Advanced vocabulary workshop: meaning-preserving revision, with worked cases, counterexamples and independent practice.

The sentence on the screen is excellent English. That is what makes the mistake difficult to see.

The source says: Hana tried to retrieve the missing file before the meeting. The revision says: Hana successfully recovered the missing file before the meeting. Both sentences are smooth. Both concern Hana, a file and a meeting. The second even sounds more accomplished. But the first sentence does not tell us whether the file was recovered. An editor has quietly supplied an outcome that the source never supplied.

Maya notices the improvement in rhythm. Jia Jun notices the more impressive verb. Ethan notices the missing evidence. He asks a question that changes the whole editing session: “Could she have tried and failed?”

Yes. That one possibility is enough to reject the revision as a faithful paraphrase. The problem is not spelling, vocabulary size or grammatical correctness. It is a change in what the sentence commits the writer to saying happened.

This is the next stage of advanced vocabulary learning: not simply finding another word, but testing the consequences of putting that word into a sentence. A student can know definitions, collocations, registers and phrasal verbs and still change the event while attempting to improve the language. The more fluent the replacement sounds, the easier the change can be to miss.

This workshop introduces a practical counterexample test. Before accepting a replacement, try to imagine an ordinary situation in which the original sentence is true but the revision is false. Then reverse the direction. When a sensible example separates the sentences, the replacement has changed some part of the meaning. The change might be unacceptable, harmless for a particular summary, or deliberately useful in a new piece of writing. The important achievement is to recognise which situation you are in.

The teaching procedure here is an original classroom application of established distinctions in semantics and writing instruction. It is not a standardised assessment, an official examination marking scheme or a claim that one exercise has been independently shown to raise grades. All student scenes, source packets, numerical examples and practice passages below are constructed for learning. They are not accounts of actual pupils, school incidents or measured tuition outcomes.

This article continues from Vocabulary and Phrasal Verbs. That guide develops control of multiword verbs. Here, learners use that control to investigate what an edit preserves, removes or invents. The wider explanation of substitution dimensions remains at Lexical Substitution Geometry on eduKateSG. This is its applied companion: a sustained editing laboratory in which students must defend decisions with examples rather than recite a checklist.

The fifty-second route: do not improve a sentence by inventing its world

Keep the source beside the revision. Underline the changed expression. Ask what the original actually establishes: who acted, what happened, whether it finished, when it happened, how many people were involved, how certain the writer is and where the information came from. Now ask whether the replacement adds a result, motive, judgement, quantity or condition that was not there.

The quickest test is often one short continuation. She tried to recover the file, but could not is coherent. She successfully recovered the file, but could not recover it on that occasion is not coherent under the same ordinary reading. The difference shows why the two expressions are not interchangeable in this sentence.

After checking factual meaning, check grammar, voice and audience. A revision can preserve the event but still change a polite request into an order, a neutral report into an accusation, or a child’s dialogue into official prose. Meaning preservation is therefore not one yes-or-no question answered by a thesaurus.

Use the full workshop in manageable parts. The early sections build the method. The laboratories test different kinds of change. The extended case files make several changes interact. The final exercises remove the hints. A learner does not need to finish the entire article in one sitting to benefit from one carefully explained distinction.

Reader routes: The counterexample method · Meaning ledger · Outcome and event laboratories · Scope and conditions · Source and perspective · Extended case files · Independent practice · Teaching and assessment · Sources and next reading.

First decide what the edit is allowed to change

An editor needs a brief before needing synonyms. “Make it better” is too vague because different kinds of improvement permit different kinds of change. A faithful paraphrase should preserve the relevant information and the source’s position. A summary intentionally removes selected detail. An explanation may add background, but should distinguish that addition from what the source said. A creative rewrite may change viewpoint or emotional intensity because that is the assigned task. These are different jobs.

Consider a source that says, The cyclist whispered an apology to the pedestrian. A shorter version, The cyclist apologised, preserves an important event but removes the recipient and the manner of speaking. That may be adequate in a one-sentence summary about whether an apology occurred. It is inadequate in an analysis of how the writer presents embarrassment or intimacy. Neither judgement can be made from word count alone. The question being answered determines which details remain essential.

Now consider The cyclist muttered an apology. This is not merely a shorter or longer version. Depending on the context, muttered can suggest low, unclear speech and may invite a different interpretation of attitude. The writer may have supplied a vocal manner that supports a different character reading. Calling both words “ways of speaking quietly” does not settle whether they are interchangeable in the passage.

Before revising, write a one-line brief: Preserve the reported events and uncertainty; make the paragraph clearer for a parent reader. Or: Reduce the passage to its explanation of the delay; retain the cause and the exception. Or: Change the message from a peer request to a respectful request to an organiser; do not add urgency. Such instructions make the comparison fair. Without them, one student may optimise brevity while another protects voice, and both may believe the other has misunderstood the task.

Purdue OWL’s paraphrasing guidance recommends understanding a source, producing a fresh formulation and checking the result against the original for essential information. It also preserves the obligation to acknowledge borrowed ideas. The counterexample routine below adds a particular comparison question to that broader writing practice; it does not replace source attribution or permission requirements.

An important boundary follows. A false original should not be preserved as if it were verified reality. When editing a source-based assignment, distinguish faithful reporting of what the source claims from independent endorsement of that claim. When correcting a factual mistake in your own writing, make the correction because you have better evidence, not because the new wording sounds more authoritative. The brief should identify whether you are preserving a statement or repairing it.

The counterexample method: a small imagined world can expose a large change

Let A be the original sentence and B the proposed revision. Start with a fixed reading: the same people, the same occasion, the same relevant time period and the same intended senses of the words. Otherwise the comparison becomes unfair. You cannot prove that two versions differ by imagining that bank means a riverbank in one and a financial institution in the other when the passage clearly describes a savings account.

Ask first: Could A be true while B is false? The file example gives an easy answer. Hana attempted recovery but failed. A remains true; B does not. The revision adds success. Next ask: Could B be true while A is false? Some pairs support an inference in one direction but not the other. Do not force a second counterexample when the ordinary reading does not support one. The first counterexample has already established that this revision is not equivalent to the source. The reverse check is especially useful when a shorter rewrite removes detail rather than adding a new result.

A useful counterexample does not have to be bizarre. In fact, ordinary examples are preferable. If a student needs an elaborate fantasy involving disguised people, impossible machines or unusual meanings to separate two everyday paraphrases, the proposed distinction may not matter for the task. Keep the scenario small enough that a classmate can repeat it in one breath. “She searched but did not find it” is more useful than a page of invented complications.

The method works because similarity is not equivalence. Two sentences may share most of their words and still make different claims. They may also use very different words while communicating the same event in the given context. The discussion was put off until Monday and The discussion was postponed until Monday can be a reasonable local match. Their acceptability comes from the selected sense, timing and event, not from a rule that every phrasal verb has one perfectly interchangeable formal partner.

One counterexample disproves a proposed universal equivalence under the tested reading. The reverse is not true: failing to think of a counterexample does not prove that none exists. A learner may have overlooked a sense, a presupposition or a social implication. The test is a practical error detector, not an automatic certification machine. That asymmetry is valuable. It encourages students to reject clearly unsafe edits while remaining appropriately modest about subtle ones.

A helpful classroom sentence is: This replacement is unsafe because the original could describe ___, while the revision would require ___. That structure forces the student to name the difference. “It sounds stronger” becomes “It turns a possibility into an actual outcome.” “The tone changed” becomes “It changes a neutral statement into an acknowledgement of responsibility.” Vague discomfort becomes an explanation another reader can inspect.

The logical distinction between paraphrase and one-way inference is also used in language research. Androutsopoulos and Malakasiotis’s survey discusses paraphrasing alongside textual entailment and describes the relevance of testing inference in both directions. Classroom editing remains broader than formal truth conditions, however: politeness, focus and narrative voice still require separate attention.

Three possible relationships, not just right or wrong

Sometimes A and B contradict each other: None of the boxes were opened cannot describe the same completed inspection as All the boxes were opened, assuming the boxes exist and the time is fixed. Sometimes A permits B but does not establish it: Some boxes were opened permits a situation in which every box was opened, but does not tell us that all were. Sometimes A establishes a less detailed B: Three labelled boxes were opened by Hana supports Hana opened boxes, but the latter does not retain the number or the labels.

These distinctions matter during feedback. A teacher who calls every unsafe replacement “false” may teach another mistake. The new statement might happen to be true in the real world. The problem is that the supplied source does not warrant it. Use unsupported addition, loss of required detail or contradiction according to the actual relationship. Students then learn why a revision fails rather than merely learning that it displeased the marker.

There is also a fourth useful judgement: not enough context. A bare sentence may leave the reference of a pronoun or the sense of a verb unresolved. Good editing sometimes means asking for the previous sentence, not choosing a synonym more confidently. Uncertainty should have a place in the workflow so that students do not treat every task as a forced guessing competition.

Build a meaning ledger before opening the thesaurus

A meaning ledger is a short record of what must survive the edit. It is not a linguistic diagram that every child must draw for every sentence. Use it when an important revision has several moving parts. For a simple sentence, the ledger may be spoken in ten seconds. For a dense paragraph, writing it down can stop an attractive rewrite from taking over the task.

Take this constructed source: According to Hana, two volunteers may have moved the smaller boxes before the organiser arrived. The ledger needs several entries. Hana is the information source. Two volunteers are the possible actors. Movement of the smaller boxes is uncertain rather than confirmed. The comparison smaller restricts which boxes are involved. The reported movement would have occurred before the organiser’s arrival. Nothing says that the volunteers acted without permission, that they stole anything, or that they moved every box.

Now compare: Two volunteers secretly removed all the boxes before the organiser arrived. This revision changes more than vocabulary. It removes attribution, removes uncertainty, adds secrecy, changes the object set and shifts movement towards removal. A student might improve its rhythm again and still be repairing the wrong sentence. The ledger makes the unlicensed additions visible before style distracts attention.

The most useful ledger headings are participants, action, outcome, time, quantity, source, certainty and restrictions. For a narrative task, add viewpoint and manner. For a message, add purpose and relationship. For a technical task, add the operational definition of the term being edited. Do not force every heading into every task. The ledger should expose the actual risks, not become another worksheet completed without thought.

Participants: keep the direction of the event

Who gives, who receives, who asks and who is asked? A replacement can preserve the topic while reversing the relationship. Maya lent Hana a notebook and Hana borrowed a notebook from Maya can describe the same transaction, provided the same notebook and occasion are intended. Maya borrowed Hana’s notebook describes the opposite direction. A quick arrow between the two names can be more effective than memorising another definition of lend.

Pronouns make this harder. Replacing a repeated name with she may reduce repetition while making the reference ambiguous. An elegant sentence is not an improvement when a reader can no longer tell whether Hana or Maya made the decision. The ledger should track people across sentences, not only inside one sentence. This is particularly important when two participants share the same pronouns.

Outcome: distinguish an attempt, a process and a completed result

Looked for, was looking for and found belong to the same situation but do different work. So do started repairing, was repairing and repaired. The event ledger asks whether the source commits to completion. A student should not supply completion because it makes the narrative feel satisfying. Unfinished action may be the exact point of the scene.

The same care applies when a revision removes an outcome. Changing managed to unlock the door to tried to unlock the door preserves effort but no longer establishes success. In a summary about how the group escaped a locked room, that loss is not minor. A shorter version may be acceptable only if it retains the fact that they got out or the door opened elsewhere in the paragraph.

Source and certainty: keep the wrapper around the claim

Hana said that the file was missing is a report of what Hana said. The file was missing presents the missing file directly. The second may be supported by other evidence, but the first alone does not give the editor permission to discard the source boundary. It is possible for someone to report a mistaken belief sincerely.

Similarly, may have moved is not a decorative hedge around moved. It changes whether the event is established. Do not remove uncertainty merely to make writing sound confident. Equally, do not insert uncertainty into a settled measurement simply because cautious prose seems more mature. The target is faithful calibration, not maximum confidence or maximum hesitation.

