Obj.3p.an. and 'person' (Re: Lakota wa- 'variety object')

[Koontz John E]

On Thu, 11 Dec 2003, REGINA PUSTET wrote:
> It seems particularly significant that two of the three examples involve
> reduplication. However, all of these examples involve at least an
> implicit 'them', also marked with wa- in Dhegiha.
> ======
> That's interesting because in Lakota, wa- seems to have inanimate
> reference only. So OP wa- seems to have developed differently,
> semantically speaking, than Lakota wa-. (Which of course leaves the
> chicken-and-egg question open).

Bob mentioned this in passing in an earlier part of this thread.
Mississippi Valley is the only branch of Siouan that has special third
person animate object markers.  Within this, Dakotan has wic^ha- in this
function, apparently a cut down variant of the noun that appears as
wic^ha's^a or wic^ha'sta in the various dialects of Dakotan.  But Dhegiha,
Ioway-Otoe, and Winnebago all have a prefix wa- in this capacity.  This is
usually assumed to be a specialization of the detransitivizer wa-, but it
might also be a cut down version of a some noun - considering the number
of nouns and nominalizations that start with wa-.

Your chicken or egg comment is quite apt here in several senses.  We might
wonder which use of wa- comes first, for example, which I assume is your
implication.  I tend to suspect the detransitivizer, because it is more
widespread in Siouan, but that kind of a distributional argument is not
especially strong.  I don't know what the crosslinguistic evidence is for
directionality in indefinite <=> third plural development.  We do have
third plural => indefinite in English constructions like indefinite they -
e.g., "they never learn."  But I seem to recall reverse developments in
Chukchi paradigms, with passives or antipassives serving as third person
forms.

On the other hand, we could also wonder whether wic^ha- and wa- are
independent developments in this capacity, and, if not, which came first.
One might hypothesize that PMV had *wa- in this capacity and that the
incorporated noun(s) wic^ha'S(ta) replaced this *wa- in Dakotan.  In any
terms, wic^ha- is a sort of vicar (in the ecological sense) of wa- in
Dakotan.

Here it may be useful to know that the 'person' nouns in MV are not
especially similar looking and are certainly not regularly correspondent.
In Dhegiha, 'person' is generally like OP nikka(s^iNga).  This would match
something like Dakotan (?) nic^ha(s^iNc^a) < PS (?) *riNhka(s^iNka) or
maybe (?) lic^ha(c^hiNc^a) < PS (?) *Rihka(yiNka), or you can mix and
match the two parts.  Now (?) nic^ha or (?) lic^ha are somewhat
reminiscent of wic^ha-, but they are nowhere near right on the mark, so
that my inclination is to see nikka and wic^ha- as unrelated.

The closest thing to wic^ha's^a in Dhegiha is actually waz^a'z^e 'Osage'
(a clan name found in most of the Dhegiha groups).  Dakotan wic^ha's^a
suggests PS *wihkas^- or *wiyas^-, while Dhegiha waz^a'z^e suggests PS
*was^as^- or *wayas^.  The second consonant in both cases is ambiguous as
to antecedant.  We're still uncertain about voicing of fricatives, and the
a:e final vowel alternation is standard for "consonant-final" stems in
Dakotan vs. Dhegiha (and Ioway-Otoe).  So it is really only the first
syllable vowel that's perhaps different, if we assume *wiyas^- and
*wayas^- as the hypthetical sources.  That's a much closer match than we
get with Dakotan wic^ha- and Dhegiha nikka.

I tend to think that Dhegiha waz^a'z^e 'Osage' is related to *(i)yas^-
'name', cf. Da c^haz^e' and OP iz^a'z^e.  So Dhegiha 'Osage' and perhaps
Dakotan 'person' might originally been a trope on the order of
'name-bearers'.

It's true that Dakotan has a band name waz^a'z^a which is an even better
match with Dhegiha waz^a'z^e, but this is rather too good a match.  I
suspect this is actually a loan from Dhegiha - perhaps the name came along
with some originally Dhegiha-speaking waz^a'z^e clansfolk.  If waz^a'z^a
were inherited, it would support the PS *was^as^- hypothesis.  But
pretonic z^ is unusual, except in Dhegiha as a reflex of *y, and the
Winnebago form of 'Osage' is waras^, which, if inherited, is from PS
*waras^- or *wayas^-, which (weakly) supports the PS *wayas^- hypothesis.
So the easiest way to get a nice non-canonical match for Dhegiha waz^a'z^e
like Dakotan waz^a'z^a is to borrow it.

I don't mean to suggest that the ethnoym 'Osage' is inherited, of course,
but only that the Winnebago form may be a valid calque, an etymologically
correct match, for waz^a'ze, while waz^a'z^a is a phonologically slightly
adapted (-e > -a) borrowing.  Of course, the Winnebago calque could be
based on a false folk etymology of what waz^a'z^e means (cf. Wi raa's^
'name'), and so irrelevant.  That's a decision folks will have to make for
themselves. (Now there's an indefinite from a plural.)

JEK