More regarding "wa"

[Koontz John E]

These are excellent questions, actually, and issues like this have been
puzzling me for a long time, too.

On Wed, 17 Dec 2003, Tom Leonard wrote:
> In Ponca we have "ni'the" -to heal. "Wani'the" translates as "THE
> healer" (Jesus).  In this sense, it seems to me, "wa" acts like a
> "personifier" (correct term?).

Personifier, in other words, a marker that means 'a person who ...'?  I
suppose maybe the traditional term might be something like "agentive."
In this particular case, I believe that the standard analysis would be
that wa refers to unspecified people who are healed or to the unspecified
existence of people or things healed.  So the form is in effect a sentence
'he heals people' being used as a noun.  Sentences being used as nouns is
sort of the core concept in Siouan derivational morphology.  I think that
the wa- is needed here because without it the form would be 'he heals him'
which would nominalize more or less as 'his healer'.

> We have "xu'be" -holy or mysterious. "Waxu'be" translates as "sacred
> thing", "sacred bundle", etc. We find "ni ni waxu'be" for "sacred pipe".
> ...

But with waxu'be, wasa'be, and was^a'be its a bit more difficult to see
what's happening - for me, anyway.  Presumably with these one-argument
verbs wa makes that one argument nonspecific, yielding an underlying or
literal meaning of 'something that's holy', 'something that's black',
'something that's dark', but it's less clear to me what the unmodified
stem is not adequate.  Clearly it isn't adequate, but I definitely feel
that I grasp the mechanism less certainly.  It appears that patient
arguments - things that would take aN, dhi, etc., as pronouns, rather than
a, dha, etc. - require filling with wa or an incorporated noun to make a
nominal, whereas agent aguments do not.  I still have an uncomfortable
feeling, however, that there are exceptions to this rule, though I'm not
remembering one at the moment.

Notice that Dhegiha does allow wa with animate reference.  I was
momentarily taken aback by Regina's comment yesterday that Dakotan wa was
necessarily inanimate, because of that.  Somehow I had always assumed that
wa could have a non-specific animate reference, too.  Would a Dakotan
nominalization require wic^ha- or something like that if the inspecified
argument was animate?