Re: Split/Fluid-S systems: A 3-way split

Chris Bates wrote:

>
>  If you put all experiencers in the dative case, then (I think) this
> would include the arguments of most stative verbs (eg to be red, to be
> ill, ...) wouldn't it?
Personally think I'd want to limit dative-experiencer to animates (maybe
just humans). Inanimates aren't really "experiencing" the qualities they
exhibit...no?  I'm not sure what case they ought to be in-- absolute?
instrumental? locative? It might vary according to the semantics.

Seems not unlike the Kash system, where human direct objects are in the
accusative only if the verb action brings about a change, or a physical
affect-- hit, bite, kill etc. (and special case, give {in marriage] and
archaic, buy/sell [a slave]). Otherwise human direct objects are in the
dative. Non-human animates, and inanimates, are always in the accusative. I
wonder if any natlang has such a system.

(Other special cases of acc. human objects might be: abandon ~adopt ~give (a
child), exile ~expel s.o.-- and 'give birth to' probably ought to, too--
perhaps historically they did, but modern-day speakers (and myself) are so
used to using the far more frequent dative that the system probably has a
lot of variation.)


So you could get a three way instead of two way
> split in intransitives:
>
> stative verbs/verbs with experiencer as single argument (argument in
> dative case)
> active verbs when action is volitional (argument in nom/erg case)
> active verbs when action isn't volitional (argument in acc/abs case)
>
>  Do you think this is a realistic system?
Yes, and I suspect there are ANADEWisms.