How to Improve Vocabulary | Learn Word Boundaries — Examples, Non-Examples and Near Misses

How to improve vocabulary becomes easier when you stop treating a definition as the finish line. A learner can recite that reluctant means unwilling or hesitant and still misuse it in a sentence about someone who is simply slow. The missing knowledge is the boundary: what counts as the word, what nearly counts, and what looks similar but belongs somewhere else.

This guide teaches vocabulary through examples, non-examples and near misses. It builds precise English word meaning by comparing cases that differ in one important feature, then transfers that precision into comprehension, speaking, academic writing and creative writing. Instead of memorising longer definitions, learners learn where a word stops.

Within eduKate’s How to Improve Vocabulary ecosystem, this article owns word-boundary training. Semantic Maps organise neighbourhoods; the Word-Choice Decision System selects among candidates. Here we deliberately construct examples and near non-examples to discover the feature that separates them.

Quick Read: The Near Miss Often Teaches More Than the Easy Example

If a learner sees only obvious examples of fragile—thin glass, a delicate ornament, a cracked shell—the word can become tied to breakable objects. Add a fragile agreement and the concept expands. Add a sturdy plastic cup as a non-example and the boundary becomes clearer. Add an old ceramic mug that looks delicate but survives ordinary handling and the learner must decide which feature actually matters.

The fictional eduKateSG workshop cast—Adrian, Jo, Ben, Aisha, Ryan, Mira, Clara and Ethan—appears throughout. Their cases, dialogues and writing tasks are original teaching material.

1. The Four-Case Boundary Test

For each target word, build four cases: a clear example, a clear non-example, a near example and a near non-example. Then ask which feature changes at the boundary. The answer may involve degree, cause, intention, duration, evidence, social attitude, physical property or grammatical sense.

Do not assume every word has a perfectly sharp boundary. Natural language can be graded, context-sensitive and contested. The purpose of the test is not to pretend language is mathematics; it is to make the learner’s uncertainty visible enough to investigate.

2. Eighty Word-Boundary Laboratories

Boundary Lab 1: reluctant

Clear example: Write a short situation where reluctant is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so reluctant is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 2: fragile

Clear example: Write a short situation where fragile is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so fragile is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 3: generous

Clear example: Write a short situation where generous is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so generous is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 4: stingy

Clear example: Write a short situation where stingy is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so stingy is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 5: brave

Clear example: Write a short situation where brave is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so brave is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 6: reckless

Clear example: Write a short situation where reckless is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so reckless is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 7: confident

Clear example: Write a short situation where confident is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so confident is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 8: arrogant

Clear example: Write a short situation where arrogant is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so arrogant is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 9: curious

Clear example: Write a short situation where curious is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so curious is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 10: nosy

Clear example: Write a short situation where nosy is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so nosy is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 11: persistent

Clear example: Write a short situation where persistent is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so persistent is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 12: stubborn

Clear example: Write a short situation where stubborn is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so stubborn is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 13: careful

Clear example: Write a short situation where careful is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so careful is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 14: cautious

Clear example: Write a short situation where cautious is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so cautious is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 15: timid

Clear example: Write a short situation where timid is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so timid is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 16: efficient

Clear example: Write a short situation where efficient is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so efficient is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 17: effective

Clear example: Write a short situation where effective is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so effective is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 18: reliable

Clear example: Write a short situation where reliable is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so reliable is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 19: accurate

Clear example: Write a short situation where accurate is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so accurate is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 20: precise

Clear example: Write a short situation where precise is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so precise is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 21: relevant

Clear example: Write a short situation where relevant is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so relevant is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 22: significant

Clear example: Write a short situation where significant is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so significant is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 23: essential

Clear example: Write a short situation where essential is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so essential is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 24: optional

Clear example: Write a short situation where optional is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so optional is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 25: plausible

Clear example: Write a short situation where plausible is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so plausible is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 26: possible

Clear example: Write a short situation where possible is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so possible is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 27: likely

Clear example: Write a short situation where likely is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so likely is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 28: certain

Clear example: Write a short situation where certain is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so certain is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 29: evidence

Clear example: Write a short situation where evidence is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so evidence is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 30: proof

Clear example: Write a short situation where proof is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so proof is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 31: cause

Clear example: Write a short situation where cause is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so cause is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 32: factor

Clear example: Write a short situation where factor is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so factor is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 33: pattern

Clear example: Write a short situation where pattern is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so pattern is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 34: coincidence

Clear example: Write a short situation where coincidence is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so coincidence is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 35: observe

Clear example: Write a short situation where observe is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so observe is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 36: infer

Clear example: Write a short situation where infer is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so infer is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 37: assume

Clear example: Write a short situation where assume is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so assume is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 38: persuade

