Serpentine Squiggles

I don’t know how to start this essay, which is why I think now is the perfect time to write it.

I’ve teased this one a few times. At this point, my coverage of writing theory is becoming almost comprehensive‍ ‍—‍ I’ve discussed some principle of writing sentences and paragraphs, and some broad points of structuring plots, even touched briefly on characters, so the biggest hole remaining is that middle ground: scenes, composed of those reified beats progressing the ur‍-​development.

Of the three, I’d argue this is the most difficult task. It’s not that hard to write a one‍-​off fragment of striking, vivid prose‍ ‍—‍ so often, they just come to you. Nor is it that hard to string together some vague outline of a plot. No, what’s hard is executing both in a single motion.

Once that snippet of prose runs out of steam, how do you continue it? What’s the next sentence, the next paragraph? You might have ideas, but nothing flows right, does it? And once you’ve outlined all the broad bits of plot, how do you, uh, actually start writing it? Most writers who find themselves “blocked” in some sense know what needs to happen next (in an abstract way, at least), but somehow the first line of the next scene isn’t coming.

Let’s do something about that.

Introduction 

Ultimately, the limitation of all previous writing essays is though they may be useful tools, I’d now say that they weren’t truly about writing. You could go on for chapters about color theory and brushstrokes, but that’s not how you draw a figure. It’s all ancillary.

So let’s try to talk about real writing, the actual process. Where to begin?

We can back up, try analyzing this from first principles. When you break down this problem, what do the core components look like? If someone reads a good piece of writing, they want to read it, necessarily. That’s our goal. Breaking that down further, they’d need a reason to start reading, need reasons to keep reading, and when they’re done, they’ll have reasons to think they got something out of it.

Let’s capture this all with a little metaphor. Imagine a bit of text you write is some rope used to capture a target. A cowboy’s lasso, maybe.

So, every effective passage in a story will hook the reader, pull them in, twist the line, tie it into other threads while leaving some to hang. What does all that mean, in practical terms? You catch the reader’s attention with some striking description, a unique scenario, then keep them interested by expanding on something they’re invested in, then surprise them by delivering novel, interesting information, then connect the new information to what they’re invested in, then leave them wanting more, so they keep reading

None of this is new insight (though I haven’t seen a metaphor as clean as mine; the freytag pyramid is close, if less compelling). Indeed, the middle three of this quintet (pull, twist, tie) is just synonymous for the presentation, juxtaposition, and generation steps of my ur‍-​development. The outer two simply relate the core triad to the rest of the piece.

Indeed, passages framed this way are fundamentally shaped around the larger context, the way contour lines points toward a vanishing point by the laws of perspective. The hook, the hang, it all amounts to a promise: that the story is setting up something to pay it off later.

We could call that distant payoff, what every scene is gesturing toward, the thesis‍ ‍—‍ a fairly stock standard term, as these things go.

So, if we’re to get started working out scenes from first principles, maybe we could begin by writing down this thesis, and then…

Well, there’s two directions I feel pulled to take this essay. There’s a cute parallel between them, because each of them entail asking:

What more do you or I even need to write?

If that hook/pull/twist/tie/hang metaphor were the main things this essay offers, well, I’ve just described it. Can I just say the rest is an exercise for the reader?

If you have your thesis, can’t you just write that? Why construct whole scenes to say, for instance, a hero’s anger problem is becoming a hindrance on their path to justice? Why not just say “The hero’s anger was growing too violent and uncontrolled, and needed to be reined in.”

Something you’ve noticed if you spent any time on this blog is that I take a while to get to the point. Even in my essays, my prose is florid and meandering, dotted with cutesy analogies, structured to build things up, draw them out instead of just getting on with it. Do I just like to hear myself talk, or does this serve a purpose? Permit me a mathematician’s answer to that. Still, I have a reason not to want to put things in the simplest way I can:

Concise communication is not effective communication.

To put something concisely, to distill a essay or scene or book down to a thesis statement, grants you a sentence made almost mathematically absolute‍ ‍—‍ take the sentence right above as an example. I stand by it, but the tone is so definite. It’s a concise statement of what these next few paragraphs are about to belabor, but it achieves that through a sleight of hand. You get these mathematical assertions by packing a lot of meaning tightly into individual words. What do I mean by “concise” or “effective”? I’m right mainly by defining these words in a particular way that rubs against conventional language.

A clearer statement, then, would take some time to lay out what kind of conciseness and efficacy I’m talking about, pin it all down. Words are vague tools, and it takes a lot of them to truly disambiguate.

Still, this isn’t very controversial‍ ‍—‍ sure, let’s not be so concise that we just deliver a pompous pronouncements without any explanation, but what’s stopping you from laying out an essay as merely a series of definitions that clarify subsequent assertions?

That, of course, isn’t enough, because anyone can define and assert anything. That doesn’t make it true. You need evidence, or at least convincing rhetoric. More than evidence, you need illustrations, grounding theories in specific instances. Now, while one must believe the goal of conciseness is slipping us by a little, we can augment our definitions and assertions with arguments, and make those as clean and simple as we can.

But the test of communication isn’t whether you can read something and understand it, it’s whether you can read it and understand it tomorrow, next week, a month from now.

A concise thesis statement packs so much meaning into individual words that if you forget a word, or the exact word used blurs, suddenly the whole assertion might be compromised. Avoiding that, then, is the strength of metaphors, of striking language which ensnares the attention‍ ‍—‍ by making numerous associations between ideas, by making the presentation interesting, it makes it memorable. That, I say, is the secret behind memorability: we remember things we care about, so every successful essay tells the reader why they care, or makes them care.

So, to really convey an idea, one needs motivations, definitions, arguments, illustrations, assertions, and we need to wrap it all up in an evocative presentation, all to lodge the idea in the readers’ brains.

If you take a look at the scrollbar, you can see you’ll be here a while. So forgive the verbosity, but I’m not just here to describe an idea or offer a take. The goal is to illustrate my thought process in such detail as to truly communicate it.

When you come down to it, summary is just the illusion of meaning. Or, at best, the ghost of a meaning found elsewhere.

Now, I included this long digression for a reason‍ ‍—‍ it’s clear I haven’t just been talking about my essays there, have I? But that, I think, I’ll leave to the negative space.

Part 1: The Formal Art 

Of People doing stuff in places 

A scene is what matters. I mean that both as an imperative and a description. You condense or outright ellide the stuff that doesn’t matter, and render as a scene the stuff that does. And what a scene does, fundamentally, is convince the reader that what’s important to the plot matters‍ ‍—‍ that is, it gets them invested.

Still, what is a scene? Where does one end and another begin? For the purpose of figuring out how to write a scene, it’s worth defining and knowing what we’re thinking about.

Scene is a medium. It is a form that limits what you can do, much in the way that, say, a script is a particular form with rules for what’s admissible and valid. In a scene, you’re often limiting yourself to a perspective, the knowledge contained in the skull of a single character and an account of their senses. And further, you’re limiting yourself to a particular slice of time and space on top of that, but despite those limitations, what results is a flexible and fruitful basis for storytelling. It achieves so many narrative ends‍ ‍—‍ revealing character, progressing plot, establishing themes‍ ‍—‍ through the simple means of people doing stuff in places.

Where does one scene end and another begin? If scenes are defined by people doing stuff in places, then one scene is a certain combination of people, stuff, and place. A scene lasts until there’s a notable change in the place (they go somewhere else) or the stuff (they finish what they’re doing), or the people (someone or several people dramatically arrive or depart).

