[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [lojban] measuring semantic space
At 03:37 PM 5/6/03 -0400, MorphemeAddict@wmconnect.com wrote:
In a message dated 2003-05-06 5:56:54 AM Eastern Daylight Time,
lojban@yahoogroups.com writes:
How does one demonstrate that? (Or, for starters, measure semantic space?)
my first thought is: by starting with everything and successively dividing
it into smaller spaces. rick morneau (creator of katanda/nasendi) does
this or very similar.
second thought: use concept lists (e.g. roget's thesaurus).
A Roget's analysis of the gismu list was performed in the early history of
the language
http://www.lojban.org/publications/wordlists/roget.txt
http://www.lojban.org/files/history/ROGETNEW.TX1
i don't know what the 35 minimal concepts that have been mentioned here
are (any pointers, anyone?), but if i choose 5 concepts, something is
likely to get left out. if i choose 50 or 500, i think the same is
true: something will get left out. btw, toki pona comes to mind as a
minimalist language.
The concept of "blanketing semantic space" was important when we designed
the gismu list, and several approaches were used to ensure completeness,
erring on the side of redundancy. On the other hand, new concepts are
always being invented, and thus there is the potential need to cover an
expansion of semantic space, so we have fu'ivla which can be used for
borrowings and coinings, as well as additional potential to add gismu (but
hopefully not for a generation or so).
There has been at least one attempt to map semantic space with only 30
roots. Somewhere in my deep archive (paper only), I have a letter
describing this. The guy who devised it was selling his tome defining all
English words in terms of these roots for something like $100. (I may have
mentioned it in an early JL - likely before issue 7) It was pretty gross
(and rather naive in its interpretation of English words as if they had
only one meaning).
lojbab
--
lojbab lojbab@lojban.org
Bob LeChevalier, President, The Logical Language Group, Inc.
2904 Beau Lane, Fairfax VA 22031-1303 USA 703-385-0273
Artificial language Loglan/Lojban: http://www.lojban.org