PS *waNtho' 'grizzly'

[Koontz John E]

The CSD, following Matthews 1958, reconstructs a form for PS grizzly bear.
Matthews suggests *matho', but the CSD, following the trend initiated in
Hollow's 1970 Mandan dictionary, of seeing nasal stops as nasalized
resonants, offers *waNtho'.

The forms collated under this reconstruction, with some slight
alterations in notation, are

Ma  maNto?    (underlying waNto?) (Carter)

Te  maNtho'

OP  maNc^hu'  (with diminutive c^h for th)
Ks  miNc^ho'
Os  miNcho'   (with c = ts, irregular before o)
Qu  maNtho'

IO  maNtho'

Wi  maNc^o'   (Pre Wi maNtho')

Of  *uN.'thi

Tu  (ha)nuB'thih (Mithun) (cf. *muNt[h]i WM)

This is one of the rather rare (post)aspirate t sets, but it is irregular
in having the *th as th, not h in Dakotan.  The Of and Tu forms are also
listed in 'black bear' *huNt-, withthe suggestion that there may be some
confusion of the two stems there.

The vowel iN in Ks and Os in Dhegiha is also irregular for *aN, but the
same thing occurs in the 'bow' set, which is a loan from Algonquian.

La  (i)ta(zipa) (taking earlier *maNta(zipa) as a first person)

OP  maN'de
Ks  miN'j^e
Os  miNce
Qu  maN'tte

IO  maN'hdu     (Pre IO *maNhtu < *maNtku with regular metathesis)

Wi  maNaNc^gu   (Pre Wi *maN'tku)

Cf. Proto-Algonquian me?tekw-a 'wood' + ANIMATE

I recently occasion to notice again Proto-Algonquian *ma0kw-a 'bear' +
ANIMATE, in which 0 = theta, initially an abstract symbol for unknown
cluster initial leading to preaspirates some palces and fricatives
others.  I think it is sometimes suggested it might be a voiceless l.
Under this are collated (per Aubin)

Fox      mahkwa
Cree     maskwa
Odawa    mko, mkwa (these from Rhodes)
Shawnee  mkwa

etc., including Arapohoe and Eastern Algonquian reflexes.

My main source Aubin also lists Proto-Algonquian maxkw-a, which I take it
means that there are some awkward reflexes around for 0k, though these are
not indicated.

Aubin also offers *naapee0kw-a 'male bear', in which I think *naapee- is
'male', and a medial *a0kw-a 'bear', underlying these two forms.  I
believe that from the Algonquian perspective m- is essentially an initial
for making medials into forms capable of standing alone./

I have taken the liberty here of segmenting -a, which I hope is correct.

What I've noticed is a dgree of parallism with the 'bow' term.  Thjis is
in the nature of exploring a hypothesis.  I am not yet ready to insist
that PA *m-a0kw- is the source of PS *waNtho'.  It's just that there are
some similarities, and both forms appear to be somewhat vexed.

- In both cases we have Siouan forms lacking an element that might match
the animate marker -a.

- In both cases we have PA mV lining up with maN most places in Siouam,
but miN in Ks and Os.

After that it's a bit more variable.

- PA ...?tkw appears as Siouan ...t-e or ...tku (in Winnebago-Chiwere), in
which -e is a Siouan noun forming suffix, but

- PA ...0kw lines up with  ...tho in Siouan.

It's not clear if there's any connection, but if there is,

- Why should ?tk should match t ~ tk, but 0k match th.

- Why should we  keep -u in one case but -o in the other.

One possible reason why 0k became th would be the following vowel -o.
While this was reduced to w we would have a cluster kw which seems to be
unstable in Siouan.  In 'cat' and 'squash' it gets this treatment.

Da  *km     (kw before a nasal vowel)
IO   dw     (earlier tw)
Wi   c^(V)w (earlier tw)
Dh  *kdh    (essentially kr)

Of course it's not clear why *tw would be preferred across the board
in bear.  Later reduction to -to is not too different from -ku with 'bow'.