Master Art of Translation | The Information Structure Fidelity System — How Topic, Focus, Given and New Information Preserve Emphasis

How do you preserve emphasis in translation? Accurate translation of topic and focus means preserving which idea the text is about, which information is new, what is being corrected, and which alternatives the speaker excludes. A translation may keep every noun and verb yet shift the answer to a different implicit question. The practical remedy is to identify the question under discussion, mark given versus new information, then rebuild the message using the target language’s own word order, particles, intonation, punctuation or constructions.

People searching for information structure in translation, topic and focus translation, theme and rheme, given and new information, contrastive focus, word-order translation, preserving emphasis, sentence focus, translation accuracy and AI translation quality need a diagnostic method rather than a universal rule about placing a word first or last. Languages package prominence differently; a word-for-word rewrite may alter who sounds surprised, which fact is the answer, what was already agreed, and how a reader links one sentence to the next.

This world-facing guide explains the hidden problem, the linguistic mechanisms, a repeatable professional workflow, common failure modes, original worked examples and transfer exercises. It distinguishes discourse topic from grammatical subject, new information from merely unfamiliar vocabulary, and focus from typographic boldness. It applies to human translation, machine translation, AI-assisted drafts, subtitles, interface copy, instructions, technical writing, journalism, stories and ordinary conversation; the examples use English mini-texts as transparent analytical models, not as claims that every other language uses English structure.

Fast answer: Read the entire exchange. Ask what question the sentence answers. Underline the words already recoverable from context. Circle the answer or contrast. List plausible alternatives. Recreate that pattern naturally in the target language, then back-test with a reader who sees only the target passage. If the reader answers a different question, the translation needs revision.

This chapter belongs to a larger translation system. Return to the Master Art of Translation root for the complete map. The conversation repair chapter covers how people correct misunderstandings in exchanges. This chapter owns the separate problem of how an utterance distributes attention and information. Vocabulary acquisition remains with the Vocabulary Learning Hub; English-internal grammar and discourse remain with How English Works.

Quick map: what information-structure fidelity protects

• Topic: what the current utterance is about, not necessarily its grammatical subject.
• Focus: the part that supplies the answer, contrast or most relevant update.
• Given/new: what the audience can retrieve versus what the utterance contributes.
• Question under discussion: the implicit or explicit question that makes one answer relevant.
• Alternatives: the other people, items, times, reasons or actions that a contrast distinguishes.
• Prominence: stress, particles, constructions, position and discourse organisation that direct the audience’s attention.
• Reader test: whether the target reader knows what has changed without being told the source wording.

Contents

• Search intent and the hidden problem
• The mechanisms of information structure
• The repeatable translation method
• Failure modes and their repairs
• Worked translation diagnostics
• Transfer training and four-week practice
• Quality assessment and glossary
• Frequently asked questions
• Research notes and next routes

Search intent and the hidden problem

A reader asking how to translate emphasis usually has a real problem: the output is grammatical but somehow answers the wrong question. Consider a report saying, “The design team approved the blue version, not the red one.” A literal draft may retain both colour words yet bury the fact that the issue was a colour choice. Another draft might sound as though the design team, rather than the approval, is the surprising news. Neither error is necessarily captured by spelling or terminology checks. What failed was the message’s informational architecture.

An utterance is always situated against a mental background. Someone has asked a question, made a guess, described a problem, introduced a referent, or set up several alternatives. The words that follow update that background unevenly. Some are connective tissue; others carry the news. A translator who sees only an isolated string must often reconstruct the missing conversational question. If the context is unavailable, the correct action is to record uncertainty or choose a minimally committal rendering, not invent focus with an assertive cleft.

Do not conflate ‘topic’ with what a webpage is about, an academic paper’s subject area, a semantic role, or a marketing keyword. The linguistic topic of a clause is its local point of departure in discourse. Nor should ‘focus’ be confused with a camera focal point or the writer’s general purpose. In linguistic analysis, focus concerns the information distinguished from live alternatives, which may be a noun, a time, an entire event, or a correction.

The central paradox is that the safest translation often changes surface order. The same content can be packaged through stress, particles, clefts, dislocation, passive voice, an introductory phrase or a new sentence boundary. Faithfulness demands that the target reader receive the correct update, not that the target text display the same sequence of source words.

There is no single global theory or mapping that makes topic, theme, focus, rheme, givenness and contrast identical. Different linguistic traditions place boundaries differently. The operational definitions here are deliberately explicit: annotate what the utterance is about, what is recoverable, what answers the active question, and which alternatives are contrasted. When a technical analysis depends on a disputed classification, document it rather than presenting one school’s account as a universal fact.

The central diagnostic: one proposition, several meanings

The words ‘Mira sent the draft on Tuesday’ express a basic event. In reply to ‘Who sent the draft?’ the focus is Mira. In reply to ‘What did Mira send?’ the focus is the draft. In reply to ‘When did Mira send it?’ the focus is Tuesday. In reply to ‘What happened?’ the focus may be the entire event. The truth conditions may overlap, but conversational relevance, correction and natural prominence differ. A target that makes Tuesday the strongest news after a question about the sender can sound evasive even when it is literally true.

Try changing the question and leaving the words fixed. Read the sentence aloud each time. Notice how stress moves. Now imagine a target language that relies on a focus particle or a different constituent order. A translator cannot assume that the source script’s visible order already captures the stress. The source’s informational role must be carried by whatever reliable resources the target language actually possesses.

The mechanisms of information structure

1. Aboutness topic versus grammatical subject

Mechanism. The sentence can be about a customer even if a delivery driver is its grammatical subject. In “That customer, the courier called twice,” the opening phrase maintains a discourse referent; it is not the agent of calling. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Preserve an accessible topic through natural dislocation, a topic particle, or coherent pronoun reference rather than mechanically copying the source subject position. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A subject-first rendering can shift attention from the established customer to a newly salient courier. Ask what both sentences before and after are talking about, and check whether the target keeps that thread. When the test fails, return to the source context and record the intended question before trying a new target formulation.

2. Broad focus

Mechanism. A whole event can be new when the audience has asked what happened rather than who performed one step. “The equipment stopped responding” can be an all-new event report after a silent monitoring screen. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Use an unmarked event presentation when the target language supports it; do not insert an exclusive particle or artificial contrast. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The translator singles out ‘equipment’ as if selecting it from other machines, although the entire failure is news. Ask the reader to describe the update without choosing a person, object or time from an invented list. When the test fails, return to the source context and record the intended question before trying a new target formulation.

3. Narrow focus

Mechanism. One component is new against a familiar event frame. If the question is ‘Who replaced the seal?’ then the operator’s identity carries the update. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Choose target emphasis that distinguishes the replacement agent while preserving the settled verb and object as background. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A translation foregrounds the seal and answers what was replaced rather than who acted. Replace the highlighted name with an alternative and see whether that is the contrast the dialogue licenses. When the test fails, return to the source context and record the intended question before trying a new target formulation.

4. Contrastive focus

Mechanism. Focus can reject a live alternative rather than simply provide a new fact. “We requested the smaller case, not the lighter one” contrasts dimensions, not merely objects. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Retain the paired alternatives and the intended correction, even when target syntax reorders the adjectives. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The output implies that the case is both smaller and lighter, eliminating the intended exclusion. Ask which candidate the utterance rules out and whether the translation rules out the same one. When the test fails, return to the source context and record the intended question before trying a new target formulation.

5. Corrective focus

Mechanism. A speaker can repair a prior false assumption with special prominence. After a colleague says the meeting is Thursday, ‘It is on Friday’ corrects the day. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Represent the correction so Friday, rather than the meeting itself, becomes the revised fact. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. An all-new presentation makes the sentence sound like a fresh announcement, obscuring the interpersonal repair. Give the reader the incorrect claim first and test whether the target is naturally heard as its correction. When the test fails, return to the source context and record the intended question before trying a new target formulation.

6. Given but accented information

Mechanism. Repeated material can receive prominence when contrast or emotion makes it relevant. “I said the EAST door” may repeat a previously mentioned door while correcting an orientation error. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Treat givenness and focus as separate annotations rather than equating repeated words with background. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The output de-emphasises ‘east’ merely because it appeared earlier. Check what competing alternative is still live even though both words are familiar. When the test fails, return to the source context and record the intended question before trying a new target formulation.

7. New but backgrounded detail

Mechanism. A fact new to the listener need not be the main answer to the current question. A new model number may appear in a subordinate clause while the main news concerns a safety halt. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Keep the main update prominent; let incidental novel details remain supporting information. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A conspicuous introduction makes the model number appear to be the author’s chief claim. Ask the reader to state the paragraph’s decision in one sentence and note if the technical detail takes over. When the test fails, return to the source context and record the intended question before trying a new target formulation.

8. Question-under-discussion alignment

Mechanism. Utterances are designed to address questions that may be implicit. “Only the second sample passed” often answers which samples passed, not whether testing took place. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Restore the set of eligible answers and the selected member before choosing a target construction. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The translation says that a second sample passed but no longer excludes the others. Write the question the target seems to answer and compare it with the source question. When the test fails, return to the source context and record the intended question before trying a new target formulation.

9. Focus alternatives

Mechanism. A focus interpretation depends on a set of candidates, explicit or recoverable. “Lena reviewed the charts” can contrast Lena with other reviewers, charts with other documents, or reviewing with filing. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Represent the intended alternatives using context, particles, stress or syntax. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The target language chooses an exhaustive construction without evidence that all other candidates were excluded. List the alternatives supported by context; refuse to introduce an additional exclusion. When the test fails, return to the source context and record the intended question before trying a new target formulation.

10. Exhaustive versus non-exhaustive focus

Mechanism. Not every focused item means that it is the only item. Answering ‘Who called?’ with ‘Jules called’ does not necessarily assert that nobody else called. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Reserve exclusive words and exhaustive clefts for evidence that supports that force. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. An AI draft adds ‘only’ for fluency and turns a truthful answer into a false exclusivity claim. Ask whether the source licenses ‘and perhaps someone else’. When the test fails, return to the source context and record the intended question before trying a new target formulation.

11. Focus-sensitive particles

Mechanism. Words such as only, even and also change meaning according to their associates. “Only Hana repaired the valve” differs from “Hana repaired only the valve.” The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Track both particle and scope; the target may need different order or morphology to preserve the association. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The particle attaches to the wrong phrase and changes who or what is excluded. Paraphrase with ‘no one else’ or ‘nothing else’ and check which paraphrase fits. When the test fails, return to the source context and record the intended question before trying a new target formulation.

12. Topic continuation

Mechanism. Repeated reference can keep a paragraph coherent without repeating a full noun phrase each time. A report introduces a sensor, then says it was reset and subsequently monitored. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Use pronouns, ellipsis or repeated nouns according to target-language discourse conventions while preserving unambiguous reference. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The target restarts each sentence as a new subject, flattening progression and raising ambiguity. Trace the same referent through three clauses and check for accidental topic shifts. When the test fails, return to the source context and record the intended question before trying a new target formulation.

13. Topic shift

Mechanism. A writer can deliberately move from one referent to another and signal the turn. “As for the backup unit, its battery is intact” switches from the main unit. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Use the target’s natural topic-shift marker or a clear new paragraph when warranted. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The new unit appears merely as an object of the old discussion, weakening the intended pivot. Ask the reader which object is under discussion now. When the test fails, return to the source context and record the intended question before trying a new target formulation.

