Checked and rebuilt: 17 September 2026. This legacy Bukit Timah creative-writing URL now has one clear educational owner: high-achiever precision without performative writing. The page is a craft guide for strong Primary writers whose language is already competent but whose prose can become over-engineered, over-intense, adult-sounding or too eager to display sophistication. The historical photo archive from the original page is retained at the end for continuity; the public teaching job is the writing system below.
Advanced writing is often misunderstood as a race toward harder vocabulary, longer sentences, denser imagery and more dramatic emotion. That direction can produce prose that looks impressive at first glance while becoming less accurate, less believable and less recognisably the student’s own. A rare word can distort the scene. A metaphor can overstate a small moment. A long sentence can blur cause and time. A polished phrase can sound imported from an adult model rather than generated by the child’s actual viewpoint.
The stronger target is different: fidelity under tighter control. The writer chooses the exact intensity the event has earned, the register the narrator can plausibly carry, the sentence structure that represents the relationship among ideas, the amount of detail the reader needs, and the degree of explanation that still leaves room for inference.
Quick answer for busy parents
- Advanced writing is not harder wording. It is more accurate control over meaning, intensity, register, viewpoint and reader effect.
- Specific beats impressive. A precise ordinary verb can outperform a rare inaccurate one.
- Intensity must be earned. A small inconvenience should not automatically trigger catastrophic language.
- Collocation matters. Knowing a word includes knowing what other words it naturally works with.
- Register matters. Grammatically correct language can still sound wrong for the narrator, scene or genre.
- Figurative language should sharpen perception. If deleting the metaphor changes nothing except “style”, it may not be earning its place.
- Sentence variety is a meaning decision. Cause, contrast, time, condition, accumulation and discovery create different syntax for different jobs.
- Inference is a form of trust. Do not explain emotions that action, dialogue and context already make clear.
- Restraint is not blandness. It is refusing language that exceeds the scene.
- Voice must remain the student’s. Adult editing or AI should not replace the writer’s judgement.
- Revision should diagnose. Ask what is inaccurate, excessive, unclear or unnecessary before rewriting everything.
- The mature writer can defend a simpler choice.
The scientific job of this article
This guide helps parents and strong Primary writers answer:
- What does “advanced” writing actually mean?
- How can vocabulary become more precise without becoming unnatural?
- How should emotional intensity, register and figurative language be calibrated?
- How does sentence structure carry logic rather than decoration?
- How much should a writer explain, and what should be left for the reader to infer?
- How can a student revise aggressively without losing personal voice?
- How should high-achiever writing develop under examination time constraints?
- How can AI be used for contrast and feedback without replacing authorship?
The central principle: difficulty is not sophistication
A difficult word is not automatically a sophisticated word. A difficult sentence is not automatically a sophisticated sentence. A difficult metaphor is not automatically a sophisticated metaphor.
Sophistication is better understood as fit:
- Does the word fit the exact meaning?
- Does the intensity fit the event?
- Does the register fit the narrator?
- Does the image fit the physical scene?
- Does the syntax fit the relationship?
- Does the paragraph fit the story’s pressure?
- Does the explanation fit what the reader still needs?
A simpler choice can therefore be the more advanced choice when it produces greater fidelity.
The precision ladder
Before celebrating a “powerful” phrase, move through five gates.
- Correct: does it mean what the student thinks it means?
- Natural: is it idiomatic and plausible in this context?
- Precise: does it represent the exact action, relation, intensity or attitude?
- Necessary: does the scene benefit from this degree of specificity?
- Coherent: does the choice fit the surrounding voice, pace and reader effect?
A word can pass the first gate and fail the next four.
Precision layer 1: semantic accuracy
Strong writers begin with meaning. Near-synonyms are not interchangeable decorations.
Consider:
- walked;
- strolled;
- marched;
- staggered;
- crept;
- hurried.
Each verb encodes a different physical pattern, intention or emotional implication. A child who replaces every form of walked with the rarest verb in a thesaurus can reduce accuracy rather than increase it.
The verb test
Before accepting a stronger verb, ask:
- What did the body actually do?
- How fast?
- With what balance or force?
- Did the character intend the action?
- What emotion, if any, does the verb imply?
- Does the surrounding scene support that implication?
Precision layer 2: specificity
Specificity is useful when it changes the reader’s model.
Compare:
He put the object on the table.
He set the cracked phone beside the untouched bowl of noodles.
The second version adds useful information only if the cracked phone, untouched food or placement matters to the scene. Specificity should not become random catalogue.
The relevance test for detail
Ask:
- Does this detail reveal action?
- Does it reveal character?
- Does it establish a future object or consequence?
- Does it orient the reader spatially?
- Does it change tone?
- Does it create a meaningful contrast?
If not, detail can become clutter.
Precision layer 3: intensity calibration
Disappointed, upset, distressed, devastated and shattered do not form a ladder that students should automatically climb. They represent different psychological and situational intensities.
Before choosing an intense word, ask:
- What was lost or threatened?
- How unexpected was the event?
- What personal significance had been established?
- How responsible does the character feel?
- What behaviour would this state plausibly produce?
- How long would the reaction reasonably last?
The event-to-language calibration test
If a character misplaces an ordinary pencil, “my world collapsed into a pit of despair” is usually not advanced. It is disproportionate unless the story has established unusual significance.
If the pencil was the last gift from a deceased grandparent and the story has earned that emotional history, the intensity may be different. Context sets the scale.
Precision layer 4: collocation
Knowing a word means more than knowing its dictionary definition. It includes knowing which words commonly accompany it.
A learner may know that commit relates to performing an act, but natural English says commit a crime, not every imaginable noun combination. A learner may know utterly means completely, but it does not combine naturally with every adjective.
Collocation is hidden fluency
Strong writing sounds effortless partly because common word partnerships are automatic. Weak high-achiever writing can sound translated or assembled because each individual word is “advanced” but their combination is unnatural.
How to practise collocation
- Compare two near-synonyms in the same sentence.
- Ask which noun each verb commonly takes.
- Notice prepositions in authentic reading.
- Record useful phrases rather than isolated words.
- Study near-miss examples, not only perfect model sentences.
Precision layer 5: register
Register is the level and social character of language appropriate to a situation, speaker, relationship and genre.
A sentence can be grammatical and still sound wrong because its register is too formal, too adult, too casual or too theatrical.
Compare:
I proceeded towards my companion with an expression of profound remorse etched upon my countenance.
with:
I walked towards Amelia, apology rehearsed in my head, but the words disappeared when she looked up.
The second version is not less sophisticated. It is more plausible for a young narrator and more precise about action, timing and mental state.
