Why revision is necessary

I'm kinda stuck on this subject right now, since I'm in the middle of this big novel rewrite, and I am not thinking about much else for a change. But I never cease to be amazed how much of the important work in a novel--more so, perhaps, than a story--happens in, say, the fifth or sixth draft.

I may have posted here before about my perennial experience writing a first novel draft--I cruise along like a bastard for about 150 pages, and then I screech to a halt. Why? Because that's when I finally begin to realize what novel it is I'm trying to write, and it's never the one I've been writing--it's this other one. So the rest of the manuscript is the "right" one, and draft #2 is always about going back and fixing the first 150 pages so that they're actually part of the same novel as the rest.

Draft three is about polishing draft two. Then I pass it around to people--Rhian, Ed, Bob, Brian--and get their opinions. Then I sulk for a couple of weeks (sometimes many months) and decide who to listen to and who not to. Then I write draft number four.

At this point, I'm beginning to discover the stuff that will later hold the book together. Images, themes, recurring secondary characters; familiar turns of phrase, evolving narrative flourishes, parallel plot twists. Today, deep in draft four, I found myself pulling a neat fast one--a prop that figures prominently in a flashback episode late in the novel, I ended up planting earlier on, without explanation. Suddenly there's a line drawing those two parts together, when before there was none. And I spent yesterday eliminating a character and repurposing some of her scenes--one appears in a dream, and another is taken over by somebody else. It turns out I didn't need her after all--in fact, she was all wrong for the book (as Rhian warned me, early on). I will probably end up cutting the dream, even.

Draft five will be another major overhaul, most likely, in the wake of another peer review, and then draft six will go to my agent. Should the thing be accepted by a publisher, I'll do draft seven under an editor's pen, draft eight in the wake of a copyedit, draft nine in loose galleys, and draft ten in bound galleys. And probably another one will sneak in there, somehow--maybe my agent will have some ideas, too.

That's one draft of new writing and nine or ten of revision. The new writing takes six months. The revisions take a year and a half, at least. This is where the pile of pages turns into an actual book. Up until then, it's just a good idea that I ruined by trying to actually write it.