The heart of the debate:
> In terms of the commutation test /kWis/ and /kWys/ are equivalent,
> as are /ios/ and /yos/ for [yos]. Thus far, this language does not
> show a phonemic opposition between /i/ and /y/. Therefore, one of
> them is dispensable. If that is acceptable for Sanskrit, it is
> acceptable for Indo-European; and if it is unacceptable for Sanskrit
> it is unacceptable for Indo-European.
Of course, however... /i/, /o/, /u/ and /e/ can all be replaced with
/y/, /aw/, /w/ and /ai/ in Sanskrit. In other words, all these
vowels can be replaced with other already-existing elements that don't
contrast with the phonemes in question.
That rule being exposed, what do we replace *o with if it is a function
of *e? It can't be *e: because that already exists and a clear
distinction exists between *e: and *o already. It can't be *e and some
imaginary consonant *Q because this is desperate pleading and we aren't
using preexisting elements of the language. We're instead inventing
our own conlang.
So it doesn't seem any further analysis is truly possible like it is
with Sanskrit. I don't get this game. What are the rules, if any, to
play?
= gLeN