>There are computer programs which make music, so I will not be
>surprised if a computer program makes grammar and vocabulary
>of a language. I could not distinguish such a computer-generated
>language from a language made by a human artist. The only
>exception are philosophical languages. The compound words of
>these languages are descriptions, so the language-making
>computer would have to know as much about the world as we do.

There are programs which indeed write music that to most sounds reasonably
ok. However, these programs only run on theory of harmony and similar
theory. You can only come this far with that (and all music theory has to be
derived from the actual listening experience of human beings). This is what
makes the masters
of classical music different from any application - they knew where they
could bend the rules or even outright break them in order to achieve beauty.
The giants have something computers won't have - intuition and a truly
subjective - yet surprisingly "universal" - sense of beauty and the ability
to do a mistake and realize "hey, wow! that sounds great! let's keep it", or
even the ability to improvise their way out of a bad mistake (like Jimmy
Page). If we had computers compose all our music, we'd only have soft easily
listened sweet music, no Schostakovich encoding his anger at the soviet
system behind layers of soviet propaganda. No Schoenberg breaking his own
rules of serialism, even inventing serialism at all. No Schoenberg to
emulate quartertone music in a twelvetone system (as has been proposed that
he did). We'd have no Finale to Beethoven's 9th incorporating a choir - a
choir! - into the well established symphonic formula that very clearly
excluded such ideas. We'd probably not even have tritonus utilized as a
"harmonic engine" behind our most important chord sequences. We'd probably
never had heard the blues shuffle rhythm - for which engine is probably to
come up with the idea to use such a simple yet quirky figure as the basis of
thousands of songs? Provided the tritonus ever got to be the harmonic
engine, would we have the I7 IV7 V7 chord sequence? And what about jazz?
Hard rock? I think not. Computers cannot invent new music and assess its
quality like men can. It required a visionary like Link Wray to realize that
the fuzzy noise that guitar amplifiers cause when played at high volumes for
a long time could be utilized as a timbre of its own. A computer would
merely have proceeded to fix the problem, without thinking twice about it.

I suspect the reason you'd be unable to distinguish a computer-made language
from a man-made, is that you don't enjoy quirkiness, you don't enjoy
irregularities, so you don't search them out. Is this true? It is in the
irregularities you can find the marks of human hands.
You probably don't know what a wealth of options that semantics provide us
with. Consider the great puns any human can create into the language on
purpose, which in a computer would only arise so-and-so often as a result of
chance. Today I saw this old coca-cola commercial, "Coca-cola - the beverage
for discriminating people" - without any knowledge of old cocacolian company
culture, you'd probably not get that pun - which probably wasn't intended a
pun back when it was coined. Such puns won't ever be an integral part of a
computer-generated conlang.

And, as for papers and pens, a second benefit of using them, is that you can
write with any new-invented character or any alfabet, without having to
create a new font or track it through internet. You can apply any weird
triple-dot diacritic or squiggly descender or barred ascender to any letter
you want, and change between hundreds of alfabets on one page without having
to press any odd key-combination or look down "unicode tamil 98" in some
menu that your computer just decided not to let you open. You can write with
any slant you want, and even with bent lines, you can make text crossing
text to save paper and get a nice look. All those require extra job on a
computer.

With all due respect,
Markus Miekk-oja