Vocabulary | What Is a Vocabulary Experiment? Testing How Words Improve a Real Task

Imagine two students learning the word constraint. One writes “a limitation” beneath it and repeats that definition. The other uses the word while planning a small exhibition: the room is available for one hour, the display must fit on two tables, and the team can print only a limited number of pages. Both students may pass a matching exercise. Only the second example shows the word helping someone organise a task.

A vocabulary experiment asks whether a learner can move between the word, the idea and a situation. That movement matters because real tasks rarely announce the exact vocabulary they require. A story does not always say, “Now distinguish intention from consequence.” A group project does not always say, “Now identify the condition that prevents your plan from working.” You notice the need and bring the relevant idea to it.

The useful unit is a word in an action

Start with a sentence such as: “When I read a claim, I will use evidence to ask what supports it.” The sentence connects vocabulary with an observable action. It is more informative than “I will learn evidence,” which could mean anything from spelling the word to evaluating a complicated argument.

Other useful starting sentences are: “I will use at most to identify the largest permitted value”; “I will use draft to communicate that a piece of work is ready for comments but may change”; and “I will use exception to look for a case that a general rule does not cover.” Each sentence gives the learning a job.

You can then design a task that reveals whether that job is happening. If your target is at most, asking for its dictionary meaning is only one check. A stronger task asks which of several quantities satisfy a stated limit, and why the boundary value belongs. If your target is draft, ask a partner what action they believe you are requesting when you send one. Their response reveals whether the shared meaning is working.

Three different questions

Question	What it can reveal	What it leaves uncertain
Can I explain the word?	Accessible meaning in this moment	Whether I will recognise when to use it
Can I use it in this task?	Application under these conditions	Whether the skill will last or transfer
Can I use the distinction later in a new task?	More durable and flexible use	How far that use generalises beyond the tasks tried

The third question is demanding, but it should not make the first two feel worthless. A definition gives you a starting point. A familiar application lets you practise. A new application checks whether the learning is becoming flexible. The problem arises when one kind of evidence is used to claim another. Getting a flashcard right does not by itself show that the word will improve an essay.

A practical learning trial, not a verdict about a person

Suppose a learner misunderstands three instructions containing unless. You now have a useful teaching clue. You do not have a measure of the learner’s overall intelligence, effort or worth. The difficulty may involve the word, sentence structure, attention, unfamiliar subject matter or several things at once.

Phrase the question so that correction remains possible: “Which part of this instruction changes the action?” is more useful than “Why are you bad at instructions?” A vocabulary experiment should make the task clearer. If it turns into a contest to identify who is clever, it has lost its purpose.

All the people, task records and numerical results in this guide’s worked cases are invented teaching examples. They show how to reason about a result. They are not reports of eduKate student outcomes or promises of improvement. When you try a case yourself, your own observations may look different, and that difference is information worth examining.

Contents · Next chapter

A technology is often pictured as a device: a microscope, a camera or a computer. Here, the comparison is broader. People create and share ways of representing a problem so that something previously difficult becomes more manageable. A map helps someone plan a route. A numbered recipe helps someone repeat a sequence. Vocabulary can perform a similar supporting role by making distinctions portable.

Consider cause, condition and coincidence. The words let you point to three different relationships without rebuilding a long explanation every time. You still need evidence to decide which relationship is present. The words provide handles for the investigation; they do not supply its answer.

Four ways a word can help

It can direct attention. A reader who understands contrast has a reason to examine differences between two accounts. A learner who understands boundary has a reason to inspect the point where a condition changes. Attention becomes more purposeful because the word suggests a question.

It can organise information. Headings such as claim, evidence and question divide a page into useful roles. Without those headings, the same information may appear as an undifferentiated pile of notes. With them, a reader can see a missing connection.

It can guide an action. The difference between estimate and calculate changes how you approach a numerical task. The difference between describe and explain changes what an answer must attempt. Vocabulary can therefore influence the sequence of work, not just the style of the final sentence.

It can coordinate people. A team that distinguishes proposed, agreed and completed has a clearer way to discuss the status of a task. Everyone still needs to understand the labels and update them honestly. A status word with no shared rule may create the appearance of coordination while confusion continues underneath.

The mechanism must be visible

When you say that a word helped, try to finish the sentence: “It helped because…” Did it remind you to inspect a boundary? Did it make a missing reason visible? Did it let a partner ask a more precise question? A visible mechanism makes the result easier to interpret and the learning easier to repeat.

For instance, “My writing improved after I learned although” is vague. “I used although to connect a concession to my main point, and I checked that the two clauses expressed the intended relationship” is more specific. The second account describes something a teacher or peer can inspect. It also exposes an error if the conjunction is used incorrectly.

This is why a useful experiment records an explanation as well as an answer. A correct answer reached for the wrong reason may fail on the next task. A slightly awkward sentence that reveals a sound distinction may be a productive stage in learning. The point is to understand what the learner is doing, not merely to count polished words.

The limits matter

Vocabulary works alongside perception, experience, diagrams, memory, movement and other resources. It is not the whole of thought. In a small clinical study, Varley and colleagues described preserved mathematical abilities in three adults with severe grammatical impairments. That finding cautions against treating language and reasoning as identical; it does not settle every question about their relationship. See Agrammatic but Numerate.

A person may recognise a pattern before knowing a conventional name for it. Another may know an impressive name while misunderstanding the pattern. A bilingual learner may understand an idea clearly in one language while still learning how to express it in English. Your experiment should allow for those possibilities.

The most useful promise is therefore specific: learning and practising a distinction may improve access, precision or coordination in the tasks where that distinction matters. Test that promise at the scale of the task. There is no need to turn it into a sweeping claim that more vocabulary automatically creates more intelligence.

Contents · Previous chapter · Next chapter

Before choosing new vocabulary, watch one real attempt. A student rereads a question five times. A group keeps revising a poster because nobody knows what “finished” means. A reader describes every fictional character as “nice” or “bad”. Each situation suggests a different possible language problem. Starting with the situation makes the learning more purposeful.

Ask the learner to talk through a small part of the task. Listen for where the explanation becomes uncertain. Do they misunderstand a word in the question? Do they know it but forget to use the relevant idea? Can they explain orally but struggle to produce the written form? Are they missing the subject knowledge needed to apply the word?

Separate four common difficulties

Difficulty	Example	A useful first response
The meaning is unfamiliar	The learner cannot explain what “at least” permits	Teach the meaning with examples and boundary cases
The meaning is known but not noticed	The learner explains “unless” after the mistake but skipped it while reading	Practise locating the condition before acting
The idea is understood but hard to express	The learner can demonstrate a contrast but cannot write it clearly	Offer sentence examples and supported production
The task requires missing subject knowledge	The learner knows “evaluate” but lacks information about the options	Build the knowledge needed for evaluation

The last row is especially important. If a student cannot explain why a material suits a particular design, teaching the word suitable will not provide the missing facts about the material. Vocabulary can organise knowledge that exists or help a learner ask for knowledge that is missing. It cannot replace every kind of instruction.

Use an error that is small enough to inspect

“My essays are weak” is too large for a first experiment. “My paragraph states a claim but does not show how the example supports it” is workable. You can introduce reason or warrant, depending on the learner, and inspect that connection in a fresh paragraph.

“Our project meetings are chaotic” is also too large. “We leave without distinguishing a suggestion from an agreed action” is more precise. You can introduce two status labels and check whether each participant leaves with the same understanding. The trial becomes manageable because you have chosen one point of failure.

Write down the original difficulty before teaching. Otherwise, a successful lesson may tempt you to change the description of the problem afterwards. A simple note is enough: “In two sample instructions, the learner included values beyond the stated maximum.” That observation gives the next attempt a clear comparison.

Invite the learner’s account

Ask, “What did you think the sentence was asking you to do?” A response such as “I thought compare meant list everything I know” tells you something more useful than a mark alone. It reveals the interpretation that produced the work.

For a younger child, use objects, pictures or oral explanation. You might place six counters on the table and ask the child to show “at most four”. If the child removes two and explains the limit accurately, a written error elsewhere may involve something besides the core meaning.

For an older learner, ask for a counterexample. “Show me a response that would describe the event but would not explain it.” If the learner can create one, they have begun to separate the categories. If every example fits both categories, the boundary may need more teaching.

Keep the first observation brief. The purpose is to identify a useful next move, not to collect a permanent dossier of mistakes. A notebook line, a saved sentence or a marked task can be enough. Choose evidence you can revisit without making the learner feel that every ordinary conversation is an assessment.

Contents · Previous chapter · Next chapter

The easiest way to overload a vocabulary experiment is to teach too many words at once. If a student learns claim, evidence, reasoning, counterclaim, rebuttal, assumption and qualification in one lesson, any later change may be difficult to interpret. The words are related, but they make different demands.

Start with the smallest distinction that addresses the observed difficulty. If the learner treats a personal opinion as proof, claim and evidence may be enough for the first trial. Once the learner can distinguish those roles, you can examine the reasoning that connects them.

Choose by usefulness, not impressiveness

A familiar word can carry a difficult idea. Only, unless, some, all, before and after can change the meaning of an instruction. There is no reason to begin with an unusual term if a common word is causing the problem.

Ask three questions. Does this distinction occur in work the learner actually does? Does misunderstanding it produce a recognisable error? Can you create a small task that reveals whether the distinction is being used? When all three answers are yes, you have a strong candidate.

Suppose a group keeps saying that an idea is “good” without agreeing why. You could teach innovative, elegant and sophisticated. Those may enrich discussion, but they do not necessarily solve the immediate problem. The terms criterion and evidence may be more useful because they prompt the group to state what it is judging and what supports the judgement.

Teach the word with neighbours and limits

Give a short explanation, an example and a contrasting example. If your word is reluctant, distinguish being unwilling or hesitant from being unable. A learner may want to join a game but be unable because of another commitment. Another may be free to join but reluctant because the game feels unfamiliar. The situations are different even if both people stay outside the game.

Then inspect how the word behaves in a sentence. “Reluctant to speak” is a useful pattern. A definition alone does not supply it. Nation’s account of word knowledge includes connections among form, meaning, grammatical behaviour, collocations and conditions of use. That broader view explains why understanding a word involves more than matching it to a short gloss. See Knowing a Word.

Keep your own teaching proportionate. You do not need a full linguistic analysis before a learner can use a word. You do need enough information to prevent the particular mistake you are investigating. Add another layer when the next task makes that layer useful.

Write a testable learning intention

A practical intention might read: “Given a short account of an event, I can separate what the narrator directly observed from what the narrator inferred, and I can explain one uncertain inference.” It names the task, the distinction and the evidence of use.

Compare that with “I will have a powerful vocabulary.” The second sentence is encouraging but impossible to inspect in a short trial. The first lets you decide whether the next activity should be easier, more varied or less supported.

Finally, identify when the word would be unnecessary. A student should not force constraint into every paragraph or label every ordinary sentence as an inference. Knowing when to leave a term unused is part of practical control. The goal is a better task, not a higher count of target words.

Contents · Previous chapter · Next chapter

A learner completes a difficult passage, receives an explanation of inference, and then completes a much easier passage. The second answer is better. Did the vocabulary teaching help? It may have, but the change in task difficulty gives another explanation. A useful comparison tries to keep such differences visible.

You do not need laboratory conditions to learn something from ordinary practice. You do need a little discipline about what changed. Use tasks of similar length, subject familiarity and demand. Keep the instructions broadly consistent. Record help that was given. Then describe the result at the level your evidence supports.

The same task and a similar task answer different questions

Repeating the exact task can show whether the learner now understands that example. It is valuable during teaching. However, the learner may remember the answer rather than apply the distinction afresh. A new but comparable task gives you another view.

Suppose the first mathematics item says, “Choose a whole number that is at most seven.” A comparable item might use “at most twelve” with a different set of options. An item involving a complicated algebraic inequality would change more than the vocabulary. If the learner fails it, you would not know whether the obstacle was language or algebra.

