Students concentrating on English reading work during a small-group lesson at Singapore English Tuition Centre.

The Creative Writing Handbook Vol No.001 | Creative Writing Exercises With Constraints: How Limits Make Writers More Inventive

A blank page offers infinite freedom. Infinite freedom is often a terrible training environment.

EDK-SETC-CWH-0010 · Return to The Creative Writing Handbook

Ask a learner to “write something creative” and you may accidentally give them twenty problems at once. They must find a subject, choose a genre, invent characters, decide point of view, create a setting, establish a voice, control a beginning, decide what happens next, manage sentences, judge vocabulary, remember spelling and punctuation, and somehow work out whether the result is any good. A confident writer may enjoy that freedom. A developing writer may freeze, repeat the same safe story, or spend the entire session decorating a first paragraph.

A constraint changes the job. “Write a scene in which two people want different things, but neither may state what they wants directly.” “Describe a room without naming its purpose.” “Retell the same event from one metre away and then from fifty metres away.” “Write 150 words in which the important object appears only twice.” Suddenly the writer is not trying to solve all of writing. One relationship has been made visible.

A useful constraint removes decisions so that another decision can be practised more deeply.

This is why limits can make writers more inventive. The effect is not magical. A random restriction does not automatically create originality. The value comes from design: the constraint must be narrow enough to change behaviour, clear enough to follow, and temporary enough that the writer can later recover full choice.

The problem with “be more creative”

Creativity is often discussed at a level too large for instruction. A teacher may see a predictable story and say, “Try to be more original.” A parent may see a plain description and suggest, “Use your imagination.” A learner may read an accomplished story and decide, “I need better ideas.” These comments describe the desired direction without identifying a controllable action.

Suppose Maren writes the same type of opening repeatedly: weather, alarm clock, mirror, breakfast, departure. Her difficulty may not be lack of imagination. She may be choosing the earliest chronological point because she has no alternative decision rule. Telling her to “be creative” leaves the rule untouched. A constraint such as “begin after the most important event has already happened” forces a new structural question. Another such as “the first paragraph may contain no background explanation” makes her discover how action, object and dialogue can carry context.

Iona may produce detailed scenes but explain every implication immediately. Her reader is never allowed to infer. The useful constraint is different: “For one page, remove every sentence that names an emotion. Keep the emotion available through behaviour, setting, dialogue or selection.” Leonie may have strong ideas but spend too long choosing among them. Her exercise might be “choose the second workable idea, not the best idea, and start within three minutes.”

The same instruction should not be given to all three. Constraints are diagnostic tools. They are most useful when they target a behaviour that can actually change.

What a constraint is—and what it is not

A creative-writing constraint is a temporary rule that limits one or more choices during generation, drafting or revision. It may restrict length, vocabulary, point of view, information, structure, time, setting, chronology, sentence form, character knowledge, dialogue, description or any other controllable feature of the work.

It is not necessarily a rule for good writing. “Use no adjectives” is a useful exercise because it makes nouns, verbs and selection visible. It would be absurd as a permanent law. “Write exactly 100 words” can teach compression. It does not mean 100 words is an ideal story length. “Write the scene without dialogue” can reveal how much a writer depends on speech. It does not establish that dialogue is weak.

The distinction is essential. Training constraints are like weighted drills in sport. They exaggerate one condition so that the learner can feel and control something that was previously hidden. The athlete does not compete forever with the training device attached. The writer should not mistake the exercise for the whole craft.

The four jobs of a useful constraint

1. Reduce decision load

If the exercise gives you a place, a character relationship and a time limit, you no longer have to invent those before practising scene construction. The reduction is not cheating. It allows attention to move.

2. Expose a habit

A writer may not realise how often they explain motives until explanation is temporarily forbidden. They may not realise how much every paragraph depends on dialogue until dialogue is removed. A constraint creates contrast with normal behaviour.

3. Generate alternatives

When the first route is blocked, the writer searches elsewhere. If a character cannot say “I am angry”, the scene may use silence, misplaced politeness, repetition, physical distance, an object handled too carefully or a completely different strategy. The exercise increases the number of available solutions.

4. Make comparison possible

Rewrite the same paragraph under two constraints and the effect of a choice becomes easier to see. Comparison is powerful because writing quality is difficult to discuss in the abstract. Two versions give the learner evidence.