The narrator-fit test
Read the sentence aloud and ask:
- Would this narrator plausibly think this way?
- Would this child use this phrase?
- Does the sentence belong to the rest of the piece?
- Does the register change suddenly because the student remembered a memorised phrase?
Precision layer 6: viewpoint
Viewpoint controls what the narrator can know, notice and interpret.
A first-person narrator cannot normally describe another character’s private thoughts as fact unless those thoughts were communicated. A close third-person perspective should not jump freely between minds without a clear reason.
Viewpoint precision questions
- What can the narrator actually see?
- What can the narrator infer?
- What can the narrator know?
- What is still uncertain?
- Whose sensory experience is foregrounded?
Inference is not mind-reading
Instead of:
Mr Tan was furious and thought I was irresponsible.
a first-person narrator who cannot know his private thought might write:
Mr Tan looked at the broken model, then at me. He set the clipboard down very slowly.
The reader can infer anger and judgement without impossible access to the other mind.
Precision layer 7: physical plausibility
Figurative or dramatic writing can accidentally violate physical reality.
Students may write that a heart “stopped”, lungs “exploded”, eyes “popped out”, or the ground “swallowed” them in scenes where none of this literally occurs.
Hyperbole is possible, but the writer should understand the physical baseline before departing from it.
The body test
Ask what the body actually does under fear, embarrassment, exertion or anger:
- breathing changes;
- muscles tighten;
- hands become restless or still;
- voice changes;
- attention narrows;
- movement speeds up or hesitates.
Specific behaviour is often more convincing than a stock emotional metaphor.
Precision layer 8: figurative language
Simile, metaphor and personification are valuable when they sharpen perception or compress a relationship that literal language would handle poorly.
They become performative when their main job is to prove the writer knows a technique.
The metaphor deletion test
Delete the figure of speech. Ask:
- What meaning disappeared?
- What perception became less clear?
- What emotional relation was lost?
- What tone changed?
If nothing disappears except visible “style”, the metaphor may not be necessary.
Metaphor must respect scale
Draft:
The corridor was a ravenous monster swallowing my fragile soul into its endless abyss.
If the scene is simply a child walking nervously to the principal’s office, the image may exceed the event.
More restrained:
The principal’s door seemed farther away each time I looked up.
The second line lets physical perception carry anxiety.
Precision layer 9: sensory detail
Sensory writing is often taught as a checklist: include sight, sound, smell, taste and touch. Strong writing does not need all five senses in every scene.
Choose the sensory channel that the viewpoint character would notice and that changes the reader’s model.
Sensory relevance
A burning smell matters if it signals danger. The hum of an air-conditioner matters if silence makes it suddenly noticeable. The rough edge of a torn paper matters if the character is gripping it. Random sensory inventory slows the scene.
Precision layer 10: sentence architecture
Sentence variety should arise from varied reasoning.
Different relationships invite different structures:
| Relationship | Possible syntactic job |
|---|---|
| Cause | Make dependency explicit |
| Contrast | Hold competing states together |
| Condition | Limit when a claim or outcome holds |
| Accumulation | Build pressure through linked details |
| Discovery | Land new information cleanly |
| Delay | Withhold decisive information until later in the sentence |
| Simultaneity | Show actions occurring together |
| Interruption | Break an expected rhythm |
Short sentences are not beginner sentences
A short sentence can create clarity, emphasis, finality or shock.
Accumulation:
I checked the desk, the floor beneath it, the side pocket of my bag and finally the chair where I had been sitting.
Discovery:
The wallet was gone.
The change in sentence length follows information structure.
Long sentences need one controlling logic
A long sentence becomes weak when it contains several independent ideas joined because the student wants a “complex sentence”.
Ask:
- What is the main clause?
- What relation does each subordinate clause have to it?
- Is the timeline recoverable?
- Is the subject still clear?
- Would two sentences be easier?
Comma splices and run-ons are often thinking problems
A student may join clauses incorrectly because the relationships have not been classified. Once the writer identifies cause, contrast, sequence or independence, punctuation and clause structure become easier to choose.
Precision layer 11: information structure
Sentence order changes emphasis.
Because the road flooded, the bus changed route.
foregrounds cause.
The bus changed route because the road flooded.
foregrounds the action before explaining it.
Both may be accurate. The writer chooses based on reader attention.
Fronting should have a reason
After three failed attempts, the team changed its plan.
The fronted time phrase matters because it changes how the decision is interpreted. Fronting becomes performative when phrases are moved merely to make sentence openings look different.
Precision layer 12: paragraph movement
Strong sentences cannot rescue a paragraph with no movement. A paragraph should have a job.
Narrative paragraph jobs can include:
- orient;
- build pressure;
- complicate;
- reveal;
- decide;
- confront;
- recover;
- reflect.
Paragraph pressure
Ask what changes from the first sentence of the paragraph to the last. If nothing changes in knowledge, tension, decision, relationship or physical situation, the paragraph may be static.
Precision layer 13: scene versus summary
Scene slows down important moments. Summary compresses routine or lower-value time.
Weak high-achiever writing sometimes scenes everything because detailed writing feels “better”. The result is a story that spends equal attention on opening a bag, walking across a room and confronting the central conflict.
The significance test
Give more words to:
- choices;
- turning points;
- reversals;
- discoveries;
- confrontations;
- moments whose interpretation matters.
Compress routine transitions.
Precision layer 14: pacing
Pacing is not simply sentence length. It comes from how much story time is covered, how much detail is selected, how quickly new information arrives and how much reflection interrupts action.
Slow down for meaning, not for decoration
A turning point may deserve precise physical detail and internal hesitation. A routine bus ride may need one sentence.
Precision layer 15: dialogue
Dialogue should reveal something that narration cannot deliver as efficiently.
Useful dialogue jobs include:
- decision;
- conflict;
- avoidance;
- misunderstanding;
- relationship;
- new information;
- character voice.
Do not use dialogue to repeat narration
Weak:
I was angry. “I am very angry!” I shouted angrily.
The same information appears three times.
Stronger writing lets wording, action and dialogue distribute the work.
Dialogue tags and action beats
Students sometimes replace said with increasingly dramatic verbs. A simple tag is often invisible and effective. Action beats can reveal tone without forcing an exotic speech verb.
Precision layer 16: emotion through behaviour
Emotion words are useful, but strong narrative often lets behaviour carry the state.
Instead of:
I was extremely nervous.
the writer might show:
I read the first line of my cue card three times without taking in a word.
The behaviour gives the reader evidence.
Do not ban emotion words
“Show, don’t tell” is not a universal rule. Summary sometimes requires direct naming. The skill is choosing where evidence is worth the space and where a concise statement is more efficient.
