The Draft That Tells You What It's About

For a long time I thought a first draft was a rough copy of the finished book. Same shape, worse execution. You wrote it badly, then you tidied it in successive passes until the badness was gone and the book was there. Under that model, revision is sanding.

That is not how it has ever once worked for me, and it took writing several long things to admit it.

The first draft is not a worse version of the book. It is a different object with a different job. It is the instrument you use to find out what the book is about — and it performs that job specifically by being wrong.

The theme you declared and the theme that showed up

Every long project I have started, I started with a stated theme. I could say it in a sentence. I believed it. I built an outline out of it, and the outline was usually good enough that people nodded when I described it over dinner.

And every single time, somewhere past the midpoint, a different theme walked in and quietly took the chair.

It does not announce itself. What happens is that one scene comes out unaccountably hot — easier to write than the ones around it, longer than it needed to be, still there in your head three days later when you are doing something else. Usually it was not in the outline. Often it is a scene you will eventually cut. But the energy in it is real, and the energy is a signal: something in that passage is closer to what you actually came to say than anything in your stated theme.

Then you have a choice, and it is not a comfortable one. You can defend the plan — keep the announced theme, treat the hot scene as a digression, finish what you set out to write. Or you can accept that the book has told you something and go back to rebuild two hundred pages around a thesis you did not possess when you wrote them.

I have made the first choice. The results were competent and inert. Everything readable I have written came from making the second one.

You cannot think your way past it

The obvious objection, and the one I made to myself for years, is that this sounds like a planning failure. If you had thought harder up front, done more structural work, interrogated the premise properly, you would not need to discover the theme by accident two-thirds of the way in.

I no longer believe that, for a reason I find fairly convincing.

Planning can only operate on what you already know. You can outline the parts of a book you have already understood; that is what an outline is. But the thing worth writing is almost never among those parts. If you fully understood it in advance, the writing would be transcription, and transcription is what dead prose sounds like. The reason to write a long thing is to reach something you could not reach by sitting and thinking about it — and there is no route to that place except through pages that do not know where they are going.

The wrong draft is not the price of insufficient planning. It is the mechanism. You are not writing a bad version of the book. You are running the experiment that produces the book's actual subject, and the experiment is expensive because real experiments are.

What that means for the pages you throw away

This changed how I feel about cut material, which used to bother me more than it should have.

I have discarded a great deal — chapters, whole arcs, a character I liked more than the book could accommodate. For a long time each cut felt like waste, and specifically like evidence of my own inefficiency. If I were better at this, surely I would produce less material that turns out not to belong.

But those pages were not failed product. They were instrumentation. The chapter I cut is how I found out the story was not about what I thought; without having written it, I would have kept the wrong thesis and the finished book would have been worse in a way I could never have diagnosed. The cut pages did their job. Their job was simply not to be in the book.

That reframe is worth more than any productivity technique I have tried. It converts the most demoralizing part of long work — the discarding — from evidence of incompetence into evidence that the process is functioning.

Not only about writing

I notice the same shape everywhere in my working life, now that I know to look.

The first version of a product is not a rough version of the eventual one. It is the instrument that reveals what the product is for, and it reveals that by being wrong in an informative way — customers using the feature you considered secondary, ignoring the one you built the company around. The plan does not survive contact with reality, and I have come to think that phrase is usually said with the wrong tone, as though reality were an obstacle. Contact with reality is not what damages the plan. It is what completes it.

The uncomfortable part, in both cases, is the same. You have to be willing to have been wrong in public, on the record, at length, after considerable effort. The draft is proof that you did not know. So is the first version of the product. Everyone who has finished anything real has a drawer full of that proof, and mostly we do not show each other, which leaves each new person to conclude they are uniquely bad at this.

So this weekend, if you are somewhere in the middle of a long thing and it has stopped resembling what you promised yourself it would be, I would offer this: that may not be the project going wrong. It may be the project telling you, in the only language it has, what it was always going to be about.