Pedants among you might be tempted to probe the edges of this definition. What counts as a “different place”? What counts as “finishing what they’re doing”? But this is not a path that leads to insight. If you’re at the point of genuinely asking whether a different room of the same house counts as a different scene, you’re past the point where the label in itself is useful. Sometimes it is and sometimes it isn’t. These subtle shifts could be one beat in a larger scene, or the transition between scenes.

There’s only two places where the distinction matters. Either you’re wondering whether you should outline them as separate scenes, or you’re wondering whether you should put a scene break between them. And the answer to both is whichever gets the point across clearer.

To me, the dividing line is whether you could cut away from the scene to something unrelated and never cut back. If the scene is finished, this feels fine, but if there’s a pull to see a little more to know how things turn out, or it feels like the scene ended before anything happened, it’s unfinished.

Think of them like sentences. Where does one sentence end and another begin? There’s plenty of sentences where you could chop it off in the middle and not notice the cut; no grammar has been violated. Likewise, stories operate off of a kind of “scene grammar,” but it’s so much harder to discuss, because while you can hold an entire sentence in your hand, the only way to think about scenes this way is to abstract, omit details. (And as I said, summaries are illusionary ghosts.)

Grammatically speaking, then, what do scenes need? Where is the line between a short scene and a “scene fragment?” A lot of writing manuals will tell you the answer is “change”. That every scene needs to change something in the story‍-​world, move some part of the plot forward. I certainly think this is good form; things should be changing at every opportunity, progressing constantly, but there is some strain, in a definition like this. Do flashback scenes change things? Imagine a scene that does nothing but describe just enough to deliver a twist reveal that recontextualizes the whole plot, without actually changing anything in a material sense (other than the reader’s perceptions, I suppose, if you must contort yourself).

What carves things closer to the joints, I think, is a word I’ve seen used by a few others‍ ‍—‍ scene are defined by their “turn”. A given scene will turn on some action or reveal or what have you.

Let’s return to the five point rope model. Hook, pull, twist, tie, hang. This idea of “turns” feels a lot like the ‘twist’ step. With some effort, you can see the twist is the only necessary component needed for a scene to turn. Hooks are structurally inevitable‍ ‍—‍ what’s the difference between the absence of a hook and a bad, ineffective hook? A hook can always fail, it’s subjective, so hooks aren’t grammatically necessary, it’s just a useful concept to have when working.

A good twist has a pull and a tie by necessity‍ ‍—‍ if you convey “the Blue Hand was behind the kidnappings all along” the reader should not ask “Who’re the Blue Hand? What kidnappings?”. If the twist means anything, we care about what’s twisting, and if the twist means anything, it will have implications. If the protagonists wanted to stop the kidnappings, then this twist ties back into their motivations, recontextualizes their previous interactions with the Blue Hand. (Though, conversely, this implies a freestanding twist without a proper pull or tie somewhere isn’t quite a full turn on its own.) And hangs are, again, structurally inevitable. If there’s still story remaining to tell, you’re leaving something hanging, some question unanswered, a thread untied, and the question is how explicit you’re making that.

So all you need for a scene to turn is a twist, the rest is presentation.

Can we do a similar compression with “people doing stuff in places”‍ ‍—‍ which of those three is most important? One can imagine scenes without explicit people (a vivid description of a storm building in intensity, snow and hail piling up, until a fateful lightning strike yields enough destructive power to collapse the only bridge that could take the heroes to their destination). One can imagine scenes without explicit places (a scene consisting only of “white room” dialogue. Alternatively, there are epistolary segments or diary entries‍ ‍—‍ arguably these aren’t scenes in form, though). A scene with nothing doing anything is tricky (environmental descriptions in the style of Lost Cities (I brought it up again; take a shot)? Most of those contextualize the environments with histories of stuff done, though).

I think this is another blind alley, though. I’m only so interested in probing the technical precision of a definition of scene. The utility of reducing the five point model to a consequents of the twist is that it allows for a very quick sanity check: “Does this scene turn? If not, this is a scene fragment”. But no such value can be gained here for the simple reason that when you write scenes, they have people doing stuff in places, nearly without exception. “Oops, I forgot to have anything change” is a failure mode worth worrying about in a way “Oops, I forgot to include any characters in this scene” isn’t.

(If you must know, I suspect it’s ‘doing stuff’ that bears the load‍ ‍—‍ stuff generally must be done by someone, and it really must be done in a place of some sort to have an affect, so at its most elliptical, ‘stuff done with a twist’ constitutes the minimal scene.)

But enough theoretical discussion. How do you bring all of these elements together into a well‍-​formed scene?

Of Setting Scenes 

In my essay on outlining, I didn’t have much to say about what I called the fourth stage: arrangement. I brushed this off as being because it’s a step I elide or skip, preceding directly from an elaborated idea of a scene sequence to a finished chapter with no formal steps in between. The truth was I just didn’t know how to outline scenes. I got better, though.

I’ve found the simple formula for a scene goes like this:

In [location], [characters] are [activity]; [focal char] wants [goal] but [obstacle] so [result].

Two clauses‍ ‍—‍ the framing and the plot‍ ‍—‍ joined together into one sentence. It’s easy enough (but not too easy) to write these, and they are immediately effective. When I set down and try to formulate an entire arc this way, I soon find that my ideas for scenes are no more than scene fragments. It’ll be stuff like “characters finds the villain’s book”, or “mentor teaches character a new spell”, or “they interview a suspect.” Where is the ‘but’? Where is the ‘so’? These are floating premises or conclusions, with no meat to them.

The utility of this diagnostic must be recognized: so often have I sat on a blank page, knowing the next scene I need to write, but unable to think of something worth writing. Most writing problems are this simple to solve: all you have to remember is the fundamentals. What does the focal character want? What is the conflict? How do they overcome that? When a scene isn’t coming together, you can fix most of it by figuring out and aligning out these elements.

Aligning them is the important part. It’s easy to think you’ve got a goal for your characters, a conflict for them to work through, a solution to pull it all together, but be flat out wrong.

We have cached ideas of what conflict looks like. It’s people fighting, arguing. But look more carefully. Does the supposed conflict actually stand in the way of goal? (How do the two connect?) Does engaging in it flow out of pursuit of the goal? (Could they not just sidestep or ignore it?) Is the goal actually at stake? (What threat does it truly pose?) Or is it all just sound and fury?

It’s also easy to pick the wrong goal. Think carefully about what the reasonable thing for a character to want would be, what plan sees them exercising their entire willful, rational capacity, rather than patterns of tropes or what’s most convenient for the plot you have planned. And paradoxically, consider some clever ideas for what the character really wants. Are they arguing because they genuinely disagree, or could they just be asserting dominance, venting their frustration, or projecting another, hidden feeling? Do they even realize what their real motivation is?

When it comes to tying off a scene, writing the conclusion… the first thing to determine is what’s remotely unexpected about it. If there’s a fight and the protagonists win, what’s the point? Did they lose something? Did they learn something from the people that attacked? Did the coming of age hero finally pull off the technique they’ve been trying to master?

A degenerate triangle has two or more vertices laying on the same line, and as a result it has no area: it’s just a line segment with extra steps. Similarly, a degenerate scene is technically well‍-​formed (it’s no fragment), but the outcome lines along the same throughline as the goal or the conflict.

It’s something that instinctively bothers me. Writing what amounts to “they wanted to do a thing. And then they did it.” The reverse is only marginally better‍ ‍—‍ “they wanted to do a thing and then they can’t.”‍ ‍—‍ if only because it’s hard to write failure in a way that isn’t just setup for changing approach and then winning a different way. Yet it still can feel like a waste of time.

