Osage

[Koontz John E]

On Sat, 19 Jan 2002 rlarson at unlnotes01.unl.edu wrote:
> 1) By the Sacred Legend, the Dhegihans originally lived east of the
> Mississippi, in the Ohio Valley. At some point before contact, they
> crossed the Mississippi and differentiated into three subgroups: the
> UmaNhaN, or 'Upstream People', who went north, the Ugaxpa, or
> 'Downstream People', who went south, and the NiukaNska, or 'People of
> the Middle Water', who lived in between, and became the Osage and Kaw.
> (I think this is the standard popular assumption.)

I agree, except that I think that the specifics of the Osage/Kaw as the
Niu(s)kaNska are actually your interpretation, this term being specific to
the Osage and, as far as I know, not previously attached to this story.
In fact, I think you said the Osage and Kaw were missing in the Sacred
Legend in your summary of the 18th.  I mention this only because it shows
how easily stories like this grow.  Anything that makes sense becomes part
of the canon.  I may have misparsed the scope of "In the Sacred Legend,
..." failing to properly perceive your summary here as going intentionally
beyond the legend to encompass your own hypothesis.

I omit Rory's summarization and elaboration of his own version (# 2),
which, if he will work it up, is certainly worth pursuing.

> 3) The Sacred Legend is a nineteenth century concoction designed to
> explain how linguistically related tribes such as the Osage and Quapaw
> came about with respect to the Omaha.  It's claims of an earlier
> homeland in the Ohio Valley are false, as is its account of the
> crossing.

False as presented.  If true in any sense, they would presumably apply
only the the Dhegiha, and then a Dhegiha population that was
linguistically undifferentiated.  A systematic tripartite terminology like
Omaha/Niu(s)kanska/Quapaw, if the system isn't a figment of our ex post
facto analysis of the pieces, only makes sense after emplacement in the
Transmississippi.  It seems to be a fallacy to assume that any part of it
existed before that, like assuming that the 13 colonies were founded by an
immigration of 13 tribes named Hampshirites, Massachusettsians, etc.  I'd
say that a sure sign that a story is made up after the fact to explain an
existing situation is an attempt to explain the whole cast of nations at
the time the story is first known by importing them as a set from
someplace else.  If an immigration occurred, it would probably have been
made by an undifferentiated ancestral nation.

I might add that as soon as one starts thinking in terms of a story that
recounts the immigration of undifferentiated Mississippi Valley Siouan
people or even merely undifferentiated Dhegiha people, one is probably
talking about something so long ago that it becomes rather difficult to
credit there being a useful memory of it.

Of course, if this story were available in independent versions from all
or even several of the Dhegiha groups, e.g., not just the Omaha, Ponca,
and Otoe (a local group in the historical period), but also the Osage,
or, particularly, the Quapaw, it would have considerably more weight.  But
when the sources are all centered on the Omaha Agency area and differ
rather widely in detail, I tend to see it as a local (and recent)
tradition.

> In fact, the Dhegihans differentiated in about the area of
> southwestern Minnesota and northwestern Iowa. From here, the Osage,
> Kaw and Quapaw moved south, leaving the Omaha-Ponca in their ancestral
> homeland.  (This is John's view, as I understand it.)

I used to get much more specific about this, but as Oneota archaeology and
my own understanding of Mississippi Valley Siouan get more complex, I've
retreated to a sort of core of two things that I take to be facts:

A) A linguistic entity like Mississippi Valley has to originate in a
relatively restricted area.  It cannot have come out of the ground over
the entire territory attested for MVS languages at contact, let alone the
more extended range that developed subsequently.

B) Several MVS groups have been assigned by archaeologists to Oneota and
"quasi-Oneota" entities (i.e., Psinomani for Dakota) in the Minnesota
(Dakotan) - NE Iowa (Ioway) area.  Others have been assigned to Oneota in
Wisconsin (Winnebago) or the lower Missouri River area (Omaha, Kansa,
Osage), though these assignments have been hotly disputed in various
quarters.  I suspect there is something in the general association.

I'd go further, with a sort of syllogism:  given A, either B is correct.
including some of the disputed assignments (in general, if not in
specific), or B is wrong.

I should emphasize that this sort of argument is pretty fuzzy.  I'm not
sure how big the area involved would be.  Maybe it could extend from
Minnesota to the lower Ohio, though I'd be surprised if the cultures
involved were as divergent as this would imply.  And, perhaps B really is
wrong.  I've sometimes suggested that the southern influences Bob Rankin
sees in Dhegiha could be explained in Minnesota in terms of the known
Mississippian settlements in the area, contemporaneous with Oneota,
Aztalan being the best known, though not the only one and not in what I'd
assume to be Dhegiha territory.  But maybe the influence I attribute to
these settlements was more intense even than I've assumed, and included
introduction of some of the more northerly of the Mississippi Valley
Siouan languages, and not just some areal influence on them.

