The scene card, and why a synopsis is not one

Most outlines are lists of what happens. Chapter 4: Ada searches Tomas's room and finds the letter. It is accurate, it is the thing you will write, and it tells you nothing about whether the scene works — which is why a novel can have a good outline and forty flat scenes, and why the flatness is invisible until a reader says I put it down around chapter four and cannot say why.

A synopsis describes an event. A scene is not an event. A scene is a change: someone wants something, something is in the way, and by the end the situation is different from how it began. If nothing is different, what you have is a summary sentence wearing a scene's clothes, and the reader feels the costume before you do.

The method here is old, and it fits on an index card. One card per scene, with five questions on the front. The questions are the ones every craft book eventually asks; the card's contribution is that it makes you answer them before you write the scene, in one line each, where the blank is visible. You do not need any software to do this. If you use software, the steps are the same.

Why the synopsis fools you

Because it is written in the past tense of a thing that already happened. Ada finds the letter. The sentence is complete; the scene is not. Whose scene is it? What was Ada trying to do when she found it — was she looking for the letter, or for something else, so that the letter is an interruption? What did finding it cost her? What is different for her when she puts it down?

A synopsis answers none of that, and it does not need to, because a synopsis is for someone else — an agent, an editor, the version of you who has forgotten the plot. A scene card is for the writer about to write the scene, and it asks the questions the writing will have to answer anyway. Answering them on a card takes ninety seconds. Discovering the answer is nothing on page 212 of a draft takes a month.

The card

Take a card. At the top, the scene's title — not its summary, its name: The letter, Down to the harbor, Tomas's room. Under it, five questions, one line each.

1. Whose eyes are we seeing this through? One name. A scene has a point of view even when the narration is omniscient; it is the person whose situation changes. If you cannot choose, you have two scenes, or a scene that belongs to nobody.

2. What does she want here? Not in the novel — in this scene, now, in the next twenty minutes of story time. Ada wants to find the pawn ticket before her father gets home. The want in the scene is small and concrete. The want in the novel (find Tomas) is what the small wants add up to.

3. What's in her way? Something, or someone, or herself. Her father is home early. The drawer is locked and she is the locksmith's apprentice, which is its own kind of shame. A scene with a want and no opposition is a scene where she walks across the room and picks the thing up. That is a sentence.

4. How does it end? Four answers, and only four:

• She gets it.
• She doesn't.
• She gets it, but it costs her.
• She doesn't, and it gets worse.

5. What's different for her by the end? From what, to what. From certain Tomas ran, to certain he was taken. From trusted, to suspected. From safe, to watched. One pair of words, and they must be different words.

That is the card. Five lines. Below them, if you want, the moment it turns and what cannot be undone afterward; those two are the questions a scene that works can always answer, and a scene that does not work cannot.

The fourth question is the one that teaches

Look at the four endings again. The first — she gets it — is the one most first drafts are made of, and it is almost always wrong. A scene where the character wants a thing and gets it is a scene the reader could have skipped, because nothing was at risk. It is a summary sentence: Ada found the letter. Write that sentence, put it at the top of the next scene, and move on.

The two that make stories are the third and the fourth. She gets it, but it costs her: Ada finds the letter, and her father finds her with the drawer open, and now he knows she is looking. She doesn't, and it gets worse: the drawer is empty, and the emptiness means someone was here first. Each of those ends the scene in a different place from where it began, and each of them makes the next scene necessary — which is the only thing a scene has to do.

Forcing the choice on the card, before writing, does something a craft book cannot: it makes the choice yours, scene by scene, on a Tuesday, instead of a principle you agree with and forget. If you lay forty scene cards in a row and thirty of them say she gets it, you have learned in four minutes what a reader would have taken four chapters to tell you.

The fifth question is the one that finds the dead scenes

What's different for her by the end? If the answer is the same as the scene before — she was afraid, and she is still afraid; she suspected, and she still suspects — the scene has not moved. It may be beautifully written. It is not earning its place, and two of them in a row is the exact spot where the reader put the book down.

On the table, this is easy to see. Write the from-and-to on each card, lay the cards in order, and read down the right-hand edge. Safe → watched. Watched → afraid. Afraid → afraid. There is the scene to cut, or combine, or give a turn it does not yet have.

What the card is not

It is not the scene. The card holds the scene's shape, so that you can see the shape apart from the prose. The prose goes on the page, later, and it will change everything except the answers to the five questions — and if it changes those, update the card, because the card is the version you can see forty of at once.

It is not a beat sheet. A beat sheet says where the midpoint goes. A scene card says what this scene does. You need both, and they are different objects; guide 19 is about the other one.

It is not a promise to keep the scene. Half the value of a card is that it can go in the cut pile without the prose going with it. A scene that answers all five questions badly is a scene you have learned something from, for the price of a card.

Doing this on the table

A stack of cards, one per scene you have or intend. Lay them in the order the book will go. Now read across the top: the names. Then read down question 1: whose eyes, scene by scene — six of Ada's, then one of her father's, then eleven of Ada's; you will see the shape of the point of view before you have written a line of it. Then question 4: the endings. Then question 5: the changes.

Each pass takes a minute, and each one tells you something the outline could not. That is the whole method. The rest of this guide is what happens when the cards live in software instead of on a table.

Doing this in software

Software built on cards will give a scene card these five questions as fields — and, if it is careful, will ask them as questions rather than labeling them with the craft vocabulary (value shift, scene goal, opposition), because a field a writer does not understand is a field left blank. The four endings should be four sentences to pick from, not four codes. The from-and-to should be two fields, not one, so that the same-as-last-time case is visible to the software as well as to you.

Then two things become possible that the table cannot do. A pass over the cards can say, in a sentence and without judging the writing, nothing changes for anyone in this scene — at least, nothing you've written down, or either nobody wants anything in this scene, or nothing is in their way — one of the two is blank, or she gets what she wants here, and nothing is linked that made it cost her anything. Each is a fact about the card, not a verdict on the scene, and each is the question the card was asking anyway. And the order of the scene cards can become the order of the draft: put the prose on the back of each card, arrange the fronts, and the manuscript compiles from the arrangement — so that moving a scene moves its prose, and the reverse outline is never out of date because it is the outline.

The scene card does not make a scene good. It makes a scene's shape visible before the prose hides it, and it makes a flat stretch something you can point at. What you do about it is the writing.