Quantity, time and restriction: protect the small words

Two, smaller, before, only and unless are easy to overlook because they are familiar. Yet they often carry the strongest constraints. Replacing two boxes with the boxes can suggest a larger set. Replacing before the meeting with during the meeting changes the timeline. Removing only when supervised turns a conditional permission into a broader one.

When a learner is tired, these small items deserve deliberate attention. It is tempting to spend the entire editing session replacing adjectives while missing the word that reverses the instruction. The ledger redirects effort towards the information with the greatest consequence. The visual size of a word is not a measure of its importance.

A worked demonstration: three attractive rewrites, different consequences

Here is the source: The team planned to set up two tables outside, but moved the session indoors when rain began. Jia Jun offers: The team established an outdoor event before heavy rain forced everyone inside. Maya offers: Rain disrupted the team’s event. Hana offers: The team had planned to set up two tables outside; when rain began, they moved the session indoors.

Start with Jia Jun’s version. Planned to set up has become established. Intention has become accomplished preparation. Rain began has become heavy rain. Intensity has been added. Moved the session indoors has become forced everyone inside. The revision adds coercive force and a claim about every person. None of those changes follows from the short source. The version is fluent but cannot pass the meaning ledger.

Maya’s version is not necessarily false. It may be a plausible summary of an inconvenient change. But disrupted adds a judgement about the effect, while the original could describe an orderly contingency plan carried out without difficulty. It also loses the intended outdoor arrangement and the movement indoors. Whether that loss matters depends on the question. If the task is to report the adjustment, the summary is too vague. If the task is to identify the weather event that prompted a change, the focus still needs clarification.

Hana’s version is closer for a meaning-preserving revision. It retains intention, the two tables, the outdoors-to-indoors change and the sequence. Even then, a careful editor checks the causal connection. The source’s when rain began naturally invites a connection in this context, but the rewrite does not need to strengthen it to because severe weather made continuation impossible. The defensible version says enough without supplying an emergency.

Now create counterexamples. The team might have moved before any table was set up. The rain might have been light. Everyone might already have been under shelter. These are ordinary possibilities, not tricks. Each one exposes a different addition in Jia Jun’s sentence. The lesson is not that his vocabulary is too advanced. It is that his revision makes more commitments than his evidence can carry.

A more compact alternative is The team moved the session indoors when rain began, despite its plan to set up two tables outside. Notice what neither faithful version settles: whether any tables had already been set up before the move. Adding instead of setting up the tables could imply that the planned setup never happened. The source does not establish that. The discussion demonstrates why editing is a sequence of local decisions, not a race towards the shortest sentence.

Reference tools: what each one can settle, and what it cannot

A learner dictionary can confirm a word’s senses, common constructions, usage labels and typical examples. A grammar reference can clarify why one complement pattern is acceptable and another is not. A corpus can show how expressions occur in a selected collection of texts. A thesaurus can suggest candidates. None of these tools, used alone, decides whether a new sentence preserves the particular event described in your source.

Suppose a thesaurus connects postpone and delay. That is a useful lead, not the final answer. They postponed the meeting until Friday identifies an intended change in schedule. They delayed the meeting may leave the new date unstated and may describe a different kind of postponement or hindrance. The student must inspect the sentence, not merely the relationship between two dictionary headwords.

The same principle applies to grammar. British Council guidance on reporting verbs shows that different reporting verbs take different following patterns. That is useful when changing suggested that we leave to suggested leaving, or contrasting a suggestion with an instruction. The grammatical pattern is one check; preserving who made the suggestion and whether anyone followed it remains another.

Use a narrow question when consulting a reference. Instead of asking, “Is this better English?”, ask, “Does recover in this sense require successful retrieval?” or “Can suggest take this object-plus-infinitive pattern?” or “Does the source establish that the request was accepted?” Specific questions produce information that can be compared against the ledger. Broad requests invite broad rewrites, and broad rewrites often change several dimensions at once.

For tool-assisted practice, keep the original attempt. Ask the tool for one candidate at a time and request a list of possible changes in meaning rather than a confident approval. Then inspect those changes yourself. A second fluent explanation is still another text to evaluate. The counterexample is useful because it gives the learner something concrete to check rather than asking them to trust the tool’s tone.

The next laboratories apply this procedure to success, timing, scope, conditions, source, grammar and social meaning. Each begins with a limited source. Do not secretly supplement it with a story that makes the revision work. A hypothetical rescue can show that two statements are compatible. It cannot show that the source guarantees the revision.

Laboratory A: verbs that change whether the event happened

Case 1. “Tried to” does not give an editor permission to write “managed to”

Source: Ethan tried to open the jammed drawer before the lesson. Revision: Ethan managed to open the jammed drawer before the lesson. The revision has not merely made the verb more precise. It has changed the outcome. In the ordinary affirmative reading, managed to open presents the opening as accomplished, often against difficulty. Tried to open presents an attempt and leaves the outcome unresolved.

The counterexample can be very small: Ethan pulled the handle, checked the lock and gave up because the drawer would not move. The source fits that scene; the revision does not. No amount of shared vocabulary rescues the substitution. The jammed drawer actually makes failure a particularly natural possibility, but the distinction would remain even if the drawer had not been described as jammed.

Now reverse the proposed edit. If the source says managed to open and the student changes it to tried to open, the revision removes success. That does not necessarily make the sentence false. It makes it less informative. In a summary whose point is that a missing worksheet was finally retrieved, losing the successful opening would be a material loss. The editor must preserve the outcome somewhere, even if the exact expression changes.

A usable repair is Ethan attempted to open the jammed drawer before the lesson. This keeps the attempt and the unresolved outcome in a somewhat more formal form. A different repair is simply to keep tried. The shorter common word was already doing the correct job. Vocabulary improvement sometimes consists of rejecting a replacement rather than inserting one.

For an advanced discussion, ask students what follows from Ethan did not manage to open the drawer. Under the ordinary same-occasion reading, successful opening is denied. But do not turn this into an unrestricted rule for every sentence containing manage. If Ethan manages to open it does not tell us that he opened it. Neither does Did Ethan manage to open it? Karttunen’s work on implicative constructions examines precisely how the surrounding construction affects such inferences. In classroom practice, fix the sentence type before drawing the conclusion.

Transfer challenge: The source says that Maya hoped to finish a report. A revision says that she completed it. Give a counterexample, then produce a faithful version without using hoped. An acceptable answer must preserve an attitude towards a possible outcome rather than announce the outcome itself.

Case 2. Plans are not completed events

Source: The students planned to put up a display near the classroom entrance. Revision: The students installed a display near the classroom entrance. Here the attractive word is installed. It sounds concrete and capable. Yet the source contains only a plan. The materials may not have arrived. The teacher may have chosen another location. The plan may have been abandoned for a reason that the passage has not yet revealed.

The counterexample is not a challenge to the students’ sincerity. They could have planned earnestly and still not acted. Intention, preparation, attempt and completion are separate stages. A rewrite that moves from one stage to another is not neutral. Notice how this connects directly to phrasal verbs: planned to set up is not equivalent to set up, even though most of the phrase remains unchanged.

A faithful revision might say The students intended to create a display near the classroom entrance. Depending on the passage, intended and planned may differ in how much preparation they suggest, so retain planned when the organised preparation matters. Another option is Their plan was to put a display near the classroom entrance. This changes the structure without manufacturing completion.

Now make the task a summary rather than an exact paraphrase. The source paragraph may later say that the display was completed. In that larger context, a summary can report the completed display. The permission comes from the later evidence, not from treating planned as a synonym for installed. This distinction is useful when students think an editor is being excessively literal. The editor is allowed to use context; they are not allowed to invent it.

The same problem appears in project reports. We proposed, we prepared, we tested and we implemented are not interchangeable markers of progress. The more important the report, the more important it becomes to name the actual stage. A polished account that promotes a proposal into an implementation misleads its reader even when the project is worthy and everyone involved means well.

Transfer challenge: Rewrite The group was preparing to present its findings without claiming that the presentation happened. The safe revision must keep the group on the preparation side of the event boundary.

Case 3. Looking for information is not finding it

Source: Hana looked for the receipt in the folder. Revision: Hana found the receipt in the folder. The two verbs belong to one familiar script: a person searches for a missing object. Because that script often ends in discovery, a reader may mentally complete it. Editing requires separating the expected ending from the stated ending.

Imagine the folder contains old worksheets and a timetable but no receipt. Hana searches every pocket. The source is true; the revision is false. This counterexample reveals a result added by the new verb. A student who says, “But she was looking for it, so she probably found it,” has moved from language evidence to a guess about events. Probability does not convert the source into a statement of discovery.

A safer revision is Hana searched the folder for the receipt. It retains the search, the location and the target. It also changes the grammatical frame. Look for the receipt in the folder becomes search the folder for the receipt. The editor must move the objects and prepositions correctly, rather than inserting searched into the old structure mechanically.

Now compare Hana found out why the receipt was missing. The particle changes the kind of discovery. Found the receipt identifies an object located; found out why it was missing identifies information discovered. A student who treats find and find out as interchangeable may produce a sentence whose topic is similar but whose achievement is different. The edit could even turn a still-missing receipt into a recovered one.

This case is particularly useful because the words are not rare. Advanced vocabulary control is often visible in ordinary verbs. A learner who can keep the search, the object and the discovery of information separate has more practical control than a learner who merely collects formal synonyms for look. The sophistication is in the boundary, not the number of syllables.

Transfer challenge: The source says Maya looked into the missing delivery. Does Maya located the delivery preserve the meaning? It does not. An investigation may end with an explanation, an unresolved question or a located item. The source alone has not selected the result.

Case 4. “Put off” and “call off” make different changes to a plan

Source: The organisers put off the rehearsal until the following Tuesday. Revision: The organisers called off the rehearsal. In this context, the source states a postponement with a new time. The revision states cancellation and omits the rescheduling. Although cancelled activities can later be reorganised, the revision does not preserve what this source says about the intended Tuesday rehearsal.

Imagine the same participants attend the same planned rehearsal on Tuesday. The source describes that arrangement naturally. The revision could lead a reader to believe no rehearsal is now planned. The practical consequence is clear: a student relying on the rewritten message might fail to prepare for Tuesday. A vocabulary distinction has become an action difference.

A faithful version is The organisers postponed the rehearsal until the following Tuesday. Here a single-word verb can replace the phrasal verb without losing the key scheduling information. The successful match is local. Put off has other uses, including discouraging or repelling someone, so a student should not store “put off equals postpone” as an unrestricted translation for every sentence.

Consider a different source: The smell put Hana off the food. Postponed Hana would plainly fail. The relevant sense is discouragement or loss of appetite, not rescheduling. This is why the editor first chooses the sense and then evaluates substitution. A dictionary relationship attached to the wrong sense does not help.

The case also teaches a useful restraint in summaries. Removing until the following Tuesday may be acceptable if the question asks only what the organisers did when the room became unavailable. It is not acceptable in a message intended to tell participants when to return. The required detail depends on the reader’s job. The editor must preserve the new date when the date is the reason for writing.

Transfer challenge: Compare The speaker put off answering the question with The speaker refused to answer the question. Delay is not refusal. Supply a scene in which the speaker answers later, and the difference becomes easy to defend.

Case 5. An action in progress is not automatically an action completed

Source: At four o’clock, Maya was writing the final paragraph. Revision: Maya had completed the final paragraph by four o’clock. The source locates an ongoing activity at a time. The revision places completion no later than that time. This is a change in aspect and timeline, not merely a more efficient way to say that writing occurred.

A counterexample needs only an unfinished sentence. At four o’clock Maya is halfway through her paragraph; she finishes at four fifteen. The source fits; the revision does not. The fact that the paragraph is described as final tells us its position in the report, not that its writing has already ended. Students often mistake a noun’s place in a sequence for evidence that the sequence is complete.