For reading, keep the inferential demand similar. If the first paragraph requires distinguishing visible behaviour from an uncertain motive, choose another paragraph with the same kind of distinction. Avoid making the second paragraph supply the motive explicitly. That would test a different reading problem.

Order creates its own effects

The second attempt often benefits from practice, familiarity or simply knowing what the adult expects. It can also suffer from fatigue. These possibilities do not make the activity useless. They limit the conclusion. Write “The learner performed more accurately on the second set after teaching” rather than “This word caused a permanent improvement in reasoning.”

If you are comparing two forms of an instruction with several willing classmates, you can vary which version people see first. That may help reveal an order effect. It does not automatically create a rigorous study. Different participants bring different knowledge, and a small classroom activity cannot control everything that matters.

For one learner, a practical alternative is to collect several brief observations over time. Look for a pattern across different examples rather than celebrating a single favourable result. Keep the burden low: two or three carefully chosen tasks can be more informative than a large pile of rushed exercises.

Keep the teaching honest

If you teach the meaning, give a worked example, provide a diagram and prompt the learner during the next attempt, you have changed several things. That may be excellent instruction. The mistake would be attributing the whole change to the word alone.

Record the package plainly: “Short explanation, two examples, a number line and one reminder to inspect the boundary.” On a later attempt, reduce the help and see what remains possible. The comparison becomes a way of planning instruction rather than a performance advertisement.

Sometimes the practical question is precisely whether a support helps. A learner may need to use a glossary or a task card in normal work. You can compare performance with that support, but call it supported performance. Later, if independence matters, check what happens when the support is absent.

A modest conclusion can still be useful

“The revised instruction reduced requests for clarification in this group” is a useful observation. So is “The learner explained the distinction accurately but did not notice it without prompting.” Neither statement needs a claim about everyone, every subject or every future task.

A practical experiment earns its value by helping you choose the next step. Perhaps the word needs a contrasting example. Perhaps the task needs a clearer label. Perhaps the learner is ready for a new context. An honest small conclusion is enough to guide those choices.

Contents · Previous chapter · Next chapter

Before the next attempt, decide what you will inspect. If you wait until afterwards, it is easy to select whichever result looks best. A learner might answer faster but make more mistakes, or write a longer paragraph without improving its reasoning. A single positive-looking number can hide the actual trade-off.

Choose a primary outcome that matches the task. For boundary words in mathematics, accuracy may matter most. For a group handover, shared understanding of the next action may be central. For descriptive writing, the relevant question may be whether a chosen word fits the evidence and tone.

Keep accuracy, speed and quality separate

A faster answer is useful only when speed serves the task. If the learner is learning a subtle distinction, slower but more accurate work may be welcome. If a team spends an extra minute clarifying an instruction and avoids a longer repair later, the extra discussion may also be worthwhile.

Write two short observations rather than forcing everything into one score. For example: “Correctly identified four of five boundary cases; took longer than the first attempt because each answer included a reason.” This gives you a much clearer picture than “Performance improved.”

Quality can be assessed with a small set of visible features. A reading explanation might need a claim, a relevant detail and a statement of how certain the interpretation is. Avoid turning a nuanced response into a complicated scoring system when a brief discussion would reveal more.

Purpose	Useful observation	Weak substitute
Read a condition accurately	Correct inclusion and exclusion of boundary cases	Number of advanced words used
Explain a claim	Relevant evidence connected to the claim	Paragraph length alone
Coordinate a task	Participants identify the same next action and owner	How professional the message sounds
Use a word flexibly	Appropriate use in an unfamiliar example	Reciting yesterday’s sentence
Retain a distinction	Accurate use after a gap, with support recorded	Immediate confidence after reading the explanation

Count opportunities as well as successes

If a learner uses a distinction correctly three times, you need to know whether there were three opportunities or ten. “Three correct applications in four opportunities” is clearer than “Used the word three times.” The denominator matters because the task creates the opportunities.

It also matters whether the learner needed prompting. You can mark an attempt as independent, lightly prompted or fully guided. These labels are descriptions for your local activity, not standard scores. Define them simply enough that you can apply them consistently.

For example, “lightly prompted” could mean that the adult said, “Check the condition,” without identifying the relevant word. “Fully guided” could mean that the adult explained the condition and worked through the decision. Both forms of help can support learning. Recording them prevents you from confusing assistance with independence.

Inspect costs as well as benefits

A new term can make a discussion more precise while also adding explanation time. A detailed checklist can prevent omissions while becoming too cumbersome for a simple task. A child might understand the idea better when the adult uses everyday wording before introducing the formal label.

Ask whether the improvement is worth the effort in the setting where the tool will be used. If a three-word label requires a ten-minute explanation every time, it may not yet be useful shorthand. The next move may be to practise the concept in a familiar context or choose a clearer phrase.

The goal is useful control. A vocabulary experiment succeeds when it reveals how to improve the match between the learner, the language and the task. A result that shows a word was unnecessary can be just as valuable as a result that shows it helped.

Contents · Previous chapter · Next chapter

A good record can fit on one page. It should preserve the question, the conditions, the learner’s response and the next step. It does not need elaborate charts or a large spreadsheet. If recording takes longer than the useful learning, simplify it.

Begin with the task and the target distinction. Then record what the learner did before teaching, what support was introduced, and what happened in the later attempt. Include one quotation or concrete example where possible. “Wrote that the character was angry because he closed the door” is more informative than “Inference weak.”

A reusable record

Field	What to write
Question	Will distinguishing observation from inference help me qualify a reading claim?
Task	Two short fictional accounts with uncertain motives
First attempt	Quoted the action accurately but treated one possible motive as certain
Teaching and support	Explanation, one contrast, and a two-column note
Next attempt	Separated the visible action from a possible explanation
Remaining uncertainty	The second account may have been easier; the note was available
Next step	Try a new account later without the note and ask for an explanation

This record captures a process. It also makes your uncertainty explicit. You can see what improved, what was supported and what needs another look. A row for uncertainty is particularly useful because it gives doubt a place to go. You do not have to choose between exaggerated celebration and dismissing the activity entirely.

Record errors by their cause when you can

A wrong answer may arise from a misunderstood meaning, a missed condition, an unsupported assumption, a weak example or a slip while writing. Do not assign a cause just because it seems plausible. Ask the learner to explain, or mark the cause as uncertain.

Suppose a student selects thirteen when asked for a number “at most twelve”. The student might believe the phrase means “more than”. They might have misread twelve as twenty. They might understand the meaning but select the wrong option accidentally. Those possibilities call for different responses.

Use a correction sentence: “The error happened because I…” followed by a specific action or interpretation. If the learner cannot yet explain, the adult can offer two possible accounts and ask which fits. Keep the tone investigative. The record is a tool for finding a repair, not a list of personal defects.

Keep useful failures

A failed attempt can reveal the boundary of the learning. A student may use evidence well in a story but treat any numerical statement as proof in a persuasive article. That tells you where the concept still needs work. Deleting the failure would make the learning record less accurate.

Similarly, a group may agree on the meaning of draft but still send a file without saying what feedback is wanted. The vocabulary helped with status but did not solve the request. The next experiment can examine the difference between “Please review” and a more specific request to check claims, organisation or spelling.

Keep the original response alongside the revision when possible. The comparison lets you see exactly what changed. A neat final answer alone can conceal the reasoning that produced it and the assistance that made it possible.

End with one next action

Every record should lead somewhere. “Practise vocabulary more” is too broad. “Try three new boundary cases tomorrow, including the exact limit” is useful. So is “Ask a partner to restate the handover request before revising the labels.”

Choose the smallest action that addresses the uncertainty you discovered. This keeps the experiment from becoming paperwork. You are building a cycle of noticing, teaching, applying and checking. The record supports that cycle by preserving what you would otherwise forget.

Contents · Previous chapter · Next chapter

Read this invented scene: “When Jo entered the classroom, the conversation stopped. Adrian turned a sheet of paper face down. Jo placed her bag on the floor and sat by the window.” What can you say with confidence, and what remains an interpretation?

You can report that the conversation stopped when Jo entered and that Adrian turned over a sheet of paper. You might infer that the group was hiding something from Jo. You might also imagine that they had finished discussing a private matter, were preparing a surprise or simply noticed the class beginning. The scene does not establish a single motive.

The first attempt

Ask the learner to write two sentences explaining what happened. A response might say, “They were gossiping about Jo. Adrian hid the evidence because he did not like her.” This answer contains a possible interpretation, but it presents several unsupported details as settled facts.

Do not begin by banning inference. Reading often requires it. The useful distinction is between what the text directly supplies and what the reader constructs from it. Teach observation as a report of something noticed or recorded, and inference as a conclusion drawn from information. In a fictional reading task, the “observation” is usually a detail explicitly provided by the text.

Make two columns. Put “The conversation stopped” in the first. Put “They may have been discussing something private” in the second. Ask what connects them and what other explanation remains possible. The columns make the roles visible without requiring a sophisticated vocabulary list.

The next task

Use a different invented scene: “Mira read the message twice, put her phone in her bag and walked towards the teacher’s desk. Her unfinished worksheet remained open on the table.” Ask the learner to identify two explicit details and one reasonable inference, then state what cannot yet be known.

A sound response might say that Mira may want to ask the teacher about the message. Another reasonable response might suggest that the message reminded her of something urgent. The learner should recognise that the text does not prove the message was upsetting, that the teacher sent it or that Mira intended to leave school.

The success criterion is not whether the learner guesses the author’s imagined story. It is whether the explanation distinguishes evidence from interpretation and keeps uncertainty proportionate. Several answers can satisfy that criterion. Asking for the relevant detail helps you judge the reasoning without insisting on one unsupported conclusion.

What to record

Record whether the learner quotes or paraphrases a relevant detail, labels the inference appropriately and avoids adding invented facts. Also record whether the two-column support was available. A learner who performs well with the columns may still need practice noticing the distinction in an ordinary paragraph.

Suppose the learner writes, “Mira definitely received bad news because she read the message twice.” The evidence is relevant to attention, but it does not establish bad news. Ask for another possible reason to reread a message. The learner might suggest confusion about a time or a pleasant surprise. That contrast reveals why definitely is too strong.

A useful revision would be: “Mira’s repeated reading suggests that the message mattered to her, although its content is not given.” Even this interpretation should be discussed rather than treated as a compulsory sentence. The purpose is to make the relationship between detail and claim inspectable.

Try it beyond a story

Later, use a neutral everyday example: “The classroom clock shows 8:10, and the teacher has not arrived.” Which statements are observations? Which are inferences? “The teacher is late” depends on the scheduled start time. “The teacher forgot” introduces an explanation for which the observation supplies little support.

This transfer task changes the setting while preserving the distinction. It helps you see whether the learner has memorised a story exercise or gained a useful question: “What do I know, and what am I adding?” Avoid turning the exercise into speculation about real classmates’ private motives. Invented situations keep the attention on reasoning.

For a younger learner, replace written scenes with a picture and allow oral answers. For an older learner, add a second piece of evidence that changes which inference is most reasonable. In both versions, inspect the same underlying skill: keeping the source of a claim visible.

Contents · Previous chapter · Next chapter

A student writes, “Our school should have more reading corners because reading corners are good.” The sentence has a clear preference, but its reason repeats the preference in different words. Adding an impressive adjective would leave the problem intact. The useful learning concerns the roles of claim, evidence and explanation.

Use a fictional decision so that everyone has access to the same information. A class is choosing between a reading corner and an extra display board. It has a small unused area, several boxes of books, and a weekly period during which students may choose quiet activities. The teacher also reports that some students prefer drawing or puzzles. These are the facts available for the exercise.

Teach the roles through the actual task

A claim states what the writer wants the reader to accept. Evidence is relevant information offered in support. An explanation makes the connection clear. These brief working descriptions are enough to begin. The learner can refine them as tasks become more demanding.

