The process of revising effectively tends to vary from writer to writer just as the first-draft writing process varies, and it’s not necessarily connected to the way one writes your first drafts. In fact, often (though not always) the revisions process seems to need to be the opposite of the writer’s writing process in some way: writers who are very methodical and who do outlines and character sketches and so on for their first drafts find themselves winging their revisions, while those who write things in order, front-to-back, find themselves skipping all over the book while revising.
Revising is a separate skill from writing it down in the first place — related, but still different. And like writing it down, revising is a skill that gets better with practice. By the time one gets to the end of the first draft, one has definitely had a novel’s worth of practice at getting the words down on paper, and a lot of writers expect this to translate into ease of revision. If you haven’t been revising-as-you-go, however, it is highly unlikely that your revisions skills will be up at the same level as your first-draft skills…and an awful lot of writers cannot revise as they go without killing the story.
One could, of course, try revision someone else’s terminally bad piece of prose for practice and hope that the techniques one figures out will be applicable to one’s own work. It’s not hard to find examples of bad prose on the net; the trouble is finding some that makes the same mistakes you do without also making you feel as if your stuff is too horrible to contemplate.
So most of us are left with getting to the end of the book, right about the point where we feel as if we know what we’re doing, and then starting over again trying to boot up an entirely new skill (revising). The first step is always, always, always figuring out what the problem is. Diagnosis is key; if you can’t see what’s wrong, and you try to fix it anyway, it’s like trying to fix a delicate piece of electronics blindfolded and wearing oven mitts. Don’t. Just don’t.
Figuring out the problem isn’t as easy as it sounds – after all, if you’d known it was a problem, you wouldn’t have written it that way in the first place. There are various ways of going about this. Some writers lean heavily on first readers and crit groups to point out problems; others swear by the “cold box” method (stick it in a drawer for a couple of weeks or months, until it’s “cooled off” and you don’t remember what you meant to say quite so clearly). Some find that just making the manuscript look different is enough to do the trick, which these days is a simple matter of changing the font and the margins. I have friends who swear that they get this effect from looking at hard copy (as opposed to seeing tings on screen), even though nothing else changes.
Or you sit down and analyze. This means approaching the work coldly and intellectually, looking for places that don’t work and (even more important) for why they don’t work. It means avoiding the trap of getting lost in the fun, brilliant bits that you just love, and equally avoiding the trap of deciding every word, every comma, is trash and utterly without merit. It means learning the difference between fixing a problem and second-guessing a decision.
A word about this bit: the common advice to “murder your darlings” does not mean that you are supposed to go through your manuscript and take out every single thing in it that you actually like. If you don’t like what you write, why should anyone else like it? What it means is that if the only reason a particular sentence is in there is to show how clever the author is…take it out. You can save it for some other book if you like, somewhere that it will add to the characterization or the plot or the setting or something story-related, rather than author-related.
When you’re analyzing your own work, you generally need to look at both the macro and the micro level. The macro level is stuff like structure and pace and flow and tension. First you look for where things seem to be not-working; then you look for why they’re not working. In the first draft of The Far West, for instance, I had three scene in a row of studying a critter in the lab, followed by three scenes in a row of reunions with old friends/family returning from elsewhere. I hadn’t noticed when I wrote them; once I saw the problem, it was obvious that I needed to move things around so that I had some critter-studying followed by a reunion followed by more critter-studying, instead of having my heroine do the same thing over and over with different people.
Sometimes it’s not the content of the scenes that’s the problem. Sometimes it’s a lack of transition between two bits, or the fact that something wasn’t set up properly two or three scenes or chapters earlier. Sometimes the macro fix is down at the micro level. The first editor who saw Talking to Dragons told me that the pace was too slow (a macro-level problem); I fixed it by cutting roughly 5,000 words…two or three words at a time. (Basically, I figured out that I needed to cut three lines per manuscript page, and then spent three weeks going through the ms. a page at a time, crossing out words and rephrasing sentences so they’d be shorter, until I got three lines out of each and every page. It was a horrible job, but I learned a lot.)
The micro-level revision is down at the scene-to-sentence level – getting rid of ambiguous phrasing and tongue-twisting dialog, spotting the places where you over-use a particular sentence structure or a particular word. (I recall one ms. in which the student had learned to use partial parallel repetition to emphasize a point. Had learned it too well. Had become vastly fond of it. Had used it over and over. Had driven me crazy with the particular tic…which took forever to make her even see, let alone fix.)
The micro-level is where one sometimes has to dismantle and reassemble a paragraph or a scene, or rewrite it wholesale. Sometimes several times. Occasionally, a sort of reverse-layering technique is useful here, especially if there’s a scene where one can’t figure out what the problem is. You take the scene and hide everything except the dialog, so it’s just talking heads, and then you look at the flow of the dialog and whether it makes sense as a conversation without all the emotion and internal dialog and stage business that it has in the scene. Then you do the same thing with the physical action, and then the descriptive bits. It’s a bit tedious and too labor-intensive to use on every scene, but it can be really useful when one hasn’t a clue where the problem is.
Some writers find that their prose hardens into concrete at some point, and chipping out the rough spots leaves visible seams. There are two approaches to this problem: one, get to the revisions soon, before the prose sets up (for some writers, this means the same day it gets written); two, figure out how to either delay the hardening-up or soften up the prose once it’s gone hard. One writer I know with this problem prints out her ms. formatted the way her page proofs look; since she’s used to fixing things in page proof, she can see and fix them on the printout when she can’t on the screen. Another writer is fine as long as she doesn’t print out the final draft of a chapter – as long as it’s all pixels, it stays workable for her. Still another has to set aside the written scene and re-imagine the whole thing from scratch, then write a whole new version. It depends, as usual, on how your particular mind works.