14. Contrastive topic

Mechanism. An utterance may compare what holds for separate subtopics without asserting a single full alternative answer. “The paper copies were approved; the digital copies remain under review” divides outcomes by format. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Preserve the two topic domains and their distinct predicates. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A merged sentence treats the digital documents as an afterthought and obscures the split decision. Create a two-row topic-by-result table and compare source and target. When the test fails, return to the source context and record the intended question before trying a new target formulation.

15. Theme and rheme as reading aids

Mechanism. Some traditions speak of a clause’s point of departure and what follows from it. “Regarding the schedule, revisions begin tomorrow” sets a frame then delivers the update. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Use these labels operationally, recognising that theme/rheme is not one universal synonym pair for topic/focus. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The translator labels every initial word ‘old’ and misses a fronted focus. Check both grammatical starting point and actual audience knowledge instead of inferring one from the other. When the test fails, return to the source context and record the intended question before trying a new target formulation.

16. Prosodic nuclear prominence

Mechanism. Speech can mark informative emphasis without visible punctuation. “I ordered GREEN labels” may reject a previous proposal for blue labels. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Consult audio where possible; a transcript alone cannot always resolve stress. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The transcript is translated with flat wording that misses the correction. Listen to stress peaks, pauses and intonation, then compare the repaired script against the recording. When the test fails, return to the source context and record the intended question before trying a new target formulation.

17. Cleft constructions

Mechanism. A language can isolate an answer using a marked construction. “It was Noor who found the defect” presents the finder as prominent against a salient event frame. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Choose a cleft only when its pragmatic effect fits; a normal clause may be better in another language. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. Overusing clefts makes neutral reports sound corrective or contrastive. Compare the target with and without the cleft, asking which sounds like an answer to the actual question. When the test fails, return to the source context and record the intended question before trying a new target formulation.

18. Pseudo-clefts and identifying clauses

Mechanism. A ‘what … is …’ structure can signal a property or action as the answer. “What matters is the second measurement” selects a criterion rather than a person. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Preserve the selection of criterion and the rhetorical set-up, not just the repeated noun. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A flat translation says merely that the second measurement matters, losing the question-like reveal. Ask whether the final clause supplies the resolution promised by the opening frame. When the test fails, return to the source context and record the intended question before trying a new target formulation.

19. Existential and presentational openings

Mechanism. Some sentences introduce new entities into a scene rather than resume established ones. “There is a crack along the upper panel” may be an initial discovery report. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Use a target presentation structure that treats the crack as newly introduced. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A topicalised crack implies the reader already knows which crack is meant. Ask whether the source is introducing an entity or reporting something about a known entity. When the test fails, return to the source context and record the intended question before trying a new target formulation.

20. Passive and active as information packaging

Mechanism. Voice can align familiar material with the clause’s starting point. “The samples were sealed by the technician” can keep the samples topical after several sentences about them. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Preserve participant roles while allowing voice choice to serve target-language cohesion. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A mechanical active rewrite makes the technician the default topic, derailing the paragraph. Follow the referential chain across adjacent sentences, not just a single isolated clause. When the test fails, return to the source context and record the intended question before trying a new target formulation.

Discourse packaging and the target language

21. Dislocation and resumption

Mechanism. Speakers may place a topic outside a clause and resume it with a pronoun. “That old printer, we finally repaired it” signals a familiar object and a new outcome. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Use an idiomatic topic construction where available, or a normal clause if the target lacks a comparable pattern. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. Literal imitation creates ungrammatical duplication or an unnatural colloquial voice. Ask whether the target speaker would actually phrase the topic this way in the stated genre. When the test fails, return to the source context and record the intended question before trying a new target formulation.

22. Ellipsis licensed by context

Mechanism. Given material can disappear while the informative piece remains. “Who signed?” “Ari.” is complete as an answer despite lacking a finite verb. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Do not fill every gap if the target naturally permits a fragment in that communicative setting. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The translated answer expands into a formal sentence and sounds strangely detached. Check whether the answer remains recoverable and appropriate to medium and register. When the test fails, return to the source context and record the intended question before trying a new target formulation.

23. Anaphora and deictic accessibility

Mechanism. Pronouns and demonstratives rely on what the audience can identify. “That one passed” requires a visible or previously named candidate set. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Restore enough identifying context in the target without inventing an object. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The demonstrative refers to the wrong candidate after word order or layout changes. Ask a target reader to point to the referent and record any disagreement. When the test fails, return to the source context and record the intended question before trying a new target formulation.

24. Definiteness and introduction

Mechanism. Articles and determiners can track whether an entity is identifiable. “A package arrived. The package was damaged” moves from introduction to established reference. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Use the target language’s actual resources for accessibility rather than mechanically adding or deleting articles. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A target that marks the initial package as already known creates an unsupported presupposition. Read the passage from the start and ask at which sentence the reader can identify the package. When the test fails, return to the source context and record the intended question before trying a new target formulation.

25. Frame setters

Mechanism. Time, place, condition and domain phrases can establish the scope in which news applies. “For online orders, the return period is shorter” narrows the proposition to a channel. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Keep the domain restriction salient enough that a reader cannot apply it to all orders. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A target buries the frame and makes the limitation look universal. Test the same advice on a different channel and see whether the target warns the reader it does not apply. When the test fails, return to the source context and record the intended question before trying a new target formulation.

26. Focus in questions

Mechanism. Wh-words and question structure specify which slot the answer should fill. “Which file did Mara delete?” requests a file identity, not an account of deletion generally. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Preserve the answer variable, including whether the question presupposes an event. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The question becomes ‘Did Mara delete a file?’, changing the type of answer sought. Ask whether a yes/no response would count as sufficient; if so, the question has drifted. When the test fails, return to the source context and record the intended question before trying a new target formulation.

27. Negative focus and scope

Mechanism. Negation can deny one focused option or the entire proposition. “I did not ask for the RED cover” can deny the colour choice without denying that a cover was requested. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Retain scope and available alternatives, especially where stress is only in audio. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A translation says no cover was requested at all. Ask what affirmative claim still remains possible after the negative statement. When the test fails, return to the source context and record the intended question before trying a new target formulation.

28. Scalar focus

Mechanism. Even, at least, at most and similar expressions invoke ordered expectations. “Even the reserve unit failed” signals a surprising outcome relative to expected reliability. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Reconstruct the scale and the marked point rather than treating the particle as decoration. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The target renders it as ‘also,’ losing the unexpectedness. Ask what the audience was expected to believe before hearing the sentence. When the test fails, return to the source context and record the intended question before trying a new target formulation.

29. Additive focus

Mechanism. Also and too connect a focused member to at least one other. “Ravi also checked the backups” can add an actor or an activity depending on the context. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Identify what is added to what and choose target placement accordingly. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The translated ‘also’ appears to add a different action, not a second reviewer. List the already-true parallel proposition. When the test fails, return to the source context and record the intended question before trying a new target formulation.

30. Parallelism across clauses

Mechanism. Repeated shape can distribute attention among compared alternatives. “One screen displays alerts; the other stores logs” contrasts functions neatly. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Keep comparable phrase lengths and corresponding slot positions where idiomatic. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. One branch becomes dense and the other vague, implying unequal certainty. Read the clauses side by side and compare which variable changes in each. When the test fails, return to the source context and record the intended question before trying a new target formulation.

31. Information structure across paragraphs

Mechanism. Prominence accumulates across several sentences rather than resetting at periods. A paragraph introduces a defect, examines causes and ends with the remedy. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Let sentence topics advance in a controlled chain so that the remedy lands as the culmination. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. Every sentence starts with a fresh category and the reader loses the argument’s trajectory. Summarise the purpose of each sentence and see whether the last one answers the opening problem. When the test fails, return to the source context and record the intended question before trying a new target formulation.

32. Headlines and microcopy

Mechanism. Short text has little contextual room and easily misplaces a contrast. “New password required for shared accounts” can mean a policy about account types, not all users. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Preserve the affected subgroup, and test the headline with its surrounding interface context. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A target headline implies every account must change passwords. Ask an unaffected user whether the translated headline wrongly seems to address them. When the test fails, return to the source context and record the intended question before trying a new target formulation.

Technology, safety and review

33. Subtitles and timing

Mechanism. Audio focus can be flattened by screen-space limits. A spoken correction may need one visible contrast word retained while filler is compressed. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Keep the information-bearing correction and pair it with scene knowledge. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. Compression drops ‘not’ or the alternative and reverses what viewers should believe. Pause after the subtitle and ask what just changed in the plot. When the test fails, return to the source context and record the intended question before trying a new target formulation.

34. Term consistency versus information function

Mechanism. The same technical term can be topic in one sentence and focus in another. A report may introduce ‘baseline drift’ as new, then use it as a stable topic for the next four paragraphs. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Maintain term identity while varying grammatical packaging according to its role. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The term is translated consistently but repeatedly introduced as if new. Trace first mention, later reference and final conclusion separately. When the test fails, return to the source context and record the intended question before trying a new target formulation.

35. Machine-translation fluency illusions

Mechanism. A smooth sentence can conceal a prominence error invisible to lexical overlap. A model may keep all nouns but silently convert a corrective reply into a neutral announcement. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Audit the discourse question and rejected alternatives independently of style. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. Reviewers approve fluency without checking what the source was trying to correct. Compare two target readings after supplying the context, not simply two strings. When the test fails, return to the source context and record the intended question before trying a new target formulation.

36. Multimodal grounding

Mechanism. Pictures, pointing, gesture and layout can determine which candidate is salient. A speaker says ‘this one’ while touching the lower of two switches. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Link target deictics and emphasis to what the audience sees, then check caption positioning. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The words alone identify a different switch in the target layout. Have a reviewer view the final screen or video, not an isolated text export. When the test fails, return to the source context and record the intended question before trying a new target formulation.

37. Ethical restraint against invented emphasis

Mechanism. Strong target syntax can assert exclusion or certainty not supported by the source. A neutral ‘Mina contributed’ must not become ‘It was Mina alone who contributed’ without grounds. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Prefer semantically conservative wording when context is missing; flag ambiguity for review. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The output creates sole-credit or blame claims absent from the evidence. Try appending ‘others may have contributed too’ and observe whether the target contradicts it. When the test fails, return to the source context and record the intended question before trying a new target formulation.

38. Cross-lingual naturalness

Mechanism. Languages can express the same focus through different combinations of syntax and discourse markers. A language may prefer a focus particle where English relies on pitch, or a presentation verb where English uses an existential. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Begin with the target speaker’s natural response to the same situation, then verify semantic constraints. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. A formally aligned target text feels translated and directs stress to the wrong constituent. Use a native-context judgment plus a source comparison; neither test alone is sufficient. When the test fails, return to the source context and record the intended question before trying a new target formulation.

39. Discourse-level reviewer agreement

Mechanism. Two fluent translators may disagree about the focus because the brief is incomplete. A sentence extracted from a table may be about a row label, a category, or a recent change. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Record the question and context used for each decision; escalate irreducible ambiguity. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. Review becomes subjective style negotiation with no shared hypothesis. Compare annotations first; if they disagree, seek the preceding turn or source creator. When the test fails, return to the source context and record the intended question before trying a new target formulation.

40. Controlled emphasis in high-stakes text

Mechanism. Warnings and obligations require exact restriction of what changes and who must act. “Only authorised technicians may reset the alarm” limits the permitted actor. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Protect restrictions before polishing; test boundary cases explicitly. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. An elegant target permits any user to reset the alarm, or implies technicians may perform only that action. Apply counterexamples: unauthorised person, authorised person, other action. When the test fails, return to the source context and record the intended question before trying a new target formulation.

