Re: [tied] Re: o-grade thoughts

On 2006-08-29 20:53, tgpedersen wrote:

> Erh, and? How does the first statement follow from the last?
> You're saying that since the o-grade of perf sg is accented
> and that of derivations isn't, they are the results of different
> processes?

At least, they inhabit different accentual environments. The o-grade of
the perfect is rather clearly an ablaut variant of the root vowel. The
O-fix has some peculiarities pointing to its original independence: in
some root structures it doesn't move into the root-vowel slot but
remains in the prefix position; in others (some roots containing *u/w)
it doesn't seem to have been vocalised at all. It also fails to attract
accent the way true vowels do.

>> Had it been a real full vowel originally, we would
>> expect accent retraction in all O-fixations.
>
> Not if the prefix vamoosed fast enough, see below.

What I mean is: there's no accent retraction in the *tomh1-ó-s and
*tomh1-á-h2 types, or in *-éje/o- causatives. Normally, the root vowel
would have drawn the accent from the suffix. Here it didn't which
probably means that there was a time when this infixed *o didn't brhave
like a vowel (i.e. the structure of the stem was something like *tRmh1-ó-)

>> A _reduplicated_ stem by definition provides more room.
>
> Erh, meaning what?

A reduplication is by definition longer than the morpheme it is based
on. Why shouldn't it contain two vowels if that's what reduplication is
all about?

> But semantically, reduplication, meaning plurality, made no
> sense in the sg and must have been introduced analogically.

Reduplication may also express intensity, repetition and the like ("It
rained and rained as he rode and rode"). It doesn't necessarily take two
people to do bang-bang. As a matter of fact, I don't know _any_ examples
of IE reduplication expressing plurality, in verbs or in nouns.

> I note with relish that the PPIE *-a-a- -> PIE *-i-o- tendential
> pattern seems to be as general as PPIE *-aCa- -> PIE *-iCo-.
> In the sg of the perfect, that would mean that PPIE *aman- ->
> *imon- which could go either way of -> *mon- (with a rule giving
> loss of "i-grade" in anlaut) or -> *mimon-/*memon- by analogy with
> the plural. Now I have a theory of the o-grade of the sg of the
> perfect, and that theory reduces, not increases the number of entia
> in PIE. What do you have? Your move.

See above. Where is the evidence that reduplication was _ever_
restricted to the plural in (pre-)PIE? The only sure case of an
unreduplicated perfect (both in the singular and in the plural) is
*woid-/*wid-, and there are reasons to believe that it represents an
exceptionally early case of de-reduplication (due perhaps to its
frequency of use).

Piotr