Re: Long-range protolinguistics (was Re: Q (Caucasian Elf))

On Mon, 26 Feb 2001 21:44:09 -0600, Danny Wier <dawier@...> wrote:

>Anyway, the case system of IE, Kartvelian, Uralic, Altaic and even Sumerian
>show likely genetic relationships for what it's worth, but AA case marking
>is greatly simplified; Arabic marks nominatives with -u(n), accusative
>with -a(n), and genitive with -i(n).  No eight-case Sanskrit system or
>ninety-two whatever cases of Hungarian...
Akkadian has a more complex case system than Arabic.

>> Regarding the noun cases of Kartvelian, there seem to be some odd shifts
>> and reversals.  How did it come that -m marks the accusative case in IE
>> and Uralic, but the *ergative* in Kartvelian?
<scissors>
>Well ergative and accusative are both more marked than nominative and
>absolutive.  Also, remember that Georgian is really a split-ergative
>language, and absolutive case is used for subjects of intransitive verbs.
>Only certain screeves of transitive verbs (a screeve is an array of tense,
>aspect, mood and voice) use the ergative case in the subject slot; some
>screeves use other cases and I'll have to look up which ones...
>
>> PIE might once have had something similar, if one accepts Lehmann's idea
>> that it also had active alignment earlier.
>
>I think the theorized process is active > ergative > nominative, which has
>precedent in real world languages.
IIRC in Akkadian -m is something like an actualizing particle which can be
added to different case endings. And there have been attempts to connect
it with the tanwins in Arabic (i. e. suffixed indefinite article). IMO the
state of affairs in Akkadian (and therefore the oldest reconstructable
state in Semitic, and thus in AA) allows to deduce all other variants
from it. But all this remains guessative while a deeper AA reconstruction
is close to absent.


Basilius