obviation in Siouan languages

[Bryan Gordon]

How about a giant summary? Little of what follows is my original work;
I'm sure you've all heard much of this before but it's just too much
information to keep in your head at once.

I believe that there is some sort of historical areal influence that
links the Algonquian phenomenon with the Siouan. It has been proposed
before that the Algonquian phenomenon preceded the Siouan, and that
makes sense to me considering the hit-and-miss dispersion of obviation
in Siouan compared to Algonquian.

There are certainly some serious differences, though. I'll use the
pair Omaha-Ponca/Ojibwe to explicate:

The Ojibwe obviative canonically serves as a part of the theme/animacy
system in such sentences as:
Onaagaans egaasaanig obi-dakonaan.
Dish.DIM P.small+OBV+P 3s.TS-hold.DIR.OBV
She carried me a little bitty dish.
(DIM - diminutive, P - participial, TS - towards speaker, TS - towards speaker)
The covert subject is proximate, so the dish and its "small"
participle are both marked obviative. The verb is marked Direct
because the argument structure is the same direction as the animacy
hierarchy (2 > 1 > 3Prox > 3Inan/Obv).

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

While I haven't seen Ojibwe referents switching from obviative to
proximate or vice-versa within a short span of text, I have noticed
obviative referents switching relatively rapidly (and, unsurprisingly,
have not noticed that for proximates). For instance:
Mii iw animoshishan wegitiziimimaad anindowa gaa-aawid.
CNF CMP dog.CTP.OBV P.have.for.parent.DIR.3s.P such P.PST-be.sth.3s.P
Gii-bezhigowan oniijaanisiwaan akwezensan.
PST-sole.OBV 3.POSS.child.3p woman.DIM.OBV
One such [farmer] was one of those who had a dog for a parent.
They had an only child, a daughter.
In the first sentence the obviative is a dog; in the second, it is the
child; the proximate is the same for both, although the number
changes. (This is a good example of another areal phenomenon, in which
a singular can become plural when it includes "implicit" others such
as, in this case, a husband.)

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.

The Omaha-Ponca obviative, like the Ojibwe obviative, occurs with
animate referents alone. However, Ojibwe marks animate obviatives and
inanimate plurals with much of the same morphology, which suggests
that the effect starts out somewhere pretty close to VP or
TP/IP/whatever-you-like in the syntax.

This is not the case at all for OP. As mentioned earlier in this
thread, there are a couple of different phenomena that have to do with
obviation in OP. The most noticeable is that there are separate groups
of articles. Some of the obviative articles encode the same positional
information as the inanimate articles:
kHe for horizontally scanned inanimates vs. kHe for dead animates
tHe for vertically scanned inanimates vs. tHaN for standing animates
ge for scattered/diffuse inanimates vs. ma for multiple animates or
generic classes
Others do not: there is no inanimate analogue to "thiN" - moving; nor
any animate analogue to "thaN" - symmetrical/round.
The proximate animate articles are akHa and ama, and seem to have
nothing to do with the obviative articles historically,
synchronically, positionally, pragmatically, semantically or anyhow at
all. The obviation system of OP, in short, is epiphenomenal compared
to the well-integrated obviation of Ojibwe.

Also, OP animate objects are almost never marked with a proximate
article, and animate subjects likewise are only marked obviative when
they occur in an intransitive sentence. In Ojibwe, on the other hand,
an object is never marked obviative unless there is another animate
referent competing for attention. I believe there must be two
referents in short-term memory (activated) in order for obviation to
occur in Ojibwe; this is not the case for OP, which has encoded
obviation in a fashion more similar to Indo-European case-marking.
We've all seen a few examples of how Siouan split transitivity
intersects with obviation.

The other way that obviation is marked in OP is actually quite the
opposite: proximacy is marked with the "plural" morpheme at
clause-end! So it is quite possible that in OP proximacy is the marked
phenomenon, in direct contrast to marked obviation in Ojibwe.

A potential gold mine, which I would love to hear others' thoughts on,
is the Legend of Ukiabi (Dorsey 1890 pp 609-612).

In this story there's an action sequence in which Ukiabi and his son
are constantly being referent-tracked. The following differences are
apparent, some of which have been analysed as obviation effects, and
others which I've never heard mention of:

Ukiabi gets the proximate -bi and -i ("plural") affixes. His son does not.
Ukiabi gets the proximate a- prefix on motion verbs and "have". His
son does not.
Ukiabi gets the proximate continuatives ama and akHa. His son has only
one sentence marked with a continuative, and it is kHe
(lying/dead/horizontal).
Ukiabi's sentences regularly end with evidential tHe. His son's do not.
His son's sentences regularly end with reportive/dubitative ama. His do not.