--- In
cybalist@yahoogroups.com, "Daniel J. Milton" <dmilt1896@...>
wrote:
> Rejecting laryngeals is maybe too heretical to be considered
> seriously, but this [Wyatt's book] seems to be something you might
> want to read.
I've read parts of it. The basics of his theory are that a and schwa
are found in complementary distribution. Scwha never occurs word
initially, for example, but a is frequently found there. Internally the
only languages that show any difference between a and schwa are Greek
and Indo-Aryan. I don't know if he ever addresses the triple
representation of scwha in Greek. However for Indo-Aryan he claims that
a (< *a) and i (< *@) are in complementary distribution as well. Thus
he assumes [a] and [@] are just allophones of /a/. I think a pretty
solid case can be made for PIE /a/, but there's got to be at least one
laryngeal (though I'm skeptical of there being more than one).