This is the nature of scene as a medium. Consider the principle of sensory reification. If we are to present a character’s emotions and other abstract qualities, convention demands that we express this in the form of sensory images, lines carefully picked to reflect and imply the notion we want to articulate. We show instead of telling.

There is a similar principle of, let’s call it scene dialectics. In its ideal form, it holds that neither a character’s goal nor its obstacle can stand unchanged at the end of a scene. Imagine them as reactive chemicals, each transforming the other, combusting to generate the heat that is dramatic tension.

There’s something almost puzzle‍-​like about this principle. (Indeed, it calls to mind my old sudoku metaphor.) If my characters walk into a scene wanting something, I try find some way for them to not get it without the plot breaking. From the opposite end, if my plot needs something to happen in a scene, then I need to set up my scene so that it’s somehow not what my characters want. Now, it’d be hard to sustain anything more serious than a comedy or melodrama if nothing the characters want ever happened, if everything that happened was unexpected and undesired. So the true statement of scene dialectics must be in fact subtler, less surprising, less controversial.

When a character seeks a particular result with particular consequences by a particular approach, they cannot get all three. This is flexible enough you can already see the possibilities it parameterizes. You can have a scene where the protagonist gets what they want, but only by doing it in a way they don’t want, or (a classic) what they want actually has unexpected consequences. You can have disastrous scenes where the protagonist stubbornly sticks to their preferred approach, and it just doesn’t work. The last permutation is curious, but familiar with some thought. A scene where the approach doesn’t work and they don’t get what they want… but somehow they get the consequences they want? But it’s the familiar case of a “not what you want, but what you need” aesop. Or an example: gearing up to beat the villain’s henchmen in order to foil their a new scheme, only to end up swaying the henchmen, gaining a new ally instead of beating them up.

A useful technique for avoiding degenerate scene is what I call overshooting. If characters want something, and according to the plot you have planned, need to get it, there may be no way to have them “partially” succeed and still have the plot work. What then? Instead of changing the result, you can change the goal. Have the character hope for more, have plans more ambitious than what they’ll achieve. Then it looks like you’re having them fail, when they’re simply getting the victory you originally planned.

By contrast, the risk with degenerate failure is that it just feels like a waste of time, a detour for the sake of a detour. The heroes’ want to do something, but there’s an obstacle and oh no, there’s nothing they can do about it. And sometimes that’s how the plot has to go, the heroes can’t overcome that obstacle and need to try something else. There are easy ways to mitigate this‍ ‍—‍ perhaps they gain some valuable information other than “this won’t work lol”. Maybe they impact the obstacle even though they can’t fully overcome it. But if nothing else, you can try the inverse of the overshooting technique, and change the stakes. Make the obstacle more dangerous, threaten greater peril, such that merely getting away having achieved nothing is a good outcome the heroes are glad to have managed.

Practically, the takeaway here is that if you wish to establish something via scene, do not introduce it as an inert fact‍ ‍—‍ demonstrate it in action, reveal it via the development of a conflict. Consider the scene ideas I mentioned earlier:

A character finds the villain’s book. This can be the goal: the character wants to find the book, but the villain’s lair is filled with traps. This can be the obstacle. Character wants to wash their hands of this plot and leave the villain to be taken down by other, more capable people‍ ‍—‍ but they find the villain’s book, so they feel obligated to bring them to justice. And this can be the result. Character has tracked the villain down to their lair, but the villain got forewarning and has escaped, so the character only finds their book.

Now consider the second: the mentor teaches character a new spell. Perhaps the possibilities are already dancing before your eyes: the mentor wants to teach, but the student resists learning, so the mentor has to convince them. Maybe you’re getting bored of this, so we can try being a slight bit more imaginative with it. Character wants the mentor to teach, but they’re being hunted by assassins, so barely have any downtime to focus on it. A character’s love interest wants to go on a date, but the mentor wants to teach the character a new spell, so they have to chose between or balance the two options. A monster is rampaging through the countryside, but the character has no means of defeating it, so the mentor teaches them a new spell.

There’s been a few missteps among those more “imaginative” suggestions. Can you spot the problem?

Call these blurry scenes. Something that rears its head if you chew on those proposed scene plots is that they don’t quite work. Try to imagine a single scene that introduces wanting to learn a spell, establishes assassins hunting the protagonists, and faithfully demonstrates a state of “no downtime.” Try to walk through a dialogue where one character is asked for a date, and then their mentor suggests a new a lesson, and they’re left deciding between the two. These aren’t impossible asks (indeed, limitations require creativity). But it’s a lot to handle in a single scene. Remember, a scene is specific people doing certain stuff in a particular place.

So I’d argue, for instance, the date vs lesson conflict is closer to two: they could work up the courage to ask for and work out the details of a date in one scene, followed by another scene where the mentor makes their demand.

The final scene idea I proposed, I think is flat out infeasible without remarkably compressed pacing. How do you present a monster rampaging through town, demonstrate that the main characters can’t do anything about it, then show them learning a new spell, all in one scene? (Perhaps if the monster is a nuisance, and the spell is a mere cantrip, it would be appropriate.)

To my ear, this is three scene ideas wrapped together, waiting to blow up on you when you dig into them. In this case, it’s a fairly workable example‍ ‍—‍ these three scene would share a lot of DNA when it comes to goals and motivations, so this sentences gives you most of what you need to work with. But it’s easy to imagine if the hidden complexity were a little bit greater, the idea of sitting down to write this scene would be more daunting‍ ‍—‍ or worse, you’d find yourself stumped at where to start.

Think about it. I estimate this idea could be executed in three scenes, but look closer, really imagine how you’d actually write it, and suddenly that seems naively optimistic. Is a monster rampaging through town one scene? Would we not need to establish why the readers should care about what happens in this town? Might we need to establish some lore about how this sort of monster works in this setting, that might be awkward to introduce in the middle of a rampage? When do the characters learn about the monster and how do they arrive at the site of the rampage and how do they escape the rampage after they realize they’re outclassed? If the mentor knows the spell, why can’t they just cast it themselves‍ ‍—‍ are they at the site of the rampage, or do the heroes need to travel to them?

(Stories always seem to turn out longer than you plan them. It’s this unexpected complexity that reigns as the devil in the details.)

So, how do you avoid packing unexpected complexity into a scene?

Well, the first thing to note is that what I’ve been spitballing here aren’t scene formulae as I’ve defined them, and the fact that they aren’t is important‍ ‍—‍ a scene is people doing stuff in a place. You must write your scene formulae by defining the location, the characters involved, and the activity performed over the course of the scene. Were I this rigorous in my spitballing the ridiculousness would be readily apparent. I would be forced to describe the “go on a date or learn magic” scene as though the love interest and the mentor are in the same room, looking expectantly at the character while they decide who to reject. I would be forced to frame the village rampage as though the characters are struggling to learn new magic even as a monster is still maiming children and women can be heard screaming in the distance.

Thinking about the space and time you need to block off for a scene to be effectively communicated is the key to formulating scenes that can be painlessly executed.

