From: "Lars Henrik Mathiesen" <[log in to unmask]>
> > So anyway, is *ph_2te:r the kind of lenghtened grade word you're talking
> > about?
> It is --- though Beekes' explanation (which I've now found) is
> different: The nominative -s, where found, is a later development, but
> there's a general sound law that lengthens the last vowel in a word if
> followed by a sonorant other than m (i.e., i, u, l, r, n). And since
> many athematic noun stems end in sonorants, we see this effect in the
> nominative where the ending (is zero or) has no vowel.

Huh.  I figured that looked like a rule, from looking at the Greek
declensions, but I didn't know it was one that old.

I had it inherited into Hadwan anyway, er, that whenever nominative -s had
to be added to an athematic stem where it couldn't phonotactically go,
the -s wasn't added and the vowel was lengthened instead.

And, heh, those would be the stems ending in sonorants...