Back to the table

16 Intermediate Tn Tension Story Foundations

Tension

The reader's worry about what will happen next.

What it is

Tension is the low, steady worry about how something will turn out. It runs under the sentences the whole time you read. You feel it in your body, in tight shoulders and in pages you turn faster than you meant to.

Two elements sit closest to tension and get mixed up with it more than any others. Conflict is the opposition inside the story: a person pushing back, a storm, a lie the character can't stop telling. Stakes are what that opposition threatens to cost. You arrange both of those inside the story. They are facts about the book. Tension happens somewhere else. It happens in the reader's nervous system, while the reader is reading. So a chapter can have good conflict and painful stakes and still give the reader nothing to worry through. When that happens, the setup is fine. The fault is in the delivery.

The central point is that tension runs on anticipation. Action has little to do with it. The tense scenes readers remember are often loud, but the noise isn't what made them tense. Picture a chase where the hero plainly can't be caught. Things happen one after another, and the scene reads like a list of events. Now picture a woman alone in a kitchen, waiting for a phone call. The reader already knows what the message will be. Nothing moves, and the scene is unbearable. The difference between those two scenes is anticipation. Whether anyone moves doesn't matter.

Tension needs two things, and it needs both of them:

Take away either one and the tension is gone. Uncertainty without investment is trivia. You don't know what a character will order for lunch, and you don't care. Investment without uncertainty is sentiment. You're fond of a character, you have no doubt about what will happen, and you get pleasant prose about a nice person. Put both in place and the reader leans forward. That leaning is the element.

Suspense and surprise often get used as if they meant the same thing. They name opposite techniques. A surprise depends on information you keep from the reader. It arrives, it lands, and then it's over. It also works backward, making the reader see what they already read in a new light. Suspense depends on information the reader has about something that hasn't happened yet. It works forward, and it lasts as long as you let it.

Both are fair tools, and neither one is better in the abstract. They don't cost the same, though. A surprise works once, and you pay for it with every page that came before it, because those pages had to be spent hiding something. Suspense can run for a hundred pages on one fact the reader holds. For that whole stretch, the pages themselves are the product, and nothing is wasted. So when you face the choice, give the reader the information. The trade nearly always pays more. Telling the reader things is the move with the biggest return, and hoarding works against you.

The most dependable engine for tension in fiction has a plain name: dramatic irony. The term means only that the reader knows something a character doesn't. Dramatic irony turns ordinary material into loaded material, and it costs you nothing. Say a character walks into a room and makes a joke. Nothing about that is tense. Now tell the reader, a page earlier, that the person she's joking with has already decided to destroy her. The same joke hurts to read. You added no threat, no weapon, no deadline and no chase. You moved one fact from later in the book to earlier. Dramatic irony makes the reader a co-conspirator, holding something the character doesn't have. That feels like standing in the room with her, and the reader responds with their body as well as their mind.

Three elements on this table describe the pull a story has on a reader, each at a different size. A hook is a pull the size of a point: a line, an image or a chapter ending that makes you need to turn the page. The dramatic question is the one big question a whole book is built around. Tension is the continuing state between the two, a steady worry running through the paragraph, the scene, the chapter and the novel all at once. A hook uses that worry. A dramatic question shapes it. In practice the three come apart. You can write a hook that runs on pure curiosity, in a scene with no tension at all. You can have a strong book-length question and still write a middle where nobody worries about anything for forty pages. When a reader says they couldn't put a book down, they're telling you about tension.

The most common damaging mistake with tension is believing that more is always better. Tension is a change in the reader's state, and any state held long enough stops feeling like a change. Hold the reader at maximum worry for ten chapters and you don't get ten times the worry. You get numbness, then tiredness, then skimming, and then the reader quits the book.

Relief is part of how tension works, because it lets the next rise happen. The scene and sequel pattern is built on this: a scene of pressure, then a stretch where the character reacts, regroups and decides. The pacing page covers the same back and forth. Relief can be a joke, a meal, a small kindness, or a chapter where a character sleeps. Each one lowers the baseline so the next rise can be felt. For a long book, the shape to aim for is a saw-tooth that climbs over time. A smooth, straight climb is the shape to avoid. A book readers praise as relentless has a saw-tooth with short teeth. A book they call exhausting has a flat line stuck at the top.

