Water monsters

[Koontz John E]

On Fri, 25 Jan 2002 Rgraczyk at aol.com wrote:
> Actually, I think that you could make a case that Crow buluksa'a is
> cognate--or partially cognate--with Dakota unktehi.  In Crow t can become s
> before a low vowel, and sometimes VhV > VV.

But I think bulu isn't a very good match for uN, u after denasalization.
And what usually happens to kt?  Isn't the k reduced to preaspiration? On
the other hand, wablus^ka has an extra wa over anything exhibited in
bulukta'a, and I'm not sure that uNktehi is terribly canonical in form for
Dakotan, either, though it falls within the general range of the trickster
term variants.

Of course, I guess we should really be wondering what happens in some of
the connected languages, too.  So far I haven't thought of anything that
matches uNktehi in MV, and I don't know anything about Hidatsa or Mandan.
You'd expect maybe *oNtte((h)i) in Dhegiha.  The -hi might be analyzed off
as 'leg' in OP, where *hu becomes hi.

Looking around, I've found in LaFlesche for Osage a reference to a creek
called Wagthushka iabi /walus^ka iapi/ 'where a strange animal was seen' -
warriors were crossing a stream on a log and when all but two had crossed,
it turned its head downstream and went away.  Notice that when citing
forms from a context LaFlesche uses the proper plural form.  I'm not sure
the form is to be translated 'where a walus^ka was seen'.  It looks like
iapi is 'he speaks', but I could be wrong.  Anyway, it looks like Osage
bases its term on the word for 'bug, lizzard, worm', in Osage walus^ka (<
*wakrus^ka).  It seems to me that this was pointed out already.

For Ioway-Otoe I find isc^exi, is^c^eximi(N) (female) for 'water spirit,
horned underwater panther (a malevolent spirit monster) in Good Tracks.
This looks like it would be something like *is^texi, which doesn't seem
helpful in this context.

All I could find for Winnebago was nuNuNjake' 'water snake'.  This would
be from something like *ruNuNtaka, perhaps (irregularly, I think)
*ruNuNtka, which is actually not a bad match.  I've assumed *ka > ke,
which is regular, but it might also be *ruNuNtke, which is a fairly good
match for uNkt^ehi.  One language or the other has metathesized the
cluster, and the initial *r is lost in Dakota (it would be y, as in
*yuNktehi).  That r would actually help explain the Crow bulu-.  Crow
would have added an additional *wa-.  However, I feel like I'm sort of
herding the form in the direction I want it to go.  I'd feel better
looking a bit further at snake and insect terms closer at hand.

The IO bug term is wagri'.  Winnebago has wikiri'.  Together these suggest
*wakri', maybe *wikri', which is in the ballpark with Dhegiha *wakrus^ka
and Dakota *waprus^ka, though none of these match regularly.  The -ka are
probably noun formants and the prefixal wa- could be either the indefinite
object wa- or one of the old fossilized noun classifying prefixes that
Rankin has discovered across Siouan-Caddoan and Yuchi.

JEK