Jens:
> How's this for help: I want equal rights for the two systems. One
> should not erase from the range of typologically acceptable
> solutions one that is actually known to us from Sanskrit.
But I don't. I 'erase' what is not attested: a _true_ one-vowel
system which not even Sanskrit has. No language has that, in all
honesty. However, if you mean an _abstract_ monovocalic system for
IE as we find for Sanskrit, a kind of pseudo-monovocalism, I still
have problems, _not_ because Sanskrit isn't a sufficient example to
show that this can take place but because I don't think IE has the
same qualities that would make this analysis transferrable to IE.
I don't even see the point of it in either Sanskrit or IE.
I'm not convinced that *e and *o can be reduced to one vowel because,
unlike /o/ being replaced with /aw/ in Sanskrit, there is nothing
with which you can replace IE *o without reinventing IE altogether. This
is why your analyses get strange and exotic with double-long vowels
-- It's because your analysis is flawed. It isn't possible to reduce
IE to monovocalism like it is in Sanskrit without theorizing unicorns.
This is exactly what you do. I choose not to. That's where our
difference seems to lie.
> Up to a point yes. Most instances of PIE /o/ are variants of the
> same morphophoneme as /e/; others are underlying consonants; and
> there may be unknown sources too.
And this is crazy because there is nothing in this assertion that
is inevitable logic. We can't be sure that *o derives from a consonant
in some or any instances. You only have your *o-infix theory but this
can still be modified to 'a-Epenthesis' which doesn't require the
assumptive consonant at all. The opposition of *e and *o is not just
an idle alternation but is fully phonemic. There are noun roots with *o
and noun roots with *e. There are verb roots with *o and verb roots
with *e.
> I haven't seen any such yet "equally satisfying" solutions yet.
You don't want to.
> Everything we do in analysis of this kind is based on
> presuppositions that should be specified for the statements to be
> honest at all. I do that, and this is the thanks I get.
Presuppositions, assumptions, whatever. We don't start with an
unlikely idea or one that could be equally replaced with another.
We start with ideas that are sure-bets or at least the surest we
can find. You start with complete chaos and end up with chaos. You
need an initial conclusion to base all others on. This is logical
deduction: If this, then that. If that, then so forth, etc. You
however are stubborn and wish to follow the model of "I'm gonna
side with this at random, and then base this on that and so forth".
Random is random. Your theory is random. It's chaotic because there
is no initial conclusion to base your others on.
> In _roots_, not just verb stems (and there not always, by the way).
> I am not forced to do it, I could also stay ignorant.
Why just verb stems? The nominal root *kwon- shows *o, not *e. The
root *nepot- also shows *o. There are also verbs with *o that you
just deny because you want them to go away. They don't though and
it's stupid to think that a natural language would only have roots in
*e.
I don't think IE was such a language nor does it seem to show this.
At best, it has a "preponderance of *e" and that's all that can be
safely said. It also has verb roots with *o whether you like it or not.
So we have to accept that and stop ripping apart IE or roboticizing it.
> A single example is always exceedingly rare. This is the story about
> a single language, i.e. in essence a single example.
It's a lovely story, I'm sure, but with an artificial plot.
= gLeN