Piotr:
>We also have the unalternating *-tor- suffix of agent nouns and the
>heteroclitic *-tr/*-t(e)n- of neuters (like *h1ei-tr 'route'), very
>popular in Hittite . The less securely reconstructed "woman" suffix
>*-sor-/*-sr- and the conjectural kinship suffix *-h2ter- have no *-n-
>alternants. If heteroclisy ever occurred in active/animate
>paradigms, it must have been eliminated very early.
Perhaps my question wasn't specific enough. I meant to enquire into
the existence of an inanimate *-r suffix derivable from Mid IE *-r
alone... not *-er- (< *-ere), *-tor- (*-ta-(e)re) or -sor-
(< *esare) suffixes.
The existence of *-r is apparently rare to non-existant in MidIE.
Afterall, words often terminated in vowel according to the
penultimate accent rule. Since *-er- comes from *-ere (again,
based on the PA rule), we don't really have a (pre)historical *-r
in IE, do we? We only have the choice of the inanimate *-r which
we know derives from *-n. This is in part why, if we must divide
*kWetwores up into smaller roots, the suffix *-r<*-n appears to be
required.
We know that, by analysing *kWetwor- as **kWet- + *-wr, we are
immaturely introducing a slew of unsupportable assumptions such as
the unattested existence of a verb **kWet- and a curious semantic
shift from the resultant tentative verbal noun into the numeral
"four". These assumptions are simply too fantastical to enable us
to salvage this theory. Logically, this view must be abandoned.
I do not propose any odd verbal root. I simply assert that
*kWetwor- is derivable from an earlier form meaning the exact same
thing. This fact in itself shows greater Occam compliancy. I
claim that *kWetwor- is derived from Mid IE *kWetW�res based on
concrete, easy-to-understand and ordinary sound changes that are
readily applicable to the rest of the Late IE lexicon in order to
derive the Mid IE stage. From this, I understandably say that
*kWetWares is simply *kWatWen + *-es (plural), based on
vocalic-stress constraints that explain later *e-*o alternations
very well and on the *-n>*-r rule that helps to explain the origin
of the strange alternations of the heteroclitic inanimate
declension. We might also reconstruct **kWatWer, however, we are
then unable to rip a non-existant *-r suffix from the Mid IE root
to help explain the *r-less *kWetWa-xe "eight" (>*ok^to:u) and the
unaccented *kWetWe- in compounds.
By taking away *-es, which is most evidently a plural suffix, we
are also forced to reconstruct *kWatWen as opposed to **kWetWare,
which remains the other possibility.
You see, if we felt compelled to dissect **kWetWare, we could
only divide such a word into a Mid IE **kWetWa- (some verb?) +
*-ere (actor suffix) but this would make our theory as unlikely
as the previous **kWet-wr theory. Yet again, we would be forced
to illogically assume a verb or noun root **kWet-. If we assumed
that the word **kWetWare was indivisable, we'd also reach a dead
end because no etymological sources for the word can be seen
outside of IE, whether via early loans, or by inheritance within a
larger Nostratic context.
The most efficient solution then is to reconstruct *kWatWen with
an *-n suffix (functioning as an inanimate singular) that is absent
not only in MIE *kWetWa-xe "eight" as well as in compounds like
*kWetWe-esare (> *kWet(w)esor). Further, the *n-less parallel
with *xotta in the likely-to-be-closely-related-to-IE Tyrrhenian
language (nb: Etr huth, Ytte:nia) is readily apparent and
attractive.
Beyond IndoTyrrhenian, we can trace the numeral in other near-by
language groups like Uralic *kutte "six" and Altaic *�t� "five"
(personal reconstruction) based on grounded sound changes (Steppe
*k becomes Uralic *k, Altaic NULL and IndoTyrrhenian *k)
Now to your theory, Miguel. A long list of problems might include
the following:
- *kW > *p(W) is much more attested and intuitive
than *pW > *kW
- strict explanation and proof of proposed early sound
changes from Nostratic to IE and from Nostratic to
AfroAsiatic (both vocalic and consonantal) remain
undefined, although very necessary for your proposals
- it is generally agreed that, if IE and AfroAsiatic are
related, they are seperated by many moons (probably as
much as 15,000 years). Therefore, any direct comparisons
between IE and AA without more closer-to-home attestations
in languages such as Uralic, Altaic, EskimoAleut or even
Dravidian are hardly credible in both IE and Nostratic
circles
- Reliance on Dolgopolsky's reconstructions which are in
themselves filled with linguistic "variables", parentheses
and unestablished and/or unlikely sound changes with
general disregard for internal classification and
reconstruction of intermediate stages of the language
groups involved
Should I stop, or go on?
In contrast, the above critiques are not relevant to my etymology
of *kWetwores. Your theory that *kWetwor- derives from *pwat(w)-,
then, is markedly inferior in comparison because of the above
problems. BTW, this supposed AfroAsiatic *p could just as well
come a Nostratic *kW-... Who's to say different without solid
sound rules to back all this up?
Conform to Occam's Rasor. You must comply. Resistance is futile.
- love gLeN
_________________________________________________________________
Get your FREE download of MSN Explorer at
http://explorer.msn.com/intl.asp