Greek question

[George Hinge]

Even though I suppose that the Moderator is right in rejecting
*nog(w)h-t-, I am not sure that it can be ruled out completely on the
basis of Bartholomae's Law.

The law is restricted to Indo-Iranian and is a late procedure. It is
at least possible that the assimilation of dh-t bh-t gh-t has
happened twice (or over a longer period): the Proto-Indo-European
assimilation was limited and did not cross the morpheme-boundary,
while the post-Indo-Iranian was complete. Then, one has to argue,
that there is a closer affinity between -t- and the stem in *nogwh-t-
than -to- and the stem in, say, *bhudh-to-.

Or, one can explain the problem away generatively, arguing that
Bartholomae's Law was a synchronic rule in Proto-Indo-Iranian, an
expression of the relationship between cognate stems with a
particular segment at the root-boundary (*baudh- ~ *buddha-). The
only derivation of the stem *negwh- in Indo-Iranian was *nogwh-t-,
and there was no "need", then, for marking original root-structure,
ergo > *nakt-.

In spite of all these badly phrased and idle considerations, I'm
still sceptical towards *nogwh-t-. Maybe Greek nykh- is a false
interpretation, not of the nominativ nyks (cf. Chantraine), but of
nykt|, on the analogy of lektron ~ lekhomai, thaptos ~ taphos etc.
The forms ennykh(i)os and pannykh(i)os are poetic and may be
artistic creations of the epic tradition.

George Hinge
oldgh at hum.aau.dk