Some sixty-one years ago, the American philosopher Willard Van Orman Quine wrote a famous essay, “On Simple Theories of a Complex World.” Actually, referring to this as a “famous essay” is a tad redundant, since Quine is one of those people who only ever wrote famous essays. But setting that observation (bordering on sour grapes) aside, Quine goes on to observe the difficulty in saying just what does qualify as simplicity. He further observes the legitimate psychological and formal reasons while theory builders so ardently crave simple theories: the simpler the theory, the more readily it can be employed in our various cognitive activities. Of course, too simple a theory leaves us with no purchase on the world what-so-ever. “God willed it” is about as simple a theory as you can come up with, but it is also as singularly useless a theory as anyone could ever imagine; it provides absolutely no insight, a complete absence of predictive power, and only an illusion of emotional comfort for those readily distracted by vacuous hand waving.
Quine was writing more than a decade before the emergence of computational complexity as a sub-field of abstract Computer Science, in which upper and lower bounds for kinds of complexity (and thus, conversely, forms of simplicity) was even formulated. But we do now have a variety of ways to address Quine’s concerns about how to characterize complexity and simplicity. I’ll say more about this in a moment. What I want to start with a more controversial proposition: Namely, Quine got it backwards. In a very real sense, it is the world that is fundamentally simple and our theories that are complex.