Precision layer 17: inference
Inference is what the reader can conclude from evidence the writer supplies.
Strong writers leave some work for the reader when:
- action already reveals emotion;
- dialogue reveals conflict;
- objects carry established significance;
- contradictions imply uncertainty;
- the viewpoint character does not fully understand yet.
The deletion test for explanation
Delete the explanatory sentence. If the intended inference survives accurately, the explanation may be redundant.
If meaning becomes genuinely ambiguous, restore the minimum necessary guidance.
Precision layer 18: productive ambiguity
Not every uncertainty should be resolved immediately.
Productive ambiguity gives the reader several evidence-supported possibilities. Confusion gives the reader insufficient structure to interpret.
A strong writer can preserve uncertainty honestly when the character also lacks information.
Precision layer 19: cohesion
Cohesion is how sentences and paragraphs connect.
Useful cohesive devices include:
- reference;
- repetition of key nouns;
- logical connectors;
- parallel structure;
- lexical chains;
- cause-effect continuity;
- time continuity.
Synonym cycling can weaken cohesion
Students sometimes avoid repeating a noun at all costs. They replace dog with canine, creature, animal, four-legged companion and other substitutes. The result can be less clear than strategic repetition.
Precision layer 20: openings
Strong openings orient or create a meaningful question quickly. They do not need to begin with weather, a dream, a proverb or an explosion.
Possible opening jobs:
- place the reader inside an unstable situation;
- establish a relationship that will change;
- introduce an object whose significance will matter;
- begin at a decision point;
- create a precise unanswered question.
Opening restraint
If the opening uses the story’s maximum emotional and stylistic intensity before the central conflict has been earned, the rest of the story has nowhere to grow.
Precision layer 21: endings
An ending should resolve the story’s most important movement, not merely announce a moral.
Possible ending jobs:
- show consequence;
- reframe an earlier object;
- show changed behaviour;
- leave a bounded uncertainty;
- echo the opening with altered meaning;
- compress reflection into one earned observation.
Do not bolt on a lesson
I learnt that honesty is the best policy may be true, but if the story has already demonstrated the cost and repair of dishonesty, a direct moral can flatten the ending.
Precision layer 22: theme
Theme is not a slogan inserted at the end. It emerges from repeated choices, consequences, images and relationships.
A story about responsibility becomes more convincing when responsibility shapes actions throughout the narrative rather than appearing only in the final reflection.
Precision layer 23: originality
Originality does not require bizarre plots. It often comes from specific observation, an honest viewpoint, unusual but plausible relationships, and decisions that belong to the student rather than to a model-answer template.
Template risk
Memorised openings, universal metaphors and stock reflections can make many students sound the same. Templates can scaffold structure temporarily, but they should not replace situation-specific judgement.
Precision layer 24: revision
Revision should not mean “make every sentence better”. That instruction is too vague.
Use separate passes.
Pass 1: meaning
Does each sentence say what the writer intends?
Pass 2: intensity
Where is language stronger than the event?
Pass 3: viewpoint
Does the narrator know only what they can plausibly know?
Pass 4: physical action
Are movements possible and spatially clear?
Pass 5: sentence logic
Are clauses joined because of real relationships?
Pass 6: inference
What explanation can be deleted because evidence already carries it?
Pass 7: voice
Which phrases sound unlike the student?
Pass 8: economy
Which detail, metaphor, adverb or sentence can disappear with no loss?
The overwriting deletion test
Delete one phrase or sentence. Ask what useful meaning disappears.
If the answer is “it sounds less impressive”, the deleted material may have been performative.
The one-new-information rule
Every sentence should contribute something: new action, new evidence, new decision, new relation, new perception, new consequence or a deliberate rhythmic effect.
Repeated emotion with no new evidence creates verbal inflation.
The performative-writing test
Ask:
If the reader could not see how difficult this word, metaphor or structure is, would I still choose it?
If not, the choice may be aimed at displaying ability rather than serving the scene.
The adult-editor test
Ask:
- Would the student generate this phrase independently?
- Can the student explain why every word was chosen?
- Can the student produce a similar decision in a new story?
- Does the final paragraph still sound like the same writer?
If not, the page may be polished while the writer remains unchanged.
The capability atlas for high-achiever writing
The following capabilities allow parents and teachers to diagnose a strong writer more precisely than “needs better vocabulary”.
Lexical precision
- Chooses verbs by physical action.
- Distinguishes near-synonym intensity.
- Uses collocations naturally.
- Controls register.
- Avoids thesaurus substitution without context.
- Uses specific nouns when specificity matters.
- Uses modifiers selectively.
- Recognises when ordinary language is more precise.
Emotional calibration
- Matches emotion to event significance.
- Builds intensity gradually.
- Uses behaviour as evidence.
- Avoids redundant emotion labels.
- Distinguishes internal uncertainty from certainty.
- Connects emotion to character history when relevant.
- Allows emotional change across the scene.
Figurative control
- Uses imagery to clarify perception.
- Maintains physical plausibility.
- Avoids mixed metaphors.
- Matches image scale to event scale.
- Uses recurring imagery coherently.
- Deletes decorative figures.
- Can explain the exact relationship a metaphor creates.
Sentence architecture
- Chooses structures by logic.
- Controls clause boundaries.
- Uses short sentences deliberately.
- Uses long sentences with one controlling relationship.
- Controls sequence and simultaneity.
- Uses fronting for emphasis rather than variety alone.
- Avoids run-ons and comma splices.
- Varies rhythm according to information load.
Paragraph architecture
- Gives each paragraph a job.
- Creates change across the paragraph.
- Controls scene versus summary.
- Builds causal continuity.
- Uses transitions when relationships require them.
- Maintains spatial and temporal orientation.
- Places turning points with enough space.
Viewpoint and inference
- Knows what narrator can know.
- Separates observation from inference.
- Preserves uncertainty when appropriate.
- Uses action/dialogue as evidence.
- Avoids mind-reading.
- Controls narrative distance.
- Lets readers infer without losing clarity.
Voice and register
- Maintains narrator age and personality.
- Uses dialogue consistent with character.
- Avoids imported model-answer phrases.
- Controls formal/informal shifts.
- Maintains stylistic continuity.
- Can defend personal wording choices.
Revision control
- Diagnoses before rewriting.
- Separates meaning revision from proofreading.
- Deletes redundant explanation.
- Repairs ambiguity.
- Calibrates intensity.
- Replaces weak abstractions with useful specifics.
- Preserves student voice.
- Can transfer a repaired principle to new writing.
