Information / nouns vs. verbs

[Koontz John E]

> haNhepi is L. form for D./L. hanNyetu.
> While I cannot analize haNhepi itself, a-haNhepi (amahaNhepi) is "it is
> night on..",
> a-haNhepi-pi is "it is night on them", ~"they are on-nighted"
> (likewise, a-bloketu-pi, a-wetu-pi, a-chagxa-pi, etc.)

Ah, why is it plural, or nominal, as the case may be?  It looks like the
seasons are still locative.  What about aNpetu?

I think I can explain haNhepi, in a manner of speaking, apart from the
-pi.  I think the Proto-Siouan root is *haNh-, which becomes (underlying
root) haNhA in Dakotan terms, and occurs in the -e ablaut grade before
-tu.  It's interesting that it would still be e-grade before pi.  I guess
maybe it's not ablauting, really, in Dakotan.  The vowel is at least, I
think epenthetic (or Dakotnaists might prefer to see it as deleting).  I
don't recall incorporated examples with 'night' in Dakotan, but I think
I'm correct that aNpetu alternates with aNb-.  Dhegiha has haN, which,
like OP aNba 'day' and the seasons, doesn't require a locative.  I recall
that OP has haNhe- in one compounded form, I think haNhewac^hi 'night
dancer(s)'.  This could be a fossilized relict of the fuller form, or a
borrowing from Dakotan.

That a- construction is like be- in English benighted, in the more
traditional sense of 'caught en route at night' as opposed to the derived
sense 'foolish, unprepared'.  The ac^haghapi is analogous to bedewed, but
the productivity of the pattern is rather restricted in English, e.g.,
besprinkled, and some others, but not *be-iced or *be-springed.

JEK