Siouan root constraints

[R. Rankin]

I think we'd probably all agree that a root is always a
single morpheme and is the semantically central
morpheme in the word.  A stem may contain more than one
root (as in compounds) and is "everything you have in
the word up to the point where you want to add
something new."  There are a few instances in which a
single root cannot be isolated synchronically in
Siouan.  These include 2-part verbs like /ma-ni/ 'walk'
or, in Dhegiha, /koN-ra/ 'want'.  Historically, I can
identify the roots here, but synchronically, no.  And
in some, both parts of the verb are inflected and you
just have to say they are a single verb.

As for Ch and C?, there are definite trends but no
absolute rules.  By and large both types of complex
stops are restricted to roots.  (Grammaticalization and
a couple of phonological processes belie this as
anything absolute.)  Ch is, for the vast majority of
cases, historically linked to accent.  Therefore you
expect to find these consonant types as the onset to
accented syllables, and this is indeed true.  Lack of
accent accounts for why you so seldom find them in
enclitics and affixes.  From this, you would expect
that accented syllables with a plain C onset would not
exist, and, indeed, they are pretty rare.  But they do
exist, and I don't think we are sure why that is so.
John listed several important examples.  The historical
relationship between accent and aspiration is no longer
in operation except in Ofo.

As I recall, Mandan, Hidatsa and Crow all have the
reflex of glottalization in syllable codas.  In Mandan
and, I think, Hidatsa the corresponding form to Dakotan
C?V is CV?.  In Crow the same thing happens, but
apparently Crow has then vocalized the ? as vowel
length.

Bob

>   I was wondering if anybody has ever taken a look at
> the root
> constraints on stops in any of the Siouan languages.
> What I want to
> know is: are there very many roots with two voiceless
> unaspirated
> stops?  The nomenclature for this type of root might
> be something like
> TVT(V).  It seems to me that a lot of roots have one
> aspirated and one
> unaspirated stop.  Is this accurate?  Also, would it
> be fair to say
> that ejectives are relatively rare in the Siouan
> languages that have
> them?