The Structure of Scientific Revolutions is one of those perennially recommended books for anyone interested in the philosophy and history of science. Structure’s language of “paradigms” and “paradigm shift” has been so thoroughly appropriated and bastardized in business book circles that encountering that language in its original context prompted a little bit of an eyeroll from me. But the ideas here really are worth thinking about, and they were one of the first steps I took down the road that has led to Quasioptimal.

Because I already had a sense of the direction of the argument and the conclusions it would come to, I decided this book would be best to consume as an audiobook. I was doing a solo 11ish hour trip from Pittsburgh to St. Louis on a hot August day in my little silver Toyota Corolla, and I plowed through most of Structure along the way. I went for the Dennis Holland narration, which was neutral and therefore unmemorable.

To write this annotation, I’m drawing on a couple of summary sources as well as some sparse notes I made at the time.¹

Physics Can’t Be Your Metaphysics; or, There Is No Perfect Theory

I came into school with a generative mindset toward theory. How can we lay down new ground? How can we connect theories together? How can we explain everything? I wanted to work either in theoretical physics or theoretical economics, and I chose economics because the potential value added seemed higher.

I remember feeling distinctly uncomfortable as I sat in my ECON 2 lectures in Freshman year. The temperature of the room was pleasant, and the chairs in which I sat were usually good enough. My discomfort was distinctly mental. The models being used were a bit too simple, and surely we could do better. How could all of these brilliant economists not realize that these models were grossly oversimplifying? And might not all of the issues in economics be solved by simply including more factors in the models? Surely we could do better.

I look back on that view as rather quaint and naive now, and reading Structure was the death knell for the last vestiges for my hope that “objective reality” would be computable or describable in through any kind of neat or systematic theory. Before it, I really was looking looking for a perfect theory, one with perfect explanatory power. Afterward, I grew more sensitive to the models and “paradigms” I was relying on. I began to notice how these models elide certain details of reality, how they interact to give an incomplete picture that feels dangerously complete from the inside.

By the time I left school, my orientation toward theory was one of profound skepticism, especially at the boundaries. If a perfect theory is too much to hope for, then we must think seriously about what theory has to offer and what its limits are. It’s almost obvious when you say it aloud, but this really changed how I approach thinking and reasoning at a deep level.

A Certain Conservative Sensibility

Structure takes a position that describes “normal science” not as the quest for a perfect model, but as “puzzle solving” within an accepted model that is good enough. Good enough is, in usual times, good enough.

But, as Kuhn shows through historical examples, and as I felt distinctly in my introductory economics course, our models and theories tend to reify and even replace our perception of reality if we spend too much time operating inside of them. We become so accustomed to our models that, when we encounter evidence that they are no longer good enough, we impulsively dig in our heels. We want to “[stand] athwart history, yelling Stop” when faced with a new model or explanation that threatens our old, familiar ones.²

Kuhn’s diagnosis of this conservative impulse among practicing scientists really stuck with me. It’s just intuitively true: we can all recognize it in ourselves. Models are sticky, and we feel a certain resistance to updating our beliefs. Structure left me feeling a certain anxiety about avoiding this state of entrenchment, which I think is a particular danger for someone with my inclinations toward conceptualizing and philosophizing. That anxiety has made me hypervigilent: one way of phrasing the entire endeavor of quasioptimal might be a kind of hypervigilence around models, concepts, “paradigms,” and ways of knowing.

The Path-Dependence of Thought

Structure prompted me to think more seriously about the history of thought. In particular, it got me thinking about how the notion of rational scientific enquiry is distinctly inflected with Western biases and makes a lot of invisible assumptions that are historically contingent and open for debate.

As I’ve mentioned many times in this space, I got my start as a physical reductivist steeped born of a blind scientism. In my teenage years, I was that hubristic sort of interlocutor who would frequently appeal to “objective reality” and reject outright any appeals to the authority of tradition.

Kuhn’s work forced me to confront just how dynamic scientific understanding is, and just how much of an “authoritative tradition” science really is. You can’t make an appeal to “objective reality” based on science in good faith if, as Kuhn claims, the development of science is not simply the filling out or resolving of a stable representation of reality.

In Kuhn’s account, the “filling in the picture” metaphor describes the “normal” or “puzzle-solving” phase of the scientific process. It’s no accident that that was what I imagined science to be: that portrayal is dominant in popular representations, and I’d never really lived through a big, paradigm shifting “revolution” in physics, which the discipline with which I was best acquainted.

1. Summary Sources: Wikipedia and Stanford Plato ↩

2. William F. Buckley, Jr. (1955), “Our Mission Statement” for National Review. This is an absolutely fascinating document to read in 2023 because of the insight into the deep intellectual and emotional currents underpinning the American right to this day. There are sentences in this piece that could be taken out of a Tucker Carlson vamp. ↩