Nominal Ablaut, Noun Theme Formants, and Demonstratives

[Koontz John E]

Kennard's view of -e on Mandan nouns is:

p. 26, "The indefinite article is the suffix -[eta] (i.e., a short e)."

p. 6, "Whenever a suffix beginning with a vowel is used with a stem ending
in a vowel, an r is inserted between the two vowels."

I don't see any sign of a -re verbal suffix.  I think Catawba has one.
And Biloxi has a -di on verbs and I think nouns that would match.

===

As far as e vs. a demonstratives, I think that the e form is clear and
reconstructable for Proto-Siouan, if evidently not Proto-Siouan-Catawban.
The a or ha form (?*Ha) is more obscure, but here's the data, including
the initials of 'day' to show the *H correspondence.  I'm not sure what *H
represents.  I don't mean to imply it's a new proto-phoneme.  It seems to
be an h that's there is some languages and not in others.  I think that
Chafe reconstructed *rh for this set ('day' that is), though it's not the
same *rh that Allan Taylor reconstructed for quite a different kind of set
(like 'arrive here').  The r comes from reflexes in Southeastern.  It's
not absolutely clear whether the *Ha demonstrative is the same set as
*HaNp- 'day', because it has more restricted distribution.  Some sort of
grammatical alternation may be at work in 'day', perhaps one of Bob's
obsolete classifier prefixes.

The *Ha demonstrative is primarily found active in Dhegiha, though perhaps
it also occurs in Ioway-Otoe and Winnebago.

Dakotan - I'm not really sure of anything for the *Ha demonstrative.  The
usual interrogative root is tV.  For the initial correspondence cf.
aNpe'=tu 'day', aNb=wa's^te 'pleasant day'.

Omaha-Ponca - a'=gu=di 'where', a'=thaN 'when, how far, how long',
a'=na(N)ska 'of what size, how big', a'=na(N) 'how many', a'=xt(i)=aN 'how
possibly' = 'how on earth' (?), e=a'=thaN ~ e=a'=c^haN 'how', a'=daN
'therefore'.  All these forms are indefinite as well as interrogative.
Some of them have parallels with e= as the demonstrative root, like
e=na(N) 'that many'.  Some do not and some seem to have both e and a,
interestingly enough.

For the initial correspondence in OP cf. aN'ba 'day'.

There is also a form a'wa- that seems to mean something like 'which of two
(or several?)' that can serve as a demonstrative base that might be
connected.  This has a variant or maybe just a resemblant alternative
wiaNwa that is better distributed in the rest of Dhegiha.  Dorsey claims
that a=the=di, a=khe=di, etc., exist as '(some)where' forms, but I've only
seen a'wa=the=di, etc., in the texts.

I'll skip Osage in favor of Kaw, since they're fairly similar.

Kaw - hago'j^idaN 'when', hago'ha 'where, whither', hakhaN' 'when, how
far, how long', ha'yoNska 'how big', ha'naN 'how many, how much', hago',
ha'go 'why', hago'daN 'why, how come'.  For the initial correspondence
haN'ba 'day'.

I've omitted boundaries here and in Quapaw, though they're probably quite
analogous with OP.   I'm not absolutely positive about an analysis like
ha=go=j^i=daN, especially given the Quapaw form ha'kidede (modified from
ha'kudidaN?), but I make the analysis of OP a'gudi in internal terms
'something' = gu 'yonder' = di LOCATIVE or something like 'where away'.

Quapaw - ha'kidede- 'where', haki'ttaN 'whence', hathaN'- 'when', ha'naska
'how big/small, what size', hana(N)' iNke 'some' ~ hana (h)itte (or hide)
'how many', hanaN' 'how much'.  For the intial correspondence haN'ba
'day', haNba' 'light'.

Ioway-Otoe - JGT lists ha, ha?e 'that', but the indefinite demonstrative
is ta(N) and I'm not at all sure of the status or usage of ha.  For the
initial correspondence, cf. haN'we 'day'.

Winnebago - The usual interrogative demonstrative is j^aa.  There are a
few possiblities for ha in the form of hac^aNga' 'which one', hac^iNiNj^a
(< hac^aNiNj^a, per Lipkind) 'where', haga' 'time, occurrence, intance'.
For the correspondence, cf. haN'aNp 'day'.  Hawa?uN' 'that's why' looks
like hawa + uN.  These could be fossils of *Ha and *Hawa, but maybe not.

By way of a demonstration that the initial of 'day' is not the usual h,
compare it with 'night', which is consistently haN(he).

So there's definitely a Proto-Siouan demonstrative e, and there's
definitely a Dhegiha, possibly a Proto-Mississippi Valley, "demonstrative"
(indefinite/interogative root) *Ha.  The real issue is whether they have
anything to do with the finals of nouns - or in some restricted ways - of
verbs.

JEK