👋 Hey Friends,

Some kids grow up wanting to be firefighters, astronauts, or teachers … I wanted to be Laplace’s Demon. Let me explain.

Pierre Simon Laplace hypothesized:

if we knew both the laws of physics and the location of every particle in the universe, we would be able to predict everything that would happen in the future. — Paraphrasing Pierre Simon Laplace on Laplace’s demon

Laplace attributed these powers to an imagined demon with almost infinite intellect. That’s what I wanted to be. Oh the ego on that kid.

Over the last 3 years I’ve wrestled with two of the pillars of my world-view: Reductionism and Determinism. From an angsty highschool kid who saw the world in black and white, to my recent research into more human aspects that have sprinkled in a *touch of grey *between the black and the white. Let’s chat a while on Computer Science ego, reductionism, and where to go from there.

🪴 My Reductionist Roots

Reductionism is the belief that complex phenomena can be explained by reducing them to their fundamental parts or components. It is a way of thinking that has been influential in many fields, including science and philosophy. — ChatGPT

Thanks ChatGPT! One day you’ll replace us all 😊

So that’s the foundation I built my world view on. And how could I not? It feels so scientific.

To understand why this view was so seductive, you’ll have to learn a little about who I was in highschool. I was, well, a pretentious prick. I had convictions about the way the world was, which beliefs were held by smart people and which ones by the … other people. Among the good beliefs was that real science (read STEM) could explain everything. Leaving anything to the supernatural, the spiritual, or god forbid, the ambiguity of sociology and philosophy was an admission of naivety.

I was above all that. I believed in the cold hard facts and physics was the ticket to those facts. Most importantly, I believed all the world could be reduced and explained by its components and its processes. I was a hardcore reductionist supported by my scientific heroes: Neil DeGrasse Tyson, Richard Dawkins, Carl Sagan and so on

Have I painted an adequately science-bro douchebag picture yet? Great. I’m in pain just thinking about it.

💽 Computer Science and Its Perfect Reductionism

So, naturally, that pretentious kid went into Computer Science and minored in Mathematics — two realms where reductionism is not just effective but beautiful. And I really do mean beautiful.

In science, there’s a tendency to view a theory as more elegant the more phenomena it can explain with as small an ontology as possible. Sound familiar? Einstein, Feynman, and so many of the greatest scientists took this stance: elegance = simplicity = reduction. It’s baked into scientific culture and it was baked into me. Nowhere did I find this more poignant than in Computer Science.

There’s two homes of reductionism in CS (at least as far as I can tell) — an applied sense and a more philosophical sense. In the applied sense, algorithms and codebases are more elegant when they’re shorter and more efficient. Makes sense. There’s practical reasons for this field-wide drive toward code minimalism. Software works better and is easier to work with when it’s written that way.

What got to me was the beauty of understanding computers through reduction. Let me give you a probably mostly accurate example.

I’m writing this article on Notion. Notion is software that some team wrote in some programming language.

💡 That’s one layer — user interface → programming language.

That code gets compiled or interpreted into some more basic language that the computer can execute.

💡 Human legible code → machine legible code.

That machine legible code is then processed as a series of operations or instructions

💡 Machine legible code → Computer Operations

The Computer Operations are carried out by hardware — the CPU, GPU, Memory, etc.

💡 Computer Operations → Hardware carrying out those operations

That computer hardware is built up by transistors, wires, circuit boards, etc.

💡 Computer specific Hardware → simpler hardware

That hardware is designed and explained in physics and mathematics

…

You get the picture.

So there’s this gorgeous reductionism that’s deeply embedded in Computer Science. It moves from the every day software you and I use all the way down to the physics of electricity. I came to love it … probably because it reaffirmed my zealously held beliefs. While most students hated assembly language (a low level language close to what the computer understands), I dove headfirst into it. It made me feel closer to the true workings of my computer.

❓ If Computers work this way, everything should … right?

Oh, CS-ego. You seductive silver-tongued bastard.

Everything I’d known and been successful at reduced. And every time I reduced things further, I understood more and was given a gold star for my collection. How could I not believe the whole world worked this way? My self-worth depended on it. If physics and computer science — homes to the most challenging problems (self-aware sarcasm intended) — advocated for a reductionist philosophy, why shouldn’t that apply to every other domain?

So I did just that. I applied reductionism everywhere. Worse than that, I believed that every hard problem should be reduced to physics as the underlying explanatory framework, and computer science as our engine for creating solutions to them. We see this kind of thinking in the crypto-bros who believe that financial and governing systems, flawed but fundamentally human systems, can be fixed or replaced by code. The arrogant maxim — *Code is Law — *succinctly captures CS-ego.

➕ Greater than the sum of its parts

It would be possible to describe everything scientifically, but it would make no sense; it would be without meaning, as if you described a Beethoven symphony as a variation of wave pressure.” — Albert Einstein

In the last 3 years, my work has shifted into a field called Human-Computer Interaction. Note the human part comes first. Specifically, I build tools to improve how students learn computer science. This is where my beliefs have been most ardently challenged.

I started with the reductionist question: What are the processes that make up expert programming? I wanted an essence of problem solving so I could teach it to students. 2 years later and dozens of papers deep, I’ve found nothing so succinct and satisfying. As it turns out, people can’t be reduced to vectors of attributes. There’s no set of properties or behaviours that perfectly captures an individual. There’s no platonic form, no essence of what it is to be human or to solve problems like one. Not yet at least. Yet everything I had learned up until this point pushed me to find it. I yearned for simplicity. I wanted to reduce human interactions like I did computers because that’s where I felt I’d find success.

As a disclaimer, there’s plenty of brilliant work on this topic. Insight, transfer appropriate processing, computational thinking. Many minds more experienced and intelligent than mine have been knocking on this problem’s door for many more decades than I’ve been around. But the problem isn’t solved.

That brought me to a deeper question. Even if it were solved, even if we could reduce the ways we solve problems — the ways we think — down to the firing of neurons, the neuro-chemistry, the physics of those electrical impulses … would that be the best way to make sense of it all? Like Beethoven’s symphony, would it be meaningful to describe it in reductionist terms? I don’t think so. Useful? Probably somehow. Meaningful? Probably not.

Recent movements in physics push for non-reductionist approaches too. I could try and likely butcher an explanation of dynamical systems theory but I won’t. The important part is, even the most steadfast reductionist fields are embracing analyses at multiple levels. No matter where I look, the whole seems to be greater than the sum of its parts.

👿 Now What?

We built computers from the ground up. Layer after layer of abstraction we compiled ways to make computers easier to use and develop on. Of course it reduces — we made it that way. Intentionality is not so easily found everywhere else. Whether you believe in a God or not, the clean deconstructible designs we build into our systems aren’t as obvious, if they’re there at all, in the mind or the universe.

Am I saying that reductionism is dead? That all things should be described by the systems they’re in or maybe, not at all? Nah. What I am saying is that maybe, our views of the world should be a little more nuanced than those of a 16 year old with a hard on for his Laptop. Maybe there’s a little room for ambiguity, for nuance, for art and poetry. Maybe the whole world doesn’t have to sit neatly in categories. Maybe it’s not all just particles bouncing about. Maybe I’ll never be Laplace’s demon predicting the universe’s next steps with my infinite intellect. Maybe I just have to be okay with that.

---

Cheers,
Nathan Laundry

✉️ Join my Email Newsletter #GuidingQuestions here