Dave Wolf wrote [on meta-composition]:

>As the poster, I wanted to address one of the comments on this idea: that
>it was similar to an outline for a writer ...  I think what I have in mind
>is more the emotional architecture of the piece; this would include some
>specific ideas as to how that architecture would be accomplished, even to
>the point of sketching out certain motifs, and the way they would
>transform--the emotional highpoints, perhaps.  ...
>
>If my understanding of this is close to Beethoven's intention, it seems
>unlikely he BEGAN with that idea in mind-- but "discovered" this direction
>as he composed.

I agree with all that.  He probably did both.  Some compositions would have
a greater degree of preplanning (sketching) than others.  Some themes just
naturally flowed on their own, where others took some thinking and
planning.  (ref my previous post from his letter)

As a progrmmer, which is very close to composing, I sometimes need to draft
something in advance because of its complexity.  Usually I design on the
fly as I go.  Often I realize a better way to do it, and rewrite the whole
routine again, forgetting my design entirely.  (I redid the theme generator
several times) Sometimes I find I must redesign something as I begin coding
it - coding is where the rubber meets the road.  The actual hearing of the
composition is where the rubber meets the road as well, probably causing a
composer to redesign now that his has heard it as an implemented whole.
Since Beethoven composed after deafness, he could probably hear from his
design.  But I'm sure his impulsive, improvisorial nature influenced his
music as he actually got into writing it down..

Since writing the above I took a break to watch the pianist Martha Argerich
play the 3rd movement of Schumann's A min on the ARTS channel, which I keep
running in the background.  That inpired me to go and practice the cadenza
to Beethoven's 4th PC which I am learning (just the cadenza, not the PC) -
a great cadenza - planned or impulsed, I wonder.

>I'd think that the ability to "shape" the composition (before the
>themes and all the writing is begun) would be a very important parameter.
>An interesting adjunct of this would be if your program had a database of
>the "shapes" of a number of specific movements

There is a library of forms to choose from, i.e.  rondo, sonata, rhapsody,
etc.  but they are not emotion based.  The computer can't actually "hear"
the music so must 1) generate themes wihch "should" sound decent based on
the rules employed, and 2) organize these themes and their motifs into a
structure that has some organization.  Your's is a good suggestion, perhaps
for version 2.

I once start a thread about emotions, hoping to get some insight on this
but the general opinion of the list was that a list of emotion could not
even be made, much less an analysis of the underlying musical aspects.  I
am not sure the computer can deal with emotions until they can be defined
and quantified since the computer can't feel.  That's not to say that the
software isn't fun to play with and as such has some value.  Perhaps the
list might be motivated to try again sometime.

Bill Pirkle