Quick answer: narrative logic improves when the writer treats knowledge as a changing state. A character can act only on what they know, believe, suspect or misunderstand at that point in the story. If information arrives later, the character’s decisions should change with it.
This historical Bukit Batok Creative Writing page owns one specific craft problem: character information-state control. It is not a claim of a current eduKateSG Bukit Batok centre. Its job is to help students track who knows what, when they learn it, what they only assume, what remains hidden, and how new evidence changes action, suspense and plausibility.
The one-sentence answer
At every important turn, separate knowledge from belief, keep hidden information genuinely hidden, reveal new evidence through a believable route, and make later decisions respect the updated state.
Four information states
- Known: supported directly by evidence available to the character.
- Believed: the character accepts it but may be wrong.
- Suspected: plausible but not established.
- Unknown: not yet available to the character.
Many plot holes begin because the writer silently upgrades belief into knowledge or gives a character access to author-only information.
Writer knowledge, reader knowledge and character knowledge are different
The writer knows the whole story. The reader may know more or less than the viewpoint character. The character knows only what the narrative has allowed them to observe, infer or receive.
Keeping these layers separate produces suspense, misunderstanding and dramatic irony without cheating.
Arrangement 1: reader and character learn together
This is the cleanest arrangement for discovery. The missing object is noticed at the same moment by narrator and reader. Both now face the same uncertainty.
Arrangement 2: reader knows more
The reader may see that a bag was left behind while the character walks away. Tension comes from waiting for the character to catch up with what the reader already knows.
The character’s actions must still remain reasonable from their own information state.
Arrangement 3: character knows more
This can create mystery, but first-person narrators should not hide facts they are actively thinking about merely to manipulate the reader. Delayed disclosure needs a believable narrative reason.
The knowledge ledger
For a complicated story, students can use a temporary ledger:
| Stage | Character A | Character B | Reader |
|---|---|---|---|
| 1 | knows key is missing | does not know | knows key is missing |
| 2 | suspects friend | knows friend borrowed it | sees both possibilities |
| 3 | sees evidence | explains | understands misunderstanding |
The scaffold should fade once the student can maintain continuity mentally.
Belief updates should have causes
A character should not suddenly stop suspecting someone because the writer needs the story to end. New evidence, testimony, contradiction or memory should alter the belief.
A useful pattern is: belief → evidence → conflict → revised belief → new action.
Misunderstanding is an information structure
- incomplete evidence;
- reasonable but wrong interpretation;
- action based on that interpretation;
- new evidence;
- belief revision;
- relationship or goal consequence.
This is stronger than characters misunderstanding one another merely because the plot requires conflict.
Secrets need reasons and leak paths
A secret should have a reason for remaining hidden and a plausible route by which it might be exposed. Fear, embarrassment, loyalty or uncertainty can motivate silence. Evidence, another witness, an object or a contradiction can reveal the truth later.
The longer the secret persists, the more pressure should accumulate.
Worked example: broken trophy
Three students are near a display cabinet when a trophy falls. One sees the cause. One hears the crash but not what happened. A teacher arrives afterwards.
The story becomes structurally interesting because knowledge is uneven. Later accusation, defence and apology must follow the actual evidence each person possesses.
Worked example: the half-read message
A student sees only part of a notification and assumes friends are excluding them. Their behaviour becomes colder. Later, the full message reveals a surprise plan.
The misunderstanding works because the original inference was reasonable but incomplete. The later reveal reinterprets earlier evidence.
Worked example: object evidence
A torn label is found near a missing bag. One character treats it as proof; another knows the label came off earlier. The object does not “solve” the plot by itself. Its meaning depends on who knows its history.
The impossible-knowledge audit
Underline every major decision. Ask what information justified it. If the necessary fact was never observed, inferred or communicated, the decision contains an information leak.
The belief-evidence audit
For each strong belief, identify the evidence available to the character. If the evidence is weak, the character may still believe it—but the story should show why the person is prone to that interpretation.
The reveal fairness test
After the truth appears, can a careful reader see how earlier clues fit? If the reveal depends on brand-new information, plant a fair route earlier.
Pronouns are part of information control
“She told her that she had seen it” may be grammatical but unusable if several female characters are present. Unclear reference damages who-knows-what tracking. Names and specific noun phrases are sometimes necessary for narrative logic.
P3–P4: track who saw what
Middle-primary students can begin with simple knowledge differences and misunderstandings based on partial evidence.
P5–P6: layer belief, secret and delayed disclosure
Upper-primary writers can manage several information states provided the logic remains readable. Current examination details should always be checked against official MOE and SEAB materials for the relevant cohort.
AI can repair continuity invisibly
AI can fill missing information links and make a plot appear coherent. That can hide the student’s real tracking weakness. A bounded use is: student writes a knowledge ledger → AI may challenge one decision for evidence → student repairs → tool removed → new draft tracked independently.
What progress should look like
- Impossible knowledge errors reduce.
- Belief changes follow visible evidence.
- Misunderstandings become plausible rather than convenient.
- Secrets have clearer reasons and reveal paths.
- Twists become fairer in hindsight.
- Pronoun reference is easier to follow.
- Students can explain what each character knows at major turns.
- Unseen prompts produce new information structures without plot leakage.
Questions parents and teachers can ask
- What does this character know?
- What do they only believe?
- What evidence created that belief?
- Who knows the secret?
- How can the information realistically travel?
- What new evidence changes the next decision?
- Could the reader reconstruct the reveal fairly?
Current eduKateSG class information
Selected Primary English and Creative Writing students are taught in focused 3-pax small groups, typically 1.5 hours. Current centres listed on eduKateSG are 83 Punggol Central, Singapore 828761 and 8 Fourth Avenue, Singapore 268674. This historical Bukit Batok URL is not a claim of a current Bukit Batok centre.
2017 photo archive
The photographs below were part of the original 2017 Bukit Batok page. They remain as historical archive material and do not establish a current Bukit Batok centre or current class configuration.
The deeper principle
Story logic depends on information boundaries. Characters should know only what the story has given them, believe only what the evidence can plausibly support, and change their minds when new evidence arrives. When knowledge states remain honest, suspense, misunderstanding and surprise can become complex without becoming arbitrary.