41. Accessibility and emphasis

Mechanism. Screen readers, captions and display modes may not convey bolding or visual layout. A colour-coded table may indicate the changed row without textual marking. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Make essential contrasts explicit in accessible language or markup. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The translation relies entirely on colour, typography or spatial position. Read the plain text or listen to a screen-reader rendition and test whether the contrast remains. When the test fails, return to the source context and record the intended question before trying a new target formulation.

42. When the best fix is a query

Mechanism. Some source ambiguity cannot be repaired responsibly from the segment. A line reading ‘Only approved on Tuesday’ could scope approval to date or describe a lone approval. The translator’s task is to preserve the role of the expression in the surrounding message. The words alone do not prove which reading is intended; the prior exchange and the alternatives available to the reader are evidence.

Reconstruction. Ask for the surrounding sentence or brief rather than selecting a dramatic target structure. This may require a visible grammatical change. A good change preserves what the reader is supposed to notice, while leaving the underlying participants, facts and constraints intact. Do not trade factual fidelity for a display of elegant emphasis.

Failure and test. The final target invents an exclusivity claim and becomes hard to correct after release. Write two plausible readings and identify the minimum missing evidence that would decide between them. When the test fails, return to the source context and record the intended question before trying a new target formulation.

A repeatable translation method: twenty-four passes

Pass 1: Read the whole exchange

Collect the preceding turn and the reply, plus any relevant following reaction. A line-by-line translation is blind to which material has already been introduced. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. What changed in the audience’s mind during this exchange? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 2: Determine the communicative job

Classify the source as answer, correction, contrast, introduction, comparison, warning or continuation. A statement can be factually stable and pragmatically wrong. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. What action or understanding must the target reader reach? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 3: Write the question under discussion

Record the literal or inferred question, without forcing an unsupported interpretation. This is the most compact way to prevent the target answering the wrong question. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Which answer would have been irrelevant here? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 4: Mark topics

Identify the current aboutness referents and mark topic shifts explicitly. The topic may not equal the grammatical subject. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Would a reader know what the paragraph is talking about? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 5: Mark givenness

Underline recoverable people, objects, events and propositions, not just repeated words. A mention can be given through sight, situation or inference. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Could the target audience identify this item before this utterance? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 6: Mark new contribution

Circle the fact supplied or revised by the utterance. Newness is assessed against the discourse, not general world knowledge. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. What exactly was learned now? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 7: Mark alternatives

Write the plausible competing answers that make a focus relevant. Implicit alternatives drive meanings of only, even, instead and not. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Which alternative would make this reply false or corrective? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 8: Check exclusivity

Separate ordinary selection from claims that no other candidate qualifies. An automatic cleft or ‘only’ can add a damaging restriction. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Can another candidate still qualify without contradicting the source? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 9: Trace negation

Identify whether a negation denies the action, actor, object, time, degree or claim. Wrong scope can invert instructions and legal obligations. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Which positive statement is still possible? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 10: Trace additive particles

Find the prior true parallel proposition for also, too and as well. The same particle can attach to different phrases. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. What was already true before the new addition? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 11: Trace scalar meaning

Write down the expectation behind even, at least or at most. The scale often supplies the real rhetorical effect. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. What ranking of likelihood or importance is implied? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 12: Check the audio

If spoken material exists, note peaks, pauses, stressed syllables and corrective intonation. Transcripts often erase the clearest evidence. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Does the voice single out a different word than the typography? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 13: Check gestures and layout

Resolve demonstratives, pointing and screen-position references using the final medium. The focus may be visually supplied rather than lexically stated. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. What did the audience actually see? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 14: Plan target information order

Draft a map of the target’s topic frame, new contribution and contrast. Choose an idiomatic ordering rather than duplicating source positions. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Would a natural target speaker deliver this answer in that order? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 15: Select target resources

Use word order, particles, clefts, voice, pronouns, omission or stress appropriately. No device is universally interchangeable across languages. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Does the chosen device add correction, exclusivity or formality? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 16: Draft without embellishment

Create a plain target version that preserves reference, predicate and scope. Avoid adding emphasis merely because the source sounds expressive. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Did the target introduce a stronger claim? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 17: Run the answer test

Show the target and prior context to a reader; ask what question it answers. A fluent sentence can answer a different question. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Is the same slot filled in source and target? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 18: Run the alternative test

Ask what candidate is excluded, added or treated as surprising. This catches particle association errors. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Which counterexample remains possible? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 19: Run the newness test

Ask what knowledge the passage adds at this point, then compare with source. A salient but incidental item can displace the real update. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Would readers write the same one-sentence takeaway? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 20: Run the cohesive reading

Read three sentences before and three after the target segment. A locally elegant sentence may break the paragraph’s referential chain. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Where does the reader’s attention move between sentences? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 21: Check register and medium

Test the draft in dialogue, subtitle, instruction panel or article format as relevant. Some marked constructions sound natural in one mode and theatrical in another. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Would the intended audience plausibly use this phrasing? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 22: Tag residual ambiguity

Mark unresolved cases with concise notes and source questions. False certainty is worse than an honest query when the brief lacks context. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. What evidence would decide the case? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 23: Review high-risk focus claims

Prioritise exclusions, permissions, safety warnings, negation and actor attribution. Not every emphasis mistake has equal consequences. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Can the target cause the wrong person to act? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Pass 24: Release with a record

Store the chosen interpretation and the reason for it in the project handoff. Future reviewers and revisions need a stable rationale. Record the source evidence and the target-language choice in one or two sentences. A useful work record is not a long theoretical annotation; it should let the next translator identify why a seemingly awkward surface change actually protects meaning.

Checkpoint. Could someone recheck the decision without guessing? If you cannot answer with confidence, do not use a marked construction to conceal the gap. Retrieve context, produce two clearly labelled candidate readings, or ask the source owner for clarification. The objective is an auditable meaning decision rather than cosmetic fluency.

Failure modes and their repairs

Failure 1: A literal first-word rule

Trigger. Every sentence-initial phrase is treated as old topic. Consequence. A fronted focus is mistaken for background and loses its corrective bite. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Separate initial position from discourse function by checking the question. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 2: A literal last-word rule

Trigger. Every final phrase is treated as new focus. Consequence. A final afterthought receives unjustified main emphasis. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Determine focus from context before appealing to end position. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 3: Grammatical-subject substitution

Trigger. The reviewer labels the subject as the topic by default. Consequence. An object topic disappears in an active-voice rewrite. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Trace what prior sentence establishes as the aboutness entity. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 4: Pronoun-repetition bias

Trigger. The translator keeps a full noun wherever the source has one. Consequence. Repeated noun phrases imply unnecessary reintroduction or contrast. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Use target pronouns or ellipsis where identification remains safe. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 5: Invisible topic shift

Trigger. A new frame is inserted without a discourse transition. Consequence. Readers think the old decision still applies to the next object. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Mark the change with an idiomatic frame setter. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 6: Overenthusiastic clefting

Trigger. A neutral answer is rewritten as ‘It is X that …’. Consequence. The target suggests correction or exclusivity without evidence. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Use an ordinary sentence unless the source licenses marked focus. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 7: Exclusive-particle hallucination

Trigger. A drafting tool adds ‘only’ for punch. Consequence. An additional option is excluded incorrectly. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Probe whether the source allows other participants or outcomes. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 8: Missing exclusive restriction

Trigger. Only is removed as repetitive. Consequence. A prohibition, permission or selection expands unintentionally. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. List allowed and disallowed cases and retest. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 9: Misattached only

Trigger. Only moves near another constituent. Consequence. The restricted actor becomes a restricted action or object. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Paraphrase the claim as ‘no one else’ or ‘nothing else’. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 10: Misattached also

Trigger. An additive particle is repositioned casually. Consequence. The text adds a task rather than a participant. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Name the prior parallel proposition and align the added slot. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 11: Even translated as also

Trigger. Surprise is flattened into neutral addition. Consequence. A carefully calibrated expectation disappears. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Ask which outcome would have been least expected. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 12: New item treated as topic

Trigger. A newly introduced entity is topicalised too early. Consequence. The reader is made to recognise an entity not yet introduced. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Use a presentation structure where natural. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 13: Old item treated as discovery

Trigger. An established entity receives an existential reintroduction. Consequence. The paragraph sounds as though the same object appears twice. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Preserve reference continuity across the paragraph. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 14: Correction rendered as fresh statement

Trigger. The translated reply ignores a preceding error. Consequence. The speaker sounds evasive or oblivious. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Preserve the rejected day, number or name in the correction context. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 15: Contrast rendered as conjunction

Trigger. Not X but Y becomes X and Y. Consequence. Mutually exclusive alternatives are combined. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Explicitly test what the source rejects. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 16: Contrast rendered as chronology

Trigger. A correction is interpreted as a later event. Consequence. The target says Y happened after X, not instead of X. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Distinguish replacement from temporal sequence. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 17: Scope of negative moved

Trigger. Negation is placed near another clause. Consequence. The translation denies the entire act rather than a detail. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Write both scopes as full paraphrases and select from context. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 18: Questions changed to confirmations

Trigger. A wh-question becomes a yes/no question. Consequence. The reader supplies assent instead of the requested identity. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Check what answer type completes the question. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 19: Answer fragments overexpanded

Trigger. A one-word reply becomes a formal independent sentence. Consequence. The dialogue loses immediacy and can misframe the new item. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Use a fragment if target dialogue permits it. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 20: Target fragments underexplained

Trigger. A fragment is copied into a language or document style that requires more structure. Consequence. The reader cannot reconstruct the missing proposition. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Restore necessary grammar without adding an unsupported claim. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 21: Audio stress ignored

Trigger. The text-only draft overlooks pitch and pause. Consequence. A spoken correction selects the wrong alternative. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Review the source audio at reduced speed with context. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 22: Typography mistaken for semantics

Trigger. Bold text is interpreted as exclusive focus by default. Consequence. A design cue is over-read as a categorical statement. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Check editorial style and whether bold denotes navigation only. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 23: Screen position changed

Trigger. The same deictic phrase is retained after UI rearrangement. Consequence. A user selects a different option. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Compare translated copy with the final interface layout. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 24: Subtitle emphasis compressed away

Trigger. A short subtitle removes the contrast word. Consequence. The audience misidentifies the decision or culprit. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Compress filler first; protect the information-bearing distinction. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 25: Headline scope broadened

Trigger. A subgroup restriction moves to a remote clause. Consequence. Readers apply a rule to everyone. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Keep the domain frame adjacent to the affected rule. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 26: Headline scope narrowed

Trigger. An optional restriction is rendered as a universal qualifier. Consequence. Readers miss an announcement that applies broadly. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Back-translate the affected population explicitly. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 27: Passive voice treated as mere style

Trigger. The target is forced into active voice automatically. Consequence. A stable topic becomes an unexpected actor focus. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Check thematic continuity before changing voice. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 28: Agent hidden unnecessarily

Trigger. A passive is introduced to sound formal. Consequence. Responsibility that mattered in the source becomes obscure. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Do not sacrifice accountable agent information for surface elegance. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 29: Newness confused with difficulty

Trigger. A rare technical term is always highlighted as the news. Consequence. A complex but established term dominates the sentence. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Measure novelty against discourse, not lexical rarity. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 30: Frequency confused with givenness

Trigger. A common word is assumed to be given. Consequence. The actual new action receives too little prominence. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Use scene knowledge and prior statements, not word frequency. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 31: Contrastive topic collapsed