There’s another principle to keeping things simple: scenes are about action, not abstraction. It’s bad form to write the conflict of a scene as “a monster is rampaging through a village”. It’s too blurred, accurate but too encompassing. What is a rampage? It’s not something that happens in a scene, it’s something that happens throughout a scene. I find it more helpful to engage in metonymy: say the monster destroys the tavern, or that it kills the blacksmith. Sure, the threat a monster poses is to the entire settlement‍ ‍—‍ it’ll surely destroy more buildings, kill more people. But imagine if you had to represent it with a single panel of a comic, a single, brief shot in a movie. What would be contained in it? What’s the most striking, the most characteristic, what would be enough to give the viewer the general idea?

Let’s be more specific. For the goal, what’s the first action the character takes toward that goal? What action makes it clear they’re committed to this goal in the scene? Do they throw the first punch, voice the first accusation, pull out their magnifying glass in the hunt for clues?

When does it become obvious the obstacle needs to be dealt with, when the nature of the adversity is clear and undeniable? And what demonstrates the conclusion most generally‍ ‍—‍ the enemies lying dead or defeated on the ground, the hero limping away dripping out lifeblood, the suspect storming off with tears in their eyes, the murder weapon lying hidden with a fingerprint traced in the dust?

I find there’s more direction to be had in outlining scenes as a series of landmarks to pass by, rather than a vague semantic mist that abstractly floats around the route you’ll take. There’s so many ways to rampage, but only a few of them see the roof torn off a tavern, monstrous jaws closing around a blacksmith, and if you know that happens, you’re already thinking about the blocking‍ ‍—‍ where does the viewpoint need to be the see that as it happens, where are the crowds of other patrons running? The tavern is surrounded by other buildings‍ ‍—‍ which route does it take?

That’s all it takes to start stirring up a scene.

Of Stirring Scenes 

We’ve covered dialectics enough that the last scene idea can easily be couched in such a form‍ ‍—‍ Max knows Daisy is the culprit, but has to interview her to proof it, and he doesn’t get the info he needs to. Dante wants to interview Luca, but Luca fears he’ll be falsely suspected due to his lack of alibi, so Luca evades and dissembles, looking all the more suspicious for it. Michelle had gotten a distress call from a storefront, but arrives after the criminals have departed, so she interviews Janus as a witness.

We’ve learned that to make these proper scene formulae they need not just plot‍ ‍—‍ goal/obstacle/result triads‍ ‍—‍ but framing. But the benefits we’ve thus far outlined‍ ‍—‍ awareness of the logistics of what happens where, grounding scenes so their complexity isn’t hidden‍ ‍—‍ can be achieved by just setting these encounters in nondescript rooms or random fields. And indeed, this is the tendency I’ve noticed easily emerges in plans, when the only thing on your mind is the plot and its progression.

Even in scenes like these that are just people talking, picking the right place informs the tone of the scene. Consider whether Max confronts Daisy in her own home, showing up at her doorstep, or summons her to his office. Imagine Dante interviewing Luca in a traincar, where there’s no room for Luca to escape; or in the corner of a large party, where there’s a risk anyone might overhear.

It’s easy to pick a location sufficient for a scene to take place. Sometimes, there’s a location that’s necessary (the villain only has one lair). But when there’s flexibility (and usually there’s flexibility), you should think about what makes for the best scene.

Easy to say “put some thought into it.” What sort of thing are you thinking about? There’s a number of heuristics I could give‍ ‍—‍ pick locations that are striking, are cool to describe, pick locations that heighten the drama of the conflict about to take place, pick locations that define how the scene could play out.

But the real principle here is that you don’t pick locations, you bind them. You formulate a scene as a whole, and the location isn’t an independent variable‍ ‍—‍ that is to say, the location is entangled with the characters, the activity they’re performing, the conflict.

Of course, the practical reality is that one of them comes first. Perhaps you’ve got a plot beat to render and you wonder where it happens, or you know the characters are passing through a major location and something happens and the details need to be fleshed out.

When deciding locations, consider what locations have significance to the characters. Where would they want to handle this conflict. What props does this set offer them? In her home, Daisy could pour tea to calm herself; or perhaps in his office, Max has a giant grandfather clock that ticks ominous doom throughout. Having a heated questioning take place during a party could entail the characters making a scene at the climax‍ ‍—‍ a crowd watching, the drama ruining the reputation of one or both of the characters. Sometimes, a location is good just for novelty. Perhaps Max shows up to Daisy’s job, questions her while she’s busy administering a company’s computers (updating their software one by one, perhaps?) and this activity gives you something to describe throughout, intercutting the dialogue.

When crafting a conflict to fit a location, think about how characters would judge the location. They travel through a swamp, and does the viewpoint love the vibe of swamps? They’re stuck in traffic‍ ‍—‍ will they join the drivers in rubbernecking? What station are they listening to? Altogether, what knobs does this location offer you to twist? What resolutions to the conflict do they incentivize, and which become nonstarters?

The target you’re aiming for is stirring the different elements of your scene until they dissolve or bind together into a inseparable mix. I find that a scene is ready to write when free variables have been bound, and there’s only one way things could really play out. There’s nothing left to decide, only things to express.

And that’s when you really begin to spin scenes.

Sidebar: Of Weaving Scenes 

I leave this here not to inform you, but as a disclaimer of incompleteness. If this article were repackaging the thoughts of others, dispensing conventional wisdom, I would have something to say about weaving scenes. I might introduce that common concept of scene polarity, that scenes go from positive to negative, or negative to positive (or negative to more negative, or—), and how you should balance scene moods to avoid fatiguing the reader, or to heighten tension. I might have something to say about managing scene length, and how that impacts a story’s pacing. I ought to have anything to say about the emotional impact of scenes on the reader, the rise and fall of passions it produces.

But I don’t know how to do any of that. It doesn’t reflect in how I think about scenes, and the concepts have never cropped up in my musing. I’ve seen enough writing manuals discuss these concepts to suspect there’s something there, but I’m not going to proselytize about something I don’t know for sure works. So in the interest of honesty, I have nothing to say in particular about how you arrange your scenes, and that absence is worth noticing.

[Since first writing this post, I have gone on to dissect the concept of pacing.]

Interlude: A Worked Out Example 

Part of the specificity I’m aiming for‍ ‍—‍ really communicating my thought process‍ ‍—‍ cries out for a thorough example. It’s something I avoided at first‍ ‍—‍ I used to feel I was clear enough without examples, and I actually tend to skip them in articles I myself read‍ ‍—‍ but the comet girl story in my previous article was pretty well received, so let’s see some scene formulae in action as I try to make that fic come to life in more detail.

For those who don’t have the context from the essay on outlines, the concept of the story is something like this:

When magical storms besiege a kingdom, a swordswoman born of a comet demands to know why her newfound power lies entangled with them. But will she live to find answers when ignorant clergy call for an inquisition against her? Only when she accepts her nature as a comet will she find answers lies in her mother’s past and the stars beyond.

(I included a much longer overview afterward, I won’t eat up space quoting it.)

So, how would we go about formulating a couple scenes for this story?

The first obstacle I encounter is (as I’ve said in correspondence to those asking if I planned on actually writing it) is that this story was made to be generic. The core concepts are a bit neat‍ ‍—‍ comet lore really gets my imagination going‍ ‍—‍ but the prospect of trying to belabor it in prose just seemed boring.

But as of this article, my ambitions are pretty high‍ ‍—‍ I want to outline the whole thing, just to flex my skills. So my first step is to get myself interested in it. So I took a walk and brainstormed some cool details.

It started with merely evocative details. What if, when Aurora is using her magic, she’s enveloped in mist like the tail of a comet? What if the faith of earth and sun have their own magic for growing crops, and that’s where their legitimacy comes from? What if they worship dryad‍-​like nature spirits who have suspicious parallels with comet spirits?

