10.720, Qs: Proximate/obviative, ...

[regina pustet]

Just an afterthought: Of course, the Latin iste - ille contrast is,
functionally speaking, by no means comparable to obviation in Algonquian
languages. The former implies deixis proper, the latter pragmatic deixis,
or discourse perspective. But this functional aspect would, at least
theoretically, not preclude the occurrence of coordinated NPs which differ
with respect to obviation status. Intuitively, I'd say, something like
this is unlikely to occur, but the more I think about it, the more
fascinating I find this possibility. However, the STRUCTURAL facts
connected with Algonquian obviation/inversion systems probably rule out
this possibility. In such a system, marking for discourse perspective is
inseparably linked with marking for semantic role -  perspective markers
cannot be shifted around as freely as demonstratives without affecting the
structure of a clause as a whole. In particular, the structure of
Algonquian obviation systems basically requires that in a given clause, a
proximate and an obviative participant fulfill different semantic roles.
However, coordinated NPs, as in the Latin examples, fulfill the same
semantic role; and this is where the contradiction lies. But this state of
affairs describes the situation for transitive clauses only. Coding
restrictions might be a bit more relaxed in intransitive clauses. Very
much to my surprise, in Uhlenbeck's Peigan Blackfoot texts I have
discovered examples of simple clauses in which the very same participant
is coded as a proximate on the verb, but as an obviative on the NP, or
vice versa. This is the reason why I'm sympathetic to the idea that it
might be possible to play around with obviation markers in coordinated NPs
as well, at least in intransitive clauses -- although Im unable to come up
an example at this point.
R.P.