Trigger. Two groups with different outcomes are merged. Consequence. The split decision becomes an ambiguous global rule. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Keep a parallel topic-result structure. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 32: Parallel clauses deparallelised

Trigger. One clause is heavily expanded and the other shortened. Consequence. Comparative weight becomes misleading. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Match variable slots rather than word counts. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 33: Anaphoric chain broken

Trigger. A pronoun picks up a nearer but wrong antecedent. Consequence. The translation attributes an action to the wrong object. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Annotate every referent over the passage. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 34: Ambiguity falsely resolved

Trigger. The source fragment lacks a focus cue. Consequence. The target chooses one precise exclusive reading. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Request context; otherwise avoid a stronger commitment. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 35: Back-translation overtrusted

Trigger. A reverse translation repeats the same words. Consequence. Reviewers conclude that information structure is identical. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Test question-answer alignment with real target readers. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 36: Fluency score substituted for fidelity

Trigger. A smooth AI output is approved without discourse comparison. Consequence. An error in relevance survives automated checks. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Use separate adequacy, focus and alternative-set checks. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 37: Paragraph order flattened

Trigger. Every sentence is treated as an isolated unit. Consequence. The destination of the argument disappears. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Read the whole paragraph and outline what each sentence contributes. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 38: Termbase used as sentence grammar

Trigger. Consistent dictionary equivalents dictate the entire clause. Consequence. Terminology is right but the utterance sounds unnatural. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Keep term identity stable while rebuilding discourse packaging. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 39: Source defect silently repaired

Trigger. A contradictory passage is rewritten into one clear focus. Consequence. The translator inadvertently becomes an editor or fact author. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Record source errors and seek authorisation before changing meaning. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 40: Accessibility contrast lost

Trigger. Colour or boldness alone carries an essential distinction. Consequence. Screen-reader and plain-text readers receive no correction. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Use accessible textual contrasts where necessary. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 41: Wrong candidate set

Trigger. The target contrasts two objects but the source contrasts two times. Consequence. The same vocabulary produces a different communicative act. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Write the live alternatives and compare their semantic type. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 42: False exhaustive answer

Trigger. A focus marker implies ‘and nobody else’. Consequence. An innocent attribution becomes exclusive credit or blame. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Use a non-exhaustive target reading unless the evidence is stronger. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 43: Excessive stylistic emphasis

Trigger. A translator adds capitals, italics and exclamation marks. Consequence. The target sounds aggressive and raises unjustified certainty. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Preserve rhetorical strength proportionately to the source. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 44: Lost topic after segmentation

Trigger. Translation memory splits two linked sentences into separate units. Consequence. The second unit gets a fresh topic not supported by discourse. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Pass context notes with the segment and review combined output. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 45: Register conflict

Trigger. The target uses spoken topicalisation in a formal instruction. Consequence. The message sounds informal or unserious despite correct facts. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Choose a target construction that fits audience and domain. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 46: Omitted frame-setter

Trigger. A time or condition phrase is deleted as redundant. Consequence. A limited instruction becomes universal. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Test whether the claim remains true outside that frame. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 47: Scope confusion in tables

Trigger. A column heading is detached from a cell during translation. Consequence. A result is associated with the wrong row category. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Inspect rendered tables and reconstruct heading-cell relationships. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 48: Unnecessary topic restatement

Trigger. Each sentence reintroduces a referent with full description. Consequence. Long documents feel repetitive and hide the actual updates. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Use the target’s cohesive devices judiciously. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 49: Premature summary

Trigger. The target opens with the conclusion before the source builds contrast. Consequence. The author’s intended reveal and reasoning order change. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Preserve pedagogical sequencing where communicatively important. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Failure 50: Review without reader questions

Trigger. Proofreading checks grammar but not audience understanding. Consequence. An elegant mismatch remains invisible. This is a discourse-level risk even if grammar and ordinary word choice remain acceptable. The source may still appear recognisable in a side-by-side comparison, which is why general proofreading alone often misses the problem.

Repair procedure. Ask readers what was assumed, what changed and what is excluded. Write down the relevant earlier context, then compare two target versions: one using ordinary syntax and one explicitly marking focus. Reject the version that adds an exclusion, changes an answer variable, or pushes a supporting fact ahead of the real message. For consequential texts, include an example of a reader action that would be wrong under the defective interpretation.

Worked translation diagnostics: original cases

The following examples are invented training material, deliberately written as short English source passages so that the logical change is visible even to readers who do not share a particular language pair. A ‘repair’ is an instruction for an idiomatic target-language solution, not a prescription to copy English word order. Each case supplies an implicit question, the hidden risk, a corrective strategy and a reader-level check. Replace the prototype with real source and target sentences in language-pair practice.

Diagnostic 1: Document approval

Source prototype: “The board accepted the revised appendix, not the original.” Discourse context: The committee has already discussed both appendices. The question under discussion is: Which version was accepted? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the revised version is the accepted alternative; the original is excluded. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: replace the correction with a neutral statement that the appendix was accepted. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep both alternatives and the rejection visible. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the reader know that the original appendix was not approved? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 2: Warehouse scan

Source prototype: “The second scanner failed after the update.” Discourse context: The first scanner has already been discussed as functioning. The question under discussion is: Which scanner failed? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the second scanner is the new member of a set of devices. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: introduce a fresh equipment failure without distinguishing devices. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: make the second-versus-first contrast natural. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could the reader identify the newly affected scanner? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 3: Recipe adjustment

Source prototype: “We reduced the salt, not the sugar.” Discourse context: The cook is responding to a proposal to reduce sugar. The question under discussion is: What ingredient was adjusted? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the recipe correction targets an ingredient identity. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: describe both ingredients as reduced. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: contrast salt with sugar and preserve negation. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Does the target accidentally instruct someone to change the sugar? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 4: Product return

Source prototype: “For refurbished items, the warranty starts on delivery.” Discourse context: The document has discussed both new and refurbished products. The question under discussion is: Which group does this rule apply to? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the frame restricts the warranty rule to refurbished goods. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: write the rule as if it applies to every purchase. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: foreground the affected category in idiomatic target order. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a new-item buyer wrongly believe the rule addresses them? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 5: Research result

Source prototype: “Only sample C showed contamination.” Discourse context: All three samples were tested under comparable conditions. The question under discussion is: Which samples showed contamination? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, exclusivity excludes samples A and B. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: say that sample C was contaminated without excluding the others. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain exclusivity over the sample set. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Does the target permit sample A also to be contaminated? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 6: Service alert

Source prototype: “The backup server is operational; the primary is not.” Discourse context: Users already know there are two systems. The question under discussion is: Which system is available? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, availability splits by system. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: replace a contrast with a vague overall service status. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: pair each system clearly with its condition. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Can the reader decide which service to use? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 7: Customer support

Source prototype: “I requested a refund, not a replacement.” Discourse context: A support agent has offered a replacement. The question under discussion is: What outcome did the customer ask for? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the desired resolution is refund and the rival is rejected. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: phrase the request as acceptance of either option. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve corrective contrast without sounding unnecessarily rude. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the agent send a replacement after reading the target? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 8: Travel confirmation

Source prototype: “The departure time changed, not the platform.” Discourse context: Travellers have been told an update has occurred. The question under discussion is: Which field changed? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the changing field is time, while platform remains stable. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: present both details as modified. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: contrast attributes explicitly. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could a traveller believe they must switch platforms? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 9: Laboratory note

Source prototype: “The technician measured the voltage yesterday.” Discourse context: The lab has already identified the voltage measurement. The question under discussion is: Who measured it? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the technician identity is the answer. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: highlight yesterday as the main update. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: give agent prominence and keep time as context. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Does the target answer ‘who’ rather than ‘when’? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 10: Meeting record

Source prototype: “Nadia sent the notes on Monday.” Discourse context: The team knows Nadia and the notes were sent. The question under discussion is: When were the notes sent? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the day is the new contribution. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: promote Nadia as unexpected sender. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: mark time naturally without adding exclusivity. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would ‘Monday’ be the first useful answer from the target? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 11: Museum label

Source prototype: “The frame dates from the later period; the painting does not.” Discourse context: The label distinguishes an artwork from its mounting. The question under discussion is: Which component belongs to the later period? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the two entities must not inherit the same dating claim. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: merge them under a single historical date. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the contrast between components. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could a visitor infer the painting has the frame’s date? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 12: Repair log

Source prototype: “The motor was replaced; the cable was cleaned.” Discourse context: Two components had different maintenance needs. The question under discussion is: What was done to each component? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the two interventions are paired by topic. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: combine both actions into a generic repair. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain one-to-one component/action mapping. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Can an engineer tell which part was replaced? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 13: Grant announcement

Source prototype: “Applications open tomorrow for first-time applicants.” Discourse context: Two applicant groups are defined elsewhere on the page. The question under discussion is: Who can apply tomorrow? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the eligibility group is a condition, not a decorative aside. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: make opening date seem universal. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the applicant restriction. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would an ineligible reader attempt an application? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 14: Classroom question

Source prototype: “Which chapter did you finish?” Discourse context: The teacher already knows a chapter was completed. The question under discussion is: What detail must the student supply? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, chapter identity is the answer variable. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: turn it into a yes-or-no check of completion. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain a genuine wh-question. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could the student reply merely ‘yes’ and satisfy the translation? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 15: Repair conversation

Source prototype: “The left hinge, not the right hinge, needs oil.” Discourse context: Both hinges were examined. The question under discussion is: Which hinge needs treatment? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, direction distinguishes candidates. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: omit left-right opposition as obvious. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep directional contrast. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the wrong hinge receive oil? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 16: Medical instruction

Source prototype: “Only the prescribing clinician may change the dosage.” Discourse context: The text establishes a permission boundary. The question under discussion is: Who may change the dosage? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, only applies to the actor authorised to decide. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: change the restriction to which dosage can be changed. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain the permitted-actor scope. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could an unauthorised person infer permission from the target? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 17: Safety note

Source prototype: “Do not open the outer panel while the unit is powered.” Discourse context: The guide distinguishes outer and inner panels. The question under discussion is: Which panel must remain closed? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, outer panel is the relevant object under a condition. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: turn the warning into a ban on opening any panel ever. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve both object and temporal condition. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Does the target instruct the wrong action during maintenance? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 18: Status dashboard

Source prototype: “Even the redundant sensor reported an error.” Discourse context: The secondary sensor was expected to protect against faults. The question under discussion is: What unexpected event occurred? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the surprising failure concerns the component expected to be robust. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: translate even as a neutral ‘also’. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: express the expectation mismatch without inventing impossibility. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Does the reader understand why this error is noteworthy? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 19: Budget discussion

Source prototype: “We also need to review staffing.” Discourse context: Equipment costs have already been reviewed. The question under discussion is: What additional category is under review? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, staffing is an added subject in the agenda. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: make ‘also’ suggest staffing is the second reviewer. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: attach the additive relation to the category. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Can the reader identify the prior category being extended? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 20: Software bug

Source prototype: “The issue affects export, not import.” Discourse context: Users have reported data transfer trouble. The question under discussion is: Which function is affected? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the contrast distinguishes two directional operations. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: render a broad data-transfer failure. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: name the impacted operation and exclude the other. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Will a user avoid a functioning feature unnecessarily? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 21: Scientific abstract

Source prototype: “The increase was observed only at the higher concentration.” Discourse context: The experiment included lower and higher concentrations. The question under discussion is: Under which condition did the increase occur? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, only scopes the concentration condition. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: make the increased result exclusive to one researcher instead. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve restricted concentration range. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the target falsely imply a lower-concentration effect? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 22: Shipping label