Constraint design begins with the question: what are we trying to notice?

A poor exercise starts with novelty. “Write a story without the letter E” is amusing and can produce surprising language, but it should not be assigned merely because it sounds creative. A stronger teaching sequence begins with a learning problem and chooses a restriction that changes the relevant behaviour.

If scenes feel static, constrain the character: they must want something specific before the first line ends. If dialogue sounds like information delivery, constrain the exchange: neither speaker may directly state the subject they are arguing about. If description is decorative, constrain selection: every descriptive detail must alter a later action, judgement or possibility. If endings become moral summaries, constrain the final paragraph: no abstract nouns such as lesson, friendship, courage, kindness or responsibility.

The better the diagnostic question, the less arbitrary the constraint becomes.

Constraint family 1: time

Time limits are often misunderstood as speed training. They can be, but their more interesting use is decision training. A five-minute exercise prevents the writer from treating every sentence as permanent. It encourages commitment, produces material quickly and makes hesitation visible.

Try three versions of the same prompt: three minutes, ten minutes, thirty minutes. Do not compare them only for polish. Compare the decisions. Does the three-minute version enter the scene later? Does the thirty-minute version accumulate unnecessary setup? Does extra time produce depth or simply more editing of the opening?

Leonie discovers that she is strongest at twelve minutes. At three she writes before thinking; at thirty she begins replacing perfectly functional sentences. That observation creates a training question: can she use the first twelve minutes to discover the scene, then deliberately switch modes rather than continue polishing locally?

Constraint family 2: length

Word-count limits train selection. The writer has to decide which action, detail, line of dialogue or piece of context earns space. The smallest useful exercise is not necessarily “write a very short story”. It may be “write this 300-word scene in 180 words without changing what the reader needs to understand”.

Then reverse the task. Expand 180 words to 300 without adding a new event. This exposes another skill: deepening an existing moment through perception, tension, sensory evidence, internal conflict or consequence rather than multiplying plot.

Maren learns something important during expansion. Her instinct is to add a flashback. The exercise forbids new events, so she has to stay. She notices a hand that does not release a cup, an answer that arrives too quickly and a chair angled away from the table. The scene acquires pressure without acquiring another incident.

Constraint family 3: information

Information constraints teach control of reader knowledge. They are among the most powerful exercises because point of view, suspense, inference and narrative distance all depend on what the reader receives and when.

Write a scene in which the writer knows why a character is leaving but the narrator may not state the reason. Write another in which the viewpoint character misunderstands the reason. Write a third in which the reader knows something the viewpoint character does not. Compare the resulting tension.

Iona enjoys this family because it turns reading discipline into writing craft. She has to separate what happened, what the viewpoint character knows, what the narrator tells and what the reader infers. A single careless sentence can collapse the design by revealing information from outside the chosen perspective.

Constraint family 4: point of view

Point-of-view exercises are most useful when the underlying event stays fixed. If everything changes, the writer cannot see what the viewpoint itself changed.

Take one small event: a student receives a message and places the phone face down. Write it in first person from the student, third-person limited close to a friend, and an external narrator who reports only visible action and audible speech. Do not change the actions. Notice what becomes easy, difficult or impossible in each version.

The purpose is not to decide which point of view is best. The purpose is to feel the trade-off. Intimacy gives one kind of access; distance gives another. A first-person narrator can reveal a private rationalisation but may know nothing about the friend’s expression once they turn away. An external narrator preserves uncertainty but must create interest through selection, action and speech.

For a deeper craft route, continue into the established eduKate point-of-view materials rather than treating this exercise as a complete theory of narration.

Constraint family 5: narrative distance

Narrative distance is easier to feel than to define when learners are beginning. Give them a single event and ask for three camera-like positions—not because prose is literally a camera, but because the spatial metaphor makes attention visible.

Version one stays across the street: bodies, movement, weather, no private thought. Version two stands at the table: gesture, voice, objects, small reactions. Version three sits inside one character’s attention: associations, bodily sensation, fragments of thought, selective noticing. Afterward, underline details that could exist only at that distance.

Maren finds that her “close” version still describes the whole room evenly. The constraint has exposed a hidden habit: she understands closeness as adding thoughts rather than narrowing attention. The next exercise limits her to details the character would plausibly notice while wanting something urgently.

