Story Under a Vocabulary Ceiling

A vocabulary ceiling is a fixed word list that a story may not exceed. It can be used as the story’s schedule for what gets revealed, instead of as a restriction to be written around.

Wanting something and not getting it can be written without any feeling-words. A complete arc therefore fits inside the first few hundred words: a person wants a thing, moves toward it, and gets something else. Interior life is added at the rate the word list grows. The two schedules can be made to line up, so that the word a reader is about to learn and the thing a character is about to admit arrive in the same beat.

Core takeaways

• Hold the peak of a scene near 98% known words. Spend the new words in the calm beats instead.
• Below about 95% known words, which is one unknown in twenty, the text stops teaching and starts frustrating.
• A want, an obstacle, a reversal, and a change are not words, so a ceiling does not block them. They can be staged at the first-book floor of 250 字.
• The external Want can be said early, in concrete nouns. The internal Need waits for a later book’s vocabulary.
• The book floors are 250, 450, 750, 950, and 1000 字. A marquee, a deep two-hander, and a special have no floor.
• The coverage figures tell a writer where feeling is affordable. They do not promise that the text will teach the language.

How much of the text the reader already knows

A 2000 fiction-coverage study asked university students to read the same kind of story at four known-word levels: 80, 90, 95, and 100 percent. Adequate unassisted comprehension sat near 98% of running words. The sample was small, sixty-six people, and the thing measured was comprehension, not whether new words stuck. Later coverage work treats the same 98% as a useful target for comprehension, and keeps uptake in a separate literature. Incidental learning is a stronger claim than that paper can carry. In this technique, 98% is a comfort target: the share of running words already known when the reader should be able to move without help.

The wider coverage literature puts a lower figure near 95% and calls it minimally adequate. Comprehension still holds there. Effort rises. Below about 95% the text stops teaching and starts frustrating. The 95% point is approximate: some readers and some tasks cross it sooner, some later. The practical design rule is the arithmetic. Ninety-five percent known is one unknown in twenty, which on a short page is one or two new words. A writer who plants three new items on a page of forty has already gone below 95%.

Coverage is one of three conditions, held as a design order and not as measured stages. First the input has to be comprehensible: i+1, input just past the current level, useful as “slightly beyond” and not as a precise number. Then it has to be compelling: the reader forgets the language is foreign. Then the affective filter has to stay low. Anxiety, boredom, and embarrassment are the named barrier. These are hypotheses, and they are contested.

A reader decoding hard words during a hard beat learns nothing. That is why a scene needs warmth and momentum. The peak of a scene is where coverage has to be highest. Whether the affective filter works as a mechanism is disputed. The part that still holds is the quitting: hard decoding plus a hard beat, and nothing is learned.

The figures are design guidance, revised as evidence comes in. Ninety-five and ninety-eight shift by task, text type, and reader. i+1 has no clean definition that can be measured. The input hypothesis itself is contested. A 2025 critique argues that comprehensible input is necessary but not sufficient, and that interaction and production sit outside the job of a vocabulary ceiling. That paper does not say stop making input comprehensible. The numbers stay because they tell a writer how much to spend on a page. They stay only while later evidence still agrees.

The parts of a story that are not words

A want, an obstacle, a reversal, and a change all survive the ramp, because none of them is a word. What carries a first-book reader is the drama. It can be staged at this project’s first-book floor of 250 字, characters rather than English words, and paid off at the last-book floor of 1000 字.

The gap is expectation minus result, widening across acts. It needs only actions. A reader who can watch someone reach and miss does not need the word for disappointment. The miss is the gap. A wider miss one act later is the same gap, larger.

The premise is one arguable sentence of the form X leads to Y. Held fixed, it keeps a multi-book arc coherent in the years before the theme-words exist. A first book can stage the X in objects. A last book can pay the Y in words the first book did not have. Premise and Controlling Idea covers the language-independent through-line that holds that arc when the vocabulary cannot yet name it. Bone structure means physiology, sociology, and psychology. It is a separate tool from the premise. Treating the two as one loses both: a filled-in character sheet is not an argument, and an argument with no character in it has no one to make it. A controlling idea, a value plus its cause, does the same work as the premise under a different name: holding it fixed lets a plant at the first-book floor pay off books later.

The story circle is eight beats that travel out and back: comfort, want, threshold, search, find, pay the price, return, changed. The same eight carry a single first-book episode at 250 字 and nest across a season. The circle is also claimed to scale to one scene, or to a whole book from arrival to belonging, at whatever size the vocabulary allows.