Clear example: Write a short situation where persuade is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so persuade is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 39: inform

Clear example: Write a short situation where inform is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so inform is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 40: warn

Clear example: Write a short situation where warn is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so warn is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 41: request

Clear example: Write a short situation where request is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so request is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 42: demand

Clear example: Write a short situation where demand is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so demand is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 43: glance

Clear example: Write a short situation where glance is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so glance is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 44: stare

Clear example: Write a short situation where stare is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so stare is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 45: peer

Clear example: Write a short situation where peer is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so peer is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 46: wander

Clear example: Write a short situation where wander is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so wander is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 47: stroll

Clear example: Write a short situation where stroll is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so stroll is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 48: trudge

Clear example: Write a short situation where trudge is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so trudge is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 49: whisper

Clear example: Write a short situation where whisper is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so whisper is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 50: mutter

Clear example: Write a short situation where mutter is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so mutter is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 51: announce

Clear example: Write a short situation where announce is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so announce is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 52: murmur

Clear example: Write a short situation where murmur is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so murmur is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 53: irritated

Clear example: Write a short situation where irritated is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so irritated is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 54: furious

Clear example: Write a short situation where furious is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so furious is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 55: resentful

Clear example: Write a short situation where resentful is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so resentful is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 56: indignant

Clear example: Write a short situation where indignant is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so indignant is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 57: relieved

Clear example: Write a short situation where relieved is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so relieved is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 58: content

Clear example: Write a short situation where content is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so content is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 59: anxious

Clear example: Write a short situation where anxious is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so anxious is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 60: alarmed

Clear example: Write a short situation where alarmed is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so alarmed is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 61: vivid

Clear example: Write a short situation where vivid is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so vivid is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 62: vague

Clear example: Write a short situation where vague is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so vague is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 63: coherent

Clear example: Write a short situation where coherent is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so coherent is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 64: complex

Clear example: Write a short situation where complex is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so complex is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 65: complicated

Clear example: Write a short situation where complicated is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so complicated is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 66: adaptable

Clear example: Write a short situation where adaptable is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so adaptable is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 67: flexible

Clear example: Write a short situation where flexible is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so flexible is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 68: robust

Clear example: Write a short situation where robust is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so robust is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 69: delicate

Clear example: Write a short situation where delicate is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so delicate is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 70: scarce

Clear example: Write a short situation where scarce is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so scarce is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 71: rare

Clear example: Write a short situation where rare is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so rare is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 72: unique

Clear example: Write a short situation where unique is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so unique is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 73: ancient

Clear example: Write a short situation where ancient is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so ancient is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 74: obsolete

Clear example: Write a short situation where obsolete is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so obsolete is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 75: innovative

Clear example: Write a short situation where innovative is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so innovative is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 76: novel

Clear example: Write a short situation where novel is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so novel is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 77: transparent

Clear example: Write a short situation where transparent is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so transparent is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 78: ambiguous

Clear example: Write a short situation where ambiguous is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so ambiguous is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 79: explicit

Clear example: Write a short situation where explicit is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so explicit is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 80: implicit

Clear example: Write a short situation where implicit is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so implicit is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 81: literal

Clear example: Write a short situation where literal is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so literal is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

Boundary Lab 82: figurative

Clear example: Write a short situation where figurative is strongly justified. Do not use the target word inside the situation. Make the relevant evidence observable enough that another learner can infer the label.

Clear non-example: Change the central feature so figurative is no longer appropriate. Keep other surface details similar. This prevents the learner from memorising topic, setting or emotional tone instead of meaning.

Near miss: Construct a case that tempts the label but lacks one necessary feature or activates a neighbouring word. State the feature that decides the boundary: degree, cause, intention, duration, evidence, relationship, register or viewpoint.

Writing transfer: Use the successful case in a sentence, then rewrite it for a different genre. Check collocation and word form after the meaning decision. The boundary comes first; sentence polish comes second.

3. Forty Worked Boundary Contrasts

Contrast 1: brave or reckless?

Boundary feature: risk plus judgement. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with brave and the other with reckless.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 2: confident or arrogant?

Boundary feature: self-belief versus inflated superiority. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with confident and the other with arrogant.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 3: curious or nosy?

Boundary feature: interest versus intrusive interest. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with curious and the other with nosy.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 4: persistent or stubborn?

Boundary feature: continued effort versus resistance to change. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with persistent and the other with stubborn.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 5: careful or cautious?

Boundary feature: attention to accuracy versus avoidance of risk. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with careful and the other with cautious.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 6: timid or cautious?

Boundary feature: fearfulness versus considered restraint. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with timid and the other with cautious.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 7: efficient or effective?