Symptom-to-cause diagnostic table
| What you see | Likely mechanism | First repair |
|---|---|---|
| Many rare words, prose sounds unnatural | Collocation/register weaker than vocabulary recall | Near-synonym and phrase comparison |
| Every emotion is extreme | Intensity calibration weak | Event-significance ladder |
| Metaphor in nearly every paragraph | Technique used as display | Metaphor deletion test |
| All sentences are long | Complexity confused with sophistication | Relationship-first sentence architecture |
| All sentences are short for drama | Surface rhythm imitation | Match sentence length to information job |
| Writer explains every feeling | Inference trust weak | Delete explanation after behavioural evidence |
| Adult-sounding phrases appear suddenly | Voice ownership weak | Read aloud and rewrite in narrator register |
| Rich description, little plot movement | Detail not tied to narrative pressure | Paragraph job and significance test |
| Strong opening, flat middle | Maximum intensity spent too early | Pressure curve planning |
| Ending states moral directly | Theme explained rather than embodied | Consequence/change ending |
| AI-revised paragraph is excellent, next paragraph weak | Tool replaced judgement | Contrastive feedback then independent transfer |
| Many grammar corrections recur | Revision fixes product, not writer system | One recurrent mechanism per revision pass |
Error taxonomy for strong writers
Semantic inflation
A word is stronger, rarer or more dramatic than the intended meaning.
Register mismatch
The sentence sounds unlike the narrator, genre or situation.
Collocation mismatch
Individual words are correct but combine unnaturally.
Metaphoric overreach
The image exceeds the physical or emotional scale of the scene.
Explanatory redundancy
The writer tells the reader what existing evidence already makes clear.
Viewpoint leakage
The narrator knows thoughts or facts they cannot plausibly access.
Sentence overloading
Too many independent relationships are forced into one sentence.
Paragraph stasis
Many sentences add texture but no change.
Template intrusion
A memorised phrase or structure overrides the specific scene.
Adult-edit dependence
The final prose improves while the student’s independent decision system does not.
The first-wrong-craft method
When a paragraph feels wrong, find the earliest craft decision that distorts the scene.
Example:
- Character is mildly embarrassed.
- Writer chooses “devastated”: first intensity distortion.
- Later metaphors, trembling and dramatic reflection are built to support that inflated word.
Repair the calibration at Step 2 rather than polishing the later overstatement.
The high-achiever stop rule
Stop adding craft techniques when:
- the scene is already clear;
- the new word is less natural than the old one;
- the metaphor adds no perception;
- the sentence is longer only to sound complex;
- the child cannot explain the revision;
- the tutor or AI is generating more language than the student;
- the revision is erasing voice;
- time spent polishing is preventing story development or completion.
Practice library: lexical precision
Practice 1: verb microscope
Replace one generic verb with three possible precise verbs. Describe the physical difference among them.
Practice 2: intensity ladder
Order near-synonyms by intensity, then create situations that genuinely earn each level.
Practice 3: collocation sort
Choose which noun–verb or adjective–noun pairs sound natural and explain why through authentic examples.
Practice 4: common-word defence
Give the student permission to keep a common word and require a reason why it is the most accurate choice.
Practice 5: specificity budget
Allow only three highly specific details in a paragraph. Choose the details that change interpretation most.
Practice library: intensity and emotion
Practice 6: event scale
Rank several events by likely emotional cost before choosing language.
Practice 7: behaviour before label
Write three observable behaviours, then infer possible emotion.
Practice 8: emotion deletion
Remove the emotion word and see whether the scene still communicates it.
Practice 9: understate deliberately
Rewrite a melodramatic paragraph using quieter language without making the character emotionless.
Practice library: figurative restraint
Practice 10: metaphor deletion
Delete the image. Identify what useful meaning is lost.
Practice 11: literal first
Describe the physical scene accurately before adding any figurative transformation.
Practice 12: scale match
Compare a small event with an enormous metaphor and repair the mismatch.
Practice 13: one-image paragraph
Use at most one figurative image and make it carry a clear perceptual job.
Practice library: sentence logic
Practice 14: relationship before connector
Name cause, contrast, condition, time or result before combining clauses.
Practice 15: short-sentence landing
Build an accumulating sentence and then use a short sentence for discovery.
Practice 16: de-load the sentence
Split an overloaded sentence, then recombine only clauses with a real dependency.
Practice 17: emphasis swap
Move a causal or time clause from front to back and discuss reader focus.
Practice library: viewpoint and inference
Practice 18: know/see/infer
For each statement, classify whether the narrator knows, observes or infers it.
Practice 19: mind-reading repair
Replace another character’s private thought with observable evidence.
Practice 20: explanation deletion
Remove one explanatory sentence after dialogue/action and test whether inference survives.
Practice 21: productive uncertainty
Write a moment where the narrator is unsure, but the reader has enough evidence to understand why.
Practice library: paragraph movement
Practice 22: paragraph job label
Write one word above each paragraph: orient, pressure, reveal, decide, consequence.
Practice 23: first-to-last change
State what changes across the paragraph. If nothing changes, revise the paragraph job.
Practice 24: scene-summary switch
Expand one turning point and compress one routine transition.
Practice 25: pressure curve
Draw a simple line showing story pressure across paragraphs. Check whether the prose intensity matches.
Practice library: dialogue
Practice 26: remove the tag
Use action or wording to carry tone where possible.
Practice 27: dialogue job
Label each line as conflict, decision, avoidance, revelation or relationship.
Practice 28: no-repeat dialogue
Delete narration that merely repeats what the dialogue already states.
Practice library: revision
Practice 29: eight-pass revision
Use meaning, intensity, viewpoint, physicality, sentence logic, inference, voice and economy as separate passes.
Practice 30: one mechanism only
Revise a paragraph for one recurring weakness rather than “improve everything”.
Practice 31: reverse the edit
Compare the student draft and edited version. Ask whether the student can explain every changed decision.
Practice 32: new-task transfer
After correction, write a new paragraph that requires the same principle without copying wording.
Twenty-five myths worth retiring
Myth 1: Better writers use harder words.
Better writers use more accurate words.
Myth 2: A composition needs many idioms to score well.
Memorised idioms can become unnatural or irrelevant. Meaning and control matter more.
Myth 3: Figurative language automatically improves description.
It helps only when it sharpens perception.
Myth 4: Every paragraph needs sensory detail.
Use the sensory channel that serves the moment.
Myth 5: “Show, don’t tell” is universal.
Both showing and concise telling have jobs.
Myth 6: Longer sentences are more mature.
Sentence length should follow information structure.
Myth 7: Short sentences are childish.
They can create precision and force.
Myth 8: Strong emotions need dramatic vocabulary.
