Torsten:
> I understand that your objection is that genitive *-s
> is intrinsically singular and -m is intrinsically plural
> and therefore they can't have been derived from something
> non-distinguished for number, whereupon you launch a more
> optimal theory, based on the fact the pre-PIE did not
> distinguish for number?
No, close but no cigar. It's not the same thing.
You are claiming a shift of two case endings at the same
time (which requires a pattern between the two to make any
sense at all and a motivation to theorize it). I am
claiming a shift of a _single_ case ending (which only
requires that the shift in question makes semantic sense
and that such a change is required to explain something
in IE).
I object to the fact that there is no discernible pattern
between the connection of *-s and *-os on the one hand
(where a nominative singular supposedly relates to a
genitive SINGULAR) and a connection between *-m and *-om
(where an accusative singular supposedly connects with a
genitive PLURAL). Here, there is simply no organized
evolution of functions to speak of. Thus, we cannot account
for both changes at the same time. This contradiction
proves that this simply cannot be so unless the case
endings are considered in _isolation_ as two seperate
innovations perhaps. Even in isolation, these proposed
changes are highly suspect.
To add, your theory is not only untenable, it's even
unmotivated. There is nothing intrinsically problematic
with *-s and *-os or *-m and *-om to require that it be
explained. There's nothing even guaranteeing to us that
these endings are the same except amateurish comparisons
of coincidental phonetic similarity.
On the other hand, if the nominative plural is *-es and
the accusative plural is *-ns, derived from *-m + *-(e)s,
we should surmise a genitive plural **-os-(e)s > **-o:s,
not a completely unexpected *-om. This clearly needs to
be explained and is an undeniable motivation for a theory.
I can easily support the idea that the genitive plural is
derived from a locative that was originally unmarked for
plurality because, in contrast to the above unnecessary
hodgepodge of erratic case shifts that defy ordered sense,
I'm merely theorizing a case shift in isolation that does
not demand anything more than that such a shift is
semantically feasible. Since Turkish uses a locative to
indicate possession, the shift is indeed grounded in
reality.
The original plurality of the case ending is perhaps
merely added triviality to the above theory but the
pattern seen in the IE case system suggests that indeed
plurality was originally not marked for weak cases
since those endings are invariably piecemeal and prone
to dialectal variations.
To some, these plural endings almost seem to have
merged together. Logically however, it would appear
that the plural weak case endings were being slowly
derived, either from the correlating singular endings
or from other pre-existing particles and endings once
IE began to demand that plurality (and duality even)
be marked for indirect objects as well. This process
would have begun in the Late IE period (circa 4500 BCE)
when the linguistic area was already expanding and
diversifying and it must have continued into Post-IE
times, in order to explain sufficiently the dialectal
variations, gaps and oddities of the nonsingular
system.
= gLeN