Responding to a bunch at once:

On Fri, 7 Sep 2001 01:44:24 -0400 Roger Mills <[log in to unmask]> writes:
> Jesse Bangs wrote:
> >Hiksilipsi contains three grades of phonemes: a radical grade, a
y-grade,
> >and a w-grade.  The y- and w- grade phonemes are formed by adding a y
[j]
> >or w [w] glide to the radical phonemes....
>
> Interesting.  I played around with a system like that once.  What
> about the vowels?  will they also have three grades?

No, the vowels are a little simpler, and the only suprasegmental
affecting them is tone.

> My version had only 2 vowels, high/low, central i.e. /@ a/, with
> allophones
> e.g.
>     /p@/ [p@],    /py@/ [pi],  /pw@/ [pu]
>     /pa/ [pa],  /pya/ [pe],  /pwa/ [po]
> One could get quite baroque if one were so inclined.......

That is weird.  How would you arrive at the consonant grade system to
describe this, instead of just positing six vowels?

> >There are a couple of odd things about this system....There are the
> clusters [?j] and [?w], which I've never seen in a real language.
>
> IIRC there's an article by Sapir (in his Collected Writings, ??) on
> a Native American lang. with [?j], in fact I think it patterned with
the
> ejective system.  I also recall a single word (in the very complete
> dictionary) in Sa'dan Toraja (South Celebes, Indonesia) with
> medial -?j- cluster, totally anomalous for that language.

That's good.  The cluster isn't hard to pronounce at all, so I would be
surprised to hear that it didn't ever exist.

> It kind of surprises me that /sj/ does not become [S_j] as in Japanese,
> else everything seems OK.

Well, I didn't want [S] in this lang, so I didn't do that.  It's my _fiat
dei_ as a conlanger.

> > The nasals and nasal occlusives do not participate in the mutation
> > scheme, so I'll leave them out of the next charts.
>
> No ny?

Actually, they *do* take the glides, but they never mutate, so I left
them out.  The only process that affects the nasals is dissimilation, so
you get /my ny nw ñw mpy ñkw/

> The change of y to w seems odd to me (I would have /k/ + /y/ > /ky/ and
/p/
> + /w/ > /p/), but the f change seems very likely as far as articulation
is
> concerned.

I said it was dissimilation--what you have is assimilation.  It was
important to keep the glide-grades distinct from the radical grades, but
distinguishing between /y/ and /w/ wasn't as important.  Thus, to avoid
assimilation we get dissimilation.  Besides this results in lots of /kw/,
which I like.

> >There are a couple of odd things about this system.  Everything's
pretty
> >regular underlyingly, but the surface forms are really weird.  There
is
> >no plain [w], but there is a [w_0].  There are the clusters [?j] and
> >[?w], which I've never seen in a real language.  Then /psy/ and /ksw/
> >mutate to assimilate the /s/, but plain /sy/ and /sw/ are unchanged.
>
> I also have [?j] and [?w] and no [j] or [w] outside of clusters --
there
> must be something going around.

Really?  How odd.  BTW, there were independent /y/ and /w/ at various
stages in the history of the language, but they were turned into /ty/ and
/kw/ respectively.

> >Do these things seem reasonable to everyone?  Or is it hopelessly
> weird?
> My though for the day:  Reasonableness in a language is abomination.
> Very little is weird.

:-)


Jesse S. Bangs     Pelíran
jaspax@ juno.com
"There is enough light for those that desire only to see, and enough
darkness for those of a contrary disposition."  --Blaise Pascal