Lakota wa- 'variety object'

[R. Rankin]

See below . . .

> You brought up something that I haven't thought
of yet
> -- explaining the structures I have via the
> detransitivizing function of *wa-*. But let's
see how
> this works in detail in my examples. In
>
> (1) thi-w-í-wa-'uN
>         house-things.PAT-paint-1SG.AG-paint
>
> the starting point is the transitive verb
*i'u`N* 'to
> paint', which has valence slots for AG and PAT.
*wa-*
> as a detransitivizer functions to eliminate the
PAT.
> So *i'u`N* plus *wa-* yields 'to paint' minus
PAT,
> i.e. detransitivized 'to paint' (or, with
incorporated
> *thi-*, detransitivized 'house-paint'). But in
(1)
> there are two *wa-*affixes. Where is the
additional
> PAT slot that is eliminated by the second *wa-*?
In
> your equation, we'd get detransitivized 'to
paint'
> minus PAT, i.e. a verb with a valence of [-1]
for PAT.
>
> But we can take it even further. My speaker also
gave
> me
>
> (2) w-í-wa-w-iyuNg^a-pi
> WA-LOC-WA-WA-ask-PL
>         'they ask around about him'
>
> and this structure contains three *wa-*s. The
base
> verb *iyúNg^a* 'to ask' is transitive, so that,
after
> three detransitivizations or PAT-eliminations,
we'd
> get a valence of [-2] for PAT. So I conclude
that
> rather than taking away valence slots, *wa-*
functions
> to add slots, at least in some cases.

OK, I think this is where your grammaticalization
theory comes in.  But, personally, my tendency is
just to look at it in terms of linguistic change.
Grammar change is primarily analogical in its
mechanics, which means that not all aspects of the
grammar are changed at once -- it operates
piecemeal and often lexeme-by-lexeme.  And it may
well be the case that you'll be forced to analyze
some instances of wa- as valence reduction and
some as adding slots.  Functionally, this would
mean the morpheme has split and you now have two
wa- prefixes where there was one before.  I
haven't tried to trace all the developments of the
(various??) WA- prefixes in Mississippi Valley
Siouan, although John's note immediately preceding
this one, does part of that job.  This is why I
say we have to be very careful about just how many
WA's we posit and have to pay close attention to
which occurs in what verb.  it sounds like a good
dissertation topic to me.  :-)

In any event, these are all very interesting
forms.  Several of the 'paint' forms have
interestingly strange Dhegiha analogs, and it will
be interesting to see how it all works out.  My
tendency is to TRY to analyze each instance of
/wa-/ in terms of valence reduction, following
Mary Haas's dictum "We mustn't translate the
(target) language into English and then analyze
the English".  I think paying attention to that
explains the house and the plates in the painting
sentences.  Beyond that, my Dakota fails me, and
I'm giving up for the evening.  :-)

Bob