etymology of MANDAN

[Koontz John E]

This is very long.  If you're not interested in the Mandan name issue,
delete it quickly and escape while you can.

On Tue, 17 Aug 1999, Alan H. Hartley wrote:
> Could it be that both MINITARI and MANDAN stem ultimately from a
> Hidatsa/Crow or Mandan word for 'ford'? Moulton (Jrnls. of Lewis & Clark
> Exped. III. 206), presumably from Hollow, gives Mandan mingintari 'water

I suspect the g here shouldn't be there, i.e., I'd expect miNniNtari, or,
as Hollow would write the underlying form in his dictionary, wiriNtari.
(I'm writing VN for nasal V as an email expedient.  He uses the nasal hook
under the V.)

The form is from Hidatsa wiri-tari 'to cross water' (minitari in
pronuciation as initial w is typically nasalized), because Mandan doesn't
attest tari 'to cross', as far as I can tell from Hollow.  However,
miNniN-tari has been Mandanized to the extent of substituting Mandan
miNniN 'water' for Hidatsa wiri (miri in pronunciation)(miNniN from
underlying wiriN from still more underlying wriN, in Hollow's analysis).

By way of background/review, note that as far as I can recall Wes Jones
writes m and n for initial w and r in Hidatsa as an orthographic cue to
the typical initial nasalization of w and r, but without arguing that m
and n contrast with w and r.  Harris & Voegelin wrote w and r everywhere,
without, as far as I can recall without looking, implying that the surface
phonetics were any different from what Jones reports.  Hence Jones's miri,
H&V's wiri, and (pretheoretically) Matthews' midi, bidi, mini, etc., are
all orthographic variants.

Hidatsa lacks nasal vowels, at least within recent times.  Siouan
comparativists assume that this is because it lost them at some point in
the past. Conceivably it was in the process of losing them when first
observed, but Matthews' experience suggests they were pretty much gone by
his time, at least 150 years later.  I think this would be consistent with
what Randy Graczyk has argued for Crow and I hope he'll correct me (again)
if not.

I mention it because the nomenclature for these groups may well originate
in a period before nasal vowels and vowel-conditioned sonorant
nasalization were lost in Hidatsa.  On the other hand, as this example
illustrates, Siouan languages are pretty slapdash about adapting the
phonology and lexicon of loans from other Siouan languages to their own,
in whole or in part.  This Mandan form would be analogous to using a
hypothetical Wasser-gun for 'watergun' in German.

In Mandan Hollow argues that all m and n derive from w and r before nasal
vowels (nasal vowels do also occur independently of m and n after stops,
etc.), and that nasality spreads to the left (toward front of word)
across w and r. He also argues that all wV1rV1 sequences derive from
underlying wrV1, which is probably true enough in most cases.  Hollow
tends to write very "underlying" forms in his dictionary, but I think he
preferred a more surfacy orthography in texts.  Hence wriN > wiNriN >
miNniN in Hollow's usage, depending on context.

Hollow lists the root of Mandan 'to cross' as kxaNh, while Matthews lists
tadi ~ tari (i.e., tari) for Hidatsa.  Note that the final h in Mandan
means that the declarative -?os^ manifests an intrusive h, i.e.,
kxaNho?s^, suggesting that an otherwise deleted final h is part of the
underlying form of the stem.  (The stem would appear as kxaN if nothing
followed it.)  I'm not sure if the x here represents Hollow's hearing of
Mandan's (rare) phonemic aspiration (noted subsequently by Carter), or
what.

> ford' (< Hidatsa) as the etymon of MINITARI. Clark (1804 ibid. 207) has
> "Manodans", representing the "pre-syncopated" version. (The Eng. forms

Variations with back vowels instead of i(N) are unexpected, but not
impossible, once the form is opaque to the speakers.

> The further derivation would be mVNi 'water' (Hidatsa midi,
> Mandan mani) + a verbal form (? taNi) 'cross over' (Hidatsa tadi/tari).

I believe that Mandan has miNniN for 'water' normally, so the a has to be
explained, if only by hand-waving.

> Clark (op. cit. 209) has "Manetarres", showing V = a. I don't know what
> the topographic referent would be; I assume the Missouri was nowhere
> fordable in that area.

I believe that's true.  However, the village people tended to live on the
western tributaries of the Missouri.  The archaeological classification of
the area is based on these tributaries, which, from south to north, are:
Niobrara, White, Bad, Cheyenne, Moreau, Grand, Cannonball, Heart, Knife,
and Little Missouri.  The Niobrara and Little Missouri bound the "Middle
Missouri"  region.  The subunits between the Niobrara and Little Missouri
(exclusive)  are:  Big Bend, Bad-Cheyenne, Grand-Moreau, Cannonball, and
Knife-Heart. I suspect these streams were at least potentially fordable,
though not necessarily at their mouths.

Jimm Good Tracks:
> > > It was explained to me (by VYB, KYB &
> > > CB) that "Hira:'tsa" is short form for "Wirahatsitati' (Willow Tree
> > > Houses)" village.