The safe repair may be Maya was working on the last paragraph at four o’clock. That version broadens writing slightly into working on, which could include revising or planning. If the exact activity matters, keep writing. A closer structural change is The final paragraph was still being written by Maya at four o’clock, although still adds a continuation perspective that should be justified by context. The original may therefore remain the best sentence.

Now ask students to separate Maya wrote a paragraph for fifteen minutes from Maya wrote the paragraph in fifteen minutes. These are constructed contrasts, not universal claims about every possible use. Under their ordinary readings, the first specifies duration of activity; the second presents the time taken to complete a bounded task. A changed preposition can therefore alter the interpretation of completion even when the main verb is identical.

This laboratory is valuable in reports of experiments, building projects and revision work. The team was testing the model does not establish that the test was finished, that it was successful, or that the model was adopted. Keeping those stages separate protects both factual accuracy and the reader’s understanding of what remains to be done.

Transfer challenge: Rewrite The students were working through the questions when the bell rang. A faithful answer should not quietly turn working through into completed.

Case 6. “Stopped checking” and “stopped to check” describe different interruptions

Source: Ethan stopped checking the figures when the teacher entered. Revision: Ethan stopped to check the figures when the teacher entered. The source says that the checking activity stopped. The revision says that Ethan interrupted some other activity for the purpose of checking. One small change creates a different event structure.

Imagine Ethan had been checking for ten minutes and put the paper away as the teacher arrived. The source fits. The revision suggests that the arrival prompted him to begin or undertake checking instead. The two versions point in opposite directions with respect to the activity that matters. A student who looks only at the unchanged words stopped, check and figures can miss this.

The repair should preserve the relationship, not merely the vocabulary. Ethan ceased checking the figures when the teacher entered is a possible more formal version, but it may be unnecessarily stiff for a story. When the teacher entered, Ethan stopped his check of the figures preserves the cessation but may make the checking sound like a more formally bounded inspection. Keeping the original is again a legitimate editorial decision.

For a new scene, Ethan stopped to check the figures is entirely useful. Perhaps he was carrying the report to the office and paused in the corridor. The sentence is not bad English. It is wrong only as a replacement for the specified source. This distinction should become routine: reject the substitution, not the expression everywhere.

Students can extend the inquiry to remembered sending and remembered to send. The first concerns memory of an event; the second ordinarily presents remembering to perform an intended action. The exact inferential pattern depends on the surrounding sentence. A question, condition or denial can change what is established. Advanced learners should inspect the full construction rather than assign permanent outcome labels to individual verbs.

Transfer challenge: Write two short scenes, one ending with Hana stopped talking and the other with Hana stopped to talk. The scenes should make the difference visible without relying on a definition. This is a useful test of whether the learner controls the construction rather than merely recognises an explanation.

Case 7. “Did not prevent” does not establish either permission or occurrence

Source: The organiser did not prevent the students from entering the room. Revision: The organiser allowed the students into the room. These may fit the same scene, but the first does not necessarily establish an act of permission. The organiser might have been absent, unaware or unable to intervene. Failure to prevent is not automatically an affirmative decision to allow.

There is a second possible addition. A reader may interpret allowed the students into the room as presenting entry as having occurred. The source itself does not establish entry. The students might have changed their minds outside the door. No prevention occurred because no attempt to enter proceeded. The safe interpretation should therefore remain narrower until the surrounding passage supplies more information.

The counterexample test works well here because students can imagine a familiar classroom. The organiser is in another room; the students approach, notice the lights are off and leave. The source can be true. A revision claiming authorised entry would not be warranted. The vocabulary problem is not politeness. It is a change in agency, permission and possibly outcome.

A faithful revision might be The organiser did not stop the students from entering the room, provided stop has the prevention sense in context. Even this should not be converted into The students entered unhindered unless the passage establishes that they entered. The latter statement reports an event with a manner; the former reports the absence of a particular intervention.

The broader lesson is not to interpret every negative sentence as the positive opposite. Did not refuse is not always the same as accepted. Did not object is not always the same as endorsed. Did not disprove is not the same as proved. Different verbs introduce different relationships. A tidy two-column antonym exercise can conceal distinctions that matter greatly in reporting what people did.

Transfer challenge: Explain why The teacher did not forbid the suggestion cannot automatically become The teacher approved the suggestion. Identify an ordinary neutral position between prohibition and approval.

Case 8. A question or condition does not report a completed event

Source: If Hana finds out where the file is, she will tell the group. Revision: Hana discovered the file’s location and informed the group. The first is conditional. The second reports two completed events. The vocabulary has become more formal, but the grammatical environment that made the original uncertain has disappeared.

A simple counterexample is that Hana has not yet found anything. The original promise or prediction can remain in place while the search continues. The revision would be premature. Another possibility is that the condition never occurs. In that case, the original does not by itself tell us that the group receives a message. The conditional structure is part of the meaning, not material an editor can remove for pace.

A faithful version might say Hana will inform the group if she discovers the file’s location. This preserves the conditional connection. It changes the order of information, which may affect emphasis, but it does not turn a possible discovery into a completed one. The editor should still check that discovers and finds out fit the intended information-search sense.

Now compare Did Hana find out where the file was? with Hana found out where the file was. The first asks a question; the second makes an assertion. The question may be asked precisely because the answer is unknown. Students should not treat an embedded expression as true merely because it appears in a sentence. Whether the expression is asserted, questioned, imagined, denied or attributed matters.

This laboratory provides a stopping rule for overconfident semantic analysis. First identify the sentence’s overall job. Then examine the lexical item inside it. A useful dictionary explanation of discover cannot override if, a question mark or might. The surrounding structure controls how much of the event the writer has committed to.

Transfer challenge: The source asks, Could the students have worked out the answer without the diagram? Write a faithful report of the question. Do not claim that the students did, or did not, solve it independently.

Laboratory B: scope, quantity and conditions

Case 9. “Some” leaves possibilities open that “all” closes

Source: Some of the students checked their answers before handing in the work. Revision: The students all checked their answers before handing in the work. The revision turns a limited claim into a universal claim. Imagine a group of six in which two students checked and four did not. The source is true; the revision is false. The counterexample is sufficient without knowing which actual students checked.

There is an advanced complication worth keeping. In ordinary conversation, some can suggest “some but not all”, but that suggestion is not always a strict part of what the word establishes. Some students checked—in fact, all of them did can be a coherent clarification. Therefore, do not repair the first error by teaching that some always means a precisely limited minority. It certainly does not automatically mean “fewer than half”.

When the intended meaning is explicitly “some but not all”, preserve those words or an equally clear formulation. When the source gives no upper limit, do not invent one. Several students checked may also introduce a quantity expectation not exactly matching some. The safest replacement may involve the sentence structure rather than the quantifier: Before submitting their work, some students reviewed their answers.

This distinction becomes especially important when students summarise survey results. A sentence about a responding subset should not become a claim about the whole class. If a survey records views from twelve volunteers, the editor must not silently promote those volunteers into every student at the school. The changed population can matter more than every adjective in the paragraph.

For a useful challenge, ask students to invent two worlds compatible with the source: one in which only a small number checked, and another in which everyone checked. The exercise trains them to see a statement as permitting a range of possibilities. A paraphrase should not close that range without evidence or an explicit change in the task.

Case 10. “Not all” is not “none”

Source: Not all of the windows were locked. Revision: None of the windows were locked. Imagine five windows: four locked and one unlocked. The source is true and the revision false. The original denies that the whole set satisfies the condition; it does not deny the condition of every member.

A faithful, explicit revision is At least one of the windows was not locked, assuming the same finite set and ordinary meaning of locked. This version makes the minimum information clear without claiming how many windows were unlocked. It is not as stylistically compact, but it can be useful during learning because it exposes the logical structure. Once students understand it, they can return to the shorter form.

The reverse mistake also appears. Changing None of the windows were locked to Not all the windows were locked usually weakens the information. The latter would remain true if four out of five were locked, so it does not tell the reader the full extent of the original problem. A summary may deliberately generalise, but it should not do so when the extent of the problem is central to the task.

Ask what happens when a negation moves. All the students did not agree can be less clear than either None of the students agreed or Not all the students agreed, because readers may resolve its scope differently. When writing for clarity, choose the version that makes the intended distribution explicit. Advanced editing is not merely defending an interpretation that is grammatically possible; it is helping the intended reader reach the right interpretation reliably.

This case is a useful bridge between language and mathematics, but do not turn every ordinary discussion into symbolic logic. The five-window example is enough for most learners. What matters is the ability to produce a counterexample with a mixed set and then explain why the source permits it while the revision does not.

Case 11. Moving “only” can move the boundary of the claim

Source: Only Maya checked the labels. Revision: Maya checked only the labels. Under the intended ordinary readings, the first restricts who checked the labels. The second restricts what Maya checked. The same words appear, but the exclusion has moved.

Imagine that Maya checked the labels, the dates and the quantities, while nobody else checked the labels. The source is true. The revision, understood as saying she checked no relevant item other than the labels, is false. Now imagine that Maya checked just the labels while Hana also checked the labels. The revision can be true while the source is false. Both directions expose the difference clearly.

A safe edit should preserve the focus. Maya was the only person who checked the labels retains the first restriction. The labels were the only items Maya checked retains the second. Each is longer than its source, but the expansion is useful for understanding. Once the intended restriction is clear, the writer can decide whether the compact original is sufficiently unambiguous.

Spoken stress can complicate the interpretation of only, so the workshop fixes the reading rather than claiming that written position mechanically resolves every possible utterance. A speaker can stress different elements, and a surrounding contrast can influence what alternatives are relevant. In a written school answer, it is often better to remove uncertainty than to rely on imagined intonation.

The editorial lesson is practical: small focus words deserve their own comparison pass. Students sometimes preserve nouns and verbs perfectly while moving only, even or also as if they were decorative rhythm markers. Such movement can change who is included, what is contrasted and what the reader is invited to assume. The same-looking sentence can therefore fail the meaning test before any advanced synonym is involved.

Case 12. “If” and “only if” do not offer the same guarantee

Source: A student may enter the workshop if the supervisor approves. Revision: A student may enter the workshop only if the supervisor approves. In a simplified rule-reading exercise, the source gives approval as a sufficient condition for permission. The revision makes approval a necessary condition. The difference concerns whether other routes to permission remain possible and whether approval alone settles the matter.

Construct a policy with two routes: supervisor approval or a pre-issued access pass. The source can be compatible with both routes. The revision, read as requiring supervisor approval in every case, excludes entry based solely on the pass. The added only has narrowed the permission. It may be a sensible policy, but it is not a neutral edit of the supplied sentence.

Now test the reverse direction with another simplified policy. Approval is required, but the student must also complete an induction. Saying approval is necessary does not make it sufficient. A student who has approval but no induction may still lack permission. This scenario helps learners separate “you need this” from “this is enough”. The phrases are often confused because everyday instructions leave other background requirements unstated.

For that reason, the classroom case deliberately specifies a simplified rule. In real life, an isolated sentence about permission may rely on many understood conditions. Do not use this exercise to infer an actual institution’s access rules from one sentence. The language lesson is about preserving the stated relationship, and a faithful revision may need the full notice or policy context.

A clear teaching paraphrase of only if is approval is required. A clear teaching paraphrase of the simplified if rule is approval is enough for permission under this rule. Those expansions reveal the direction of the condition. They also show why a writer should avoid switching connectors just to vary vocabulary. A connector is part of the claim’s architecture.

Case 13. “Must not” is not “do not have to”

Source: Students do not have to submit a printed copy. Revision: Students must not submit a printed copy. The source removes an obligation. The revision introduces a prohibition. Imagine a student who is allowed to bring a printed copy but is also allowed to submit electronically. The source fits; the revision does not.

This is not a subtle stylistic difference. It changes the student’s available choices. A family reading the revised instruction might discard a useful printed copy or worry that bringing one breaks a rule. The language has moved from optionality to forbidding. Small grammatical changes can therefore have consequences even when the subject matter is ordinary classroom administration.

