Fricatives (was Re: Hda / Sna)

[Koontz John E]

On Wed, 22 Oct 2003, Louis Garcia wrote:
> The name used here for Jingle dress is xdaxda sugsugnica.
> I hope I spelled sugnica correctly. The 'x' is an s with a dot over it.

Usage varies, historically, especially for Dakotan and Omaha and
Winnebago - a function of the number of sources available.  Under the one
source - one orthography principle, the better documented a language is
the more orthographies it has.

There are three fricative series s, s^, x like English s, sh, and German
ch, approximately.  These also occur voiced (or less strident) z, z^, gh,
like English s, English ge in garage (for some) or z in azure, essentially
like French j.  I can't think of any convenient widely known examples for
gh!  Like gh in ghage 'to cry' in OP, I guess.

The s/z set is pretty much always written s/z, except in languages that
change it to th/dh (like Ioway-Otoe), where theta and edh become more
common.  This is also what underlies LaFlesche's use of c-cedilla for both
s and z in Omaha (and Osage).  Apparently he spoke a variant of Omaha with
th/dh for s/z.  This is mentioned in once place in Dorsey's notes, as
characteristic of Frank LaFlesche, and examples from Fletcher show it was
fairly general in Bikkude (a/k/a Village of Make-Believe Whitemen).
Dorsey used c-cedilla for theta and LaFlesche seems to have learned this
from him.

For sh/zh you see those diagraphs (not always convenient), or s^/z^
(representing s and z with a little "vee" (hacek) over them), the usual
recent Americanist linguistic convention, or s-dot/z-dot, or
s-accent/z-accent, or c or z for sh and j for zh or some combination of
the last few.

For x/gh (gh being a way to write gamma when you haven't got a gamma) you
might see x and gamma (Greek g), the usual recent Americanist linguistic
convention.  The Colorado Lakota Project uses h^/g^ (same explanation of
^).  Various sources use h-dot/g-dot.  I haven't seen accents with h and
g.  Dorsey used q and x (q for x, x for gh!).  LaFlesche crunches them
together as x.  Dakotanists are perfectly happy with g for gh, since
contexts where gh occurs are contexts where g can't occur.  You might find
r for gh (based on what French r sounds like).  Dorsey used this at one
point.

Of course, particular sources have some particular way of doing things,
i.e., Riggs is internally consistent, Buechel is, Dorsey is, LaFlesche is,
and so on.  That is, these folks are consistent in a particular source.
They often differ between publications, or between the draft in the
archives and the published version, and so on.  I haven't tried to list
these particular schemes, however, but only to give a general idea of the
variation you will encounter.

I believe Linda used s, s^, x for Riggs's s, s-accent, h-dot and Buechel's
s, s-dot, and h-dot.  I'm not absolutely positive I remember these two
gentlemens' practice!  At this point I sort of see what it is when I open
one of their books and go on without much of a pause.  A Siouanist who
isn't flexible in this respect is in a bad way ...

As far as the precise senses of the sound symbolism sets for sounds, it
wouldn't surprise me to see huge differences from one place to another,
maybe from one family to another, maybe even from one idiom to another.
One man's rattle is another man's buzz or ringing.  Actually, in the high
pitches a whistle in my left ear is a rattle in the right, I've
discovered.  Apparently something is broken in there.  Particular names
will probably always have the same sound value and same gloss, but might
be wildly out of step with another person or place's preferred gloss of
that sound value.

This problem is even worse between languages.  The CSD folks didn't want
to allow themselves to freely compare any fricative to any other fricative
- what we might call the Greenbergian approach - and had many anguished
debates over this issue.  They eventually developed some practical
protocols - informed by the patterns of Siouan languages - under which
they would look at sets with essentially a reconstruction of *S (a cover
symbol for *s/*s^/*x) as opposed to *s, *s^ or *x.  In such sets *S
represents a set in which sound symbolism has been at work leaving
non-corresponding fricative grades behind, e.g., forms that suggest *sra
in one language, *s^ra and *xra not being in use; and forms that suggest
*s^ra in another, *sra and *xra not being in use, and so on, but with
clearly related glosses.  These protocols interact with the requirement
that a form be attested in at least 2 branches of Siouan as opposed to
just Mississippi Valley.  Actually, requiring attestation outside MV tends
to clean up a lot of the worst cases of this sort of thing, I think.