During my daily dose of reddit I saw an article on the artifically-intelligent software composer, Emily Howell, and the story of David Cope and his first creation Emmy (or EMI, for Experiments in Musical Intelligence).
Algorithmic music has always interested me, but not for the sheer pleasure of music appreciation; it’s the way you can create atmosphere on-demand, and given my love of computer games - the most interactive art form yet - the ability to make music defined by your actions raises the experience to a whole new level.
Just imagine. You become one with the violins. Horns wait for their cue while the bassoon annotates your every move. While you were busy playing the conductor with your existence, the timpani has tirelessly amplified your heartbeat a million times. Every little bit so seamless and unobtrusive, so perfect, that the music feels like a natural extension of you. It literally is the soundtrack of your life, and just thinking of the potential gives me goosebumps.
Anyway.
That’s all well and good, but Emily isn’t that kind of composer.
She writes music for the sake of music. And in contrast to her forebears, it’s all original. (Naturally, with the necessary qualifications that any jazz man will give you - numerous quotes about everybody stealing from everybody else spring to mind, unattributed in my disorganised mind yet universally acknowledged.)
I honestly have no problem with this, since anyone can write a song. Writing a good song, well, that’s something else.
What I’ve heard so far (excerpt 1, excerpt 2) is certainly pleasing, but it’s far from Chopin. That’s not a performance issue, either - Chopin sounds great even in the hands of an amateur - but something more fundamental. I’m definitely impressed that a computer wrote it, but as a piece generally? Meh.
I’m not here to write a review, though. What I found especially interesting was the reaction of critics against Emily’s precursor, Emmy. David Cope created Emmy in an attempt to solve a severe case of writer’s block; her task was to create chorales in the style of Bach. She passed with flying colours, and musical scholars hated him for it.
(Cope also created more scores in the style of other composers, with varying success.)
The funny thing is, being able to compose music to sound like a long-lost Bach piece shouldn’t be as amazing, disturbing, or even as revolting as the critics have made out. Our definition of a Bach piece is based solely on his repertoire as a reference point; the computer doesn’t have to capture the essence of Bach at all, merely convince us that it is Bach. Something entirely different.
Conversely, suppose you find a long-lost Bach piece that sounds unlike his other work. You have two choices: destroy the evidence, or move the goalposts to retain consistency, calling it a blue period, or hampered by external creative interests. (I particularly like that one - it’s like saying he’s on drugs without the stigma of an outright accusation.)
Either way, perception and reality are not necessarily aligned, which is why I consider mimicking specific things a bit of a red herring when asking whether music has ‘soul’. The illusion of soul, perhaps; a snapshot of it. But that doesn’t make it the real deal.
“The question,” Cope says, “isn’t whether computers have a soul, but whether humans have a soul.”
Emily Howell doesn’t tickle my soul - at least, not with what I’ve heard so far. However, even this is not the most important part of the question. So far, Emily relies on Cope for control, and to assess whether the results are any good. We have no idea what the success rate is, or how much intervention and selection is required. Even now, the output isn’t spectacular.
Liken it to the infinite monkey theorem; a tool like Emily can speed up the process, but we decide whether the output is correct or has artistic excellence, not the ignorant monkeys. The milestone lies beyond this, in generating works of art every time, and I’m just not convinced we’re even near that yet.
A threat to contemporary composers? Only when the results are consistently good without the need for manual selection. But even if Emily accelerates the creative process as David Cope indicates, this should be seen as a step forward for everybody. After all, nobody’s complaining that the Industrial Revolution happened.
