What she wants and what she needs

Character sheets are long. Height, birthday, favorite food, the name of the dog she had at nine, a paragraph on her childhood, a Myers–Briggs type. Writers fill them in because filling them in feels like knowing the character, and a week later the character is still doing whatever the plot needs and nothing she does surprises anyone, including the writer.

The sheet fails for the same reason the synopsis fails in guide 17: it describes, and a character is not a description. A character is a distance — between what she is chasing and what would actually help her — and a story is what happens when the two pull apart. That distance can be written on a card in five lines, and once it is on a card you can hold it next to the scene cards and see, exactly, which scene makes her choose. If no scene does, you have found the sagging middle, and you have found it on a Tuesday afternoon instead of in the third draft.

You do not need any software for this. If you use software, the steps are the same.

The card

One card per character who matters — which is fewer than you think. Her name at the top. Then five questions.

1. What is she chasing? The external thing, the thing she would say if you asked. Ada wants to find her brother. Concrete, nameable, the engine of the plot.

2. What would actually help her — even if she doesn't know it? The internal thing, which she would not say, because she cannot see it. Ada needs to stop covering for the people she loves. If the answer to this question is the same as the answer to the first, you do not have a character yet; you have a goal with a name.

3. What happened to her before the story starts? One event, the old injury. Her mother left, and Ada, at eleven, told the neighbors she was visiting an aunt, for a year.

4. What does she believe that isn't true? The lie the injury taught her. If I keep the story straight, nobody gets hurt.

5. What does she need to figure out? The truth that answers the lie, which is usually the need in different words. Keeping the story straight is how people get hurt.

Under those, one more: does she change? Three answers: she changes for the better; she changes for the worse; she does not change — she changes the people around her. The third is not a cop-out. It is the shape of most detectives and several saints, and knowing you are writing it saves you from forcing a transformation on a character whose job is to be the fixed point.

That is the card. Voice, secrets, what she looks like, where she starts and where she ends: those go on the back, and they are useful, and they are not the engine.

Want against need is the engine

Read questions 1 and 2 together. Find her brother and stop covering for people. Now ask: is there a scene in which she cannot do both?

There has to be. That scene is the book. Ada finds the pawn ticket, and it tells her where Tomas is, and it also tells her what he did, and the only way to reach him is to tell the harbourmaster the truth about her brother. To get what she wants she must give up what she has always done — and if she keeps the story straight one more time, she keeps her brother lost. Yes, but or no, and; there is no she gets it available, and that is what makes it the scene the whole book has been walking toward.

If you cannot find that scene in your outline, one of three things is true:

• The want and the need do not actually conflict. Find her brother and learn to trust people can both be satisfied by the same act, so she never has to choose. Rewrite the need until it costs the want something.
• The scene exists and she gets away with it. She finds a third option, a friend intervenes, the harbourmaster already knew. The collision is set up and then dodged, and the reader feels the dodge as a cheat even if they cannot name it.
• The scene is missing. The outline goes from Ada finds the ticket to Ada finds Tomas, and the choice happened between chapters. This is the most common one, and it is the sagging middle: everything before it is setup and everything after it is consequence, and the thing they were setting up and following from was never written.

On the table, hold the character card beside the row of scene cards and read down question 5 of each scene — what's different for her by the end? — looking for the scene where the from-and-to is the lie → the truth. If there is not one, mark the gap with a blank card and write its name.

The lie needs a moment

Question 4 is a belief, and beliefs do not change by argument. They change by a moment in which the belief fails in front of her. If the card says she believes keeping the story straight protects people, there must be a scene in which keeping the story straight visibly hurts someone she was protecting, and she sees it, and the seeing is on the page.

That moment is a card too — a reveal, in the sense that something is disclosed to her, not to the reader — and it should be linked to the character whose lie it breaks. If the lie is on the card and no scene breaks it, she reaches the end believing what she believed on page one, and the change you wrote in the last chapter is asserted rather than earned.

Voice, briefly

On the back of the card: how she talks, and what she never says. The second half matters more. Every character has a sentence they will not say — Ada will never say I don't know where he is — and a scene that forces them to the edge of it is a scene with tension the reader can feel without knowing why.

The test for voice is not a paragraph of description. Lay the character cards side by side and read the how she talks line of each. If two say roughly the same thing — short sentences, dry, deflects with a joke — then on the page a reader will not be able to tell those two apart in a room together, and no amount of hair color will fix it.

What the card is not

It is not a biography. The event before the story is one event, chosen because it produced the lie. The rest of her childhood is yours to know and the reader's to infer.

It is not fixed. The want on the card at chapter one is not the want at chapter twenty; it sharpens, or turns out to have been a stand-in for something else. When it changes, change the card. The card is the version you can hold beside the scenes.

It is not for every character. The harbourmaster gets a name, a job in the story, and the sentence he will never say. He does not get a wound. A card that asks all five questions of a character with two scenes produces five answers you will never use.

Doing this in software

Software built on cards will give a character card these questions as fields, and — if it is built well — ask them as questions rather than as the craft terms (want, need, wound, the lie, the truth, arc type), which are precise and exclude anyone who has not read the books they come from. The stored answers are the same either way; the toggle between the two vocabularies is a relabeling, not a migration.

Then the collision becomes checkable. A scene card says whose eyes it is through and what is different for her by the end. A character card says what she is chasing and what she needs. A pass over both can say: what Ada is chasing and what she actually needs never pull against each other in any scene you've written down — so they never have to choose. Or: you've written what Ada believes that isn't true and what she needs to figure out — but no moment where she finds out is linked to her. Or, reading the voice lines across the cast: two of your characters talk the same way. On the page, how would a reader tell them apart?

None of these is a judgment. Each is the software noticing that a card's answer has no scene to land in — the same thing you noticed on the table with the character card in one hand and the row of scenes in front of you. The software's advantage is that it notices every time the scenes change, and says so in a sentence, and never gets tired of the question. What the scene should be, when it comes, is yours.