Ugh whatever. I should be beaten with a stick for trying ;)
Jens:
> You mean some other languages having some other traces of course,
> that's fine. But the last sentence is a plain non sequitur. If there
> are traces of the augment in languages other than those you grant
> membership of the augment club, that club must have been bigger at
> one time. If it comprises Celtic, Italic and Balto-Slavic, what more
> is needed to make it Indo-European?
An innovation can spread like a wave to other dialects that were
previously unaffected at the time of Proto-IE. I don't think of
the evolution of a language, including that of Proto-IE, as being
straightforward and linear. Did satem not start in one location
and spread into other dialects too? Same thing with the augment.
Of course, on the edges of this wave, there can only be a few
traces of such an isogloss. It doesn't automatically mean that
this isogloss should be considered pan-IE.
So I don't think this is a non-sequitur although I can see how
it can be misunderstood as such.
> If a collocation of adverb + verb can grow into a one-word-form
> in the case of the augmented verbal forms, be it in PIE or
> later in a separate subsection, surely an exactly comparable
> process can be accepted to fuse a preceding "/O/" [...]
The objection I have is that you're taking functional analysis
way too far. It's like assuming that Sanskrit derives from
a one-vowel system because the analysis shows this. Uh, no.
The analysis may be appropriate for that stage of the language
at the time, and it may be cool to talk about it at parties with
one's friends from mensa... but it doesn't necessarily yield
a proper solution. Most likely you'll be strayed completely
off course. Here, I agree with the analyses behind your O-fix.
I don't think I can avoid that no matter how stubborn I feel.
However the solution that you eke out of the analyses is based
on nothing but an unwillingness to consider a simpler more
sensible solution from what I can see. There is only one infix.
Your new infix doesn't conform to the pattern seen with *-n-.
It also doesn't conform to any of the other phonemes known to
exist in IE and rather than accept these unlikelihoods you'd
rather take up arms against a sea of troubles to little avail.
Perhaps even introduce MORE phantom constructs that IE doesn't
otherwise allow.
What DOES your theory conform to? Not to IE as we know it and
that's the big problem. Ah, but then you'll no doubt whisk
us away to a Pre-IE stage which "cannot be controlled", as
you say. I dare say that it can be controlled because it is
WE who are in control of our own theories and their logic.
At least I would hope so!
I still don't see why you turn a blind eye to phonotactics
to explain your O-fix. But enough of my rant. I'm making
another go of a-Epenthesis, t'hell with everybody, and I'm
starting to see something interesting. Ignore everything
I said earlier about a-Epenthesis. We'll start afresh.
Get out the popcorn...
My new hot-off-the-press definition of a-Epenthesis:
-----------------------------------------------------
In early Late IE, after the events of Syncope,
Clipping and Nominative Misanalysis, *a was prefixed
by default to all "overloaded" vowel-final stems
or, due to association with its parent stem, *a
was infused into the initial weak syllable.
"Overloaded" just defines stems that have clumsy initial
consonant clusters and/or initial weak syllables. "Weak"
syllables are syllables containing *i, *u, *& or syllabic
resonants (*m, *n, *r, *l) at the time of the rule. They
had developped, as you might have guessed, from the
previous Syncope. Also in the latter case of infixing,
it seems that this was invariably with already polysyllabic
stems.
The new order of the rules is:
Syncope
Clipping
Nominative Misanalysis
a-Epenthesis
Saussure
Acrostatic Regularization
Vowel Shift
So it would be fair to go over the examples again and
make damn sure that the rule makes sense...
*kwan > *kwon-
---------------
We have MIE *kawana becoming *kwan. It is not a vowel-
final stem and so we don't alter it further. It
regularly becomes *kwon- after Vowel Shift. Yippy!
*sdas > *osdo-
---------------
MIE *sat:asa becomes *sdas after syncope. It is animate
and therefore is *sdas-s& in the nominative. After clipping
of the nominative *-s& to *-s, *sdas-s is misanalysed as
*sda-s. Since the stem is now vowel-final, a-Epenthesis
ensues because of clumsy *sd-, creating *asda-s and
finally expected *osdo-s. Perfectly normal.
