Three-Way Contrast of Secondary Articulations in PIE

[David L. White]

[ Moderator's note:
  The following may be a duplication for some readers.  I apologize for the
  inconvenience.  A note by Pat Ryan was apparently sent twice in place of
  this posting.
  --rma

> A way to reformulate what I've said about a three-way contrast in
> pre-PIE consonants that does not involve secondary articulations,
> multiplication of the consonant inventory, or typological objections,
> would be to focus on what became of the vowels.  So, instead of
> reconstructing, say, three nasals **n (>*n, *-r), **n^ (>*n ~ *i) and
> **nw (> *n ~ *m ~ *u), one would simply reconstruct **na, **ni and
> **nu.  When the vowels disappeared (when unstressed), or merged (when
> stressed), there may well have been a brief and unstable period where
> the three-way contrast was transferred, in the shape of secondary
> articulations, to the consonants, leading to a separate labio-velar
> series in PIE, and to certain transformations of or variations in the
> other consonants (e.g. *n ~ *i, *n ~ *m, *l ~ *i, *t ~ s, *t ~ *i, *p
> ~ *kw, *m ~ *u).

        Well, getting that off the ground would take a lot of work.
        But it should be noted that labialized velars are dramatically more
common than labialized non-velars, in languages of the world generally.  All
this seems to be a lot of trouble to go to just to 1) wind up with a
garden-variety labio-velar series, and 2) avoid having to posit some dialect
mixture in Balto-Slavic.  Labio-velars and dialect mixture are fairly
well-known things after all.
        But granted that the existence of three series of velars in PIE has
supposedly recently been proven on the evidence of some things (which I
forget) in (non-Hittite) Anatolian, are the facts of Balto-Slavic such a
problem after all?  Is it true that even with three series (palatal, velar,
and labio-velar) we must still posit a troubling degree of dialect mixture?

Dr. David L. White