Things are always a little bit slippery when we try to assign languages to
pots.

> First, I think John's view has much to recommend it in terms of
> geographical parsimony.  If we can place the Dhegihans here, with the
> Dakotans in south-central Minnesota, the Winnebago in eastern
> Wisconsin, and the Chiwere in eastern Iowa, then we have all four
> branches of MVS neatly wrapped up in a compact area.  Within this
> area, dialect gradients can nicely explain features that link Dakotan
> with Dhegihan and Winnebago with Chiwere, and that link Dakotan with
> Winnebago and Dhegihan with Chiwere.  Further, we might also be able
> to recognize a common MVS archaeological culture in the Effigy Mound
> culture, which covered a good section of that territory.  Is this your
> thought too, John?

That's pretty much what I like about it, though I'm thinking in terms of
Oneota and I'm not sure to what extent archaeologists are willing to
identify Effigy Mound as the burial manifestation of Oneota or some of
Oneota.  It does seem that there is a historical tendency to perceive the
settlements and the burial sites as separate, perhaps erroniously.

> > There are difficulties here for a modern student.  For example, why are
> > all these tribes together, though already differentiated?  There are ways
> > that this might happen, but perhaps it represents a literary encoding of
> > the recognition that they are lingusitically related, coupled with a lack
> > of realization that such a relationship normally arises due to
> > differentiation from a common source language.
>
> Why should it be a difficulty that these tribes should be together?
> (I assume you're referring to the Ioways being with the Dhegihans.)

Yes, and to the Dhegiha groups being together, too.  I know of some
limited cases of fragmentary groups associating with others or several
fragments of larger groups coming together in larger settlements, but most
of these associations seem to have been rather temporary.  I think that
seeing families of nations travelling together smacks of rationalization.
Relatives may well travel together, and undifferentiated ancestral
nations, but not usually entire fully differentiated related nations.
Linguistic differentiation implies separation.

Something else that may earmark this as a rationalization is that there
are no specific individuals named.

> We don't know enough about the history of the tribes before contact
> even to be sure where they were, much less to rule out possible social
> relationships between them.  Even if we assume that the Ioways were
> Oneota and based up in Iowa, just one group of Ioway visitors that
> made an impression at the time could easily have ended up in the
> Sacred Legend.

I guess it just seems to me more plausible to see the presence of the
Ioway as a flaw in the account than to take this line of reasoning.

> And if the Sacred Legend is no more than a naive rationalization of
> language differentiation (which seems intrinsically unlikely to me),
> why should it include the geographically distant Quapaw and the
> linguistically rather distant Ioways while ignoring the Osage and Kaw,
> and the geographically proximate Dakotans and Winnebagoes?

I've seen a version that includes the Winnebago, and, I think, the Osage
and Kansa, but never one that includes the Dakota.  I suspect some special
knowledge is involved.  For one thing, in the mid to late 19th century the
Quapaw were largely resident with the Osage, if I recall.  I've actually
seen a proposal (from Ludwickson and/or Shea?) that Omaha familiarity with
the Ohio Valley was acquired during steamboat and/or rail visits of
chiefly delegations to Washington for treaty negotiations.  I deduce that
speculations on the etymology of "Ohio" might date to then, though such
speculations might well occurred anytime and be secondary to the story.

> Thanks to you and Wally for this information! This is definitely
> relevant, but not necessarily a counter-argument.  So English got Ohio
> from the French, who got it from the Seneca, whose name for the entire
> Ohio River, including the Allegheny, was ohi:yo?, probably originally
> meaning something like 'Great River', or 'Rio Grande'.  Now, am I
> understanding correctly that Osage Opha=p=a is fully cognate with
> Omaha Uha=i, and is also the name for the Ohio River?

The Osage for 'to follow' is opha.  This is one of the stems where
Omaha-Ponca have h for ph.  The Omaha plural is =i.  The Osage plural is
=pi.  I believe the analog of the Omaha =hau (earlier =ha) male
declarative is =a, in Osage, and that this merges with =pi as =pa.  The
older Omaha-Ponca female declarative is =he.  Osage =e also merges with
the plural =pi, as =pe.  From this I reconstruct *opha=p=e as the Osage
equivalent of uha=i=hau.  This is not attested, as far as I know, and
presumably does not exist.

