Suggestions for (no) change (was: [lojban] To clarify...)

["James F. Carter" <jimc@MATH.UCLA.EDU>]

Not to clutter up the Lojban list with non-Lojban stuff, and not to be seen
as aligning myself with a rather obnoxious troll, and not to take the word
"unsatisfactory" too seriously, but I couldn't resist a reply...

On Thu, 29 Nov 2001 pycyn@aol.com wrote:
> In a message dated 11/29/2001 6:09:42 AM Central Standard Time,
> thinkit8@lycos.com writes:
> > the words themselves are
> > unsatisfactory (both in form, and what was chosen to be a gismu)...

> Well, it would help it you did make clear just in what ways they are
> unsatisfactory. that would lead almost directly to suggestion about how they
> could be better done.

I think you've left a change category off your list: the word formats. I'm
not the first person to point out that it's kind of a crock to have
separate gismu and rafsi. In -gua!spi I used C^nV^n words, with a
practical maximum of 4 letters, and got pretty well separated word
assignments.

> This limits the possibilities for change to 1) a new choice of
> languages (but the stats on language use have not changed much in the
> last 10 years),

Due to effort issues I limited myself to Chinese, Latin and English
(avoiding Latin borrowings where possible).

> 2) a different algorithm (really?) for constructing Lojban words from
> those languages,

The algorithm can be tweaked to maximize separation between words and to
maximize the use of euphonic patterns as judged in the source languages.
Totally random assignment turned out to be unacceptably ugly.

> 3) a different choice of words in various languages for representing
> the concepts (I suspect there is now the collective expertise to do
> something useful in that respect),

But again, that's a lot of work, and unlike Esperanto where there's the
strong Latin influence to promote cognate formation, I've found the cognate
relations in Lojban (and -gua!spi) to be relatively unuseful in learning
vocabulary.

> 4) choosing different concepts to begin with (maybe more
> semantic primes rather than high frequency -- but complex -- words).

I agree: we've never been able to sink our teeth into the idea of a
semantic basis set, but it's clear that some of the gismu are a lot more
"equal" than others. But if we ever got into that kind of discussion it
would make today's arcana-fests pale by comparison. It's also
doubleplusungood to limit the available semantic space so as to make the
language less expressive, and one also can't be sure if the selected basis
truly covers the desired semantic space.

It would be a very bad idea to actually think of renovating the Lojban
vocabulary in any of these ways, because a lot of people have invested a
lot of time in learning *these* gismu and rafsi. I have my ideas for
improvement, as people know, but I also have confidence in the semantic
coverage of the Lojban gismu, and even if not perfect, they're really not
bad at all.

James F. Carter Voice 310 825 2897 FAX 310 206 6673
UCLA-Mathnet; 6115 MSA; 405 Hilgard Ave.; Los Angeles, CA, USA 90095-1555
Email: jimc@math.ucla.edu http://www.math.ucla.edu/~jimc (q.v. for PGP key)