The English Tutor Handbook · Volume 0006 · Series ID SETC-ETH-0006
Two wrong answers can look identical on the page and require completely different teaching.
Two weak paragraphs can receive the same mark and fail for different reasons.
Two grammar errors can share the same correction while coming from different mechanisms.
If a tutor classifies every English problem by surface appearance, the tutor will eventually prescribe the wrong repair.
The English Error Map exists to increase resolution.
This volume does not replace SETC’s existing learner-facing examination articles such as PSLE English Comprehension Error Patterns. It also does not replace the generic Diagnostic Tutor owner on eduKate Sengkang. Its job is practitioner-side and cross-skill: give the English tutor one working classification for mistakes appearing in reading, vocabulary, writing, editing, listening, speaking and timed performance.
Quick Read
- An error is evidence, not a diagnosis.
- Classify the earliest useful failure, not merely the last visible mistake.
- Distinguish task, access, meaning, reasoning, selection, expression, verification and performance errors.
- Separate knowledge gaps from retrieval failures.
- Separate misunderstanding from misapplication.
- Separate one-off slips from stable patterns.
- Record error conditions: when, where and under what load the problem appears.
- Use the map to choose the next test and the smallest useful repair.
1. Wrong Is Too Low-Resolution
“Wrong” tells the tutor the response failed a criterion.
It does not tell the tutor what the learner needs next.
Suppose Iona answers a comprehension question incorrectly.
Possible causes include:
- she misread the task;
- she did not know one word;
- she lost track of a pronoun;
- she built the wrong passage model;
- she inferred too early;
- she found the right evidence but chose the wrong relationship;
- she knew the answer but wrote it incompletely;
- she failed to check a near-correct option;
- she rushed because time was low.
One incorrect answer can sit at the end of nine different routes.
2. The Eight-Layer Error Map
A practical tutor can begin with eight layers:
- Task error: the learner solves the wrong problem.
- Access error: the language cannot be decoded, recognised or retrieved reliably.
- Meaning error: the learner builds an inaccurate or incomplete model.
- Reasoning error: the learner makes a weak inference, comparison, causal claim or judgement.
- Selection/design error: the learner knows more than the response should contain, but chooses or organises poorly.
- Expression error: intended meaning is damaged during wording, sentence construction or delivery.
- Verification error: a correctable problem survives because checking is absent or badly targeted.
- Performance error: an otherwise available capability collapses under time, pressure, fatigue, switching or workload.
The map follows the English Learning Map from Volume 0001 but turns it into an error-classification tool.
3. Task Errors — The Student Solves the Wrong Problem Well
Task errors are often expensive because everything downstream can be competent and still irrelevant.
Examples:
- explaining what happened when the question asks why;
- giving one point when two are required;
- writing generally when source evidence is required;
- using an informal tone for a formal audience;
- answering the topic rather than the exact question;
- describing a language feature without explaining its effect.
The correction should not begin with sentence polish. It should begin with task reconstruction.
4. Access Errors — The Meaning Is Locked Behind Language
Access errors occur when the learner cannot reliably get through the language layer.
In younger learners this may include decoding and fluency. Across ages it can include vocabulary, syntax, reference and retrieval.
- unknown word;
- known word used in an unfamiliar sense;
- slow decoding that consumes attention;
- misread punctuation;
- failure to parse a long sentence;
- pronoun confusion;
- failure to retrieve a known word while writing or speaking.
Access is not the same as intelligence or reasoning. A student may reason well once the language barrier is removed.
5. Meaning Errors — The Learner Builds the Wrong Model
A meaning error happens when enough words are accessible but their relationships are integrated incorrectly.
Examples:
- events placed in the wrong sequence;
- cause confused with coincidence;
- speaker A’s opinion attributed to speaker B;
- a concession treated as the main claim;
- a later correction ignored;
- a paragraph read as support when it is actually a counterpoint.
The tutor needs to rebuild the model, not simply supply the answer.
6. Reasoning Errors — The Model Exists but the Decision Is Weak
Reasoning errors appear when the learner has enough information but makes an unsound decision from it.
