i- in Dhegiha i-POSITIONAL=...CAUSE (RE: Word for 'prairie'?)

[Koontz John E]

> > khe  ihe=...dhe
> > (The i here is probably not a locative, however.)
>
> I'm surprised.  It certainly "feels" like one to me.
> Is there evidence from other languages against the
> i- dative interpretation?

There are two sorts of evidence against i as a locative - instrumenal,
dative or otherwise - though I admit that I initially took this as a
locative myself, and we are lacking evidence of the most desirable sort -
a case of, say, A12 or P1 aN coming up against the i and not doing the
standard aNdhaN thing that characterizes the locative i.

The first kind of evidence is internal (and applies throughout Dhegiha, as
far as I know).  It happens that idhaN, ithe, and ihe occur frequently in
causatives (form=...dhe) in the sense 'put and object of such and such a
shape down' or 'put an object positioned in such and such a
configuration'.  However, they also occur alone and with causatives in the
sense 'be positioned; begin; do suddenly; do suddenly and repeatedly', and
in all these situations they alternate with thi-forms like thidhaN,
thithe, or thihe.  And we also see some cases with hi initial in such
sequences.  So it appears that i alternates with thi and hi, which are
clearly motion verbs ('to arrive here' and 'to arrive there'). Thus, it
seems likely that it is a motion verb itself, and, of course, there is i
'to come'.  However, dhe 'to go' never occurs in this context, so there
are some oddities to the slot.

The second kind of evidence is comparative.  In Dakotan equivalents of the
i forms you generaly see uN, e.g., uNgnahela 'suddenly', where I think the
initial uNgna matches igdhaN.  I am not positive of this, and the OP i
matches i across Dhegiha and in IO, all places where u might be expected,
if i is from *u.

In general, however, I think the first set of evidence is fairly
conclusive.