A faithful version might be Submitting a printed copy is optional. That works if the source context really means there is no requirement and no separate prohibition. The editor should check the whole instruction rather than assuming permission from the absence of obligation in every conceivable setting. In this constructed case, optional submission is the intended reading; in a real notice, another sentence could add a restriction.

British Council’s reference on permission and obligation explicitly distinguishes lack of necessity from prohibition. Use that distinction as a starting point, then return to the source. The learning goal is not to replace every do not have to with optional. It is to preserve the difference between an action that is unnecessary and one that is not allowed.

A useful follow-on is You need not answer every question versus You must answer no questions. The second makes the prohibition extreme and exposes the first sentence’s more limited meaning. Students who can explain the available choices are showing understanding. Students who only memorise a modal table may still miss the choice structure when the wording becomes longer.

Case 14. Possibility, likelihood and obligation are different kinds of meaning

Source: The new timetable may reduce waiting time. Revision: The new timetable will reduce waiting time. The revision turns a possibility into a prediction stated without that possibility marker. It might eventually prove correct, but the source alone does not establish it. A counterexample is a timetable that is capable of helping but fails to reduce waiting because another bottleneck remains.

Do not respond by assigning universal percentages to may, might, could and likely. Their interpretation depends on context, speaker and construction. A classroom chart saying that one word always means exactly a particular numerical probability would create false precision. The safer question is whether the revision has narrowed uncertainty or increased the strength of the commitment.

Now consider The supervisor must be in the office and The supervisor must report to the office. In the first, must can express a strong inference. In the second, it can express an obligation. Replacing both with certainly would fail because an obligation is not a statement that the action has occurred. Replacing both with is required to would also fail when the sentence is about an inferred location.

The counterexample method helps students avoid sorting words by form alone. A person can be required to attend and still fail to attend. A speaker can infer that someone is present and be mistaken about the evidence. The two sentences concern different relationships to the proposition. The editor must preserve whether the sentence is reporting reality, estimating it or prescribing behaviour.

For an advanced revision, write the meaning plainly before selecting the modal. It is possible that the timetable will help does one job. We expect the timetable to help does another. The timetable is required to meet this waiting-time target does a third. Keeping these separate makes both school writing and everyday planning more honest.

Case 15. “By Friday”, “on Friday” and “until Friday” protect different timelines

Source: The group will submit the report by Friday. Revision: The group will submit the report on Friday. The original permits earlier submission. The revision locates submission on Friday. A group that submits on Wednesday satisfies the source but not the revision. The added precision is not earned merely because both versions mention the same day.

A faithful expansion is The group will submit the report no later than Friday. The precise boundary may still depend on the context: a school deadline could specify a particular time on Friday. Do not invent that time. If the original notice gives a time, preserve it. If it does not, avoid turning by Friday into a made-up midnight deadline.

Now compare The room is available until Friday with The room becomes available on Friday. One concerns the end of an availability interval; the other concerns its beginning. A room usable on Monday but unavailable from the weekend fits the first and not the second. The same word available does not compensate for the changed temporal relation.

The particle in a phrasal verb can interact with this timeline. The session was put off until Friday identifies a revised schedule. The session was held off until Friday may suggest delaying or restraining its occurrence and can sound different depending on context. A simple formal replacement should retain the date and the intended scheduling action, not merely preserve the noun session.

Ask learners to draw a rough timeline rather than argue from impressions. Mark the source’s permitted interval and then the revision’s interval. If one allows Wednesday and the other does not, the difference is visible. The drawing is a reasoning aid, not a replacement for reading the sentence. It helps students see where a supposedly elegant time phrase has closed an option that the source left open.

Case 16. An average improvement is not an improvement for every student

Source packet: In this invented four-student example, scores change from 40, 50, 60 and 70 to 30, 50, 70 and 90. Source sentence: The group’s average score increased. Revision: Every student improved. The mean rises from 55 to 60, but the first score falls and the second remains unchanged. The revision does not follow from the group-level statement.

This case includes explicit numbers so that learners can construct the counterexample rather than merely trust the explanation. The values are teaching data, not results from any class. The arithmetic is simple enough that the language remains the focus. An aggregate statistic summarises a group; it does not describe every member’s change.

A faithful revision might be The mean score for the group was higher in the second set of results. If the source uses average without defining the measure, the editor should not automatically choose mean in every real-world context. Here the packet explicitly defines the arithmetic mean, so the substitution is supported. The supplied data determine which technical term is justified.

The same error occurs without numbers. The route is usually faster does not mean it was faster today. Most participants preferred the revised wording does not mean every participant did. The school’s results improved does not specify the result of one child. A writer can move between levels of description, but the transition requires evidence.

For advanced learners, ask whether the stronger statement could happen to be true in another dataset. Yes. All four scores could rise. But that compatibility is not enough. The original statement permits both mixed and uniform improvement, while the revision selects only the uniform case. This is the central counterexample principle again: a plausible stronger story is not a faithful substitute for the more limited evidence.

Laboratory C: source, background assumptions and perspective

Case 17. “Believed” and “realised” position the narrator differently

Source: Maya believed that the cupboard was empty. Revision: Maya realised that the cupboard was empty. The first reports Maya’s belief. It allows the cupboard to contain something she has overlooked. In the ordinary reading of the second, the narrator presents the cupboard’s emptiness as a fact that Maya comes to recognise. The revision therefore changes more than the strength of her feeling.

Imagine a packet hidden behind a stack of folders. Maya looks quickly and believes the cupboard is empty, but Hana later finds the packet. The source fits. The revision, under its ordinary straightforward narrative reading, would present an emptiness that the scene does not support. A writer could create an unreliable narrator or an ironic correction, but the editor should not silently invent that special context to defend a routine substitution.

A safer revision is Maya thought the cupboard was empty. Depending on the passage, believed and thought may differ in emphasis or strength, so they are not universally identical. But this change can preserve the crucial distinction between a person’s view and the narrator’s presentation of a fact. The editor must still consider whether the source intended a firmly held belief or a passing impression.

This is related to factivity, a topic explored in the Stanford Encyclopedia of Philosophy’s entry on presupposition. Some expressions ordinarily treat embedded information as taken for granted. Context can affect how that background information behaves, so the classroom rule should remain careful: inspect what the narrator presents as established, rather than teaching that a word magically verifies reality.

A useful reversal is Maya did not realise that the cupboard was empty. The ordinary reading denies her awareness while continuing to present the emptiness as background. Compare Maya did not believe that the cupboard was empty, which reports her attitude without the same straightforward commitment. The word choice changes the relationship between mind and world. That is why advanced vocabulary requires more than an intensity ladder from weak to strong.

Case 18. “Again” and “stopped” can add a past the source never gave

Source: The class was quiet when the visitor entered. Revision: The class stopped talking when the visitor entered. Both can describe a quiet class at the relevant moment, but the revision ordinarily introduces an earlier talking activity and a transition out of it. The source does not establish either. Perhaps the class had been reading silently for twenty minutes.

That silent-reading scene is a counterexample. The original description is accurate; the revision manufactures a previous activity. A student may prefer stopped talking because it creates movement and causation. But the added liveliness comes from a changed event, not merely improved wording. The editing brief has to authorise that kind of creative invention.

Now compare The class became quiet again. The word again ordinarily introduces a previous quiet period, though particular contexts can affect the exact repetition or restoration being presented. A source that only describes the current silence does not automatically provide that history. The editor should ask what earlier event the new sentence invites the reader to take for granted.

The repair can be simple: The visitor entered a quiet classroom. This preserves a quiet setting but changes the sentence’s perspective and may leave the exact timing slightly less explicit. When the visitor entered, the class was quiet is an even smaller structural revision that preserves the original relation. A student does not have to manufacture action to avoid repeating a sentence pattern.

This case demonstrates why two sentences can agree about the present yet differ in background meaning. A basic truth comparison at one instant may miss the added history. Extend the ledger backwards. What must have happened before this sentence can be used naturally? Has the revision supplied earlier noise, repeated failure, a resumed action or a former state? These questions make presupposition practically useful without requiring learners to memorise an entire theory.

Case 19. A reporting verb can turn a statement into a confession

Source: Hana said that she had moved the folders. Revision: Hana admitted that she had moved the folders. The content about moving folders remains. The reporting verb changes how the speech act is presented. Admitted commonly frames the statement as an acknowledgement, often of something awkward, contested or unfavourable. The source’s neutral said does not by itself establish that framing.

Imagine that the organiser asks who placed the folders on the clean table and Hana replies cheerfully because she is explaining an ordinary helpful action. The source fits naturally. Admitted may add an implication that there was a reason to withhold or resist acknowledging the action. That could be the writer’s intention in another scene, but it is not supplied here.

Do not reverse this lesson into “admit always means guilt” or “claim always means lying”. Those would be inaccurate absolutes. Reporting verbs are sensitive to context and can be used without those simplistic interpretations. The useful question is narrower: what attitude, evidential position or speech-act character has this particular replacement introduced?

A faithful formal revision might be Hana stated that she had moved the folders. Even stated can sound more formal or deliberate than said. If those differences matter to the character’s voice, keep the common verb. Said is not a weak word that every advanced writer must replace. Its relative neutrality is often precisely the feature an evidence-based report needs.

Extend the exercise with confirmed, announced, insisted and explained. A confirmation relates to whether information is established or acknowledged; an announcement foregrounds making information known; insistence presents persistence or force; explanation presents a relationship or reason. The same quoted proposition can be introduced in several ways, but the source must justify the chosen reporting act. A synonym list cannot supply a speech event that never occurred.

Case 20. Attribution is part of the claim, not a disposable introduction

Source: According to the notice, the room was expected to reopen on Monday. Revision: The room reopened on Monday. The revision removes the source, removes expectation and changes the event from anticipated to accomplished. This is three changes in one short sentence. It can look like excellent concision while being materially unfaithful.

Imagine that the notice was printed on Friday, further repairs were needed on Sunday, and the room actually reopened on Wednesday. The original report of the expectation remains accurate. The revision is false. The example shows why a plan or forecast must not be written as a completed fact merely because its date has arrived in the narrative.

A safe revision is The notice gave Monday as the expected reopening date. The source and expectation remain visible. The wording is compact but does not claim that the forecast came true. If the larger passage later confirms Monday’s reopening, the editor can report it, but should recognise that the evidence now comes from an additional statement.

Attribution can be subtle even without a named source. The room is believed to be safe and The room is safe are not equivalent assertions. The first reports a belief held by someone or some group, possibly unspecified. The second asserts safety directly. For school practice, use low-stakes invented situations; do not infer actual safety from wording. The language distinction is about who is committed to the claim, not whether a real room should be entered.

British Council’s advanced passives reference explains impersonal reporting patterns such as is believed to and is reported to. These constructions are useful because they separate the report from a direct assertion. When editing them, preserve both the event’s time and its source status. Removing the reporting frame can be a much larger change than replacing a dramatic adjective.

Case 21. Sequence is not the same as cause

Source: The queue shortened after a second desk opened. Revision: The second desk caused the queue to shorten. The source establishes sequence and, in ordinary context, may invite a causal interpretation. The revision makes causation explicit. Whether that strengthening is justified depends on the rest of the evidence.

Construct an ordinary counterexample: several people leave the queue for an unrelated reason just as the second desk opens, but the new desk has not yet served anyone. The queue becomes shorter after the opening, yet the opening did not produce the reduction in this scene. The revision has selected a mechanism that the original short statement alone does not establish.

This does not mean a reader must distrust every natural causal connection. Context can make the explanation clear. Perhaps the passage states that the second desk began serving waiting customers and records the corresponding flow. Then a causal summary may be well supported. The key is to identify the evidence that licenses the move, rather than assuming that after and because are interchangeable connectors.

A faithful concise version is After a second desk opened, the queue became shorter. It changes word order and phrasing while preserving the limited relation. If the task specifically asks for the cause and the evidence is insufficient, the answer should acknowledge that limitation or seek the missing context. A more confident sentence is not a substitute for a better-supported explanation.

