From: cbmvax!uunet!math.ucla.edu!jimc
Message-Id: <9108062337.AA29542@luna.math.ucla.edu>
To: gryphon%dino.ulowell.edu@ulowell (Coranth)
Cc: lojban-list@snark.thyrsus.com
Subject: Re: neat thought that came up... 
In-Reply-To: Your message of "Mon, 05 Aug 91 16:38:37 EDT."
             <9108052038.AA02760@dino> 
Date: Tue, 06 Aug 91 16:37:58 -0700
Status: RO

Coranth d'Gryphon writes:
> Now, if this is true, then it would be worth investigating so that we can
> generate a vocabulary that falls (in item count) BELOW the average max
> vocabulary of our target audience.

I think we may have already done this -- in Lojban (and the various
other Loglan varieties) (as well as Basic English and probably others).
The set of gismu is well under any suggested limit.  Current doctrine
is that Lojban lujvo are to be constructed in the field, rather than by
a central authority or "Word Makers Council", yet they are supposed to
have a unique meaning.  One can therefore deduce that there must be an
algorithmic procedure to map from rafsi to meanings, even if we don't
yet know what that procedure should be.  

Conclusion -- and a powerful one it is, too:  given the algorithm, to
understand any Lojban word you only need to know the gismu / rafsi; the
lujvo don't count towards vocabulary count.  What power this gives to a
beginner!  That's my main motivation for pushing so hard on diklujvo.  

Of course users will quickly memorize their favorite lujvo and so if
you deduce vocabulary from memory space occupied, these lujvo will
count.  But the main conclusion remains: if lujvo are handled
aggressively, Lojban can become uniquely powerful, uniquely precise,
and uniquely learnable.

		-- jimc