A possible claim is: “A reading corner would be a useful option during the quiet-activity period.” The existing books and available space are relevant evidence about feasibility. An explanation might say that the class could create the corner without first obtaining a large new collection or finding another room.

Notice what the evidence does not establish. It does not show that all students prefer reading, that a reading corner will improve everyone’s attainment or that the display board has no value. A careful argument can recommend an option while keeping those limits visible.

The vocabulary experiment

First ask for a recommendation using the supplied facts. Save that response. Then teach the three roles and show a short worked paragraph on a different classroom choice. Ask the learner to write a fresh recommendation from another small set of facts, perhaps choosing between a shared stationery tray and an additional noticeboard.

Inspect whether the later paragraph states a clear recommendation, uses a relevant fact and explains why that fact matters. Do not award success merely because the words claim and evidence appear. The labels can be used in planning and disappear from the finished paragraph while their organising function remains.

You can use three colours to mark the roles, but treat the colours as a teaching support. On a later task, ask the learner to identify the roles without colour prompts. If the learner can do so, the distinction is becoming available beyond the original worksheet.

A worked revision

Initial version: “We should choose a reading corner because it would be amazing and everyone would love it.”

Revised version: “A reading corner would give students another quiet activity during the weekly choice period. The class already has spare books and a small unused area, so the idea could begin on a modest scale. Because some students prefer drawing or puzzles, the corner should remain one option rather than replacing the other activities.”

The revision is stronger because its claims match the supplied information. It identifies a practical reason and acknowledges a relevant difference among readers. The improvement comes from the structure of the reasoning, not from the length or rarity of the vocabulary.

Ask the learner to explain which sentence changed most and why. If they say, “I made it sound better,” invite a more precise answer. They might say, “I removed the claim about everyone because the facts did not support it.” That explanation shows a usable distinction.

When the experiment reveals missing knowledge

A learner may understand all three labels but still struggle to write an argument because the information is too thin. If the class has no idea how many students would use either option, the learner cannot honestly invent that evidence. The appropriate next step is to identify the unanswered question.

This is a valuable result. Vocabulary has helped reveal a gap in knowledge. A student who writes, “We would need to ask classmates which activities they use” is doing something useful, even if the paragraph remains provisional. Rewarding invented certainty would undermine the purpose of the exercise.

For transfer, ask the learner to inspect a short fictional advertisement with a claim and weak support. Then ask what evidence would be relevant. This task checks whether the learner can use the distinction as a reader as well as a writer. Keep the advertisement harmless and invented so that the focus remains on the reasoning.

Contents · Previous chapter · Next chapter

Some of the most powerful vocabulary in mathematics consists of ordinary words. At least, at most and exactly determine which values a condition permits. A learner who treats them as interchangeable may perform the arithmetic correctly and still answer the wrong question.

Begin with a visible collection. Place eight counters on a table. Ask the learner to show a group containing at most five counters. Then ask whether four, five and six are allowed. A learner who understands the condition should accept four and five and reject six. The boundary value is particularly informative.

A first set of tasks

Use this invented club rule: “A game needs at least three players and at most six players.” Ask whether groups of two, three, five, six and seven can play under the rule. Ask for an explanation of three and six, because those values test whether the learner includes the boundaries.

The permitted groups have three, five or six players among the listed options. Two is below the minimum; seven is above the maximum. The rule also permits four, although four was not included in the options. Asking for an unlisted permitted value checks whether the learner understands the range rather than memorising a list.

Now change the rule to “The game needs exactly four players.” The permitted number is four. A group of three does not qualify merely because it is small, and a group of five does not qualify merely because it has enough people to form a group of four. The task concerns the stated group size.

Teach the distinction with a boundary

For at least, identify the minimum and show that the minimum itself is included. For at most, identify the maximum and include it. For exactly, specify the one required quantity. Use words, objects and a number line together if that helps the learner connect the representations.

Ask the learner to invent an example for each phrase. “Bring at least two pencils” is different from “Bring exactly two pencils.” More than two satisfies the first instruction but not the second. Discuss why an organiser might choose one instruction rather than the other. The wording expresses a purpose as well as a numerical relationship.

A useful contrast is more than versus at least. More than five excludes five; at least five includes five. Likewise, fewer than five excludes five; at most five includes it. Introduce these neighbours after the first distinction is stable, so that the experiment remains manageable.

The comparison task

Use a new context with similar numerical demand: “A display can contain at least two photographs and at most seven photographs.” Offer groups containing one, two, six, seven and eight photographs. Ask which groups satisfy the rule and why.

Record boundary errors separately from other mistakes. If the learner accepts six but rejects two and seven, the main difficulty may be inclusion at the limits. If the learner accepts every number larger than two, the upper limit may have been overlooked. Those are different patterns and need different practice.

Then give a condition without multiple-choice options: “Write three possible whole-number answers to a task that allows at most four.” Clarify whether zero is permitted in the scenario. If the task is counting items that must be present, a minimum may need to be stated. The language and the situation together define the answer space.

A transfer check with two conditions

Try: “Choose an even whole number that is at least four and at most ten.” The permitted numbers are four, six, eight and ten. A learner who gives five may have satisfied the range while overlooking even. A learner who gives twelve may have used the even-number condition but ignored the maximum.

This task introduces coordination between conditions. It is a useful next step, but record that the demand has increased. Failure here does not erase earlier understanding of at most. It may reveal difficulty combining requirements.

Finally, ask the learner to explain the rule to someone else without giving the answers. If the listener can generate valid examples and reject invalid ones, the explanation is doing useful work. The vocabulary now supports reading, reasoning and communication within a clearly defined mathematical task.

Contents · Previous chapter · Next chapter

Some distinctions become easier when you begin with an invented rule rather than a complicated real-world argument. Imagine a puzzle club with this rule: “To enter the blue room, a player must have a blue card and know the password.” Having a blue card is necessary under the rule, but it is not sufficient. The card alone does not satisfy both requirements.

Having the card and knowing the password is sufficient for this invented entry rule, provided there are no additional conditions. The phrase under this rule matters. In real situations, a short description may omit relevant conditions. A vocabulary experiment can use a simplified world to teach the relationship without pretending that every real decision is equally simple.

Start with examples before labels

Show four players. Ben has the card and knows the password. Aisha has the card but does not know the password. Clara knows the password but has no card. Ethan has neither. Ask who can enter and why.

Ben qualifies. Aisha and Clara each satisfy one requirement but not the other. Ethan satisfies neither. If the learner can explain these cases, introduce the labels. A necessary condition must be present for the outcome under the stated rule. A sufficient condition is enough to establish the outcome under that rule.

These descriptions need examples because the words can otherwise become empty vocabulary. Ask, “Can the outcome happen without this condition?” to investigate necessity. Ask, “Does this condition guarantee the outcome in the rule we have been given?” to investigate sufficiency.

Use a second rule to avoid a memorised pattern

Now imagine a different activity: “A player may join the green game if they have either a green card or a yellow card.” A green card is sufficient under this rule, but it is not necessary, because a yellow card also qualifies. The learner must adjust to the changed relationship.

This contrast is important. If every example uses two required conditions, the learner may memorise that a single condition is always necessary but never sufficient. The second rule shows why the relationship depends on the actual rule.

Ask the learner to design a new rule and explain one necessary condition and one sufficient condition. They may choose a drawing challenge: “A valid entry must contain a circle and a triangle.” The presence of a circle is necessary; having both shapes is sufficient for the requirements stated. If the learner adds a hidden condition later, pause and make it explicit.

Test a common mistaken inference

Give the statement: “Every completed task in this activity has a tick.” Does seeing a tick prove that a task is completed? Not necessarily, unless the rules also say that ticks are given only to completed tasks. The initial statement establishes one direction. It leaves the reverse direction open.

For some learners, this is an appropriate extension. For others, it adds too much logical demand to the first lesson. Use it after the simpler card examples are secure. Record whether the learner is working with the everyday meaning of a tick or the exact invented rule. Both may be reasonable interpretations in different contexts, which is why the stated conditions matter.

A correct response should identify the missing information rather than merely say “No”. For example: “The rule tells us that completed tasks receive ticks, but it does not tell us whether unfinished tasks can also receive them.” That explanation demonstrates the distinction more clearly than repeating the words necessary and sufficient.

Transfer without overgeneralising

Use a school planning example: “Reading the instructions is necessary for following them accurately.” Even this everyday statement deserves care: someone could happen to act correctly without reading. The intended advice concerns a dependable method, while the logical wording makes a stronger claim. Ask the learner to revise it to fit the purpose: “Read the instructions before beginning so that you can check the requirements.”

This moment shows a valuable boundary. Technical vocabulary can sharpen reasoning, but it can also encourage overstatement if applied casually. The experiment should teach the learner to inspect the claim, not simply insert the formal terms.

Keep the first conclusion narrow: “The learner distinguished necessity and sufficiency in the two invented rule systems.” Then try another clearly specified situation. A successful puzzle task is a promising sign of understanding; it is not evidence that every complex argument will now be handled correctly.

Contents · Previous chapter · Next chapter

A learner may know a topic and still answer a different question from the one asked. The difficulty can sit in the task verb. Describe, explain, compare and evaluate invite different kinds of work. Their precise expectations also depend on the subject, the surrounding question and the teacher’s guidance.

Use the following activity as a language exercise rather than an official examination marking scheme. The aim is to connect each verb with a useful action and then check that action against the full task. There is no universal formula that replaces reading the question.

Keep the content simple

Give the learner two invented study spaces. Room A has individual desks, bright lighting and frequent corridor noise. Room B has shared tables, softer lighting and a quiet location. Both are available for the same period. The learner has enough information to make simple comparisons without needing specialist background knowledge.

Ask four questions: “Describe Room A”; “Explain why corridor noise might make a listening task difficult”; “Compare the rooms for a group discussion”; and “Evaluate which room would suit silent reading, using the information given.” Save the answers before discussing the verbs.

The first response should identify relevant characteristics. The second should connect a feature to a possible effect, with appropriate caution. The third should examine similarities or differences relevant to the stated purpose. The fourth should make a reasoned judgement using criteria, while acknowledging information that is missing.

Teach actions rather than slogans

For describe, ask: “What are its relevant features?” For explain, ask: “How or why are these things connected?” For compare, ask: “Along which common dimensions will I examine the options?” For evaluate, ask: “What criteria matter, what evidence is available, and what judgement does that support?”

These prompts are starting aids. A description can include relationships, and an explanation may require description. An evaluation can contain both. The verbs do not divide writing into sealed boxes. They indicate the main job the answer must accomplish.

Ask the learner to point to the sentence that performs the main job. If an “explanation” only lists features, the missing relationship becomes visible. If an “evaluation” simply announces a favourite, the missing criteria become visible.

A worked comparison

Weak response to “Compare the rooms for a group discussion”: “Room A has desks. Room B is quiet. Room A is bright. Room B has tables.”

More useful response: “Room B’s shared tables may make it easier for group members to face one another and share materials. Room A’s individual desks may need rearranging for the same purpose. Room B is also described as quiet, while Room A has corridor noise, so Room B appears more suitable for a discussion that requires everyone to hear clearly.”

The stronger response uses common dimensions: arrangement and noise. It connects both rooms to the purpose of group discussion. The vocabulary has guided an operation on the information, rather than merely changing the adjectives.

The new task

Provide two invented ways to organise a class book collection. One sorts books by author; the other sorts by broad topic. Ask the learner to describe one system, explain a benefit, compare both for a stated reader and evaluate a choice under a clear condition.

Look for transfer of the task action. Does the learner compare the systems using common criteria, or return to disconnected lists? Does the judgement depend on the reader’s purpose? Can the learner say what further information would change the recommendation?

If the learner uses the verb correctly only when the prompt card is visible, retain the card while practising and then gradually reduce reliance on it. The goal is to recognise the job from the full question. Removing all support immediately can conceal progress by increasing several demands at once.

