a meander about writing and language

Recently I wrote about how an essay I read, 28 Slightly Rude Notes on Writing, made me want to write more. The post didn’t have a really sharp thesis, and it didn’t have an intended audience. This flew in the face of I’d been led to believe is “good writing,” where it’s important to know your audience, and to write for that audience, and to have an objective, and to focus your writing such that it attains that objective concisely. But in attempting to find links to justify my claim that this is what’s commonly considered to constitute “good writing,” I found a different source saying “if it’s not professional, it’s fine to just write for yourself,” so: I didn’t have an audience or an objective for that last post and I’m sure as hell not gonna have one here.

But. But but but. When I say I want to write more, but I don’t have a topic or an audience in mind, what is it I actually want? The phrase that leaps to my mind is “putting words together.” It is—or can be—a fun, brain-stretching challenge to choose words, assemble them into sentences that flow smoothly (or clunk joltingly), stuff sentences full of digressions in subordinate clauses. To cut long sentences with short ones. To wrangle paragraphs, each with an arc that feels inevitable, into a journey that goes somewhere surprising, or at least a little interesting.

I’m not a professional writer and I don’t claim to be great at this; it’s still a fun challenge, though. As with all art, the space of possibilities is too large to navigate any way but intuitively, and I’ve done enough reading and writing to have some trust in my intuition.In the middle of writing this post, I came across the aphorism “intuition fails in high dimensions.” Because I know just enough math to be dangerous, I tend to think of a lot of things as high-dimensional spaces, so I was somewhat gratified to see that the phrase refers specifically to machine learning.

But! But! The question remains: what’s the point of writing if it’s not about something? What does “writing for its own sake” actually look like?

The train of thought that led to this post occurred while I was in the back seat of a car returning from a family vacation, and somewhere in my musing I looked out the window at the sky and had what felt like a slightly sophomoric thought: how could a piece of writing be like a cloud?

Bear with me. I mean like a cloud in that it doesn’t have a purpose or a goal: it’s just there, doesn’t need to justify its existence, doesn’t mean anything. Let me get the obvious out of the way first: it is not possible to write something like this. There is too much meaning and human intentionality in language to use it to create something truly meaningless. But still. What kinds of writing might come closest to this? A few things come to mind:

1. Aleatoric writing, by which I mean writing produced in part by chance. This approach to writing poetry is not uncommon. Legend has it that many of Radiohead’s Thom Yorke’s lyrics for Kid A and Amnesiac were “written” this way, by pulling words or phrases out of a hat (whether the hat is literal or figurative is not known to me). Legend further has it that the bridge of System of a Down’s “Chop Suey!” came from Serj Tankian’s pulling a book off Rick Rubin’s shelf, opening to a random page, and using the first line his eye fell on.I confess I personally find it a bit farfetched that Tankian just happened to flip to what is probably one of the Bible’s most recognizable turns of phrase, but the point is that this is a done thing.
   
   

   This kind of writing dates back at least to the Dada art movement, as documented in Allison Parrish’s “Language models can only write poetry,”Allison Parrish has thought a lot more about the kinds of things I’m talking about in this post, and she’s written about them more sharply than I could hope to. I really recommend reading this.
   
   which has an example formed by the “words from a hat” method (and argues that the large language models of 2025 are a sophisticated version of same). Even with this method, it’s impossible to excise intentionality from the process entirely: you need to choose the words that you put into the hat. Even if you take them out of a random book on your shelf, you still (presumably) employed your own judgment in acquiring that book; even if you use a book you don’t own, it’s probably in your native language, and so forth. And with a poem constructed this way, we kind of can’t help but try to ascribe meaning to it (well, I can’t, anyway). But still, there’s something cloudlike about a poem constructed this way. After all, clouds aren’t completely random there are, after all, physical constraints on their formation that just lack anything we’d call “intentionality.” And a viewer can still ascribe meaning to a cloud. But there are a couple other interesting cases I wanted to explore.

2. Zen koans. My (Western layperson’s) understanding is that many koans are self-contradictory, paradoxical, intended to thwart the reader’s understanding. Of course, even acknowledging that they are “intended” to do anything means recognizing how they are un-cloudlike. Text that refuses its own meaning does not become meaningless; the refusal is layered on top of the meaning rather than nullifying it. But there is still a cloudlike unconcern to a koan. It may be intended to help guide the reader towards enlightenment, but it’s not trying to convince. I think a koan is less like a cloud than a poem drawn from a hat. But it seems to have a similar indifference to the reader’s comprehension.

3. Haiku. I get the sense that they are generally sort of intended to convey something about the natural world: often sort of a snapshot in time, always self-contained, rarely “meaning” anything. Like koans, I do think haiku are intentful, but the best of them do seem to float like a cloud, to have been pulled from the air rather than constructed.

If texts from east Asia spring to mind more readily for me here, I suppose it’s because many seem less intent on persuading or informing than Western work. I may also have Ken Liu’s translation of the Dao De Jing on the brain, having read it recently. Much of the Dao De Jing has the flavor of Zen koans, to me: another recent reader described its tone well as “deliberately aggravating and provocative, but in a gentle way.” But I bring it up here because in some of his interstitial commentary, Liu mentions the inadequacy of language to explain the Dao and how Laozi basically just wrote it down because his followers asked him to.

I have other thoughts about language churning around and this post didn’t get to half of what I wanted to set down, but it’s long enough already. So I’ll leave the rest, ambitiously, for a future installment.