Mixed Conjugations (Re: Irregular "to eat" in Dakotan)

[Koontz John E]

In regard to the case of Crow and Hidatsa having what looks like a
*V-initial 'eat' stem in the first person and an *r-initial stem in the
second and third persons, this is a kind of suppletion that occurs
elsewhere in Siouan.

Suppletion in the more general sense is also widely found, of course.  As
an example of that it would be hard to beat OP 'to say', which has
something like e=...he in the first and second persons, e in the third
person (usually a=i or a=bi=ama), and dhaN in the inclusive.  The first
two are probably related historically, with the second a simplification of
the first - either a contraction or just the e(=), but treated as
ablauting like =he.

Anyway, for the simpler type where the stems are essentially identical,
but treated as if they have different initial sounds, and thus different
conjugations, consider the Dakotan glottal stop stem pattern, e.g.,:

A1   m-uN
A2   n-uN (probably from *sh-nuN)
A3     uN

Compare this with the Dhegiha pattern:

A1   m-aN
A2  zh-aN
A3     aN

and with the corresponding forms in the *r-stem paradigm, cf.
Dakotan

A1  b-l...
A2    l... (probably from *sh-l...)
A3    y...

and Dhegiha (this is the OP version)

A1  b-dh...
A2 sh-n... (becoming just n since 1870 or so)
A3    dh...

This suggests that maybe the Dakotan glottal stop stem second person (n,
i.e., l before a nasal vowel) is from the *r-stems.

Another example occurs with the inflection of Dhegiha *i...aNghe 'to ask
someone (a question)'.  This varies among reflexes of idhaNghe and iwaNghe
for the third person, where the dh and w are plainly epenthetic.  The
first person is either imaNghe (like a glottal stop stem) or ibdhaNghe
(like an *r-stem).  The second person is usually inaNghe (earlier
ishnaNghe) (like an *r-stem)  (details vary for different Dhegiha
languages, depending on how they treat *sh-r).  Another verb like this is
the sitting definite article:

A1     m-iNkhe
A2  (sh)-niNkhe
A3       dhiNkhe

How *i...aNghe 'to ask someone (a question)' gets a mixed paradigm is
fairly clear, but the mechanism in Dakotan glottal stops stems or with
Dhegiha *dhiNkhe (< Proto-SIouan *riNkhe) isn't as clear.

Incidentally, though glottal stop does appear in some persons of some
languages' glottal stop stems, this kind of stem usually looks more like a
special kind of V-initial stem.

JEK