Bring it back to schoolwork

Choose one actual assignment and ask the learner to underline the task verb, circle the subject of the question and mark any limiting condition. Then compare the planned answer with the teacher’s instructions. This final check respects the local expectations of the course.

A useful record might say: “The learner now distinguishes listing features from explaining a relationship, but needs help choosing comparison criteria.” That points to a teachable next step. “Needs better vocabulary” would leave the real difficulty hidden.

Contents · Previous chapter · Next chapter

A group is planning a class display. One student says, “It has to be colourful.” Another says, “It has to fit on the table.” A third says, “It has to be ready tomorrow.” The same phrase, has to, is being used for requirements of different kinds. Until the group separates them, discussion may circle without a decision.

For this invented exercise, the teacher has provided two firm conditions: the display must fit on one table and must be ready by the next afternoon. The group would like it to be colourful, interactive and easy to carry. These preferences matter, but they may be adjusted if they conflict with the firm conditions.

Introduce three useful labels

A constraint limits the available options under the current task conditions. A preference expresses something a person or group would like. A trade-off arises when improving one valued feature involves giving up something else. These are working descriptions for planning, not a claim that every real-world boundary is fixed forever.

The labels help the group ask better questions. Who set the condition? Is it negotiable? What happens if we choose an option that violates it? Which preferences matter most? What would we accept losing to gain something else?

Give the group two proposals. Proposal A is a large moving model that would take several days. Proposal B is a compact board with lift-up explanation cards that can be completed within the available time. The group may prefer the ambition of A, but under the stated conditions it is not currently feasible.

Run the first attempt

Before teaching the labels, ask the group to choose an option and explain the decision. Listen for whether they distinguish the conditions from their wishes. A discussion dominated by “I like A” and “B is boring” may reveal that no shared decision structure is available.

After introducing the words, ask the group to place each feature under a heading: required, preferred or uncertain. “Uncertain” is important because the group may not know whether a condition is genuinely fixed. Instead of guessing, they can identify a question for the teacher.

Ask them to revise the recommendation. A useful response might be: “We choose B because it meets the size and time requirements. We will make the lift-up cards colourful to preserve part of our preferred design. If the deadline changes, we can reconsider a moving component.”

What counts as improvement?

Do not judge success solely by whether the group chooses B. A group could choose B by chance or because one person insists. Inspect whether the explanation identifies the constraints, acknowledges a preference and states the relevant trade-off.

Also watch whether the language improves participation. Does a quieter student gain a way to question an assumption? Does someone ask whether “colourful” is a requirement or a preference? A useful word can make the structure of a disagreement easier to discuss.

However, the labels can also be misused. A forceful student may call a personal preference a constraint to stop discussion. Ask for the source of each supposed requirement. Vocabulary supports judgement only when people apply it honestly.

A changed-condition task

Now change the fictional brief. The display has two weeks, but it must be carried by two students and set up in five minutes. Ask whether the original recommendation should remain. The learner should revisit the decision rather than treat the earlier choice as a permanent answer.

This task checks whether the words are functioning as tools. A memorised conclusion says “B is always better.” A usable distinction says “The relevant constraints have changed, so the comparison may change.” The reasoning remains stable while the answer adapts.

For an individual study example, ask a learner to plan a revision session with a fixed end time, two required topics and a preference for colourful notes. The task is not to shame the preference. It is to decide how the preference fits within the purpose and available time.

Keep values visible

Some preferences express important values, such as including every group member or making a display readable from a reasonable distance. Calling something a preference does not make it trivial. The label describes its role in this decision; it does not assign its moral worth.

End by asking, “What did the vocabulary help us see?” A strong answer might be: “We were arguing about the design before agreeing on the conditions.” That is the mechanism worth carrying into the next task.

Contents · Previous chapter · Next chapter

Four students are preparing a short presentation. One says, “My part is done.” Another opens the file and discovers notes, unfinished sentences and a question for the group. The first student meant that the initial writing was finished. The second understood “done” to mean ready to present. Both used a familiar word, but the shared task remained unclear.

A useful vocabulary experiment can examine whether more precise status words improve the handover. The target is a small coordination problem: when a person sends work to someone else, can the recipient identify its status and the next required action?

Observe one handover

Give a group an invented assignment to make a two-minute introduction to a class exhibition. Each member prepares one short section. Before teaching status labels, ask them to exchange work with a brief message. Save the messages and ask recipients what they believe they should do next.

Look for disagreement. “Here is my part” may be understood as a request for feedback, a notice that the work is finished or an invitation to rewrite it. A message can sound friendly and still leave the action uncertain. Do not judge it by tone alone.

Now introduce a small local vocabulary. Draft means an initial version open to revision. Reviewed means that a specified check has been completed. Complete means that the agreed requirements for this task have been met. Each label needs a scope. Reviewed for spelling does not mean reviewed for factual accuracy.

Use a status plus an action

Ask the sender to write a message such as: “This is my draft. Please check whether the explanation is clear to someone who has not seen the display. I will correct the wording after your comments.” The message describes both the state of the work and the recipient’s job.

For another handover: “The group has checked the order and timing. Please practise your section using this version. Tell us if you find a factual problem before we print the cue cards.” The word checked is connected to specific checks, and the remaining route for correction stays open.

Ask recipients to restate the next action in their own words. This is a practical check of coordination. If the recipient says, “You want me to check facts,” while the sender wanted comments on clarity, the mismatch becomes visible before unnecessary work begins.

What to measure

Count whether each recipient identifies the intended action correctly. Record clarification questions and misunderstandings. You can also ask whether the message contains enough information to begin without guessing. These observations are more relevant than the number of status terms used.

A message that says “Draft reviewed final complete” is not useful simply because it includes the target vocabulary. The terms must describe a coherent state. A piece of work can be a draft that has been reviewed for one purpose; it cannot become complete merely because someone attaches a confident label.

Use a fresh handover for the comparison. If the group repeats the original exchange, members may already know what was intended. A new short task, such as preparing captions for a display, gives a better view of whether the language is transferable.

Inspect the failure modes

The labels may be too numerous. A team that invents ten statuses for a small project can spend more effort maintaining the system than completing the work. Start with the distinctions that address actual confusion.

The labels may also be used inconsistently. One student may mark work complete after writing it; another may wait until it has been read aloud. The repair is to discuss the task criteria and apply them to one example together.

Finally, a label can be accurate and the file can still be wrong. Status vocabulary makes the process easier to inspect; it does not eliminate mistakes. Keep a route for reporting a problem after a review. Good coordination includes the ability to correct the shared record.

The learning is visible when a group can say, “We misunderstood the stage of the work, so we added a status and a specific request.” That explanation identifies a mechanism the group can reuse in another collaboration.

Contents · Previous chapter · Next chapter

A student changes “The character was unhappy” to “The character was devastated” because the second word sounds more advanced. The sentence becomes more intense, but does it become more accurate? That depends on the text. Vocabulary growth should increase choice, including the choice to keep a simpler word.

This experiment examines whether a learner can select a word that fits the evidence and intended tone. The target is precision. You are not testing how many unusual words the learner can insert into a paragraph.

Build a small set of contrasts

Use words such as disappointed, frustrated, reluctant, uncertain and distressed. Explain them with short situations. A person may be disappointed when an expected outcome does not happen. They may be frustrated when progress is repeatedly blocked. They may be reluctant to act because they feel unwilling or hesitant.

Acknowledge overlap. Real emotions can coexist, and a short scene may support several descriptions. The useful question is whether a proposed word has evidence and whether it claims more than the scene supplies.

Give this invented sentence: “Ethan looked at the cancelled match notice, sighed and put his boots back in the bag.” Disappointed is plausible. Devastated might overstate the evidence. Reluctant would need a clearer action he was hesitant to take. Ask the learner to explain each judgement.

Use a revision task

Provide a short paragraph with three vague descriptions. For example: “The plan was bad. The group felt bad. The final result was good.” Supply the surrounding facts: the plan required materials the group did not have, members became impatient after repeated delays, and the revised display clearly communicated its main idea.

A useful revision might call the first plan impractical under the available conditions, describe the group as frustrated by the delays, and say that the final display was clear and informative. The improved wording distinguishes different problems and achievements.

The learner should also be allowed to use a phrase. “Required materials we did not have” may be clearer for the audience than impractical. A vocabulary experiment should not force a single-word replacement when a short explanation serves the reader better.

Compare fit, not sophistication

Ask a reader to match each revised description to the supplied facts. Does the word point to the relevant feature? Does it introduce an unsupported judgement about a person? Does its intensity fit the evidence? These questions make precision inspectable.

For instance, replacing “The group was slow” with “The group was incompetent” is not a neutral upgrade. It turns an observation about pace into a broad judgement about ability. A more precise account might say that the group spent fifteen minutes waiting for a missing file. The second account identifies a problem that can be addressed.

Keep the comparison task similar in length and complexity. If the later paragraph gives much richer facts, the learner has more support for choosing precise language. Record that difference rather than assigning all improvement to the new words.

Try a counterexample

Ask the learner to invent a scene where reluctant fits but unable does not. Then reverse the task. A child who can join a performance but hesitates because the audience is unfamiliar may be reluctant. A child who wants to join but is away on the day may be unable.

Discuss how both could apply in a different situation. The purpose is not to draw an absolute border around every emotional state. It is to show that the words carry different information, and that a writer should choose intentionally.

For transfer, revisit the learner’s own writing and select just one vague sentence. Ask what the reader needs to know and what evidence the writer has. A strong revision may be shorter than the original. Precision is an improvement in the match between meaning and expression, not a requirement to make every sentence longer.

Contents · Previous chapter · Next chapter

Small words can control large actions. “Submit the form if you are attending” differs from “Submit the form unless you are attending.” The vocabulary is familiar, but the logical relationship changes. An experiment with conditional instructions can reveal whether a learner follows the relation or simply reacts to the main action word.

Use harmless invented tasks involving coloured cards or paper shapes. This keeps the focus on language and avoids consequences for a misunderstood real instruction. Ask the learner to explain the action before performing it.

Begin with one condition

For Activity A, give a complete sorting rule: “Put starred cards in the box and leave every other card outside.” Then describe the positive condition with “Put the card in the box if it has a star.” Present a card with a star and one without. The full activity rule places the starred card inside and the other outside.

For Activity B, reverse the complete sorting rule: put cards without stars in the box and leave starred cards outside. In this context, “Put the card in the box unless it has a star” uses the star as the exception. Ask the learner to restate the positive condition: “Put it in if it does not have a star.”

These activities state what happens in both cases. A standalone sentence containing if can supply a condition without settling every other possibility. Read the complete rule before assuming that an unmentioned case is forbidden. This keeps the exercise consistent with the distinction between necessary and sufficient conditions.

A correct restatement is useful evidence. It shows more than following a gesture or copying another person. If the learner acts correctly but explains the reverse condition, slow down and inspect the mismatch.

Introduce only with care

“Put only the blue cards in the box” limits which cards may be placed there. “Only put the blue cards in the box” can be less clear in ordinary conversation because the position of only may leave its scope uncertain. For teaching, choose wording that makes the intended restriction clear.

Ask the learner what the word is limiting. Is it the colour, the action, the person or the location? Compare “Only Ben may move the cards” with “Ben may move only the blue cards.” The first restricts the actor. The second restricts the cards Ben may move.

This is a productive place to examine how word order affects meaning. The learner may know every word in both sentences and still need to inspect their relationships. Vocabulary knowledge operates inside sentence structure.

A task with an exception

Give the instruction: “Place all the circles on the sheet except the red circle.” Use a blue circle, a green circle, a red circle and a red square. The blue and green circles belong on the sheet. The red circle is excluded. The red square is not a circle, so it is outside the initial group as well.

Ask why the red square was not included. If the learner says “because red is forbidden”, test that interpretation with a blue square. The real rule first selects circles and then excludes one kind of circle. The example reveals how the conditions are combined.

