English tuition for Bedok families becomes more useful when a student can look across several school papers and see patterns rather than a pile of unrelated red marks. “Wrong again” is not a diagnosis. A repeated pronoun error, unsupported inference, vague answer, paragraph-development problem or tense inconsistency is a recurring system signal.
This rebuilt legacy page now owns one distinct job inside the eduKateSG English estate: building a portable error taxonomy across school papers. It does not claim a current Bedok centre and it does not duplicate generic diagnosis pages. Its purpose is to show how parents and students can turn several months of marked work into a smaller map of recurring English error families that can actually be repaired.
eduKateSG currently teaches selected English levels in focused 3-pax small groups, typically 1.5 hours, at 83 Punggol Central, Singapore 828761 and 8 Fourth Avenue, Singapore 268674. Families from Bedok can enquire about suitable placement according to level, timetable and travel fit.
The Direct Answer
A useful error taxonomy groups mistakes by why they happened, not merely by which paper section they appeared in.
- Input error: text or question was not read accurately.
- Interpretation error: meaning or task demand was misunderstood.
- Language-resource error: vocabulary/grammar knowledge was missing or unstable.
- Organisation error: ideas were present but not structured.
- Expression error: understanding was correct but wording was vague or inaccurate.
- Execution error: method was known but not triggered under time.
Different paper section, same underlying error = one repair opportunity.
Why Paper-by-Paper Correction Can Hide Patterns
A student may lose one mark in comprehension, one in editing and one in writing. If each is corrected separately, they look unrelated.
But all three may share the same cause. For example, weak pronoun tracking can affect passage comprehension, sentence editing and the clarity of a composition. Weak paragraph organisation can affect essay development and open-ended explanations. Poor command reading can affect comprehension, Situational Writing and even oral follow-ups.
The taxonomy looks across components and asks what keeps recurring.
Error Family 1: Question-Command Errors
Symptoms include answering where instead of why, giving an example when asked for an explanation or missing a second requirement such as “and why”.
The repair is task interpretation: identify the output before searching the text or drafting.
Next-time cue: What exactly must exist at the end of my answer?
Error Family 2: Evidence-Locality Errors
The student understands the passage generally but searches in the wrong paragraph, copies too much or ignores a nearby decisive clue.
The repair is evidence search: identify the event/phrase in the question, start locally and widen only when the relationship crosses paragraphs.
Next-time cue: Where is the nearest sufficient evidence?
Error Family 3: Unsupported Inference
The student gives an intelligent but overly strong conclusion. Behaviour suggests nervousness, but the answer claims guilt. A writer’s criticism becomes “hatred”.
The repair is claim calibration: what is stated, what is implied and what does the evidence actually justify?
Next-time cue: What is the strongest claim the evidence allows?
Error Family 4: Reference Tracking
Pronouns, “this decision”, “such behaviour” and other references point to the wrong noun or earlier idea.
The repair is substitution: replace the pronoun/reference phrase with the candidate and check whether the sentence and paragraph remain coherent.
Next-time cue: What exactly does this word point back to?
Error Family 5: Grammar Knowledge Versus Grammar Transfer
Some students do not know the rule. Others know it in worksheets but miss it in long sentences or their own writing.
The taxonomy should distinguish these. Missing knowledge needs teaching. Transfer failure needs detection routines and changed-surface practice.
Next-time cue depends on the family: true subject?, timeline?, pronoun?, sentence boundary?
Error Family 6: Vocabulary Fit
The word meaning is roughly correct but the collocation, tone or intensity does not fit.
“Devastated” for a minor inconvenience, a formal word inside casual dialogue or a synonym that does not match the sentence pattern can all belong to the same family.
Next-time cue: Does the word fit meaning, phrase, intensity and tone?
Error Family 7: Paragraph Development
A paragraph contains relevant ideas but repeats instead of moving. Evidence appears without commentary. A narrative scene describes without changing the story.
The repair is paragraph function: topic → development → turn/condition → landing, or setup → pressure → decision → consequence depending on genre.