Tension works inside single sentences too. A revision pass at this level pays off faster than any other. You can do it in an afternoon, and the structure of the book doesn't change. Three moves do most of the work.

The first move is to withhold information inside the sentence by putting the reaction before the cause. Her hands were shaking before she reached the second line works, because now the reader wants line two. She read that the loan had been refused and her hands started shaking doesn't work, because the sentence answers its own question. He was on his feet before he understood why uses the same move.

The second move is delayed syntax. English sentences pay off at the end, so put the thing the reader is waiting for last. She checked the drawer: dish towels, a corkscrew, two dead batteries, the spare key, and no knife. Turn that around so the missing knife comes first, and the wait disappears. In the drawer, under the dish towels, exactly where she had left it: nothing. saves its key word for the end in the same way.

The third move is the interrupted answer. A character starts to say the thing and gets cut off. A woman leans across a restaurant table and says, "Because in 1998 your father —" and then the waiter arrives with bread, and she changes her mind. If a power cut does the interrupting, the reader feels the writer's hand. A waiter, a phone, a change of mind or another person talking all happen in real conversations, and each one buys you extra pages.

Use these moves too often and they harden into tics. Save them for the three or four moments in each chapter when the reader most wants an answer. At those moments, they make the difference between a flat scene and a tense one. In dialogue, subtext does related work. Two people talk about a train timetable while they're really settling whether one of them will leave.

Put simply, conflict and stakes are facts you arrange, and tension is a feeling you deliver. When a chapter is flat, the usual cure is to move information earlier. Adding more trouble rarely fixes it.

Why it matters

Tension is the most common reason a competent manuscript fails, and the easiest to repair. Beta readers rarely use craft terms. You will almost never hear "your dramatic irony is underused in the second act." What you hear is "I put it down around chapter twelve and didn't pick it up again," or "I skimmed the middle," or the review authors dread most: "well written, but I wasn't gripped." Every one of those is a report about tension. The good news is in the repair. A structural problem can mean rewriting 40,000 words. A tension problem often needs one fact moved earlier, one reassuring paragraph cut, or one thing handed to the reader that the viewpoint character doesn't have.

Telling tension apart from conflict and stakes stops you rewriting the wrong thing. Prolific authors gain the most here, because they have the least time to waste. The test is simple. If a reader says they never felt anything was at risk, you have a stakes problem. If a reader understood the risk perfectly well and skimmed anyway, you have a tension problem: the setup worked and the delivery didn't. Those two reviews look alike. The repair work is completely different.

Tension decides whether people buy, read on and recommend. Someone sampling your book decides whether to buy it based on the tension in chapter one. Read-through, the share of readers who go on to book two, depends on whether book one sent them away wired or merely relieved. Word of mouth runs on one sentence: "I stayed up too late finishing it." That sentence reports tension and nothing else. None of this is about the quality of the prose. Plenty of beautifully written books fail all three tests.

Every genre runs on tension, and quiet genres need it most. Thrillers have no special claim on it. In a romance, the tension is about whether the pair will finally say the honest thing, inside a genre that guarantees its ending. The stakes page covers how guaranteed outcomes work. In a cozy mystery, the pull is your guess about which suspect the book is steering you toward. A literary novel set at a family dinner can be tenser than a siege, because anticipation drives tension and peril doesn't have to. Writers in quiet genres need this skill more than anyone. They have no explosions to fall back on.

Real-world examples

Hitchcock and Truffaut (1962): the bomb under the table

The classic illustration of suspense against surprise comes from Alfred Hitchcock, talking to the French director François Truffaut. Truffaut interviewed Hitchcock in August 1962 at Universal Studios in Los Angeles, with Helen G. Scott as interpreter. The interviews became a book. In France it came out as Le Cinéma selon Hitchcock (Robert Laffont, 1966). The English edition, Hitchcock, came out from Simon & Schuster in 1967, in Scott's translation. Later editions carry the title Hitchcock/Truffaut.

