While Flannery O’Connor’s example suggests the value of holding on to and interrogating the kind of anomalous details that don’t do enough work in early drafts, there are also those cherished elements we resist cutting that should probably go.

Ben Lerner describes one such instance in his autofiction 10:04 concerning a short story of his, “The Golden Vanity,” that appeared in the New Yorker. He relates its conception and shares an extract of an early draft—“the prose I generated first, the kernel of the work”—detailing a plot to fake letters from famous writers that the protagonist hopes to sell to an archive. And yet, once the story is tentatively accepted, the editors at the New Yorker request “a major cut: to get rid of the stuff about the fabricated correspondence, the section I considered the story’s core.” Lerner is righteously outraged, refuses to make the edit, and withdraws the story in high dudgeon . . . only to comically climb down off his high horse when friends he trusts concur with the editors’ opinion.

What’s being described here, so frankly and self-deprecatingly captured by Lerner, is the way we cling to our old hypotheses. Those elements that come early in a story, that essentially enable us to write it, are often the hardest to give up. Indeed, 10:04 itself is in part a book about expanding “The Golden Vanity” into a novel—the very novel we’re reading!—during the course of which the faked correspondence is briefly restored as a plot point.

I’m inclined to think of elements like this as scaffolding—aspects of a draft as essential to its writing as a scaffold is to the construction of a building—that can be taken down after the story is built. And indeed, much of the expansionary process of revision I have described does eventually give way to a phase of contraction, of cuts and edits. Having expanded the story, we can finally be sure of what we know about it and then allow ourselves to follow Hemingway’s advice and “leave out” (or take down) some of the knowledge that allowed it to be constructed.

I use the word scaffolding above, but the more familiar term for these props and crutches may be darlings. Our dogged resistance to cutting them underlies the single most famous piece of revisionary advice there is: kill your darlings. A line so famous as to have a dozen fathers and mothers (among them Faulkner, Wilde, Welty, and Woolf—whose rather chilling variation was “kill your little darlings”). The earliest use of the phrase, though, seems to come from the early-twentieth-century British writer Sir Arthur Quiller-Couch and his 1914 lecture “On Style” warning against “extraneous ornament”:

If you here require a practical rule of me, I will present you with this: “Whenever you feel an impulse to perpetrate a piece of exceptionally fine writing, obey it—whole-heartedly—and delete it before sending your manuscript to press. Murder your darlings.”

(Note that even here, at the origin of this most famous of all pieces of editorial advice, the impulse toward expansion is still encouraged before cutting.)

While Quiller-Couch is barely remembered today, I’ve been familiar with him, in name at least, almost from the start of my writing life, when I won a prize named after him at college (one of the first encouragements I received as a writer, worth far more in validation than the twenty-five-pound award it came with). So perhaps I might, respectfully, offer a counter to his famous, murderous advice and suggest that writers rather save your darlings.

Peter Ho Davies. Photo: Lynne Raughley

There are various ways to mean this. One could imagine a revisionary strategy that works in a Marie Kondo–like fashion by throwing out everything that doesn’t “spark joy,” essentially a revision that murderously cuts everything except one’s darlings. And many writers do employ strategies like this. Philip Roth, in his Paris Review interview, describes something similar:

I often have to write a hundred pages or more before there’s a paragraph that’s alive. Okay, I say to myself, that’s your beginning, start there; that’s the first paragraph of the book. I’ll go over the first six months of work and underline in red a paragraph, a sentence, sometimes no more than a phrase, that has some life in it, and then I’ll type all these out on one page. Usually it doesn’t come to more than one page, but if I’m lucky, that’s the start of page one.

Still, in my own experience, this saving of darlings means something else. While I do often cut them, I resist the urge to kill them and instead try to save them, not in the story, but in a file of ideas, of phrases, of “off-cuts” in case I might find a later use for them. Different writers have different names for this file. Benjamin Percy calls it his “cemetery folder” (appropriately enough for our “undead” darlings); others deem it a “compost heap.” Personally, I like to think of it as a lifeboat, or a raft. God, after all, may have sent the Flood to revise his work, but he too saved the best bits of his first draft in the Ark.

And what becomes of all these darlings? What are we saving them for?

To call on another metaphor, a trick of the mind I deploy in these cases is the idea of the jigsaw puzzle. For the sake of argument, let’s say it’s a picture of a desert island. Often, we pick up the wrong piece of the puzzle—blue, but is it the sky or the sea?—and try to jam it into the corner we’re working on before realizing it doesn’t fit. But at that point we set it aside. We don’t throw it away. We hope and assume it will find its rightful place in some other part of the puzzle. The piece is a “darling,” the puzzle is the totality of our writing, not just this one story, but a lifetime of writing. And every so often one of these stray darlings, cut months or years before from one story, will find its proper home in another.

Dan Chaon speaks of such a process as he assembled his collection, aptly titled in this context, Among the Missing:

It’s weird to look back at old drafts and see that Sandi in “Safety Man” once suffered from the unexplained blackouts that now plague the narrator of “Big Me,” and that the blow-up doll in “Safety Man” actually first made his appearance as a brief image in “Falling Backwards.”

This movement of “pieces” from story to story can also apply between novels and stories. The trans character Reese in Brit Bennett’s outstanding 2020 novel, The Vanishing Half, for example, began as a figure called Reece in a story she workshopped with me in 2012. Similarly, two or three sections cut from my own novel The Welsh Girl, had afterlives as short stories; works that I think of as “Teflon” or “Velcro” stories, after those apocryphal “spin-offs” of the space program.

Excerpt from “Darlings” from The Art of Revision: The Last Word. Copyright © 2021 by Peter Ho Davies. Reprinted with the permission of Graywolf Press, Minneapolis, Minnesota.