Boundary feature: resource use versus achieving intended result. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with efficient and the other with effective.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 8: accurate or precise?

Boundary feature: correctness versus exactness/specificity. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with accurate and the other with precise.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 9: relevant or important?

Boundary feature: bearing on the question versus general importance. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with relevant and the other with important.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 10: significant or essential?

Boundary feature: importance/size versus necessity. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with significant and the other with essential.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 11: possible or plausible?

Boundary feature: can happen versus reasonably believable. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with possible and the other with plausible.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 12: likely or certain?

Boundary feature: probability versus absence of meaningful doubt. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with likely and the other with certain.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 13: evidence or proof?

Boundary feature: support versus conclusive establishment in contexts where proof is appropriate. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with evidence and the other with proof.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 14: cause or factor?

Boundary feature: producing relation versus contributing consideration. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with cause and the other with factor.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 15: pattern or coincidence?

Boundary feature: structured recurrence versus chance co-occurrence. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with pattern and the other with coincidence.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 16: observe or infer?

Boundary feature: notice directly versus conclude from evidence. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with observe and the other with infer.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 17: infer or assume?

Boundary feature: evidence-based conclusion versus taking something as given. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with infer and the other with assume.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 18: inform or persuade?

Boundary feature: provide information versus try to change belief/action. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with inform and the other with persuade.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 19: request or demand?

Boundary feature: asking versus forceful requirement. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with request and the other with demand.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 20: warn or threaten?

Boundary feature: alert to danger versus communicate intended harm or consequence. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with warn and the other with threaten.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 21: glance or stare?

Boundary feature: brief versus sustained looking. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with glance and the other with stare.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 22: peer or stare?

Boundary feature: effort to see versus sustained gaze. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with peer and the other with stare.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 23: wander or stroll?

Boundary feature: lack of fixed direction versus leisurely walking. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with wander and the other with stroll.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 24: stroll or trudge?

Boundary feature: relaxed movement versus effortful heavy movement. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with stroll and the other with trudge.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 25: whisper or mutter?

Boundary feature: low volume versus low/unclear speech often carrying attitude. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with whisper and the other with mutter.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 26: announce or murmur?

Boundary feature: public communicative purpose versus low soft speech. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with announce and the other with murmur.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 27: irritated or furious?

Boundary feature: mild versus extreme anger. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with irritated and the other with furious.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 28: resentful or indignant?

Boundary feature: lingering grievance versus anger at perceived wrong. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with resentful and the other with indignant.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 29: relieved or content?

Boundary feature: release from worry versus satisfied calm state. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with relieved and the other with content.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 30: anxious or alarmed?

Boundary feature: anticipatory unease versus reaction to perceived danger. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with anxious and the other with alarmed.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 31: vivid or detailed?

Boundary feature: strong clear impression versus amount of detail. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with vivid and the other with detailed.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 32: vague or ambiguous?

Boundary feature: insufficient specificity versus multiple plausible interpretations. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with vague and the other with ambiguous.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 33: coherent or complex?

Boundary feature: parts fit logically versus many/interacting parts. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with coherent and the other with complex.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 34: complex or complicated?

Boundary feature: structural multiplicity versus difficulty/intricacy in common usage. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with complex and the other with complicated.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 35: adaptable or flexible?

Boundary feature: capacity to adjust versus ability to change/bend across literal and figurative senses. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with adaptable and the other with flexible.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 36: robust or delicate?

Boundary feature: resistant to disruption versus requiring careful handling. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with robust and the other with delicate.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 37: scarce or rare?

Boundary feature: insufficient availability relative to demand/context versus infrequent occurrence. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with scarce and the other with rare.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 38: rare or unique?

Boundary feature: uncommon versus one of a kind in the relevant set. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with rare and the other with unique.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 39: ancient or obsolete?

Boundary feature: very old versus superseded/no longer useful or current. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with ancient and the other with obsolete.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

Contrast 40: novel or innovative?

Boundary feature: new/unusual versus introducing useful or original change. Build two situations that are identical in topic, people and setting but differ on this feature. Label one with novel and the other with innovative.

Near-miss test: Move the deciding feature gradually rather than flipping it completely. Ask when the first label becomes doubtful and when the second becomes more informative. If the boundary is graded, say so rather than inventing a false exact cutoff.

Sentence test: Put both words into natural sentences and compare their collocations, grammatical frames and connotations. A semantic distinction can be correct while a sentence remains unnatural.

Creative-writing test: Remove both labels and write behaviour that would let a reader infer the distinction. Then decide whether naming the quality adds anything. This connects vocabulary precision with scene construction rather than adjective accumulation.

