Posthuman Collective Composer

The video above is from one of my son, Aodan's, favorite DVDs. It's an Animusic DVD that Charissa, my music teacher wife, brought home for my boys to watch. The Animusic series is a collection of cartoon music videos featuring machinery of various types playing music. Some of the machines are better-mousetrap style contraptions, others are like giant wind-up toys. The one above features robots on a spaceship playing percussion instruments. As we watched the DVD yesterday (and Aodan drummed along on a toy drum), I remarked to Charissa that I was uncomfortable with the possibilities this particular video explored. I am uncomfortable with the extreme posthuman (or maybe dystopian) theme of machines playing musical instruments without apparent human involvement (this is the kind of conversation I routinely subject my poor wife to). Despite my interest in digital literacies/cultures and computer mediation of human literacy habits, I want artistic sensibility and aesthetic to belong to humans. Humans make art. They may use technology to do so (of course they do; musical staffs and symbols are a technology after all), but it is humans that control it in order to turn sound into art.

Then today, a friend and colleague of mine posted a Gizmodo article about a scholarly journal article by Imperial College-London researchers published yesterday in PNAS called Evolution of music by public choice. This study essentially studied the evolution of sound into music in order to compare it to models of evolutionary biology. In this study, researchers basically (I'm skipping important steps for the sake of space. Go read the article) started with clips of randomized noise and allowed people to rate the clips according to musical quality. The top rated clips were then combined in a semi-randomized "genetic" style creating new clips, and the process repeated itself. Over time, the clips begin to sound like recognizable beats, then melodies, then they finally become relatively complex and interesting musical strains (I listened to all the published examples. I encourage readers to at least listen to the commentary and overview offered on the Gizmodo article. It's pretty amazing).

The idea of the study was to study consumer input in the evolution of musical aesthetic by isolating it. In other words, there was no human writing the music: no experimental artistry by a person trying to play with old conventions, no producer looking for a hook, no band members wanting a solo. In removing the factors these factors so they could look at data relating only to the issue they were studying, researchers also removed the people that are typically associated with making music--that is, the composer, lyricist, and producer. Instead, these processes were automated.

Historically, music is considered a humanistic art form, not because it touches a human audience, but because it is created by a human composer. This music, then, is remarkable because it doesn't have that singular person we typically associate with musical composition. But I think the Gizmodo article oversimplifies when it says that "it's possible for digital music to evolve by itself, without creative input from a composer." It is certainly the case that there is not a single human composer. By getting rid of this figure, this experiment abolishes the concept of the genius artist, individually achieving a transcendent artistic artifact. Instead, there is something decidedly posthuman in the composition of this piece.

When I speak of the posthuman here (a term which involves a kind of spectrum of thought), I'm thinking along the lines of Donna Haraway's "cyborg." This involves the idea that the separation between our selves and the objects we make/use is, as it turns out, rather blurry. As we evolve our instruments, those instruments evolve us as well (this idea of "man as ongoing process" is the central theme of posthumanist theories). It is this interplay between our selves and our technology that make us "natural born cyborgs" (Clark). We are, by our nature, part human, part tool.

The DarwinTunes are composed through the interplay between tool (a computer designed by humans running an algorithm programmed by human researchers) and active human beings (the people voting on which clips move on).

Not only is the music produced by DarwinTunes posthuman, but it is also an example of another important element of digital culture in that the human half of the composition process is completely collaborative. Just as the singular artistic genius is replaced by a computer program, so also is he replaced by, not one musician, but thousands of consumers, all of which bring to the process their own cultural histories (in the form of chord progressions, dissonances etc. which seem "natural"), personal aesthetic sensibilities, and so on.

Perhaps, then, the most surprising thing about the DarwinTunes is that, after about 500 generations, they start to sound pretty good.  Pieces of music composed, not by an individual or small collaboration of talented artists, but by a process of negotiation between a piece of technology and collective intelligence may, after enough generations, turn out to be as complex and sophisticated as any experimental piece by Philip Glass. This brings into focus one of the fundamental questions we begin to ask when studying digital culture: Just how important is the "expert/genius/author" after all?