Message-Id: <9202272241.AB21587@relay1.UU.NET>
Date:         Thu, 27 Feb 1992 13:19:20 PST
Reply-To: cbmvax!uunet!cuvmb.cc.columbia.edu!fschulz
Sender: Lojban list <cbmvax!uunet!cuvmb.cc.columbia.edu!LOJBAN>
From: cbmvax!uunet!cuvmb.cc.columbia.edu!fschulz
Subject:      word lists
To: John Cowan <cowan@snark.thyrsus.com>
Status: RO

Chris Handley writes

>Frank writes
>
>>My list generating skills are extremely poor, so it would take me a long
>>time to generate such a list. If anyone is interested in working on this
>>project, send me mail.
>>
>About 10 minutes with the thesaurus and a spreadsheet should give you such a
>list.
>

I guess my original message was unclear. I have no trouble with the
extracting categories out of the list. This is the trivial part.
I know how to use the computer tools.

What I what is more structure of top of this, which has to be
manually generated, since the conceptual framework is undefined.
Perhaps another example will make things clear.
Say we take body parts. A possible mnemonic ordering
would be to go proximal to distal, inferior to superior as a partial
ordering. This example is not good because it is too easy. I suspect
most other stuff will require more creative solutions. To really
do this well would require familiarity with most of the words on
the list. What I want here is to put a mnemonic conceptual link
structure on the semantic space. I still have not decided on exactly
what kind of link structure is best and how many links per word.
I later expect to define operators from the linked semantic space.
I want the the link structure to depend on intrinsic properties of
the semantic space. This is still probably not clear. I will wait
for more comments before explaining further.
--
        Frank Schulz ( fschulz@pyramid.com )