolaev]

Stanley Friesen wrote:

>So, my current "best guess" for the laryngeals is something like yours,
>/x', x, x^w/, or /h, x, x^w/.  (With /x/ being a *back* fricative, and /x'/
>or /h/ being less far back).

As the last but not least argument for fricative nature of PIE
laryngeals it is commonly claimed that, being /x/s, laryngeals would
pattern the three rows of tectals (if these are in fact three :-> );
and that is why a quality of a palatized counterpart of H2 is
assigned to H1.
I take the liberty to expound some views on this tangled topic.

On the one hand, if the system is /x', x, x^w/ this would account for
the fact that H2 is the most widespread -- it is thus the least
marked, as to the set of its distinctive features. But in the course
of the development of the laryngeal theory H1 came to be held for the
least marked of the three, as it causes no colouring at all. And then
the system /h, x, x^w/ seems better: then the phonemic status of H1
as the least marked is not violated (the putative parallelism to
tectals could be omitted, being itself an obscure matter).

But -- the decision, that H1 doesn't colour an adjacent vowel largely
(if not wholly) depends on the H2O problem. And if one believes that
/o/ resisted the colouring by adjacent H2, it should then resist
being colouring by H1, too. (And discussing H3o is needless).
Then the conclusion is that /o/ is not coloured by any of the laryngeals.
And as the cardinal /a/ is extremely rare
and doesn't influence this logics much, since in the examples, where
it is present, it doesn't neighbour any of the Hs.
And then it's impossible to say, that H1 doesn't produce ANY
colouring-effect: if within the system of PIE this sound could
theoretically affect 2 vowels, and we know, that one of these
resisted being affected, and we do not know for sure, if what we write as /e/
was indeed [e] -- claiming that H1 makes no colouring is impossible,
isn't it? And then there's no need to account for its "weakness" by
assigning to this phoneme the "weakest" phonetic realisation [h].

Then /x', x, x^w/ is, finally, the one set, i prefer.

As to my position, I personally do believe, that H2o resulted in o; i
guess, there're some 12 examples, like orkh'eomai - arkh'os, 'onkos -
ank'ule:, 'ago: - 'ogmos, 'akros - okr'is... And i do believe that
the traditional PIE /e/ could have been something like [a[, /o/ being
[@], a kind of a Pulleblankianism.

	Any comments?

		Best wishes,
			Alex Nikolaev