Intracranial tectonics

Or, what I think people are talking about when they equate thinking and writing

Confession: I am the market for articles providing writing motivation and tips. Anyone sharing their writing process and talking about what writing has done for them gets my attention — especially if they end up on Hacker News. Only a small part of the enjoyment I get from reading these articles comes from discovering tips that I could actually apply to myself. Instead, I’m mostly interested in watching how people have managed to convince themselves to spend hours at a stretch manipulating symbols like bacteria creating plasmids.

What I don’t understand, though, is this idea that’s arisen, again particularly among the Hacker News set, that writing is a tool for revealing your own thinking to yourself. The clearest nidus I can come up with for this idea is Paul Graham. I love his essays, and I love how opinionated he is about writing, but when he says things like, “There is a kind of thinking that can only be done by writing,” it’s just puzzling to me. It’s not just PG either, but others in that orbit whose writing I also admire. Rohit from Strange Loop Canon says, for example, “Writing to me isn’t separable from thinking.”

Here’s what I imagine they mean by this: They have an idea for an essay. They sit at their computers and begin writing about what they think that they think about that topic. But as they do so, they discover that their thoughts are incomplete or subject to some hitherto unnoticed error — and then, through the act of writing, they correct these deficiencies. It seems like this process would be totally aberrant to most essayists, especially of the past, who tended to frame their writing as a report of a train of thought they’d had, almost like a travelogue, the writing of which is at least ideally a completely transparent container.

Precisely, I part ways with the equation of thought and writing because I find the two activities completely impossible to do at the same time. When I approach a piece of writing, I have to have at least a loosely formed idea of its structure before I touch fingers to keyboard, otherwise I’ll end up with absolute garbage in the best case scenario. More likely I’ll just end up stuck. (The old advice about writing a breezy, judgment-free first draft and then editing in subsequent drafts has never worked for me.) Within a piece of writing, too, I have to know what each paragraph does and where I need to end up in my argument to be ready for the following paragraph before I can start writing it.

That said, I still find writing to be a great occasion for thinking. PG, again, talks about how writing provides you with an opportunity “to pretend to be a neutral reader who knows nothing of what's in your head, only what you wrote. When he reads what you wrote, does it seem correct? Does it seem complete?” This is very true. When I’m able to get distance from something I wrote and try to follow the argument as a stranger might, it’s an opportunity to notice places where I’m not being explicit enough or, more likely, where I haven’t even thought through something enough to realize that there’s a leap needed to get from A to B.

But what do I do when I make that discovery? Step one is recognizing the deficiency. Step two: reading others who’ve thought about the issue. In step three, both of my brain cells resign in protest, but step four is where things get interesting, because it’s usually where I start playing a video game. Really what I need at this point is a way to occupy my hands and my attention just enough that my subconscious can start chipping away at the problem. This is why I can’t pick up a new game — it has to be something I’m pretty familiar with.

(Right now, the game is Tears of the Kingdom. Sudoku also works. Prior to Breath of the Wild, which I only started playing well into 2020, it had been the Civilization series, but I decided to uninstall it when I saw that I’d played some horrifying mid-three-digit number of hours. To be clear, these games also beckon me when I’m not trying to solve problems, otherwise I wouldn’t have an issue with appalling amounts of play time.)

This is not an entirely reliable process, but it seems to work well enough that I can do it even under a deadline. It does take time, though, which is the mysterious thing to me. What is my brain doing when it’s solving these problems?

I see two alternatives:

1. My brain is generating and testing permutations of the problem

2. My brain is waiting for me to achieve some optimal state of mind that will allow the solution to come

As an example, here’s a problem I was trying to solve late last year. I needed to devise an analysis for testing the effect of prior COVID-19 on disease severity. The previous studies I was looking at seemed to run smack into immortal time bias, meaning that the people most susceptible to severe disease — people who died during their first recorded bout of COVID — were excluded from the reinfection cohort. I had put together an initial analysis plan, but realized that it was inadequate. So, I read some similar observational studies and some literature on causal methods, and started playing Hades.

It took at least a couple hours, but at some point I suddenly realized as I was playing the game that my approach needed to be to generate a counterfactual first infection severity for the previously infected cohort at the time of their second infection. This counterfactual could be informed by the set of first infections taking place near the same time, but I couldn’t say anything at all about the counterfactual wherein the cohort of first infections were actually repeat infections. I needed to settle on generating an average treatment effect among the treated, in other words.

How did this realization pop into my awareness?

One possibility is that, as soon as I got sufficiently absorbed in the game, my subconscious created a model of the data and began pasting together simulated solutions from others’ insights that I’d read, all while testing whether the solution gave an unbiased result. There’s a biological precedent for this in our thymus, which generates permutations of T-cell receptors and tests them against antigens. I don’t know why this process would be recapitulated in the subconscious, though.

The other possibility — I don’t see these two as exhaustive, though, or even entirely exclusive — is that I have to achieve a certain mental state that allows the solution to come. For example, perhaps I need to shift into a right-brained-enough mindset that I can present the problem as a suitable metaphor to my subconscious and also interpret the metaphor it sends back.

Either way, this is the opposite of working through my thoughts in writing. Writing is more akin to arranging material that I’ve already procured, rather than going out and finding that material. With writing, though, the thought process is definitely based on permutations within a model: permutations of word choices and sentence structures based on a model of my reader. It’s medium-specific, whereas the thought process behind it isn’t. I don’t see writing, then, as an activity that is uniquely suitable for enabling the kind of slow, subterranean conceptual drift that produces ideas. It’s just the one that I’m most comfortable with.