- premature inference;
- ignoring contradictory evidence;
- selecting evidence because of a keyword rather than relationship;
- treating possibility as certainty;
- confusing example with proof;
- accepting a near-correct option without comparing alternatives;
- using background knowledge to override the text.
This category is especially important because weak reasoning can masquerade as weak comprehension.
7. Selection Errors — The Learner Knows Too Much or Chooses Poorly
English rewards selection.
A student may understand five relevant details but need only two.
A writer may have six ideas but need three distinct paragraph roles.
A speaker may know ten examples but need one clear example and a developed explanation.
Selection errors include:
- over-answering;
- under-answering;
- choosing related but incomplete evidence;
- using impressive vocabulary that weakens precision;
- including every idea instead of prioritising;
- keeping a sentence because it is beautiful rather than useful.
The repair is judgement, not more content.
8. Design Errors — The Parts Do Not Form the Right Whole
Selection and design are related but not identical.
A learner may choose good material and arrange it badly.
- paragraphs repeat instead of progressing;
- evidence arrives before the claim it supports;
- a narrative climax occurs before enough causal build-up;
- an oral answer gives the example before establishing the point;
- a situational-writing response includes all content points but in an order that obscures purpose.
Design errors belong above sentence correction because architecture controls what the sentences are trying to do.
9. Expression Errors — Meaning Is Lost in Translation
Expression errors happen when the learner knows roughly what to say but the language produced does not carry it reliably.
- tense makes the timeline ambiguous;
- pronouns lose reference;
- word choice distorts intensity;
- sentence structure creates unintended meaning;
- punctuation hides a relationship;
- spoken delivery becomes too compressed for the listener;
- register does not fit the audience.
The tutor should separate expression from idea quality whenever possible. Otherwise students can be mislabelled as weak thinkers when the real failure is linguistic control.
10. Verification Errors — The Learner Could Have Caught It
Some mistakes survive because the checking system is weak.
Verification errors include:
- not rereading the question;
- checking spelling but not scope;
- reading intended wording instead of actual wording;
- not comparing a doubtful answer with evidence;
- failing to notice a missing required point;
- spending all checking time on low-risk surface errors.
The repair is often a targeted verification routine rather than more content teaching.
11. Performance Errors — Capability Changes Under Conditions
A learner can possess a capability and still fail to deploy it under load.
Performance errors appear when the learner:
- misreads more questions near the end;
- abandons planning when timed;
- uses simpler vocabulary under pressure;
- stops verifying after one difficult item;
- freezes after an unexpected prompt;
- spends too long repairing one sentence and rushes the conclusion.
These are not excuses. They are conditions. If performance matters, the conditions must be trained.
12. Missing Node, Broken Edge, Wrong Edge, Routing Problem
Another useful way to increase resolution is to ask what kind of knowledge-network failure is present.
Missing Node: the learner never properly learned the relevant concept, word, structure or rule.
Broken Edge: both pieces are known but the relationship between them is weak.
Wrong Edge: the learner has connected two things incorrectly.
Routing Problem: the learner knows the method but does not recognise when to use it.
This language is useful because “more practice” treats all four as though they were the same.
13. Knowledge Gap or Retrieval Failure?
A word absent from a composition may be unknown.
Or it may be known but unavailable during production.
A strategy not used in an examination may be unlearned.
Or it may be understood but not self-initiated.
Test both.
Ask the learner to explain the concept directly. Then later create a task where the learner must retrieve it without being told that the concept is relevant.
Knowledge and availability are different variables.
14. Misunderstanding or Misapplication?
A student may explain a rule correctly and apply it incorrectly.
That suggests the failure is not the rule itself.
Perhaps the student misclassified the situation.
For example, the learner knows how to infer meaning from context but tries to infer when the passage directly defines the word. The method is known. The routing is wrong.
Teaching the method again may create frustration without improving selection.
15. Slip or Pattern?
Not every error deserves a theory.
People make slips.
A useful pattern generally shows some combination of recurrence, condition stability and explanatory value.
Ask:
- Has this happened before?
- Does it appear across different tasks?
- Does one small prompt reliably repair it?
- Does the same mechanism explain several visible mistakes?
