Re: [tied] Bader's article on *-os(y)o

--- In cybalist@yahoogroups.com, "Rob" <magwich78@...> wrote:

> I agree that wherever thematic vowels occur, they are the result

of

> stem-final vowels.

Most gratifying.

> However, why does verbal thematic *o only exist
> before nasals and nowhere else?

I don't think that is the case. The thematic verb has -o- in the
optative *-o-yH1- (thematic vowel + zero-grade of *-yeH1-), i.e.
before /y/ as in the nom.pl. *-oy of pronouns. There is also -o-
before /w/ in the 1du, Skt. -a:-va(s[i]). I have suggested the same
before /r/ in the 3sg middle, as OIr. -berar 'is carried' from *bher-
o-r. The last one will agree with the prnominal adverbs saeen in
Goth. hwar 'where', thar 'there'. I have assumed the regular
distribution is -o- before voiced segments, -e- elsewhere. That
seems to have been largely accepted on this list. The corollary that
the nom.sg. *-os is then from an older form *-oz, however, has
caused some debate, as you may have noticed.

> Furthermore, why do the pairs above
> exist at all? Being syllabic suffixes, one would think that they
> would *always* be accented.

In principle, IE accent assignment is not predictable, but
distinctive. There are types, however, and many conceivable things
do not exist.

> About *h2ak^mons (> Gk. akmo:n) 'stone': Why would 'stone' be an
> animate noun?

I really wouldn't know.

[...]

> So then Mr. Gasiorowski is in error?

Well, I said I have a different opinion.

[...]

> Not to change the subject, but what do you think is the origin of

the

> 1sg *-o: ending?

I have swept it under a carpet carrying the lable "unknown". I do
tend to accept Cowgill's derivation from *-o-m-i, developed much as
in the instr.pl. *-o:(:)ys from *-o-mis, only I would rather derive
the latter from *-oy-bhis. Both derivations would be largely ad hoc
and carry no weight of their own. I have drawn attention to the
imponderables: We do not know what the thematic vowel was at the
relevant stage, nor what the "primary marker" *-i or the instr. pl.
ending were really made of. That makes it anybody's guess.

> > The general explanation of this, which I accept, is that

it

> > reflects old root aorists which of course had ablaut, so that

e.g.

> > the 3pl would be *bhug-ént. Adjusted to the thematic classes

with

> > 3pl in *-o-nt this then gave rise to a type with zero-grade

+

> > accented -é-/-ó-.
>
> How did that happen? Also, are there any historical attestations

of

> *bhugént?

If the 3pl aorist ended in *-ént, it could be adjusted to the many
thematic types which formed *-o-nt, so that the new form was
reinterpreted as *-o- + *-nt, which could lead to replacement of the
1.2.3.sg *-m, *-s, *-t by *-o-m, *-e-s, *-e-t. Adjustment of an
athematic ending to the thematic type is not rare, cf. Latin sunt,
OCS soNtU vs. Oscan sent, Gk. eisí, Goth. sind, Old Welsh hint.

The form *bhug-ént is not attested, AFAIK, but the type is, cf.
Vedic ákran, ágman etc, from augment + *kWr-ént, *gWm-ént. If the
type had *-ont in IE already, so much the easier.

>
> > For a few verbs, as *wid-é-t 'saw', *H1ludh-é-t 'ascended' this
> form
> > had been reached already before the disintegration of the
> > protolanguage.
>
> What about forms like *e-we-wkW-é-t 'said'?

That is better derived from the 3sg middle form in *-é. The thing is
that the reduplicated aorist has a causative function, so that forms
with the same meaning as the base-verb are liable to reflect the
middle voice. That may incidentally also be true of *wid-é-t,
*H1ludh-é-t, cf. Gk. fut. eídomai, eleúsomai indicating that the
verbs belong to the middle voice. The process whereby *-é became *-é-
t may be seen as one of marking: Since the 3sg middle had no
consonant marking the person, the *-t of the active could be taken
over, especially in verbs that were not semantically associated with
the middle-voice anymore. The new 3sg form *-et was segmented *-e-t
and led to *-o-m, *-e-s etc.

[...]

>
> Hmm. That would seem to explain the
> acrostatic/hysterokinetic/proterokinetic distinction. What do you
> think motivated the change?

What change? I don't see any.

> But *wlkWos, like Glen said, had stress
> on the first (that is, the *zero-grade*) syllable. What happened
> there?

I wouldn't know. It's the story of a single word. Nobody can know
that. We do not know the root, so we cannot even know that the word
is genuinely Indo-European. It could be a foreign word for all I
care.

> > This may be fatal to the theory: If new phonological rules are
> > posited they should apply to the whole language, especially the
> > parts of it that *must* have existed at the time.
>
> Your theory, or mine?

Well, yours, I'm afraid. I have taken pains to avoid such clashes.

[...]

> Okay, I can see this being possible. However, I think the theory

of

> aorist subjunctives becoming presents is debatable.

Narten (the very one) wrote about it, showing how it can be traced
through textual attestations in Vedic. We had a fierce debate on
this some time ago list when I suggested that the Tocharian presents
in *-se/o- were in origin s-aorist subjunctives of the same type as
Vedic váks.ati.

> > There is the root of Lat. liqueo:, liquidus, OIr. fliuch, Welsh
> > gwlyb 'wet' which must be *wleykW-.
>
> Hmm. Is that the only one?

You're asking a lot if you demand more: another example of PIE *wl-
not by zero-grade, but as a cluster preceding the root vowel of
something that can be confidently posited for the protolanguage. We
get an outcry every time someone says what he thinks the nominative
was like.

[...]

> Hmm. Is *H2wólk-s attested in Greek and/or Latin? The examples

you

> gave do shed a different light on things.

Frisk quotes Greek áloks, aûlaks, acc. ô:lka from different dialect
sources. Schindler boiled it down to PGk. *awolks, *awolk-a, *awlak-
os from IE *H2wólk-s, *H2wólk-m, *H2wl.k-ós. The root is that of
Lith. velkù 'pull'.

[..]

> Okay, I just got an idea -- it may be a crazy one, but here it

is.

> What if the *nókWt-s reconstruction is wrong? What if it was
> actually *nókWs/nókWt-? That is, the final *-t became *-s just as
> with the 2sg marker? However, assuming a *n-kW-t root-shape, the
> expected form would be *nakWáta > *n@...@ > *nkWát > *nkWet (>
> *nkWes ?). That is hardly what we see. Another possibility is

that

> the *-t was a suffix, so the original root-shape was simply *n-

kW.

> Perhaps then, *nakWa > *nákW@ > *nakW. Then *nakW ta > *nakW-ta >
> *nákWta > *nákWt@ > *nákWt. However, it seems that the *a would

have

> had to be long to explain the o-vocalism. Anyways, just some
> thoughts.

These calculations do not operate by anything resembling rules I
would be prepared to accept. The few s/t-stems we know ('month', the
pf.ptc.act.) seem to form nom. in *-t-s. That makes it perhaps a
little less promising.

[...]

> The *-tó/-nó/-ró suffixes must have had different, though perhaps
> (probably?) similar, meanings. I have read that the *-tó suffix

was

> a "verbal collective" or something similar.

I can't see the difference between *-to- and *-nó- which just appear
to belong to different verbs. And I would really like to know what
the basis of the diagnosis 'collective' could be here. The
participial function is one thing above suspicion. Could you try
and dig up the reference?

Jens