Jens:
> Nothing tells you that Syncope has anything to do where you see only
> consonants. [...] That looks like a good reason to make the preform as
> short as can be, i.e. *-m.
My solution pays attention to not only the evidence but to phonotactics
every step of the way. There is no point adhering only to evidence if
that evidence doesn't bring us to a clear and natural _structure_.
Every single stage of my account from MIE to IE itself has defined rules
governing what a valid syllable looks like as well as a balanced
typologically plausible vowel system. I've gone to great pains to adhere
to natural structures and evidence here to has showed me the way.
In MIE, the syllable shape mirrors what we see in Uralic -- that of a
CV(C) shape. In eLIE, it allows syllables like CVCC and CCVC or even
CCV:CC because of Syncope, but it still does not yet allow what would
be later found such as CV:CCC (*bHe:rst). It is a gradual and natural
change that I observe.
Furthermore, not only do the phonotactics the way I see it help connect
IE to Uralic better, they help us etymologize some IE roots to their
appropriate source, Semitic. The large number of *s-roots can be traced
to Semitic causitive stems. The Semitic causitive prefix appears to be
*sa- and so here too, a vowel seems likely to be present even if we wish
to throw my carefully designed phonotactics of MIE out the window.
> If one wants the endings *-m, *-s, *-t to be derived from full
> syllables or indeed fully vocalized independent words, it is easy to
> imagine that such a prestage existed at a still earlier time. That
> prestage could then be a prestage of more than just IE.
Not so, thanks to Tyrrhenian, but we will have to disagree for some
time yet until that field develops further. Unfortunately, Tyrrhenian
denasalizes and so there is only /mini/ < *min-e where *-e is the only
thing left of IndoTyrrhenian *-am.
> The grand parallel case is said to be *to-d, but that is not parallel
> at all. If they were, *so should be *so-s (and *to-d should be *to-t).
No. I've already explained that *so was _undeclined_ from the beginning
while *to- was always declined. The use of *-d in *to-d emphasized the
new inanimate role of the originally gender-neutral demonstrative. The
added advantage of applying the suffix to the inanimate at that stage was
to differentiate it from the bare locative. With *sa, there could be no
confusion because it was undeclined for case anyways. So *tod developped
but **sos didn't.
> I have in fact posited *so-s as a prestage of *so, but you won't have
> that, presumably because it may backfire.
No, because it's an ad hoc rule that has no connection to anything else
in the language. To posit a one-time loss of *-s for no real reason is
hardly convincing.
> Do we have enough etymologies combining IE desinences with external
> material to show us what the genetic match of the IE nom. *-s would
> to be in the other Nostratic branches?
Yes. It would be **-sV in Uralic. There is no such morpheme and we know
that marked nominatives are rarer than unmarked ones as Uralic shows.
So the original prestage must have an UNmarked nominative, logically.
And since we can derive *-s from *so, there are no further problems. Just
solutions.
On Kabardian and antipassives/imperfects:
> I cannot follow this. Is there anyone reading this which is willing
> to explain it to me?
I'll find the article and get back to you.
>> Yes. A merger of stative into the perfect.
>
> A very unsure thing. Sihler's stative is the perfect; Oettinger's
> stative is the middle voice; the stative of most other scholars is
> the derivative type in *-eH1-. Now you are using the word stative to
> designate the presumed prestage of the hi-conjugation.
Ugh, well what I mean is that whatever verbs were marked with *-x
in the 1ps (the "stative") came to be marked with "perfect" suffixes
like *-xe. As for *-eh-, I remember it being called an inchoative.
On the derivation of the middle:
> Especially, why would an ablauting vowel *-e be added to the
> consonants of the desinential morphemes of the perfect?
As I said, the ending vowels of the perfect were applied during a
prestage of IE when the "perfect" served a different function. The
mediopassive shows the same vowel, only it was unaccented originally
and followed by a voiced consonant, usually *-r which I stated
previously was a postfixed particle at one time conveying an agentive
nuance. The *-r caused the preceding unaccented schwa of the suffix to
turn to *o instead of *e in middle forms. Eventually, the *-o itself
came to be the new identity of the middle. The change in accent though
still stumps me. It must be analogical.
At any rate, the middle dates back to phrasal patterns in MIE but this
doesn't mean that the middle can't be ancient. It's just that it wasn't
originally a verbal category like durative, aorist and perfect was.
= gLeN