MIME-Version: 1.0
Date: Mon, 20 Jan 2014 14:30:57 -0500
Message-ID: <CAAJP7Y2sgOfLqOfiozp2hgHvKw8okJOq0yJ5Ahi903GAm4AMOw@mail.gmail.com>
Subject: simpler language enjoying lojban non-ambiguity properties?
From: Warren D Smith <warren.wds@gmail.com>
To: lojban@lojban.org
Content-Type: text/plain; charset=ISO-8859-1
X-Spam_score: -0.1
X-Spam_score_int: 0
X-Spam_bar: /

I'm interested in the idea of an artificial language enabling
human-computer (H-C; also H-H and C-C) communication.
The unambiguity properties (every sentence has exactly 1 parse, the
sound-sequence uniquely specifies the word sequence) of lojban would
be very helpful.

More generally, every artificial language so far (500-1000 tries), has
failed.  Why try again for sure failure?  Well, there is this:  H-C
communication would offer a big economic
motivation, and to individuals, not just society wide.  That is the
first time ever in human history that this can be said.

----

However, it seems to me lojban is too complicated and at least some of
that is intentional.  Linguist Arika Okrent said constructing a valid
lojban sentence is "like doing long division in your head," and I
observe that a large fraction of the lojban sentences on your
lojban.org web site are incorrect (proof: run your jboski program on
them - I already did), ditto sentences on online lojban fora -- i.e.
it is so difficult that even *you* often cannot do it even with
considerably longer than real time to work on it.  If incorrect lojban
were common that would defeat the purpose (for me) of enabling easy
H-C communication taking advantage of unambiguity theorems.  Okrent
also pointed out the full specification of lojban grammar was 600
pages, a hellish amount.
As an example of "intentional" complexity: the "emotion tags" feature,
while maybe desirable for poetry & stuff, is completely
counterproductive for communicating with computers.

Another criticism of lojban is the "culture neutral" feature.
The result is that learning words is basically maximally difficult for
everybody.
The amount of (say) English in it has been diluted so much that
English speakers have essentially no start-off advantage.  In contrast
Esperanto and Interlingua intentionally tried to give euro-speakers a
start-off advantage by making the words and rules highly euro-like.
That may not have been very useful for (say) Arabs trying to learn
Interlingua, but it was a big win for euro-speakers.  For H-C
purposes, most computer users already are familiar with euro-languages
(although maybe this is less true than it used to be) so the
culture-neutral thing is a disadvantage. Also, as a matter of
marketing, culture neutral seems bad -- you want to get a large core
group of speakers fast, and being culture-biased will enable that.
Esperanto and Interlingua are the most successful attempts ever and
they went the biased route.

I saw it was claimed at least one fluent lojban speaker existed (Nick
Nicholas) with no further details (how many are there?).  That
suggests these objections are overcomable, EXCEPT that I saw, zero, I
repeat zero, data on what fraction of Nick Nicholas's high speed
utterances actually were valid lojban that passes jboski.  If, say,
only 20% do, then the claim he is a fluent lojban speaker, is kind of
debatable and for H-C purposes this would be nearly useless.

OK, this brings me to my QUESTION.
Suppose it were desired to produce a simpler language (perhaps related
to lojban the way "basic english" is related to English) still
enjoying all un-ambiguity theorems, but
with grammar describable in only 100 not 600 pages.

Do you think this would be possible?

I mean, it might be that the lojban creators did a pretty good job,
and it is just not possible to do it that much simpler.  Your guesses
solicited.


-- 
Warren D. Smith
http://RangeVoting.org  <-- add your endorsement (by clicking
"endorse" as 1st step)