Now test contributed to against caused entirely. A factor can participate in a result without being the only cause. A revision that adds entirely, solely or because of this alone closes other possibilities. Students should watch causal modifiers as carefully as causal verbs. The source may support a contribution while leaving several interacting factors in play.

Case 22. Active and passive can preserve an event while changing its focus

Source: Hana checked the labels before the boxes were moved. Revision A: The labels were checked by Hana before the boxes were moved. Revision B: The labels were checked before the boxes were moved. A retains the agent explicitly. B leaves the checker unspecified. Both can be grammatical; only A necessarily preserves the named participant in the sentence itself.

Imagine the task asks who performed the check. B has removed the answer. If the task asks whether checking preceded movement and the agent is irrelevant or established elsewhere, B may be adequate. This is a useful example of task-dependent omission rather than unsupported invention. The revision is not automatically bad simply because it is less detailed, but it must serve the actual reader’s question.

The counterexample for reverse inference is straightforward. If all you know is B, the checker might have been Maya rather than Hana. B therefore does not recover the full information in the source. The missing agent is not a matter of personal preference between active and passive style. It is information that may be essential to the report.

At the same time, A changes what appears first and what the reader is invited to treat as the topic. The core event may be preserved while the paragraph’s emphasis shifts towards the labels. That can improve coherence if the preceding sentence was about labels. It can weaken coherence if the paragraph is tracing Hana’s actions. British Council’s passive guidance explicitly connects passive constructions with changes of focus and with omitting agents when appropriate.

The advanced lesson is to stop treating passive voice as automatically evasive or automatically sophisticated. Ask what it retains, what it omits and why that information order serves the paragraph. A good editor can defend either active or passive wording when the choice preserves the required meaning and improves the reader’s route through the text.

Case 23. A new verb may require a new participant structure

Source: Maya lent Hana a ruler. Revision: Hana borrowed a ruler from Maya. Under a fixed transaction, this is a useful example of a substantial surface change that can preserve the event. The giver and receiver exchange grammatical positions, but the direction of temporary transfer remains the same. A student who merely counts changed words would underestimate this successful revision.

Now change the revision to Maya borrowed Hana’s ruler. The direction reverses. Maya becomes the borrower rather than the lender. The names and the object remain familiar, but the relationship has changed. A simple diagram, with the ruler moving from Maya to Hana, exposes the error more quickly than a long explanation about sophisticated vocabulary.

The case shows why substitution sometimes requires changing the whole frame. Different verbs make different participants prominent and connect them with different prepositions. Hana received a ruler from Maya preserves a transfer but may lose the temporary-loan relationship. The ruler could have been a gift. Maya gave Hana a ruler has a similar problem unless the surrounding context clearly marks it as lending.

A faithful edit should retain that the ruler is lent rather than permanently transferred. Maya let Hana borrow a ruler may work in a natural narrative, although the emphasis shifts towards permission. Hana borrowed Maya’s ruler is often economical if the ownership and same event are clear. The editor chooses among these formulations according to focus, while protecting the direction and nature of the transaction.

For an advanced extension, compare teach with learn, sell with buy, and show with see. Some pairs describe related perspectives on an event, but one side may add intention, outcome or transaction conditions that the other does not. Do not assume every pair is as reversible as a carefully specified lending transaction. Construct the small world and test the roles each time.

Case 24. The same facts can be used to perform different social acts

Source message: Could you send the draft this afternoon, please? Revision: Send the draft this afternoon. Both direct attention towards the same desired action and time. But the revision changes the social act’s presentation from a polite request towards an imperative instruction. Whether that is suitable depends on the relationship and the intended tone.

A purely factual ledger is not enough here. Neither sentence simply reports an event as true. The editor has to record purpose: a request is being made. The counterexample therefore becomes an audience-response test rather than a truth-value test. Could the new wording be heard as an order in a relationship where the original was intended as a courteous request? Yes. The social change matters even though the desired delivery is unchanged.

A faithful revision might be Please send the draft this afternoon when you have a moment. But the added when you have a moment can soften urgency or introduce flexibility not present in the original. Another is Would you please send the draft this afternoon? That preserves a polite request more closely while changing the form. There is no universal requirement to replace a natural request at all.

This case establishes the boundary of the counterexample method. Truth-preserving substitutions are not automatically voice-preserving, politeness-preserving or purpose-preserving. The method should be paired with a social reading: who is writing, to whom, and for what effect? Earlier guides on register and connotation develop those distinctions.

An advanced writer can deliberately sharpen a request into an instruction when authorised by the brief. The error is doing so accidentally while claiming that the meaning is unchanged. Good vocabulary control includes being able to name the social change: “I have made the message firmer”, “I have reduced distance”, or “I have made the request less urgent.” Deliberate revision is different from invisible drift.

Extended editing case files: when several changes interact

Single-sentence exercises make one distinction visible. Real editing is harder because several distinctions can move together. A changed reporting verb may also remove an attribution. A shorter sentence may turn a future plan into a completed event. A more elegant summary may replace an average with a claim about every member. The following source packets are deliberately small enough to inspect but rich enough to resist mechanical synonym replacement.

For each case, keep three separate records: the facts or claims supplied by the packet, the brief governing the rewrite, and the editor’s final choices. Do not blend them. A source tells you what information is available. A brief tells you which information matters for the task. An editorial decision tells you what wording you selected. Confusing those records is how an attractive sentence begins to look better supported than it is.

Case file 1. The revised workshop notice

The constructed source packet. An organiser’s Monday notice reads: We plan to hold the vocabulary workshop outdoors on Saturday. If heavy rain is expected, we may move the workshop indoors. We will confirm the venue by Friday afternoon. Participants do not have to print the reading passage; a digital copy is acceptable. On Thursday morning, Hana writes in a group message: I think the organiser has reserved an indoor room, but I have not seen the confirmation. No later notice is supplied.

The editing brief. Write a short update for participants using only this packet. Keep the uncertainty about the venue. Preserve the confirmation deadline and the choice between printed and digital materials. Do not turn Hana’s belief into an organiser’s announcement.

Jia Jun’s proposed update reads: Due to heavy rain, the organiser has relocated Saturday’s workshop indoors. The venue will be confirmed on Friday afternoon. Participants must bring the digital passage rather than a printed copy. It is concise, coherent and confidently administrative. It also fails the brief in several different ways.

Start with the weather. The Monday notice states a condition involving a forecast, not a report that heavy rain is occurring or certain. Due to heavy rain adds an actual cause. Then inspect the event. May move becomes has relocated, so a possible future change becomes a completed decision. The difference is not a more formal synonym; it is an invented transition in the workshop’s history.

Now inspect the deadline. By Friday afternoon establishes a latest confirmation time in the notice. On Friday afternoon places the confirmation during that afternoon. Thursday confirmation would satisfy the original deadline but not the revised promise. The new wording is unnecessarily restrictive even though both sentences mention Friday. Finally, optional printing becomes mandatory digital attendance material. A participant carrying an acceptable printed passage would now appear to violate the instruction.

The smallest counterexample world is easy to construct. The organiser keeps the outdoor plan, confirms it on Thursday, and welcomes both printed and digital copies. That world fits the Monday notice. Hana could still have believed an indoor room was reserved without having seen confirmation. Yet all three major directions of Jia Jun’s update would be wrong. One ordinary scenario exposes the accumulation of unsupported changes.

Maya proposes a cautious alternative: The workshop might happen somewhere on Saturday. Please wait for more information and bring the passage. This version avoids the invented relocation, but it discards too much. The planned outdoor venue, the Friday deadline and the permission to use a digital copy are all relevant to the participants’ preparations. Caution does not excuse vagueness. A faithful update must protect both uncertainty and useful information.

Hana’s version is more successful: Saturday’s vocabulary workshop is planned for outdoors. The organiser may move it indoors if heavy rain is expected and will confirm the venue by Friday afternoon. You may use either a printed or digital reading passage. It does not mention her Thursday belief because the brief asks for participant instructions supported by the organiser’s notice. Leaving out a weakly sourced detail can be more responsible than polishing it into authority.

The last sentence deserves its own check. Does a digital copy is acceptable establish that printed copies remain acceptable? In this packet, the explicit statement that participants do not have to print the passage presents printing as an available but unnecessary action. The either/or formulation is a reasonable local paraphrase. Were the source merely Bring a digital copy, the same either/or version would be an unsupported relaxation. The legitimacy comes from this packet, not from a universal rule about digital materials.

For further practice, change the brief: Tell a co-organiser what Hana believes, without endorsing it. A suitable sentence is Hana thinks an indoor room has been reserved, but she has not seen confirmation. That sentence now belongs because source uncertainty is itself relevant to the new task. The vocabulary has not become more impressive. The editorial reasoning has become more discriminating.

Case file 2. A narrative paragraph that acquires an invented confession

The constructed source passage. When Ethan returned, the blue folder was no longer beside the printer. Maya said that she had moved two folders to the shelf, but she could not remember their colours. Hana thought the blue folder might be among them. Jia Jun began checking the shelf while Ethan looked under the table. By the time the tutor entered, they had not found it.

The editing brief. Rewrite this passage more concisely while preserving what the narrator establishes and what each character merely says or thinks. Do not identify a responsible person, invent a motive or resolve the missing-folder question.

The attractive but unsafe version reads: Maya admitted hiding the blue folder on the shelf. Hana realised where it was, and Jia Jun searched the shelf while Ethan checked beneath the table. They failed to recover it before the tutor arrived. Much of the vocabulary sounds appropriate to a mystery. That is precisely the danger: a familiar narrative genre supplies missing events, and the editor mistakes plausibility for permission.

The first sentence contains at least four additions. Said becomes admitted, which presents Maya’s speech as an acknowledgement with a different interpersonal frame. Moved becomes hiding, adding concealment. Two folders becomes the blue folder, identifying an object that Maya could not identify by colour. To the shelf is retained, but its unchanged presence makes the invented content look anchored to the source.

The second sentence supplies knowledge where the source supplies uncertainty. Hana thought the folder might be among the moved folders. The rewrite says she realised where it was. A possible location becomes a known location, and the narrator presents the embedded information differently. The paragraph’s later failure to find the folder does not cancel that earlier overstatement; it merely produces a more puzzling story.

The last sentence is closer, but still worth examining. They had not found it by the tutor’s entrance gives a state at a deadline. They failed to recover it before the tutor arrived also reports non-recovery before that deadline and can be acceptable in a concise summary of the search. Yet recover may imply regaining possession or obtaining something lost, while find directly describes locating it. In this ordinary missing-folder scene the difference may be small. In a technical file-recovery task, it could be substantial. Do not declare a synonym unsafe merely because a dictionary gives it another sense somewhere else.

Ethan builds a counterexample world without adding dramatic detail: Maya moved a green folder and a yellow folder; the blue one had already fallen behind the cabinet. Hana’s guess about the shelf was understandable but wrong. Everyone searched sincerely and had not found the folder when the tutor entered. That world fits every sentence of the source and contradicts the revised identification and concealment story.

A faithful concise version is: The blue folder was missing when Ethan returned. Maya said she had moved two folders to the shelf but could not recall their colours. Hana thought the blue folder might be there, so Jia Jun began checking the shelf while Ethan searched under the table. They had not found it when the tutor entered. The word so deserves caution here. The source sequence makes Hana’s suggestion and Jia Jun’s search adjacent, but does not explicitly establish that her suggestion caused his choice. Under a strict preservation brief, remove it: Hana thought the blue folder might be there. Jia Jun began checking the shelf while Ethan searched under the table.

That self-correction is valuable. A repair can contain its own new inference. Editing is not complete simply because the first obvious errors have been removed. The final pass should inspect newly added connectives, pronouns and compressed descriptions as carefully as the conspicuous replacement verbs.

Now authorise a different task: Write a new fictional version in which Maya deliberately conceals the folder. The earlier sentence may become useful, provided the new narrative establishes the hiding and acknowledgement coherently. It is no longer a faithful paraphrase. It is a creative transformation. The difference between the two tasks is not whether strong vocabulary is allowed; it is whether changed facts are allowed.