Constraint family 6: dialogue

Dialogue becomes more interesting when it has a job beyond transferring information. Constraints can isolate subtext, power, rhythm and relationship.

  • Two characters discuss a broken lamp while actually negotiating whether one will leave.
  • Neither speaker may ask a direct question.
  • One speaker may use no sentence longer than six words.
  • One speaker knows the truth; the other suspects it; neither may name it.
  • The scene must contain a change of power without any insult or threat.
  • Remove all dialogue tags; revise only if speaker identity becomes unclear.

These exercises make dialogue dynamic because the writer has to control more than literal meaning. After the draft, read it aloud. Ask what each line tries to obtain, avoid, conceal, test or change.

Constraint family 7: description

Description improves when the writer stops treating it as inventory. A useful constraint gives selection a reason.

Describe a kitchen through the eyes of someone who has come to apologise. Describe the same kitchen through the eyes of someone looking for evidence. Do not change the room. The two narrators should select different details because purpose changes attention.

Now remove visual description and use sound, temperature, texture, smell and spatial relation. The goal is not to ban sight forever. It is to notice how quickly “description” becomes synonymous with visual adjectives.

Finally, impose a consequence rule: every descriptive detail must affect a later action, inference or mood. If the wet floor matters, someone slips, slows, changes route or notices the leak. If the clock matters, it changes urgency. If the open window matters, sound enters, cold enters, privacy changes or an object moves. Description becomes part of causality.

Constraint family 8: character desire

A scene often becomes active when somebody wants something that cannot be obtained without friction. The desire does not have to be dramatic. A person wants to leave without being rude. A student wants a teacher to notice an error without exposing a friend. A child wants to keep a found object. A parent wants a conversation to end before a particular question appears.

Write a scene in which the viewpoint character’s desire is clear by the end of the first paragraph, but the narrator may not state it directly. Then write a second version in which the character wants the opposite. Keep the setting and other character the same.

The exercise teaches that desire reorganises the world. The same door becomes escape or obstacle. The same pause becomes relief or threat. The same question becomes opportunity or danger.

Constraint family 9: causality

Writers often learn plot as a sequence of named stages. Stages can help, but causality is more fundamental. What happens next should often be shaped by what happened before: a decision, misunderstanding, pressure, discovery or consequence alters the available future.

Use the “because / therefore” constraint. Between each major event in a short outline, you must be able to write either “because” or “therefore” honestly. If the only available link is “and then”, investigate whether the event is merely adjacent.

This is not a prohibition on coincidence. Stories can contain coincidence. The exercise simply reveals when coincidence is doing all the structural work.

Constraint family 10: sentence architecture

Sentence exercises become useful when they connect form to effect. “Use a variety of sentence lengths” is vague. Try something testable.

  • Write a tense moment using no sentence longer than eight words. Then revise it with one long sentence at the point of greatest pressure.
  • Write a reflective paragraph in which each sentence begins from information established by the sentence before.
  • Write five sentences without using the verb “to be”; then restore it where removing it made the prose unnatural.
  • Write a paragraph with no adverbs. Recover only the adverbs that add meaning you cannot create efficiently elsewhere.
  • Write a paragraph containing one deliberate fragment. Explain why it earns the break.

The final explanation matters. A constraint without reflection can become a puzzle. Reflection turns the puzzle into craft knowledge.

Constraint family 11: vocabulary

Vocabulary constraints can counter two opposite problems: narrow language and ornamental language. For the first, create a small semantic field and require distinctions. For the second, remove the temptation to display rarity.

Exercise A: write a scene of disagreement without using angry, upset, annoyed, furious or any synonym that directly labels the emotion. Exercise B: describe movement using ten precise verbs across the draft, but no verb may be chosen merely because it sounds unusual. Exercise C: rewrite a paragraph so that every noun and verb is ordinary enough for a general reader, while preserving the exact meaning.

Iona notices that the plain-language version is harder than expected. She can no longer hide a vague thought behind a prestigious word. Precision and simplicity turn out not to be opposites.

Constraint family 12: chronology

Chronology feels natural, so writers often use it without choosing it. Change the starting point and the whole piece may change.

Write a 500-word event beginning at its earliest logical moment. Then begin again one minute before the central decision. Then begin after the consequence and let the earlier event emerge later. Keep the underlying event constant.

Ask which version creates the most useful question for the reader. “What will happen?” is only one possible engine. “Why did this happen?”, “what has this person done?”, “what are they hiding?” and “can this be repaired?” can all create forward movement.