Source prototype: “The fragile items go in the upper compartment.” Discourse context: Packers already know about two compartments. The question under discussion is: Where should those items go? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, upper location supplies the instruction. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: focus on fragility as if it is newly discovered. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep location unambiguous. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could a packer choose the lower compartment? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 23: Interview transcript

Source prototype: “What I meant was the deadline, not the budget.” Discourse context: The interviewer misunderstood a previous reference. The question under discussion is: What was the interviewee correcting? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, a reference correction selects deadline. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: turn the utterance into an unrelated new claim. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the repair and rejected referent. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the interviewer stop asking about budget? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 24: Story dialogue

Source prototype: “It was the gardener who heard the noise.” Discourse context: Characters already know a noise was heard. The question under discussion is: Who heard it? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the hearer identity is dramatically salient. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: make hearing a new event rather than revealing the witness. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: choose target focus that signals the revelation. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Does the reader hear this as a witness identification? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 25: Equipment introduction

Source prototype: “There is a fault in the upper seal.” Discourse context: The inspection report has not mentioned faults before. The question under discussion is: What did inspection discover? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, a new defect enters the discourse. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: treat ‘the fault’ as a previously known defect. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: use an introduction-like expression. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could the reader identify this fault before this line? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 26: Argument structure

Source prototype: “As for cost, the proposal is acceptable.” Discourse context: A committee compares cost, reliability and training. The question under discussion is: What is the judgement about cost? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, cost sets the domain while acceptance is qualified to it. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: render the whole proposal as universally acceptable. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the issue-specific frame. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the target imply reliability and training have also been approved? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 27: Evaluation form

Source prototype: “The second criterion passed, but the third failed.” Discourse context: Reviewers are assessing criteria one by one. The question under discussion is: Which criteria had which outcomes? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, parallel results map to distinct labels. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: state that the evaluation overall passed. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep a criterion-to-result correspondence. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Can the reader fill in both cells correctly? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 28: Checkout page

Source prototype: “For guest accounts, verification is required at checkout.” Discourse context: A product supports both guest and registered accounts. The question under discussion is: Who must verify at checkout? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, account status limits the requirement. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: drop the introductory account frame. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: make the limitation easy to see in a compact UI. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would registered users incorrectly believe they must verify? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 29: Academic response

Source prototype: “The author challenges the assumption, not the data.” Discourse context: A reviewer has claimed that the data were being denied. The question under discussion is: What is being criticised? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the contrast separates interpretation from evidence. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: translate both assumption and data as challenged. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the correction with appropriate scholarly tone. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the reviewer still accuse the author of rejecting the data? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 30: Audio caption

Source prototype: “I ordered the GREEN labels.” Discourse context: The speaker is correcting an earlier order for blue labels. The question under discussion is: What colour did the speaker actually order? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, prosodic stress identifies colour contrast. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: treat the labels themselves as an unexpected item. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain corrected colour in the subtitle. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a viewer still think blue labels were ordered? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Comparisons, interfaces and reported findings

Diagnostic 31: Public announcement

Source prototype: “The north entrance remains open.” Discourse context: Another entrance was just reported closed. The question under discussion is: Which entrance is available? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, north entrance is selected from an access set. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: turn it into news that entrances generally are open. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve directional identity. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would an attendee choose the correct entrance? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 32: Editorial correction

Source prototype: “The quotation came from the interview, not the press release.” Discourse context: Both sources were cited in a draft. The question under discussion is: Where did the quotation originate? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, source attribution is being corrected. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: report two sources as joint origins. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain one accepted and one rejected source. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the citation be fixed correctly? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 33: Accessibility help

Source prototype: “This button expands the details; that one downloads them.” Discourse context: The screen shows adjacent controls. The question under discussion is: What does each button do? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, demonstratives must map to visual controls. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: copy deictics into a reordered layout. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: check interface positions and descriptive labels. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could a keyboard or screen-reader user predict the actions? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 34: Quality report

Source prototype: “The defect occurred after packaging, not during assembly.” Discourse context: Investigators are deciding between process stages. The question under discussion is: At which stage did the defect occur? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, timeline stage is selected and an earlier stage is denied. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: describe both stages as causes. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep before-after and not-during distinction. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would investigators mistakenly change assembly? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 35: Project planning

Source prototype: “The deadline is fixed; the scope can change.” Discourse context: A team negotiates scheduling and deliverables. The question under discussion is: Which parameter is negotiable? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, two project variables have different flexibility. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: render the whole plan as flexible. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: pair each variable to its permitted status. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could the team delay the deadline based on target wording? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 36: Product description

Source prototype: “The new version adds offline search.” Discourse context: The reader already knows search is supported online. The question under discussion is: What capability was added? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, offline operation is the new feature. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: treat the existence of search as completely new. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: foreground offline availability, not generic search. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could a customer distinguish versions correctly? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 37: Financial statement

Source prototype: “Revenue rose, but operating profit fell.” Discourse context: An analyst compares two indicators. The question under discussion is: Which measure improved and which deteriorated? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the contrast is between metrics, not periods. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: collapse both indicators into one trend. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep distinct measures and directions. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a reader infer profit grew merely because revenue grew? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 38: Instruction sequencing

Source prototype: “First save the draft; only then submit it.” Discourse context: The user is about to submit an unsaved document. The question under discussion is: When is submission permitted? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, a temporal ordering restriction governs the second action. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: translate ‘only’ as limiting which document may be submitted. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve dependency between save and submit. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a user skip saving under the target instruction? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 39: School notice

Source prototype: “The assessment date changed; the venue stayed the same.” Discourse context: Families have already received an earlier notice. The question under discussion is: What part of the notice was updated? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, a change/no-change contrast reduces unnecessary action. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: list date and venue as both changed. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: pair the updated field with the unchanged one. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a reader travel to a different venue unnecessarily? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 40: Technical specification

Source prototype: “The connector carries data, not power.” Discourse context: Hardware design distinguishes two functions. The question under discussion is: What does this connector transmit? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, function contrast excludes a dangerous assumption. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: describe it vaguely as a multipurpose connection. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve data-only restriction. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could an installer use it for power? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 41: Customer escalation

Source prototype: “It was the dispatch team that called the customer.” Discourse context: An account manager has been blamed for contact. The question under discussion is: Which team contacted the customer? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the source corrects attribution. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: turn it into a neutral report of a phone call. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep the agency contrast without adding absolute exclusivity. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the target wrongly attribute the call to the manager? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 42: Cultural event notice

Source prototype: “The morning session is full; places remain in the evening.” Discourse context: Attendees can choose among two time slots. The question under discussion is: Which session still has space? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, availability differs across time frames. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: state that the event is fully booked. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: contrast morning and evening capacities. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could the reader still book the available session? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 43: Data visualisation

Source prototype: “The red series shows forecasts, not observed values.” Discourse context: A chart has red and blue lines. The question under discussion is: Which data type is encoded by red? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the legend correction distinguishes estimate from observation. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: treat both data types as actual measurements. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep colour mapping and epistemic distinction. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a reader confuse forecast with evidence? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 44: Literary introduction

Source prototype: “A small bird appeared in the doorway.” Discourse context: No bird has been mentioned in the scene. The question under discussion is: What entered the scene? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the bird is a newly introduced entity. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: write ‘the bird’ as though readers should recognise it. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: use natural presentational packaging. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the reader search backward for an earlier bird? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 45: Literary continuation

Source prototype: “The bird hesitated, then crossed the floor.” Discourse context: The bird was introduced in the prior sentence. The question under discussion is: What did the already-known bird do next? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the entity remains the topic; movement supplies the new event. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: reintroduce ‘a bird’ as if it were another animal. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve established reference and narrative progression. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Does the target imply one bird or two? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 46: Debate rebuttal

Source prototype: “The policy changes reporting, not eligibility.” Discourse context: An opponent says the policy excludes more applicants. The question under discussion is: What aspect does the policy actually change? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the rebuttal narrows the affected dimension. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: make eligibility sound newly restricted. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain the domain correction without advocacy. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a reader wrongly believe applicants lose access? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 47: Examination feedback

Source prototype: “Your conclusion was clear; the evidence needs work.” Discourse context: The writer has requested two-part feedback. The question under discussion is: Which part needs revision? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, praise and correction have distinct objects. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: present the entire essay as weak. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve strengths and improvement area. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Can the learner act on a precise revision target? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 48: Contract summary

Source prototype: “Only signed amendments take effect.” Discourse context: The parties have discussed both drafts and executed changes. The question under discussion is: Which amendments are operative? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, legal effect is restricted by signing status. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: reduce the sentence to a general statement that amendments take effect. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve signed-only scope. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would an unsigned draft be treated as binding? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 49: Weather message

Source prototype: “The wind warning ends tonight, not tomorrow.” Discourse context: An earlier message gave tomorrow as the endpoint. The question under discussion is: When does the warning end? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, time correction matters for planning. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: present both days as possible endpoints. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep the replacement date prominent. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the reader plan for an extra day of warning? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 50: Digital security

Source prototype: “The code expires after ten minutes.” Discourse context: Users know the code must be temporary. The question under discussion is: When does it expire? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, duration is the operational update. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: focus on the code’s existence as new information. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep timing unambiguous without adding exclusivity. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the reader expect the wrong validity period? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 51: Museum tour

Source prototype: “The curator restored the frame, not the canvas.” Discourse context: Visitors know conservation work occurred. The question under discussion is: Which component was restored? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, component identity distinguishes work done. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: generalise to restoration of the whole painting. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve object contrast. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could visitors infer the canvas was treated? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 52: App setting

Source prototype: “This option turns off reminders, not notifications.” Discourse context: The app distinguishes two message channels. The question under discussion is: Which type of message does the setting affect? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, channel distinction prevents wrong user expectation. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: translate both categories as alerts. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: use distinct approved terms and preserve not-contrast. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Will the user still receive notifications after using it? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 53: Community message

Source prototype: “The volunteers delivered supplies yesterday.” Discourse context: The audience knows supplies arrived but not by whom. The question under discussion is: Who delivered them? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the agent group is the actual update. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: make yesterday the sole emphasis. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: give the group prominence naturally. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could the reader identify who should receive credit? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 54: Judicial report

Source prototype: “The witness identified the vehicle, not the driver.” Discourse context: The record separates object recognition from person identification. The question under discussion is: What exactly was identified? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the scope of testimony is narrow. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: suggest the witness recognised a person. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain the explicit limitation. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a reader overstate the witness’s knowledge? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 55: Energy report

Source prototype: “Consumption decreased in winter, not summer.” Discourse context: The analysis compares seasonal periods. The question under discussion is: When did usage fall? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, season is the discriminating dimension. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: translate decrease as applying equally across seasons. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep only the supported season. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could a reader conclude summer consumption also fell? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 56: Hiring note

Source prototype: “The interview confirmed experience, not certification.” Discourse context: The recruiter is reviewing several eligibility criteria. The question under discussion is: Which qualification was verified? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, two criteria are separated by evidence status. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: state the candidate is fully certified. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve positive and unresolved/negative distinction. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would HR claim a certification that was not verified? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 57: Delivery dialogue

Source prototype: “No, the small parcel arrived.” Discourse context: A colleague asked whether the large parcel arrived. The question under discussion is: Which parcel arrived? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, a denial repairs the guessed size, not the arrival event. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: make ‘no’ deny that anything arrived. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: connect the correction to parcel size. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the colleague stop looking for the large parcel? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 58: Library return

