Between Sound and Science

Between Sound and Science

The more I hang out with scientists and engineers—and this seems to happen more and more often these days—the more I feel like an incorrigible composer. No matter how much knowledge and lingo I absorb, it sometimes seems that our goals or areas of concern are fundamentally different.

Written By

Isaac Schankler

The more I hang out with scientists and engineers—and this seems to happen more and more often these days—the more I feel like an incorrigible composer. No matter how much knowledge and lingo I absorb, it sometimes seems that our goals or areas of concern are fundamentally different. As a staunch proponent of collaboration between the arts and the sciences, this bothers me a great deal, and I’d like to get to the bottom of it if I can.

There are certainly artificial barriers between the two domains, built up over the years by mistakes and misconceptions. I’d like to place most of the blame for this on pop science journalism, usually created by people who seem to know little about science or music. For just a couple of recent examples, here’s one about pop music getting sadder and here’s one about the distribution of chords in more than 1300 popular songs. I’ll leave it as an exercise to the reader to properly deconstruct these. Okay, fine, a couple prompts: 1) Are “Paperback Writer” and “Help!” really happy songs? 2) What kind of insights could be gained from an examination of the distribution of words in more than 1300 popular novels?

I’m being a little flippant here, but the point is that most musicians are probably less dogged and stubborn than me, and have better things to do than nitpick an argument, or look up primary sources to see how they’re misrepresented. So musicians tend to believe that this is representative of how scientists look at music, resulting in a great deal of skepticism and mistrust.

Conversely, researchers in music often ignore music theory on the grounds that it’s not rigorous or verifiable. But many of these researchers have not experienced firsthand the explanatory power of theory for all kinds of musical events, and spend much of their time developing sophisticated methods to reverse-engineer theoretical understanding, occasionally with very strange results. For example, automatic music genre classifiers that do well on certain data sets can be thrown off by small tweaks in equalization, suggesting that they are paying more attention to surface features like production or mastering, and not what we actually hear when we distinguish disco from country. A little music theory here could go a long way; just because it wasn’t created with scientific research in mind doesn’t mean it can’t be incorporated into a scientifically rigorous model or experiment.

Even if we bridged this gap, however, I’m still not sure musicians and researchers would see things the same way. Geraint Wiggins argues persuasively that scientists are in fact creative, but I wonder if they’re a different species of creativity. Put simply, scientists are interested in directed creativity that tackles a particular research problem or goal, whereas artists are interested in exploratory creativity where the destination is much less certain. These are not hard and fast categories, and there’s certainly some bleed-through; it sounds like Wiggins believes that scientists could stand to be a little more broadly imaginative, and artists would never finish anything if they weren’t narrowly focused at least part of the time. (This comes through in writing style, too; I can never seem to get into the idea of stating my thesis up front, preferring for it to develop gradually over the course of several paragraphs.)

Perhaps this is the greatest benefit to be had from collaboration between scientists and artists, this productive clash of perspectives. Even now, though, it’s hard to imagine what this collaboration should ideally look like. Should scientists be technicians, making cool stuff to artists’ specifications? Should artists be subjects, providing data for scientists’ publications? Both of these approaches are valuable and valid, but neither of them builds respect and trust. Should scientists and artists educate each other until we’re all comfortable in both domains? If so, are the end products of these collaborations still research projects and/or works of art? If they are something else instead, what are they? I’d be curious to find out.