*ktwax > *okto:u
-----------------
MIE *kWatwaxa "eight" (merely the plural of "four") becomes
*ktwax. The final laryngeal is uvularized (aka lowered or
a-coloured) by preceding *a (a rule appliable elswhere that
explains the origin of *q from allophonic *k). However
uvularized *x is unstable and disappears in eLIE, producing
*ktwa: as we also see with *gla:-s (> *glo:us) from MIE
*k:alahWa. While *gl- in *gla:-s is perfectly okay, *ktw-
in *ktwa: certainly isn't... So you know what THAT means.
Yep. It becomes *aktwa: by a-Epenthesis. The only problem
here is trying to figure out when *w disappeared but mere
Vowel Shift would produce *oktwo: and that alone is a
reason for a subsequently inaudible *w to go bye-bye.
Hence *okto:u (with final pure *-o: closing to *-o:u).
*xuyan > *ohWuyom (*o:uyom)
----------------------------
While I admit that my initial explanation of the phonetically
bizarre *ohWuyom (*o:uyom "egg") was a bit messy, I think
this new explanation will now be much better.
So we start with *hWawayana being a derivative of *hWawai
"bird". After Syncope, "bird" is *hWawi and "egg" is
*hWuyan. As you can see, the simple loss of schwa buggered
up the association between the two words entirely.
Misanalysis then spliced the egg word as *hWuya-n further
divorcing it from *hWawi. Then if that wasn't bad enough,
a-Epenthesis came along to remedy the overloading. Hence
*ahWuya-n. We then get *ohWuyom after Vowel Shift.
Everything still looks regular.
*plhas > *polho-/*polh-u-/*pol-no-
----------------------------------
While the MIE verb *paleha becomes *pleh& "he fills", the
correlating derivative *palahasa regularly becomes *plhas
after Syncope with syllabic *l in the first syllable. Thus,
the first syllable is "weak". Misanalysis causes *plha-
and thus the stem qualifies for a-Epenthesis. However,
unlike *sda-, this is a polysyllabic stem and it begins
with reduced *pl-. It remains undivorced from the verb
*pleh& and so *a is _infixed_ to the first available slot,
within the syllable *pl-. Thus *palha- > *polho-. Now,
*polho- is free to create new derivatives like post-
Saussurean *polno- or *polh-u-, for example. In *sda-, the
*a is not infixed because the origins of *sda- from
fullgrade *sed- are shattered by post-Syncope desyllabi-
fication of the core root. Whereas an association between
*pelh& and *plhas is far more natural for the layman
speaker since *plh- still contains a syllable. Thus
infixing was preferred to preserve the association.
Conclusional (or possibly Dillusional) Thoughts
------------------------------------------------
Finally, a note on weak syllables. Since *kwon-/*kunos
must derive transparently from the Syncope stage, we are
forced to reconstruct eLIE *kwan-/*kunas (from MIE
*kawana/*kawanasa). While everything here is phonetically
regular, it admits to the fact that weak syllables WERE
allowed at that stage (ie: *kunas with weak *ku-). The
explanation for the a-Epenthesis rule then is that it
occured to avoid _strong_ stems with initial weak roots.
The rule would have been that only weak case forms or
other paradigmatically derived forms could contain
initial weak syllables such as kun-as "of the dog" which
wasn't a stem on its own since it derived from full
*kwan-. For hypothesis sake, a valid derivative stem then
could never be **kun-a-, but *kwan-a- was allowed.
Also, as with *sda-, the addition of *a (with initial
glottal stop implied) produced a new structure *(?)asda-
that started with CVC- (ie: *?as-). The only monosyllabic
vowel-final stems left at this point were pronominal
(*ta- "this, that" > *to-).
The implication of this new slant on a-Epenthesis only
underlines what can already be deduced otherwise. Roots
such as *dngHu-, *suxnu- and *wlkWo- cannot derive from
this early stage and were created or changed at a later
stage of Late IE when phonotactics began to allow weak
initial syllables in strong stems too and even (gasp!)
accented zero-grade syllables as with *suxnu-.
= gLeN