lujvo, tanru, and the price of fish

[cbmvax!uunet!mullian.ee.mu.OZ.AU!nsn]

Dear folks,

you will have noted in recent times the discussions on Esperanto wordmaking,
and how the aposteriori scheme for it, worked out be Rene de Sausurre (no
relation) and codified by Kalman Kalocsay (who is not considered 'some guy 
called Kalocsay in that Esp community, but then this is a forum for lojbo,
not Esperantology), did not fit the existing wordmaking system perfectly.
I fear you all have learnt the wrong lesson from this. To begin with,
Zamenhof's wordmaking was as indefinite and not-rulebound as lojban's tanru
are today. But I strongly suspect that, thirty years from now, somebody
will look at usage and say, "hey, most lujvo of the scheme {brodaske}
(something-science) can be expanded as {saske be lo broda} (science about
something) rather than, say, {saske poi broda} (science which is something).
e.g atom science is science about atoms, not a science which is an atom."
The researcher could then formulate analogous semnatic rules to those of
Esperanto, which would be more complex as the neat division into semantic
categories of brivla in Esp does not exist in Esp, and because the Esp
rules are really more to do with grouping than semantics anyway; and this
in turn could have a prescriptive effect on lojbo as has happened in Esp,
such that noone will end up saying {ni'oske} to mean {saske poi cnino}
- a new science - but {saske be lo cnino} - science about new things, the
study of innovation of something. This will have an advantage in that you
won't have to run off to the lojban academy every time you want to coin
a lujvo, becuase the semnatics of it will be kinda-sorta well defined.

But that fact is that lujvo, and often tanru, are not that advantageous,
not always worth the effort. Are lujvo shorter? With all those spare y's
lying around? Not always. Are tanru that good? Consider:

le crino gerku [ku]
le gerku poi cnino [ku'o] [ku]
le gerku poi ke'a cnino [ku'o] [ku]

The first (green. dog.) is riddled with unambiguity. The second is miles
more precise, and the third IS AS UNAMBIGUOUS A STATEMENT AS POSSIBLE.
How many more syllables? One to three. Often tanru are not necessary.

An exercise for you all (John Cowan, keep out of this one %^). Construct
a tanru for "big blond [male] lifeguard". Then, expand this tanru completely -
express everything in terms of explicit relations between single brivla.
Your brivla are:
nakni - male
barda - big
pelxu - yellow
kerfa - x1 is the hair of x2
fanta - x1 prevents the occurence of x2
morsi - dead
djacu - water

You might find the exercise instructive. Indeed, I contend that the key to
good lojban is NOT using tanru, but being able to avoid them. If lojban has
a proper semantics at all, then all tanru must be expandable.

Consider this the next time you use {mutce} in a tanru. Guess what, folks:
{mutce} IS A BRIVLA IN ITSELF! It does not exist merely to be inserted before
another brivla, like the English 'very': it has its own place structure.
Use it sometime.

Hope I've stirred y'all up enough. co'omi'e nitcion.