Bill Pirkle wrote:

>...Now back to the original point, assuming that great music could be
>written by a computer, guided by an UNtalented human -or with what Dr.
>Strangelove called "human meddling":-) - will it be accepted on its face or
>will the CM community reject it out of hand - as my query asked - "How much
>does the composer and historical significance matter?" I think that its a
>valid question and it is sincerely asked.

The first thing is whether or not there is anything to your assumption.
But granting it for the moment, I would say that the music would stand
on its own merits, if it were truly as good as what master composers have
written.  Especially if people didn't know it was written by a computer.
False identity after all has been responsible for many famous fakes.  Think
of the lovely works composed by Fritz Kreisler that he published under
various pseudonyms, like Pugnani, Tartini, Padre Martini, and so on.
People didn't stop loving them after they found out the truth.  But he
would have had a harder time selling them without his little deception.

But whether a computer can actually DO that, is another very big question.
Computers at this point only do what they're told.  That means that all
the allowable variables and their statistical occurrence in actual music
of any period and composer have to be programed into the machine.  It's
probably more possible to do more along those lines than ever before,
especially now that IBM has just built the world's fastest supercomputer
which is 3x faster than the next fastest, which may be the one that beat
Kasparov.  BUT how will a computer EVER know the EXACT combination of
variables that will make a human being weep in context? Like that
modulation in the first movement of Schubert's Quintet in C? I don't weep
because I know it's by Schubert!  It could be by Joe DiMaggio for all I
care.  All I am interested in is how my memory of what came before
interacts with that musical present of surpassing sadness and tenderness.
If a computer can learn to FEEL that, MAYBE it can find out how to express
it in music.  But that's a few dozens of teraflops away, I think.

Computer music experiments date back to the 50s, with Lejaren
Hiller's "Illiac Suite" for string quartet, which is AFAIK the first
computer-generated musical score.  I have heard this work but it has been
a long time.  My impression of it then was that it was quite mechanical.
Certain contrapuntal and formal techniques had been programmed, and I think
a theme was inputted (is that a word?) for the computer to work on.  I
seriously wonder if major strides have been made since then.

Chris Bonds