What I meant was to poke a little fun at the possiblity that Ohio /ohaio/
in English could have come from anything in Dhegiha at all.  In fact, the
Omaha version is clearly a folk etymology of the English version.  The
Omaha name for the Ohio comes from English.  If the Omaha ever had a name
for the Ohio, I very much doubt it was uhaihau or anything like that.
Before about 1880 or so, it would have been uhaiha.  In Proto-Dhegiha it
would have been ophapi, and it's getting progressively less like ohaio as
it goes.  Apart from this, English speakers did not learn the name of the
Ohio from the Omaha in the 1890s, but from the French in the 1600s.

> Of the Okaxpa and the ImahaN Quaxpa villages, which was upstream and
> which downstream of the other?

All I know is that they varied over time.  My point is that if we have to
see Okaxpa as opposed to some 'upstream' group, we can look a lot closer
than Nebraska.  In fact, you can easily be downstream of your neighbors
without the neighbors being named upstream, and if you move at a later
date to a location upstream of those neighbors you may be called
downstreamers still.  The English no longer dwell in the angle of the
northern German coast, if that's the basis of the name, and there are
Suffolks and Norfolks all over the landscape in the US, none of them south
or north of anything in particular.

> Why should the Dhegihans be more chameleon-like than the Chiwere?

I don't know.  Archaeologists are generally happy with the proposition of
identifying all three Chiwere groups with Oneota focuses or phases
(depending on the time of the attribution).  But they have been unable to
identify any consistent cultural pattern for Dhegiha.  A number of Oneota
suggestions have been made and some disputed, but otherwise they seem to
be unable to distinguish the various groups among the local phenomena in
the areas where they are first noted.  I have actually overstated things a
bit, since I think that a little hard looking will find Oneota antecedants
for all five Dhegiha groups.  The Chameleon effect actually only occurs
with the replacement of most imperishable property with trade goods and
the spread of various late cultural fashions.  The more abstract aspects
of Dhegiha culture are still present in the 1890s and even today and are
quite distinctive.  I'm thinking of things like the clan names and other
aspects of their theories of their societies.  Not to mention their
languages ...

> One possible answer to this question would be that they were refugees
> from another archaeological province who moved into Chiwere territory
> and had to adopt a Chiwere-like mode of life.

Many aspects of the patterns attested in the Lower Ohio should have worked
as well westward.  In fact, Oneota transports rather well, too, and seems
to spread progressively south, east and west along a broad front from
Illinois through Missouri into Kansas and Nebraska over time.
Archaeologists just don't have any idea who was living at most of these
sites in most cases.  Various Oneota sites in Illinois and elsewhere have
been attributed to the Illinois.  There are Oneota phases in Wisconsin
that have never been attributed to the Winnebago.  The Psinomani materials
attributed to Dakotan look like Oneota pottery combined with a Woodland
pattern of subsistance and a fair amount of less Oneota-like pottery, too,
or, to put it another way, like Oneota influences spreading into new
areas.  The last layers in the American Bottom (OK - yuck!), the homeland
of Cahokia Mississippian, are Oneota.  There are more than enough Oneota
phases around now to account for all attested Mississippi Valley Siouan
groups.

> The collapse of MVS *u into *i may technically be a relatively simple
> change, but its effects are catastrophic.  It means that every
> morpheme in the language that was formerly distinguished from another
> only by [u] vs. [i] is now indestinguishable.

There aren't many grammatical distinctions that get lost this way,
however.  Actually, I can't think of any.  Also, since both Kansa and
Osage have *u > u" (umlauted u), both have already undergone a step very
helpful in changing u > i.  If this happened in Proto-Dhegiha, it wouldn't
be surprising if two of the daughter languages independently changed u" >
i.

Actually, there are some cases in Kansa where *i appears as u" or *u as i,
so I think that the merger is already in progress.

> If a couple of rare consonants like k? and x? collapse, it's fairly
> trivial, but for two vowels to collapse, especially ones so common as
> [u] and [i], is devastating.  I agree though that this doesn't
> conclusively link Quapaw more closely to OP than to Osage and Kaw.

The morphemes with k? and x? are realtively few, but they occur fairly
frequently.  For example compare Dakota k?u 'to give' with Omaha-Ponca ?i
'to give'.  Which of the two changes involved is more of a problem?
Anyway, we don't subgroup languages based on the consequences of a change
to the structure of a language, but on the simple occurence of the change.
We try to consider only changes that are connected, but in practice this
is often a matter of judgement.  It's very difficult to control for
changes that occur in dialect continuums, especially complexly articulated
ones.  You do begin to get suspicious (or should) when your criteria lead
to overlapping subgroups, or, as it usually happens, competing schemes of
subgrouping.