News You Can Bruise for 2008 June 30

(2) Mon Jun 30 2008 23:31 Serious Review of "The Most Unwanted Song": Rather than just chortling over its existence, as often happens. You can hear the song here among other places. My thesis is that the song fails because it conflates two different kinds of badness. I guess I'm equating "badness" with "unwantedness" when they actually have a complex relationship, but whatever. Here's the homepage in case you're not familiar with the project.

I think one important aspect of badness is irreducibility. A bad thing with something good about it is generally not as bad as the same thing with the good part removed. "The Most Unwanted Song" contains many different aspects (each disliked by lots of people), and they're not all combined into one piece but arranged in a "Fingertips"-like series of mini-songs. This raises the possibility that any given person will like part of the song, effectively carving out a smaller song that they like. I believe this is what happened, and this is why the reaction to the song is more positive than you'd expect from the parameters of the experiment.

I decompose TMUS into three major songs. Let's call them "Opera Rap", "Frank Zappa Goes to Town", and "Kids Can't Sing". Only the third song is really unredeemably bad. I was enjoying the song until around the 9 minute mark when "Kids Can't Sing" started in earnest. If the song had stopped there it would have been pretty good and not particularly Unwanted.

"Frank Zappa Goes to Town" is a series of decent Zappa-esque noisefests, I can see how some people could take or leave it. But "Opera Rap" is really excellent. There's something genuinely beautiful about hearing a soprano sing to a rap meter. And just as the cowboy-themed lyrics start to get old she busts out a verse about Wittgenstein. Really, you wouldn't write those lyrics unless you wanted to hold people's attention.

So the problem, and I think this may have been the point of the project, is that badness comes from hackwork and not from a bad choice of attributes. Songs about cowboys aren't bad per se; they tend to be bad because many people like songs about cowboys so much they suspend their critical faculties. Commercial jingles and elevator music are bad because they're forced on you: the people who wanted the song to be created don't consider themselves part of the audience. Children's songs tend to be bad for the same reason, and also children haven't developed their critical faculties yet.

If you ask people what they don't like in a song they'll give you a bunch of heuristics, but treating those heuristics as constraints won't give you a bad song because constraints are a spur to creativity, bane of hackwork. At Viable Paradise I had to write a story in the genre I most hated. I wrote a bodice-ripping vampire romance that's really embarrassing but it's a pretty good story with some emotional depth. I don't like vampire stories or romance stories because I think there's a lot of hackwork there and it's not worth it to me to filter it out. But given those things as constraints I can work within them. (Actually now that I think about it, I should have written a novelization of a movie, because I hate those even more.) TMUS is bad not where it deploys some genre or instrument that everybody hates, but where it feels like hackwork.

This is also why "The Most Wanted Song", while listenable, is not very good. People are more similar in their likes than in their dislikes. The heuristics were too close together, which made it very easy to crank out some hackwork. It's much more difficult to come up with a good smooth-jazz love duet than to come up with a good operatic rap about the profession of cowboy. That fact probably surprises a fair number of people but I doubt many of them read this weblog.