- Does it disappear when conditions change?
One isolated mistake should often remain just that: one mistake.
16. Error Frequency Is Not Error Priority
A learner may make many small punctuation slips and one occasional task-reading error.
The task-reading error may be far more expensive.
Prioritise using:
- frequency;
- meaning cost;
- downstream impact;
- generalisability;
- repairability;
- performance relevance.
The goal is not to remove the largest number of red marks. It is to improve the largest amount of future performance.
17. Error Cascades
English errors often cascade.
A task error causes a selection error.
The selection error creates a design problem.
The learner then forces irrelevant material into a sentence and creates an expression error.
The tutor sees the awkward sentence and teaches grammar.
The earliest useful failure was three layers earlier.
When several mistakes cluster, ask whether they share one upstream cause.
18. Error Clusters in Reading
Reading errors can cluster by mechanism.
- Reference cluster: pronouns, speakers and entities become confused.
- Update cluster: later information fails to revise an earlier interpretation.
- Scope cluster: the learner repeatedly answers a broader or narrower question.
- Evidence cluster: relevant details are found but incomplete proof is selected.
- Inference cluster: the learner jumps beyond available evidence.
- Structure cluster: paragraph relationships are not tracked.
These clusters are more teachable than the label “weak comprehension”.
19. Error Clusters in Writing
- Idea cluster: claims appear without development.
- Architecture cluster: paragraphs repeat or lack distinct jobs.
- Cohesion cluster: references and transitions fail to connect ideas.
- Sentence-load cluster: grammar collapses when sentence complexity rises.
- Lexical cluster: vocabulary is vague, overformal or collocationally weak.
- Revision cluster: the learner edits surface features but not meaning or structure.
Again, the visible red marks may be downstream symptoms.
20. Error Clusters in Listening
- failure to predict the information category;
- overwriting a correct first understanding with a distractor;
- missing changed information;
- trying to note everything;
- losing one item and carrying the disruption forward;
- confusing uncertainty with absence of meaning.
The tutor should distinguish hearing difficulty, language-access difficulty, note-selection difficulty and decision difficulty.
21. Error Clusters in Speaking
- idea generation stalls;
- response does not fit the question;
- organisation disappears during spontaneous speech;
- vocabulary retrieval narrows;
- sentence repair becomes excessive;
- delivery fluency hides shallow content;
- memorised structures reduce responsiveness.
Fluency should not be treated as the whole speaking system.
22. Error Clusters in Editing
Editing can fail because the learner:
- does not detect the error;
- detects something is wrong but misclassifies it;
- knows the correction but cannot explain why;
- repairs locally while creating another error;
- uses memorised rules without checking sentence meaning;
- checks only familiar error categories.
Detection, classification and repair are separate capabilities.
23. Error Conditions Matter
Record not only what went wrong but when.
Does the error appear:
- only when timed?
- only in longer texts?
- only when two plausible answers exist?
- only when vocabulary is unfamiliar?
- only after several demanding questions?
- only in independent work?
- only when the learner writes more ambitious sentences?
Condition-sensitive errors require condition-sensitive repair.
24. The Error Interview
Before correcting a recurring mistake, ask:
- What did you think the task required?
- What information did you use?
- What made you choose this?
- What other possibility did you consider?
- Where were you least certain?
- What would you check if you could try again?
The interview turns a mark into evidence about process.
25. The Smallest Separating Test
When two explanations are plausible, design a tiny test that separates them.
Does Maren lack vocabulary or merely retrieve poorly during writing?
Ask for definitions first. Then create a new writing context without showing the list.
Does Iona misunderstand inference or select evidence badly?
Let her explain the inference orally, then separately choose the smallest complete proof.
Does Leonie lack the method or lose it under time?
Compare timed and untimed versions.
Small tests prevent the tutor from building a large plan on an untested story.
26. Error Maps Should Shrink, Not Grow Forever
A tutor can become so fascinated by diagnosis that the learner accumulates an enormous catalogue of weaknesses.
That is not the goal.
The map should compress complexity into a small number of current high-leverage patterns.
Old errors that no longer recur should be retired.
