Dorsey's Law

[Koontz John E]

On Tue, 2 Apr 2002, ROOD DAVID S wrote:
> A very quick answer which someone else will probably elaborate on:
> Winnebago speakers do treat the Dorsey's law sequences as single syllables
> for purposes of assigning stress.  Stress is normally on the first or
> second syllable of non-Dorsey's law words, but on the second or third for
> Dorsey's law forms.  I'm sure Ken Miner has written this up in detail
> somewhere.

There are some problems with this statement.  I think David meant to
say something like "Before the application of the Winnebago Accent
Law (shift accent one syllable to the right) but after the application of
Dorsey's Law (insert a copy of V between T and R in TRV), stress was
normally on the first or second syllable of non-DL words (words without
TRV sequences), but on the second or third for DL (TRV-including) forms."

In other words, you found words like *CV(V)'CV and *CVCV'CV, and when CV
was TRV which got expanded to TVRV, you also found words like *TVRV'CV or
*TVRVCV'CV.  Subsequently these became CV(V)CV', CVCVCV', TVRVCV', and
TVRVCVCV'.

Of course, words like *CV(V)', *CV(V)'C, and *TVRV' were unaffected.  As
were cases like *CVCV' and *TVRVCV'.  Accent can't be placed on or moved
onto a non-existant syllable.

There are some words which have two DL, followed by other syllables, in
which accent originally fell on the second DL, and which therefore are now
accented on the fifth syllable:  **TRVTRV'CV => TVRVTVRVCV'.

There are also words of the form *CV'TVRV and *CVTVRV'CV that come out
CVTV'RV and CVTVRVCV'.

Accent can only end up on RV of TVRV, if it started there before the
Winnebago Accent Law.  Otherwise it is only on TV' or TVRV.

For me the important generalizations are therefore that accent before the
effects of DL and the WAL was on the first or second syllable, and only
the separate or combined effects of DL and then the WAL (plus the
existence of following syllables) can push accent beyond these points.
The WAL always pushes accent one syllable if it can, but only the
existence of one or two DL applications can boost it beyond the third
syllable.

Turning from this to the issue of vowel length, it appears to me that all
Winnebago long vowels and/or diphthongs correspond either to (a)
monosyllables, (b) loanwords (like kearipuNuNiya' 'California'), (c)
bimorphemic syllables, or (d) old initial accents (cases like *CV(V)'CV
that appears as CVCVCV'.  Thus Siouanists are left with a situation in
which Pre-Winnebago provides no more evidence than the existing Siouan
languages to help us determine whether initial accent and initial long
syllables are independent or interdependent and how.  This does not mean,
of course, Winnebago proves there were no long initials, but only that it
doesn't do any more toward proving that there were such long syllables
than, say, properly heard Omaha-Ponca or Ioway-Otoe.  It's pretty obvious
that many initial accents fall on long vowels, and Winnebago certainly
shows that this combination of features can be resolved into separate
features.