Behaviour and context often communicate intensity more convincingly.
Myth 9: Repeating a noun is always bad style.
Strategic repetition can improve cohesion.
Myth 10: Every sentence should start differently.
Artificial variation can damage cohesion and parallelism.
Myth 11: A good opening must be dramatic.
It must create orientation, instability or a meaningful question.
Myth 12: A good ending must state a lesson.
Consequences and changed behaviour can carry theme more effectively.
Myth 13: Original writing requires an unusual plot.
Specific observation and viewpoint can make an ordinary situation original.
Myth 14: Editing means replacing simple words.
Revision begins with meaning and structure, not vocabulary decoration.
Myth 15: A polished paragraph is proof of improved writing.
Transfer to a new task is stronger evidence.
Myth 16: Adult-sounding prose is advanced prose.
Voice mismatch often reduces credibility.
Myth 17: Strong students need less explicit teaching.
They often need more precise discrimination among near-correct choices.
Myth 18: Model compositions should be copied for style.
They are better used to analyse decisions, not import phrases.
Myth 19: Thesaurus use automatically develops vocabulary.
Without collocation and register knowledge it can create errors.
Myth 20: More descriptive detail always creates immersion.
Irrelevant detail can stall the scene.
Myth 21: AI-generated alternatives teach writing by themselves.
Learning requires comparison, explanation and later independent transfer.
Myth 22: AI should never be used in student writing.
Bounded contrastive use can support judgement when authorship remains with the student.
Myth 23: Examination writing rewards maximum craft density.
Craft must remain controllable under time and support a complete coherent response.
Myth 24: Strong writers should start Secondary techniques as early as possible.
Depth of control at current level is often the stronger preparation.
Myth 25: Simpler language means simpler thinking.
Simple language can carry very complex, precise thought.
Parent case study 1: the thesaurus achiever
The child replaces ordinary verbs with rare alternatives and writing becomes unnatural.
Intervention: verb microscope, collocation and common-word defence.
Parent case study 2: the melodramatic writer
Every setback becomes devastation; every surprise makes the heart stop.
Intervention: event-significance ladder and behavioural evidence.
Parent case study 3: the metaphor collector
Every paragraph contains imagery regardless of need.
Intervention: metaphor deletion and one-image budget.
Parent case study 4: the model-answer voice
Composition suddenly contains formal phrases never used elsewhere.
Intervention: narrator-fit read-aloud and rewrite in owned register.
Parent case study 5: the beautiful static paragraph
Description is strong, but nothing changes for a page.
Intervention: paragraph job and first-to-last change.
Parent case study 6: the explanation-heavy writer
After every action, the narrator explains the emotion and lesson.
Intervention: deletion test after strong evidence.
Parent case study 7: the sentence-variety performer
Connectors and sentence openings are varied mechanically.
Intervention: relationship-first syntax.
Parent case study 8: the adult-edited child
Homework is polished, timed writing is much weaker.
Intervention: reduce editing intervention, require explanation and new-task transfer.
Parent case study 9: the AI-polished writer
An AI rewrite is excellent, but the child cannot reproduce the choices independently.
Intervention: use AI only for contrastive options, then remove tool and write new material.
Parent case study 10: the strong writer who cannot finish
Sentence-level craft is excellent but time is spent polishing early paragraphs.
Intervention: story architecture and completion floor before micro-polish.
A parent dashboard for high-achiever writing
- Are word choices accurate rather than merely difficult?
- Does intensity match event significance?
- Are collocations natural?
- Does register fit the narrator?
- Does viewpoint remain consistent?
- Does imagery clarify perception?
- Does sentence structure reflect logical relationships?
- Does each paragraph move the story?
- Are important moments scened and routine moments compressed?
- Is explanation deleted when inference is enough?
- Does the final voice still belong to the student?
- Can repaired craft transfer into new writing?
Parent operating manual: how to coach without ghost-writing
Parents often see a weak sentence and know a better one immediately. The risk is that the adult solves the craft problem before the child has had to discriminate among alternatives.
Use questions before replacements.
Instead of “Use a stronger word”
Ask: “What exact action happened? Which verb shows that action without adding an emotion the scene has not earned?”
Instead of “Make this more descriptive”
Ask: “What detail would change what the reader understands?”
Instead of “Add a simile”
Ask: “What is hard to perceive literally here? Would an image make it clearer?”
Instead of “This sounds childish”
Ask: “Is the meaning inaccurate, the register wrong, or just simple?”
Instead of rewriting a paragraph
Offer two contrasting alternatives and ask the student to explain what each changes, then write a third version independently.
Home-practice architecture: 20 minutes
- 5 min near-synonym or collocation comparison;
- 5 min one sentence-logic task;
- 5 min deletion/edit test;
- 5 min independent rewrite.
Home-practice architecture: 40 minutes
- 10 min craft discrimination;
- 15 min one paragraph draft;
- 10 min one diagnostic revision pass;
- 5 min explanation of choices.
Home-practice architecture: 60 minutes
- 10 min architecture/scene planning;
- 25 min drafting;
- 15 min targeted revision;
- 10 min independent transfer or read-aloud audit.
Fifty parent questions that reveal craft judgement
- What exact meaning do you want?
- What does this verb imply?
- Is that implication true in the scene?
- Is there a more ordinary but more accurate word?
- What intensity has the event earned?
- What behaviour would show that intensity?
- Does this word collocate naturally here?
- Would this narrator really say this?
- Does the register fit the scene?
- What can the narrator actually know?
- What is observation and what is inference?
- Does the physical action make sense?
- What does this metaphor help the reader perceive?
- What disappears if the metaphor is deleted?
- Is the image too large for the event?
- Which sensory detail matters most?
- What relationship joins these clauses?
- Would two sentences be clearer?
- What information should come first?
- What should land at the end of the sentence?
- What is this paragraph’s job?
- What changes from first sentence to last?
- Is this moment important enough to scene?
- Can this transition be summarised?
- Is dialogue doing a job?
- Does narration repeat the dialogue?
- Can behaviour carry the emotion?
- What can the reader infer already?
- What explanation can you delete?
- Is this ambiguity productive or confusing?
- What key noun should repeat for cohesion?
- Are synonyms making reference less clear?
- What pressure exists at this point in the story?
- Does the prose intensity match that pressure?
- Has the opening spent too much intensity?
- Does the ending show change or merely state a moral?
- What theme emerges from action?
- Which phrase sounds imported?
- Can you explain why you chose every revised word?
- Would you make this decision without adult help?
- Can you apply the same principle in a new paragraph?
- What is the first craft decision that distorted the scene?
- What would happen if this sentence were simpler?
