Why should it have to be elegant?
One of the best questions to ask yourself (and others) is:
what is the last (meaningful) thing you changed your opinion on?
Over the past few months, my latest answer to this question concerns whether or not explanations should or even need to be elegant. And what do we gain or lose by valuing elegance in explanation?
It was not until recently I fully appreciated the extent that my training in music and the humanities valued simple and elegant explanations for phenomena we hope to explain (I am sure this also happens in parts of the sciences, but this was not the majority of my formal education.) Off the top of my head, I can think of several papers that I read in grad school where some complex problem was “solved” (or a solution was proposed) where it felt like the answer came in the form of “so simple this whole time”, “right in front of our nose”, “the master key that unlocks our understanding of X” or “A grand unified theory of a monolithic concept, underpinned by only a few central axioms”.
I started thinking a lot about this (in a more serious way) after the workshop I attended on music, computation, and neuroscience back in January where this was a clear recurring motive. At the workshop, we were discussing the implications of these large language models as they will eventually be applied to music (which was also happening during that giant dump of new AI generative models in January 2023) and someone brought up Frederick Jelinek’s quip that a language model’s performance improves every time you fire a linguist. Will we get the parallel quip for theorists and composers? (In looking for the source of this quote, I did find a very interesting article contextualizing that quote and some of these ideas )
Then this past week Reyna Gordon came to discuss all things music and genomics with the Music Cognition Group. This is even further removed from my areas of expertise, but just reading a primer paper and the beat GWAS paper again and thinking about how many moving parts need to be considered to do any genomics research (not just musicality) made it clear how impractical and stifling it would be at this point in this field’s progress to try to restrict thinking by things that are elegant and easily explainable.
As an anything goes scientific anarchist, I realized that my valuing and learned preference for these elegant solutions has limited my thinking on a lot of these topics.
And this all feels somehow related in the context of music research to David Huron’s prediction that if people who have traditionally been considered ‘music’ researchers do not continually adapt our thinking, we (people with this kind of educational training) need to be prepared to have people without music backgrounds take center stage for the next generation of music related research. One way I think we can get stuck in our ways is not continually questioning what we value in research and why. I have of course many other strong opinions on this topic as well.
I’m not trying to extinguish a preference for simple solutions, just say that I think I need to start to more seriously consider the other side of the complexity solution spectrum. In not doing this, I think I’ve limited my thinking, especially if I want to honor a commitment to anything goes and intellectual growth.