Three-Way Contrast of Secondary Articulations in PIE

[David L. White]

> Campbell's "Compendium of the World's Languages" (a far from perfect
> book, but it's what I have here at hand), gives palatalized *and*
> labialized consonants for the very first language decribed in it:
> Abkhazian.  Labialization together with palatalization occurs in
> North-West-Caucasian in general, together with a very poor vowel
> inventory (/@/ vs. /a/).  Here too, *i and *u yielded *y@ and *w@,
> while *a > *@ (and presumably *i: > *ya, *u: > *wa, *a: > *a).
> Something similar is assumed for Proto-(North-)Afro-Asiatic.

        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.  Velarization alone produces
something that would more accurately be called "[o]-quality" (perceptually),
and not surprisingly is more than a little difficult to keep distinct from
"[u]-quality", given the very brief window allowed before any vocalic sounds
would begin to be perceived as moraic "true" vowels, or as semi-vowels.
        Given the predictability (if my understanding of my source is
correct) of [u]-quality spellings in Old Irish, their value for establishing
contrast is limited, predictability and contrast being two things that do
not go well together.  But though suspicious, I am (I hope) open-minded:  is
there any unequivocally good evidence for a three way contrast of secondary
articulations in Old Irish, or Tocharian?  Predictable spellings don't
(necessarily) cut it.

Dr. David L. White