--- In
cybalist@yahoogroups.com, "etherman23" <etherman23@...> wrote:
>
> Pretty interesting that *h2 doesn't occur before labiovelars.
So I did a little investigating about this. I don't have LIV but I do
have Starostin's Tower of Babel resource. Aside from a couple examples
of ambiguous roots (because they occur only in Satem languages and
even then not well represented) there's only 2 counterexamples. The
first one is a root represented only in Greek and Armenian, and the
second is a root only represented in Greek and Old English (the Old
English word looks to me like it should go back to a plain velar). I
think we can say we a very high degree of confidence that *h2 did not
occur before labiovelars.
Just for fun I decided to check on examples of *h3 before labiovelars.
The results were quite similar. There's one counterexample which is
confined to Greek and Baltic.
Despite the general rarity of plain velars it's not difficult to find
examples of plain velars following *h2 and *h3. Seems kind of obvious
to me that labiovelars delabialized after *h2 and *h3. I'm not sure
yet if this holds for *h1 too.