>
> 1. the vocalisation of a kind of /x/ in *ph2te'r CANNOT BE ANYTHING
>ELSE BUT /Vx/ (by adding a vowel before it /p&x-t'er/)
You are wrong. Like any other syllabic continuant, it can be
pronounced with a prop-vowel before, a prop-vowel after, or
no prop-vowel at all.
=============
Some languages exhibit very heavy consonant clusters,
apart from Salish, you can list Kartvelian
and Moksha in Uralic.
Languages which accept this feature have
complex syllable formats.
As PIE clearly accepted initial cluster
like *spl or *skr-
we can conclude synchronic PIE accepted
things like skHtlk-e-H as a phonable word,
with only one apparent true vowel.
In Moksha, long clusters usually include s s^ f
in the middle.
ARnaud
===============
The distinction between vocalized and non-vocalized
layngeals is crucial in other soundlaws as well, like for
instance Hirt's law in Balto-Slavic, where a consonantal
laryngeal causes retraction of the stress, but a vocalized
laryngeal does not.
================
This distinction means ° vocalic schwa
is a phoneme, constrasting with *e and *o.
Arnaud
=================
Or, for that matter, the French ("retard" = [Xta:X]).
Miguel Carrasquer Vidal
=======================
this is the formal situation
I think colloquial is [°Rta:R]
R is more a fricative than a
hard unvoiced X.
In my own fast speech,
I have afternoon : après-midi
which is [a-pRmi-di]
in three syllables
and short "cet aprm"
[stapRm]
It's unclear what the vowel is
in the pRm syllable.
ARnaud
=================