On Thu, 12 Feb 2004 21:58:31 -0800 (PST),
enlil@... wrote:
>Me:
>>However, as I told Miguel *septm would have to be from
>>Common Semitic since East Semitic doesn't work to explain
>>the mimation chronologically or geographically.
>
>Miguel:
>> East Semitic is the *only* solution that works to explain
>> the mimation.
>
>You lie. It is not the only solution. The clear solution
>is _Proto-Semitic_ itself. Where do you think mimation
>comes from?
In the numerals, it's only attested in Akkadian.
>The only way to deny this solution would be to claim
>somehow that mimation is purely an East Semitic thing
>but it's not last time I checked. It's in Proto-Semitic,
>since it's not only present in Akkadian but Arabic too.
Nope, there's no mimation in Arabic. There's mimation in South Arabian,
but not in the numerals.
The patterns in the singular are:
E Semitic -m
NW Semitic -0
Arabic -0 (def.) or -n (indef.)
S Semitic -m
We have essentially the same pattern in the fem. pl. In the masc. pl., we
have:
E Semitic -0
NW Semitic -m or -n (Aramaic)
Arabic -n
S Semitic -n
In the dual, the patterns are:
E Semitic -n
NW Semitic -m or -n (Aramaic)
Arabic -n
S Semitic -n
It is clear that Proto-Semitic had no mimation in the (masc.) pl. and the
dual: the original marker must have been *-n (a plural morpheme), which
only became -m analogically in Canaanite and East Semitic, and was lost in
East Semitic after the long vowel of the nominal plural (-u:, obl. -i:) [-n
is not lost after long vowel in the Akkadian dual -a:n, so we must posit
*-u:n > *-u:m after back vowel, extended analogically to the oblique *-i:m,
and then loss of *-m after long vowel].
In the singular and fem.pl., the situation is not so clear. There are
several possible solutions. One of them is that PS had mimation in the
singular, which was lost everywhere except in the outliers East and South
Semitic. But that leaves Arabic indefinite -n unexplained as an Arabic
innovation. A second solution is that Arabic continues the original PS
situation, and that there was a distinction between indefinite *-n and
definite *-0. In Sabaic and Akkadian, *-n became *-m after nominative /u/
(and analogically after Acc. -a and Gen. -i), and was generalized to be the
absolute state marker (although Sabaic has no mimation in certain
categories of nouns, such as numerals, names of places, regions, peoples
and gods, terms denoting materials, and some fixed expressions). In NW
Semitic, the unmarked definite state was generalized to be the absolute
state marker. Arabic maintained the original distribution, with indefinite
-n and definite -0.
In summary, it's far from clear whether Proto-Semitic had mimation in the
singular of numerals. I think the evidence on the whole suggests it
didn't.
=======================
Miguel Carrasquer Vidal
mcv@...