4. Forty Boundary-Building Drills

Drill 1: One-feature flip

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 2: Degree ladder

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 3: Cause swap

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 4: Intention swap

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 5: Duration swap

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 6: Audience swap

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 7: Evidence swap

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 8: Viewpoint swap

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 9: Register swap

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 10: Collocation swap

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 11: Example without label

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 12: Label without example

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 13: Sort four cases

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 14: Odd one out

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 15: Explain the rejection

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 16: Defend two answers

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 17: Find the hidden sense

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 18: Literal versus figurative

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 19: Everyday versus technical

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 20: Positive versus negative connotation

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 21: Action versus trait

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 22: Moment versus disposition

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 23: Ability versus performance

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 24: Possibility versus probability

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 25: Association versus causation

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 26: Support versus proof

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 27: Specificity versus ambiguity

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 28: Novelty versus usefulness

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 29: Age versus obsolescence

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 30: Rarity versus scarcity

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 31: Precision versus accuracy

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 32: Efficiency versus effectiveness

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 33: Persistence versus stubbornness

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 34: Confidence versus arrogance

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 35: Curiosity versus intrusion

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 36: Courage versus recklessness

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 37: Anxiety versus alarm

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 38: Relief versus contentment

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 39: Near-synonym sentence pair

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

Drill 40: Independent boundary card

Choose one target from current reading. Construct two cases that differ as little as possible while changing the feature named in this drill. Do not put the target word inside either case. Give the cases to another learner and ask which one licenses the word more strongly.

After the decision, explain the boundary in plain language. Then test a third case between the first two. If reasonable speakers could disagree, identify the contextual assumption causing the disagreement instead of forcing a false rule.

Finish with production: write one natural sentence using the target and one using its nearest competitor. Check meaning first, then grammar, collocation and register. The exercise succeeds when the distinction survives outside the original cases.

5. The Umbrella at the Door: An Original Boundary Story

It had stopped raining ten minutes earlier when Ethan found the umbrella beside the library door. He picked it up, looked at the empty corridor and carried it to the desk.

“Generous,” Ben said.

“What is generous?” Ethan asked.

“You. Returning someone’s umbrella.”

Jo looked up from her book. “Is returning property generous?”

Ben paused. “Helpful?”

“Probably. But generous usually involves giving more freely than required—time, resources, attention, perhaps forgiveness. Returning an umbrella that is not yours may be responsible or considerate without being generous.”

Ethan placed the umbrella behind the desk. “So I failed generosity?”

“You failed the vocabulary test,” Ben said. “Morally, you seem fine.”

A few minutes later a younger student arrived, wet at the shoulders. Her own umbrella had broken. Ethan offered his spare folding umbrella from his bag and told her she could return it the next day.

Ben pointed. “Now?”

Jo smiled. “Much stronger example. He is voluntarily giving access to something useful that belongs to him, at some inconvenience or risk.”

Ethan stared at them. “I am beginning to regret doing anything near vocabulary students.”

“Reluctant?” Ben asked.

“Not yet.”

“Near miss,” Jo said.

6. What the Umbrella Story Teaches

The first action is positive, but positive behaviour is not automatically generous. The near miss is useful because it shares the favourable evaluation while lacking part of the concept. That forces the learner to identify what generosity contributes beyond “good.”

The second action strengthens the label because Ethan voluntarily makes his own resource available to another person. Even then, context matters; generosity is not reducible to a numerical rule about how much must be given.

The final joke demonstrates another boundary. Regret or annoyance is not automatically reluctance. Reluctance becomes more appropriate when someone is unwilling or hesitant to perform an action they may nevertheless perform.

7. A Four-Week Word-Boundary Programme

Week one: obvious cases. Build clear examples and non-examples for five useful words. Explain the central feature in plain language.

Week two: near misses. Keep topic and setting constant while changing one semantic feature. Compare neighbouring words rather than unrelated opposites.

Week three: uncertain boundaries. Use graded cases and discuss where reasonable disagreement begins. Check reliable references and authentic examples when intuition is unstable.

Week four: writing transfer. Use the words in independent explanation, dialogue and narrative. Then run the Vocabulary Revision Pass to see whether the distinctions survived real drafting.

8. Final Principle: Know What the Word Excludes

A definition tells you where to begin. Examples show the word alive. Non-examples stop the concept from spreading too far. Near misses reveal the feature that actually controls the choice.

This is why precise vocabulary is not simply a larger vocabulary. It is a vocabulary with better boundaries. The learner knows not only what furious can describe, but when irritated is more truthful; not only what proof means, but when the evidence supports only suggests; not only what unique means, but when rare is the defensible word.

Learn the centre of the word. Then walk towards the edge. The edge is where precision begins.

Continue: How to Improve Vocabulary — Complete System