The Wound and the Lie covers the Lie / Want / Need triad. Under a vocabulary ceiling only the scheduling split matters. The external Want is sayable early, in concrete nouns: an envelope, a chair, a hand refused. The internal Need waits. A first-book reader can watch the Want. The Need is a later-book word.

Word floors for the five books

This project’s five books are named B1 through B5. B1 here is the public language ladder’s first rung, not its third. The floors below are named first-book, third-book, and last-book rather than B1, B3, and B5. The counts themselves are house defaults, set at roughly one and a half times a textbook article and mapped onto that public ladder: 250, 450, 750, 950, and 1000 字.

Book	Floor (字)	Public-ladder rung
First	250	first rung
Second	450	second
Third	750	third
Fourth	950	fourth
Last	1000	fifth

Important story pieces are exempt. A marquee, a deep two-hander, a special has no floor and no ceiling. A writer who cuts a two-hander to 250 字 because “the floor says so” has misread the table. The ceiling still governs run-ups and ordinary coverage. The one-in-twenty rule still applies to a scene that has a special name.

Putting the core material before the end of the third book reaches the median reader, and charges the cost to the readers who finish. Later books have a smaller audience. The deepest interiority is read by the fewest readers. That is a choice made on purpose with its cost known, and it is not based on a measured leave-rate.

Working structure first commits the whole arc in advance. The last-book payoff is known before the first-book page is written. An envelope counted on the table, and a chair that stays empty, are promises made at 250 字 and kept for four books. The objects cost little to place. Paying them off costs a lot. The timing is set by the reader’s capacity, not by when the writer would find it elegant to reveal a secret.

Where the new words go

The first book has words for actions only. One character’s family-hunger is shown as a flinch. Another character’s load is shown as an envelope she counts and a hand she will not take. 沈文 and 阮草 are those two characters, and the sentences work whether or not a reader opens their cards. The second book has words for reasons. Motive is stated in the same reading that names it.

Withholding a word until its level arrives is the reversal. Some content needs a register that does not exist below a floor, and forcing it early produces a sentence the reader can sound out but not hold. A wound can be withheld past the ceiling entirely and still be carried by an object. Interiority Through Action and Object covers objects and gestures: what a floor forces when feeling-words do not exist.

The scene’s few nouns are established through concrete, repeated, natural use in the calm opening beats. One teaching method repeats a single structure many times through easy questions: circling, in the practitioner sense, with the fifty-to-a-hundred counts treated as workshop lore rather than as a measurement. The transferable half is the prose equivalent: key objects appear plainly and often in quiet moments, so they are already known when they matter.

The peak is held near 98% known. The new-word budget is spent in the calm, never on the turn. A turn that also teaches a word spends the reader’s effort on the word. The feeling the turn was built to carry then has to wait. A chosen object standing in for a feeling, and a known thing left unsaid, are what the floor forces. The full treatment is in Interiority Through Action and Object.

The want itself has to be folktale-strength, or a small register cannot state it. Hunger will do. A promise will do. A gift kept hidden will do. Compelling-input is useful as a test of that strength, not as a sufficient theory of how language is acquired. This project’s first book has two load-bearing wants: chosen family, and provision made in secret. Both are fully sayable at the first-book floor and deepen from there. A first-book reader can watch someone sit down at a table that is not theirs, and watch someone count an envelope and refuse a hand. The later books add the reasons. A want that strong carries the reader whether or not they are trying to improve. That is a teaching claim, and it carries the same uncertainty as the coverage figures.

The seed of a late payoff has to be a behavior the later book can re-show, never a stated theme. Seeding and Payoff covers why a first-book behavior can pay off when the words arrive. The Iceberg and World as Pressure covers why the hidden eighth of that behavior must be known to the writer, even when it never appears in the text.

Audio-first work in this project caps a scene at three or four speakers. A large ensemble is a reason to write many short readings, three to a lesson, rotated, rather than one crowded room.

What it costs and when it fails

Every noun repeated for comprehension is a noun not spent on new texture. That cost is the priming tax, and the calm-beat budget at the first book is small. The ceiling works only for stories whose depth lives in behavior and want. Wordplay, register irony, and dense monologue cannot be written under one. A writer working on those should not use a vocabulary ceiling.

Applying show-don’t-tell absolutely is the failure that tempts a writer. A well-placed tell, “she was tired”, can do the work of a whole scene of inference. Omission fails when it is hollow. An unstated wound works only if the author knows exactly what it is. The Iceberg and World as Pressure covers why that hidden eighth has to be written down somewhere, even if it never ships. Atmosphere with no gap in it is decoration.

