Finished, and then two hundred and eighteen changes

On 21 August I wrote that Pantheon was finished. Since then the chapters have been changed two hundred and eighteen separate times.

That is the gap between calling a thing finished and it actually being finished, and I can measure it exactly, which is either useful or unkind depending on the day. Everything below is what has been filling it.

What is in the gap

Editing and rewriting, mostly. Keeping each character’s voice separate from the others so nobody drifts into sounding like everybody else, which is the failure I am most afraid of and the one that arrives most quietly.

Then everything that is not writing. Turning a manuscript into an ebook a shop will actually accept. Getting advance reader copies together and out to people. The retailer paperwork, one storefront at a time, each with its own idea of what a book is. The social posts, written, checked against what has not been said yet, and sent.

All of it is mine. There is nobody else in this to hand any of it to, and that is the actual shape of the thing people mean when they say a one-person project. Not the writing. The writing was always going to be one person. It is the forty other jobs standing behind it.

I also set myself a schedule I could not really keep. The work happens every day, in the evenings, after a full-time job has had the rest of the day. Two hundred and eighteen changes is what evenings look like.

What a model is actually good at

A model does a real part of that list well, and it is worth being precise about which part, because the honest answer is narrower than the marketing and more useful than the backlash.

It is good at what I can no longer see. A word I have used four times on one page. A chapter whose voice has drifted off true from the one before it. A construction I keep reaching for when I am tired, that I would swear I do not use. After a year inside the same manuscript I am the worst available reader of my own prose, because I am reading what I meant rather than what is there, and something that has no memory of what I meant is genuinely better at that than I am.

Two of the things I have written about this month are exactly that. A number I had turned into the name of a thing, and a book that was evenly too long. In both cases something counted, and the counting was worth having.

In both cases it also stopped well short of the answer. The counting told me four was ordinary and did not know that I had used it as five different names. It told me the book was long and had nothing to say about which sentences were allowed to go. I got the rest by reading.

Where it stops

It stops the moment the question is not what is on the page but why is this not working.

Why does this scene go flat two thirds of the way through. Whether you will believe a choice a character makes deep into the book, forty chapters after I set up the reason. Whether a reveal lands where I think it lands, or four chapters earlier, or never. Whether a person is behaving like a person.

Nothing that counts things can answer any of that, and I have stopped expecting it to. What actually helped is unglamorous: reading how other writers solved the same problem. Most of the craft advice I have picked up this year came from author accounts on Instagram, read in the gaps the evenings leave. Someone explaining how they handle a time skip. Someone else on why their middle sagged and what they did. It is not a course and it is not a mentor, and it has been worth more to me than any tool I have built.

You can lean on a model to a certain point. Past that point the only thing that gets better is you, and the way you get better is the way everybody has always got better, which is by reading people who are further along and stealing what works.

The word count, and what it bought

I went too free with the length and I do regret it. Not as a performance of modesty. I regret it because I can see exactly where the extra sits and I could have caught it a year earlier, when catching it would have cost me a conversation with myself instead of a month.

What I do not regret is what the length bought. This book had to lay a foundation wide enough to carry the two that follow it, and everything I have planned for them is already sitting on the page, waiting for me to reach it. A leaner book one would have been a better object and a worse first act.

So two things change for the second book. Time, given to it properly and in advance rather than scavenged from whatever evenings are left, which is why it is planned for 2027 and not sooner. And a word count set at chapter one and held, instead of discovered to be far too generous after sixty-three of them. That second one is going to fight me the whole way, but we will see.

Why I keep going back

Pantheon and Perpetūra are both worlds I would rather keep returning to than finish and shelve. I built them for myself first, which is the only reason either of them survived the parts where nothing was working.

I would love them to be worth your time as well, whoever you turn out to be.