Jens Elmegård Rasmussen wrote:
> Piotr Gasiorowski <gpiotr@...> wrote:
>
>> The vocalism *a can be due to a fundamental *a (at least I think
>> so), but such an *a does not alternate apophonically with *o.
>
>
> I have seen and heard that statement many times, but what is it based
> on?
Absence of evidence, which admittedly is not per se proof of absence,
but as far as I can see the only thing that certainly _can_ alternate
with fundamental *a (i.e. those instances of *a which do not seem to
result from any kind of secondary colouring) is *a:, as in *wa(:)stu-,
or *Hna(:)s-. I'm not quite certain about *a : zero, but I haven't seen
any convincing examples thereof (except, possibly, Gmc. *sultjo:, if
from *sal(d)-, but this is a bit too isolated to be decisive). As for
cases like *ko:pah2 and (perhaps) *kono- from *kap- and *kan-, I believe
the *a in these verb roots comes from *e coloured by the adjacent *k,
cf. the long-vowel grade *ke:p-, not !*ka:p-.
Piotr