Source prototype: “The blue book belongs on shelf four.” Discourse context: Librarians are sorting books by colour labels. The question under discussion is: Where should this book go? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the shelf number is the useful answer. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: focus on the colour as if it were disputed. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: state location in the target’s natural imperative pattern. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could a reader place it on the right shelf? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 59: Quality investigation

Source prototype: “What failed was the seal, not the sensor.” Discourse context: A technician initially suspected the sensor. The question under discussion is: Which component failed? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the pseudo-cleft sets up an explanatory correction. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: make the sensor appear as an additional failure. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the revealed component and exclusion. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the repair target the seal? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 60: Academic lecture

Source prototype: “First define the variable; then test the relationship.” Discourse context: Learners are building an argument step by step. The question under discussion is: What action precedes hypothesis testing? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, procedural sequence makes definition a prerequisite. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: reorder steps for stylistic fluency. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep the pedagogical dependency visible. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would students test an undefined variable? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 61: Charity reporting

Source prototype: “The campaign funded transport, not accommodation.” Discourse context: Donors are asking how resources were used. The question under discussion is: What expense was funded? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the expenditure category is contrasted. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: translate as general support for all travel needs. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain expenditure specificity. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could auditors identify the supported budget line? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 62: Factory notice

Source prototype: “The evening shift uses the updated checklist.” Discourse context: Earlier instructions described the daytime shift. The question under discussion is: Which team must use the new checklist? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, shift is the domain frame. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: announce a universal checklist change. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: restrict the instruction to the evening shift. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the day shift unnecessarily change procedure? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 63: Scientific peer review

Source prototype: “The effect was detected only after calibration.” Discourse context: A reviewer asks whether the initial measurement showed the effect. The question under discussion is: When did the effect become detectable? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, temporal condition limits the observation. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: state that detection occurred before calibration too. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve sequencing and only-scope. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the reviewer misinterpret raw results? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 64: Creative critique

Source prototype: “The joke depends on timing, not the word choice.” Discourse context: The translator is deciding which aspect of humour to recreate. The question under discussion is: What produces the comedic effect? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, mechanism focus is rhythm, with lexicon deprioritised. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: recreate a pun that changes the joke type. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve delayed reveal and natural rhythm. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the audience react at the same point in the dialogue? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 65: Conference subtitle

Source prototype: “I meant NEXT year.” Discourse context: A speaker is correcting an interpreter who said this year. The question under discussion is: Which year is intended? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, prosodic focus delivers a critical timing correction. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: remove emphasis to save subtitle space. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain next-versus-this contrast visually or verbally. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would attendees schedule the event in the correct year? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Instructions, media and specialist transfer

Diagnostic 66: Scientific figure caption

Source prototype: “The upper panel shows predictions; the lower panel shows measurements.” Discourse context: A figure has two vertically stacked plots. The question under discussion is: Which data type is in which panel? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, parallel visual labels are informative. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: swap panel order during translation. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve spatial reference and mapped data type. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Can a reader correctly identify both plots? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 67: Training manual

Source prototype: “The trainee may observe; only the supervisor may authorise.” Discourse context: Two roles participate in a procedure. The question under discussion is: Who is permitted to approve? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, permissions differ by role and action. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: make observation and authorisation equally available. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain role-specific permission scope. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could the trainee mistakenly authorise a procedure? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 68: Poetry commentary

Source prototype: “It is the silence after the line that changes its meaning.” Discourse context: The analysis has already discussed the line’s vocabulary. The question under discussion is: What feature creates the effect? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the pause is selected from possible meaning-making devices. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: focus only on unusual words. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain pause-as-answer in an idiomatic critical voice. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could the reader test the interpretation by changing the pause? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 69: Case summary

Source prototype: “The appeal concerns procedure, not the original evidence.” Discourse context: A reader expects a dispute about factual proof. The question under discussion is: What is actually challenged? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the legal dimension is restricted to procedure. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: convert the appeal into a factual innocence claim. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve procedural frame without broadening issues. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the reader understand what question remains open? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 70: Brand localisation

Source prototype: “The slogan promises simplicity, not speed.” Discourse context: The campaign brief includes two possible benefit claims. The question under discussion is: Which benefit is the proposition? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the emotional promise differs from a performance claim. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: translate into a speed guarantee. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep intended benefit and refrain from invented measurable claims. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a consumer expect faster service from the target? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 71: Animal care instruction

Source prototype: “Feed the younger animal first.” Discourse context: The caretaker recognises two animals of different ages. The question under discussion is: In what order should they be fed? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, comparative identity structures the procedure. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: focus on feeding generally while omitting age order. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: protect selection and sequence. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the caretaker start with the correct animal? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 72: Archive index

Source prototype: “The 2024 report was revised, not the 2023 report.” Discourse context: A librarian compares successive annual editions. The question under discussion is: Which edition was revised? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, year identifies one item within an ordered set. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: describe both years as revised. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep year-to-version mapping. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would readers request the correct revised edition? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 73: Music rehearsal

Source prototype: “The second verse is softer, not slower.” Discourse context: Performers have suggested reducing the tempo. The question under discussion is: What expressive change is requested? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, dynamic level is selected; tempo change is rejected. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: translate soft as a tempo instruction. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the musical distinction without assuming instrument specifics. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Will performers change volume rather than pace? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 74: Humanitarian logistics

Source prototype: “Water reached the outer shelters; blankets did not.” Discourse context: A field summary lists two supply types. The question under discussion is: Which supplies arrived? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, contrasting outcomes must be paired with items. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: write that all supplies arrived. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain the split delivery result. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Will planners know what still needs delivery? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 75: Garden notice

Source prototype: “The shaded beds need less water.” Discourse context: The guide distinguishes sunny and shaded beds. The question under discussion is: Which beds require reduced watering? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the class of beds is the frame. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: give a universal watering reduction instruction. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the subclass restriction. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would sunny beds be under-watered by the target? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 76: Food allergy menu

Source prototype: “The sauce contains sesame, not peanuts.” Discourse context: Guests have asked about two allergens. The question under discussion is: Which ingredient is present? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, ingredient presence and absence are both relevant. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: treat both allergens as absent or both present. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the yes/no map separately. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a person with a sesame allergy receive a false reassurance? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 77: Media interview

Source prototype: “The researcher questioned the forecast, not the measurement.” Discourse context: The reporter conflated two kinds of scientific statement. The question under discussion is: Which claim was criticised? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, distinguish predicted from observed information. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: suggest the underlying data were rejected. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain this epistemic object contrast. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the journalist accurately report what was disputed? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 78: Disaster drill

Source prototype: “The alert is a test, not an evacuation order.” Discourse context: Participants have just heard an alarm. The question under discussion is: What kind of signal is this? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the contrast prevents unnecessary evacuation. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: omit the ‘not’ to shorten the message. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep the status correction unmistakable. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would readers evacuate solely because of the translation? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 79: Review summary

Source prototype: “The proposal is affordable, but the timeline is unrealistic.” Discourse context: Evaluators assess financial and practical feasibility separately. The question under discussion is: Which dimension needs revision? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, mixed assessment does not license overall approval. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: summarise the proposal as simply good or bad. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep two evaluation dimensions intact. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Can the reader identify what must change? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 80: Open-source note

Source prototype: “The patch fixes display scaling, not data export.” Discourse context: Contributors have discussed two bug reports. The question under discussion is: Which issue was addressed? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the patch scope excludes export. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: write that the release fixes both bugs. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: limit the fix claim to display scaling. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would users stop reporting the export bug prematurely? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 81: Research recruitment

Source prototype: “Only participants who consented received the questionnaire.” Discourse context: Ethics procedures distinguish invited people from consenting people. The question under discussion is: Who received the instrument? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, the recipient set is restricted by consent status. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: imply every invitee received the questionnaire. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve subset and condition. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a reviewer think non-consenting people were surveyed? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 82: Messaging app

Source prototype: “Maya replied to the second message.” Discourse context: The conversation includes several messages. The question under discussion is: Which message received a reply? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, ordinal position selects the object, not the respondent. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: interpret second as meaning a second reply. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep ordinal attachment to message. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the reader identify the correct message thread? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 83: Weather forecast

Source prototype: “Rain is expected later, not during the opening ceremony.” Discourse context: Organisers are deciding whether to start on schedule. The question under discussion is: When is rain expected? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, a temporal contrast protects the current event window. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: announce rain during the ceremony. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: reconstruct the relevant window and excluded period. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a planner cancel unnecessarily? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 84: Troubleshooting

Source prototype: “The problem appears only when the battery is low.” Discourse context: A fault has been reproduced under several power states. The question under discussion is: Under what condition does the problem occur? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, low charge is the restricted trigger. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: translate a generic battery failure independent of charge. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve conditional restriction. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a technician test the fully charged state in vain? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 85: Literary focal effect

Source prototype: “Only then did she notice the empty chair.” Discourse context: A scene has concealed a detail until a specific moment. The question under discussion is: When does awareness change? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, temporal focus carries a delayed realisation. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: open with an ordinary observation that erases timing. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the moment of recognition without forced literal inversion. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Will the reader experience the reveal at the intended point? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 86: Civic guidance

Source prototype: “The appeal window closes Friday, while payments continue Monday.” Discourse context: A notice contains separate service timetables. The question under discussion is: Which process stops and which continues? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, different activities have different deadlines. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: treat both as ending Friday. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep each temporal claim paired to its process. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could a reader miss a still-available payment service? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 87: Bibliography correction

Source prototype: “The article was published in the journal, not the edited volume.” Discourse context: An editor is fixing a source type. The question under discussion is: What is the correct publication venue? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, venue category is corrective. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: list both sources as if both contain the article. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep one correct bibliographic type. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the bibliography select the right reference format? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 88: Audio editing

Source prototype: “The audience laughed after the pause.” Discourse context: The producer is studying how timing affected reception. The question under discussion is: When did the reaction occur? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, sequence, not merely laughter, is the interpretive focus. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: move the laugh before the pause for sentence smoothness. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve temporal ordering. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would an editor insert the pause in the wrong place? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 89: Home energy guide

Source prototype: “Insulate the roof before replacing the heater.” Discourse context: A homeowner is deciding a sequence of efficiency improvements. The question under discussion is: Which step should come first? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, priority is the new instruction. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: state both tasks without an order. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the recommended sequence without claiming universal savings. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the reader schedule the intended first action? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 90: Machine status

Source prototype: “The error was recorded, but the process completed.” Discourse context: A dashboard separates logging from task completion. The question under discussion is: Did the process finish? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, outcome contrasts with an error log. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: translate an error message as process failure. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve log-versus-success distinction. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the operator rerun a completed task? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 91: Stage direction

Source prototype: “She lowered her voice on the word ‘never’.” Discourse context: An actor has been asked where emotional stress belongs. The question under discussion is: Which word receives the vocal change? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, prominence is attached to one token within a line. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: describe the entire line as whispered. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the location of vocal emphasis. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Can a performer reproduce the intended effect? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 92: Agricultural update

Source prototype: “Yield improved on the irrigated plots, not the dry plots.” Discourse context: A study compares two treatment groups. The question under discussion is: Where was the improvement found? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, treatment condition bounds the observed outcome. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: generalise to all plots. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: retain group-specific outcome without inventing causation. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would a reader assume irrigation caused the difference without further evidence? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 93: Museum audio tour

Source prototype: “The inscription, rather than the object, is modern.” Discourse context: Visitors are assessing the age of two components. The question under discussion is: Which component is modern? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, a contrast fixes the scope of an age claim. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: date the object as modern. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep component identity clear. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Will visitors correctly understand which part is old? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 94: Telephone clarification

Source prototype: “I said fourteen, not forty.” Discourse context: A number was misheard over a poor connection. The question under discussion is: What number was intended? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, correction distinguishes acoustically similar numbers. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: leave only the intended number without repair context where confusion may persist. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: repeat both alternatives clearly. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the listener update the number correctly? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 95: Proofreading exchange

Source prototype: “Keep the comma; remove the dash.” Discourse context: Two punctuation marks are under discussion. The question under discussion is: Which mark stays and which goes? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, opposed imperatives map to different tokens. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: make both marks subject to removal. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: preserve the two actions and their objects. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Could the editor implement both changes accurately? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Diagnostic 96: Training assessment

Source prototype: “The answer is incomplete, not incorrect.” Discourse context: An evaluator is distinguishing quality levels. The question under discussion is: What feedback category applies? This question, not the sentence’s bare surface order, determines what information is actually being advanced.

