There's a gap in the action between 22 and the next card. Now you can dream up some scenes to fill the gap. Go back into your trance and dream two or three cards in there. At some point you're going to find yourself dreaming, on cue, scenes that you didn't get to the first time, filling gaps. You'll also find that a lot of scenes you have dreamed aren't going to make it into the final structure of the novel. And you'll find that the contradictions become reconciled as the structure takes shape and you find your way through to the end. If you're arranging your cards in this extreme way, all the way to the end of the book, number the cards, one to however many represents the whole structure.
During the brainstorming phase, do not give any consideration whatsoever to continuity. Embrace the seeming randomness. If you are tapping into your unconscious and moving around a legitimate character with legitimate yearning, it's hard to say what your unconscious is already perceiving and contextualizing. Once you finish the dreamstorming — weeks — then you try to bring order to that randomness. However, you're simply looking for continuity from one scene to the next. You are looking for the through-line among all those disparate parts. The nature of transition is totally unconsidered. There are so many potential moments and so many possibilities of scene in this book that the only way to explore them fully without willing them into a structure is to let them happen at their own seemingly random pace. Then, only after that, your job is indeed to see the sequence that makes narrative sense of the disparate pieces.
Let me digress for a moment on the subject of research. I always record the books I read on index cards. I'm looking for different things than people look for in academic research, and my cards usually represent a personal index of sense details. Sense details, scenes, images you find in your research you can record on your cards and plug them into the sequence as you arrange. For example, Countrymen of Bones was set in the Alamogordo desert during the building of the first atomic bomb. This book took some historical research, as well as reading in nuclear physics and archeology, and the cards allowed me to indicate briefly certain things I might need to know for certain scenes. For example, one of my main characters worked on the bomb, and his job was to lead the team trying to craft the lens in exactly the right shape to hold in the explosion. I didn't have to know everything about the bomb; as soon as I read about the team working on that lens, I said, "That's Lloyd." So I did a lot of indexing of that particular job, the cloud chamber where they tracked the beta particles as they flew off the explosion, the photographs they took of the smoke in the chamber, the track of the beta particle as it poops out. Such details were suggested by scenes in my brainstorming, and some suggested scenes. (For example, Lloyd rearranges a table of cloud chamber photos in a scene.) When you record such details, on your card are just six or eight words, a sense detail, an indicator of scene, or maybe a reference to a book where you can fill in a historical or professional detail. The juice is not written out of these scenes because you've made sure you didn't start writing the description or the dialogue; you just put down the bare indicator and then you went on.
When your cards are arranged, you take the first card and you start writing the novel. Here's the first scene. And it's ready for your unconscious because all you have here is eight words, a sense impression to call up that scene. You go into your trance and you write that scene, and now the second scene comes up, and you write that scene, and then you write the third card, and — guess what happens in the third scene. Something you didn't expect. You don't even know where it came from, because it's coming from your unconscious. Great. That leads you to a fourth scene and a fifth scene you didn't expect. Great.
But now you feel that the story has to be drawn back into the main thrust of everything yet to come. Now that you've got an unexpected third, fourth, and fifth scene, you go back and look at the fourth card, and it doesn't fit quite the way it did at first. Wait a minute, this changes some things. In fact, you must go back and look at all the cards, numbers 4 to 92 (or however many you have). What do you do now? The next day you go into your trance and lay out cards 4 to 92 and rearrange them. In essence, you rewrite your book structurally.
I personally write from beginning to end of a book. I don't want to be dogmatic about this, because I'm sure a particular artist may do it some other way. But it's hard for me to imagine writing very much out of sequence, because sequence is crucial in a narrative. If everything is organic in a work of art, and I skip six or eight or ten scenes to write this scene that feels hot to me right now, how do I make decisions about character, voice, image, event in that scene? And there are crucial matters of motif, of recomposition, that I will soon speak of, which are impossible to manage by skipping ahead out of context. In a contextual vacuum, you make decisions that are at best tentative. And why are you doing that? I think oftentimes the impulse is to avert your eyes: that's the scene I can do safely; the scene that is not really going to challenge me; that's a good scene to unite today.
So I would say it probably doesn't work to "write around" in a novel. Structure happens from the imperatives of the whole object you're creating. When you are driven by the desire for the organic wholeness of the object, and by the need to recompose the elements that are already in the work, and by the dynamics of your character's desire, structure will inevitably come from that.
Yet this system, if it's going to work for you, has to be totally flexible. The fact that you've got ninety-two cards in a row doesn't mean that you're going to write those ninety-two cards out one after the other and think you've got a book. No, you're rewriting and rewriting and rewriting, over and over, not on the level of phrases and paragraphs, but on the level of structure. There's a lot of rewriting going on here, but it's not in drafts, it's in following those instinctive, from-the-hot-spot surprising things, and then restructuring everything to come as a result. It's got to stay flexible, and it's got to stay from the trance, or you will pull the trigger on the shotgun and blow off the back of your novel.
It's also possible to use this system in a more limited, but still more flexible way. In They Whisper, my seventh novel, for instance, I was dealing with a very complex structure driven by the flow of delicate emotional associations. No way in hell I could anticipate what the sequence was going to be. But I dreamstormed two hundred cards, and then all I did was look for eight cards that might be near the beginning, and I strung them out and rearranged and rearranged them. Then I wrote those scenes. Once they were exhausted, I went and got six or eight more cards, and so forth. The useful thing was that the possibilities already indicated on the cards helped to guide and structure my unconscious, which was improvising the form as it went.
I've never used the cards the same way twice — maybe they operate for me like a tarot deck. But this is fundamentaclass="underline" keep it open, fluid; realize that nothing you do here is locked in, it's got to stay subordinate to the trance state in order to work.
The advantage I see of this system over multiple drafts is that in the big sprawling rough draft, no matter how open-minded the writer is, she has to make approximations in the first draft, then she must make approximations in the second, and more in the third, adding more rough, headlong stuff in the fourth. If the book is at all complex, the draft writer will hit forks in the road, over and over, and must choose this fork instead of that. If that happens early in the book, or even in the middle, by the time she gets to the end and the novel is sprawling in whatever way it sprawls, it's very difficult to go back and take the other fork she faced on page 30. With this system, all the forks are fine — you follow this one, you follow that one; you go down this fork in the sixth week of dream-storming; in the tenth week, you go down that one, as far as you want to go — because at each point you are rewriting and redreaming the book on the level of structure.