Here's the illustration, retold. Two men sit at a table having an ordinary conversation. There's a bomb under the table. In the first version of the scene, the audience doesn't know about the bomb. It goes off, and they get a jolt. In the second version, the audience watched someone plant the bomb and knows roughly when it will go off. Now the same dull conversation is almost unbearable to watch. Every time one of the men reaches for his coffee instead of getting up, the audience wants to shout at him. Hitchcock summed it up as a few seconds of shock set against "fifteen minutes of suspense."

Look at the arithmetic. The information was the same in both versions. Only the timing changed, and the effect lasted roughly a hundred times longer, or more. No other craft move gives you that much for moving one sentence.

Hitchcock also made an exception. If the unexpected ending is the whole point of the story, keep the secret. A plot twist that makes the reader see the whole book differently is a fair trade, paid for with the tension you gave up to get it. The unfair version is hiding things out of habit, where no twist ever arrives and the dull result gets passed off as mystery.

Hitchcock didn't invent the idea, and he didn't always follow it. The film scholar David Bordwell, in a 2013 blog post listed below, traced the argument back to Gotthold Ephraim Lessing, a critic writing in the 18th century. In Lessing's version, hiding information buys a short surprise, and telling buys a lasting unease. Bordwell follows the idea forward through the theater theorists William Archer and George Pierce Baker, which shows it was around long before Hitchcock. He also shows Hitchcock breaking his own principle often, with Psycho among the cases, and the films working anyway. So treat the bomb as a picture of what telling buys you. Don't treat it as a law.

The Secret History (1992): the murder on page one

Donna Tartt's first novel, The Secret History, came out from Alfred A. Knopf on September 16, 1992. The first printing was 75,000 copies. In 1992, about 10,000 was the norm for a literary debut.

The opening looks like commercial self-harm. The prologue begins by telling you that Bunny has been dead for some weeks. A paragraph later, it tells you who killed him: the narrator and his friends. By chapter one there's no whodunit left. Tartt did this on purpose, and she did it on the first page.

The payoff is roughly two hundred pages in which the reader knows what's coming. Every scene with Bunny alive, and grating, is read by someone who knows what will happen to him. A lunch becomes hard to sit through. A walk in the woods starts to feel like it's heading somewhere. The reader stops asking what happens. The new question is how these particular people, who seem merely pretentious, get from here to that, and then what it does to them afterward. That second part is what the book is really about. Wikipedia describes The Secret History as an inverted detective story for exactly this reason.

Tartt uses a second device as well. Emily Temple, in a close reading of the prologue for Literary Hub in March 2020 (listed below), shows how the suspense stays alive. The death is known from page one, but the murder scene itself is held back until far into the book. The two devices split the work. Dramatic irony supplies the anticipation, and the withheld details supply the uncertainty. Neither one could carry a novel of about five hundred pages alone. Tell the reader the outcome, hold back how it happens, and the tension lasts as long as the pages do.

The book uses relief at full scale too. Long stretches of campus comedy, full of snobbery, drinking, Greek translation and money worries, are far from padding. They work as the low points of the saw-tooth, and they keep the later dread strong. If you want a full method for keeping a reader anxious across a whole book, beyond the big set pieces, Jane K. Cleland's Mastering Suspense, Structure, and Plot is listed below. Cleland is a mystery novelist, and the book won the Agatha Award for non-fiction in 2016.

Inglourious Basterds (2009): tension at a kitchen table

Quentin Tarantino's Inglourious Basterds premiered at Cannes on May 20, 2009, and opened in the United States on August 21, 2009. It runs 153 minutes. Christoph Waltz, who plays Hans Landa, won two awards for the role: Best Actor at Cannes, and the Academy Award for Best Supporting Actor.

The film's first chapter, titled "Once Upon a Time in Nazi-Occupied France," is the plainest example of tension with no action at all. Hans Landa, an SS officer, arrives at a French dairy farm. He sits down at the kitchen table with the farmer, Perrier LaPadite. He asks if they can carry on in English. He asks for milk. He praises the farm. He fills in paperwork. He takes a very long time to reach his question. Under the floor, a Jewish family is hiding. The audience knows they're there, and so does LaPadite. What Landa knows, or suspects, is the open question.

