*G^EN-

[petegray]

I'm sorry if I was too suspicious of your motives, Pat!

You said:

>Like Pokorny, I believe that an Old Indian form like _janati_ points to
>*g^en- rather than *g^enH-.

Not necessarily.   It could have come from a zero grade *g'nH-eti (syllabic
n > an in this context.)   The same pattern is found in the zero grade Vedic
perfect plural jajanus for the expected (and later) jajnus.

I admit the accent is a problem, but this pattern of zero grade present
+ -e/o suffix is not uncommon.  Rix takes several verbs back to this
pattern, eg Latin maneo, venio, Greek baino,  Lithuanian lupu, Gothic
ga-lu:kan, Latin/Greek luo et al.

You also said:

> everything I have investigated leads me to believe that the basis
>of the CVC definition of a root is pretty universal.

Yes, but the vowel is not always -e-.  It can also be -r-, -i-/-y- -u-/-w-
etc, - I'm  sure you know the theory in PIE, whatever you think of it.   So
a root g'enH can be seen (and should be seen!) as CeVC / CVC: *g'enH/g'nH.
The unusual roots are those that are CC.

As for a root *g'en without laryngeal, there are some Sanskrit forms that
show no laryngeal, but they stand alongside forms that need it, and they can
be explained from the *g'enH root, whereas the forms with H cannot be
explained from a root *g'en.

So we're left with no evidence for laryngeal-less *g'en.  That does not mean
it did not exist!   Only that we have neither evidence for it, nor need of
it.

Peter