Do not add many conditions at once. A long instruction with colour, shape, size, order and exceptions may primarily test memory for a sequence. That can be a worthwhile task later, but it changes the question you are investigating.

Run the comparison

For the first attempt, use two short instructions and ask the learner to explain them. Teach one misunderstood relationship with a pair of contrasts. Then use two new instructions with the same structure and different objects.

Record errors by scope: Did the learner overlook the exception? Reverse the condition? Restrict the wrong part of the sentence? Forget one of several requirements? This detail helps you choose the next example.

A learner who succeeds only when the key word is highlighted may need practice locating it in an unmarked instruction. A learner who understands orally but struggles in print may need support with reading the sentence. The same final action can conceal different routes to the answer.

Transfer to ordinary study

Choose a short teacher-written instruction and ask the learner to mark its main action and conditions. Check the interpretation against the original context. If the instruction is ambiguous, the useful response is a clarification question, not an increasingly elaborate guess.

One good question is: “Does this exception apply to the whole task or only to this part?” Another is: “When you say only, which choice are you restricting?” The vocabulary experiment has now produced a practical communication skill: recognising when the words need a shared interpretation before action begins.

Contents · Previous chapter · Next chapter

Knowing a familiar word can sometimes make reading harder if you apply the first meaning that comes to mind. A learner may understand volume as loudness and become confused when a mathematics question asks for the volume of a container. The word is recognised, but its meaning has not been matched to the context.

This experiment asks whether a learner can select and justify the relevant meaning. It is especially useful when someone says, “I know every word in the sentence, but I do not understand it.” The difficulty may concern relationships or specialised meanings rather than completely unfamiliar vocabulary.

Start with controlled contrasts

Use three sentences: “Please lower the volume”; “Find the volume of the box”; and “She borrowed the second volume of the history.” Ask the learner to explain the word in each sentence and point to the clue that supports the interpretation.

The first sentence commonly concerns loudness, though additional context could clarify the device or sound. The second concerns three-dimensional space. The third concerns a book or part of a larger work. A learner who supplies the same definition every time has revealed the target difficulty.

Teach a short checking routine: identify the topic, inspect the nearby words, choose a possible meaning, and reread the whole sentence to see whether it makes sense. If uncertainty remains, consult a suitable dictionary or ask someone who knows the subject. Context gives clues; it does not guarantee a correct guess.

Use a word whose meanings are less obvious

Try charge: “The shop will charge a small fee”; “The battery holds a charge”; “The team put Jo in charge of the display.” The surrounding phrases guide different interpretations. Avoid adding legal or technical details that exceed the purpose of the exercise.

Ask the learner to explain why another familiar meaning does not fit. In “put Jo in charge”, the phrase concerns responsibility, not a price. That contrast tests control more deeply than selecting the correct answer from a familiar list.

Then ask for a new sentence that preserves one meaning while changing the setting. A learner might move from a school display to a family game. If the phrase remains appropriate, the meaning is becoming flexible within that use.

Run a fair comparison

Before teaching, use a short set containing familiar words with different meanings. Afterwards, use a fresh set of similar difficulty. Do not teach every word that appears in the second set and then interpret success as a general context-reading skill. That would primarily show learning of those particular examples.

Record both the selected meaning and the clue. A correct guess without a plausible explanation may need another check. An incorrect guess based on a reasonable but incomplete clue can reveal exactly where the learner needs to widen the context.

For example, “The plant is beside the river” could refer to a living organism or an industrial facility. The sentence alone may not resolve the ambiguity. A strong reader should be allowed to say that more context is needed. Insisting on one answer would teach false certainty.

Notice when a subject changes the vocabulary

A word such as significant can mean important in everyday discussion and carry a more specific meaning in statistical writing. A learner should notice that technical contexts may require checking a subject-specific definition. This guide does not use a vocabulary exercise to teach the whole technical concept.

Similarly, work in everyday conversation and in physics is not always interchangeable. Recognising the possibility of a specialised meaning is a useful first step. The next step is to learn the subject concept from an appropriate source or teacher.

For transfer, ask the learner to collect one familiar word used differently in two school subjects. Have them write the sentence, the meaning in context and the clue that helped. The collection should preserve the sentence rather than reduce the word to an isolated translation.

The experiment succeeds when the learner begins asking, “Which meaning is working here?” That question turns recognition into interpretation. It also reduces the temptation to treat the first dictionary entry as a universal answer.

Contents · Previous chapter · Next chapter

A learner can choose words with roughly the right meanings and still produce an awkward expression. “Do a decision” is understandable, but “make a decision” is the conventional combination. Another learner may write a grammatically correct request that sounds too abrupt for its setting. These problems involve different aspects of usable vocabulary.

Collocation concerns words that conventionally occur together. Register concerns how language fits a situation, purpose and audience. This experiment separates the two so that feedback becomes more precise.

Choose a small communication task

Ask the learner to write a short message requesting feedback on a project paragraph. Provide the audience and purpose: the message is to a classmate who has agreed to help, and the writer wants comments on whether the explanation is clear.

A first attempt might say, “Do feedback on my paragraph now.” The issue is not only politeness. The word combination is awkward, the kind of feedback is unspecified and the timing may not fit the agreement.

Offer examples such as “give feedback”, “offer a suggestion”, “raise a concern” and “make a revision”. Then discuss a possible message: “Could you give me feedback on whether this paragraph explains the idea clearly? Please mark any sentence you had to reread.” The wording directs the reader towards a specific action.

Keep choices connected to the audience

The same request may be phrased differently to a close friend, a teacher or a group working under a tight shared deadline. There is no need to rank every formal expression as better than every informal one. The useful question is whether the language fits the relationship and task.

For a familiar peer, “Can you check whether this explanation makes sense?” may be entirely suitable. In a more formal setting, the learner may choose a fuller request. Adding several elaborate courtesy phrases can make a simple request harder to understand.

Ask the learner what the recipient needs to know. Usually this includes the object to inspect, the kind of help wanted and any relevant timing. Vocabulary supports the request by making those elements clear.

Compare two outcomes

First inspect the expression itself: are the combinations conventional, and does the tone fit the stated relationship? Then inspect the recipient’s interpretation: can they identify the requested action? A message can be linguistically polished while still being vague.

Use a new communication task for the next attempt, such as asking someone to check the order of slides. Keep the social relationship similar so that the learner is practising the language without also having to infer a very different convention.

Ask the recipient to restate the request before acting. If the message asks for comments on clarity but the recipient corrects only spelling, discuss whether the task was specified clearly enough. That disagreement creates a useful opportunity to refine the wording.

A worked contrast

Message A: “Please improve this.”

Message B: “Please tell me which sentence makes the main idea hardest to understand. I will revise the wording myself.”

Message B makes the requested help and division of work more explicit. It does not assume that a helper will rewrite the paragraph. The difference is practical: the vocabulary changes what the other person is being asked to do.

Now consider whether Message B is too detailed for a particular exchange. If two classmates have already agreed on a review routine, a shorter message may be enough. Shared context can carry information that would otherwise need to be written. The experiment should respect that context rather than demand maximum explicitness everywhere.

Use a dictionary as a source of patterns

When checking a word, inspect example sentences and grammatical patterns as well as a brief definition. For a word such as responsible, the relationship with for may matter. For reluctant, an infinitive pattern such as “reluctant to speak” may be useful.

Practise the pattern in an original sentence, then ask whether the meaning still fits. Copying a polished example gives you a model; changing the situation tests whether you understand how the model works. Keep the final aim visible: a reader or listener should understand the intended meaning and action with reasonable effort.

Contents · Previous chapter · Next chapter

A learner writes, “The character is certainly hiding something.” The text shows only that the character turned away when a question was asked. The claim may be possible, but the degree of certainty is too strong for the available evidence. Vocabulary can help the learner express the difference.

This experiment concerns matching the strength of a statement to what is known. It does not assign universal numerical probabilities to ordinary words such as likely. People and contexts may use those words differently. When an exact probability matters, the numbers and their basis must be stated rather than assumed.

Begin with an invented information set

A class is choosing whether to hold a paper-aircraft activity indoors or in a sheltered outdoor area. The only information supplied is that the outdoor area is currently empty and that wind has disturbed similar paper activities there before. Ask the learner to write what is known, what is plausible and what remains uncertain.

“The outdoor area is empty now” is supplied information. “Wind may affect the aircraft activity” is a possibility supported by the history. “The activity will definitely fail outdoors” goes beyond the information. The learner should be able to state the uncertainty without becoming unable to make any decision.

For this exercise, a reasonable recommendation might be to use the indoor area if the aim is to compare aircraft designs under steadier conditions. The recommendation rests on the task purpose and known sources of variation. It need not pretend that failure outdoors is certain.

Teach a ladder without pretending it is a measuring instrument

Place phrases such as “could be”, “the evidence suggests”, and “the information establishes” along a rough continuum of commitment. Explain that the exact wording must fit the case. The ladder is a discussion aid, not a validated scale.

Give pairs of statements and ask which is better supported. “She may have forgotten” is weaker than “She forgot”. “The results are consistent with this explanation” is weaker than “The results prove this explanation”. For advanced learners, discuss why several explanations can be consistent with the same observation.

Then ask for the missing evidence that would justify a stronger claim. This shifts the activity from adding cautious words mechanically to examining what would change the judgement.

Compare revisions

Provide a paragraph containing three overconfident statements about an invented school event. Ask the learner to revise each one while preserving useful information. A weak revision adds maybe to every sentence. A stronger revision distinguishes established details from uncertain interpretations.

For example, “All students disliked the activity because three students complained” could become: “Three students complained about the activity. Their comments show some dissatisfaction, but they do not establish how the whole class felt.” The revised wording identifies the scope of the evidence.

Inspect whether the learner has weakened a claim appropriately or weakened it so much that it becomes uninformative. If a fictional notice clearly states that a room is closed on Tuesday, “The room might perhaps be closed” is unnecessarily hesitant within the exercise. Calibration includes justified confidence.

Transfer to a prediction

Ask the learner to predict which of two invented project plans is more likely to finish on time, using a short description of required tasks and available resources. After the fictional outcome is revealed, compare the reasoning rather than judging only whether the prediction happened to be correct.

A well-supported prediction can turn out wrong because uncertain events occur. A poorly supported prediction can turn out right by chance. The vocabulary experiment should help the learner separate the quality of the reasoning from a single outcome.

Record the evidence available at the time of the judgement. This prevents the later outcome from making the earlier decision seem more obvious than it was. The useful habit is to state what you know, what you infer and what would cause you to revise the claim.

Contents · Previous chapter · Next chapter

A word can feel secure while the explanation is still in front of you. The real test may arrive later, when the sentence has changed and the prompt is gone. A retention check asks what remains available after a gap. It should be planned gently enough that forgetting becomes useful feedback rather than a surprise punishment.

Begin by distinguishing recognition from retrieval. Recognition occurs when the word or meaning is present and you identify it. Retrieval requires bringing the information back with less support. Both are useful, but they make different demands.

Build a small retrieval task

After teaching constraint, close the explanation and ask the learner to describe the idea in ordinary words. Then ask for a constraint in a new fictional task. A learner who says “something that limits our options” and identifies the fixed table size has done more than recognise a matching definition.

Allow correction after the attempt. If the learner confuses a preference with a constraint, reopen the example and discuss the difference. The purpose of retrieval in a learning routine is to expose and strengthen what needs attention, not to trap the learner into making an error.

Karpicke and Roediger’s 2008 vocabulary-learning experiment found that continued retrieval after initial success supported delayed recall in the conditions tested. The study concerned learned word pairs; it does not show that flashcards alone create flexible writing or reasoning. Its practical implication here is modest: include opportunities to recall, then check. See The Critical Importance of Retrieval for Learning.

Return to the idea over time

As a starting routine, revisit a new distinction on another day and again later in the week. Adjust the interval according to what happens. If recall is unreliable, shorten the gap and simplify the support. If recall is dependable across varied examples, a longer gap may be reasonable.