The practical lesson for students is to keep genre expectations separate from source evidence. A missing object does not prove theft. An uneasy character does not prove guilt. A speaker who cannot remember a detail is not automatically lying. Advanced vocabulary often names motives and mental states efficiently, but those words should enter only when the narrative gives the writer grounds to use them.

Case file 3. A classroom report with more certainty than data

The constructed source packet. Four learners completed a ten-item exercise before a workshop and a different ten-item exercise afterwards. Their scores were, respectively, Maya 4 then 3, Hana 5 then 5, Jia Jun 6 then 7, and Ethan 7 then 9. A tutor’s note reads: The group’s mean score rose from 5.5 to 6.0. Because the two exercises were different and no comparison group was used, this small example does not establish that the workshop caused the change. It suggests questions for a more careful follow-up. These numbers are invented for this editing task.

The editing brief. Produce an honest short account of this packet. Retain the difference between the group mean and individual changes. Preserve the limits on causal interpretation. Do not market the workshop or invent a research result.

A promotional rewrite reads: The workshop proved effective: every learner improved, producing a significant rise in vocabulary mastery. The results demonstrate that the new method raises performance. The sentence is fluent and contains familiar academic words. Its scientific appearance comes from vocabulary, not from the evidence supplied.

Begin with the arithmetic. The initial total is 22, giving a mean of 5.5. The later total is 24, giving a mean of 6.0. The mean therefore rose by 0.5 points on these exercises. But Maya’s score fell, Hana’s stayed the same, and only Jia Jun’s and Ethan’s scores rose. Every learner improved is directly contradicted by the supplied scores. This is not a subtle question of register or a debatable interpretation of an emotion word.

Next inspect the measured construct. The packet provides scores on two ten-item exercises. It does not measure all aspects of vocabulary mastery. A learner’s command of meaning, pronunciation, collocation, productive retrieval and contextual transfer is not represented in the packet as a single comprehensive measure. The rewrite replaces a narrow observed result with a much broader capability claim. That expansion needs evidence, not a more confident adjective.

Then inspect significant. In ordinary prose it might mean important enough to matter. In a report using words such as results and demonstrate, readers may also hear a technical statistical claim. The packet supplies no significance test. Even a correctly calculated statistical result would not by itself settle practical importance or eliminate design limitations. Here, the safe account states the numerical change directly instead of asking an ambiguous prestige word to do unsupported work.

The causal problem is separate from the mean problem. Correcting every learner to the group on average would repair the aggregate-to-individual error but leave the workshop caused the improvement unsupported. The source itself identifies different exercises and no comparison group. A small counterexample world is that the later exercise was easier for some learners or tested vocabulary they already knew better. The reported score change could occur even without the workshop producing the claimed effect.

A responsible account reads: In this constructed four-learner example, the mean score increased from 5.5 to 6.0 across two different ten-item exercises. Two learners scored higher, one scored the same and one scored lower. The packet does not establish that the workshop caused the change; it proposes a more careful follow-up. This version is less promotional, but more informative. It preserves the units, the individual distribution and the source’s limitation.

A shorter version can also be legitimate: The group’s mean score increased slightly across two different exercises, but individual changes varied and the packet does not establish a workshop effect. This is suitable when exact numbers are not required. It is less suitable in a quantitative comparison where the reader needs the magnitude. Once again, a summary can omit detail deliberately; a paraphrase must not disguise omission as full equivalence.

For an advanced extension, ask which extra evidence would answer each unresolved question. Comparable tasks might address task differences; additional observations might clarify stability; a suitable comparison design might improve causal interpretation. These are proposed evidence needs, not promises that one simple extra step would establish a universal educational claim. The editing lesson is to make the missing evidence visible rather than manufacture it through vocabulary.

Case file 4. One phrasal verb, a permission rule and an unfinished job

The constructed source packet. A coordinator writes: Please back up the shared folder before you delete any duplicate files. You may remove a file only if another complete copy has been checked. You do not have to remove every duplicate today. Let me know by 5 p.m. whether the backup has finished. At 4:30 p.m., Ethan sends: The backup is still running. I have checked two copies but have not deleted anything.

The editing brief. Summarise Ethan’s progress for a colleague. Keep completed actions separate from ongoing actions. Do not turn permission into obligation or assume that a required condition has been met for every file.

A careless summary reads: Ethan backed up the folder and deleted the duplicates after checking them. He must finish deleting all duplicates today and confirm completion at 5 p.m. Almost every familiar word contributes to a different unsupported change. Back up names a process with a completion state, not merely an intention. Still running is not finished. Have not deleted anything is not deleted the duplicates.

The coordinator’s rule also needs careful handling. May remove only if a checked complete copy exists makes the checked copy a necessary condition for permission. It does not say that checking a copy creates a duty to delete the original. Nor does it say that two checked copies establish safety for every duplicate in the folder. The object of the condition matters. A writer who changes singular a file to collective the duplicates may silently widen the authorised operation.

The final instruction asks for a report by a deadline about whether the backup has finished. It does not require the backup to be finished by that deadline, and it does not require the report to occur exactly at 5 p.m. A message at 4:45 saying the backup is still running could satisfy the reporting instruction. The content of a required report is not the same thing as a requirement that the reported event be successful.

Ethan’s minimal counterexample is simply the packet itself. At 4:30 the backup remains in progress, two copies have been checked, and no files have been deleted. Every coordinator instruction may still be respected. The careless summary says several incompatible things have already happened and invents an obligation to delete everything. No imagined exceptional circumstances are needed.

A faithful progress summary is: At 4:30 p.m., Ethan reported that the backup was still running. He had checked two copies and had not deleted any files. The coordinator had asked for a status update by 5 p.m.; completing all deletions today was not required. The report can add the permission rule when relevant: A file may be removed only after another complete copy has been checked, and the shared folder must be backed up before any duplicate files are deleted. Both conditions should remain visible rather than allowing the shorter copy-check rule to erase the initial backup requirement.

That final point makes the case more demanding. Source packets can contain several conditions distributed across sentences. A locally accurate paraphrase of the second sentence can still misrepresent the whole instruction if it omits the first sentence’s sequencing constraint. Advanced editing therefore compares the revision with the full relevant packet, not merely with the sentence containing the target word.

For a final challenge, change the intended recipient to a younger student unfamiliar with back up. A suitable explanation might say, Make and check a safe additional copy of the shared folder before deleting duplicates. This adds explanatory wording while retaining the purpose of backup. It must not pretend that every copied item has already been checked, or that the coordinator’s remaining permissions have disappeared. Audience adaptation is compatible with precision when the editor protects the action sequence and conditions.

Independent practice: decide before reading the commentary

The next tasks mix acceptable changes with unsupported changes, incomplete summaries and genuinely ambiguous cases. That mixture matters. A learner who rejects every replacement has not learned to edit; the learner has learned to be suspicious. The aim is to identify what a particular change does under a particular brief.

For each item, write a decision, a reason and a possible repair. Use one of four descriptions: acceptable for this brief, adds an unsupported claim, removes required information, or needs more context. Where the revision contradicts the source, say so explicitly rather than treating contradiction as merely missing evidence. Several items invite more than one observation.

A strong answer names the changed commitment and gives a small example. A weaker answer says only “wrong vocabulary” or “different tone”. Do not look for the longest explanation. Look for the explanation that allows another person to check the decision. For an acceptable rewrite, identify what stayed stable; do not assume that a large change in wording must be a large change in meaning.

Practice set: twenty-four editorial decisions

1. A lower bound. Source: At least six learners returned the form. Revision: Exactly six learners returned the form. Brief: preserve the number information, including uncertainty about larger numbers.

2. An upper bound. Source: The room can hold at most six participants under this arrangement. Revision: The room can hold fewer than six participants under this arrangement. Brief: preserve the stated capacity limit, not a new safety recommendation.

3. Two boxes. Source: No learner opened both boxes. Revision: No learner opened either box. Brief: preserve what is known about opening the two boxes during the same inspection.

4. A shared or separate passage. Source: Every learner read a passage. Revision: Every learner read the same passage. Brief: preserve the source without adding assumptions about the materials.

5. A local phrasal-verb match. Source: The interview was put off until Tuesday. Revision: The interview was postponed until Tuesday. Brief: preserve the scheduling information in a neutral update, with put off used in its postponement sense.

6. An invented lifetime history. Source: Ethan no longer checks the old index. Revision: Ethan has never checked the old index. Brief: preserve the history implied by the source under its ordinary reading.

7. Work remaining. Source: Hana had three pages left to read. Revision: Hana had finished reading three pages. Brief: preserve the progress information about the same book and moment.

8. Remembering an event. Source: Maya remembered locking the cupboard. Revision: Maya remembered to lock the cupboard. Brief: preserve the kind of remembering described, not merely mention a cupboard and a memory.

9. A question with a new assumption. Source: Jia Jun asked whether the gate was locked. Revision: Jia Jun asked why the gate was locked. Brief: report the question accurately without supplying an answer.

10. An attributed possibility. Source: Hana said that the route might be shorter. Revision: The route was shorter. Brief: preserve both the information source and the degree of certainty.

11. A restrictive word moves. Source: Only Maya carried the red case. Revision: Maya carried only the red case. Brief: preserve the ordinary written focus, with no special contrastive stress supplied.

12. Obligation becomes inference. Source instruction: You must check the labels before packing. Revision in a progress report: You must have checked the labels before packing. Brief: report the instruction, not infer whether somebody followed it.

13. A deadline boundary. Source instruction: Submit the draft before noon. Revision: Submit the draft by noon. Brief: preserve the stated deadline boundary, treating noon as a specified moment.

14. Availability changes direction. Source on Monday: The room is available until Friday. Revision on Monday: The room becomes available on Friday. Brief: preserve when the room can be used, assuming the same intended booking period.

15. A passive that keeps the participant. Source: Hana checked the labels. Revision: The labels were checked by Hana. Brief: retain the event, agent and past time while making the labels the topic.

16. A passive that loses the required participant. Source: Hana checked the labels. Revision: The labels were checked. Brief: answer the question, “Who performed the check?”

17. A loan changes direction. Source: Maya lent Hana a ruler. Revision: Maya borrowed a ruler from Hana. Brief: preserve who lent and who borrowed in the same transaction.

18. A loan changes perspective correctly. Source: Maya lent Hana a ruler. Revision: Hana borrowed a ruler from Maya. Brief: retell the same transaction from Hana’s grammatical perspective.

19. A comparison becomes an absolute judgement. Source: This route is shorter than the riverside route. Revision: This route is short. Brief: retain the information supplied without assuming what counts as a short journey.

20. Cost becomes affordability. Source: The blue notebook costs less than the green notebook. Revision: The blue notebook is affordable for every learner. Brief: preserve the price comparison without introducing anyone’s budget.

21. A noun conceals an outcome. Source: The team discussed repairing the display. Revision: The team’s repair of the display was discussed. Brief: preserve whether a repair is proposed or already established.

22. One noun appears to settle a technical claim. Source: The measurements were close to one another, but the true value was not supplied. Revision: The measurements were accurate. Brief: report only what this packet establishes; do not assume that agreement among measurements establishes closeness to the true value.

23. A pronoun is made too definite. Source: Maya told Hana that she would bring the book. Revision: Maya told Hana that Hana would bring the book. Brief: preserve the intended person, but no preceding context is available.

24. A concise summary with a narrow brief. Source: Hana apologised quietly to Maya beside the classroom door after lunch. Revision: Hana apologised to Maya. Brief: answer only whether Hana apologised to Maya; place, time and manner are not requested.

Answer commentary: what a defensible decision contains

1. Preserve the lower bound. The revision adds an exact total that the source does not establish. Seven, eight or ten returns would still satisfy at least six, but would make exactly six false. A faithful alternative is Six or more learners returned the form. The source is not necessarily saying that the final count is unknown to everyone; it is simply not supplying an exact upper limit. Do not invent uncertainty about the observer’s knowledge when the relevant issue is uncertainty in the information given to the reader.

