Last time, I argued that revision is an OODA loop, and that the step most writers skip is Orient — the move where you frame what you’ve read against what the passage is actually for, before you change a single word. That’s easy to assert and harder to show. So this dispatch is one edit from my own work-in-progress, traced through all four stages, start to finish.
Here’s the part that should bother you before we even begin: the paragraph I’m about to fix was already clean. A line editor would have read it and moved on. The best edits often look like that — invisible to the kind of close reading most of us call editing, and visible only once you ask what the passage is supposed to do.
A little context, because the passage comes from a system-driven science-fiction novel and the narrator’s voice is doing something specific. Cade is an operative. His narration is clinical by design — he reads people, files what he finds, and moves on. There’s a hard rule in his voice: his emotions are never named directly in the prose. When he feels something, the text shows it in behavior, never in a label. Hold onto that rule; it does the heavy lifting later.
The scene: Cade has just sprung the trap he spent the whole book sensing. He’s extracted the person he’s guarding, and now he has to seal a woman named Mira out of the safe room — Mira, whom he failed to read for two hundred pages, who may have betrayed them and may simply have been used. Here’s how the chapter ended in the draft:
Before anyone could protest, I hit the close button, then set the security seal.
One anomaly dealt with.
Read that as prose. It’s fine — two clean sentences and a hard, terse button on the chapter. And that’s exactly the trap.
Observe
Read what’s on the page, not what I meant to put there. The sealing happens as a sequence of operations: close button, security seal, one anomaly dealt with. Mechanical. Efficient. In character for a man who files things.
And — this is the important part — there’s no prose friction here. Nothing snags. If I were running a line pass, scanning for weak verbs and clutter and repetition, this paragraph would sail through untouched. The signal that something is wrong doesn’t come from the words. It can’t, because the words are good. It shows up only at the next stage — and only if I bother to run it.
Orient
Frame what I observed against the structural model. Two facts about the larger book meet at this paragraph.
First: this is the climax of Mira’s entire arc. The book has been one long failure to read her — every chapter, Cade catalogs another anomaly and can’t resolve it. This is the moment that failure stops being a question and becomes irreversible. He seals her out, and he still doesn’t know if he’s right.
Second: the voice rule. Cade’s professional distance is built to crack in exactly two places across the manuscript, and this is one of them. The crack is structural. It’s planned. It’s supposed to happen here.
So the passage’s job is not “end the chapter.” Its job is to register the human cost of an irreversible decision, at an arc climax, inside a narrator who is forbidden from saying what he feels. Hold the draft up against that job and the problem appears instantly: One anomaly dealt with discharges the most important emotional beat in the book as a piece of paperwork. The arc lands structurally — the seal happens, the consequence is real — but the cost never reaches the reader. The beat is doing its plot work and skipping its emotional work, and the cleanliness of the prose is what hid that from me. The diagnosis came from naming the job — not from reading the sentences.
Decide
Once Orient is done correctly, the decision is usually fast — and, just as usefully, constrained. The framing tells me what I’m allowed to do.
I can’t name the emotion. The voice rule forbids it, and breaking the rule here would be louder and cheaper than the moment deserves. So the cost has to be carried by behavior, not statement. I considered the alternatives — an italicized line naming what the seal cost him, a closing beat that spelled out the failure — and they all broke the voice by over-naming the very thing the character won’t say. The disciplined choice is the smallest one the voice permits: a single physical beat, placed between the seal and the file, doing the work the label isn’t allowed to do.
Act
Before anyone could protest, I hit the close button, then set the security seal. My hand stayed on the panel a moment too long.
One anomaly dealt with.
Nine words. My hand stayed on the panel a moment too long.
And look at what those nine words do to the line that was already there. One anomaly dealt with used to read as cold — a man closing a file. Now it reads as a man composing himself: the hand betrays him, holds a beat past function, and then he files the feeling away, because filing is what he does instead of feeling. The terse line stopped being coldness and became control. I didn’t just add a beat. I changed the meaning of the sentence underneath it without touching a word of it.
That’s the loop closing on itself — the new sentence became new evidence, and it rippled backward through prose I’d already decided was finished.
Where This Gets Complicated
The honest caveat: you cannot run this loop on every paragraph. If I stopped to ask “what is this passage’s job at the arc level” on all of them, I’d never finish a draft, and most paragraphs don’t reward the question — they’re connective tissue, and a clean read is genuinely all they need.
The skill is knowing which paragraphs are load-bearing. This one was: an arc climax that doubled as a designated crack in the narrator’s armor. And here’s the tell I’ve learned to watch for — when a paragraph that’s doing structural work goes down easy, the easiness is suspicious. A clean climax, a frictionless turn, a final line that gives you no trouble at all: at the moments that carry weight, smoothness is exactly when to stop and Orient.
I’ll concede the other half too. A writer with the arc fully alive in their head might have written the hesitation the first time and never needed the loop at all. I didn’t — the draft sealed Mira out and moved on, and the loop is what caught it.
The Takeaway
A line-read asks one question: is this sentence good? The loop asks a different one: what is this sentence for? At the moments that matter most, those two questions have different answers — and the gap between them is where the edits live that no amount of careful reading will surface, because they were never a problem of the words.
That’s what the last dispatch meant by naming what a passage is trying to do before you touch it. The proof isn’t the nine words I added — it’s that naming the job was the only thing that told me to add them.
So the next time a clean paragraph is sitting at a moment that matters, don’t trust the cleanliness. Ask what it’s for. Sometimes the answer is nothing — it’s doing its job, leave it. And sometimes the answer is the most important beat in the book, hiding behind two good sentences and a hard, terse close.
Until next time, thanks for listening.
Leave a Reply