Back to the table

17 Intermediate Ce Cause and Effect Story Foundations

Cause and Effect

Each event happens because of the one before it.

What it is

Here are two sentences.

The king died and then the queen died.

Here they are again, with a comma and two words added.

The king died, and then the queen died of grief.

Nothing else changed. The order is the same, the cast is the same, and the word count barely moved. The second version is a different kind of thing. In the first, two deaths just happen to sit next to each other. In the second, one death produces the other, and a reader who finishes that sentence knows something about a marriage that nobody spelled out.

That pair of sentences is E.M. Forster's, from Aspects of the Novel, the book that grew out of the Clark Lectures he gave at Trinity College, Cambridge, in 1927. Forster used it to pull apart two words people treat as if they meant the same thing. A story, he said, is events in the order they happened. A plot is events tied together by causality, meaning one thing causes the next. Faced with a story, we ask and then? Faced with a plot, we ask why?

So cause and effect is the glue that turns a list of things that happen into a plot. It's the difference between events that are consecutive, meaning they sit next to each other in time, and events that are consequent, meaning each one exists because of the one before it.

Aristotle got there first, and he was blunter about it. In chapter 10 of the Poetics, he separates incidents that happen because of one another from incidents that just happen after one another, and he says the difference is enormous. English translators have long used two Latin tags for the pair, propter hoc and post hoc. Those are the translators' shorthand, not Aristotle's own words. The idea underneath is the one you need as a novelist. Aristotle called a plot whose scenes only share a timeline episodic, and he ranked it as the weaker kind.

The practical test is a word game. Lay your scenes end to end and try to put one word between each pair. Only four words come up often, and they aren't equal.

Cause and effect is easiest to fix when you make it concrete. A scene that's properly linked to the one before it takes at least one of these from it, and usually several:

Notice that the link runs through a person. In fiction, events don't cause other events the way billiard balls do. They cause decisions, and the decisions cause events. A chain that skips the human step feels mechanical. Problems with cause and effect often turn out to be motivation problems in disguise, and the fix is usually a scene where somebody makes a choice, not an extra paragraph of explanation.

Coincidence isn't banned. It's rationed, and how much you get depends on where it falls.

At the start of a story, a coincidence is how the problem shows up: the wrong suitcase, the overheard sentence, or the neighbor who moves in next door. Readers accept an unlikely event that creates trouble, because until the trouble exists, there's no chain of causes for it to break. An inciting incident, the event that kicks off the story, is allowed to be just that.

At the end, the same coincidence is theft. Say a reader has watched a character struggle for three hundred pages, and then the solution arrives by luck. That luck cancels the struggle after the fact. The reader cared about the character's effort, and a lucky rescue tells them the effort didn't matter.

There's a useful refinement, too. Coincidence that harms the protagonist, your main character, holds up much better than coincidence that helps. An accident that makes things worse feels like fate, or like the world not caring, and readers have a big appetite for both. An accident that makes things better feels like the author stepping in. Take Act 5 of Romeo and Juliet, where a plague quarantine stops Friar John from delivering Friar Laurence's letter to Romeo in Mantua. That still lands after four centuries. An equal stroke of luck that rescued the lovers would be unbearable.

So here's the rule of thumb. Coincidence may open a door, and it may slam one. It may not unlock the last one.

Why it matters

Cause and effect is behind the most valuable reaction a reader can have, which is "of course". It's the sense at the end that the book was always heading here, even though the reader couldn't have said so beforehand. Looking back, the ending feels inevitable. Going in, nobody could predict it. Those two sound like they contradict each other, and they don't. A reader can't predict the ending because they don't have all the causes yet. Once they do, the ending is the only thing that could have fit them. Everything readers praise as "earned", "satisfying" or "airtight" is that feeling. Everything they complain about as "contrived", "convenient" or "out of nowhere" is that feeling gone missing.

It matters for sales, and it matters most for a series. A reader who finishes book one feeling cheated by the ending doesn't buy book two, and they don't need to be able to explain why. They'll say the ending didn't work for them. What they usually mean is that the ending didn't grow out of the causes the book had set up.

It affects how fast you write, and writers underestimate this part. A chain of causes keeps producing its own next step. If scene fourteen ends with a character finding out something they can't un-know, you don't have to invent scene fifteen. Scene fifteen is what they do about it. When writers stall in the middle of a first draft, they're usually short of consequences, not ideas. The last scene they wrote didn't force anybody to do anything, so they have to pick the next one from an endless menu. Tight cause and effect shrinks the menu to two or three options, and picking between three options is fast.