Do not describe a fixed sequence of days as universally optimal. Cepeda and colleagues’ work on distributed practice examined how spacing and the later retention interval interact; larger gaps were not automatically better. See Optimizing Distributed Practice. The schedule in your notebook is a practical starting choice, not a guarantee derived for every learner.

Keep the retrieval activity connected to meaning. Spelling a word, giving its definition and using it in a new explanation are different tasks. If your goal is clearer reasoning, at least some later checks should involve reasoning.

A worked retention record

Attempt	Support	Observed use	Next move
After teaching	Example visible	Distinguishes a constraint from a preference	Try a new example without the explanation
Next day	Word given, meaning hidden	Recalls a broad meaning but misclassifies a negotiable choice	Discuss the source and flexibility of the condition
Later in the week	No definition card	Uses the distinction in a different planning task	Revisit after a longer gap

The middle row is useful. It reveals that the learner remembers a rough definition but still needs to judge whether a condition is fixed in the particular task. Repeating the definition alone would not necessarily address that uncertainty.

When retrieval becomes too difficult

If the learner cannot begin, offer a cue rather than repeatedly demanding an answer. You might provide a partial example, a choice between two meanings or the first part of a sentence. Record the cue so that later comparisons remain honest.

A successful cued response is a step in learning. It should not be reported as independent recall, but it should not be dismissed either. The learner has shown what becomes possible with that amount of support. The next lesson can aim to reduce the support gradually.

Also consider whether the task is overloaded. A learner trying to remember a word, solve an unfamiliar problem and write a polished paragraph may struggle for several reasons. Reduce one demand at a time. A clear oral explanation can reveal understanding that a difficult written task conceals.

End the retention check by using the word meaningfully. A short conversation, a new sentence or a decision about an example reconnects recall with purpose. The aim is for knowledge to remain available when it has a job to do.

Contents · Previous chapter · Next chapter

A learner distinguishes observation from inference in a story about a classroom. Will the same distinction help when reading a science observation or interpreting a group message? Transfer means using learning beyond the original example. It should be checked, because familiarity with one task can look like broader understanding.

Change the setting gradually. First vary the names and details while preserving the structure. Then change the subject or response format. Finally, ask the learner to recognise when the distinction is useful without announcing it. Each step increases the demand.

Near transfer: a different example of the same kind

If the learner practised at most with players in a game, use photographs on a display. The numerical condition remains similar while the objects change. Success suggests that the learner is not tied to the original objects.

For a writing distinction, move from choosing a classroom reading corner to choosing a way to organise materials. The learner still needs a claim, relevant evidence and an explanation. Keep the background information equally accessible so that subject knowledge does not dominate the comparison.

Ask the learner to explain what stayed the same. A response such as “I still had to check whether the boundary was included” reveals a portable structure. A response such as “I remembered that the answer was six” reveals dependence on the earlier example.

A new format

Next, change how the learner responds. After written practice, try an oral explanation. After sorting examples, ask the learner to invent one. After choosing among options, ask for a decision with no options supplied.

These changes can reveal hidden dependence on a format. A learner may recognise the correct answer in multiple choice but struggle to produce the explanation. That does not make the earlier success false. It identifies another part of the skill that needs practice.

For a word such as reluctant, ask the learner first to choose a matching scene, then to describe a new scene, and later to use the word in an appropriate conversation about a fictional character. The successive tasks examine increasingly active use.

A new subject

Cross-subject transfer requires care because words can change meaning. Evidence in a literary interpretation and evidence in a science investigation share an underlying role but differ in what counts as relevant support. The learner needs the subject’s expectations as well as the general word.

Ask what the new setting requires. In a story, a quoted action may support an interpretation of a character. In a practical investigation, a recorded measurement may support a claim about the observed trial. Neither kind of evidence automatically proves every broader conclusion.

The transfer question becomes: “Which part of the idea travels, and which part needs adjusting?” This avoids two opposite mistakes: treating every context as completely unrelated, or assuming a word has exactly the same use everywhere.

Unprompted use

A final check presents a task without naming the target word. Give a planning problem and see whether the learner independently separates constraints from preferences. Offer a short reading passage and see whether the learner qualifies an uncertain inference.

Do not require the learner to say the formal word. They may use the distinction accurately in ordinary language. “We have to finish by tomorrow, but the colour is something we can change” demonstrates practical control even without the terms constraint and preference.

If the distinction appears only after a prompt, record that. You can then practise recognising the situations that call for it. Knowing a tool and noticing when to use it are related but different achievements.

Report transfer at the scale observed

Write, “Used the distinction in a new planning task without a prompt,” rather than “Can now apply critical thinking everywhere.” Broad claims erase the conditions under which the success occurred.

Look for a pattern across tasks and time. A single transfer example is encouraging; repeated use in varied settings is stronger evidence of flexible control. Even then, a learner can meet a new case that requires further teaching. Transfer is something to keep cultivating through meaningful use.

Contents · Previous chapter · Next chapter

Two learners can understand the same idea while expressing it differently. One may explain fluently in English. Another may use a home language, a diagram or a gesture before finding the English wording. If your goal is to examine the concept, those differences matter. If your goal is specifically to practise English expression, that should be stated separately.

A fair experiment makes the target clear. “Can this learner distinguish a limit from a preference?” is different from “Can this learner explain that distinction accurately in an English paragraph?” Both are legitimate learning questions. Combining them without noticing can lead to a misleading judgement.

Allow a route into the idea

A learner might sort picture cards correctly before being able to explain the rule verbally. Another might need to hear a sentence read aloud before discussing it. A third might understand a concept in one language but need help with the English collocation used to express it.

Record the route and the support. Do not silently change the task halfway through and compare the result as though the conditions were identical. At the same time, do not deny useful support merely to preserve a neat score. Teaching and measurement can serve different purposes within the same session.

For a bilingual learner, ask whether a similar distinction exists in a language they know well. Let them explain the idea first, then work on the English form. Avoid assuming a one-to-one translation. Some concepts are expressed through a phrase, a grammatical pattern or several context-dependent words.

Distinguish access from mastery

If a learner performs a task successfully when the text is read aloud, that shows something useful about access to the task. It does not by itself establish independent reading of the printed version. Likewise, a successful oral explanation does not establish polished written expression.

Keep both observations. “Understood the condition when heard; missed it while reading independently” points towards a different next step from “Misunderstood the condition in both spoken and written forms.” The first may call for attention to reading the sentence, while the second calls for more concept teaching.

These are instructional observations, not diagnoses. A short vocabulary activity cannot determine why a learner has a persistent difficulty. If the pattern remains unclear, discuss the actual work with the learner’s teacher and use the classroom context to plan support.

Choose examples that do not depend on hidden experience

A task about an unfamiliar sport may test knowledge of that sport more than the target word. A scenario about a complex workplace process may disadvantage a young learner who has never encountered it. Use familiar objects and provide the background information the task requires.

If unfamiliarity is part of the transfer challenge, make that explicit and offer the relevant facts. You can then inspect how the learner uses the vocabulary with new information, rather than interpreting missing background knowledge as a language deficit.

Invite the learner to suggest another example from their own life. A student who cannot connect with a museum-planning task may understand the same distinction through a CCA activity, a game, a family event or organising personal materials. The concept can remain stable while the example becomes more accessible.

Protect curiosity

Keep experiments brief enough to leave room for ordinary reading and conversation. A child should not feel that every interesting word triggers a test. Sometimes the most productive response is to explain the word, enjoy the story and return to the idea another day.

Ask the learner what helped them understand and what made the task confusing. Their account is one source of information, alongside their actual work. When those sources disagree, investigate gently. The purpose is to build usable language and agency, not to reduce the learner to a score.

Contents · Previous chapter · Next chapter

It is possible to include true facts in an answer and still fail to support the point. A learner may write everything they remember about a topic without asking which information bears on the question. The word relevant can provide a useful handle for this problem, provided the learner practises the relationship it names.

The following complete case is an invented teaching demonstration. Its learner, tasks and results are fictional. The purpose is to show how to plan a trial, inspect a mixed result and choose a sensible next step. The numbers illustrate record keeping; they are not research findings or expected outcomes.

The starting difficulty

Aisha is comfortable locating details in short passages. When asked to justify a choice, however, she often lists facts that do not help decide the question. Her teacher notices that she treats “true about the topic” as equivalent to “useful evidence for this claim”.

The first learning question is: “Can Aisha choose information that bears on a stated decision and explain the connection?” The teacher chooses one target word, relevant, with the contrast unrelated to this question. The aim is not to teach a long argument vocabulary in one sitting.

The teacher also checks whether Aisha understands the task itself. Aisha explains that she must choose an option and justify it from the information given. That response makes it reasonable to investigate evidence selection. If she had misunderstood the task verb, the teacher would address that first.

The first task set

Each item supplies a question and several facts about invented options. For example: “Which room is more suitable for a group that must read quietly for twenty minutes?” The supplied facts describe Room A as quiet, with seats for the whole group, and Room B as next to a noisy rehearsal. Other facts mention the colour of the doors and the day a wall poster was installed.

The learner selects facts that help answer the stated question and explains why. The teacher avoids calling door colour universally irrelevant: in a different task, colour might matter. Here, the purpose concerns quiet reading and sufficient seating, so the noise and capacity information bear more directly on the decision.

Two further items concern choosing a container that fits a fixed shelf and selecting a meeting time when all fictional team members are available. The required facts are supplied. No specialist knowledge, current school policy or outside research is needed.

There are nine planned opportunities to identify a relevant relationship across the three items. In this illustrative record, Aisha handles four independently. She selects several unrelated true facts and once overlooks a fact that would rule out an option. The teacher saves one representative explanation: “The room has a blue door, so it is a good room for reading.”

The teaching step

The teacher explains: “A relevant fact helps answer the question we are considering. It may support a choice, weaken it or show that we need more information.” The last part prevents Aisha from thinking that evidence matters only when it agrees with her preference.

Together they examine a different example: choosing a bag to carry a large drawing without folding it. A fact about the bag’s internal dimensions is relevant. A fact about its printed pattern may matter to a personal preference, but it does not establish whether the drawing will fit. A missing dimension leaves a question open.

The teacher asks Aisha to complete the sentence, “This fact matters because…” Aisha connects the bag’s width to the drawing’s width. She also identifies a fact that would change the choice: if the bag is too narrow, its attractive pattern does not solve the fit problem.

The support is recorded accurately: a short explanation, one worked contrast and a sentence prompt. The teacher does not later describe the trial as an effect of hearing a single new word. The learning involved a word, examples and guided reasoning.

The next task set

The second set uses new situations with similar structure: selecting a box for a set of cards, choosing a place for a conversation that everyone must hear, and planning a time for a shared activity. The teacher again supplies the necessary background facts.

In the invented record, Aisha handles seven of nine opportunities accurately. Two responses use the sentence prompt. One error treats a fact that is interesting but unrelated as decisive. Another error overlooks a limitation that weakens her preferred option.

Observation	First set	Later set	Interpretation to investigate
Accurate selections	4 of 9	7 of 9	Performance was better on this later set
Support during responses	No sentence prompt	Prompt used in 2 responses	Some later success was supported
Typical remaining error	True facts treated as automatically relevant	Disconfirming information overlooked	Next teaching should include facts that weaken a preferred choice
Time used	About 4 minutes	About 5 minutes	The learner spent longer explaining connections

The teacher can say that the later performance was more accurate under the recorded conditions. The teacher cannot isolate the effect of the word from the explanation, the examples, practice or possible differences between task sets. The record is still useful because it identifies a more specific remaining difficulty.

Inspect the surprising response

In one item, Aisha chooses a colourful box because it matches the display, even though the box is too small for the cards. Instead of saying that colour never matters, the teacher asks which requirement the colour addresses and which requirement it leaves unsatisfied.