Meaning reconstruction. In this context, a less severe judgement is chosen deliberately. A translator should identify the established frame, the specific variable that changes, and any member of the alternative set that is explicitly rejected. The intended audience must receive the correct answer rather than merely recognise the same vocabulary. Different target languages may express the relationship with a normal clause, reordered constituents, a particle, a marked phrase, or a short supplementary sentence.

Likely failure: translate both labels as simply wrong. This may pass an ordinary grammar check because the individual words remain plausible. Better strategy: keep the distinction between missing detail and false content. Review whether the change is necessary to represent the discourse relationship, and avoid adding exclusivity, surprise or blame that the prototype does not support.

Reader diagnostic: Would the learner know to add evidence rather than reverse the answer? Ask a second reviewer to answer the original question using the target wording alone. Then ask them to identify what information was already understood and which candidate, if any, the sentence excludes. If their answer diverges from the source situation, revise the target packaging before polishing style.

Transfer training and a four-week implementation plan

Training should begin with simple question-answer pairs and end with naturally occurring multilingual documents. The goal is not to memorise one correspondence between English intonation and a single target particle. The goal is to become competent at recognising discourse roles before deciding how to express them. Use authentic, permission-cleared examples and consult skilled users of the target language when a case’s pragmatic interpretation is disputed.

Workshop 1: Shift the question without changing the proposition

Task: Write one event sentence and four preceding questions asking who, what, when and why. Mark the expected focus for each answer. Constraint: Do not add exclusivity merely to distinguish the answers. Success measure: Check whether four target versions would sound naturally different. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 2: Build a live-alternatives list

Task: For a sentence containing a colour, construct three plausible alternatives and ask which the source actually excludes. Constraint: Do not confuse ‘more relevant’ with ‘impossible’. Success measure: Verify the target eliminates exactly the licensed option. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 3: Test only under five scopes

Task: Place only with actors, objects, times, locations and conditions in otherwise similar statements. Constraint: Track the proposition that becomes false when another option qualifies. Success measure: Check each version with a deliberately chosen counterexample. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 4: Restore an omitted question

Task: Take five isolated replies from a conversation and reconstruct plausible prior questions. Constraint: Mark each uncertain inference rather than pretending one context is certain. Success measure: Compare with original transcript context once revealed. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 5: Track a topic over a paragraph

Task: Highlight every expression referring to one object across six sentences. Constraint: Keep one reference chain while allowing idiomatic pronouns and omissions. Success measure: Ask a reader how many objects they think were mentioned. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 6: Detect accidental introductions

Task: Write a short scene that introduces an object and then mentions it again. Constraint: Distinguish first mention from established reference. Success measure: Test whether the target suggests one object or two. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 7: Compare active and passive packaging

Task: Translate a three-sentence account with a stable patient-topic and different actors. Constraint: Preserve agency while maintaining discourse continuity. Success measure: Ask which entity the target paragraph seems chiefly to follow. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 8: Audit a translated headline

Task: Select a public-facing headline with an eligibility condition. Constraint: Write the subgroup affected and the unaffected subgroup. Success measure: Test each subgroup with a hypothetical reader. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 9: Rebuild a correction

Task: Write a claim that is wrong about a date, followed by a correction. Constraint: Emphasise the revised fact without unnecessarily intensifying tone. Success measure: Have a reader state the original error and final answer. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 10: Separate addition from surprise

Task: Construct matched sentences using also and even. Constraint: Define the prior true statement versus the unexpected assumption. Success measure: Ask what the listener was expected to believe. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 11: Work from audio, not transcript

Task: Listen to a recorded sentence with one prominent constituent. Constraint: Annotate stress and pauses before reading a transcript. Success measure: Compare target phrasing with the heard focus. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 12: Retain function in subtitles

Task: Condense a corrective exchange into a limited character space. Constraint: Remove filler rather than the information-bearing distinction. Success measure: Pause playback and ask viewers what changed. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 13: Check table relationships

Task: Translate a table containing two row labels and two outcome columns. Constraint: Ensure headings remain associated after layout reflow. Success measure: Ask a reader to reconstruct the correct two-by-two mapping. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 14: Validate demonstratives visually

Task: Translate a caption that says ‘this’ and ‘that’ alongside two illustrated options. Constraint: Resolve referents in the final visual layout. Success measure: Ask someone unfamiliar with the source which object is indicated. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 15: Identify non-exhaustive answers

Task: Write a reply naming one participant when more may have participated. Constraint: Use a formulation that does not add sole credit. Success measure: Test whether ‘and others too’ remains compatible. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 16: Practise contrastive topics

Task: Write two claims about distinct groups with different outcomes. Constraint: Keep each group connected to its own predicate. Success measure: Create a topic-result table to audit the translation. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 17: Screen for review overconfidence

Task: Compare two fluent AI translations that differ in particle placement. Constraint: Judge discourse adequacy before style. Success measure: Choose the one answering the source question correctly. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 18: Translate with missing context

Task: Supply only the line ‘Not the first one.’ to a reviewer. Constraint: Request the missing set and antecedent instead of inventing them. Success measure: Record exactly what information is needed. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 19: Assess a warning

Task: Translate a rule with an only-authorised-person restriction. Constraint: Map permitted and prohibited actors and actions. Success measure: Run two positive and two negative boundary cases. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 20: Test paragraph progression

Task: Take a short argument from problem to evidence to remedy. Constraint: Record the informational contribution of each sentence. Success measure: See whether the target delivers the remedy as the natural culmination. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 21: Rewrite a technical abstract

Task: Find a statement with a limited experimental condition. Constraint: Prevent background details from displacing the central result. Success measure: Ask whether the target overgeneralises beyond tested conditions. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 22: Protect an editorial rebuttal

Task: Translate a correction distinguishing evidence from interpretation. Constraint: Maintain the narrow scope of disagreement. Success measure: Ask whether the target falsely says the writer rejected the data. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 23: Compare register choices

Task: Create spoken, formal and interface versions of the same focused claim. Constraint: Select natural emphasis rather than copying one construction everywhere. Success measure: Have suitable target readers rank pragmatic fit. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 24: Trace output consistency

Task: Use the same key term as new information once and established topic later. Constraint: Keep the approved term while changing packaging. Success measure: Check both terminological identity and discourse function. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 25: Design a reviewer note

Task: Write a ten-word note explaining why a focus particle is required. Constraint: Anchor the note in question and alternatives, not taste. Success measure: See if another reviewer can reproduce the decision. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 26: Measure first-pass comprehension

Task: Ask five readers what changed after reading a short translated notice. Constraint: Record answers before showing the source. Success measure: Compare misunderstandings to a specific focus failure. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 27: Test the plain-text version

Task: Remove fonts, colours and styling from an important message. Constraint: Ensure meaning survives accessibility modes. Success measure: Ask whether the exclusion remains explicit. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 28: Examine source uncertainty

Task: Collect an ambiguous line with two possible focus readings. Constraint: Document both instead of selecting the most dramatic. Success measure: Ask the source author one decisive clarification question. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 29: Simulate a glossary boundary

Task: Use one glossary term in a correction and in a neutral explanation. Constraint: Avoid treating the term’s fixed spelling as fixed prominence. Success measure: Check whether both passages are natural and correct. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Workshop 30: Run an end-to-end handoff

Task: Give another translator source, target, context notes and known alternatives. Constraint: Ask them to reproduce the reasoning without access to the original drafter. Success measure: Revise the handoff if they infer a different question. For serious practice, preserve the original, draft, comments, revised target and the outcome of a reader check. Assess meaning before fluency: a natural-looking sentence that changes the intended answer is not a successful solution.

After completing the first version, change the communicative context while leaving the source clause visibly similar. Note which focus interpretation changes, which remains stable, and whether the target requires a new particle, ordering or framing phrase. This second round is essential because learning to solve one memorised example is different from learning to recognise the mechanism in a fresh text.

Four-week progression

Week 1 — Diagnosis: Work with short dialogues, highlight topic, given material, the answer and live alternatives. Do six examples each day. Do not translate until each question-under-discussion label can be justified from context. At the end of the week, ask a second reader to classify five unseen examples without hints.

Week 2 — Controlled reconstruction: Translate matched pairs that vary only in their question or scope. Compare target alternatives using native-context feedback rather than a word-for-word back-translation. Record every case where a focus particle, syntactic movement, pronoun choice or sentence boundary would change the interpretation.

Week 3 — Discourse and media: Work with short paragraphs, an audio exchange, a subtitle, a two-row table and interface labels. Protect topic chains and segment-level relationships. Audit the rendered product, because information structure can depend on sound, placement and adjacent text unavailable in a raw translation-memory segment.

Week 4 — Independent review: Choose a new 600–1,000-word source. Annotate ten consequential focus decisions, produce a target, and arrange an answer test with readers who have not seen the source. Revise only after collecting their interpretations. Create a one-page decision log and a short list of recurring errors for the next assignment.

Quality assessment, terminology and editorial checklist

Translation evaluation is multidimensional. The W3C Multidimensional Quality Metrics community taxonomy distinguishes accuracy, fluency, terminology, style and other categories; a focus shift may fall under accuracy, omission, addition, over-translation or style depending on its actual impact. ISO 5060:2024 provides guidance on analytic human evaluation of translation output, including human, post-edited machine and unedited machine translations. These frameworks inform quality design; the specific focus tests in this chapter are a practical editorial procedure, not an invented official ISO or MQM substandard.

The W3C Internationalization Tag Set 2.0 describes metadata such as localization notes, terminology, translate directives and language information. Where translation memories or content platforms lose local context, a short note on the intended question or exclusion may prevent systematic errors. Research on topic, focus and contrast demonstrates why these distinctions deserve careful analysis; scholars do not agree on every theoretical boundary, so this guide uses testable working definitions.

A practical scoring rubric

• 4 — Preserved: the target naturally answers the same question, keeps the same alternatives and excludes no unsupported option.
• 3 — Minor adjustment: the focus remains recoverable, but target-language flow or prominence needs a small improvement.
• 2 — Material ambiguity: the intended update is one possible reading, but a reasonable reader can infer an incorrect one.
• 1 — Wrong focus: the target answers a different question, changes a restriction, or shifts the contrast.
• 0 — Unsafe reversal: negation, exclusivity, permitted actor, critical time or required action is materially wrong.