It's one of the few structural faults you can diagnose in a finished draft. You can find it and name it. You can't easily check a manuscript for voice or theme. You can check it for connectors in an afternoon, with a one-line summary of each scene and a pencil. That makes the connector audit one of the most useful passes in revision. Cause and effect is also one of the least costly things to get right at the outline stage, before the scenes exist and need rewriting.

Real-world examples

Sophocles' Oedipus Rex (c. 429 BC): the investigation that causes its own catastrophe

Oedipus Rex was first performed in Athens around 429 BC, though the date isn't certain and some scholars place it a few years later. It took second prize at the festival. Its plot is the most picked-apart in Western literature, and the reason is cause and effect.

Almost everything shocking in the story happened before the play begins: the prophecy, the baby left out to die, the killing at the crossroads, and the marriage. The play itself has only one action. Oedipus investigates a plague, and he refuses to stop investigating. Every scene is caused by the one before it, and every scene brings him closer to the answer he least wants. He summons the prophet, and the prophet refuses to speak. Therefore he accuses the prophet of plotting against him. Therefore he drags Creon into it. But the argument turns up a detail about the crossroads that snags on his own memory. Therefore he sends for the surviving witness.

Aristotle singles out the moment that makes this the best-built reversal he knew. A messenger arrives with good news, hoping to free Oedipus from his fear about his mother, and by explaining who he is, he does the opposite. The relief is the catastrophe. Here cause and effect does something no coincidence can do. One event holds its own reversal.

For you as a working novelist, the useful part is what this shows about where to put luck. The backstory of Oedipus Rex is a web of outrageous coincidence (of all the men at all the crossroads), and the play's action has none. All the unlikely parts are loaded into the setup, where readers forgive them. The ending comes entirely out of the hero's own choices, where readers forgive nothing. If you want one model for how to ration coincidence, use this one.

Trey Parker and Matt Stone at NYU (2011): the whiteboard, "therefore" and "but"

In 2011, Trey Parker and Matt Stone turned up unannounced at a first-year storytelling class at NYU's Tisch School of the Arts, and the visit was filmed for an episode of mtvU's Stand In. The clip from it has become the most-quoted piece of practical advice about cause and effect out there. Go to the source, because the retellings have drifted.

What they describe is how they actually work on South Park. The three acts go up on a big whiteboard, and the beats, the story events, of each act get filled in underneath. Then they read down the list and check which word belongs between each pair of beats. Parker's verdict on finding "and then" between two beats is unprintable. The gist is that the episode is already broken. What should be there instead, he says, is "therefore" or "but": this happens, and therefore this happens; but this happens, therefore this happens. Stone sums it up by saying those words give you the causation between each beat, and that the causation is what makes it a story rather than a list.

The details tend to get garbled in the retelling. One is the date. The class visit and the Stand In episode are from 2011. The clip went around again in 2014 when craft blogs picked it up, and lots of sources now date the class itself to 2014. The other detail is that Stone's version of the rule includes "because" alongside "therefore" and "but". That matters. "Because" lets you write the link backward (this happens because that happened), and some scenes are easier to diagnose from that direction.

The idea spread far beyond TV because it's a test, not a theory. It takes twenty minutes on a novel, and it gives you a list of specific broken joints, which is just what you need when you're staring at a sagging middle.

Euripides' Medea (431 BC): the machine that gave us the phrase

Medea was first staged in Athens in 431 BC, and it came third of three. At the end, Medea has killed her rival and her own children, and then she appears above the house in a chariot sent by her grandfather Helios, and escapes. She's lifted there by the mechane, the crane of wooden beams and pulleys that Greek theaters used to fly gods and heroes in and out. Medea may be one of the earliest surviving uses of it, and it stands out because the crane carries a character who isn't a god.

That crane is the machine in "deus ex machina". The Latin phrase is a later word-for-word translation of a Greek saying about a god appearing from the crane. No surviving ancient Roman text actually uses the Latin tag, so be a little careful about how confidently the term gets credited. What the Romans do have is Horace. In the Ars Poetica, he tells poets to let no god step in unless a knot turns up that's worthy of such an untier.

Aristotle's objection, in chapter 15 of the Poetics, is narrower than the modern shorthand, and more useful. He doesn't say gods can't appear. He says the untying of the plot must come out of the plot itself, and he names Medea as a case where it doesn't. On his account, divine machinery is fine for things outside the play, like what happened long before or what will happen after, because those aren't the plot's knot. What it may not do is cut the knot the play spent its whole length tying. As one standard commentary puts it, weak playwrights used the crane to declare the resolution, instead of unraveling it by cause and effect.