Aisha recognises that she has focused on an attractive feature while missing the fit condition. She revises her explanation: “The colour matches, but the box cannot hold all the cards. For this task, the size prevents us from choosing it.” The distinction between relevant information and decisive constraints now becomes clearer.

This response also shows why a single label cannot do all the reasoning. The colour is relevant to appearance, but appearance is not the only criterion. A fact can be relevant without being sufficient to justify the final decision. The teacher introduces that additional nuance only because the case makes it necessary.

The delayed check

On another day, the teacher gives one short planning problem without displaying the word relevant or the sentence prompt. Aisha selects a useful fact and explains its connection. She then asks for a missing piece of information before choosing between two options.

That question is a positive sign. She is beginning to use the distinction without having to repeat the target word. The teacher records the exact task and response rather than writing “Mastered reasoning”. Another context may still reveal a gap.

The next step is a reading task in which some details support an interpretation and others create atmosphere. This changes the subject while preserving the need to connect evidence to a question. Because literary interpretation differs from selecting a box, the teacher discusses what counts as relevant support in the new setting.

What the complete case teaches

The experiment began with an observed pattern, chose one useful distinction, supplied a fair starting task, recorded the teaching support and inspected a later attempt. Its most valuable result was not the increase from four to seven. It was the discovery that Aisha needed practice using information that weakens a preferred choice.

A vocabulary record becomes powerful when it supports that kind of decision. Instead of assigning more random words, the teacher can choose a precise next example. Instead of telling Aisha to “think harder”, the teacher can show what to inspect.

You can adapt the case to your own learning. Choose one paragraph, one instruction or one small decision. Ask which facts matter and why. Keep the original response, teach the distinction, and return with another example. The work stays manageable because the question stays specific.

Contents · Previous chapter · Next chapter

These invented teaching cases provide limited evidence. Before reading each suggested response, decide what you would inspect next. Which conclusion is supported, which remains uncertain and what small task could reduce that uncertainty?

A learner may face several difficulties. Treat the proposed next step as a starting point and revise it when the learner’s explanation provides better information.

Case 1: The definition is remembered, but the distinction is missing

A learner repeats, “Approximate means close to the exact amount.” Asked whether an approximate answer must always equal the exact answer, the learner says yes. They can reproduce the sentence, but this response does not show the distinction that makes the word useful.

Next step: Use a small visible collection and distinguish an estimate from a count. Discuss how an estimate can be useful without matching exactly. Then offer a fresh collection. Inspect the explanation, not just the guessed quantity. Repeating the definition more loudly would leave the uncertain concept unexamined. Keep expectations appropriate to the stated task; different purposes tolerate different amounts of approximation.

Case 2: Recognition has been mistaken for retrieval

A learner selects the correct meaning of consequence from three choices. When the choices disappear, they cannot explain the word or identify a consequence in a short, familiar event. The first response shows success in that recognition task. It does not establish unsupported retrieval.

Next step: Give a simple event with an explicitly stated outcome and ask what happened as a result. Offer a cue if needed and record it. If the learner can explain the relationship in ordinary language, practise connecting it with consequence. If the relationship remains unclear even with support, return to examples. Do not assume every missing word indicates the same difficulty.

Case 3: The idea is available without its label

A learner places four picture instructions in the correct order and explains which action comes first and why. Asked to name the sequence, they pause because they cannot recall sequence. Their successful ordering and explanation should remain visible in the record.

Next step: Decide whether the target is ordering actions or retrieving the English term. For the first target, the response already offers relevant evidence. For the second, supply the label, connect it with the demonstrated idea and revisit it later. Record these as separate achievements. Asking the learner to repeat the entire ordering lesson may add little if only the label needs practice.

Case 4: Success stays attached to one example

A learner explains temporary correctly when discussing a classroom seating arrangement. In a new task, they treat a temporary label on a box as something that must stay forever. Both tasks provide enough information to establish that the arrangement or label is intended to last for a limited period.

Next step: Compare the shared relationship across the two situations. Ask what will change and what marks the end of the arrangement. Then introduce another familiar setting without repeating the original sentence. This examines transfer beyond the seating example. Check the learner’s interpretation before deciding that memory failed; the problem may concern recognising the same idea in a different object.

Case 5: Vocabulary cannot supply missing facts

A learner can explain absorbent but cannot choose which of two unnamed materials would take up more water. No observations or material properties are supplied. The task requires information that knowing the word alone cannot provide.

Next step: Supply a short invented comparison or conduct an appropriate ordinary classroom demonstration. Then ask how the observations bear on the choice. Keep the language question separate from knowledge of the materials. A useful response before that information is available is, “I need evidence about what the materials do.” Guessing the expected answer is not a stronger demonstration of vocabulary than accurately identifying the missing evidence.

Case 6: Writing hides an oral understanding

A learner explains two differences between fictional animals when speaking, but writes an incomplete sentence when asked to contrast them. The animals’ features are fully supplied. The oral response shows a comparison that the written response does not yet express successfully.

Next step: Preserve the oral evidence and inspect the writing demand. A sentence frame may help connect the two features. Record whether it was used. Later, try a fresh written comparison with an appropriate amount of support. The immediate teaching target may be expressing a contrast in a sentence, rather than understanding contrast itself. The stronger oral performance does not by itself establish independent written expression.

Case 7: Different kinds of help have been combined

During an estimation task, an adult first reads the question aloud and then says, “The answer should be around twenty.” The learner gives twenty-one. The record says “Independent use of estimate”, but that description omits help central to the answer.

Next step: Separate access support from answer support. Reading the instruction aloud and supplying the likely magnitude do different jobs. Try another suitable task with the agreed access support while withholding the answer cue, and record what happens. Keep the earlier attempt as supported practice. It may have been helpful teaching, but it cannot show what the learner would have produced without that particular cue.

Case 8: Too many things changed to isolate one explanation

A learner struggles to justify a decision about an unfamiliar historical event. After vocabulary teaching, they justify a favourite playground activity using an example paragraph and an adult’s questions. The second answer is stronger, but topic knowledge, support and task demand also changed.

Next step: Describe the supported success accurately. Then prepare a comparison with similar background knowledge and response demands, keeping the available help visible. You do not need to remove useful teaching. You need to avoid attributing the whole change to one word. The next trial should answer a narrower question, such as whether the learner now connects a reason with a choice in a fresh familiar task.

Case 9: More target words do not mean better work

One revision uses nine newly taught adjectives. Several exaggerate the supplied situation. A second revision uses three and describes the relevant details accurately. A record based only on target-word frequency would favour the first response while overlooking the purpose of description.

Next step: Check each choice against the evidence, intended meaning and audience. Count suitable applications only when counting serves the learning question. Also notice a well-chosen ordinary word that the learner decides to keep. The useful change is greater control over expression, including restraint. A longer list of words in the paragraph does not demonstrate that control by itself.

Case 10: A familiar meaning has entered the wrong setting

A learner knows mean as unkind and becomes confused by a mathematics instruction using mean as a numerical average. The familiar word has a different role in the new context. This is not necessarily a failure to recognise its printed form.

Next step: Identify the subject and inspect the surrounding sentence. Teach the relevant mathematical meaning with appropriate subject instruction, then contrast it with an ordinary-language example. Keep any difficulty with the calculation separate from selecting the meaning. A learner may understand which meaning is intended while still needing help with the mathematics. Record the particular success or difficulty rather than marking the whole word as known or unknown.

Case 11: The instruction itself is ambiguous

A sorting task says, “Choose the light objects.” One learner chooses pale colours; another chooses objects that weigh less. The materials permit both interpretations, and no context clarifies the intended meaning. A disagreement does not automatically identify a vocabulary gap.

Next step: Ask both learners to explain their choices. Revise the instruction to specify colour or weight, depending on the intended task. Keep the first responses as evidence that the wording allowed different readings. This is a useful trial of the instruction as well as the learner. Teaching someone to guess an unstated intention would hide the design problem rather than resolve it.

Case 12: A later task changes memory and difficulty together

A learner handles consecutive in short number lists during a lesson. Several days later, they encounter it in a dense paragraph with multiple conditions and make an error. The later task differs in more than the elapsed time.

Next step: Begin with a brief new task similar to the original one. If the meaning is available there, examine the added sentence demands. If it is not, offer an appropriate cue and revisit the distinction before increasing complexity. Record the gap, task and support. One difficult later response does not establish that everything was forgotten. It identifies a place where a more informative check is needed.

A useful diagnostic answer ends with an action that could change your understanding: hear an explanation, simplify one demand, supply missing facts, vary an example or check after a gap. “Practise harder” offers little guidance because it leaves both the uncertain cause and the relevant practice unspecified.

Contents · Previous chapter · Next chapter

This adaptable sequence is not a scientifically optimal schedule. Sessions can be brief, divided or repeated. Choose the spacing and amount of work in response to the learner. Record what they can do, with which support, and what deserves another check.

The worked track uses relevant: whether a detail bears on the question being answered. The learner need not begin with a formal definition. Start with, “Which information helps us answer this particular question?” Every object, task and record entry below is an invented illustration. You can substitute a different distinction while preserving the sequence of questions.

Session 1: Find a small question and preserve the first attempt

Goal: Discover how the learner currently selects information. Prepare two fictional folders and four fact cards. Folder A holds a drawing flat; Folder B requires it to be folded. A is green and B is yellow. Ask, “Which folder lets this drawing stay flat, and which information supports your answer?” The stated purpose determines which details matter.

Materials and support: Use simple drawings or read the cards aloud if needed. Record that access support. Do not teach relevant before the first response, and do not suggest which fact to choose.

Record and next step: Save the learner’s selection and explanation. If they choose the green folder because they like green, ask how that answers the question. If they already use the flatness information, keep that success and choose a more informative next task. The baseline should guide teaching, not manufacture a difficulty.

Session 2: Teach the distinction through a changed question

Goal: Connect relevant with the relationship between information and purpose. Revisit the folder facts. Explain why colour does not answer the flatness question. Then change the question to, “Which folder is green?” The same colour detail now matters. A detail is not permanently relevant or irrelevant in every possible task.

Materials and support: Keep the question visible beside the fact cards. Model one explanation: “This detail helps because the question asks about…” Let the learner move a card and explain its role. A model and visible question are teaching support, so identify them in the record.

Record and next step: Note whether the learner connects the detail with the changed question. A correct sort without an explanation may need another look. If the relationship becomes clear, prepare a fresh but comparable task. If not, simplify the facts and demonstrate another contrast before removing support.

Session 3: Check a new example under visible conditions

Goal: See whether the distinction extends beyond the folders. Supply an invented classroom-sign task: a sign must be readable from the doorway. A check found that Sign A could be read there and Sign B could not. Other cards describe their decorative borders and how many stars they contain. Ask which evidence answers the readability question.

Materials and support: Use equally short fact cards and an explicit question. Decide beforehand whether the explanation frame remains available. Keep that condition consistent while comparing responses. The fictional observations supply the facts; learners are not expected to infer readability from decoration.

Record and next step: Preserve the chosen evidence and reasoning. If the learner uses the frame successfully, record supported application. If they select the relevant observation without it, record that narrower success. Either result can guide the next session. Do not turn one correct choice into a claim about all reading tasks.

Session 4: Revisit after a gap and distinguish two questions

Goal: Check what remains available after a stated interval. Begin by asking what relevant means in the learner’s own words. Then offer a short new task, such as choosing a drawer that holds a supplied puzzle board flat. Provide all necessary fit information and one clearly unrelated detail.

Materials and support: Keep the teaching explanation out of view initially. If the learner cannot start, offer a limited cue and record its content. The time gap belongs in the record; it is not a universal pass or fail threshold.

Record and next step: Separate remembering the label from using the relationship. A learner may forget the word while selecting and explaining the useful fact. Another may recite a definition but select an unrelated detail. These patterns call for different practice. Choose the next step from the actual response, and retain successful supported work as evidence of learning in progress.