What if‍ ‍—‍ well. I don’t want to spoil everything I came up with. Suffice it to say I came up with some stuff far neater than what I’ve shared thus far. (And importantly, stuff that’s useful for working out scenes.) Enough that I was decently motivated for the story I was going to outline.

Let’s get some high level stuff out of the way first. The idea here is to plan out a novel, not a series or serial, but a quick, fast‍-​paced story that goes for a bit and then ends. I’m aiming outline less than 30 chapters, fit for a 60‍-​90k word end product.

In terms of the actual story, let’s divide it into three acts. Act 1 sees Aurora discover her power, get training, run from the inquisition, and hm, let’s cap it off with a confrontation with the friar that called for her persecution to begin with. Act 2 will have her adventures wandering as a knight errant. This is where I include (and twist) some of the ideas I floated in the original article (another comet landing; a sorcerer that reveals the secret lore). Act 3 is everything leading up to the final showdown on the moon.

Now, one place to start would be to outline scenes in order. This how I proceed with my serious projects. This is ideal for achieving a kind of flow, for really exploring the connective tissue between one event and another. That’s not quite what we’re doing here; this story is going to skip and elide a lot, and I only care about the key events. It’s possible by the end of this my outline isn’t even complete, but a mere skeleton of major scenes that a serious execution would interpolate. (Already, I can imagine this story sprawling into a trilogy if I let it.)

So instead, the first thing I want to think about is the climatic scenes of each part. I’m planning to end each part with a big fight, and this immediately present us with something instructive.

How do you outline fight scenes?

On the face of it, these are the simplest to write formulae for. The goal of every basic fight scene is to beat up the other guy, and the obstacle is the other guy is trying to beat you up, and the result is whoever wins. Add some scenery to that, and we’re done, no?

Don’t do this. A perfect example of a blurry scene, such an outline will be of no help to you actually writing it. No, the goal of a fight scene is more than “beat the other guy up”. I’ve written about this before, but I have more to say now. Key to the goal of a fight scene is the approach and assumptions that inform how each fighter approach the fight. What the fight means to them, and how they need to go after that.

Take the first major fight of this story, between Aurora and the friar. I won’t get into how or why they fight just yet, but with some noodling, I can conclude a decent way to write this is that both party walk into this fight with something to prove. The friar has staked his career on making an example of this cometspawn‍ ‍—‍ if he finally brings her in, he gets promoted to high priest, but if he fails, he’ll never be rid of the humiliation. Aurora, meanwhile, has finally realized there’s no escaping the reach of the faith of earth and sun, so instead she challenges it, wielding the gemstone sword of her late mentor‍ ‍—‍ if she can best a trained friar using magic the republic frowns upon, she can demonstrate her power and willingness to help (by sparing the friar), thus proving that she’ll be more help to people alive than dead. In a way, it’s a battle for the supremacy of two magic systems.

Earlier in this article, I floated the idea of a story’s “thesis”. In truth, this was just a setup for the ensuing ramble, but it’s worth reviving that term‍ ‍—‍ it’s fair to call these iconic fight scenes I’ve thought up the “thesis scenes.” I can formulate them in a vacuum, and the rest of the story will serve to contextualize and set up them: they are the ultimate payoff. But it’s hard to come up with that transitional connective tissue from first principles, so how do we come up with the rest of it?

I did three passes. Well, four, but one was pretty trivial. On the “zeroth” pass, I opted to save myself work, and looked up a 27 chapter novel outline. These aren’t much more than “goal, tension, resolution” triads triply nested‍ ‍—‍ which is to say, three acts, with three sub‍-​acts within them, and three chapters in each sub‍-​act. It’s not a structure I believe in, really, I just wanted to color in some lines, so I took it as a skeleton. For each chapter, I penciled in a basic ‘main event’ that’d happen. Chapter 9 was the fight with the friar, chapter 27 with the moon queen. All told, the result is less than 100 words.

For the first real pass, I got cracking on what scenes I’d actually need. I’d need a scene where the protagonist meets their mentor figure, scenes of them training, of them on the run from the inquisition. Overall, the goal was 2‍-​4 scenes per chapter. Importantly, I only sketched out scene fragments. An idea for a goal or obstacle or result, but never all three. That comes later. This wound up a bit more than 800 words. All told, what I’ve done so far could easily be knocked out in a day.

The the second pass, we plot out the scenes. They would not yet be people doing stuff in places, but a few specific details will creep in by nature of what plotting entails. This is where I hook, pull, twist, and tie, then leave it to hang.

And this is where things went off the rails for me. Remember, detail is infectious.

You see, despite my proselytizing, I don’t think in goal‍-​obstacle‍-​result formulae. Most organic scenes can’t be expressed that way, not clearly, accurately and briefly. All scenes, I think, can be framed that way, but what’s the square root of ‍-​1? As a real number, it can’t be represented, but it can be expressed as a matrix of four real numbers. Similarly, when you work with writing tools like scene formulae or the ur‍-​development, some components of the structure are actually complex semantic matrices which could best be explained in a many sentences‍-​ or paragraphs‍-​long elaboration. (The best, shortest description, of course, would be the story itself.)

So my second pass of (what was) the first scene looked like this:

It’s a hot day, and Aurora watches the sunset, she and her father pointing out stars. Every star she points to, he can name. All the planets, and their moons, too. She asks what the planets look like up close, and he can tell her. She eventually asks “How do you know so much about the heavens, daddy?” And he doesn’t answer directly. She presses, culminating in “It’s related to mother, isn’t it?” He gets quiet. Refuses to answer any other questions. Aurora’s feelings intensify, quick to anger because she feels so hot, or heating up because she’s so angry. “Why are you always like this?” she yells. “Why can’t you just tell me about her?” She runs off.

You can find the goal, conflict, and result in there pretty easily, but it flows as if paraphrased from the full story. So I wasn’t formulating scenes here, I was summarizing them, and this was less outlining than it was taking notes with style.

If all of my scenes looked like this, it’d be fine. Although you may be getting uneasy‍ ‍—‍ that paragraph above was over a hundred words long, for one scene. Do some math, and you’d expect the full thing to be over eight thousand words. Is that why this essay is so long?

The first act of the story undershot these estimates, mainly because most chapters only had two scenes and maybe a little scene fragment as a cliffhanger. It was “merely” 1.7k words. That was all I managed to get done that day. I took a brief break, and returned to it two days later to knock out at least the second act. I got through the first couple of chapters‍ ‍—‍ over 2k words. Part of this was passion. I had grown to genuinely like this project, enjoying and wanting to make it good. Character voices were coming to me, and I was adding plot beats not for necessity but because it improved the result. I had also grown more ambitious: chapters were becoming more like episodes than installments. (or in the terminology of Brandon Sanderson, closer to epic pacing rather than thriller pacing.) All to say, I started writing entire micro‍-​arcs progressed and resolved within, and once I wrote one, it would feel weird of the other chapters were more straightforward.

Is this how bloat starts? :c

After almost a week of work, I powered through to the end, transcribing entire conversations at several points. The result was… 25k words.

And this was the second pass. These still weren’t scene formulae‍ ‍—‍ I planned to include framing in the third pass.

The essay is going to be memetically long, but I don’t intent to include me rambling about my novel for several times the length of this essay. It’s good, and you want to read about it, but it wouldn’t actually convey my points better than a more judicious excerpt. So let’s scale things back, and look at the first arc.