- Can you defend the simpler version?
- What is the minimum detail needed?
- What is the minimum explanation needed?
- What did the revision improve: meaning, rhythm or display?
- Does the piece still sound like the same writer?
- Can you finish under time before polishing?
- What should you leave alone because it already works?
AI-era authorship: a bounded workflow
AI can produce fluent adult prose instantly. That makes ownership more important, not less.
A useful bounded workflow is:
- Student drafts first.
- Student identifies one craft problem. Example: intensity, collocation, sentence logic.
- AI may provide two or three contrasting alternatives.
- Student compares them. What changes in meaning, register, rhythm and viewpoint?
- Student rejects or adapts.
- Tool is removed.
- Student writes a new sentence or paragraph independently.
What AI should not own
- the story decision;
- the narrator’s voice;
- the emotional scale;
- the final wording of every sentence;
- the interpretation of the prompt;
- the student’s reflection on why a choice works.
The authorship test
After AI or adult feedback, ask whether the learner can:
- explain the change;
- recreate the principle without copying;
- produce comparable quality on a new topic;
- reject a polished alternative that does not fit the intended voice.
Progression by Primary level
P3–P4: natural accuracy before display
Focus on clear event order, specific verbs, reference, sentence completeness, paragraph jobs, concrete observation and age-appropriate voice.
P5: calibration and control
Develop collocation, register, sentence logic, selective imagery, inference and stronger scene–summary decisions.
P6: craft under time
Maintain precision while planning, drafting, completing and revising under examination constraints. Sophistication that prevents completion is not functioning well enough.
The examination completion floor
Before polishing, the writer should have:
- a coherent response to the prompt;
- a complete narrative arc;
- clear paragraph movement;
- controlled viewpoint;
- an ending;
- enough time for high-value correction.
Sentence polish comes after the piece is structurally alive.
The advanced depth gate
Before pushing a strong Primary writer toward older-level techniques, ask whether the learner can:
- defend precise ordinary words;
- calibrate intensity;
- use collocations naturally;
- maintain register;
- control viewpoint;
- use imagery selectively;
- choose sentence structure by logic;
- create paragraph movement;
- use scene and summary deliberately;
- trust inference;
- revise without losing voice;
- transfer craft decisions into new writing under time.
Thirty-day high-achiever writing plan
Days 1–3: baseline
Take one recent composition. Mark only recurring mechanisms: lexical precision, intensity, imagery, sentence logic, paragraph movement, inference and voice.
Days 4–7: repair one high-value mechanism
Do not rewrite the whole composition. Use short contrastive exercises.
Week 2: new-paragraph transfer
Write fresh material using the repaired principle without copying model phrases.
Week 3: combine two craft systems
For example, intensity + viewpoint, or sentence logic + paragraph movement.
Week 4: timed independent piece
Use the same revision language after the tool, tutor or parent has stepped away.
The transfer test
A writing improvement is durable when it survives:
- new topics;
- new characters;
- different emotional scales;
- different paragraph jobs;
- timed drafting;
- reduced adult prompting;
- several days of delay.
Twelve signs deeper repair is needed
- rare vocabulary regularly sounds unnatural;
- emotion is over-intense;
- metaphors dominate the scene;
- narrator register shifts suddenly;
- viewpoint leaks into other minds;
- physical action is unclear;
- sentences are complex without clear logic;
- paragraphs are static;
- dialogue repeats narration;
- explanation duplicates evidence;
- adult/AI wording dominates revision;
- timed independent writing is far weaker than edited work.
Twelve signs deeper extension is reasonable
- word choice is accurate and natural;
- intensity is calibrated;
- register remains consistent;
- imagery is selective;
- sentence structure carries logic;
- paragraphs move;
- scene/summary balance is deliberate;
- inference is trusted;
- viewpoint remains controlled;
- student can diagnose own excess;
- revision preserves voice;
- craft transfers under time.
Expanded frequently asked questions
1. Should strong writers learn more vocabulary?
Yes, but vocabulary growth should include collocation, register, intensity and usage—not just definitions.
2. Are “powerful words” useful?
Only when they are the most accurate words for the scene.
3. Should children use a thesaurus?
It can generate candidates, but every candidate needs meaning, collocation and register checking.
4. How many similes or metaphors should a composition contain?
There is no useful quota. Use only those that earn their space.
5. Is simple language bad?
No. Simple language can be exact, powerful and mature.
6. How do I improve emotional writing?
Calibrate the event, use specific behaviour, and let intensity build from context.
7. How do I protect my child’s voice?
Use questions, contrastive examples and principles rather than replacing whole sentences.
8. Should parents correct every grammar error?
No. Prioritise recurring errors and high-value craft mechanisms.
9. What if my child loves big words?
Keep the curiosity, but require the precision ladder before deployment.
10. How do I reduce overwriting?
Use deletion tests, paragraph jobs and a one-new-information rule.
11. How do I teach sentence variety?
Start with the relationship among ideas—cause, contrast, time, condition, result—then choose syntax.
12. Should every paragraph have a metaphor or idiom?
No. Paragraphs need jobs, not decoration quotas.
13. How do I improve story pacing?
Give more space to meaningful choices and turning points; compress routine transitions.
14. How do I improve dialogue?
Give each line a job and avoid repeating the same information in narration.
15. How do I improve endings?
Show consequence, changed behaviour or reinterpreted meaning rather than bolting on a moral.
16. How do I know whether a metaphor is good?
Delete it and identify what useful perception disappears.
17. What if the student’s natural voice is simple?
Develop precision and range without forcing an adult register.
18. Is model-composition reading useful?
Yes when the student analyses decisions rather than memorising phrases.
19. Can AI help creative writing?
It can support comparison and feedback if the student drafts, evaluates and then transfers independently.
20. Should AI rewrite my child’s composition?
A full rewrite may produce polished prose without improving the child’s decision system. Contrastive, bounded use is more educational.
21. How do I know AI use is becoming dependency?
Independent writing does not improve, the child cannot explain revisions, or the final voice changes dramatically when the tool is used.
22. Should strong P6 writers start Secondary writing techniques?
Only when current-level control is robust. Depth in viewpoint, inference and sentence logic can provide substantial challenge.
23. How do I know revision is working?
The same error recurs less often and the repaired principle appears in new writing.
24. What if a child writes beautifully but cannot finish exams?
Protect the completion floor: plan, draft the full arc, then polish selectively.
25. Should every sentence be revised?
No. Leave strong sentences alone. Revision should be diagnostic, not compulsive.
26. What is the strongest sign of mature writing?
The writer can choose the degree of language needed—and stop when more would reduce fidelity.