Next-time cue: What changes because this paragraph exists?
Error Family 8: Answer Expression
The student can explain the correct idea orally but writes a vague answer. Pronouns become unclear, a cause is missing or the sentence is too broad.
The repair is oral-to-written transfer and answer architecture.
Next-time cue: Did my sentence make the relationship explicit?
Error Family 9: Execution Under Time
The student can perform correctly when untimed or prompted, but the skill disappears inside full papers.
This belongs to execution rather than knowledge. We practise triggers, pacing, interleaving and targeted checking after the method is accurate.
Next-time cue: What condition should make me launch this method?
Build the Taxonomy From Three Papers, Not One
One paper may be unusual. Three or more papers reveal recurrence.
We can review school tests, homework, composition drafts and oral feedback. Each error is assigned to a family. We then count frequency, mark cost and persistence.
A useful priority score can be informal:
- high frequency + high mark cost: repair now;
- high frequency + low mark cost: build routine;
- low frequency + high mark cost: watch and prepare;
- low frequency + low mark cost: defer until larger issues stabilise.
This stops revision from treating every red mark as equally urgent.
A Simple Error Ledger
| Date/task | Error family | Cause | Next-time cue | Retest |
|---|---|---|---|---|
| Comprehension | Inference | Claim too strong | Evidence boundary | New passage |
| Editing | Agreement | Nearest noun trap | Find true subject | Changed sentence |
| Composition | Paragraph movement | No consequence | What changes? | New theme |
The ledger should remain small enough to use. It is a learning instrument, not paperwork for its own sake.
Patterns Can Cross Primary and Secondary English
The surface changes with age, but some error families persist. Weak reference tracking begins with simple pronouns and later affects dense comprehension. Weak evidence boundaries begin with character inference and later affect argumentative claims. Poor paragraph movement begins in composition and later affects discursive writing.
A portable taxonomy therefore gives continuity across school years.
The Taxonomy Should Evolve
As a student improves, old error families should shrink and new, more advanced ones may appear. A P4 learner may move from sentence accuracy to inference precision. A Secondary 3 learner may move from basic paragraph structure to argument qualification.
Progress is partly visible when the error profile changes upward rather than repeating the same foundation forever.
Do Not Turn the Child Into a List of Defects
An error taxonomy is a tool for teaching, not an identity label. We also track strengths: fluent reading, good vocabulary retrieval, strong oral reasoning, reliable checking, creative idea generation.
Strengths can support repairs. A student with strong oral explanation can use speaking to organise written answers. A strong reader can use wide reading to grow vocabulary and writing.
Why Three Students Helps
Three students may make different errors on the same question. This prevents the class from assuming one generic explanation fits everyone.
Over several lessons, the tutor can see each learner’s recurring taxonomy. One may have an inference pattern, another a grammar-transfer pattern, another a writing-development pattern. Shared tasks remain efficient while repairs stay individual.
A Typical 1.5-Hour Error-Taxonomy Lesson
- Retrieve: retest one previous error family.
- Attempt: new mixed English task.
- Classify: identify cause rather than section label.
- Repair: teach one high-value relationship.
- Cue: create a next-time trigger.
- Transfer: changed surface.
- Ledger: update only if the pattern is meaningful.
Current English Context
MOE’s Primary and Secondary English syllabuses treat English as an integrated set of language capabilities, not disconnected worksheet chapters. A portable error taxonomy supports that integrated view by locating recurring processes across reading, writing, grammar, oral and viewing. Official references: MOE English Language Syllabus 2020 Primary and MOE English Language Syllabus 2020 Secondary.
When Bedok Families May Consider This Support
- Many marked papers exist but priorities are unclear.
- The same mistake appears in different English components.
- Parents see “careless” comments without knowing what to repair.
- Tuition and school corrections feel disconnected.
- The child understands feedback but the pattern keeps returning.
- Exam revision feels like trying to fix everything at once.
- The family wants a smaller map of high-value English weaknesses.
