Revise From the Outside In

Benjamin Franklin dated his improvement as a writer to a quarrel conducted in letters. As a young apprentice printer he had argued with a friend, John Collins, until his father found the papers and read them. Franklin spelled and pointed correctly, the father observed, as the printing house had taught him, but fell far short “in elegance of expression, in method and in perspicuity.” The remedy Franklin described in 1771 came in a sequence. He took papers from the Spectator, wrote a short hint of the sentiment in each sentence, set the hints aside for days, then wrote the papers out again and compared them with the originals. In a second drill he jumbled the hints on purpose and tried to restore the best order before forming any full sentences, “to teach me method in the arrangement of thoughts.” Words got a drill of their own, because he found he “wanted a stock of words.” He was practicing on another writer’s finished essays, not revising a manuscript, so the purpose was fixed. He still put arrangement before sentences, which is the order this chapter recommends for revision.

The case for that order in revision is ours, and it rests on dependency. Each scale rests on the one above it. When the purpose changes, so does the structure, and the new structure moves or removes scenes, each of which takes its sentences with it when it goes. Effort spent lower down is less likely to be undone by a decision not yet made when the work runs from the top. Writers often start where control feels easiest anyway, since a sentence can be fixed in a minute and a structure can’t. Line work produces visible progress, and it can also attach a writer to passages whose place in the Work hasn’t been settled. Two weeks of line work on a chapter that later turns out to belong in a different book is a real cost, and it makes the next structural decision harder to take honestly.

Chapter 7 puts the principle in a line, revise from the largest unresolved decision downward, and here it becomes one pass at the top scales in which you touch no sentence. To revise structure you have to see it, and few writers can hold a long manuscript of their own in view all at once. The usual instrument is the reverse outline (Practice 28), a list of what each section or scene does, written from the pages rather than from the plan. In fiction, ask what changes between the start and the end of the unit. In nonfiction, ask what the reader now knows, or has been asked to accept, that they did not know before. A unit that cannot be given a function is a candidate for cutting, merging, or relocation, and two units with the same function are a candidate for fusion. A function the whole requires and nothing supplies is a gap. Gaps show up only at this scale, and a copyedit reads straight past them.

A discovery writer often sees the true outline for the first time in this pass. A planner can set the pages beside the plan and decide which of the two is wrong.

Revision can also run upward. A voice found at the level of the sentence can reveal what a piece is actually about, and then the purpose has to be rewritten. Hemingway’s habit of rereading and rewriting each day up to where he had stopped is line-level work that doubles as orientation, and a writer who works that way still has a use for this practice at the end of a draft. A poem of fourteen lines has no separate structural scale at all, because there the sentence is the structure, and a short essay may pass through all six scales in one sitting. The order is a default for manuscripts large enough that early decisions carry a cost a long way downstream. Ask whether the work you’re doing could be undone by something decided later. If it could, decide the later thing first.

Practice. Give ninety minutes to one chapter, story, or essay. Before beginning, write one paragraph on what the piece is for. Then list every unit with a one-line function, and mark each unit keep, move, merge, cut, or missing. Do not edit a sentence; if a typo bothers you, copy it to a list for later. Finish with the annotated list and, under it, a short to-do in order, with the largest decision first.