*rh1 > ra in Greek
> what about other zero grade forms? For *wreh1g^, we have rhagás 'tear,
>chink'. Piotr offered the schwebeabluat pair *terh1b, *treh1b 'building,
>dwelling'. If the Greek forms trápe:x 'post' and téramna, téremna
>(pl.) 'chambers' are genetically related as Pokorny suggests, they
>show *rh1 > ra just as in the zero grade of *wreh1g^-.
I don't want to argue too hard against rh1 > ra in Greek, because there has
been a reasonably powerful argument that all the vowel reflexes of
laryngeals in Greek are conditioned by analogy, vowel harmony/dissimilation,
or other factors, and don't reflect the actual nature of the laryngeal at
all. Personally, I'm not yet convinced of this. The usual reflexes (e, a
o) seem too well established. But we should expect a high number of cases
where something else happens. Some of these are on awkward roots (like
onoma < h1nom-n).
However, despite that, there is an argument in this particular case. It is
likely (I think) that rhagas gets its form from rhage:nai (L&S support me
here, which is nice of them, but they're hardly the most reliable source).
The alternation between teramnon and teremnon/teremnos is much more
interesting. -ere- is a regular reflex of rh1; we must then ask
whether -era- is an alternative phonetic development, or an alteration
from -ere- for some reason. Because of the existence of the usual reflex,
I go for the alteration, but I'm only guessing.
Peter