Mandan (again)

[Koontz John E]

Alan Hartley asks:
> Is it correct to say that, in both Mandan & Dakotan, -awa- (reduced to
> open back -a-) is a variant of -ã-? In other words, that the modern
> Dakota forms (e.g., mawa'tadã) could represent mãta + diminutive
> (whereas an older Assiniboine form was mãta + locative).

No, only that maNwa might easily reduce to something that could be
perceived as maNaN or maN in Omaha-Ponca, and, I think, Dakotan.  I can't
say for Mandan, and you've stated things the reverse of what I would have.
aN is the variant of aNwa.

Incidentally, I've seen the diminutive reduced to =n in Stoney, e.g.,
wasabe=n for 'blackbear'.  I believe this form was in the CSD from Pat
Shaw. I don't recall seeing =l in Teton for this.  Of course, a Teton
speaker might produce =l by way of adapting =n without knowing it was a
diminutive.

So, what I'm saying is that a Dakotan speaker saying maNwata=... might be
perceived by somewhat with less than complete facility to be saying,
maNaNda=... or maNaNta=...  Whatever followed (the ...) might have been a
locative or a diminutive or might have varied with the time and place.

> Following is a revision of the etymology, correcting the enclitic forms
> to -l/-n and making it clear that the French name was borrowed from
> Assiniboine. (Please let me know if the tildes don't display correctly
> for anyone.)
>
> MANDAN < Fr. Mantanne (1738, explicitly as an Assiniboine ethnonym; and
> cf. late 18th-cent. Sp. Mandana), or directly from its etymon
> Assiniboine mãtan < Mandan mãta 'Missouri River' + Assiniboine locative
> enclitic -n. The English variants in -l (e.g., Mandal, Mandelle) reflect
> the Lakota form of the name, in which the locative enclitic is realized
> as -l.

This is close, but I'd say:

---
MANDAN < Fr. Mantanne (1738), explicitly from an Assiniboine ethnonym
which could be reconstituted as *MaNtan or perhaps *MaNwatan; and cf. late
18th-cent. Sp. Mandana, from maN(wa?)ta + Assiniboine diminutive or
locative enclitic -n. The English variants in -l (e.g., Mandal, Mandelle)
reflect a Lakota form of the name, in which the enclitic is realized
according to phonetic correspondences between the dialects as -l.  Compare
modern [Santee?] maNwata-daN '?' + diminutive enclitic -daN (ref).  For
maNta, cf. perhaps Mandan maNta 'Missouri River' (ref).  The derivation of
modern attested maNwata- is obscure.
---

It's pretty clear from the way the form gets adapted from one dialect or
language to another that the morphology of the form is obscure to speakers
and has been for some time.