(In)dependent body parts in Dakotan?

[Koontz John E]

On Wed, 28 Apr 1999, Robert L. Rankin wrote:
> I noticed a couple of years ago that a number of Siouan languages split
> their inalienably possessed nouns in the 1st sing. into a group with a
> reflex of ma- and a group with a reflex of mi- (Crow ba-/bi-, Mandan
> ma-/mi-, etc.).  The interesting thing is that there does NOT seem to be
> any semantic congruity from language to language in the groups of nouns
> that take one over the other prefix.  And if there were a semantic split
> in the 1st person, why not in the 2nd person?

For what it's worth there is a small opposition in the first person
possessive in Dhegiha (Omaha-Ponca).  It doesn't involve vowels, but it is
an opposition in the first person, and it might shed some light on the
mechanics of first person oppositions.  Essentially, inalienable
possession in Omaha-Ponca is restricted to kinship terms.  There are some
terms that might be construed as kin terms that are alienably possessed,
e.g., nu 'man' in the sense 'husband'.

The only traces of possession with body parts, etc., are the initial i- in
iz^a'z^e 'name', and perhaps iNkhede ~ ikhede 'shoulder', and the frequent
occurrence of a fossil wa-, apparently marking originally 'unpossessed',
with body parts (wahu 'bone', wahe 'horn') and foodstuffs (wahaba 'ear of
corn', wattaNzi 'stalk/plant of corn').  One term, 'dwelling' has an
irregular paradigm:

Px1  tti' wiwi'tta
Px2  tti' dhidhi'tta
Px3  e' tti'=i            cf. PpaNkka tti 'Ponca lodge'
                            NiN'kkagahi tti 'chief's lodge'

This is essentially the alienable pattern, but with the third person in
verbal form 'he dwells (there)', etc.

With kinship terms the prefixes are normally wi Px1, dhi Px2, i Px3.
There is no inclusive form.  Instead a first person or second person,
generally the latter, is substituted.  I am not sure what factors,
presumably pragmatic, result in the first person being used instead.

Two terms, idha'di 'father', and ihaN' 'mother', are exceptional in having
iN Px1, to wit, iNda'di 'my father', and iN'naNhaN 'my mother'.  Note that
with these stems the first person (and vocative) use alternate stems
dadi (vs. 2nd and 3rd adi) and naNhaN (vs. 2nd and 3rd haN).  The
paradigms here are:

Voc   dadi'=ha=u (modern m spkg),
      dadi'=ha   (modern f spkg)(formerly either sex speaking)

Px1   iNda'di
Px2   dhia'di
Px3   idha'di   (dh is probably historically epenthetic)

Voc   iN'naNhaN=(ha)u (modern m speaking)
      iN'naNhaN=ha    (modern f speaking) (formerly either sex speaking)

Px1   iN'naNhaN
Px2   dhihaN'
Px3   ihaN'

The change from former (c. 1880s) to modern (1900s on) usage in the
vocatives occur throughout the kinship system, and a part of a
readjustment in the sex of speaker markers that was incipient at the time
of Dorsey's work, and is well established in the work of LaFlesche.  It
appears that Dorsey preferred older, conservative speakers, as LaFlsche
and he were, of course, contemporaries.

There is one further term that I believe includes a fossilized iN prefix,
which is iNs^?a'ge 'elder, old man', often used, of course, as a vocative,
though this form is also used in reference.  There is some tendency for
the first person possessive to appear with with vocatives in Omaha-Ponca.

The iN form is identical to the dative variant of the patient pronominal
aN + gi => iN 'to me' and might arise from it.  That is, perhaps
iN-possessives reflect benefactives of verablized noun stems:  'iNda'di
'my father' < 'he is a father to/for me'.  Notice that Dakotan mi(N)  is
amenable to the same analysis.  If this analysis is deemed plausible then
it might account for one source of variant possessive prefixes.