The method in this chapter uses four strands of theory, and each answers a question a writer meets with a draft and its notes in front of her. The strands come from research on how writers revise, from editorial practice, from the psychology of reading, and from writers’ own accounts of how they work.
Why do writers revise the wrong things first?
The practical question is why a writer with a structural problem spends her revision on sentences. Nancy Sommers put a version of it to two groups in 1980, twenty first-year university students and twenty experienced adult writers, and compared what each did when revising its own essays. The students revised mainly by substituting words and phrases, a strategy Sommers called the thesaurus philosophy, and they described revision as cleaning up. The experienced writers revised to find the form or shape of their argument: they added, deleted and above all reordered, worked on the whole and the parts in turn, and expected the whole to change. Sommers’s subjects wrote essays, not novels, and her finding is about what the two groups did, not about what made their texts better; but it describes the first failure pattern of Failure Patterns closely, and it locates the difference in the writer’s idea of what revision is for.
Linda Flower, John Hayes and their colleagues (1986) proposed a working model of revision in three steps: detecting a problem, diagnosing it, and choosing a strategy. A diagnosis names the kind of problem, and only a named problem can be mended by a strategy that reaches beyond the sentence where it was found; an undiagnosed problem tends to be met by rewriting the passage in which it was noticed. A common gloss on the model, which this chapter adopts as a working assumption rather than as a finding, explains why a writer finds her own problems late: she knows what she meant, and reads her meaning into the text. The model is the reason this chapter separates the place a problem is felt from the place it is caused, and the reason it asks a writer to diagnose before she repairs.
What does an editor do first?
The practical question is which part of a book to look at before the others. Editorial practice has an answer older than the research. Maxwell Perkins’s letter to F. Scott Fitzgerald of 20 November 1924, about the typescript of The Great Gatsby, praises the book’s design and makes one structural objection, that Gatsby is somewhat vague: The reader’s eyes can never quite focus upon him, his outlines are dim. It says nothing about sentences. The editor Susan Bell, who uses Perkins and Fitzgerald as her main case in The Artful Edit (2007), divides a writer’s self-editing into the macro-edit, which reads the whole for its shape, its proportions and its sense, and the micro-edit, which works line by line, and puts the first before the second. The publishing sequence of structural edit, line edit and copy-edit is the same order at an institution’s scale.
The same practice supplies the map. Shawn Coyne’s The Story Grid (2015) asks a writer to fill in a spreadsheet with one row per scene, recording among other things the story event, the value that shifts in it, the viewpoint, the time, the duration, the location and the characters on stage, and to state the whole book on a single page before revising any of it. Writers of non-fiction have long used a reverse outline, a summary of a draft paragraph by paragraph written after the draft, for the same purpose. The book map of The Mechanism is a version of both, with columns chosen so that the instruments of this book’s other chapters can be read from it.
What should the map record?
The practical question is which facts about each scene a writer must have in front of her to see the book. The psychology of reading suggests an answer. Rolf Zwaan, Mark Langston and Arthur Graesser (1995) found that readers keep track of five dimensions of a story as they read: time, space, the protagonist and other entities, causality, and the characters’ goals. A sentence that breaks continuity on any of them takes longer to read, and events that share more of them are more closely linked in the reader’s memory. A reader’s grasp of a book, on this evidence, is built on those five, and a map that records them for every scene records what the reader is building. The book map adds what the novel’s own structure needs: the thread, what changes, what the scene deposits for later, and what the reader learns. The Transition (Chapter 5) uses the same five dimensions to price a transition, which is the same finding at the scale of a gap.
What is a reader’s note evidence of?
The practical question is how far to trust a note. Neil Gaiman put the usual working answer in one of his rules for writers, published in 2010: when people tell you something’s wrong or doesn’t work for them, they are almost always right. When they tell you exactly what they think is wrong and how to fix it, they are almost always wrong. The rule is advice, not research, but it agrees with Flower and Hayes’s model: a reader detects, and seldom diagnoses. Her note is good evidence of a symptom and of where she felt it, and poor evidence of its cause, which may lie hundreds of pages from the place she marked. The Ending (Chapter 16) gives the clearest case: an ending that passes every test of the ending and lands on nothing has its fault in the middle of the book, which has not built a want urgent enough to conclude.
Is the order a description of how writers work?
The practical question is whether a writer who does not work in this order is doing it wrong. Kurt Vonnegut, in Timequake (1997), distinguished two kinds of novelist: swoopers, who write a story quickly, any which way, and then go over it again, and bashers, who go one sentence at a time and do not move on until the sentence is right. He counted himself a basher. The ladder is not a description of either habit. It is an order of decisions, and a basher who settles each sentence before the next has settled, in her head and before the sentence, what the book is and where the sentence stands in it; when she cannot, she stops. The chapter’s order describes dependence, which holds however the decisions are reached.
The period caveat applies as it does throughout the book. The ladder assumes a novel of the kind the earlier chapters model, built on wants, decisions and causes, in which the shape of the whole can be read from a sequence of scenes. A book whose unit is the voice rather than the scene, or whose meaning lies in a pattern rather than a change, still has an order of dependence, but its rungs hold different decisions, and its sentences may sit higher in that order than this chapter puts them.
See also: the full treatment, section 2, with each strand at greater length and its sources.