Locative Postpositions (was Re: Mandan (again))

[Koontz John E]

> In fact, the postposition is -n/-l.  As it cannot occur as a free form,
> there is a tendency to cite it with some neutral demonstrative base like e
> 'that (aforesaid)'.  Add to this that V + e might develop as V and even
> linguists tend to lean on this crutch.  Still, I'm pretty sure the e is
> spurious in forms like til.

Of course, I should have said thil, since, it's thi(=pi) in Dakotan, the
stem being aspirated.

The constuction is thi=l, with = indicating the enclitic boundary.  The
corresponding Omaha-Ponca form would be tti=adi, in which =di after i
requires the intervening formant -a-, an archaism with parallels in
Dakotan, too.  In Omaha-Ponca the rule appears to have been, when adding
=di or =tta (or various other postpositions) to a stem that ends in -e,
change e to a first.  If the stem ends in some other vowel, append a
first.  Most of the surviving examples involve -i finals but other vowels
occur.  Many of the forms alternate with ones that lack the -a-, but
tti=adi and tti=atta are consistent.

I suspect that the -a- in all these cases is essentially an absolute
marker/nominalizer.  These usually derive historically from old articles
or demonstratives.

The following is reworded for clarity and complete sentences!

> I've always associated this truncated postposition -l with the full form
> -tu, though it probably also reflects at least the -c^a alternant of -ka
> ~ -c^a, too.  This -c^a alternant occurs after e, and would also reduce
> to -l in dependent contexts where there is truncation of the final
> vowel.

> The -tu source corresponds reasonably well with Omaha-Ponca -di,
> hypothetically from Proto-Dhegiha (and Proto-Mississippi Valley) *-tu.
> However, in other Dhegiha languages like Osage and Kansa, where
> unrounding of fronted *-tu" shouldn't occur, it does.  Either it just
> unrounds irregularly, perhaps because it is final and unstressed, or
> Proto-Dhegiha had really had *-ti, which wouldn't be cognate with
> Dakotan -tu after all.

> > Might this postposition be cognate with the Mandan directional
> > suffix -t (Kennard)? And is it related to the Dakotan & Mandan locative
> > -ta ? (There is no Eng. form of the type *Mandat.)
>
> It might be, but unfortunately there's another suffix to work with, cf.
> Dakota -(k)ta, Omaha -tta (where the -t-t- matches the -k-t-).  As I
> recall Dakotan has -ta, but develops an intrusive -k- in forms like
> e-k-ta.  In Dakotan truncations -ta might also alternate with -l.  In
> Omaha-Ponca terms -tta is 'to(ward)' and -di is 'in, to'.