Incorporation

[Robert L. Rankin]

You probably should look at any Siouan verb in which the person/number
marking appears in the middle of the root.  This pattern almost invariably
shows that historically there were two roots, i.e., a compound root.  mani
'walk' is one such. The conjugation is ma-wa-ni, ma-ya-ni (and, I think,
u~-mani in the inclusive person, but I'm not sure).  In this instance,
several of us thought that ma- was perhaps the root meaning 'earth,
ground'.  More recently it was discovered that there was a Catawba root of
approximately that shape meaning simply 'go', so the etymology is unclear,
but what is clear is that these "split" roots (stems, actually)  have
ancient incorporated elements.

Take for example 'cough', a verb I've been examining for possible stative
conjugation.  In Kansa at least it is split:  ho-a-xpe, ho-ya-xpe in 1 and
2 sg.  Obviously ho- is in some sense 'incorporated'.  One of the meanings
of ho- is 'voice', and I'd guess that's what we see in hoxpe 'cough'.  

I'll have to check for the 1st dual form because it often is found to the
left of everything else when the 1st and 2nd sg. are infixed.  I'd sort of
expect to find o~-hoxpe 'we-2 cough', but I'll have to search and see if I
elicited the form or if Dorsey has it.  This pattern typical throughout
much of Siouan, so 1ST DUAL/PL VERB FORMS MAY BE CRUCIAL EVIDENCE FOR
INCORPORATION VS. MERE COMPOUNDING in Siouan verbs.

Now, note that in Quapaw, all the pronominals have migrated to the left
edge of the verb:  a-hoxpe, da-hoxpe in 1&2 sg. So "incorporation" is more
convincing here.

Lastly, let me make clear that I am not talking about verbs with known
prefixes like locative a-, o-, i- or instrumentals -- that may be
incorporation by some definitions, but it is definitely different from
incorporation of full lexemes.  You should concentrate on "infixing" verbs
where the identity of the first element is unknown or is clearly a noun or
other a member of some other major lexical class. 

Bob