(Update: As one might expect from how much effort I’ve poured into it, I’ve become more serious about making Aurora Moonrise into a real fic. You can read a teaser here, a working draft of the first few chapters. I think it’s quite nice! Reading the summary included as part of this essay will spoil some of the fic. Though given the contrast you can already find between the earliest conception and the current outline, the finished draft will be richer and divergent from this summary.)

Aurora Moonrise, Arc One 

What Was the Point of That? 

Did I just make you read four thousand words of meandering fantasy with the thinnest veneer of being relevant to what this essay is about? Yes, but there’s value, I feel, in seeing organic scene formulae in the flesh. I could have cooked up some one‍-​off examples for this article, but there’s something misleading about that. I presented scene formulae as one sentence with four clauses. A concocted example would cleanly fit into that mold, but it would serve a purely pedagogical purpose. And I’m the one who is writing essays‍ ‍—‍ you’re here because you’re writing stories. So I showed you the gorey details of how long‍-​running context gets worked into scene formulae, how the excess of detail strains and buckles of the form I’m casting them into. (Nine chapters is excessive, but multiple chapters were necessary, and it didn’t feel right stopping in the middle.)

Point is, scene formulae aren’t more than a starting point‍ ‍—‍ what I’ve written above is perfectly serviceable synopsis. If you showed it to someone out of context, I’m not sure they’d see the structure, think there is anything special about these summaries.

The real utility of scene formulae became clear to me as I wrote, because several of those scenes, even after my second pass, were blurry, or I made the creative decision on the spot to invent new scenes. (Examples: meeting the archer was initially just a quick conversation; most of the scenes between finding the cabin and the archer leaving were improvised on the spot; and finally, the fight with the jeweler was originally a simple conversation, and the memory of cassiopeia was also improvised. Overall, the version I pasted into this document was initially a reasonable six chapters.)

Writing is hard, but formulas can be practiced. Several times I sat stumped at how to cut a particular scene, and just thought: okay, where is this happening? Who all is there? What can they do there? What’s the conflict?

In [location], [characters] are [activity]; [focal char] wants [goal], but [obstacle], so [result].

You can fill in the blanks and have something viable. There’s a power in that. I wrote 25k words in 6 days pumping out partial formulas mixed with detailed notes‍ ‍—‍ skill and experience (and a dash of hyperfocus) played a role in that, but how much? Give a try yourself. Perhaps go for a nine chapter outline, instead of this ever‍-​bloating 27 “chapter” mess. Dream up a thesis scene, sketch out some fragments that chart a path that builds to that, then see how much trouble a couple dozen scene formula give you. I’d love to have more data points here.

If I’m right, and you’ve outlined a sequence for a new or existing work, your work clearly isn’t done. I’d hazard my outline above wasn’t unpleasant to read, but it’s not really a conventional story, is it?

It’s time to draw the rest of the owl. How exactly do you turn hundreds words of scene fragment into thousands of words of scene?

Part 2: Of Spinning Scenes 

Wherever you are, tap out a little rhythm. One, two, three, four. Go for a few rounds to entrain it, then with your other hand add another hit once per measure, doubling a beat or perhaps accenting an ‘and’ (one and two and three and four). This is completely unnecessary, but what I find compelling about this demonstration is how, once you really lock into the groove, it doesn’t feel conscious, but your other hand moves infrequently enough that you can pay attention to it, feel the rhythm just… leap out of you.

When the writing is going good, that’s what it feels like. And when the writing is good, that’s what reading it should feel like. Rhythm features heavily in my writing instruction‍ ‍—‍ I seem obsessed with this idea of conjunctions: scenes must be “this but that so thing”, plots must be alternating ‘but’s and ’so’s, I clearly find it deeply compelling. And I should: writing in this fashion produces flow, and an interesting, inevitable flow at that. It’s a recognizably poor form of writing that merely goes “this and that and thing”, or “then this happened, then that happened, then…” without tying then together tighter than a list of things or sequence of ’well that just happened’ events.

But for all I’m so in love with it, actual prose isn’t structured like this. You’re reading it right now: most of my sentences don’t start with ‘and’ or ‘but’, and those conjunctions aren’t present in all of them. I’m belaboring the obvious‍ ‍—‍ “The wooden door was locked. Micheal had a lighter.”‍ ‍—‍ the connection between sentences can be obvious, implicit. It tends to be, in fluent writing.

Prose has rhythm. I’m not talking about scansion, the way stress syllables create iambs and anapests, but that’s an important deep art for writing good prose. No, I mean the rhythm of ideas between one sentence and the next.

I’m focusing on this point, because it’s important. This is what flow is. Prose flows when there’s an organic pulling and twisting rhythm between the underlying ideas of a piece. You don’t hop randomly just from topic to topic, you don’t skip steps, every sentence feels like it adds onto a previous sentence somewhere, answering questions already raised, or providing contrast and challenging established ideas, or tying together two unresolved ideas left hanging, and with each motion dragging you further into the piece.

Reading and writing are often united in this. Except, only sometimes. That’s the rub, isn’t it? Obviously, there’s inevitable gaps in your first draft flow you only noticed on reread, sometimes you skip steps just to get the words out, tying them together later on. But this isn’t an article on editing. No, the beast we’re here to slay is the most pressing.

Sometimes, you write and then stop as if on the edge of a cliff. There’s nowhere further to write. Just blank space that must be filled, but with what?

I’d go as far as to say writer’s block‍ ‍—‍ when it persists in spite of having energy, rest, and the desire to write‍ ‍—‍ always comes from asking the wrong question, struggling to solve a problem by approaching it from an impossible angle. If it’s unclear how to start the next paragraph, if nothing flows out, you may have just written yourself into a corner.

So instead of trying to fill that void, crossing that cliff’s edge, you should backtrack. This doesn’t mean total rewrites, per se; it can be as simple as planting a thread a few paragraphs back that you’ll pull on right when all else runs out of steam.

But this is pretty low level advice, when the problem we’re solving is the general method of going from formula to scene. So I think I started this section in the wrong place, actually. Let’s back up.

A Mind in Dialogue With the World 

Far back at the beginning of this essay, I remarked that previous efforts on this blog produced useful tools to help writing, but fell short of actually addressing the process of writing itself. Outlines aren’t writing, characters models aren’t writing, and revising telling into showing is close, but not writing.

But all I’ve sold you thus far in this article is more of that exactly. Writing scene formulae isn’t writing either. All of this talk of goals, obstacles, and results, but those aren’t the anatomy of a scene, they’re representations of its dynamics. I said it at the end of the interlude: going from formula to scene is drawing the owl, and no organ in an owl is a guideline.

So, if a scene isn’t a formula, what are the real anatomical units? A scene is recounting of a mind in dialogue with the world. A remark, an argument, a plea is made, and it prompts a response, a reply, a counter, a demand. This chain has no end in its form, only in its content.

Here’s the term you’ll see other authors use: motivation‍-​reaction units. Character reactions carry the pulse of a scene’s rhythm, chaining naturally almost like dialogue. The perspective character is faced with a motivation, and they have a reaction: immediate thoughts followed by involuntary physical responses followed by conscious action and perhaps dialogue‍ ‍—‍ then those actions have consequences, and the character will perceive those consequences, leading into another motivation‍-​reaction unit. That’s a bit of a mouthful, so we’ll just fuse this with another common scene writing term, and call these things beats.

This blog is at home in rarefied abstraction, and beats can get dangerously practical. More importantly, motivation‍-​reaction units well‍-​trod ground, and I’m not running a content mill here. These essays for original thoughts. If you just want to read about MRUs, try this piece or this one.

