Message-Id: <199601110527.HAA28865@xiron.pc.helsinki.fi>
Date:         Thu, 11 Jan 1996 00:25:57 -0500
Reply-To: Logical Language Group <lojbab@ACCESS.DIGEX.NET>
Sender: Lojban list <LOJBAN@CUVMB.BITNET>
From: Logical Language Group <lojbab@ACCESS.DIGEX.NET>
Subject:      Re: <stizu>
To:         sbelknap@UIC.EDU
Cc:         lojban@cuvmb.cc.columbia.edu
Content-Length: 3972
Lines: 67

There have been several semantic maps of gismu made over the years - each
person's semantic map has its own idiosyncracies.

In your case, yourt complaints seem to have little to do with semantic
maps, and more to do with your personal definitions of (I presume) the
English words.

But a chair is a chair regardless of whether any x3 sits in it.  And a
table likewise if nothing is upon it.  The x3 of table is used to distinguish
a table from any flat surface - tables, whether furniture or mesa in the
desert, project upward above the surrounding, as well as have a flat top,
and it is fairly irrelevant whether anything is on the table as to whether
it is a table.

The exclusion of the sitter from the chair, also distinguishes stizu from
zutse.  But we were looking specifically to deal with things like
window seats and car seats and toilet seats as well as bean bag chairs.
We could have put in a material place, but in this case the focus was
on function rather than on form.  For jubme, the focus is on form.

ckana is not really in the same category as the otehr two, since it is not
merelyrestricted to pieces of furniture or parts thereof.  A sleeping mat is
not really a piece of furniture, nor is seed bed for a flower garden, nor
is (harkening to the event usage in x3) any particular place a couple might
choose to "go to bed together", which might be a spot on the floor.

When we defined the gismu, we took the English (and other-language where we
knew them) metaphors based on the word, and tried to make the gismu as
broadly cover the range of meanings as possible, because it seems easier to
restrict meaning in tanru than to extend it.  Yes, we went through phases
where we tried to bring the places structures into semantic comformity
with each other, but these efforts seldom worked well, and often stalled
out - having done it at least 4 times myself, i assure you that a complete
and thorough look at gismu place strutures is immensely time consuming, and
of course biased by your semantic prejudices of the day so that the changes
you make at the end of the alphbet tend not to be consistent with the ones
at the beginning of the alphabet, etc.

We have had thesaurus or cross-reference surveys of the gismu list from
Paul Doudna (on TLI Loglan), Athelstan, Jim Carter (both TLI Loglan and
Lojban) John Cowan, Veijo Vilva, and myself (really as a modification of
Athelstan's, after looking at the others, and ioncorporating some of my
own understandings).  The cross-reference lists in the rightmost vcolumns
of the gismu list serve as a gross-level compendium of all the semantic
linkages identified by the various people which I accepted as being valid
(sometimes the linkage is contrast or opposition though, so these aren't
necessarily thesaurus listings.  You can also get some semantic linkages in
su'oremoi places especially by looking at multiple entries for the same
English headword in the draft dictionary - I did a lot of weeding, but
many headwords still have a dozen or more entries, and this still weeds
out things like all the standards places, which in effect makes all of
the gismu that have them semantically linked.

In any event, now that the draft dictionary is done to this point, we are
accepting no more major restyudies of the gismu list - the last one took a
year to complete, and we want to be published before that much more time
passes, at which point we will be out of the prescription business at least
in theory.  Anything more than an occasional single word place change
has SIGNIFICANT effect on the dictionary - it can take me an hour to
edit all entries for a single gismu even with some systematic procedures and
tools, but this does NOT include the possible impact on any/all lujvo built
 according to dikyjvo rules that Nick developed, whcih someone else has to
 analyze since I have yet to learn much about Nick's system (and disagree with
 it
philsophically enough that I may never master it).

lojbab