Kaw and Osage 'stand'

[Koontz John E]

On Mon, 1 Nov 2004, Rankin, Robert L wrote:
> There are a lot of remaining questions.  A full fledged reanalysis of
> accent and length has to be done.  The answers won't be simple exactly
> because of verbs like naaNz^iN 'stand'.  Historically it is probably
> bimorphemic, or, at the very least, susceptible of being folk
> etymologized as such, since the two parts have discernable meanings: na-
> 'on foot' and z^iN 'erect'.  How are such compounds treated accentually?
> Or if they're thought of as monomorphemic, is the accent different?

I can answer part of this on a basis of past experience in Dorsey and in
person, though the answer is innocent of a current appreciation of length.
Anyway, in virtually every form that is either a transparent compound or
arguably so from its morphosyntax, where the initial element is
monosyllabic (length not necessarily clear) the accent is initial.  So,
forms that have the structure CV=xxxx, have the accent pattern CV'=xxxx,
where CV is a proverb (any element that consistently precedes the
pronominals, i.e., not a locative) and xxxx the root, or CV is a "main
element" and xxxx is proclitic.  I didn' recognize length in coming up
with this observation, but in most cases length seems not to be a factor.
Rather, it is the enclitic boundary that is.

There are a limited set of exceptions.

- Initial s^u with motion verbs is not accented, and contracts with a
following a-initial, e.g., s^u=bdhe' 'I am approaching you', but
s^=adha'=i 'he is approaching you'.  Presumably this is a short vowelled
form.

- The demonstratives are stressed before most enclitic "tight"
postpositions, e.g., e'=tta 'to(ward) it', e'=di '(up)to it', but not if
=thaN is added after a such a postposition, e.g., e=tta'=thaN, e=di'=thaN
'from it'.  I suppose one could argue that =thaN causes preceding e'=di
and e'=tta to be treated as bases in their own right, suppressing their
internal boundaries as it were.

- Some (but not all) of the monosyllabic animal terms that serve as
possessors or whole-denominators in body-part compounds, e.g., tta=he'
'ruminant-horn' seem not to be accented.  This is not consistent, and I
speculate that it may depend on the length of the stem, although the stems
may appear long in some environments, e.g., as monosyllables.

- In many cases where the second (proclitic base or enclitic) element is
vowel initial, the accent falls on it, e.g., e=(?)aN' 'how'.  This is not
always consistently marked in Dorsey.  My suspicion is that this is
related to the tendency of forms like mu(u)'=ase 'I cut' to be pronounced
mw-aa'se.  In other words, V'=V => [glide]=VV'.  The quality of e in
spoken instances of e=a' sequences is more or less lax e or epsilon and
rather different from e in other contexts.  Inconsistency of marking could
reflect careful pronunciations.

- Something similar, perhaps at an earlier stage of the language, may
account for i-(dh)a'- as the first person of i-locative verbs , but
i'-dha- as the second person.  The parenthetical dh of the first person is
present, but historically epenthetic.  The shift of the accent may be
explained by still earlier i-a'- where "a" is the first person, cf.
Dakotan wa-.  The second person is "dha," cf. Dakotan ya-.  This is where
Winnebago has y-aa- < *i-a- in the first person of (h)i-locatives, and
w-aa- < *o-a'- in the first person of (h)o-locatives.

- The verbs of 'saying' e=...(h)e' and 'thinking' e=...dhe' and 'doing'
e=...(?)aN' (but see above) treat e= as unaccented except in the datives:
e'=g(i)-e 'to say to', e'=gidhe (?) 'as expected', e'=g()i)aN 'like'.
However many uses of egaN as a conjunctive particle seem to stress it
depending on the foot structure of what precedes.

- Certain forms are presented by Dorsey as accented on a third syllable in
a second or third element in a clitic conplex, e.g., e(=)bdh=e'(=)gaN 'I
think'.  I assume this is an artifact of looking at pitch accent as stress
and trying to work out e=bdh(e)'=e=gaN.  Thus these forms may not be
exceptional in any real way.  Another example like this is Dorsey's waiiN'
'blanket'.  (I think Dorsey writing a'=i 'he said it' as ai' is along the
same lines.)