Date:         Fri, 16 Oct 1998 19:27:34 +0100
Reply-To:     George Foot <george.foot@MERTON.OXFORD.AC.UK>
Sender:       Lojban list <LOJBAN@CUVMB.BITNET>
From:         George Foot <george.foot@MERTON.OXFORD.AC.UK>
Subject:      Re: vals,r,tesoroso
To:           Multiple recipients of list LOJBAN <LOJBAN@CUVMB.BITNET>
In-Reply-To:  <E0zU9rh-0002Ws-02@sable.ox.ac.uk>
Message-ID: <908564893.105065.0@listserv.cuny.edu>
Content-Length: 1640
Lines: 39


On Wed, 14 Oct 1998, Logical Language Group wrote:

> The problem withg a thesaurus of Lojban is the place structures.  the
> thesaurus classification done in 1989-91 was based primarily on the x1 place
> and not on the gismyu itself.  You end up classifying words based on their
> keyword, and not on their true meaning.

Wouldn't the `best' way to arrange a thesaurus be to stick to x1
classification, but shuffle the place structure where it's
useful; for instance, "le se casnu" and "le te tavla" are both
subjects of discussion.  Maybe you would list the lujvo
"selcasnu" and "tertavla" if not the two-word selbri or three
word sumti (since the `le' is pretty much always going to be
there).

That said, I haven't yet looked at the current thesaurus -- I'm
just responding to what lojbab said.  I think the results of the
recent word list will also need to find their way in at some
stage.

> I'll se if I can find info on how we did the thesaurus.  I know it was
> mostl;y based on the original Roget.

I agree with your other point that a Lojban thesaurus can't be
based solely upon direct translation from an English thesaurus;
though that's probably a good place to start.

--
george.foot@merton.oxford.ac.uk

xu do tavla fo la lojban  --  http://xiron.pc.helsinki.fi/lojban/lojban.html