Constraint family 13: object

Objects are useful because they make abstraction physical. Give the scene one ordinary object that must change function three times. A key begins as access, becomes evidence and ends as a burden. A receipt begins as rubbish, becomes proof and later becomes something someone cannot bear to throw away.

The object constraint can teach motif without requiring the learner to begin with the word symbolism. The object gathers meaning through use. Later, the writer can decide whether the pattern is subtle, excessive or accidental.

Constraint family 14: omission

Omission is one of the hardest forms of control because the writer knows what has been removed. The reader does not.

Write a scene about a farewell without using leave, goodbye, miss, last, gone or any direct statement that the characters are separating. The purpose is not to create a guessing game. It is to explore how action, object and dialogue can carry knowledge.

Then give the draft to a reader. Ask what they understood and which details produced that understanding. If nobody recognises the farewell, the exercise has not “failed”; it has shown that the selected evidence was insufficient. Revision now has a concrete job.

Constraint family 15: transformation

Transformation exercises teach adaptability. Take one piece of material and change its form while preserving a central relationship.

  • Turn a diary entry into a scene.
  • Turn a scene into a monologue.
  • Turn a monologue into a poem.
  • Turn a news item into a fictional situation without copying identifiable people.
  • Turn a photograph into three incompatible story premises.
  • Turn a description into an argument between two observers about what the place means.

The material is not sacred. The exercise shows that form changes what can be foregrounded. A diary can state what a scene must dramatise. A poem can compress what narrative explains. A monologue can reveal a voice that third-person summary kept distant.

Constraint family 16: genre

Genre constraints are useful when they become expectation experiments rather than costume changes. Give the same premise—two siblings open a locked box after a funeral—and write the first 300 words as mystery, comedy, horror and reflective literary fiction. Do not merely swap adjectives. Change what the opening promises, what information matters, how quickly the reader receives it and what kind of question pulls the piece forward.

Genre is a relationship with readers. The exercise helps learners see that expectation is part of craft. A locked box in comedy and a locked box in horror can be physically identical while the surrounding signals make different futures plausible.

Constraint family 17: research

Creative writing is not limited to what the writer already knows. Research expands possibility, but uncontrolled research can postpone writing indefinitely. Use a research constraint: twenty minutes, three reliable facts, one sensory or procedural detail, one question that remains unanswered. Then draft.

The restriction forces research to serve the piece rather than become a substitute for it. Later drafts can identify what still needs verification. This is particularly useful for historical settings, professions, technologies and places the writer has not experienced directly.

Constraint family 18: revision

Revision constraints can separate levels of editing. One pass may change only structure. Another may remove explanation. Another may target reference clarity. Another may test sentence rhythm aloud.

Try the deletion pass: reduce a draft by ten per cent without removing an event. Then try the restoration pass: restore only material whose absence creates a real loss. The comparison reveals which sentences were merely attached to the draft by familiarity.

Maren finds this painful because she has several lines she likes. She creates a “spare lines” document rather than forcing them to survive in the story. The psychological change matters. Deleting from the current work no longer feels like destroying the sentence forever.

The one-variable rule

When you want to learn from comparison, change one major variable at a time. If the second version uses a new point of view, a different setting, a new ending and a comic tone, you may prefer it but learn little about why.

The one-variable rule is not a law for all drafting. It is a laboratory rule for training. Keep the event stable and change distance. Keep the words stable and change punctuation. Keep the scene stable and change what the character wants. Keep the plot stable and change who knows the secret.

Creative work is normally interconnected. The training exercise temporarily separates the wires so the learner can see what one wire does.

The contrast pair: one of the highest-value writing exercises

Instead of writing ten unrelated exercises, write two deliberately contrasted versions. Contrast creates a local experiment.

  • Tell everything / imply the central fact.
  • Very close / deliberately distant.
  • Dialogue-heavy / no dialogue.
  • Chronological / aftermath first.
  • Fast scene / slowed scene.
  • Character succeeds / character refuses the opportunity.
  • Decorative description / consequential description.
  • Maximum certainty / carefully qualified uncertainty.

Afterward, the learner should write three observations beginning with “When I changed X, the reader…” or “When I changed X, I had to…”. This forces craft language to emerge from actual writing.

