From: cbmvax!uunet!mullian.ee.mu.OZ.AU!nsn
Message-Id: <9104300418.25228@mullian.ee.mu.OZ.AU>
To: lojban-list@snark.thyrsus.com
Cc: nsn@mullian.ee.mu.oz.au
Subject: This Chemical Element stuff
Organisation: Department of Electrical Engineering, University of Melbourne
Smiley-Convention: %^)
Date: Tue, 30 Apr 91 14:18:09 +1000
Status: RO

>djan. kau,n. pu cusku
>iVAN. derjanskis. writes:
>> Who cares for
>> metaphorical uses of words?  I really hope you don't intend {nikle} to be
>> used for the US 5c coin, which most of the world has never seen!

Ivan is so right on this one, I can't begin to express it. But then I don't
need to, because his correctness is self-evident. Let me say what I understand
has gone on. There are some basic semantic elements which can be used to form
a vocabulary - semantic prims, if you will. Basic English would like to think
it is based on them, as does the similar Interglossa. Esperanto started with
a similar methodology, but like so much else in Esperanto, usage got the
better of that idea; none the less, the Academy of Esp did draw up a Basic
Set of Official Stems (Officialisation of Stems has been very haphazard in
Esp) in the seventies, which the community has ignored anyway.

There are three types of gismu in lojban: prims, abbreviations (cmavo etc.)
and names (cultures, animal names, such like). The prims I don't think many
will have trouble with; the "how come you don't print these" place structures
are there to guarantee (we wish!) cultural neutrality or at least indicate
bias. That's another topic. What people are rightly indignant about is the
name gismu: {navni} is a classic example. Why have them, and what do we do
about them?

For better or worse, presumably for purposes of economy, Jim Brown apparently
decided some names-of-objects will be used so often in word formation, they
should be assigned gismu. This, regrettably, is where the concepts "metaphor
as the real world understands it" and "metaphor as JCB construes it" get
mixed up. Accept the argument that 'rose' is used in metaphor (a la real
world), as in a Burns poem or something? Then hey presto, give it a gismu,
because some lojbani schmuck will one day make a metaphor with it.

Bullshit, if you'll pardon the vernacular (and even if you don't). I challenge
anybody to make a tanru with {rozgu} that has anything to do with this
proclaimed metaphorism of JCB. {rozgu melbi}? (Pretty as a) {na rozgu frili}
(life's no bed of)? These aren't tanru, these are malglico crudities, which
have no meaning in lojban. So the rationale for having a {rozgu} gismu
fails (unless someone comes up with a half decent tanru which isn't a dumb
literal translation of anglicism, something I don't think any of us are up
to right now). Tanru IS NOT METAPHOR! And Metaphoric use of semantic elements
in langauges, however widely, has nothing to do with tanru.

Not all name gismu deserve immediate death. Strangely enough, those "cmene
in gismu clothing", the cultural names, deserve gismuification, because
they are used a hell of a lot in tanru (France country, Germany langauge,
Spain literature). The criteria for selection of cultures suck, and I suggest
someone propose a batch more of 'em for next Logfest, but gismu themselves 
are alright.  Not indspensable: anything a name gismu can do, a cmene can do 
and a le'avla can do, as far as semnatics is concerned. But convenient and 
lujvoable.

Turning to the chem elements: don't worry Ivan, John can't supply you with
non-malglico tanru from Nickel or Neon, becuase there are none. There isn't 
even a malglico tanru for it, apart from "nickel coin".

>There is undoubtedly a great deal of glico bias (and even merko bias) in
>the gismu list.  It's the product of raw empiricism, nothing more.

True prim gismu do display a bias; but I can live with that. Any system of
semnatic prims would.

>However, "metaphor" is the very basis of including a word as a gismu rather
>than allowing it to remain a le'avla.  Borrowings do not enter into lujvo,
>but otherwise they are full-fledged Lojban brivla.  In general, a word
>should be a gismu if good and useful tanru can be made from it.

There is nothing wrong with this, don't get me wrong. I simply contend that
good useful tanru cannot be made with many of the name gismu. Some, p'raps.
NOT {nikle}.

So what to do? Recall: 5 years from publication till final baseline. Pre-
diction: at that time, people go over the gismu in use and find that tanru
are not being made at all with {rozgu},{navni},{nikle} - or at least not
significantly more than with {xukrbromidi},{xukrurani}, whatever. If they
have any brains, they bow to consensus and either:

shove the unused gismu into le'avla space (which is sensible, though heavy-
handedly prescriptive), or:
leave it in as an archaism, and recommend against it (which is more or less
what Esperanto does: the Russian premetric length (?) unit Versht is an
official stem, and people are quite free to ignore it, which indeed they do).

Gismu are not carved in stone: I am confident they will gradually be added
to and decremented from (VERY gradually, if I have anything to do with it;
the highest danger for any AL is the "I'll just correct this bit for now,
and that bit for later" syndrome: reformitis. Lojbab's "chuck another cmavo
on the barbie" reminds me of this: debate against/for new cmavo should be a
lot more vocal, and a lot better considered). We don't want Lojban to end
up like Esperanto (Waringhien's version, vide his dictionary) or Ido, dragging
in a neologism for anything you can think of. Furthermore the morphological
barrier between gismu and le'avla will keep things under control.

But in general: if a gismu decision is stupid, this will be reflected in usage.
You COULD still plea for the removal of {navni} and {nikle} now, but I don't
think this is necessary: we CAN guess where JCB was being idiotic, but usage
will sort things out much more authoritatively.

In summary.
Gismu are: prims, and often tanru'd cmene.
Therefore, Gismu are not just prims (see Lojbab's response to jyjym., JL13)
Gismu like {rozgu} originated from a misunderstanding of tanru.
Gismu like {rozgu} deserve to die in the arse.
Gismu like {rozgu} will die in the arse, as far as tanru construction is
concerned, because we (Ivan, me, likeminded progressives %^) believe such
tanru construction will/should not happen.
Let lojbo give the coup de grace to these Gismu.
These gismu can then either be deallocated or fossilised.

Oh, and another thing. I've heard it claimed that BAI are there to, interalia,
allow all place structures to be expressed with BAI. Prove it. Do it for
just five brivla, and I'll believe it. Because BAI looks awfully diletantish
to me. ("Superinclusive"? Really? If there's a semantic theory on prepositions
or something such out there, avail yourselves of it. A language like lojban
cannot be designed with the happy-go-lucky empiricism of a Zamenhof. You
need to know exactly what you're doing.)

Apologies to Leyzer Zamenhof, James Cooke Brown ("Ow! I feel good! %^),
John Cowan, Bob LeChevalier (whom I've glancingly flamed in the above without
naming him), Nora LeChevalier (becuase I'm flaming her now: Nora, you let
Helsem get away with {zgikei} for "to play music". It happens that the
tanru 'to play music' exists in English, French, German, Greek, and Russian
[and Esperanto, surprise, surprise]. This is irrelevant. zgike kelci can only
mean to play a musical game, NOT to play music, which would be zgike rinka,
or zgike pilno for to play an instrument), and Ivan Derzhanski, becuase I've
presumed to speak for him. I've been very bitchy in the above, and don't
mean to offend. My excuse is simply that I don't want Lojban to fail like so
many ALs before it. It's a damn good concept, it should be allowed to grow
unhampered by technical stuffups. And it should be allowed to grow irrespective
of the hypothetical testing of a hypothesis which might never occur. Lojban
already means a lot more to me than SW. May it be used with pleasure by you
all too.

ki'eki'emi'e nitcion.