???*yaS and Woraxe as Potawatomi

[Koontz John E]

David Costa recently asked me:
> BTW, I know I asked you this years ago but maybe you have some
> new thoughts on it now -- do you have any hunch as to what the
> Siouan etymology of that name for the Potawatomi is, that one
> that starts with /w/, such as Chiwere /woraxe/? Or is it still
> just unanalyzable?

Here's my answer:

Still no clue.  Sorry!

I'll mention also Winnebago woora'xe, which seems a bit unusual to me in
having a final -e here where it could perfectly well be deleted.

It does occur to me - and this is very speculative - that what a Siouan
morphological analysis would have as the root is rax(e), which could be a
fricative grade variant of *ras^(e) ~ *raz^e.

- The voicing of fricatives presents some problems that Siouanists haven't
  really solved.

- The wo is presumably a prefix contraction wa-o- (standard in MVS)
  'something in (which), something wherein is'.

This fricative grade *ras^(e), etc., could be potentially the form in
Siouan languages that merge *r and *y, e.g., Ioway-Otoe, Winnebago, Crow,
Hidatsa, Mandan, of *yas^(e) ~ *yaz^e 'name'.  Now, a root of the latter
form appears in the word Osage (*wa-z^a'z^e, Dhegiha having z^ for *y, cf.
Omaha-Ponca iz^a'z^e '(his) name').  Winnebago has wara's^ for 'Osage',
which demonstrates at least a perception on someone's part that the root
is *yaz^e (> ras^ in Winnebago), not hypothetical *z^az^e, even though
both would merge as a nicely reduplicated z^az^e in Dhegiha.

Further afield, I've recently realized that another similar form seems to
appear in the term Hidatsa, in which -datsa (/raca/), if it is segmentable
and of Siouan origin and goes back that far, would be from Proto-
Crow-Hidatsa *-rasa.  The fricative grade is off again here:  s, not s^.

On the other hand, the root here, even if there's a pattern, may not be
'name', but something homophonous, perhaps 'person'?  The hypothetical
cognate of waz^az^e in Dakotan, the other MVS branch that distinguishes *y
and *r, would be *wac^ha'z^e (c^haz^e' is 'name').

This is similar to the actually attested wic^ha's^a ~ wic^a'sta 'man',
and, of course, I got to this point by working backwards from that.  The
first vowel in this form is wrong (wi < *wa-i 'wherewith to', perhaps,
though that seems unlikely in the context)  and there's some waffling on
the fricative:  s ~ s^.  The usual assumption is that the root here is
wic^ha-, since that appears as 'them' in the verbal paradigm.  No one
accounts for -sta or -s^a, however, as far as I know, so wic^ha's^(a) ~
*wic^ha's(a) might be the stem, leaving -ta unaccounted for.

In summary, there are widespread occurrences of roots of the form *yaS (S
= s/s^/x) or *raS (where y and r merge) in ethnonyms and terms for
'person', and the same root occurs as PS 'name', but I can only manage to
associate them by heavy use of modals and a lot of handwaving over
non-corresponding vowels in initial prefixes of the form wV.

The only other possibility that occurs to me is that perhaps the
Potawatomi speak for the twees.  (Thanks, and a nod to the late Ted
Geissel.)