What I want to talk about is the significance of beats to structuring scenes. They reflect the underlying causal flow of good writing. This is more general: suppose a character walks up to a balcony and witnesses their two best friends turned arch‍-​nemeses argue and then brawl in the streets, recounting this over the course of paragraphs. The viewpoint character specifically isn’t going to have many beats here, not in the usual sense, but throughout the dialogue and the ensuing fight, there is an obvious back and forth of actions with consequences that prompt counteractions.

Wait a minute, is that right? “The viewpoint character isn’t going to have many beats here”‍ ‍—‍ but what about when a punch to the jaw resounds an audible crack and they wince? That’s a reaction. A reaction with meaningful consequences is good form, but not necessary.

This is the first principle to remember, when writing narration. Writers often complain about writing descriptions as though it’s tedious, boring, and difficult to come up with. As always, this is approaching the problem from the wrong angle. Don’t write descriptions. Write how your character reacts to the environment.

The party journeys through a forest. Sure, you can talk about the birds singing, the deer prancing in fright, the pine needles blanketing the ground, the pit‍-​and‍-​mound topology from centuries of fallen trees, the fractal patchwork of shade cast over all of it… But what good is this inventory of stimuli? You can generate all of this with enough imaginative work, but the key question to contend with is: why is the narrator describing it? What actually catches their attention and how do they judge it?

The advice some people pick up from “show, don’t tell” is that explicit judgments are bad, that saying what a character directly thinks is bad, and this is a concerning application of the advice. When you “show, don’t tell” a character’s inner state, you still need to portray that inner state. Suggest a characters judgment (whether by reaction, word choice, contrast, etc.), but avoid just presenting the thing judged without comment.

Ultimately, people want to read about what characters think, so this application of “show, don’t tell” reminds me of cutting adverbs‍ ‍—‍ a draconic overcorrection that happens to teach a valuable lesson. Here, the lesson almost approximates motivation‍-​reaction‍ ‍—‍ the worst “telling” is underdetailed reactions, and perhaps lacking motivation.

So, how do you detail a journey through a forest? Depends on the character. If they see a frog jump out of a puddle, are the going to chase it to get a closer look? If they see a bird crowing in a treetop, will they throw a rock just to make it quiet down? Are they picking flowers or shrooms? It doesn’t need to be as dramatic as direct interaction, it can be as subtle as the satisfying crunch of pine needles underfoot, the wind making a scarf wave like a flag or forcing the character to turn their head to catch a breath. Perhaps it doesn’t leave their head‍ ‍—‍ what is their opinion on the temperature, the visibility, are they worried about the wildlife or the elements or something more pernicious?

The character doesn’t need to interaction or react or inwardly comment on everything, but I find I’m so much more motivated to talk about the rocky trail ground, the painfully sheer wind, the thin air and the endless aching climb up, if it’s all wind up for me to say this character fucking hates climbing mountains.

Everyone seems to hate descriptions, but I can see how you could love them. When you really understand how they work, they’re a delightfully expressive tool for revealing character‍ ‍—‍ the key appeal and virtue of literature.

All of that is well and good for writing description, which, let’s be honest, is still ancillary to why we read most stories and what we remember.

“Scene is a mind in dialogue with the world” is a revision, ain’t it? You remember the first thing I said: a scene is people doing stuff in places. I’ve talked a lot about how to portray places. How do you portray stuff?

Piloting Across Every Dimension 

I put it best in my essay on fight scenes, and you should remember it whenever you feel your scenes need to be longer. Longform content is not just a shortform piece stretched out or filled in, but a intrinsically different thing. A proper story told in 10k words isn’t shaped like a story told in 1k words, the same way a cube is not shaped like a square.

Consider two scenes. A conversation between Madison and Elijah just went south, and Elijah is trying to leave. Where are they? Say they’re sitting in a car in a packed parking lot. Elijah gets out of the car and walks away. There’s stuff to dramatize here‍ ‍—‍ you can detail the expressions on Elijah and Madison’s faces, you can describe his posture as he walks away. What does Madison say as he leaves‍ ‍—‍ pleading for him to stay? Angry, insults hurled? Calm, reasoned arguments? Still, there’s only a few beats for you to hit here.

Consider a similar scene in a different setting. The conversation goes south, but the two of them are in Madison’s house. Now, what can Madison do to stop him? She can interpose herself between him and the exit, she can lock the door. She can chase him through the house, she could find a prop to threaten him with… Entire dimensions of possibilities have opened up. And you can already see how you might weave these beats into an escalating scene, starting with Madison asking him to stay, then taking increasingly drastic measures to keep him here.

I may have been too hard on scene formulas at the start of this part. What are beats except the unfurling of the “this but that” part of a scene formula? What do they build to, except the “so”? The sceneplot of the scene above might read “Madison wants Elijah to stay, but he tries to leave, so she traps him there.” What makes a long scene different from a short scene‍ ‍—‍ is just chaining related formulae (as you no doubt saw abundantly in action in the Aurora Moonrise outline)‍ ‍—‍ “this but that but this so this but that so…”? Close, but not quite.

The lie at the heart of the scene formula, the simplification that allows it to mostly work, is the idea that a scene is only one ur‍-​development, that it only turns once. Correcting this, we might coin the term “compound scenes” for when one development segues into another, not unlike a compound sentence.

But… let me ask something basic. What is a story? I’ve answered this question before‍ ‍—‍ storytelling is “fruitful juxtaposition”, i.e. an implied ur‍-​development. Now, I’ve been writing about craft theory for long enough that I’ve just about hypnotically entrained myself to think anything with a “but” followed by “so” really is technically a story when you think about it, but anyone not maddeningly trapped in these weeds will find something… weak to most of these one liners. Maybe they have the beginning, middle, and end you’d expect of a story, but it’s so inadequate on its own.

I’ve run into similar problem defining what constitutes a game‍ ‍—‍ is picking ‘win’ or ‘lose’ enough? Flipping a coin?‍ ‍—‍ and I’ve decided it’s simply a heap problem. Art in a certain medium isn’t doing something categorically different from nonartistic failures in the same medium, it’s a matter of doing more of it. There is no minimal story or minimal game because by minimizing it, you make it less of a story.

We defined “compound scenes”, but how about “complex scenes”? Imagine two collaborating characters, each pursuing their own related and non‍-​conflicting goals that develop in parallel across the scene. How about “stacked scenes”? I see no reason why these developments can’t happen hierarchically. Consider a scene where a character argues, (“but”) meets resistance, (“so”) changes their argument. (“But”) this still doesn’t work, and (“so”) they change methodology entirely, and slaps some sense into their adversary.

So your scene’s logic can unfold back to back, in parallel, hierarchically, and we can imagine further convolutions, where the overlaps and interferences becomes even more subtle. But we can section off this dizzying excess, and just call these beasts “organic scenes”.

But good scenes are all organic, aren’t they?

Every one of those component developments turns in some small way, and this is what it means to spin scenes.

My idea of the ur‍-​development can be analyzed as the bedrock of storytelling, but what makes a story a story, is having so many ur‍-​developments that it overwhelms the analytical faculties, becomes something intricate enough you experience it as what it’s supposed to be.