I get:

hiraa  tsa
wira   hatsi     t(a)  ati
wood   ?modifier POSS? village
scrubwillow
'Willow (people) their-village'

I'm not positive what the -t- is, actually.  It's probably a morpheme,
though, not anything epenthetic.  The alienable possession formant ta- is
the only thing I could come up with.  Unfortunately, I couldn't find any
references to a *-t(V) locative postposition, which would be another
reasonable possibility:  *'willow(s)-in village'.

The hatsi plainly modifies wira, but I could't determine a gloss for it.

Given this, this is not a (set of) of contractions and substitutions I can
explain. That is, 'willow' is plainly wira 'wood' plus some modifier
hatsi, and hira(a) doesn't match wira in any obvious way, assuming that's
the match; nor does tsa match hatsi.  The problem is that though the form
hiratsa can be divided up in various ways into meaningful subcomponents,
the results don't make any particular sense.  So we're left trying to
compare an analysable wira-hatsi with an unanalysable hiratsa.  This
yields a match *ira*ts*, where *'s are things that just don't match.
This makes for a very unpromising comparison.

Thus, I suspect the association here is just that the Hiratsa people (a
subdivision of the present "Hidatsa") lived in a village called 'Willow
village' at some point, leading to these terms becoming synonymous without
being in any way related.  Of course, hiratsa might be just another
village name, from an earlier or later period.  Village and ethnic names
are theoretically distinguishable, but they are interrelated.

If hiratsa really means something like what wirahatsi means, then it
doesn't mean it in Hidatsa, and I have no idea what language it would be.
One interesting observation here, though, is that the Hiratsa are the
Hidatsa segment/village who say that they originated on the Missouri,
while the other segments, the Awaxawi and the Awatixa, say they arrived
from the east.

This sort of thing can happen in origin stories easily enough, because
they fuzz over issues like the independence of ethnicity and language.
However, it can't happen in historical linguistics.  People who come from
different places speak different languages, leaving aside the possibility
that the same language is spoken over the entire territory in question,
which merely refers the spread of the language/people to an earlier
period.  Presumably either the Awaxawi and Awatixa brought the Hidatsa
language and the Hiratsa adopted it instead of something else, or the
Hiratsa had it already and the Awaxawi and Awatixa adoped it from them.
I've tended to assume - Occam's Razor - that the Hiratsa were originally
Mandan speakers, who adopted the Hidatsa language from the
Hidatsa-speaking Awaxawi and Awatixa, but maybe they spoke something other
than Mandan, something that has disappeared more or less completely,
except for possible relicts like the word hiratsa.

Incidentally, hiratsa would come from earlier *hirasa, if it is really
Siouan, and I can't make anything of that in Mandan, where it would come
out as hiras^a.

Village culture on the Missouri River goes back to c. AD 1000, and seems
to expand out of the NE corner of Iowa up the Missouri.  The prehistory of
the Middle Missouri seems to be fairly complex and to involve several
different ethnic groups that interacted extensively, somehow producing the
Mandan on the one hand and Hidatsa on the other (and presumably whatever
ethnic group brought Crow, if the Crow-Hisatsa split antedates the arrival
on the Missouri).  (The Arikara seem to arrive c. 1400 from eastern
Nebraska, so they don't need to be accounted for in the early
complexities.)  Recently a number of phases in SW Minnesota, including the
one that is associated with the precontact Cheyenne, have been added to
Middle Missouri.  Some of these might be the Awaxawi and Awatixa (and
Crow, etc.), if they arrived fairly late, but before contact.

I don't think that archaeologists claim to have any idea when and where
the Mandan and Hidatsa emerge from or arrive in the complexities of Middle
Missouri. By the time Europeans arrive, the Arikara have introduced all
kinds of Central Plains tradition practices and Mandan, Hidatsa, and
Arikara villages look pretty much the same, except, possibly, for central
open areas for Okipa in Mandan villages.  In fact, cultural hardware is
pretty uniform from Nebraska north by this point.

Middle Missouri is interesting in a way that archaeologists haven't
actually addressed, in that it's closely associated with the use of
age-grading societies, which in North America occur only there and in the
Blackfoot and Gros Ventre/Arapahoe north of Middle Missouri.

> Note Matthews (1877, p. 35): "It is said by some to mean willows; but I
> know of no species of willow that bears this name. By a few of the tribe
> it is pronounced Hidaa'tsa, and in this form bears a slight resemblance
> to the word midaha'dsa, the present Minnetaree generic name for all
> shrub willows." (midaha'dsa is, I presume, identical to Jimm's Wirahatsi
> cited above. Is the -t- before -ati epenthetic for euphony?

hidaa tsa
mida  hadsa (i.e., hatsa?)

The same thing in a different orthography, though shorn of the -t-atsi, as
Alan notes.  The t is probably a morpheme, as I noted above, but I'm not
positive what it is.

JEK