Revision: Making It All Better

Years ago, in school, I didn't need to revise what I wrote. Oh, you were supposed to, of course--hence the requirement to produce an outline and first draft for term papers. I wrote them backwards. First the term paper, then the draft, then the outline...then I turned in the outline, then the first draft, then the term paper. Teachers never figured that one out, and it is easier to write an outline that fits the paper than a paper that fits the outline. It was hard to write a first draft with mistakes in it, but I managed. I was a good natural speller; I had a good grasp of grammar and syntax; I could write well-constructed paragraphs that led to a conclusion--and that's all the teacher wanted, and graded on.

More advanced writing requires more, and until I learned what revision really is, and how to do it, my writing didn't progress. Even if a first draft--the real first draft--has no spelling errors, no grammatical blunders, and no sequencing errors, it may not be publishable (or anything you'd want under your name ten years later even if it is published.) Hence, revision.

Revision, I've said before, is actually re-vision...you have to re-see the work, see it as an outsider sees it, see it the way you wanted it to be, and then do what's necessary so that the reader, the outsider, sees what you wanted them to see in the first place. And hears what you want them to hear, as the words run through most brains as a voice--as your voice, the storyteller's voice.

The hardest revisions are those which you have to do immediately after writing the piece, by yourself, in a hurry. It's hard to see the piece clearly if you've just written it, hard to be that stranger coming to the work you know so well, hard to see the gap between what is on the page (or screen) and what you hoped to convey. The easiest are those allowing sufficient time and the help of a good alpha reader.

Let's start with what can go wrong with something written by a writer who is competent in the basics--who, most of the time, doesn't confuse homonyms, doesn't misspell words, doesn't have trouble with subject/verb agreement. Someone who knows what they meant to write (length, topic, tone, basic structure) and knows what belongs to those parameters.

For me, the architectural metaphor works: in building a structure, you can have problems at the design phase (something's in the design that just can't work--a bathroom with no water supply, for instance), in the construction phase (a good design wasn't followed, or sloppy construction left the staircase hanging by one bent nail), and in the "finish" phase (although there's good design and sound construction techniques, there's a heap of construction debris by the front door, half of one wall never got paint, the handles in the kitchen cabinets don't match.)

So every piece of writing intended for publication has a design (whether you make it all first, or work as you go along) and if the design has flaws, those flaws will ruin it despite excellent craftsmanship in the writing: perfect spelling and grammar can't save a lousy plot. On the other hand, good basic designs can be ruined by lack of craftsmanship in execution. Bad writing (which usually means bad craftsmanship in constructing sentences, paragraphs, scenes) is hard to read. And finally, when editors are looking at stories to publish, and readers are comparing writers to follow, there's the finish level...does the craftsmanship of construction continue to the finish, to making an attractive, desirable interface for the reader, the book equivalent of "curb appeal"?

Many people--including some of your friends and relatives--focus immediately on the finish level: misspelled words, typos, the odd grammatical blunder. This is what they were graded down for in school, and this is what they see as the first thing to fix. But it's not. Smoothing and polishing the top layer is the last thing to do...other revision steps may eliminate that misspelled word altogether. More experienced readers, and editors if they want to bother, will see problems at many levels, and hand you a confusion of comments that relate to different levels of revision.

Depending on circumstances, you now have two ways to go: 1) start at the design level, fix all the design flaws, then do to the construction level and fix all construction flaws, and finally do the top level (spell-check, etc.) or 2) start at the front of the book or story, and work through it front to back, sequentially. Both approaches can produce a good final result, but they feel different while you're doing them.

Within each of the three basic levels of a work--design, construction, finish--are sublevels, and writers may choose to combine them into one revision pass or make separate passes for each potential problem type. That decision will depend on the writer's individual preference, but also on the writer's skill. People whose first drafts are a mess (and there's nothing wrong with a chaotic first draft as long as you can untangle it, beat it into shape, and polish it) will benefit from taking their problems sequentially and making separate passes for each class of problem. People whose first drafts are less tangled can often combine several sublevels into one revision pass.

Let's look at the two main approaches--front to back, and bottom to top--in more detail for some hints of which might work best in a given situation.

For me, bottom to top is the best method for the writer who must work alone, soon after the initial writing, with little or no input from alpha readers or editor. It requires reading clear through a minimum of three times: once for the design, once for the construction (how the design was executed) and once for finish, but if done well it produces a final version that is completely coherent (within the writer's own ability to see the work clearly) and feels "organic." Each level of revision is done sequentially, front to back (since everything early in the work affects later sections.)