A parent decision tree for high-achiever writing
- Is the meaning accurate? If no → semantic repair.
- Is the language natural? If no → collocation/register repair.
- Is the intensity earned? If no → calibration repair.
- Does the image clarify? If no → delete or replace.
- Does viewpoint permit this knowledge? If no → observation/inference repair.
- Does the sentence structure match the relation? If no → clause architecture repair.
- Does the paragraph change something? If no → paragraph-job repair.
- Is detail proportional to significance? If no → scene/summary repair.
- Can the reader infer without this explanation? If yes → consider deletion.
- Does the sentence sound like the student? If no → voice repair.
- Can the student explain the revision? If no → feedback may be replacing learning.
- Can the principle transfer to new writing? If yes → craft is becoming independent.
Current learning reference
Families who want the broader curriculum context can consult MOE’s English Language Syllabus 2020 Primary. This page deliberately focuses on craft judgement rather than reproducing a syllabus checklist.
Final synthesis: precision is the courage to stop
High-achiever writing improves when the student stops asking, “How can I make this sound more advanced?” and asks instead, “What does the reader need to understand, feel or infer at this exact moment?”
That shift changes everything. Vocabulary becomes a tool for exact meaning. Figurative language becomes a tool for perception. Sentence variety becomes a tool for logic. Detail becomes a tool for significance. Revision becomes a tool for removing distortion. Voice becomes something to protect rather than polish away.
Mature writing is not a performance of difficulty. It is fidelity under tighter control.
Historical 2017 photo archive
The photographs below were part of the original 2017 Bukit Timah page and are retained for historical continuity. They are archive material and are not evidence of the current teaching format or current programme availability.
Longform transfer appendix: how high-achiever writing becomes independently controllable
The strongest evidence of writing development is not that a revised composition looks polished. It is that the student can make comparable decisions again when the topic, characters, emotional scale and time conditions change. This appendix therefore focuses on transfer: moving from a beautifully corrected page to a writer who can diagnose excess, calibrate language and preserve voice without waiting for an adult, model composition or AI system to supply the final phrasing.
The four receipts of high-achiever writing mastery
- Explanation: the student can explain why a word, sentence or paragraph choice works.
- Delay: the principle remains available several days later without the original correction in view.
- Variation: the same judgement survives a different scene, genre or emotional intensity.
- Transfer: the student notices and repairs the issue independently in new writing.
A polished draft with no transfer is editing success. A new draft showing the same principle independently is writing development.
The high-achiever near-miss laboratory
Strong writers benefit from comparing choices that are almost correct. The task is not merely to identify “good” and “bad” sentences, but to discriminate among plausible alternatives.
Near-miss 1: technically correct, wrong intensity
Devastated, I stared at the stain on my worksheet.
The vocabulary is grammatical. The likely problem is scale. Unless the worksheet carries unusual significance, devastated may exceed the event.
Near-miss 2: precise meaning, wrong register
I regarded my classmate with profound consternation.
The words can be correct, but the register may not fit a young first-person narrator.
Near-miss 3: vivid metaphor, weak physical fit
The whisper sliced through the room like a bulldozer.
The simile is vivid but the physical qualities conflict: a whisper and bulldozer do not share the relevant perceptual dimension clearly.
Near-miss 4: varied sentence opening, no reason
With a heart pounding wildly, I opened the door. Slowly and cautiously, I stepped inside. With trembling hands, I reached for the switch.
The openings differ superficially but repeat the same participial-fronted pattern and emotional information.
Near-miss 5: good description, poor narrative allocation
A paragraph may beautifully describe every object on a desk even though the story’s real turning point is a missing envelope. The problem is not sentence quality; it is significance allocation.
Near-miss 6: strong emotional evidence plus redundant explanation
I folded the letter twice, then twice again until the paper would not bend. I was extremely angry.
The second sentence may be unnecessary if anger is the intended inference and the action already carries it.
Near-miss 7: sophisticated connector, invented relation
Although I had forgotten my umbrella, the bus arrived.
The clauses are both true but the concession relation may be meaningless unless the forgotten umbrella created an expectation about the bus.
Near-miss 8: beautiful phrase, wrong viewpoint
Across the room, Mrs Lim silently regretted doubting me.
A first-person narrator cannot know this private regret unless Mrs Lim reveals it.
Near-miss 9: natural phrase, wrong story pressure
A powerful short sentence such as Everything changed. can be weak when nothing immediately changes in the plot, relationship or understanding.
Near-miss 10: strong ending, imported moral
From that day onwards, I learnt that perseverance is the key to success.
If the story actually concerned honesty or friendship, the line may be generic rather than earned.
Near-miss 11: advanced vocabulary with vague referent
The situation was utterly unprecedented.
Ask: what situation, and unprecedented in what way? Abstraction can sound sophisticated while remaining imprecise.
Near-miss 12: sensory detail with no viewpoint relevance
A narrator running from danger probably does not catalogue six colours of flowers unless those details matter to orientation or contrast.
Near-miss 13: clever wordplay that breaks tone
A joke can be individually witty but still weaken a tense or reflective scene if the narrator would not plausibly make it then.
Near-miss 14: paragraph-level repetition disguised by synonyms
Three sentences can repeat the same idea with different vocabulary. The problem is not lexical repetition but informational repetition.
Near-miss 15: adult-perfect sentence that the student cannot reproduce
The prose may be excellent, but if the student cannot explain or transfer the decision, the final product is stronger than the writer.
Thirty troubleshooting scenarios for strong Primary writers
1. The student insists the rarest synonym is best.
Run the precision ladder: correct → natural → precise → necessary → coherent.
2. Every emotion is described as extreme.
Create an event-intensity ladder and require behavioural evidence before choosing the word.
3. Every composition begins with weather.
Replace opening-template practice with opening-job practice: orient, destabilise, introduce relationship or create a meaningful question.
4. Every composition begins with dialogue.
Ask whether the spoken line changes the reader’s model immediately. If not, begin closer to the actual instability.
5. The student uses many idioms.
Check literal fit, register and whether the idiom adds meaning unavailable in direct language.
6. The student uses many similes beginning with “as…as”.
Reduce the quota. Ask what perceptual relationship genuinely needs comparison.
7. Strong vocabulary appears only in memorised phrases.
Change the context and require the student to generate a new collocation independently.
8. The student writes beautifully but slowly.
Separate drafting from polishing. Protect a completion floor before micro-editing.
9. The student plans for too long.
Use a minimal plan: conflict, turning point, decision, consequence. Draft before polishing the plan itself.
10. The story has many events but little emotional movement.
Track how the character’s understanding, goal or relationship changes—not only what happens physically.