Very little happens in the scene, which runs about twenty minutes, roughly the length of a short story. For most of it there's no music telling you to be afraid, no weapon on show and no raised voice. There are only two polite men at a table. The effect is close to unbearable.

You can carry four reasons for that into your own work. First, the audience's knowledge makes ordinary objects threatening: the glass of milk, the pipe, the floor. Second, the scene is built from near-misses. Every polite exchange is a moment when the truth could come out, and doesn't. Third, Landa stalls on purpose. The stalling is technique and characterization at once, because a man who could ask his question in one sentence and chooses not to is scarier than a man who shouts. Fourth, the audience has no outlet. All they can do is wait, and waiting is anticipation.

The lesson for quiet books and quiet chapters is that they already have what they need. A chase and a kitchen table have the same tension available. The difference is how much anticipation the writer built in.

Columbo (1971 to 1978), Dr. Thorndyke and Ripley: the crime shown first

Columbo, created by Richard Levinson and William Link, ran on NBC from 1971 to 1978, and later revivals followed. Every episode opens by showing the killing and the killer. The audience then watches a detective in a raincoat close a trap they can already see. The format is called a "howcatchem," as against a whodunit. The form itself is far older than the show. It goes back to R. Austin Freeman's Dr. Thorndyke stories, around 1912.

Patricia Highsmith's Ripley novels use a related setup, starting with The Talented Mr. Ripley in 1955. The reader knows exactly what Ripley has done and spends the book afraid he'll be caught. Fearing a killer's capture is a strange place to put a reader, and it works. Highsmith also wrote Plotting and Writing Suspense Fiction, first published in 1966 and revised in 1981, and listed below. In a short book, she explains how a working suspense novelist builds dread into everyday scenes, with no template or system.

An everyday scenario: the fight nobody worried about

Say you're writing book four of an urban fantasy series. In chapter twelve, your lead takes on three of the antagonist's henchmen in a parking garage. The prose is fine and the fight is easy to follow. Your two beta readers both skimmed it.

Apply the test. The reader's question is will she win? Readers know she will. She's the lead in book four, and she's on the cover of book five. The uncertainty is missing. The investment is there, but nothing is using it. A harder fight won't help, because readers have correctly decided the outcome isn't in play.

The fix is to move the uncertainty onto something that really is undecided. One option: can she get out before her sister turns up for the meeting they planned in chapter nine? Another option: will she have to use the forbidden act she swore off in book three? Then give the reader a fact only the reader has. The sister is already in the building, waiting two floors up. Now the fight has a clock, and it has dread.

Then you go too far. You apply the same fix to every chapter, and all forty-one chapters end on a cliff edge. The beta comments flip. Now the book is exhausting, the readers wanted a break, and somewhere between seventy and eighty thousand words they stopped caring. The diagnosis is a flat line at the top. The cure is to put back three real pockets of relief: a scene that resolves, a night's sleep, a conversation with a joke in it. The lesson is that a chapter that lets the reader breathe out has a structural job to do.

Try it yourself

Steps 1 to 5 work on one chapter you suspect is flat, and they take about an hour. Steps 6 and 7 look at the whole manuscript, so allow about forty minutes for those on their own. Step 8 needs your first chapter and your voice. For a longer checklist on a flat chapter afterward, the 2013 Writer's Digest piece listed below is built on Cheryl St. John's methods. Her book Writing With Emotion, Tension, and Conflict, also listed below, treats tension at three levels: the chapter, the scene and the sentence.