In the design level reading, you're looking for design flaws in what I call "deep logic" (why things work as they do), motivation, actions, etc. In other words, does the story make sense in terms of your story-universe's reality? Do the characters' actions make sense in terms of the personalities you've given them? Are events in the right order? Do causes come before effects, and are there causes for each effect, and effects from every cause? Even if you don't outline (I don't), by the end of a story or book you should be able to outline it, and the outline should make sense. Does the tone make sense (if you set out to write a funny love story, is it still a funny love story all the way through, or did part of it suddenly turn into a political rant?) Do the emotional highs and lows come where you need/want them, and is the final climax/emotional payoff really the big one? (It's easy to overbuild an earlier one and have insufficient "oomph" in the end.) Since nothing else will be right if the design isn't right, you really do need to work on that first. The "outline" doesn't need to look like the outlines you learned at school, or the one that more organized minds can do up for a book proposal...but it should have "flow" in terms of cause and effect, logic, emotional push/pull, suspense/revelation, etc. If you find a design flaw, note what you wanted to have accomplished there, as well as what you hope will accomplish it. This will help in tearing out and reworking the right part. Design fixes usually require removing and adding material--sometimes a lot of it. You are putting the structure of the story back into the fire, making it malleable again, even completely remelting/remaking it. Write all those new additions in first-draft mode (in other words, you can do it fast and rough if you first-draft that way.)

Once the design level has been reworked, the next step in bottom to top revision is the construction level. Again, this will take at least one reading straight through with construction problems only in mind. This is where you deal with rough transitions, clumsy sentences, out-of-order sentences (which happens to me a lot--when I'm writing fast I may reverse sentences between my brain and my fingers and scenes that don't do what they're meant to do.) You can also catch additional extraneous scenes in this step...something that seemed important in an earlier draft may now not fit, or be too long, because what you added in the design level revision changed things. Some writers can do quite a bit of the finish work at this level (it depends on how good the construction was in the first place) but it will still need at least one detailed check on that afterwards.

The finish passes--however many there are--deal with all the obvious surface details. Spelling, of course. Words repeated too often, too close together. Personal quirks--starting sentences with gerunds, or writing in too obvious a rhythm or the use of a pet word or phrase that shows up in everything you write. "Fossils" from earlier drafts (the sentence into which you put a new detail, making it ungrammatical, for instance.) This is the best time to read the work aloud (even mumbling quietly to yourself) because that will reveal flaws you didn't see as your eye skimmed the familiar material. You will find the best word, not just the adequate word; you will add the jeweler's rouge to the tools you've used so far, and the story will shine, then sparkle.

And because it was worked from the raw material up--because it's all cast as one piece and worked as one piece--it will look and sound and feel all one piece to the reader.

Next time--the front to back approach. But note: this was front to back too, only one layer at a time.

Revising in sequence, front to back.

For the writer who has input from multiple readers or from an editor who's already committed to the work, it's sometimes more efficient to deal with multiple layers of problems at one time than to tease apart the various comments into their "layer" references and put them in order. However, the goal is still to produce a work that feels all one piece, that feels as if it had arrived on the page intended and crafted as one. Doing this while working from the front of the book to the back has a different feel from working through level by level, but as long as you don't hop around in the sequence, it will still work.

Looking over the comments for the first few chapters of a novel, for example, you might find one or two design-level comments, another few on construction, some on finish (typos circled, etc.) and some whose level you aren't sure of. Your alpha readers and editor have given you page references (or marked the manuscript) so you know what comes first...that embarrassing typo right in the middle of page one. You fix that one. Next is a comment on motivation, a design-level problem...probably. It might be, or it might be constructional. First look at the design-level possibility....does the motivation not make sense because you set it up so it's senseless (design level), or because you didn't express something you knew about that character (construction level)? Fix it at the level where the problem is. Move on to the next comment in order...being aware that design still underlies any higher-level problem, and you might spot one the readers didn't...if so, deal with it. You can't afford to depend entirely on any reader or group of them...even a good editor will occasionally miss something.

Comments like "unclear" may be either construction or design problems (saying the wrong thing more clearly won't help!) Comments like "awkward" may be either finish-level or construction problems. Think down a level when you see a problem--could it have arisen from a deeper-level problem? A cracked sidewalk isn't fixed by plastering over the crack, if the cause is a tree root swelling beneath it.

When you see multiple comments, and multiple layers of problems, in one section, tackle them from the bottom up, always. For me, a front to back sequential revision still always requires a complete pass (at least one) to follow at the finish level--again, ideally, reading it aloud--to be sure that the whole now reads as one.

Both these methods use elements of the other--there is sequential work in the bottom to top, and bottom to top work in the front to back--so obviously they can be combined. If you know an entire section needs work, you can do everything else in the book up to that in sequential mode, then rewrite that section bottom to top, and they should fit together seamlessly...and then you can do the rest of the book in either mode.

What doesn't work is hopping around doing design/construction work in random order through the book. Revisions dropped in randomly stick out like boulders in dirt...they don't fit in. Lots of people don't understand this, and will say "All you need to do is change this sentence (or paragraph, or event)"...but you can't change anything bigger than a typo without affecting other things, and to seem whole at the end, the work must be re-visioned whole.