Don’t Look Back
When you’re in the initial phase of researching a project, you inevitably battle the limits of your memory: forgetting hunches, or failing to make connections between ideas or stories that you’ve uncovered in your research. It’s those limits that make it so helpful to keep spark files, or use outboard memory tools like Devonthink or Findings. But once you enter the writing process, your memory becomes a problem for the opposite reason: it’s too powerful. Like Funes the Memorious in the classic Borges story, writers often get into trouble because they remember too much.
When I was in my twenties, writing papers in grad school and then writing my first two books, I had a default routine for working on a big project that I know many other writers share. I’d sit down at my desk, launch my word processor, and then start reading the chapter or the essay I was writing from the beginning, all the way up to the point where I’d stopped the night before. Sometimes this would entail reading five thousand words of material before I actually thought about typing out some new words. And inevitably, in reading through that older material, I’d find some passage that needed tweaking, and I’d spend a half hour getting that section fixed. And then the next day, I’d sit down at my desk and repeat the whole process, reading six thousand words before getting started on the new material.
Eventually I realized that there was something fundamentally flawed about this approach. For starters, it was a seductive procrastination tool; generally, when you’re writing, your brain is doing everything it can to justify not having to come up with new sentences, and so if you can convince yourself that reading the whole chapter will better prepare you to write those new sentences, it’s hard to resist. But before you know it, you’ve lost an hour of work time reading through the back pages.
Procrastination, however, isn’t the real problem. The real problem is what happens to your reading memory when you do this backtracking routine every single day for the life of the book. Let’s say it takes you twenty days to write a chapter, and each day you go back and read the previous material before you start writing. That means when you actually sit down at the end of the process to read through the entire manuscript, you have read every chapter at least twenty times, maybe more if you’ve gone back and re-read earlier chapters while you’re in the middle of the book.
Those twenty re-readings create a massive problem when you finally sit down to do the most important edit of all: reading through the completed manuscript for the first time. This is your one best opportunity to read with fresh eyes, to imagine what your reader will think as she’s following (or not following) your narrative or arguments. This is your best chance to sense how the overall flow of the book is working: this section moves too quickly and risks confusing the reader; this one drags on way too long. But if you’ve read each sentence two dozen times, it’s almost impossible to make those judgments, because everything seems incredibly predictable and flat to you, for the simple reason that you’ve ingrained each twist in the argument, each rhetorical turn of phrase, in your memory. Your words are like a pop song that you’ve overplayed; you literally can’t hear the pleasure of it any more.
That’s why I adopted a strategy about ten years ago where I do everything in my power to limit the number of times I read a manuscript when I’m in the middle of writing it. Each day I sit down at my computer, open up the chapter I’m working on, read back one or two paragraphs to remember where I left things off — and then I start writing. (Often times I leave myself a little shorthand description of what’s coming next so I know exactly where to start.) When I’ve finished a chapter, I print it out and do a single pass edit on it — and then I put it away and move on to the next chapter. When I finally sit down to read the first draft of the full manuscript, I’ve only read most of the sentences in the book three or four times, as opposed to twenty.
Now, this approach can make that first read-through fairly embarrassing. On several occasions, I’ve found that I have written the same paragraph twice in the book, often nearly word-for-word, forgetting in chapter four that I already explained how dopamine works in chapter two. And because I haven’t been editing the chapters compulsively in the middle of writing them, the manuscript is rougher around the edges.
But the benefits of this approach greatly outweigh costs in my experience. For starters, you write much faster and you’re less prone to the seductions of endless re-reading. But the main advantage is that you get to experience that first manuscript as a reader, and not an author who is sick of his own voice on page. You genuinely get to enjoy the bits that are really working, and you can sense immediately the parts that aren’t. You don’t have to think about an “imagined reader,” because you are that reader.