Re: Babel Text in Obrenje

--- In conlang@y..., jesse stephen bangs <jaspax@U...> wrote:

> Besides, Star Trek langs are bland because they're mostly ad-libbed
> nonsense, and any language with some structure will do better.
Often, we only hear a few sentence of a given language, which doesn't
give us much chance to see a structure.  At least the phonologies
could be more creative.  They all sound like Spanish.

Then again, the concept that aliens speak Spanish is just natural to
Americans.  ;-)


> > > Voiceless schwa?
>
> This is as good of a description as I can think of, given the information
> that you have.  Under a more sophisticated phonological analysis it might
> turn out differently.  With that in mind, though, consider where the
> voiceless schwa can occur--all of your examples in the Babel text are
> word-final.  What, then, is its phonemic status?
It's not phonemic, and neither is the voiced schwa.  The letter <e>
is simply pronounced as schwa in unstressed syllables, and naturally
loses its voice in a voiceless environment.


> Then think about
> syllable onsets and codas, to decide whether it's systemically a consonant
> or a vowel.
I see it as a vestigial vowel.  It used to be pronounced as /E/ in
Ancient Obrenje, and may still be nowadays if one wants to sound
"Shakespeary".  Even using the voiced schwa /@/ throughout will
make you sound posh or overly formal.


> Actually, looking at the data in the Babel text, I have an analysis to
> make: there is no underlying voiceless schwa.  Rather, it's a rescue
> strategy for taking care of unsyllabifiable final consonants.  Based on
> the examples in the Babel text, it appears that the only valid syllables
> are CV, CV:, and CVC; there are no CV:C syllables and no coda
> cluster--i.e. no syllable may have more than two morae.
Obrenje syllables are strictly (T)(S)V(C), where T is a stop, S a
non-stop, V a vowel and C any consonant (though the syllable-final C may
only be a stop if it's word-final or geminate).  Vowel length is
assigned to stressed vowels in open syllables, it's not phonemic.

Minimal pairs like *mon /mOn/ and *mone /mo:n/ might appear to suggest
a phonemic distinction, but the o in the second word is really only
long because it's in an open syllable (*mo.ne).

The final e may frequently not be pronounced, but still counts as a
syllable of its own.  In Obrenje, syllables and vowels have a
one-on-one correspondence.  Diphtongs are formed with the consonants
j and w.


> However, there
> are words in Obrenje that violate this principle in their underlying
> lexical forms, and on the surface these are rescued with the voiceless
> schwa:
>
> Underlying   Surface
> /s/          [s#]
> /tS/         [tS#]
> /varo:S/     ["varo:S#]
> /ni:s/       [ni:s#]
> /vikk/       ["vikk#]
Hmmm, I guess it's a valid theory, but not what I intended.  I've
always seen final -e as a residue from an older age where it was
still pronounced as /E/ (which also explains why it palatizes
consonants: *kos /kOs/ versus *kose /ko:S/).

So the underlying form of /varo:S#/ is really /va.ro:.S@/ -- the
schwa losing its voice due to the unvoiced /S/.

Anyway, thanx for the interest.  It sure helps against the feeling of
talking against a wall that I get sometimes.  ;-)


-- Christian Thalmann