2. Keep the boundary value. At most six permits six; fewer than six does not. The smallest separating case is a capacity of exactly six participants. The revision makes the capacity limit stricter than the source. A faithful version is The room can hold no more than six participants under this arrangement. A writer may recommend a lower limit for another reason, but must identify it as a recommendation, not a paraphrase. The context phrase under this arrangement also matters: a different furniture layout is outside the claim being preserved.

3. Distinguish a ban on both from a ban on either. The source allows one learner to open the first box and another to open the second, as long as nobody opens both. The revision rules out either opening by any learner. That is a stronger claim. It is not enough to notice that both sentences contain no learner and box. The scope of negation differs. A clear faithful version is During the inspection, no learner opened both boxes. Avoid replacing both with together if that would make the reader think only simultaneous opening was excluded. The same person opening the boxes at different moments would still violate the source claim.

4. Do not invent shared material. The source can describe four learners each reading a different passage. The revision requires one passage shared by all. The two sentences are compatible in a shared-passage situation, but the source alone does not establish that situation. A safe rewrite is Each learner read a passage. This item also warns against a common editing shortcut: replacing an indefinite expression with a definite one often sounds smoother because it treats a reference as settled. Smooth reference is not permission to settle a reference that the source left open.

5. Accept a genuinely local match. Given the specified scheduling sense, was put off until Tuesday and was postponed until Tuesday preserve the relevant event and date. The single-word verb may sound slightly more formal, but it fits the neutral-update brief. A learner should not reject the revision merely because put off has other senses in other contexts. The meaning under examination has already been fixed. Good caution is local and relevant; indiscriminate doubt is not deeper semantic knowledge.

6. Preserve the previous practice. Under the ordinary reading, no longer contrasts the present with an earlier period when Ethan checked the index. Never denies that history. The revision therefore conflicts with the source’s backgrounded information. A suitable paraphrase is Ethan used to check the old index but does not check it now. As with other background assumptions, unusual metalinguistic or corrective contexts can affect the interpretation. This exercise supplies no such context. Analyse the ordinary reading rather than importing exceptional dialogue to rescue the revision.

7. Remaining work does not identify completed work. Hana could have read one page of a four-page text, leaving three. Or she could have read twenty pages of a twenty-three-page text. Both fit the source, but neither requires that she has finished three pages. The revision might happen to be true for a six-page text; it is still unsupported by the supplied sentence. A faithful version is Hana still needed to read three pages. Do not label the two sentences contradictory in every possible situation. The error is replacing information about the remainder with information about completed quantity.

8. Preserve the memory relation. Remembered locking presents recollection of an event. Remembered to lock presents successful remembering of an intended action, typically involving carrying out the locking. The source can describe Maya recalling an earlier locking much later; the revision focuses on not forgetting to perform the action. Those are different events in the mental timeline even when the cupboard ends up locked in both stories. A faithful paraphrase is Maya recalled locking the cupboard. Keep ordinary narrative truth assumptions distinct from special stories about false memories, which require extra context.

9. A different question carries a different starting point. Whether asks if the gate is locked. Why ordinarily treats the locking as given and asks for a reason. A speaker unsure whether the gate is open can sincerely ask the original question but not intend the revised one. The editor has changed the information sought and introduced a background assumption. Retain whether, or write Jia Jun asked if the gate was locked. The answer to the question remains unknown. Reporting a question is not an opportunity to answer it implicitly through word choice.

10. Restore the source and the possibility. The route might turn out longer despite Hana’s statement. The original would still accurately report what she said; the revision would not. A safe paraphrase is According to Hana, the route might be shorter. A summary that needs only the existence of a suggestion could say Hana suggested that the route might be shorter, but suggested can add an interpretation of the speech act. The strongest preservation keeps the simple attribution unless a more specific reporting verb is justified by the full conversation.

11. Keep the restriction in the right place. The source excludes other people from carrying the red case. It does not exclude Maya from carrying a second object. The revision restricts Maya’s carried objects but may allow others to carry the red case too. Imagine Maya carrying the red case and a notebook while nobody else carries the case: the source fits and the revision does not. A faithful version can retain the original or say Maya was the only person who carried the red case. Spoken stress could complicate the comparison; none is supplied here.

12. Do not replace a rule with a conclusion about compliance. The instruction obliges the recipient to check labels. The revision sounds like an inference that checking occurred before packing. A person could receive the obligation and fail to follow it. That situation fits the existence of the instruction but does not support the compliance inference. A progress report should say The instruction required the labels to be checked before packing. It can report compliance only if the packet supplies evidence that the person actually checked them. The same modal form participates in different meaning systems.

13. Inspect the exact boundary. Under the specified deadline reading, submission at noon meets by noon but not before noon. The revision is therefore slightly more permissive at the boundary. In casual conversation the practical difference may be negligible, but the exercise explicitly asks the writer to preserve it. Keep before noon. This is not a lesson that every use of before and by requires legalistic analysis. It is a lesson that the brief determines whether a boundary difference matters, and that a seemingly tiny change can alter permission.

14. Availability before the date is not availability starting on the date. The source on Monday presents an available period extending towards Friday. The revision says availability begins Friday. Tuesday use supplies a simple separating case: it can fit the source while being excluded by the revision. A careful paraphrase is The room remains available until Friday. Depending on the booking task, the writer may need clarification about the exact Friday endpoint, but that uncertainty does not make a reversed start date acceptable. Preserve what is clear and ask only about what is genuinely unresolved.

15. Accept the participant-preserving passive. The checked objects, checker and past event are retained. The labels become the grammatical topic, which is precisely what the brief asks. This is an acceptable local rewrite. It demonstrates that meaning-preserving editing can change syntax substantially. A prohibition on passive voice would be a poor substitute for reasoning about the actual task. The remaining question is whether the sentence fits its paragraph, not whether the word order resembles the source closely enough.

16. Reject an omission that defeats the question. The agentless passive preserves that checking occurred but removes who did it. Since the brief specifically asks for the checker, the revision loses essential information. It is not necessarily false; it is insufficient for this purpose. Hana checked them could be a concise answer if the labels are already established in the question. This is why editing decisions cannot be graded solely by general grammaticality or truth. A true sentence can still fail to answer the required question.

17. Restore the transaction direction. The original moves temporary possession from Maya to Hana. The revision moves it from Hana to Maya. A single arrow from lender to borrower exposes the change. It is a contradiction of the intended participant roles for this same transaction, not merely a stylistic variation. The repair can keep Maya lent Hana a ruler or use the perspective change in the next item. Both names appearing in both sentences does not guarantee that the same event has been preserved.

18. Accept the correctly re-framed transaction. Hana as borrower corresponds to Maya as lender in the supplied event. The revision changes grammatical perspective while preserving the direction and temporary nature of transfer. This is the contrast partner to item 17. A learner who rejects both has noticed that borrow and lend differ but has not yet learned how their participant structures correspond. Require the learner to point to the source, recipient and object, rather than memorise that one verb must never replace the other.

19. A relative comparison does not establish an absolute category. A twelve-kilometre route can be shorter than a twenty-kilometre riverside route without being short for a particular walking task. The source supplies a comparison; the revision supplies a context-dependent absolute judgement. A faithful alternative is The riverside route is longer than this route. That substantial reordering preserves the comparison. The editor must keep the same routes and distance dimension fixed; comparing travel time in one sentence and distance in the other would change the task.

20. Relative cheapness is not universal affordability. A notebook priced below another notebook may still exceed one learner’s budget. The source supplies two prices in relation, not the resources of every learner. The universal every makes the added commitment especially strong. A safe rewrite is The blue notebook is less expensive than the green notebook. Cheaper may also preserve the price comparison in context, though it can carry a quality association in other uses. Do not remove a precise comparison merely to obtain a favourable evaluative adjective.

21. Nominalisation can make a proposal look established. The team can discuss whether to repair the display and decide not to. That fits the source. The team’s repair of the display more readily presents an identifiable repair as the topic, potentially suggesting an event or undertaking already established. The exact interpretation depends on context, so do not say the noun phrase always proves completion. Under the strict brief, avoid introducing that ambiguity: The team discussed a possible repair of the display. The word possible keeps the proposal status visible without inventing the outcome.

22. Agreement among measurements does not establish accuracy. Several readings can cluster tightly around a value that differs from the true value. That situation fits the source and defeats the revision. The source withholds the reference needed to assess accuracy. A faithful account is The measurements were close to one another; the packet does not establish how close they were to the true value. Avoid turning this into a claim that a formal precision estimate has been calculated. The exercise supplies qualitative agreement, not a complete measurement study.

23. Ask for context rather than choosing a person. The source pronoun could refer to Maya or Hana, depending on the preceding discourse and intended report. The revision chooses Hana without sufficient evidence. A defensible answer is needs more context. Once the intended person is known, using that person’s name may improve clarity considerably. The mistake is not repetition of a name or avoidance of a pronoun; it is pretending that an ambiguous source has already resolved the reference. Editorial confidence should stop where the source’s information stops.

24. Accept a deliberately limited summary. The revision retains the apology and its recipient, which are the only facts requested. It omits manner, place and time deliberately. It would not be a complete paraphrase of the source, but it is an acceptable answer to this narrow question. Ask the learner to explain why the same sentence might fail a different brief, such as analysing the embarrassment suggested by quietly. Understanding controlled omission is as important as detecting unsupported additions. Otherwise the counterexample method becomes a rule against all summarising.

Two transfer passages without highlighted danger words

Transfer passage A. The organiser asked the four volunteers to inspect the signs before the visitors arrived. Maya checked the entrance sign, and Hana checked the two signs by the staircase. Ethan said he thought the final sign had already been checked. Jia Jun began looking for the checklist. At opening time, the organiser still did not know whether the final sign had been inspected.

A proposed summary says: All four volunteers completed the sign inspection before opening, and Ethan confirmed that the final sign was safe. Identify at least three unsupported changes, then write a replacement of no more than two sentences. Do not add details about who should be blamed. A useful answer will distinguish checking from establishing safety, a belief from confirmation, and several described actions from completion by every volunteer.

One defensible repair is: Maya checked the entrance sign and Hana checked the two staircase signs, while Jia Jun began looking for the checklist. Ethan thought the final sign had been checked, but the organiser still lacked confirmation at opening time. This summary omits the original instruction because its focus is progress. A version intended to explain compliance should retain that the inspection had been requested before visitor arrival. Name the brief before deciding which version is sufficient.

Transfer passage B. According to a tutor’s note, three learners said the second passage felt easier. The note did not report their scores. The tutor planned to compare the passages and said the difference might be related to topic familiarity. No conclusion about teaching effectiveness was recorded.

A proposed summary says: The tutor’s analysis proved that familiar topics improved all learners’ scores. The invented elements include a completed analysis, proof, a causal effect, a universal learner group and scores that the packet explicitly does not report. A careful summary is: Three learners reportedly found the second passage easier. The tutor planned a comparison and suggested topic familiarity as a possible explanation, but the note reported no scores or conclusion about teaching effectiveness.

These passages test more than whether the learner remembers a forbidden pair. The danger words are embedded in a small information structure. Successful editing requires tracking who knows what, what remains unfinished and what sort of evidence exists. That is the transfer from vocabulary trivia to usable reading and writing judgement.

Teaching the method without turning every sentence into an interrogation

The workshop is long because it is a reference and practice bank, not because one learner should complete it in one sitting. Choose the smallest case that reveals a real difficulty. A student who changes tried to succeeded does not first need a lecture on presupposition projection. A student who understands outcomes but repeatedly removes according to needs source-tracking practice rather than another list of emotion adjectives.

Begin with concrete events. Searching without finding, planning without completing and postponing without cancelling can be enacted or described in ordinary language. Move to quantities and conditions when the learner can hold the relevant alternatives in mind. Introduce reporting, background assumptions and social acts when the task calls for those distinctions. These are suggested instructional choices, not fixed age thresholds or a prescribed sequence for every child.

