Funny W

[Koontz John E]

On Mon, 13 Nov 2006, Rory M Larson wrote:
> > I'm thinking more along the lines of
> >
> >    Nasality  Tier       N
> >                         |
> >    Segmental Tier     # C            V
> >                         [+ resonant]
> >                         R            V
> >                         n            u
>
> John, could you give a legend for this?  Obviously, we're using a somewhat
> different tier structure here, and I'm not sure I can follow your argument
> at this point without more of an explanation for the symbols you're using.

It's just a stab at representing the notation of what used to be
contemporary phonology, I'm afraid.  It may not be current any more.

The | is a link between tiers - just a line, really.

The [ =/- feature] notation indicates presence (+) or absence (-) of a
phonological feature.  At one point all segments (more or less phonemes,
but not arrived at by contrastive analysis) were resolved into sets of
features, and there were knock-down, drag-out arguments over which ones to
use and in how many languages.  In this context I am just using it as a
way of abstracting some common "feature" from several sounds, e.g., [+
nasal] characterizes nasal stops (m, n, etc.) and nasal vowels.

Goldsmith's autosegmental arguments in the mid to late 1970s prompted
interest in the idea of what he called autosegments.  I believe he meant
"sounds autonymous from - or independent of - regular segmental
phonology."  Goldsmith was interested in treating things like pitch
accent, vowel harmony, and nasal spreading in terms of these autosegments
- segments that spread across regular segments or occupied the same space
as them but on a different tier or plain.  He talked about non-linear
phonology, meaning it seems to me "multi-linear" or "multi-tiered"
phonology, not anything like what a mathematician would mean by
non-linear.  (Goldsmith seemed to have a genious for inappropriate
and confusing coinages.)

This was right about the time I took my last phonology course and
privately abandoned any notion of looking for work that required me to
keep up with this paticular kind of competitive dogma development.  It
seemed to me glancing up from Siouan at intervals that things rapidly went
from the idea of using tiers inhabited by special autosegments that
accounted for special "spreading" or "zonal" phenomena to treating all
phonological systems as comprised of a theoretically defined set of tiers.

I think people liked this approach because it made it easier to talk about
stress accentuation in terms of feet, and also because it provided a great
way of reducing things like coda and onset simplications to problems in
symbolic logic.  Modern linguists have always had a serious weakness for
the argument from the logic of the notation.  "The notation makes it
really easy to express this, so the fact that it happens is a natural
consequence of the fabric of the universe.  QED."  Maybe the Latin of this
would be "argument ex notatio"?

I've always preferred a simpler, conceivably heretical approach, which is
to use the notation that makes it easy to describe what is happening and
let someone else worry whether this meant it was a law of nature.

In more formal terms (and I wasn't bent on achieving that above) there are
certain set tiers in all segmental systems.  A very tongue-in-cheek
version of them would be:

- syllables (little rows of sigmas linked to stuff above them)
- syllables have onsets, cores (resonance peaks - usually vowels), and
codas ('tails' - colas in the Teton dialect of Romance)
- C V skeletons (which link up to the onsets, cores, and codas)
- hand-waving tiers (this is where your spreading nasality goes)
- the tier in which you place the actual segmental notation
- the tier in which you write acute accents and breves if you are
doing stress accent or H and L and maybe M if you are doing pitch accent

Any professionals who are standing back gasping in horror or at least
peeling their eyebrows out of their hairline should feel free to jump in
and clarify this at any point.  If you feel it should be done offline, and
perhaps much of it should, please include both Rory and myself.  I for one
would be grateful.  A reference or two and a basic synopsis wouldn't hurt
on the list.

===

So, to get back to the actual point I was trying to make:

Actually, what's missing here is a non-N (an O?) following the N and
linking to the V.  In essence, I'm suggesting that all words get an
automatic initial added N that only comes home to roost on r and w.  If
the word has an N attached to the following vowel "organically" then the
effect is that r and w become n and m.  If the first vowel is oral then
the orality and nasality engage in a struggle for the soul of the r or w
and settle on a compromise that sounds like nd or mb.  Not "n-dah" and
"m-bah" to parody the syllabic nasals of Swahili, but "raised n"-d and
"raised m"-b - a much more subtle affect akin to speaking with a cold. The
initial b of "bandaid" reduced to "mbadaid."

In intervocalic positions there is no stray N of initial nasalization, and
you get just plain r or w.

I don't insist on either the details or the actual argument.  I'm just
looking for ways to explain what we see.  I assume Proto-(MV)Siouan was a
real language with a plausible, real phonology.  I don't want to mistake
PMV for one of the current languages or their phonologies, even if I
appeal to those for cognate sets or as instances of plausible phonologies
and plausible rules.