So why do readers feel cheated by a deus ex machina? Name the feeling carefully, because it has little to do with whether the rescue is believable. Readers accept dragons. What they won't accept is being told that the effort they cared about didn't produce the outcome. A rescue from outside the chain of causes solves the plot. In the same stroke, it turns every earlier struggle into filler after the fact. What gets thrown away is the reader's own investment, which is why they react with anger and not mild disappointment.

You can see the modern version of this worry clearly in Tolkien. On June 9, 1958, he wrote to Forrest J. Ackerman about a proposed film treatment of The Lord of the Rings that used the Eagles freely as transport. Tolkien objected that the Eagles are "a dangerous 'machine'", that he had used them sparingly, and that sparingly was the absolute limit of their credibility. He put the word machine in quotation marks himself. In a letter of July 26, 1956, about the destruction of the Ring, he insisted that he had not arranged that rescue. It followed from the logic of the story, where Frodo's earlier mercy toward Gollum is the cause that puts Gollum on the mountain at all. Tolkien is defending a thousand pages of chain against the charge that his ending came from outside it, and his whole defense is about cause and effect.

Breaking Bad, "Phoenix" and "ABQ" (2009): a fuse two episodes long

For a modern example of cause and effect over a long stretch, the second season of Breaking Bad is hard to beat. In "Phoenix" (first aired May 24, 2009), Walter White stands over Jesse's girlfriend Jane while she chokes, and he chooses not to roll her onto her side. Then comes the season finale, "ABQ" (first aired May 31, 2009, written by Vince Gilligan). Jane's father Donald, an air traffic controller, a job the show set up earlier in the season, is back at work and falling apart with grief. When a charter plane's callsign catches his ear because it sounds like his daughter's initials, he loses his focus for a moment and sends traffic the wrong way. Two planes collide over Albuquerque, and debris falls into Walt's swimming pool.

Follow the connectors. Walt protects his secret, therefore Jane dies. Jane dies, therefore her father is destroyed. Her father is destroyed, therefore he makes an error at work. Therefore a great many people die, and the man responsible is standing in his own backyard with no idea that he's responsible.

What makes it useful to study, and not just clever, is that every link was planted before it was needed. Donald's job was set up in an earlier episode, in a scene that plays as simple character color at the time, so nothing is invented at the moment of payoff. The whole technique of foreshadowing, when it serves cause and effect, comes down to this. Afterward, the reader must be able to rebuild the chain using only material they already had. The example also shows that cause and effect doesn't have to be tight in time. The gap can be two episodes, or four hundred pages, or a whole book of a series. What it can't be is missing.

An everyday scenario: the audit that saved chapter twelve

Say you have a finished draft of a thriller. Your readers all like it, and they all stop caring somewhere around the middle, and nobody can tell you why.

You put the scenes in a spreadsheet, one line each, and write the connector between every pair. Most of the first third comes out therefore, but, therefore, which is why the opening tests well. Then you hit a stretch of eight scenes and write and then six times. Reading them back, you can see the pattern. Your detective goes to an address and learns something, then goes to the next address and learns something else. The scenes are full of information and empty of consequence. Nothing she learns at address one is why she goes to address two. She goes because you had a list.

Better prose in those eight scenes won't fix it. The fix is to make each discovery close a door. She learns at address one that the witness lied. Therefore the alibi she was relying on collapses. Therefore she has to go back to the person she's been avoiding, but that person now has a reason not to help her. You keep the same information and the same locations, and maybe much the same dialogue. You've welded six broken joints.

Try it yourself

Run this as a full pass on a draft, an outline or a synopsis. Give it an hour for a whole novel's chain, or twenty minutes for a single act.

