Symmetry of Lojban phonology [ was Re: [lojban] on Lojban pronunciation ]

[Brook Conner <nellardo@concentric.net>]

"Bob LeChevalier (lojbab)" wrote:
[ ... on /x/ being in a lang designed by English speakers ... ]
> With that collection of languages (Greek, German, Russian, Celtic,
> Arabic/Semitic), as well as the linguistic analysis, it is very
> unsurprising that 'x' showed up.  The real debate was whether we should
> include for symmetry the voiced and unvoiced pair of velar fricatives, just
> as we had pairs for all the rest of the unvoiced consonants, but we decided
> that this would indeed be difficult to teach to the poor English speakers
> %^).  

%^) indeed. I noticed this particular lack of voiced/unvoiced symmetry
when laying out an alternative lojban orthography using Tolkien's
Tengwar (originally done independently of Eric Raymond's, now done in
light of it, but with some differing decisions). If you ignore any
"traditional" letter assignments to tengwar, and simply follow a grid
arranged (roughly) so:

                   labial  dental palatal  velar
stops
   voiced            b       d       g
   unvoiced          p       t       k
fricatives
   voiced            v       z       j
   unvoiced          f       s       c      x
nasals               m       n
liquids/semivowels           l              r

Seventeen consonants, plus five vowels, plus y, is the 24 regular
letters. Then the pause ".", and the vowel separator "'", and finally,
the "," for removing vowel dipthongs. Use the tehtar for the vowels. A
medial dot is consistent with tengwar punctuation for a pause, the long
carrier for "'", and the short for ",", and two dots for "y".

So you find that the consonants, with very few gaps, neatly fill the
space. The lack of a voiced velar fricative is the one noted as a
concession to English speakers :-) I suspect the use of /ng/ as
allophonic for /n/ is another one, and there goes the palatal nasal. As
for the lack of a labial or palatal liquid or semivowel, as I recall,
Loglan had a "w" and a consonantal "y", which would have filled in those
spots. I suspect that those were dropped to maximize separation between
the liquids, as several major languages treat them as allophonic.

> So I believe we defined the language to allow the voiced velar
> fricative as an allophone for the unvoiced one, though I haven't ever heard
> anyone use it.

Doesn't Klingon have a voiced velar fricative or two? :-)

As for the difference between my Tengwar orthography and ESR's, I find I
esthetically prefer the use of tehtar for vowels from the approaches
typified by the Mode of Beleriand and other human modes. But then I have
Quenya tattooed on my arm, so maybe I'm different that way.

A couple of further notes: This particular mode does provide for some
visual distinction between cmavo and gismu, even compound cmavo and
lujvo - it uses the vowel-following convention for tehtar location - as
most lojban words end in a vowel, this seemed correct. A second vowel
would go under the tengwar, unless preceeded by ', in which case it
would go over a long carrier (no problems with missing a ' there!)

A greater visual distinction (and more visual variety in the script)
could be provided by using different tengwar for cmavo and for brivla.
With the compactness of the consonantal mapping above, there are enough
letters even in Tolkien's attested tengwar for *two* sets of consonants.
Pick one for cmavo, and use the other for brivla. 


Brook