OP u- and udhu- Verbs

[Koontz John E]

Commentary, u-verbs:

On Sat, 13 Nov 2004, Koontz John E wrote:

> u'agittaN 'I put on my own (shoes)' 90:43.9
> u'dhattaN=z^i 'you have not put them on (shoes)' 90:45.6
> u'na=i 'they sought them' 90:419.18
> aN'guna=i 'we hunted them' 90:434.2

Except for the anomalous first person example, which I omitted here, these
plural third person object forms all have initial accent and presumably an
initial long vowel for the locative or inclusive pronoun reflecting
contraction of wa- with the initial followed by loss of w before the
rounded vowel of u- IN or aNg- A12.  The latter counts as rounded because
it is from *uNk- (cf. Dakotan).  If we ever conclude that there are three
nasal vowels in OP (or Dhegiha in general), then perhaps the one in aNg-
is back/rounded anyway.

Of course, the rules given here reflect the history of the forms.  I think
that a description of contemporary OP has to think in terms of initial
accent/length as a reflex of wa- 'them' and wa- INDEF with
u-stems.

The same thing happens wrt nominalizations in wa- of u-verbs, whether we
treat this as a third wa- DETRANS or a variant on wa- INDEF.

> Sample u'-nouns, presumably u'- < *wo'-:
>
> u'z^iha 'sack' 90:17.10

> Inflection of udhu-verbs:
>
> udhu'ahe 'I followed her' 90:199.18
> udhu'dhahe 'you follow her' 90:194.6
> udhu'ha=bi=ama 'he followed them (elk)' 90:72.7
> udhu'ha=bi=ama 'she followed it (a trail)' 90:290.7
> aNdhaN'guhe=tta=i=the 'we will follow it (the trail)' 90:438.17

The udhu- here is from *iro-, in other words the combination of the *i-
and *o- locatives.  These seem to have been formed, if the paradigm is
taken into account, by adding i followed by epenthetic dh and then
assimilating i across epenthetic dh to the following u or aN.  Except for
assimilating i to u across dh, this is basically the assimilative behavior
exhibited with the i- and u-locatives throughout Dhegiha.  The result of
this pattern of formation is that the inclusive aNg- A1 appears between
the i-locative and the u-locative, resulting in aN-dh-aNg-u- <
*i-r-uNk-o-.

My impression is that in Dakota the inclusives of these forms inflect by
treating corresponding iyo- as a chunk and inserting uN(k) Pro12 after it:
iyo-uN-.

I don't know how the inclusives work in IO or Winnebago.

In Dakotan and WInnebaog the first person follows all of the locatives as
in OP.

> wiu'akkie" 'I spoke to him (?) [(?) in orig.] about it' 91:120.13
> wi'udhakkie 'you talked to them about it' 90:484.3
> wiu'ha=i 'they followed them' 90:440.1
> wiaN'guha=i 'we followed them (trails)' 90:419.14

In the wa- third person plurals, the epenthetic dh is either not produced
or is "squeezed out" and wa-i- contracts to wi- or probably wii-.
Another way to look at is that the "pre-composed" contraction wii- is
added.  "Precomposition" is basically a way of talking about analogical
spread.

All of the u- and udhu- forms are essentially regular for accentuation if
we assume that accent is on the second mora and that contractions of wa-u-
or wa-i- yield two mora (long) results.  The only exceptions to regularity
concern the VVV sequences that arise in the third person plurals of the
udhu-forms.  Either the way in which accent is perceived in wiiV- is a bit
variable, or there are some actual variations.  I don't know which, but my
suspicion is that all onset diphthongs - uV, iV, eV - are treated as onset
+ VV in less careful speech and that even if the onset arises from a long
vowel, it can lose accent to the following vowel as that becomes the
sonority peak of the syllable.  As a result VVV sequences tend to collapse
to VV sequences.  This probably accounts for accent shifting in other VV
sequences, e.g., muu'=ase > mw=a'ase, or ee'(=)(?)aN > eaNaN', and so on.
I think I'm correct in saying that e in ea and eaN sequences is not
[ey(a)], but a sort of open e or aesc that proceeds transitionlessly into
the following vowel, and is less prominent (maybe shorter?) than the
following vowel.  It has been a while since I heard the examples, however.

A parallel development occurs in Winnebago, where hi-a- (*i-locative +
first person agent) > yaa-, and ho-a- (*o-locative + fist person agent) >
waa-.  I assume this is an areal tendency, not an inherited one!