What Progress Should Look Like
A stronger student can name recurring error families, use a small next-time cue and demonstrate the repair in a different task. Old errors become less frequent and less expensive.
The pile of papers becomes a learning map.
Current class enquiries: selected eduKateSG English 3-pax small groups, typically 1.5 hours. Current centres: 83 Punggol Central, Singapore 828761 and 8 Fourth Avenue, Singapore 268674. Contact +65 8823 1234.
A useful taxonomy needs confidence, not just labels
One mistake should rarely create a permanent error family. We distinguish a candidate family from a confirmed recurring family. A candidate appears once or twice and deserves observation. A confirmed family returns across changed tasks, dates or components with the same underlying decision failure.
This matters because the wrong taxonomy can waste weeks. A student who misses one inference question after rushing should not automatically be labelled “weak at inference”. The tutor should ask whether the same boundary failure appears elsewhere.
The recurrence-confidence ladder
- Signal: one observed error; record lightly.
- Candidate: similar failure appears again.
- Recurring family: the same underlying process fails across changed surfaces.
- High-confidence family: the pattern survives attempts to disconfirm it.
- Retiring family: the error has stopped recurring and now needs only delayed maintenance checks.
Do not confuse neighbouring families
Two errors can look alike while requiring different repairs. A vague comprehension answer may come from weak understanding, weak evidence selection or weak written expression. A grammar error may come from missing knowledge or from failing to trigger a known rule under time.
Before assigning a family, ask what the student could do correctly immediately before the failure. The earliest unreliable decision is usually the more useful classification point.
The false-family test
After proposing a family, deliberately test another explanation. If the student succeeds when the task is untimed but fails under time, the family may be execution rather than knowledge. If the student explains the idea correctly aloud but writes it vaguely, the family may be output expression rather than comprehension.
Priority needs more than frequency
A frequent one-mark spelling leak and an occasional five-mark task-fulfilment failure should not automatically receive equal treatment. A useful priority model considers:
- frequency: how often does it recur?
- cost: how many marks or how much meaning does it damage?
- spread: how many components does it affect?
- dependency: does repairing it unlock other skills?
- repairability: is there a clear intervention available now?
The highest-value repair is often the family with moderate-to-high recurrence and wide downstream effect.
Cross-paper recurrence is stronger evidence
If reference tracking fails in comprehension, oral explanation and composition, the pattern is more convincing than three mistakes inside one similar worksheet. Cross-paper recurrence tells us the problem is portable—and therefore the repair must become portable too.
Repair receipts should be observable
A family is not “fixed” because the learner understands the correction while it is being explained. We look for a receipt:
- student names the trigger independently;
- student applies the repair on a changed surface;
- student explains why the tempting wrong route fails;
- student succeeds again after a delay;
- the error rate falls inside mixed work, not only topical practice.
Taxonomy retirement prevents permanent weakness labels
Once an error family stops recurring across several delayed checks, move it out of the active list. Keep a light maintenance probe, but release attention to the next bottleneck. The learner should not carry last term’s weakness as a permanent identity.
The smallest useful active register
Too many active families create another form of overload. For most students, one to three high-value active repairs are enough at a time. Everything else can remain on watch or maintenance status.
AI can classify errors too confidently
AI can label a marked answer instantly, but one answer rarely contains enough evidence to identify a stable learner pattern. A bounded use is: student/tutor proposes the family from several examples → AI may suggest an alternative classification → evidence is compared → changed-task retest decides.
What stronger transfer looks like
- “Careless” comments become specific error families.
- One-off mistakes are not overdiagnosed.
- Recurring families are confirmed across changed tasks.
- Repair priority reflects cost, spread and dependency.
- Students can name their next-time cue.
- Old families are retired when evidence says they are stable.
- The active repair list becomes smaller as independence grows.
The deeper principle
An error taxonomy should become more accurate and smaller over time. Classify the cause, test competing explanations, confirm recurrence across changed surfaces, prioritise by downstream value and retire families when the evidence changes. The aim is not to catalogue every mistake; it is to find the few recurring decisions that are worth repairing next.