Keep the independent attempt visible

Show the original and the proposed revision without highlighting the changed phrase at first. Ask the learner whether the revision is faithful and why. Record the explanation before helping. This first attempt distinguishes several states that a polished corrected answer would conceal: the learner did not notice the change; noticed but could not explain it; explained it but could not repair it; or defended the decision independently.

Then provide only the prompt that is needed. A broad prompt is What does the new sentence require to be true? A more specific prompt is Do we know whether the action finished? A stronger prompt supplies a contrast: Could she have searched without finding it? The learner who succeeds after the last prompt has still learned something, but has not supplied the same evidence of independence as the learner who generated the example unaided.

Do not turn the prompt sequence into a guessing ritual. When the learner lacks the meaning of postpone, explain the word directly. When the learner understands the word but confuses it with cancellation, use a timeline. When the obstacle is a pronoun’s reference, supply the missing paragraph. The support should match the obstacle rather than reward persistence through an arbitrary number of hints.

Separate noticing, explanation and repair

A useful feedback record contains three short observations. Noticed the change in outcome. Explained a failed-attempt scenario independently. Repaired the sentence but removed the deadline. This profile is more actionable than “vocabulary weak”. It shows that the learner understands one distinction while another remains unstable.

A simple formative rubric can award credit separately for locating the changed expression, naming the changed commitment, giving a valid counterexample and producing a repair that respects the brief. Such a rubric is a teacher-created observation tool, not a standardised score. Do not convert it into a predicted examination grade, a claim about intelligence or a diagnosis of language development. Its purpose is to decide what to teach next.

Include correct rewrites in the practice. Otherwise students can succeed by answering “not equivalent” every time. Include cases in which a summary is acceptable under one brief and inadequate under another. Include at least one ambiguous source where asking for context is the correct decision. A balanced set tests discrimination rather than obedience to a universal suspicion rule.

Revisit the distinction, not the memorised sentence

After teaching tried versus managed, return later with a new situation: a learner attempted to contact an organiser, tried to lift a box or attempted to solve a puzzle. Do not announce that it is the same lesson. Ask for an editing judgement. The aim is to see whether the learner recognises the outcome distinction outside the original missing-file example.

The later check should also vary the direction. Instead of always strengthening an attempt into success, replace a successful outcome with a mere attempt and ask what has been lost. Instead of always replacing some with all, ask whether all can be shortened to some for a particular summary. A learner who can reason both ways is less dependent on a memorised warning pair.

A correct response on the familiar exercise is useful but limited evidence. A correct explanation on a fresh passage, given without the old cue, is stronger evidence that the distinction is becoming portable. This is an assessment interpretation, not a claim that a particular schedule or number of repetitions guarantees transfer. Adjust the return interval and complexity to the learner’s actual performance.

Four learners, four different repair needs

Maya’s first response in these fictional workshops is often fast and plausible. Her useful next step is not necessarily more rare vocabulary. It is a brief pause at outcome and quantity words. Ask her to name one ordinary world that the source still permits. That makes room for alternatives before the fluent first answer becomes fixed.

Hana often sees several interpretations. Her next step may be a stopping rule: identify the reading supported by the supplied context, explain the relevant difference and stop searching for remote possibilities. Accuracy does not require discussing every imaginable sense of every word. The counterexample should answer the actual editing question, not demonstrate how many complications she can invent.

Jia Jun’s attraction to expressive language can be used productively. Ask him to write two labelled versions: a faithful report and an intentionally dramatic story. Then have him mark exactly which changes create the drama. He keeps creative freedom while learning to distinguish creativity from source-based reporting. The lesson should not teach that enjoying unusual vocabulary is embarrassing; it should teach that expressive choices have consequences.

Ethan tends to value precision. His challenge is to keep that precision proportionate to the task. A summary does not need every source detail, and a friendly message does not need a formal semantic analysis. Ask him which information the brief actually requires and which can be omitted honestly. Skilled editing includes releasing unnecessary detail as well as retaining necessary distinctions.

A parent conversation can remain a conversation

At home, use the method selectively when a real misunderstanding appears. A child says, “Everyone was upset,” after describing two classmates. A parent can ask, “Everyone, or the two people you spoke to?” That is enough. There is no need to convert the rest of dinner into a test of quantifier scope.

Avoid using precision as a way to dismiss feelings. A child may exaggerate while distressed. First understand the experience being communicated; later, in a suitable moment, distinguish frustrating from unfair or some people from everyone. Vocabulary instruction should support communication, not become a technical device for winning an argument against a child.

Likewise, do not treat every ordinary request as a statement awaiting logical cross-examination. The social-act case showed why truth conditions are only one part of meaning. A parent can model accurate reporting, clear questions and honest uncertainty without correcting every word the child tries. The useful destination is a learner who communicates more responsibly, not a learner afraid to speak before checking a ledger.

Advanced limits: when the counterexample test needs another tool

The test is strongest when the relevant reading and the editing brief are clear. It becomes less decisive when a source is ambiguous, figurative, ironic, incomplete or dependent on shared background. Naming those limits makes the method more useful, not less. A technique that pretends to answer every language question will eventually teach students to manufacture certainty.

Fix the sense before testing the replacement

Suppose the source concerns a discussion being put off. Testing it against postponed is reasonable in the scheduling sense. It is not helpful to reject that local match because put off can elsewhere describe discouraging or repelling someone. That would compare different readings rather than test the specified reading. First settle which event the passage describes, then compare the chosen expressions.

The opposite error is to settle an ambiguous sense without evidence. The match was off might describe cancellation in one context or a physical object’s position in another. A fuller passage could resolve it. Until then, an editor should state the uncertainty rather than produce a confident formal replacement that silently selects one interpretation.

Do not use impossible worlds as counterexamples

A useful separating scenario must respect the original’s own commitments. You cannot test Hana completed the task against Hana finished the task by imagining that she never did the task at all. That world makes the original false too. Similarly, a proposed counterexample to a statement about four specified learners must not replace them with a different group halfway through the explanation.

Check the scenario aloud: In my example, is the original still true? Is the revision false? Have I kept the people, time and sense fixed? If the answer fails, revise the counterexample rather than declaring victory. This is one reason a small ordinary scenario is educationally valuable: its consistency is easier to inspect than an elaborate invented story.

Background assumptions are not always simple assertions

Factive verbs, change-of-state expressions and words such as again can present information as already accepted. Their behaviour is sensitive to embedding and discourse. A narrator can distance a claim, quote a mistaken speaker or explicitly challenge an assumption. Advanced learners should therefore say under the ordinary reading here where the analysis depends on a background assumption, rather than treating a single trigger as an infallible truth machine.

The Stanford Encyclopedia of Philosophy discussion of presupposition is useful precisely because it examines projection and context, not merely a list of words that supposedly guarantee facts. The workshop’s ordinary examples are starting points. More complex literary or conversational cases need a wider discourse analysis and may support more than one defensible reading.

Necessary repetition is better than misleading variety

A student may replace a repeated noun with a synonym and accidentally create a new referent. If a report distinguishes draft, copy and original, cycling through those words for variety can destroy the document trail. When the same object must remain identifiable, repeating its established name may be the clearest choice.

The British Council’s guidance on avoiding repetition explains options including reference words, substitution and synonyms. Those devices still require a recoverable reference. It, this and that do not automatically improve a sentence when several possible antecedents compete. In an editing brief concerned with accountability, clear repetition can outperform elegant ambiguity.

Automated agreement is not the same as an inspected reason

A language tool can be asked to compare a source and rewrite, but the output should contain an inspectable explanation. Ask it to identify the changed commitment, construct an ordinary separating case, and propose a minimal repair. Reject an explanation that misquotes the source or invents a condition. Also reject an unsupported assurance that two sentences are equivalent merely because they share most of their words.

There is a historical research example of why that caution matters. McCoy, Pavlick and Linzen’s 2019 HANS study tested whether then-current natural-language-inference systems relied on shortcuts such as word overlap rather than the needed sentence structure. Its findings concern the systems and evaluation conditions studied in that paper, not a benchmark of every present-day language model. The transferable editing question is methodological: what evidence shows that this particular comparison was understood?

A useful tool instruction is: Compare these two versions under the stated brief. Separate factual changes, unsupported additions, omissions and changes in social purpose. For each proposed difference, give a small scenario that keeps the original true while making the revision false. Do not invent missing context. State when the evidence is insufficient. The instruction does not guarantee correctness. It asks for outputs the learner can actually inspect.

The final editorial decision: keep, change, qualify or ask

At the end of a careful pass, there are several legitimate decisions. Keep the original because it is already clear and accurate. Change the wording because the new structure preserves the required information more effectively. Qualify the sentence because the evidence supports less certainty than the draft claims. Ask for context because a reference, source or condition remains unresolved.

Do not force every successful editing session to produce a more impressive sentence. Sometimes the best result is one restored word: may, some, before, according to. Sometimes it is an ordinary verb such as said replacing an unsupported confession verb. Sometimes it is a repeated name replacing an ambiguous pronoun. Those small choices can protect more meaning than a paragraph of ornamental vocabulary.

The four students finish their workshop by comparing two versions of their own progress note. The first says they have mastered advanced vocabulary. The second says they have practised identifying changes in outcomes, scope and attribution, and can now explain several of the cases without prompts. Jia Jun prefers the rhythm of the first. Maya likes its confidence. Hana asks whether every new passage would be easy now. Ethan asks what the exercise actually demonstrated.

They keep the second. It does not make their achievement smaller. It makes their description of the achievement more exact.

That is the mature vocabulary habit this casebook is intended to practise: a writer who can enjoy precise language without allowing the language to outrun the evidence. Better words are not merely words that sound more educated. They are words whose meanings, conditions and consequences the writer is prepared to defend.

Sources and next reading

The references below support the underlying distinctions and writing practices. The student scenes, source packets, practice questions, counterexamples and suggested lesson procedures in this article are original constructed teaching material. They have not been presented as research data or as a validated intervention.

Purdue OWL: Paraphrase—Write It in Your Own Words. A practical reference for understanding a source, reformulating it and checking that essential information and attribution remain intact.

Ion Androutsopoulos and Prodromos Malakasiotis: A Survey of Paraphrasing and Textual Entailment Methods. A research overview used here for the distinction between paraphrase and directional inference, not as evidence for a particular classroom routine.

Lauri Karttunen: Simple and Phrasal Implicatives, 2012. A primary research reference on inference associated with implicative expressions and their behaviour under polarity. The article’s examples simplify selected ordinary readings rather than reproducing the paper’s formal inventory.

Stanford Encyclopedia of Philosophy: Presupposition. Background on information treated as given, projection and the role of context. Useful when a reporting or change-of-state expression does more than assert an obvious event.

British Council: Reported speech—reporting verbs. Grammar patterns that help distinguish appropriate complements and uses of reporting verbs.

British Council: Modals—permission and obligation. Reference support for distinctions including prohibition and lack of obligation.

British Council: Advanced passives review. Reference material for passive structures, reporting constructions and the role of focus and agent expression.

British Council: Avoiding repetition in a text. Options for managing repeated material; this workshop adds the requirement that reference and source commitments remain recoverable.

R. Thomas McCoy, Ellie Pavlick and Tal Linzen: Right for the Wrong Reasons—Diagnosing Syntactic Heuristics in Natural Language Inference, 2019. A historical controlled evaluation illustrating why apparent success can depend on shortcuts. It is not used here to claim a current accuracy rate for language tools.

For the underlying vocabulary map, return to Education: Regarding Vocabulary. For the preceding article, continue with Vocabulary and Phrasal Verbs. For the broader theory of word replacement, read Lexical Substitution Geometry on eduKateSG.

For a learner whose main difficulty is detecting their own errors, use Vocabulary Self-Monitoring. For the social effect of an otherwise accurate rewrite, use Vocabulary and Register and Vocabulary and Connotation. These routes separate the next learning need instead of asking every reader to repeat the whole casebook.