Fifty compact constraint exercises

The exercises below are not a checklist to complete mechanically. Choose the one that matches the current writing problem.

  1. Write for five minutes without deleting.
  2. Write 100 words in which only one action occurs.
  3. Begin after the event you first planned as the climax.
  4. Use no direct emotion words.
  5. Use no visual description.
  6. Describe a place through sound only.
  7. Describe the same place through touch and temperature.
  8. Write a scene in which nobody says what they want.
  9. Give each speaker a different sentence-length pattern.
  10. Remove all dialogue and preserve the conflict.
  11. Remove all internal thought and preserve the motive.
  12. Write from someone who misunderstands the situation.
  13. Write from someone who knows the ending.
  14. Write from someone who knows less than the reader.
  15. Write the event from outside the room.
  16. Write the event from inside one body.
  17. Write the same 200 words in present and past tense.
  18. Use exactly three paragraphs.
  19. Write one paragraph containing one decision and one consequence.
  20. Make an ordinary object necessary to the plot.
  21. Make the same object emotionally important without explaining why.
  22. Use one repeated image no more than three times.
  23. Cut ten per cent of the draft.
  24. Cut the opening paragraph entirely and repair the transition.
  25. Cut the final explanation and test whether the ending still lands.
  26. Write the scene with no adjectives.
  27. Restore only the adjectives that change meaning.
  28. Write the scene with no adverbs.
  29. Restore only the adverbs that cannot be replaced cleanly.
  30. Use only ordinary vocabulary and maximise precision.
  31. Write one deliberately long sentence at the moment of pressure.
  32. Write one deliberately short sentence after it.
  33. Use no rhetorical question.
  34. Use one rhetorical question and make it necessary.
  35. Write a scene where the important answer is “no”.
  36. Write the same scene where the important answer is “yes”.
  37. Keep every event and change the genre.
  38. Keep every event and change the relationship between the characters.
  39. Keep every event and change the setting so one action becomes impossible.
  40. Write three openings for the same story: action, image, dialogue.
  41. Write three endings: resolved, open, reversed.
  42. Choose one ending and make the earlier scene prepare for it.
  43. Write a 50-word character portrait without physical appearance.
  44. Write a 50-word character portrait using only physical behaviour.
  45. Write a conversation where one speaker changes the subject twice.
  46. Write a scene in which silence changes the outcome.
  47. Write from a photograph but invent everything beyond visible evidence.
  48. Write from a memory, then alter one factual condition and follow the consequences.
  49. Turn a scene into a letter written the next day.
  50. Take the skill from one exercise and use it in a completely new piece without the original restriction.

How to choose the right difficulty

A constraint should create productive difficulty, not paralysis. If the learner spends all available attention remembering the rule, the craft target disappears. Begin with one restriction. Add another only when the first has become manageable.

For a novice, “write a scene without direct emotion labels” may be enough. For an advanced writer, the same exercise can add a second layer: the viewpoint character misidentifies their own emotion. The principle is scalable because complexity can be added to the thinking, not merely the vocabulary.

Leonie uses a simple calibration test. If she cannot explain the rule in one sentence before starting, the exercise is probably carrying too many conditions. If she finishes without ever noticing the restriction, it may be too easy or irrelevant. The useful zone is where the constraint changes decisions repeatedly but does not dominate every thought.

Do not grade the weirdness

Constraint exercises can produce unusual writing. That is not the same as successful writing. A bizarre sentence may be inventive and unclear. A conventional sentence may solve the chosen problem perfectly.

Assess the exercise against its training question. If the aim is subtext, can the reader identify a tension not directly stated? If the aim is selection, did the shorter version preserve the essential movement? If the aim is viewpoint, does information remain inside the chosen access? If the aim is descriptive consequence, do details affect what happens?

Originality is difficult to score directly and easy to distort into performance. Better to evaluate decisions the learner can understand and repeat.

Feedback after a constraint exercise

Do not begin with “I liked this one better.” Preference may be useful later, but first recover the mechanism.

  • Where did the constraint change your normal decision?
  • What became easier?
  • What became harder?
  • What did you discover that you would keep after removing the constraint?
  • Where did the writing become artificial because you obeyed the rule too literally?
  • Which sentence or moment demonstrates the learning most clearly?
  • What new piece could test the same skill without repeating the same exercise?

