Jens:
> But that's the *same* root! So that cannot be an example of an "o-
> root".
Erh, but no. The words *genhos and *gonhos are not exactly the same.
They are two seperate words that happen to be from the same root. If
we say that *genhos and *gonhos are from the same root and therefore
that *e and *o are the same, then we can use that same absurd arguement
in English and say that "spin" and "spun" show that /I/ and /^/ are
the same vowel. I'm not willing to do that. So let's just accept the
phonemic status as we see it. Let's accept *e and *o as seperate vowels.
> We do not know IE all that well. [...]
Yes, the 'in our ignorance' speech :) I was waiting for that.
> I have myself posited IE *lowH1- 'wash' and *ghos- 'eat as a guest'
> as roots with lexical //o//. The main basis is *low&1-tro-m 'wash
> basin' and *ghós-ti-s 'guest' which are word formations of a type
> generally showing the fundamental vowel of the root.
Well then!
> Would you mention an example? I really do not know exactly what you
> have in mind.
From what I understood, you were trying to get *e out of verb roots
which appear to be *o-grade and the perfect reduplication had something
to do with it as with Hittite /sakki/?
>> Just because there is a
>> predominance of *e-grade verb stems doesn't mean that every stem
>> must be *e-grade. It might in fact be the opposite for all we know
>> or something that we haven't even considered!
>
> What would be the opposite of "every root having e-vocalism"?? No
> roots having e?
No, no. I mean, it could be that many vocalisms were once much more
common than they appear to be later in IE.
> That's not a fair statement. You cry for the consideration of
> further possibilities. So when I give you possibilities, that is
> wrong.
No, I cry to eliminate existing possibilities through reasoning. I
for one don't like lingering possibilities because that equates to
unknowns and there's no reason why we can't distill the existing
facts to arrive at an appropriate solution. Unknowns defeat the purpose
in the pursuit of knowledge, don't they?
> Not everybody's praise is worth coveting.
I'm touched.
> So you define an o-root. What is that, and where do we find one?
You mentioned examples like *gHos- and *louh- yourself and we can
always cite more familiar roots that defy splicing like *pod-, *kwon-
and *nepot-. However, I object to calling them some special name
like 'o-roots' since they are not any bit different or stranger from
their 'e-counterparts'. I don't know why you want to pick on them.
> Are we to specify elements we do not see? And as what? The vocalism
> of the perfect is /o/ with all roots, so that is not diagnostic for
> the identity of the root vowel. How can you know that an isolated
> case of -o- where almost all other roots have -e- is not due to
> transfer from the perfect or from some other o-category that just
> neutralizes the root vocalism?
We don't, I suppose. But then, we don't know whether there are
elves in your teacup either. We don't side with the theory that
requires more effort. It's easier to take the simplest approach: that
the *o is supposed to be where we find it, even if it doesn't suit
our whims. Sure, many forms with *o come from perfects and statives
but this still musn't always be the case. We have to avoid
overgeneralizations if they aren't warranted. You still don't show
how they are. You're insisting again.
> It is a stem. That is what you see, and that is what should be
> accepted.
No. We simply see *kwon-. It is _presumed_ that it is **pku-on-
but nothing compelling in way of evidence.
> *pod- is not the invariable full grade form of that root. There is
> also *ped-.
Ah, so you're doing it that way, hunh? So be it. I'll accept that.
>> It will always be a root until somebody finds **nep-. The end.
>
> It's a stem.
Well I don't see **nep- so I don't know how you can possibly call
it a stem if there's no way to splice it in the manner that you
suggest. People often will splice words in a ridiculous fashion.
I remember Larry Trask splicing Basque /zazpi/ 'seven' into *zaz-
and *-pi instead of accepting the obvious: the word is an old loan
ultimately from an AfroAsiatic language. I really hate that
splice-n-dice game.
>> In most, but not necessarily in all. You don't establish why it
>> MUST be in all stems.
>
> No, for I haven't said that.
Erh? So you accept roots in *-o- afterall?
> So you do not want the language to be analyzed?
Only where valid.
>> I guess what the problem is is that I don't see a need to wonder
>> why some verbs are *e-grade and others *o-grade.
>
> Well, fine help you are!
You have to understand that the way I see it, the common use of *e
in verb stems is the result of early IndoTyrrhenian vowel harmony. So
the *e-vocalism we see is from the transitive conjugation while the
*a-vocalism (later *o-vocalism) was used for intransitives. Since
the durative sprang from the transitive and the aorist-perfect
came from the intransitive, we ended up with what we see. We have
*e largely in the durative with a few oddities in *o, root aorists
in *o, and finally perfects with reduplication also showing *o as its
core vocalism.
To be clear on where I stand so far, I'd say that the *o-vocalism
seen in some duratives is simply the result of aorists or perfects
migrating to the durative. That's all. This could have happened at
any point between Late IndoTyrrhenian and Late IE.
So, in my mind, there is nothing terribly odd about the distribution
of *e versus *o in verbs, being that the durative is the default
aspect anyways, and even so, there were still always verbs with
other vocalisms in the durative throughout the development of IE
from IndoTyrrhenian times.
Does anything in my account sound linguistically odd. If so, what
and why?
> It *is* regular.
Yes, but there's no particular reason why the accent should oscillate
between the singular and plural. Therefore the accent is _unintuitive_.
In other words, the accent placement has to be _learned_ and is not
automatic. There must be regularity underlying this unautomatic system.
This is inevitable common sense. So I'm afraid the accent alternation
in *esti/*?senti most certainly should be explained first before
dabbling in the always-nebulous vowel system. Vowel are too, shall we
say, liquid to ascertain their exact origins without help from other
facts.
> IE languages are very close to actually showing that.
Yes, I know. But there's a difference between "precisely" and "close
but no cigar". If one can only find "very close" examples, it shows
all the more that reducing a system to monovocalism is more futile
than trying to solve the New York Times Crossword in the allotted
time. (Does anyone else have trouble with that damned puzzle?)
> Sort of.
At least you keep me on my toes and make me think :)
= gLeN