The pedagogy is contested. The coverage figures tell a writer where feeling is affordable. They do not promise that a text built to those figures will teach the language. A well-engineered 98% is not proof of acquisition. The test is what real readers do.

The fixes are these. When the reading has become a worksheet, the beat is rebuilt from the want. When the reader stalls on a new word at the turn, coverage has dropped: the new words move to the calm, and the peak is held near 98% again. When a first-book hint is a stated theme, it is replaced with a behavior the later book can re-show. When an omission reads empty, the backstory is written even if it never ships. When more than three or four people speak in one scene, interiors are cut and the rest of the cast rotates through later readings.

The checks are visible in the text. A character refuses help, and the reader infers she carries the load alone. A second-book motive reading yields the fact and the reason after one pass. A peak reveal lands in the next sentence, with no stall and no narration line to explain what just happened. A late rupture reads as something that was already latent in the early objects.

The word list is the instrument. The want is staged early, in nouns a first-book reader already has. Interiority arrives on the schedule the list sets. The numbers are not a promise that the language will be acquired.

How to practice this

1. Count the unknown words on one page of your draft. Three new items on a page of forty puts you below 95%. Notice whether a reader would stall there.
2. Mark the peak beat of a scene. Move every new word out of it and into the calm beats. Check that the reveal lands in the next sentence, with no stall.
3. Put the scene’s few key nouns into the calm opening beats, plainly and often. Then check that each one is already known by the time it matters.
4. Take one first-book hint and check that it is a behavior, not a stated theme. A later book has to be able to show that behavior again.
5. Write down the hidden part of an unstated wound, even if it never ships. Then reread the scene and check whether the omission reads empty.
6. Count the speakers in one scene. If more than three or four speak, cut interiors. Move the rest of the cast into later readings.

Related pages

• Interiority Through Action and Object: the object-and-gesture half of a vocabulary ceiling, what a floor forces when feeling-words do not exist.
• Seeding and Payoff: why a first-book behavior can pay off when the words arrive.
• Premise and Controlling Idea: the language-independent through-line that holds a multi-book arc when theme-words will not exist for years.
• The Iceberg and World as Pressure: why the hidden eighth must be known, and why hollow omission fails.
• The Wound and the Lie: owns Lie / Want / Need. A vocabulary ceiling uses only the scheduling split, Want early and Need waiting.

Open questions

Whether 98% as a peak-scene design target survives a larger, non-fiction sample.

How a marquee with no floor still holds the one-in-twenty rule in its run-ups.

Whether a first-book tell is cheaper than a hollow object.

Sources

• Hu, Hsueh-chao Marcella, and I. S. P. Nation. “Unknown Vocabulary Density and Reading Comprehension.” Reading in a Foreign Language 13, no. 1 (2000). Fiction texts at 80 / 90 / 95 / 100% coverage; N = 66 university students; ~98% for adequate unassisted comprehension.
• Laufer, Batia. “What Percentage of Text-Lexis Is Essential for Comprehension?” In Special Language: From Humans Thinking to Thinking Machines, edited by Christer Laurén and Marianne Nordman, 316–323. Clevedon: Multilingual Matters, 1989. The ~95% “minimally adequate” figure and the later Nation / Laufer coverage reviews.
• Nguyen, Q. N., and D. T. H. Doan. “Beyond Comprehensible Input: A Neuro-Ecological Critique of Krashen’s Hypothesis.” Frontiers in Psychology 16 (16 October 2025): 1636777. Necessary-but-not-sufficient; does not argue that input should stop being comprehensible.
• McKee, Robert. Story: Substance, Structure, Style, and the Principles of Screenwriting. New York: ReganBooks, 1997. The gap as expectation minus result.
• Egri, Lajos. The Art of Dramatic Writing. New York: Simon & Schuster, 1946. Premise (X leads to Y) held apart from bone structure.
• Harmon, Dan. Channel 101 / Community lectures on the story circle. The eight-beat nest; claimed to scale from scene to season. Lineage through Christopher Vogler, The Writer’s Journey, and Joseph Campbell, The Hero with a Thousand Faces.
• Eliot, T. S. “Hamlet and His Problems.” In The Sacred Wood. London: Methuen, 1920. Objective correlative: a chosen object standing in for a feeling.
• Hemingway, Ernest. Death in the Afternoon. New York: Scribner, 1932. Omission, and the warning that it fails when hollow.