Score examples with context visible and record severity separately. An informational blog sentence and a safety-critical procedure cannot be treated as equal merely because each contains one error. Sampling should deliberately include high-risk particles, corrections, domain restrictions and ambiguity notes. The rubric is a local teaching tool and does not claim to reproduce an ISO scoring protocol.

Essential glossary

Information structure. How an utterance organises information relative to discourse expectations. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Topic. The entity or domain about which a claim is being made in the current discourse. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Focus. The element or proposition presented as the relevant answer or contrast. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Aboutness. The relation between an utterance and the item it is understood to discuss. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Grammatical subject. A syntactic function, not automatically identical with topic. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Theme. A term used in several traditions for the clause’s departure point; precise definitions vary. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Rheme. What is predicated or advanced beyond the theme in certain traditions. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Given information. Material recoverable from prior talk, shared knowledge or immediate situation. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

New information. The contribution that the utterance makes relative to its discourse context. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Question under discussion. The explicit or implicit question that a contribution addresses. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Alternative set. The contextually plausible competitors to a focused value. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Contrastive focus. Prominence identifying an option against salient alternatives. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Corrective focus. Prominence replacing a previous incorrect or disputed value. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Broad focus. Focus encompassing a whole event or large proposition. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Narrow focus. Focus concentrated on one constituent or small span. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Exhaustivity. A claim or implication that no other candidate qualifies. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Additivity. A relation adding one true item or proposition to another. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Scalarness. The ordering of alternatives by expectation, degree or another scale. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Particle association. The relation between only, even, also and the expression they affect. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Focus projection. The possibility that prominence on a word supports focus on a larger constituent. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Prosody. Rhythm, pitch, stress and timing that help structure spoken messages. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Nuclear accent. A prominent pitch event that can locate information focus in speech. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Cleft. A marked construction that isolates or identifies one part of a message. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Pseudo-cleft. A structure such as ‘what … is …’ that presents an identifying answer. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Dislocation. Placement of an expression outside the normal clause frame, often for topic management. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Resumptive pronoun. A pronoun linking a clause back to a displaced topic. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Topicalisation. A construction making an element available as the point of discourse departure. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Presentational construction. A structure that introduces an entity or event into the discourse. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Anaphora. Reference back to material already accessible in a discourse. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Deixis. Reference anchored to speaking context, time, viewpoint or location in a scene. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Definiteness. A grammatical or pragmatic resource for identifiable reference. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Scope. The part of meaning controlled by operators such as negation or only. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Frame setter. An expression delimiting domain, condition or time for a claim. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Cohesion. Linguistic connections that link sentences and references. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Coherence. The interpretable overall progression of ideas and relations. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Discourse marker. A cue such as however or as for that signals how utterances connect. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Contrastive topic. A topic framed relative to alternative topics; the exact theoretical analysis is debated. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Segment context. Surrounding discourse needed to interpret an extracted translation unit. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Adequacy. A quality concern about how accurately target text reflects source meaning. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Fluency. Naturalness and grammatical acceptability of the target wording. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Over-translation. Adding a meaning or distinction not justified by the source and brief. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Under-translation. Failing to carry a distinction required by the source and task. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Review query. A question requesting missing source context rather than inventing a resolution. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Target-reader test. An assessment of what a reader actually infers from translated material. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Functional equivalence. An adequate match of communicative role, where warranted by the translation brief. For practical use, identify a real source occurrence, its preceding context, and the equivalent target-language role rather than memorising this definition without an example.

Frequently asked questions

What is information structure in translation?

It is the organisation of topic, focus, assumed knowledge and informative updates when meaning moves between languages. The issue is not simply word order: it is how a reader knows what the speaker is answering, contrasting or correcting. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How do I keep emphasis when translating?

Recover the source question under discussion, mark the focus and its alternatives, then choose an idiomatic target-language mechanism. Confirm with a target reader who has the necessary context but not the source wording. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Are topic and subject the same?

No. A grammatical subject is a syntactic role. A discourse topic concerns what the statement is about. They often coincide but can separate, particularly in dislocated or passive constructions. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Are theme and topic identical?

Not in every linguistic tradition. ‘Theme’ may refer to clause-initial point of departure, while ‘topic’ has a pragmatic aboutness interpretation. State the working definition when using these terms technically. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Does focus always appear at the end of a sentence?

No. Position is one possible resource, but intonation, particles, morphology, syntax and context vary across languages. A universal end-position rule misanalyses many utterances. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Is new information always focused?

No. Novel details can be subordinate or incidental, and a repeated expression can still be contrastively focused. Givenness and focus should be annotated separately. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Why can a literal translation be wrong?

Literal word order may redirect attention or reassign particles. The target reader might then answer a different question or hear an unsupported exclusive claim despite seeing correct individual words. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How does ‘only’ change focus?

Its associated constituent determines what is excluded. Restricting the actor differs from restricting the object, time, action or condition. Write a full paraphrase to check scope. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Can I add ‘only’ to sound more decisive?

Not unless the source and brief support exclusivity. A name supplied as the answer to ‘who’ does not automatically mean nobody else was involved. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

What does ‘even’ contribute?

It often marks an alternative that is relatively unexpected on a contextual scale. Translating it merely as ‘also’ can erase the speaker’s surprise or comparison. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Why does ‘also’ sometimes change meaning?

An additive particle can associate with an actor, object, action or condition. Its placement determines what new parallel fact the sentence contributes. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How do I translate a correction?

Identify the earlier mistake, the replacement value and any rejected alternative. Preserve that repair relationship in a natural register without unnecessarily escalating tone. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How should a translator handle clefts?

Use marked identifying constructions when they serve the relevant contrast or question. Avoid inserting a cleft into neutral prose merely to sound emphatic. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Can passive voice help preserve topic?

Yes, in some contexts passive voice keeps a familiar patient or object at the centre of the paragraph. The translator must still retain agency and information needed by the audience. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

What if my source has no prior context?

Avoid fabricated precision. Choose a conservative target rendering, or document plausible readings and ask for the missing question, scene or previous sentence. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How can I test a translated headline?

Identify the targeted reader group and the important changed fact. Ask readers from both affected and unaffected groups what they think the headline requires of them. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How do subtitles preserve focus?

Protect corrections, crucial alternative words and action-changing contrasts first. Use scene context and audio stress; compress fillers before cutting informative distinctions. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Is a word-for-word back-translation enough to validate emphasis?

No. Back-translation may reconstruct the same vocabulary while missing the conversational answer. A context-based reader test is better for focus adequacy. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Can machine translation preserve topic and focus?

Sometimes. Its output still needs review when context is absent or particles, word order and discourse prominence are consequential. Fluency alone does not prove equivalence. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How should AI translation be checked?

Give it the surrounding context and intended question, then independently verify alternatives, exclusion, negation and target naturalness. Flag unresolved ambiguity rather than asking a model to invent certainty. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

What is contrastive focus?

It identifies one value against salient competing values. The contrast may be explicit with ‘not X but Y’ or implied by a correction in conversation. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

What is a contrastive topic?

A way of managing statements about alternative subtopics, as in different outcomes for two groups. The precise theoretical status is debated; practical translators should preserve the comparison structure. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Does bold text establish focus?

Not reliably. Bold can signal navigation, formatting or editorial priority. Check context and audio before treating typography as a semantic operator. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How do I handle focus in tables?

Treat row and column labels as part of the sentence-like context. Verify that each outcome remains attached to the correct item after translation and layout. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How does the reader’s knowledge affect translation?

An entity can be known from an earlier sentence, shared environment or cultural experience. That accessibility influences whether a phrase should introduce or resume a referent. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

What is the difference between lexical knowledge and focus?

Knowing a word is a vocabulary task; knowing why that word is the answer at this point is a discourse task. Both matter, but they require different checks. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How does this relate to How English Works?

English grammar and discourse theory belong to the English owner. This translation chapter concentrates on preserving the informational role when moving across languages. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Is this method limited to English pairs?

No. The method treats English examples as analytical prototypes and permits each target language to use its own natural resources. It makes no universal claim about exact particle or word-order equivalents. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

What are the highest-risk focus mistakes?

Mistakes affecting permissions, negation, actor identity, critical timing, product eligibility, quantities or the exclusion of dangerous alternatives should receive priority human review. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How can teams avoid repeated context loss?

Attach short question-under-discussion notes and source screenshots to translation units, and review linked sentences together when the system permits it. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

When should a translator query an author?

Ask when competing readings remain genuinely plausible and selecting one would materially change blame, exclusion, timing, rights, obligations or instructions. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Can there be more than one focus?

Yes, some utterances support layered or multiple contrasts and complex discourse organisation. Do not force every sentence into a single artificially highlighted word. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How many words do I need to change to preserve emphasis?

There is no universal count. A single target particle may suffice, while another language requires a clause rewrite. Choose the minimal idiomatic change that protects the required interpretation. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

What does an information-structure audit record?

It records context, question, aboutness topic, new answer, alternatives, target mechanism, remaining uncertainty and reviewer outcome. For major risks, add an explicit counterexample. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How can a beginner practise?

Start with one sentence and four different preceding questions. Translate the answers, test them in context and compare how the target naturally handles focus. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

How should success be measured?

Ask target readers what the sentence is about, what changed, and which alternatives it excludes. Compare their answers with source-supported interpretations, not just lexical similarity. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

What should I read next?

Return to the complete translation root for neighbouring mechanisms. Use the Vocabulary Learning Hub for lexical acquisition and How English Works for English-internal language study, without duplicating their roles. In a practical assignment, confirm the interpretation with the preceding discourse and intended audience before choosing any marked target construction.

Research sources, interpretation notes and next routes

The research for this article was checked in November 2026. It uses authoritative public materials for terminology, translation quality and internationalisation rather than pretending one short online page resolves all linguistic debates. The following resources are starting points, not a claim that the authors endorse this particular teaching workflow.

• Neeleman and colleagues, A syntactic typology of topic, focus and contrast (UCL repository) — research on information-structural features and their typology.
• Titov, Do contrastive topics exist? (UCL repository) — a reminder that the status of contrastive topic is debated.
• W3C MQM Community Group — issue categories relevant to accuracy, omission, addition, fluency and style.
• ISO 5060:2024 — translation output evaluation guidance, including human and machine output.
• W3C ITS 2.0 — translation and localisation metadata relevant to providing context to downstream tools.
• Song et al., comparative judgement in human evaluation of machine translation (2025) — research on reviewer comparison methods; not a direct focus-specific standard.
• Magdy et al., linguistically motivated multidimensional quality metrics (2026) — an example of research distinguishing pragmatic and other language-specific errors.

A reliable research interpretation keeps three things distinct: a scholarly concept, an established evaluation standard and this article’s suggested classroom or editorial procedure. The theory informs what may go wrong; the quality frameworks make accuracy visible; the practical examples turn the problem into a test that a translator can actually perform.

Return to Master Art of Translation | The Complete System for Moving Meaning Between Languages to find the other specialist chapters. Continue backward to The Conversation Repair Fidelity System when the main problem concerns clarifications, replies and turn-taking. Use the Vocabulary Learning Hub when readers need stronger word recognition, recall and lexical precision. Use How English Works V1.1 when the need is to understand English internally rather than reconstruct meaning across languages.

Final principle: A translation does not merely carry a statement. It carries the way a statement changes a reader’s understanding at that moment. Preserve who or what the message is about, what the reader already knows, the exact new answer and the alternatives that remain possible. Then the language can change without quietly changing the conversation.