“Madison wants Elijah to stay, but he tries to leave, so she traps him there.” That’s not a story. It doesn’t even feel like a chapter. But how do you weave in explanations of the history between the two of them, how do you portray Elijah’s anxiety boiling over into unease and then need to escape, while Madison’s desire to control constricting tighter and tighter like a vise? How will you describe the house as Elijah walks in, mentioning the doors he’ll later run through and all the props Madison will pick up, laying them out like Chekhov? How will you introduce their interests, where they’re overlap and where there’s friction (they have to talk about some something, before things go south, after all)? Done deftly enough, that starts to seem like a story.

Writing a scene is like piloting a path across a dozen dimensions‍ ‍—‍ at least when you’re really on the ball. But how do you chart the path?

Maybe the Way We Made Was the Friends We Met 

Anyone who’d be willing to sit through all of this is probably running out of patience, so I need to start wrapping up. How do we tie this all together? Let’s start with what the takeaway for this part of the essay actually is.

We describe scenery by recounting a mind in dialogue with the world, this back and forth described by motivation‍-​reaction units, our vehicle of choice for piloting across several dimensions as our scenes turn organically, chaining and layering different development throughout the runtime.

In other words, we’ve covered how to describe places, and how to describe doing stuff, but how do we describe people?

It’s not a secret third art unrelated to the two that came before. People have appearances which you should describe only by how it catches the narrator’s attention, and what they think about it. People do stuff and it should operate by the causality easily represented by beats. What else is a man, but a pile of these two things?

But they do merit their own remarks, I think, because here we come to the other half of storytelling. I’ve defined it as a marriage of two media‍ ‍—‍ sensory images, and character drama‍ ‍—‍ and everything explained so far excels in stocking your scene full of sensory and procedural details.

We’ll begin with the dialogue, one of the purest expressions of a character. How do you write dialogue, though?

The fuel of dialogue is difference. Characters speak because they have something to say, and what gives them something to say? Three things: questions, arguments, and performance.

Questions: What does one character know that another does not? What does one character want to know‍ ‍—‍ or better yet, what does the other character think that one needs explained to them? This doesn’t need to be a literal question‍ ‍—‍ a definite statement could reveal ignorance that can’t be left unanswered, or an oblique statement might be a subtextually clear demand for enlightenment.

Arguments: What does one character intend, believe, desire about the world that another character can’t let stand? What reasoning, what threats, and what emotional appeals might one character think adequate to sway the other’s mind? If you’re particularly skilled, two characters can appear to be chatting, while a war is fought in the subtext, with friction rather than contradiction.

In an older draft of this section, I thought this was enough. These two principles certainly seem to encompass why a character might speak to another‍ ‍—‍ but suppose a character boasts of their wealth or skill, hoping the flex impresses someone. Suppose a character verbally beats another into the ground for daring to question them. Suppose one opens up to a friend about feelings they’ve kept secret for so long.

You can frame these in previous terms‍ ‍—‍ in a way, these characters perceives a disagreement about whether to respect the speaker, no? But that’s weak, far from the joints. Questions can be framed as disagreements too, differences of knowledge. So it’s far easier to label the differences: sometimes, characters speak to put on a performance, express their emotions and influence the emotional expression of others.

(For a somewhat advanced technique, imagine how you might write a conversation where characters are speaking on different terms‍ ‍—‍ one character making heated arguments when another is honestly just asking questions, one character explaining their beliefs to one trying to have a heart to heart.)

If you can figure out a single difference between two characters‍ ‍—‍ a question, a disagreement, a performative remark‍ ‍—‍ then you’ve got two lines of dialogue right there. And we already know how beats can chain. An argument gets a counterargument then a countercounterarguments (and then nothing because infinite recursions are at most three levels deep).

Characters rarely differ in just one way‍ ‍—‍ often, that first difference you hit upon is just a star in a constellation. Differences are the fuel of dialogue, and a conversation set up just right can burn through them all in a great chain reaction.

Dialogue isn’t just dialogue‍ ‍—‍ a huge portion of what happens in a conversation is and should be nonverbal, small gestures and shifts of body language worth entire lines on their own. But this isn’t a down‍-​to‍-​the‍-​sentence guide to writing scenes, it’s a guide to the structure. There’s so much to truly say about dialogue‍ ‍—‍ word choice, the rhythm of interruptions and utterance length, the concept of “lay‍-​up” lines‍ ‍—‍ and I don’t want this essay to be even longer than it is.

But that remark hits on something. If dialogue is partially nonverbal, can we apply these dialogue concepts to other things?

Say we wanted to convey a maid’s opinion of the young prince via inner monologue, perhaps as she’s washing his garments or cleaning his room. People contain multitudes, so what would we get if we conceptualized the maid as different people who differ in their opinion of the prince‍ ‍—‍ part of her that finds him arrogant, part of her regards him with some fondness. The parts don’t have to have equal rhetorical weight‍ ‍—‍ maybe he obviously is arrogant, but the point of this scene is to convince the reader of that, so how would that part of her explain that to the other part? She doesn’t need to hear voices or even have the different parts demarcated in the text, but there’s a certain utility to asking rhetorical question, prompting oneself by imagining how one’d explain it to a merely hypothetical audience (this is, really, what I’ve been doing this whole essay).

There’s an idea, I first echoed in my essay on fight scenes, of combat as a conversation by other means. It’s a cute idea, but I still feel I haven’t really done it justice in a story proper (Aurora Moonrise might be an opportunity, given that it has fight scenes so frequently you’d wonder if it has a shounen quota).

If a fight is a conversation, you can imagine the arguments equivalent to certain moves. A punch thrown with all one’s weight says something about a character much in the way a shouted bit of invective does. A cautious dodge or a parry, the choice flows out from a character’s ethos. The combinations of attack and defense, the tactics employed… Where you can think of a difference between two characters in their approach to fighting, there’s an image, a moment of combat, and where there’s one difference there’s more. Chain them together, make them spin, and you’ve got a scene. You can frame so much of this as expressions of character, write the whole world in dialogue with itself.

Here is where I’d say something cheesy like maybe the real way of writing scenes was the characters we made along the way.

I keep saying I’m trying to be exhaustive, writing this big huge essay to convey my thought process in totality, render every facet of how I think about scenes, and I feel like I could go further, refine what I’ve said, stud it with more examples and more clarity, but I think I’ve said enough.

You get the point, don’t you? Hook, pull, twist, tie, and then let it hang. What more do I need to say?

It’s all wheels within wheels within wheels that keep turning. Propellers we keep spinning on this craft we pilot through a hundred dimensions.

It’s just people doing stuff in places, with a twist.

Postface 

This essay was helped made viable by our generous supporters on patreon, who received earlier access to a semifinal draft of this essay.

The idea of this essay has been kicking around in my head for over a year, conceived almost jokingly on the heels of finishing my ur‍-​development essay. I thought of a cute title and grew more serious about, began promising it more heavily. In short, I hyped it up. The scene‍-​writing essay. The hype was all in my head, but it justified ambition: I had to make this essay big, I had to make it thorough, I had to make it worth the wait.

I’m happy with the result, but I suspect it wouldn’t be a smart stunt to repeat, with how much of my energy his been poured into making it reality. So I really shouldn’t tease that my next major essay might be on the subject of worldbuilding and brainstorming in general, something I’m even more adept at than scene‍-​writing, brimming with more opinions. (Definitely shouldn’t add that I’ve even outlined some preliminary models.)

No, but writing essays isn’t my main thing‍ ‍—‍ editing the remaining 24k words of Aurora Moonrise is liable to keep my busy for a while. (If there’s interest, patrons might get early access to the rest, when it’s available).

Still here? Go write something. Reading these essays will do you no good if you don’t put them to use!