Session 5: Change the setting without hiding the demands

Goal: Examine a more demanding application. Provide a fictional invitation containing a start time, location, decoration description and organiser’s name. Ask which details help someone arrive at the right place and time. Several details can now be relevant, and choosing one may leave the task incomplete.

Materials and support: Use a short paragraph rather than separate cards. Let the learner mark details and explain how each answers the question. If paragraph reading becomes the main obstacle, read it aloud and record the change. You are still learning something useful, but under different conditions.

Record and next step: Notice whether the learner selects both place and time and avoids treating every interesting detail as necessary. Ask what extra information would be needed if the location were incomplete. Relevance does not guarantee sufficiency: a useful detail may still leave a question unanswered. Keep that boundary visible without introducing a large new word list.

Session 6: Use the distinction in an ordinary task and decide what follows

Goal: Look for purposeful use without announcing the target word. Choose an ordinary low-stakes reading or planning task with enough supplied information. For example, ask the learner to answer a teacher’s specific question using a short fictional passage. Observe how they select details before mentioning relevance.

Materials and support: Preserve the question, passage and response. If a prompt is needed, give it after the first attempt and identify the change. Ask the learner which part helped and which remained difficult; their account can suggest a better next task.

Record and next step: Compare the evidence across sessions without assuming a smooth upward path. You may decide to practise retrieval of the label, revisit selection in paragraphs, check another subject or simply use the distinction during normal learning. Finishing six sessions does not certify mastery. It gives you a more informed basis for choosing what to do next.

Read the progress record as a pattern, not a score ladder

If the record shows	What it supports	A useful next action
Correct choices with a visible question and explanation frame	Application under those supports	Try a comparable task with one support changed
A remembered definition but unrelated details selected	Recall of wording without secure application in this task	Compare two details against one explicit question
Useful details selected, but the word itself forgotten	The relationship may be available without the label	Reconnect the label and check it later
Success with cards, difficulty in a paragraph	A response-format boundary worth examining	Inspect reading demands and record any access support
A fresh task handled without a target-word prompt	Unprompted use in that particular task	Choose a different relevant setting or return to ordinary practice

A completed entry should name the task, response and conditions. For example: “On the invitation task, selected the place and time after the paragraph was read aloud; explained both choices without the sentence frame.” That invented entry is more useful than “Relevance mastered”. It preserves what happened and leaves independent reading as a separate question.

Keep the record small enough to use. One saved response and one next action can be sufficient. The programme should help you choose better teaching and help the learner use a distinction beyond the exercise.

Contents · Previous chapter · Next chapter

An unclear result is common in ordinary learning. The learner improves on one task and struggles on another. A new word makes a discussion more precise but also slower. A child can explain the distinction orally and forgets it in writing. These results are reasons to inspect the conditions, not reasons to abandon vocabulary teaching or claim success too quickly.

The learner knows the definition but does not use it

Check whether the learner recognises the situations where the idea matters. Knowing what constraint means does not automatically make someone look for constraints at the start of a task. Give two short situations and ask which one calls for the distinction. Then ask the learner to explain the cue.

Reduce the prompt gradually. Begin with “What are the constraints here?”, move to “What should you inspect before choosing?”, and later offer the task without either question. Record the level of prompting. This sequence practises recognition of a need as well as recall of a meaning.

The learner uses the word everywhere

Overuse can be a sign of enthusiasm or incomplete boundaries. Ask for a case where the term would be unnecessary or misleading. A learner who calls every preference a constraint needs examples of genuine choice. A learner who adds therefore between unrelated sentences needs to inspect whether the conclusion follows.

Let the learner replace the word with an explanation. If the explanation does not make sense, the formal label is not yet helping. Returning to ordinary language can expose the problem without discouraging the learner from attempting new vocabulary.

The later task is better, but the reason is uncertain

List what changed: the vocabulary, examples, background knowledge, adult support, format, task difficulty and familiarity. You may not be able to separate their effects. State the practical finding accurately and design a smaller next check.

For example, if a diagram and a definition were introduced together, try a new task with the diagram available but no spoken prompt. This will not perfectly isolate every influence, but it can reveal whether the learner can use that support more independently. Keep the question local and the conclusion proportionate.

The learner is slower after teaching

Inspect what the extra time contains. The learner may be checking a condition that was previously ignored, explaining an answer more carefully or handling a genuinely harder task. Slower work may accompany a useful improvement in accuracy or reasoning.

Do not make speed the main measure unless speed is part of the learning purpose. First establish accurate use. Later, if appropriate, practise the same distinction in short familiar tasks and see whether it becomes easier to access. Avoid promising a particular timetable for fluency.

The task becomes a performance for the adult

A learner may guess what answer the adult wants instead of applying the distinction. Use questions with more than one defensible response and ask for reasons. Occasionally say, “I am not sure; let us inspect the example together.” That makes evidence, rather than agreement with authority, the centre of the task.

For group work, let recipients explain how they understood a message before the sender reveals the intended meaning. This preserves the evidence of misunderstanding. If the sender explains first, everyone may retrospectively agree that the message was clear.

The exercise keeps growing

If one trial expands into a long assessment of every aspect of language, return to the original question. Choose one target, one meaningful task and one next step. Other issues can be noted for another day.

A practical stopping point is reached when you have enough evidence to choose the next teaching action. You do not need to keep testing until every uncertainty disappears. Ordinary learning benefits from cycles of useful practice, not endless measurement.

Contents · Previous chapter · Next chapter

Is this the same as a vocabulary quiz?

A quiz can form one part of an experiment. It might check recognition, spelling or recall. The broader experiment connects that knowledge to a task and inspects what happens when the learner uses it. If your goal is clearer instructions, the useful evidence includes whether someone can act on the instructions. A definition quiz alone leaves that question open.

Do I need advanced words?

No. Begin with the words that affect the learner’s actual work. Only, unless, some, because, draft and complete can make substantial differences. An unfamiliar technical term is useful when it gives the learner access to a needed distinction. Its length or rarity does not establish its value.

How many words should one experiment include?

Use the smallest set that makes the target distinction understandable. Often that is one word and a useful contrast, such as observation and inference. A later task may need a small related group. The number should follow the task and the learner’s readiness. There is no universal quota that makes an activity effective.

Can a child take part without writing long answers?

Yes. Sorting objects, choosing a picture, explaining orally or showing an action can reveal useful understanding. Make the target clear: a successful oral explanation is evidence about oral understanding, while independent written use remains another question. Keep the response form manageable so that unrelated demands do not hide the idea you are trying to teach.

What if the learner already understands the idea in another language?

Use that understanding as a starting resource. Ask the learner to explain the distinction in the language or form that is most accessible, then connect it to the English expression needed for the task. Do not interpret a search for an English word as proof that the underlying concept is absent. Also check whether the translation preserves the same meaning in this context.

Should I make the learner use the target word in every answer?

Only if producing that word is part of the specific task. A learner may use a distinction accurately in ordinary language. “This is what we saw; this is what we think happened” can demonstrate the observation–inference distinction. Later, practise the formal labels because they provide useful shorthand. Understanding and conventional expression should support each other.

How do I tell whether a new word has become useful?

Look for an action it enables or improves: a more accurate interpretation, a relevant question, a clearer explanation, a better distinction or a shared understanding of a task. Ask the learner to explain the connection. Then revisit the idea after a gap or in a new context. A single polished sentence is a starting piece of evidence, not the whole picture.

What if a simpler phrase works better?

Use it. A formal term can be introduced alongside the phrase when it helps the learner communicate efficiently with a particular audience. For example, “the thing that limits our choices” can lead into constraint. A useful vocabulary includes the ability to explain a term clearly, not just the ability to produce it.

Should every result be numerical?

No. A saved sentence, a corrected explanation or a recipient’s account of an instruction may be more informative than a number. Counts are useful when the opportunities and criteria are clear. If you record a score, preserve enough context to interpret it: the task, support, relevant conditions and the kinds of error. Avoid false precision in a small informal activity.

Can I use the same question before and after teaching?

Yes, if you want to check whether the learner now understands that example. Recognise that memory for the answer may contribute. To investigate flexible use, add a new but comparable task. Keep the demands similar and explain any differences that could affect the result. The two checks answer related but different questions.

Does improvement prove that the word caused the change?

An ordinary before-and-after activity usually cannot establish that by itself. Teaching, examples, practice, feedback, task difficulty and other factors may change together. You can still learn from the observation and use it to plan the next lesson. Describe what happened and the conditions under which it happened, then avoid claims that your design cannot support.

Can an AI language tool help create practice examples?

It can be used to propose examples, but a knowledgeable person should check the meanings, instructions and suggested answers before the learner relies on them. A fluent explanation may still contain an error or ambiguity. Keep the activity within material you can verify, use invented situations where suitable, and ask the learner for reasons rather than treating an automatically supplied answer as the final authority.

How should I respond when the learner says the activity is boring?

Ask whether the task has an obvious purpose and whether the response demand is too large. Choose a real difficulty the learner would like to solve, such as understanding a game instruction or making a project message clearer. Shorten the recording. A useful experiment does not need to interrupt every enjoyable reading experience. Sometimes the better next step is a brief explanation followed by ordinary use.

What if the results do not improve?

Inspect the match between the target and the problem. Perhaps the main difficulty is subject knowledge, reading the sentence, recalling the word or recognising when to apply it. Perhaps the examples are too similar to teach the boundary, or too different to compare fairly. A lack of improvement can still narrow the next question. Change one part of the teaching thoughtfully and try another small task.

When should I stop testing and simply practise?

Stop once you know the next useful action. If the learner needs more examples of a boundary, teach those examples. If the distinction is working, use it in reading, discussion and ordinary tasks. Return to a brief check when there is a reason, such as a new context or repeated error. The record should serve learning rather than become the centre of it.

Contents · Previous chapter · Next chapter

The teaching cases, records, diagnostic tasks and programme in this guide are original illustrations. They draw on general principles of careful instruction and comparison, but they are not a validated assessment instrument. The primary and official sources below support particular background ideas; they should not be read as direct tests of every activity proposed here.

• I. S. P. Nation: Knowing a Word — a framework for the several kinds of knowledge involved in learning vocabulary.
• Karpicke and Roediger: The Critical Importance of Retrieval for Learning — a primary experiment on continued retrieval and delayed vocabulary recall.
• Cepeda and colleagues: Optimizing Distributed Practice — primary research on the relationship between spacing and later retention.
• Varley and colleagues: Agrammatic but Numerate — a small primary study illustrating why language and reasoning should not be treated as identical.
• Institute of Education Sciences: Teaching Academic Content and Literacy to English Learners — official guidance recommending sustained work on selected vocabulary through varied activities. Its intended setting is English learners in US elementary and middle schools; it is not a Singapore examination rule.

Working terms used in this guide

Term	Working meaning here
Target distinction	The particular difference in meaning or use that the activity is intended to teach
Baseline	A recorded starting performance, with its conditions described
Prompt	A cue or reminder that supports a response
Supported performance	What the learner can do with the recorded help available
Independent use	Use without the particular assistance being investigated, under stated conditions
Retention	What remains available after a gap
Transfer	Use beyond the original example, format or setting
Confounding change	Another change that makes the source of an observed difference uncertain
Counterexample	A case that challenges a proposed rule or shows where a claim does not apply
Revision	A deliberate change made after inspecting the meaning, task or evidence

Choose the next useful question

Take one task you already need to do. Name the moment where meaning becomes unclear. Choose a word or contrast that might help. Write what you would expect to see if the distinction became useful, and keep one short record of the attempt.

You might ask whether at most helps you inspect a numerical boundary, whether relevant helps you choose evidence, or whether draft helps a partner understand a handover. Teach the idea, use it, inspect the result and adjust the next step. The vocabulary earns its place through the work it helps you understand and complete.

For a wider learning route, return to the Vocabulary hub. For the foundations of meaning, form and use, continue with What Is Vocabulary?.

Contents · Previous chapter · Continue at the vocabulary hub