Iona often asks one extra question: “What did the reader actually know at this point?” That single question catches many failures in point of view, suspense and implication because the writer unconsciously fills gaps from their own private plan.

When constraints go wrong

Too many rules

“Write 300 words in second person, no adjectives, exactly five paragraphs, three similes, one flashback and no word longer than eight letters” may be entertaining, but the writer spends attention on compliance rather than craft. If several constraints are used, they should cooperate around one training purpose.

Rules treated as doctrine

A learner completes the “no adverbs” exercise and concludes that adverbs are bad. The teacher must explicitly reopen the choice: restore one where it improves precision or rhythm. Training is complete only when full control returns.

Novelty without diagnosis

The exercise is memorable but unrelated to the learner’s problem. This is not harmful occasionally—play has value—but a training programme should know when it is playing and when it is repairing.

Constraint dependence

The learner becomes excellent at responding to highly specified prompts but still freezes when choice returns. The remedy is fading: strong constraint, lighter constraint, self-chosen constraint, independent piece.

The fade sequence

Every important constraint should eventually disappear. Use four stages.

  1. Imposed: the teacher or guide selects the restriction because the weak link is known.
  2. Chosen from options: the learner selects one of several constraints and explains the choice.
  3. Self-designed: the learner diagnoses the current piece and invents a useful temporary rule.
  4. Independent: no explicit constraint is required; the writer applies the underlying principle when needed.

The fourth stage is the real target. A mature writer may still use constraints voluntarily, but now they are tools chosen from a larger repertoire.

A classroom protocol: write, compare, name, transfer

A 45-minute session can be surprisingly productive when it does not attempt a complete story.

  • 5 minutes: read a short stimulus and define one craft target.
  • 8 minutes: write version A under the first condition.
  • 8 minutes: write version B with one variable changed.
  • 8 minutes: compare in pairs; readers describe effects before giving advice.
  • 8 minutes: revise one version.
  • 8 minutes: begin a new piece using the discovered principle without copying the original scenario.

This structure produces less finished writing than a full-essay lesson and often more learning. The transfer attempt prevents the session from ending inside the exercise.

A home protocol: ten-minute craft

Parents and independent learners do not need elaborate worksheets. Choose one small constraint and one neutral prompt. Ten minutes is enough.

For example: “Write a person entering the kitchen. Do not state how they feel. Afterward, underline the details that carry emotion.” The parent can ask what each detail suggests. Avoid rewriting the paragraph in adult language. The learner needs ownership of the next decision.

Keep a folder of these miniatures. Over time, the collection becomes evidence of how the writer solves different problems.

A four-week constraint foundation

Week 1: attention and selection

Use sensory restrictions, 100-word scenes, one-object exercises and description-with-consequence. The goal is to make choice visible.

Week 2: character and dialogue

Use hidden desire, subtext, no-emotion-labels and dialogue-without-direct-questions. The goal is to create pressure without overexplaining.

Week 3: point of view and information

Rewrite the same event from different access levels. Track what each narrator can know. Introduce misunderstanding deliberately.

Week 4: revision and transfer

Cut, reorder, change distance, then remove all explicit exercise rules and write one independent scene. Diagnose whether the learned choices appear without prompting.

A twelve-week progression

For a deeper programme, spend two weeks on production habits and idea generation, two on character and conflict, two on viewpoint and distance, two on dialogue and scene, two on structure and pace, and two on revision and transfer. Each week should contain at least one contrast pair and one new-piece transfer task.

Do not make every week harder by adding more rules. Difficulty should come from subtler decisions. An advanced point-of-view exercise may contain fewer explicit restrictions than a beginner exercise because the writer must diagnose and manage the boundary independently.

Using constraints with model texts

Published writing can support constraint practice without becoming a sentence bank. Read a short passage and identify one technique at the level of decision: perhaps the writer withholds the name of an object, shifts into scene after summary, lets dialogue contradict narration, or uses a repeated concrete image.

Then write new material under an analogous constraint. Use different characters, situation, imagery and language. The learner is studying architecture, not copying voice.

This is also safer for originality. The most valuable thing to borrow from strong writing is often a question: what happens if I withhold this information? What happens if the scene stays inside one perception? What happens if the important fact appears as an object before it appears as an explanation?

Constraint journaling

After each exercise, record four lines:

  • Constraint: what rule did I use?
  • Effect: which decision changed because of it?
  • Discovery: what technique became available?
  • Transfer: where else might this help?