11. The story has much reflection but little action.
Require each reflective paragraph to connect to a concrete decision, event or consequence.
12. Dialogue dominates the page.
Label every line by job. Remove lines that repeat known information or do not alter relationship, decision or knowledge.
13. Dialogue sounds like adults speaking in essays.
Read aloud. Shorten, simplify and restore the character’s natural register.
14. The student narrates every movement.
Compress routine actions and scene only movements that affect tension, orientation or meaning.
15. Every paragraph ends with an emotion sentence.
Let action, image, dialogue or consequence close some paragraphs instead.
16. The student repeats “I felt…” frequently.
Use behavioural evidence, physical perception or thought selectively. Do not ban the phrase; diversify the evidence system.
17. The student overuses adverbs.
Ask whether the verb can carry the manner more precisely or whether the adverb adds necessary calibration.
18. The student writes complex sentences with weak punctuation.
Return to clause relationships and boundaries before adding more complexity.
19. The student uses fragments for drama repeatedly.
Require complete-sentence control first and ask what specific rhythmic effect each fragment creates.
20. The student uses rhetorical questions in every reflection.
Ask whether the question opens real uncertainty or simply announces emotion.
21. The student writes generic morals.
Replace moral statement with changed action, consequence or reinterpreted object.
22. The student rewrites every sentence after feedback.
Protect strong sentences. Revision should be selective and mechanism-driven.
23. The student copies a model phrase accurately.
Ask for the principle, then use a new topic where the original phrase cannot fit.
24. Parent editing makes the writing dramatically better.
Reduce direct rewriting. Use questions and two contrasting alternatives instead.
25. Tutor feedback is extensive but the same issues recur.
Reduce the number of simultaneous targets. Track one or two recurrent mechanisms across several pieces.
26. AI suggestions become the default first draft.
Restore student-first drafting. AI may enter only after a specific craft problem is named.
27. The student rejects ordinary language as “too basic”.
Use published-quality examples where common words achieve greater precision than rare substitutes.
28. The student’s writing is accurate but emotionally flat.
Add significance through viewpoint, consequence, choice and specific behaviour rather than simply stronger adjectives.
29. The student’s writing is vivid but confusing.
Prioritise spatial orientation, reference and event sequence before imagery.
30. The student is strong but anxious about every sentence.
Separate drafting mode from editing mode. Fluency requires permission to write imperfectly before evaluating.
High-achiever writing glossary for parents
| Term | Parent-friendly meaning |
|---|---|
| Collocation | Words that naturally occur together in standard usage. |
| Register | The level and social character of language appropriate to a speaker, situation or genre. |
| Viewpoint | The position from which the reader receives information and perception. |
| Narrative distance | How close the narration feels to a character’s immediate thoughts and perceptions. |
| Scene | A section that slows down and represents events in detail. |
| Summary | Compressed narration covering less significant time or repeated action. |
| Inference | Meaning the reader derives from evidence rather than direct statement. |
| Cohesion | The linguistic connections that make sentences and paragraphs feel linked. |
| Information structure | How a sentence orders old/new or background/focus information. |
| Calibration | Matching language intensity to the actual scale of the event or emotion. |
| Overwriting | Using more language, detail or intensity than the meaning requires. |
| Voice | The recognisable pattern of wording, viewpoint, rhythm and judgement belonging to the writer or narrator. |
Genre transfer: narrative writing
Narrative precision depends on:
- event order;
- viewpoint;
- scene/summary allocation;
- dialogue purpose;
- emotional calibration;
- turning-point space;
- ending consequence.
The central craft question is: what changes in the character’s situation, knowledge, goal or relationship?
Genre transfer: situational/functional writing
High-level control here looks different. Precision means:
- complete task fulfilment;
- accurate details;
- appropriate tone;
- clear reader action;
- no decorative complexity that obscures information.
A strong creative writer must still know when plain clarity is the advanced choice.
Genre transfer: reflective/expository responses
Where school work asks for explanation or reflection, precision includes:
- defining claims clearly;
- using relevant examples;
- qualifying broad statements;
- separating evidence from assertion;
- maintaining an appropriate formal register.
Craft under examination time
A high-achiever system must work under finite time. The writer needs a hierarchy of value.
First: answer the task
Prompt relevance and complete communication come before stylistic density.
Second: preserve structure
Complete the arc, paragraph logic and ending.
Third: protect high-risk language
Correct unclear reference, tense, sentence boundaries and meaning distortions.
Fourth: polish selectively
Improve a few high-value verbs, images or sentence rhythms where time allows.
This hierarchy prevents a beautiful opening from consuming the time needed to finish the piece.
The 70/20/10 revision allocation
This is not a rigid formula but a useful thinking model:
- 70% of revision attention on meaning, structure, clarity and completion;
- 20% on recurring language control such as reference, tense, sentence boundaries and collocation;
- 10% on optional stylistic polish.
The percentages can change by learner, but the principle remains: polish should not outrank architecture.
The independence ladder
Level 1: adult identifies problem and supplies correction
Useful for modelling, but highly supported.
Level 2: adult identifies problem; student generates alternatives
More judgement moves to the student.
Level 3: adult names category only
Example: “Check intensity in this paragraph.”
Level 4: student uses a revision checklist independently
Adult reviews after self-edit.
Level 5: student notices the problem during drafting
The craft principle has begun to internalise.
Level 6: student prevents the problem before it appears
The judgement has become part of fluent writing.
The portfolio test
Instead of judging one composition, compare three pieces across several weeks.
Look for:
- fewer repeated excesses;
- more owned phrasing;
- better completion under time;
- more selective imagery;
- stronger paragraph movement;
- less adult dependence;
- transfer of repaired mechanisms.
Long-term success criteria
A mature young writer can:
- choose accurate words without chasing rarity;
- calibrate emotional scale;
- use figurative language selectively;
- maintain a believable voice;
- control viewpoint;
- choose sentence structure by meaning;
- give important moments enough space;
- compress routine material;
- trust readers to infer where evidence is sufficient;
- revise diagnostically rather than compulsively;
- explain major craft choices;
- produce comparable quality independently on a new task.
Closing principle: the writer is stronger when the prose needs less proof of strength
Young high achievers often begin by showing everything they can do. Mature control arrives when they no longer need every advanced word, image, sentence pattern or reflection to appear on the page.
The craft becomes quieter because the judgement becomes stronger. The writer knows when to intensify, when to simplify, when to explain, when to trust the reader, when to polish and when to leave a good sentence alone.
That is the endpoint this page is designed to protect: not a child whose composition looks professionally rewritten, but a child who can make increasingly professional decisions while the writing still unmistakably belongs to them.