Differential case marking in Algonquian (was: obviation in Siouan languages)

[Heike Bödeker]

Dear Bryan,

> It is not clear whether two referents are "stuck" with
> their obviation/proximacy assignments once assigned. Any

No, they definitely aren't stuck with any of that. There's multiple 
proximatives (all NPs marked PRX instead of an PRX/OBV gradation), as 
well as both proximate and obviate shifts attested. A good account 
probably still being Goddard (1984) for Fox. Whereby Fox seems to not 
make much use of absolutives, i.e. zero-marked NPs, which are 
pervasive e.g. in Blackfoot and Potawatomi (Hockett 1948, text I, b):

nušnape-cu      ?o-weni?ke-t           ?esp:un
indian:AN-TOP   DEM-go.trap:AI-3:PTC   racoon:AN

ki-yawu.
PT-accompany:AI

"The Indian, who (=> when) he went trapping, the raccon was with him.

Instead of otherweise expected either:

nušnape-n-cu niw-weni?ke-nu-t ?esp:un ki-yawu

or:

nušnape-cu ?o-weni?ke-t ?esp:un-un ki-yaw-nu-t.

> Algonquianists know the answer? If so, then certainly
> there is some referent-tracking behaviour going on here;
> if not, though, it's not clear that this is any more than
> a weird sort of case-marking.

Partly it certainly does serve referent-tracking functions, but I for 
one wouldn't hesitate to call that differential case marking 
actually. It's not all that weird to have both differential subject 
as well as differential object marking, c.f. e.g. Burmese.

> I'd be interested in whether there are any cases of
> obviation in Ojibwe sentences with inverse thematic
> morphology (where the argument structure is the reverse
> of the animacy hierarchy - this has sometimes been called
> passive, but is not). I've never seen any such cases. Of
> course, it's not even theoretically possible in OP, as
> we'll see.

I'm not sure I understand this correctly. Just why should an inverse 
verb form not be able to occur with an OBV marked NP? Apart from the 
fact it indeed is not a passive because it is not detransitivized 
(Alg. lgs. additionally have such options, too, btw) but transitive, 
horribile dictu like an ergative construction.

Anyway, if Cree textbook examples (taken from Ellis ²1983) will do as 
a first starter:

Moos     kii-kiisis-ok-ow      ni-kaawiy-a
elk:AN   PV:AFF-cook-INV-PRX   1:POR-mother-OBV

"The elk (PRX) was cooked by my mother (OBV)"

That also works in the conjunct order, btw (Ellis ²1983: 504):

Tantee   kaa-ataawee-t
where?   CO:REL-sell:AI-3:S

kaa-iši-nakat-ikot           o-kosis-a
CO:REL-ANA-leave:TA/TI-4>3   3:POR-son:AN-OBV

"Where did the trader's son leave him?"

References:

Ellis, Clarence Douglas (²1983): Spoken Cree. West coast of James 
Bay. Edmonton, AB: Pica Pica

Goddard, Ives (1984): The obviative in Fox narrative discourse.PAC 
15: 273-286

Hockett, Charles F. (1948): Potawatomi, IV: Particles and sample 
texts. IJAL 14/4, 213-225

All the best,

Heike