These notes become more valuable than a score. They build metacognitive language for craft. A writer who can say “I move too quickly into explanation when I am uncertain whether the scene is clear” has a solvable problem. “My writing is not creative enough” is much harder to act on.

The writer eventually becomes the constraint designer

Professional and experienced writers often create their own restrictions without calling them exercises. A novelist decides a chapter will remain inside one character’s perception. A poet chooses a formal pattern. A memoirist decides not to reconstruct dialogue that cannot be remembered responsibly. A short-story writer limits the action to one room and one afternoon. A journalist works inside evidence and length constraints. Form is full of chosen limits.

The developing writer should gradually gain this power. When a draft feels diffuse, they may impose a temporary central question. When a character feels overexplained, they may remove direct interior labels for one pass. When a scene feels visually flat, they may revise from sound and touch. When an argument between characters becomes too explicit, they may forbid the subject for two pages.

At this point the constraint has become a diagnostic instrument rather than an assignment.

How Maren uses one constraint to repair a real story

Maren has a draft about two friends clearing a bedroom before one family moves away. The first page explains the friendship, the move, the sadness and the importance of a cardboard box. All the information is relevant. None of it is alive.

Her constraint is severe but temporary: for 400 words she may not mention moving, friendship, sadness, memory or goodbye. She must keep both characters in the room with the box.

At first the exercise feels artificial. Then one friend keeps putting objects into the box and the other keeps taking them out under practical pretexts. A train ticket is “rubbish” until somebody reaches for the bin. A cheap keyring is “not worth packing” until neither person will throw it away. The central fact has not vanished. It has moved into action.

During revision, Maren is allowed to restore direct context. She adds one plain sentence establishing that the room must be empty by tomorrow. She does not restore the original paragraph explaining the sadness. The constraint has shown her that much of the emotion already exists in the objects and choices.

The final story is not “a story with no emotion words”. It is a story in which direct explanation and enacted detail now have a better balance.

How Iona uses one constraint to repair point of view

Iona writes a scene from a younger sister’s viewpoint. Halfway through, the narration explains exactly why the older sister avoids a question. The explanation is psychologically sensible—and impossible for the viewpoint character to know.

Her constraint: revise the scene using only what the younger sister can observe, remember, infer or be told. Every sentence containing inaccessible knowledge receives a mark.

The repair creates uncertainty. The older sister folds a receipt twice before answering. She responds to a different part of the question. She asks whether the kettle has boiled even though it has already clicked off. The younger sister begins to interpret these details, but the narration does not certify her interpretation.

Iona discovers that point-of-view discipline has not reduced information. It has created a different kind of information: behaviour that can be read.

How Leonie uses one constraint to protect completion

Leonie’s problem appears during longer work. She writes slowly because she revises each paragraph before moving. By the time the central scene begins, the session is nearly over.

Her constraint is procedural: during the first twenty minutes, she may add text and make only corrections that protect meaning. She places a simple marker beside sentences she wants to revisit. No stylistic rewriting until the timer ends.

The exercise reveals that several sentences she would normally polish disappear when later events change the scene. She has been spending high-quality attention on unstable material. In the second phase she revises with more context and less attachment.

Again, the rule is not permanent. Some writers draft slowly and successfully. Leonie uses the constraint because it interrupts a specific process failure: local perfection blocking global completion.

Creative freedom returns at the end

The strongest argument for constraint practice is not that limits are inherently creative. It is that temporary limits can increase future freedom. A writer who has practised five ways of handling emotion has more options when no rule is present. A writer who has rewritten distance deliberately can choose closeness rather than defaulting to it. A writer who has compressed a scene learns what can be removed; a writer who has expanded one without new events learns how depth differs from plot multiplication.

Constraint practice should therefore end in open work. Write a new scene with no explicit restriction. Afterwards, diagnose which decisions from the exercises appeared naturally. If none appeared, the exercise may not have transferred yet. Return with a lighter scaffold, not a louder instruction.

Where to go next

Constraints work best when the writer has material to work on. The next volume builds that material deliberately: observation, fragments, questions, overheard rhythms, remembered objects, possible conflicts and small records of the world that can later become stories.

Give yourself less to decide for ten minutes. Learn something about the decision that remains. Then take the rule away and write with more choices than you had before.