Order of verb affixes

Koontz John E John.Koontz at colorado.edu
Wed Nov 10 18:15:25 UTC 2004

On Tue, 9 Nov 2004, Rory M Larson wrote:
> [IO] ... for '(black) pepper', which is glossed as "Mixed Herbs", or
>      mankan iki'rohan (LR)/
>      mankan' wi'kiruhan (JY)
> In OP, the word for 'pepper' is
>      weo'kkihaN
> which seems to be formed from
>      wa-i-u-kki-haN

> I don't know much about IOM, but it looks to me like the words
> cited for 'pepper' are essentially the same except for the
> preceding 'medicine'/'herb', but in a somewhat different order
>     (wa)-i-kki-u-haN
> with the reciprocal pronoun kki brought out in front of the u-.
> Am I correct in this analysis?  If so, is it normal for IOM to
> put affixed pronouns in front of the u-/o- in such verbs?  And
> if it is, then how do Hochunk and Dakotan compare on this?

I think you are correct, except that from the presence of iro- ~ -iru-,
which is the compound of the i and o locatives, I think that the IO
structure may be slightly more complex:


The inflection and further derivation of compound locatives is messy in OP
and I haven't a clue how it might work in IO or Winnebago.  This is an
area that hasn't been much looked into, I think.  The same is true of the
inflection of reflexives in those languages, though hints in the available
dictionaries and sketch grammars suggest it may be complex.

So, the repetition of i on both sides of khi may be part of the handling
of khi with iro, and it is also quite possible - we'd have to check for
hints in Whitman - that pronominals before khi all behave as if there were
an inserted -i-.  I actually think that might be the case, but it wouldn't
account for the form ik(h)i'rohaN where nothing precedes the khi.
Incidentally, I'm writing kh for aspirates and g for nonaspirates, just to
emphasize the distinction, following a practice Bob Rankin suggested for
the CSD, but JGT writes k and g, which would obviously be the preference
in any practical orthography.

As far as the OP forms, and looking ahead to Ardis's much appreciated
comments, it is my impression that Dorsey writes we- for underlying wa-i-
and wa-gi- sequences, but wiu- for underlying wa-i-u- sequences.  I've
always assumed there was some real phonetic basis for this, at least in
the 1890s.  I don't recall at the moment if I ever encountered a wiu-
surface sequence personally.  I did encounter we- when the i-locative was
present by itself.

As far as the inflection of i-u- compounds, the OP surface form for the
historically underlying *iro- (Da iyo-, Wi hiro-, IO iro- ~ iru-) is
udhu-, e.g., *udhukkihaN 'to cook for oneself by means of something'.  An
example that definitely occurs in this form would be udhuhe 'to follow by
means of something (like tracks)' vs. uhe 'to follow (a path, etc.)'.
When you add wa- (any kind) to udhu- you get wiu-, in Dorsey and Hahn
anyway, i.e., the epenthetic -dh- disappears and the assimilation of i to
u across it is undone.  This is a fairly drammatic pattern of allomorphy,
but it nicely demonstrates that udhu- is from *iro-.  I think this
alternation is a pan-Dhegiha pattern, i.e., an isogloss that would mark a
Siouan language as one of the Dhegiha group.

What I find interesting here is that it appears that the compound
locatives *iro- and *ira- were already part of Proto-Mississippi Valley
Siouan morphology.  I don't know if the semantics of these forms have ever
been systematically investigated in any of the languages, let alone
comparatively.  I have a rough notion of OP udhu- as 'through, along', but
in many specific cases the 'by means of' sense of the outer i- is more or
less evident, and I suspect that the 'through, along' idea is a spurious
consequence of looking at common English equivalents of udhu-forms, as
opposed to looking at the underlying morphology and how the forms are used
in clauses.

In line with Bob's remarks on whether *o- in 'cook' is a locative, I seem
to recall that in some Dakotan o-verbs there are morphological anomalies
that suggest o is not the locative.  I can't remember if the relevant
example was 'cook' or 'help'.

More information about the Siouan mailing list