New evidence should replace old labels.
The map exists to guide intervention, not to preserve a permanent history of failure.
27. From Error Category to Intervention
The classification should change what happens next.
- Task error: teach task parsing and answer requirements.
- Access error: teach or strengthen vocabulary, decoding, syntax or retrieval.
- Meaning error: rebuild text or sentence models.
- Reasoning error: compare interpretations and demand evidence.
- Selection/design error: practise prioritisation, scope and architecture.
- Expression error: practise controlled language production around known meaning.
- Verification error: build targeted checking routines.
- Performance error: train timing, recovery and pressure conditions.
The map earns its place only when it improves intervention precision.
28. From Intervention Back to Error Map
Teaching itself generates evidence.
If a task-parsing routine removes several downstream mistakes, the tutor’s classification gains support.
If vocabulary teaching changes nothing, perhaps vocabulary was not the bottleneck.
If performance improves immediately when time is removed, the condition matters.
A tutor should be willing to revise the error map when the repair behaves differently from prediction.
29. Maren, Iona and Leonie on the Error Map
Maren
Visible symptom: polished writing that feels repetitive.
Likely map: selection/design error rather than expression error.
Useful test: give her excellent sentences and ask her to organise them into two paragraph structures with different jobs.
Iona
Visible symptom: inconsistent comprehension marks.
Likely map: reasoning plus evidence-selection error.
Useful test: separate interpretation from proof. Ask her to explain the meaning before choosing evidence.
Leonie
Visible symptom: strong untimed work, weak test execution.
Likely map: performance and verification error under time.
Useful test: compare the same decision with and without time pressure, then observe which process disappears.
30. A Compact Error Ledger
For recurring issues, record:
- Observed error: what happened?
- Likely layer: task, access, meaning, reasoning, design, expression, verification or performance?
- Condition: when does it appear?
- Alternative explanation: what else could produce the same symptom?
- Separating test: what small task would distinguish them?
- Intervention: what is the smallest repair?
- Retest: did the error pattern change?
The ledger is not a score sheet. It is a memory aid for better teaching decisions.
31. Research Anchors
The English Error Map is a practitioner synthesis rather than a claim that educational research endorses one universal taxonomy. Several evidence-based guidance sources support the underlying principles.
- The US Institute of Education Sciences’ Teaching Secondary Students to Write Effectively recommends using assessment of student writing to inform instruction and feedback.
- The IES Secondary Writing Toolkit includes formative assessment and analysis of student work as a basis for future instruction.
- IES reading practice guides recommend explicit comprehension practices, purposeful text work and literacy interventions tailored to identified needs.
- The Education Endowment Foundation’s work on checking for understanding and feedback emphasises using evidence of learning to determine the next instructional response rather than treating checking as an end in itself.
32. How this guide fits
This article is deliberately broader than any one examination component and narrower than generic tutor diagnosis.
- PSLE English Comprehension Error Patterns helps learners recognise recurring PSLE comprehension mistakes.
- The English Tutor Handbook Vol No.0002 covers evidence collection before intervention.
- The Tutor Handbook Vol No.0004 | Diagnostic Tutor explains the diagnostic tutor’s role.
This guide focuses on the English-specific cross-skill error classification used by tutors after evidence is collected and before intervention is selected.
33. Final Compression
Do not mark only the symptom.
Ask which layer failed first.
Task?
Access?
Meaning?
Reasoning?
Selection?
Design?
Expression?
Verification?
Performance?
Then ask what kind of failure it is.
Missing knowledge?
Broken relationship?
Wrong relationship?
Retrieval problem?
Routing problem?
Condition problem?
Build the smallest separating test.
Choose the smallest useful repair.
Retest.
Update the map.
The point of an error map is not to give mistakes more names. It is to make the next teaching decision more precise.
That is the English Error Map.
Next: The English Tutor Handbook Vol No.0007 | Practice Engineering — Building Retrieval, Variation, Discrimination and Mixed English Practice Without Worksheet Volume.
Explore the English Tutor Handbook. Related teaching guides: Vol. 0012 — Tutor Calibration; Vol. 0025 — Marked-Paper Forensics.