  1. Make the spine. Write one line per scene or chapter, in order, in a document or spreadsheet. Each line says what happens, not what it's about: "she confronts her brother and he denies it", not "family tension". Keep it to one sentence.
  2. Write the connectors. Between every pair of lines, write one of these: therefore, but, because, and then, meanwhile. Don't argue with yourself, and don't write the one you wish were true. Write the one that actually fits.
  3. Count. Add up how many of each of the five you have. A healthy commercial novel runs mostly on therefore and but, with the buts bunched around the midpoint and the low point. Any run of three or more and thens is the sag your readers felt.
  4. Run the deletion test on every weak joint. Take each and then scene and ask: if you cut the scene before it, does this scene still work? If it does, there's no causal link, just two scenes sitting side by side. Then ask the harder version: if you cut this scene, does the next one still work? A scene that passes both tests isn't attached at either end, and it's usually the one to cut or merge.
  5. Do a coincidence count. List every event in the book that happens by luck, accident, timing, or a stranger showing up. Mark where each one falls: first quarter, middle or last quarter. Coincidences in the first quarter are fine. Look at the ones in the last quarter one at a time. If a coincidence helps your protagonist in the final act, either remove it or turn it into something they caused.
  6. Test the ending backward. Start at the last scene and write the chain in reverse: this happened because…, and keep going. You should be able to walk from the ending back to the inciting incident without a gap. Wherever you can't answer "because", you've found the spot where the climax is floating free of the book.
  7. Check that the solution was on the page. Whatever your protagonist uses to win, find the earliest page it shows up on. It might be a skill, an object, an ally, a piece of information, or a flaw they finally overcome. If the answer is "in the last chapter", you have a deus ex machina in ordinary clothes. Move its first appearance early, and make it look like something else at the time. See foreshadowing.
  8. Check that the links run through a person. For each therefore, name the decision in the middle: event, decision, event. If you can't name a decision, the chain is running on coincidence or on the author's hand, and the fix belongs in motivation.

Here's a copy-and-paste template for your notes:

Scene N: __________ Connector to the next scene: therefore / but / because / and then / meanwhile Decision in the middle: __________ Door this scene closes: __________

Common mistakes and misunderstandings

Confusing causality with predictability. The most common reason writers resist this work is the fear that a tightly linked plot will be easy to guess. That fear gets things backward. Predictability comes from familiar causes, not strong ones. What makes an ending hard to guess is that the reader doesn't have all the causes yet. Maybe you've held one back or disguised one. Maybe you've let them draw the obvious conclusion from a cause that will turn out to mean something else. Cause and effect is what makes a surprise stick once it arrives. Without it, you get a twist nobody believes.

Mistaking incident for consequence. A chapter can be full of action and still be an and then. Explosions, chases and arguments are events, and they only become links when they change what the next chapter has to contain. Plenty of manuscripts called "slow" are actually very busy, and nothing in them leads anywhere.

Using information as the link instead of consequence. Mysteries, procedurals and any plot driven by an investigation fall into this trap easily. The detective learns something. It doesn't cost her anything or close any door, so she moves on to the next appointment. Information that doesn't limit her choices isn't a cause. Make each discovery rule out an option, create an obligation, or put something she values at risk. See red herring for the version of this that's done on purpose.

Forcing a link by breaking a character. Say the only way to get from scene eight to scene nine is for a careful person to do something reckless. You haven't built a causal link. You've built a hole and painted over it. The reader will object with "she would never do that", which is a complaint about cause and effect aimed at the character instead of the plot. Fix it earlier. Give her a reason in scene seven, or change what happens in scene nine.

Over-welding. Not every scene has to be a domino. A nonstop chain of consequences with no room to breathe wears readers out. Reaction, grief, planning, comedy and rest all do work. They're part of the chain when they hold the decision that produces the next scene, which is what the sequel in scene and sequel is for. Cause and effect is about the joints, not speed. Speed is a question for pacing.

Assuming a non-linear structure gets a pass. Dual timelines, braided stories, frame stories and novels full of flashbacks all still need a chain of causes. The order you tell events in is a separate choice from the order in which they cause each other. If you reorder a timeline and nothing breaks, the order was never the problem.

Treating a deus ex machina as only a fantasy problem. These days it rarely shows up as a god. It shows up as a phone signal in the one place there wasn't any, or a minor character with a skill nobody mentioned before. It comes as a rescue by the authorities the book spent two hundred pages showing to be useless, a villain's sudden change of heart, or a piece of forensic evidence that turns up in the last chapter. The test is not "is this supernatural". The test is "did the protagonist cause this, and was the means of it on the page earlier".

Letting a subplot pay the main plot's debts. A subplot that resolves the main plot is a deus ex machina with better manners. The romance may soften the cost of the ending. It shouldn't be the thing that defeats the antagonist, the story's main opponent, unless the book has been building toward that all along.

Thinking this can only be fixed at the outline stage. It costs less to fix it there. The connector audit works just as well on a finished manuscript, and it usually means adding or moving a handful of paragraphs, not rewriting whole scenes. Most weak joints get fixed by changing the last page of one scene and the first page of the next.

Related elements