Quill

The companion is the essay's working file: its subject, sources, decisions, and open questions, kept beside the draft. It comes before the essay because the essay will change, and the companion is how every session, and every agent, starts where the last one stopped. Developed at length, it becomes the agent's extended context, so the agent already knows the work before the first sentence of prose.

Make Believe began on a night we watched the Disney film Coco. I noticed the deep themes the film was carrying on top of the children's story, and how quick people are to call films like that just a movie. The idea waited about a month in my notes, then got a companion note, opening with one provisional sentence about what the essay might be.

Developing is a conversation, not reading at a desk. I read widely, and the agent takes me deeper into what I have read and brings in what I haven't; having to test each new thing, whether it really fits, is part of how the work forms me. I do the judging, and the essay finds its real subject.

I talked the idea through, and the agent came back with research, like a lead about Joseph Campbell's influence at Disney, which went into the companion marked to be checked. Just a Movie became an essay about the stories we tell children.

With a rich companion note now carrying weeks or months of deliberations and decisions, only now do I start the essay itself. I write by voice and edit by ear. Listening to a draft will show me where it is wrong, and a listen often sends me back through it; the ear catches what the eye forgives.

When I said “Let's draft the intro,” the agent was working from the companion we had been developing together, where the opening was already worked out: the boy who cried wolf, told in disguise. I gave it more context, a family that kept its sheep after the neighbors sold theirs, and it rewrote the scene around them. Later, listening to another passage, I wanted a different image: she doesn't cry, she weeps.

When a draft is whole, other agents read it cold, with no memory of the drafting, and see what I can't. Nothing they say is adopted by default: every finding is ruled on by what the essay is trying to be. A finding can send the work back to composing, or as far as developing.

One cold read of Make Believe sent back eight findings. I was sure they were wrong; three were right, including Pinocchio's donkey ears, which he loses in the end. I also fixed another paragraph by hand.

Every name, date, and quotation is checked against a source before an essay goes out, and the checks often make a sentence more exact, not just correct. When a check overturns a claim, the work goes back to developing.

The draft said the Grimms made the mother a stepmother by their second edition. The editions show it happening in Snow White first and in Hansel and Gretel two decades later, so the sentence became “from edition to edition.”

Every essay here is narrated, and I listen to it the way a reader would. The recording shows whether the essay holds when it is heard. If a listen shows the argument itself is unclear, the essay goes back to composing.

In the recording of Make Believe, the narrator voiced a boy who tears out the roses as if he were in tears. The fix went into the narration alone; the text stayed as written.