  1. Write the chapter's question. In one sentence, write the question in the reader's head during this chapter. It has to be a real question, with at least two possible outcomes and an answer nobody knows yet. Will she get the file before he comes back? passes. What will happen is too vague to count. If you can't write one at all, you've found the problem, and no amount of line editing will fix it.
  2. Check that the question is really open. Write down the outcome a reader would bet on. If the reader would bet right and feel sure about it, because she's the series lead, or the genre guarantees the ending, or you've foreshadowed it heavily, the uncertainty is fake. Move it onto the cost, the timing, who else pays, or what she'll have to become. The stakes page helps with this.
  3. Audit who knows what. Make three columns: what the reader knows, what the viewpoint character knows, and what the other characters know. Tension lives in the gap between the first two columns, and the third is there for reference. A thin reader column explains a flat chapter. Hand the reader one fact the viewpoint character is missing. No other step rescues as many scenes.
  4. Turn a surprise into suspense. Pick the biggest thing you're holding back. Test what happens if you reveal it in chapter three instead of chapter thirty. Usually you lose one jolt and gain a dozen charged scenes. If the jolt is the point, keep it, as one deliberate decision you've made on purpose. Hiding things by reflex is the habit to break. Foreshadowing is the middle road: tell enough to make the reader worry without telling everything.
  5. Do a sentence pass on three paragraphs. Ignore the rest of the chapter. Find the three spots where the reader is hungriest for an answer. At each spot, try each of the three moves in turn and keep the ones that earn their place: put the reaction before the cause, put the key word at the end, and let something ordinary interrupt an answer.
  6. Chart the saw-tooth. List your chapters. Score each one from 0 to 5 for how worried you think the reader is at its last line. Then look at the shape. A saw-tooth that climbs is healthy. A flat line at any height is a sag. A flat line at the top for more than four chapters in a row means exhaustion, and the cure is more relief. More threat will make it worse.
  7. Plan relief on purpose. Pick three points where a reader can put the book down without worrying. Write them with care and don't rush them: a subplot that resolves, a meal, a win. Readers forgive a lot when you trust them to come back.
  8. Read your first chapter aloud. Put a pencil mark at the sentence where your own worry slipped. Usually you'll find a patch of reassurance, backstory or scenery there, written for a reader who didn't need it.

Chapter ___ Reader's question: ___________________ Could go either way? Y / N. If N, the uncertainty actually lies in: ___________________ Reader knows and POV character does not: ___________________ Worry score at last line (0-5): ___ Relief chapters in this book: ___, ___, ___

Common mistakes and misunderstandings

Thinking action equals tension. Fights, chases and disasters often produce pages with no tension at all, because the reader has already guessed the result. Whether anyone moves doesn't matter. What matters is whether the reader is anticipating something.

Hoarding information. Writers usually do it to protect a reveal. Holding back builds tension only when the reader knows something is being kept from them and knows why it matters. Without that, you get obscurity, and readers take it as vagueness. Picture an opening that spends its first three chapters being coy about who the narrator is, why she's in town and what she once did. Readers leave.

Expecting style to stand in for a question. Clipped sentences, fragments, present tense, gloomy weather and a one-line chapter ending can only make an existing worry bigger. A flat scene written in clipped sentences stays flat, and now it reads oddly as well.

Mistaking confusion for tension. A reader who can't tell what happened has lost the thread, and anticipation dies with it. Clarity comes first.

Reassuring the reader by accident. Accidental reassurance is the most common leak in drafts. It slips in through an all-knowing narrator's aside, a line of hindsight, or a viewpoint character who stays calm about something that should alarm her. A lead who isn't worried gives the reader permission to relax. The fix is the easiest one there is: delete the sentence.

Using cliffhangers in place of tension. Put a cliffhanger at the end of every chapter with no steady worry in between, and readers learn that the endings are a device. A ticking clock carries the same risk. A deadline works only while the reader believes it's real and could be missed.

Treating tension as thriller business. Tension is hardest to create, and most needed, in genres where readers don't come expecting danger. A quiet novel with sixty pages of no worry is in worse trouble than a thriller with the same sixty pages. The thriller's reader is still waiting for the danger the genre promised.

Raising the threat when you should deepen the consequence. The urge is to keep making the menace bigger: more enemies, more dead, wider destruction. Readers get used to scale quickly. They don't get used to cost that can't be undone. Louder reads as repetition, and costlier reads as relentless. The try-fail cycle is a structure that lets cost pile up.

Forgetting to re-ask the question. A question raised in chapter two isn't still alive in chapter twenty unless you touch it again. Nest your questions at three timescales: the book, the stretch and the scene. The book-wide one, your dramatic question, has to come back every few chapters or it fades.

Blaming the reader for going numb. Readers sometimes say a relentless book got boring. The instinct is to decide they weren't paying attention. They were. Steady maximum pressure stops being felt, the way a constant noise does. Give them somewhere to stand.

Related elements