Three-Way Contrast of Secondary Articulations in PIE

[David L. White]

>> Fair enough, but what we need is a THREE-way contrast of palatalization,
>> labio-velarization, and whatever "[a]-quality" would be called, probably
>> "uvularization", a problem being that no such thing as uvularization (as far
>> as I know) occurs.

> This sort of thing occurs in Abaza/Abkhaz, Adyghe/Kabardian, and Ubykh.  One
> reasonably accessible account is Aert Kuipers' monograph on Kabardian in the
> _Janua Linguarum Series Minor_, in which he first analyses three series in
> the obstruent system (palatalized, labialized, and plain) which cause
> rounding and fronting in the single vowel.

> He then proceeds to re-analyze the plain series as having a feature "open",
> which he symbolizes with a superscript <a> to match the superscript <j> and
> <w> of the palatalized and labialized series respectively, and postulates
> that with the addition of this feature, we need not have any phonemic vowel
> at all in Kabardian.

> This analysis flies in the face of everything we know about permissible
> phonological systems and naturalness.

> Later studies of the language have found that the proper analysis requires
> that there be two vowels, /a/ and /I-/ ("barred i").

        Now I am thoroughly confused.  The analysis with three contrasting
consonantal qualities is seemingly presented as a fact at first only to be
dismissed as wrong later.  If "[a]-quality", which leads to the conclusion
that there is no /a/, is an illusion, then how are there three qualities?

Dr. David L. White